Thesis ADikfidan2013

Thesis ADikfidan2013
Dissertation submitted to the Combined Faculties for the Natural Sciences and Mathematics of the Ruperto‐Carola University of Heidelberg, Germany for the degree of Doctor of Natural Sciences presented by Diplom‐biologist Aytac Dikfidan born in Bremen, Germany Oral examination:
____________________
Structural and functional characterization of Clp1, a eukaryotic RNA‐specific polynucleotide kinase Aytac Dikfidan 2013 Thesis Reviewers: Prof. Dr. Ilme Schlichting Dr. Anton Meinhart Table of contents
4
Table of contents
Tableofcontents
Table of contents ........................................................................................................................ v
Summary..................................................................................................................................viii
Zusammenfassung ...................................................................................................................... x
Publications .............................................................................................................................. xii
1
Introduction ........................................................................................................................ 1
1.1
Importance of Polynucleotide kinases in the nucleotide metabolism......................... 1
1.1.1
PNKs in DNA repair pathways .......................................................................... 2
1.1.2
PNKs in RNA repair pathways........................................................................... 4
1.1.3
PNKs in RNA maturation and RNA turnover .................................................... 5
1.2
2
Clp1, a novel eukaryotic RPNK ................................................................................. 7
1.2.1
hsClp1’s involvement in RNA metabolism ........................................................ 8
1.2.2
Structural characterization of scClp1 ............................................................... 12
1.2.3
The RNA-specific Clp1 protein family ............................................................ 13
1.3
The phosphoryl transfer reaction mechanism........................................................... 14
1.4
Scope of this thesis ................................................................................................... 16
Materials and Methods ..................................................................................................... 19
2.1
Materials ................................................................................................................... 19
2.1.1
Chemicals ......................................................................................................... 19
2.1.2
Crystallization screens ...................................................................................... 19
2.1.3
Buffers .............................................................................................................. 20
2.1.4
Growth media ................................................................................................... 23
2.1.5
Bacterial strains ................................................................................................ 24
2.1.6
Plasmids ............................................................................................................ 24
2.2
3
Methods .................................................................................................................... 25
2.2.1
Molecular Biology ............................................................................................ 25
2.2.2
Biochemistry..................................................................................................... 33
2.2.3
Biophysical Methods ........................................................................................ 40
2.2.4
Crystallographic Methods ................................................................................ 40
2.2.5
Bioinformatic and -computational methods ..................................................... 42
Results .............................................................................................................................. 43
v
Table of contents
3.1
Structural characterization of ceClp1 ....................................................................... 43
3.1.1
Bioinformatical characterization of the Clp1 protein family ........................... 43
3.1.2
Purification of ceClp1 ...................................................................................... 46
3.1.3
Co-crystallization of ceClp1 ............................................................................ 47
3.1.4
De novo phasing of ceClp1 by single-wavelength anomalous diffraction
experiments ...................................................................................................... 51
3.1.5
Phasing and refinement of the different ceClp1 complexes ............................ 52
3.1.6
Crystal structure of apo-ceClp1 ....................................................................... 52
3.1.7
RNA-binding site ............................................................................................. 56
3.1.8
ATP binding site .............................................................................................. 61
3.1.9
Phosphoryl transfer reaction mechanism of ceClp1 ........................................ 63
3.1.10
Summary of the structural results .................................................................... 67
3.2
Biochemical characterization of Clp1 from C. elegans ........................................... 69
3.2.1
Substrate specificity of the Clp1 protein family .............................................. 69
3.2.2
Minimal substrate requirements of ceClp1 ...................................................... 70
3.2.3
Characterization of ceClp1’s 5’-kinase activity ............................................... 74
3.2.4
Reverse reaction of ceClp1 .............................................................................. 76
3.2.5
Summary of the enzymology studies ............................................................... 78
3.3
4
Mutagenesis studies on ceClp1 ................................................................................ 78
3.3.1
Site directed mutagenesis of Lys127 and Trp233 ............................................ 78
3.3.2
Deletion of the N-terminal and C-terminal domain ......................................... 80
3.3.3
Summary .......................................................................................................... 83
Discussion ........................................................................................................................ 85
4.1
The Clp1 protein family ........................................................................................... 85
4.1.1
Sequence conservation of the Clp1 protein family .......................................... 85
4.1.2
Putative regulatory function of the additional NtD and CtD ........................... 86
4.1.3
Structural difference between scClp1 and ceClp1 ........................................... 89
4.2
RNA-recognition and RNA-specificity in PNKs ..................................................... 90
4.3
Phosphoryl transfer reaction .................................................................................... 92
4.3.1
A molecular model of the phosphoryl transfer reaction in PNKs .................... 93
4.3.2
Putative function of the non-canonical Walker A lysine Lys127 .................... 96
4.4
Conclusions and Outlook ......................................................................................... 98
4.4.1
Clp1 a novel eukaryotic, RPNK....................................................................... 98
4.4.2
Functional diversity within the Clp1 protein family ........................................ 99
vi
Table of contents
4.4.3
5
The central dogma of the Walker A lysine has to be reconsidered .................. 99
Acknowledgement .......................................................................................................... 101
References .............................................................................................................................. 102
6
Appendix ........................................................................................................................ 113
6.1
Appendix of the Materials and Methods section .................................................... 113
6.1.1
List of primers ................................................................................................ 113
6.1.2
Oligonucleotide sequences. ............................................................................ 115
6.2
Additional Tables ................................................................................................... 116
6.2.1
6.3
Proteins showing a conventional Walker A lysine residue. ........................... 116
Abbreviations ......................................................................................................... 120
vii
Summary
Polynucleotide kinases (PNKs) are crucial enzymes involved in DNA and RNA repair, RNA
maturation, as well as in RNA degradation processes. These enzymes have a conserved PNK
domain, showing structural homology to the classical fold of P-loop kinases. PNKs catalyse
the transfer of the γ-phosphate group of ATP molecule to the 5’-hydroxyl group of
polynucleotide substrates. Depending on their in vivo function, PNKs show different substrate
specificity. Interestingly, eukaryotic PNKs have recently been identified that specifically
phosphorylate RNA substrates. These novel RNA-specific PNKs constitute the Clp1
subfamily of PNKs named after their first identified member. Human Clp1 was shown to
participate in various RNA maturation pathways: (i) cleavage and polyadenylation of RNA
polymerase II pre-mRNA transcripts, (ii) tRNA-splicing, and (iii) phosphorylation of
synthetic siRNAs. Despite extensive studies on Clp1, the structural elements involved in
RNA-specificity and the mechanism of the phosphoryl transfer reaction have remained
elusive so far.
This thesis, therefore, aims for the structural and functional characterization of the
RNA-specific Clp1 PNK. During this work, the three-dimensional structure of Clp1 from
Caenorhabditis elegans was described for the first time at atomic resolution. Clp1 is a
multi-domain protein that consists of a central PNK domain sandwiched by additional N- and
C-terminal domains. Clp1 was crystallized with various RNA substrates that differ in length
and sequence. Based on these results, the structural features of Clp1’s RNA-specificity were
elucidated. Clp1 uses an “RNA-sensor” that recognizes the 2’-hydroxyl group of the RNA at
the ultimate position. Additionally, Clp1 was also crystallized in enzymatically relevant states
such as an inhibited substrate bound state, a transition state analog and a product bound state.
A general model for enzyme catalysis of PNKs was derived from these structures. In contrast
to other described PNKs, Clp1’s ATP-binding site within the PNK domain is obstructed by
the N-terminal domain. The crystal structures as well as activity assays with truncated
variants of Clp1 showed a contribution of the N-terminal domain to ATP-binding by
interactions with the nucleobase. The phosphate groups of the ATP molecule are anchored in
the active site tunnel, which is formed by the common P-loop motif, a divalent metal cofactor
(Mg2+), and an α-helical LID module. Structure-guided mutational analysis identified the
essential role of the Walker A lysine for enzyme catalysis. Moreover, Clp1 crystal structures
revealed a non-canonical Walker A lysine in an “arrested” conformation that acts as a
molecular switch. Activation of the switch is only achieved in the transition state complex. In
viii
Summary
contrast to other nucleotide kinases, Clp1 seems to apply a substrate-gating mechanism that
prevents futile ATP hydrolysis. In this context, the classical Walker A lysine seems to have a
so far underestimated regulatory function. Such a molecular switch mechanism of the Walker
A lysine is not restricted to Clp1 exclusively, since the PDB database provides a significant
number of crystal structures showing a similar “arrested“ conformation of the Walker A
lysine. Thus, an additional function of the Walker A lysine as a molecular switch in enzyme
catalysis is suggested. In conclusion, this thesis provides the first crystal structures of Clp1,
elucidating its RNA-specificity as well as the phosphoryl transfer reaction mechanism.
ix
Zusammenfassung
Polynukleotidkinasen (PNKasen) sind essentielle Enzyme, die in DNA- und RNA-Reparatur-,
RNA-Reifungs- und RNA-Abbau-Prozessen beteiligt sind. Diese Enzyme besitzen eine
konservierte PNKase Domäne, die strukturelle Homologie zu den klassischen P-loop Kinasen
aufweist.
Der
Reaktionsmechanismus
der
PNKasen
beinhaltet
den
Transfer
der
γ-Phosphatgruppe von ATP zu der 5’-Hydroxylgruppe eines Polynukleotid-Substrats.
Abhängig von ihrer Funktion können PNKasen unterschiedliche Substratspezifitäten besitzen.
Interessanterweise wurden erst kürzlich eukaryotische PNKasen identifiziert, die spezifisch
RNA phosphorylieren. Diese einzigartigen RNA-spezifischen PNKasen sind Teil einer
Subfamilie der PNKasen, der sogenannten Clp1 Proteinfamilie. Clp1 war die erste bekannte
RNA-spezifische PNKase in Eukaryoten und wurde zum Namensgeber dieser neuen
Proteinfamilie.
Es
wurde
gezeigt,
dass
humanes
Clp1
an
verschiedenen
RNA-Reifungsprozessen beteiligt ist: (i) dem Schneiden und die Polyadenylierung von RNA
Polymerase II unreifen mRNA Transkripten, (ii) dem tRNA Splicing, und (iii) der
Phosphorylierung synthetischer siRNAs. Trotz intensiver Studien an Clp1 ist nur wenig
Mechanistisches
über
die
RNA-Spezifität
und
den
Reaktionsmechanismus
des
Phosphoryl-Transfers von Clp1 bekannt.
Es war daher das Ziel dieser Doktorarbeit, Clp1 als RNA-spezifische PNKase sowohl
strukturell als auch funktionell detailliert zu untersuchen. Während dieser Arbeit wurde die
erste dreidimensionale Kristallstruktur von Clp1 aus Caenorhabditis elegans mit atomarer
Auflösung bestimmt. Clp1 ist ein Multidomänenenzym mit einer zentralen PNKase Domäne,
flankiert von einer N- und C-terminalen Domäne. Auf der Grundlage diverser RNA
substratetratgebundener Strukturen, die sich in Länge und ihrer Seqeuenz der RNA
unterschieden,
wurde
ein
„RNA-Sensor“
identifiziert,
welcher
spezifisch
die
2’-Hydroxylgruppe des endständigen Nukleotids erkennt. Zusätzlich konnten in dieser Arbeit
die enzymatisch relevanten Konformationen des substratgebundenen Zustandes, des
Übergangszustandes und des produktgebundenen Zustandes beschrieben werden. Mit Hilfe
dieser
Strukturen
wurde
ein
allgemeingültiges
Modell
des
Phosphoryl-Transfer
Reaktionsmechanismus von PNKasen erarbeitet. Interessanterweise ist im Unterschied zu
bereits bekannten PNKasen die ATP-Bindungsstelle innerhalb der PNKase-Domäne durch die
zusätzliche N-terminale Domäne abgeschirmt. Sowohl die strukturellen Ergebnisse als auch
biochemische Analysen in dieser Doktorarbeit konnten zeigen, dass die N-terminale Domäne
entscheidend für die ATP-Bindung in Clp1 ist. Die Phosphatgruppen des ATP Moleküls sind
x
Zusammenfassung
im katalytischen Zentrum verankert. Das katalytische Zentrum wird von dem bekannten
P-loop Motiv, einem divalenten Metall-Cofaktor (Mg2+) und durch ein sogenanntes
α-helikales „LID-Modul“ geformt. Mittels ortsgerichteter Mutagenesestudien konnte die
essenzielle Rolle des Walker A Lysins für die Enzymkatalyse gezeigt werden. Darüber hinaus
zeigt die Kristallstruktur von Clp1 eine „arretierte“ Konformation des Walker A Lysins. Das
Lysin fungiert in Clp1 als ein molekularer Schalter, der erst im Übergangszustand vollständig
aktiviert wird. Im Gegensatz zu anderen Nukleotidkinasen scheint Clp1 eine Art
„Gating-Mechanismus“ zu besitzen, welcher nicht-prozessive ATP-Hydrolyse verhindern
kann. Diese „arretierte“ Konformation des Walker A Lysins übernimmt damit eine
zusätzliche regulatorische Aufgabe in der Enzymkatalyse. Eine detaillierte Analyse der PDB
Datenbank identifizierte eine signifikante Anzahl an Kristallstrukturen die ebenfalls ein
Walker A Lysin in „arretierter“ Konformation besitzen. Mit den Ergebnissen dieser
Doktorarbeit wird die Funktion des klassischen Walker A Lysins auf die eines molekularen
Schalters in der Enzymkatalyse erweitert. Insgesamt führten die strukturellen und
biochemischen Analysen über Clp1 zu neuen Erkenntnissen hinsichtlich der RNA-Spezifität
als auch des Phosphoryl-Transfer-Reaktionsmechanismus.
xi
1 Introduction
Publications
The results of this dissertation are provided in the following manuscript:
Dikfidan A, Loll L, Clausen T, and Meinhart A (2013). Mechanism of RNA
5’-phosphorylation by eukaryotic RNA Polynucleotide Kinases. Manuscript in progress.
Additional publications:
Lüddeke F, Dikfidan A, Harder J (2012). Physiology of deletion mutants in the anaerobic
β-myrcene degradation pathway in Castellaniella defragrans. BMC Microbiol. 4;12:192.
Lüddeke F, Wülfing A, Timke M, Germer F, Weber J, Dikfidan A, Rahnfeld T, Linder D,
Meyerdierks A, Harder J (2012). Geraniol and geranial dehydrogenases induced in anaerobic
monoterpene degradation by Castellaniella defragrans. Appl Environ Microbiol.
Apr;78(7):2128-36.
xii
1 Introduction
1 Introduction
In nature, the chemistry of the 5’-terminus of nucleotides has a strong impact on the fate of
DNA and RNA molecules. DNA and RNA synthesis reactions rely on nucleoside
triphosphates, the building block of the nucleotide metabolism1,2. Nucleoside triphosphates
originate from precursor nucleosides that are activated in a cascade of phosphorylation
reactions3. Thus, newly synthesized polynucleotides typically bear a triphosphate at their
5’-termini. The 5’-triphosphate is chemically stable and persists until its removal or
enzymatic modifications4,5. Interestingly, the phosphorylation state of the 5’-terminus is
important in various cellular processes. In RNA degradation, the 5’-terminus is specifically
recognized by both prokaryotic and eukaryotic exonucleases6,7. In both cases mRNA decay
was initiated after conversion of the 5′-triphosphate (or m7GpppN (7-methylguanosine) cap
in eukaryotes) to a 5’-monophosphate (5’-phosphate)8,9, whereas in case of mRNA with a
5’-triphosphate or a 5’-hydroxyl group, mRNA decay was impaired10-16. Furthermore, the
5’-terminus is also important for enzymes of DNA and RNA repair pathways17,18. These
repair enzymes require a 5’-phosphate group for an efficient enzymatic activity. Damaged
nucleotides, however, are characterized by a 5’-hydroxyl terminus that needs to be
enzymatically processed for efficient repair. Because of the biological relevance of a
signature 5’-phophate for the nucleotide metabolism, a protein family involved in the
processing of the 5’-hydroxyl groups is emphasized in this thesis. Polynucleotide kinases
(PNKs) enzymatically alter the 5’-phosphorylate state of polynucleotides to a
5’-phosphate. In contrast to the knowledge about PNKs in the DNA metabolism, our
understanding of their function in RNA maturation and degradation pathways still needs to
be improved considerably. However, the recent identification of novel RNA-specific PNKs
(RPNKs) has led to new insights into the field of RNA research. Structural and functional
characterizations of these RPNKs are very limited and need to be the subject of future
work.
1.1
Importance of Polynucleotide kinases in the nucleotide
metabolism
In general, PNKs have an evolutionary conserved reaction mechanism that catalyzes one of
the most frequent reactions in biological systems19: a phosphoryl transfer reaction by
1
1 Introdduction
which tthe γ-phospphate group of a nucleeoside triphosphate is transferred
t
to a pletho
ora of
acceptor moleculess20. In casee of PNKs, the γ-phossphoryl gro
oup of an A
ATP molecu
ule is
transferrred to DN
NA and RN
NA polynuccleotide sub
bstrates (Fig
gure 1.1.1)). As mentioned
previouusly, the phhosphorylattion state oof nucleotiides is imp
portant in various ceellular
processees and the following sections w
will describ
be in detaill the role oof PNKs in the
nucleotiide metaboolisms and the impo rtance of the signatu
ure 5’-phossphate in RNA
degradaation.
Figure 1
1.1.1 Schem
matic repres
sentation off the 5’-pho
osphoryl tra
ansfer react ion catalyze
ed by
PNKs. T
The γ-phosphate of an ATP
A
molecul e is transferrred to depro
otonated 5’-hhydroxyl term
mini of
DNA or R
RNA polynuccleotide subs
strates.
1.1.1 P
PNKsinDNArepairpathwayss
Accuratte maintenaance and errrorless reprroduction of the genom
me is a prerrequisite fo
or cell
survivall21. Howeveer, damage of
o cellular D
DNA is an unavoidable
u
e threat to ggenomic inteegrity
that is iimplicated in the etiollogy of maany diseasess22-25 and reepresents a major facttor in
aging26. DNA dam
mage is causeed by variouus physical and chemiccal stress coonditions su
uch as
2
1 Introduction
ionizing radiation27 or reactive oxygen species28. However, DNA damage is not necessarily
linked to stress conditions, since it can also arise during regular cellular processes such as
DNA replication, recombination, or differentiation29-31.
Damaged DNA is characterized by a heterogeneous class of lesions including base
modifications, base excision, and strand breaks29. Exposure of cells to ionizing radiation
for instance induces elevated levels of hydroxyl radicals, which are chemicals suspected to
generate DNA single strand breaks (SSB; Figure 1.1.2 A)32. Without end-healing, SSB can
turn into the most lethal form of DNA damage, the double strand break (DSB; Figure
1.1.2 A)33. Other sources for DNA lesions are a stalled topoisomerase I after camptothecin
treatment (Figure 1.1.2 A)34 or nucleases, such as DNase II35,36. Many of these lesions
leave DNA termini that are unsuitable for subsequent extension or ligation reactions
mediated by DNA polymerases and ligases. Incompatible DNA-termini are composed of
3′-phosphate and free 5′-hydroxyl termini17. Repair enzymes however, require a free
3′-hydroxyl group for a subsequent elongation reaction and, in case of a ligation reaction,
additional 5′-phosphate termini17.
To ensure genetic stability, cells rely on a battery of different repair mechanisms
that are counteracting DNA lesions (Figure 1.1.2 A). The mammalian PNK (mPNK) is a
multi-functional enzyme that possesses the capacity to both phosphorylate 5′-hydroxyl and
dephosphorylate 3′-phosphate termini37,38. The mPNK is considered as a key component
for DNA repair contributing to three different pathways: Base excision repair39, DNA SSB
repair40 and non-homologous end-joining41.
Structurally, mPNK is a multi-domain enzyme consisting of an N-terminal FHA
(forkhead-associated) domain and two spatially separated catalytic domains; a
3’-phosphatase and a PNK domain42. The FHA domain recognizes phospho-threonine
residues on specific target proteins42-44 and thereby recruits mPNK to distinctive repair
complexes45. This active recruitment of mPNK and other DNA repair enzymes protects
cells against the release of premature intermediates. Interaction partners of the FHA
domain are the two scaffold proteins, XRCC1 and XRCC4. The SSB repair complex
(mPNK, DNA polymerase β, and DNA ligase III) is organized by XRCC143, whereas DSB
repair complex (mPNK, DNA end-binding protein Ku70/80, protein kinase DNA-PK and
the XRCC4/ligase IV heterodimer) depends on XRCC444. As a key factor for DNA repair,
the mPNK appears as the recurrent linker in the distinctive DNA repair pathways.
3
1 Introduction
1.1.2 PNKsinRNArepairpathways
The first PNK was discovered 1965 in bacteriophage-infected Escherichia coli 46-48. Since
then the T4 PNK has become one of the best-characterized PNKs and it is widely used for
applications in molecular biology49,50 Similar to mPNK, the T4 phage-encoded PNK is a
multifunctional enzyme composed of an N-terminal PNK domain and a C-terminal
phosphatase domain51,52. Together with the T4 ligase53, it is shown to be involved in RNA
repair, enabling the T-even phage to overcome the host’s suicide defense mechanism18
(Figure 1.1.2 B). In detail, if bacteria encounter a T4 phage infection, the Stp DNA
restriction inhibitory peptide is activated, which subsequently activates a latent tRNA
anti-codon nuclease54-56. The active nuclease triggers the lesion of the bacterial lysine
tRNA, inhibits protein synthesis and thereby phage propagation18. The cleaved tRNA
molecules have a characteristic free 5’-hydroxyl terminus and, in contrast to DNA repair
processes, a 2’,3’-cyclic phosphate group18. In order to provide acceptable substrates for
T4 ligase, T4 PNK phosphorylates the 5’-hydroxyl terminus46 and removes the cyclic
phosphate group at the 2’,3’-terminus in a two-stage process18,57,58. Subsequently, T4 ligase
can seal the broken 5’- and 3’-termini of the tRNA59. Notably, T4 PNK and mPNK are
functionally similar, but they have different polynucleotide specificity. The enzyme
displays the broadest substrate specificity on DNA and RNA molecules compared to other
described PNKs60-63.
Recently, another RNA repair system was also discovered in bacteria repairing
ribotoxin-cleaved RNAs64. Ribotoxins are proteins responsible for the site-specific
cleavage of RNA molecules involved in transcription and transition65-67. In contrast to
classical RNA repair pathways relaying on 5’-phosphorylation and 3’-dephosphorylation,
this novel system additionally methylates the 2’-hydroxyl group of RNA64 (Figure
1.1.2 B). Methylated RNA sites are thereby protected against repeated cleavage. The
enzymatic reaction involves the bacterial Pnkp/Hen1 complex68. Bacterial PNK, Pnkp, is
composed of three catalytic domains, an N-terminal PNK domain, a central phosphatase
domain, and a C-terminal ligase domain69, all being required for RNA repair70. The
presence of RPNKs in bacteria emphasizes the relevance of RNA repair in the prokaryotic
nucleotide metabolism. This fact might also be true for a eukaryotic pathway. This
hypothesis became evident just recently with the discovery of several RPNKs, which are
presented in section 1.2.
4
1 Introduction
1.1.3 PNKsinRNAmaturationandRNAturnover
Besides their functions in DNA and RNA repair mechanism, RPNKs are also involved in
tRNA maturation71. Intron sequences in tRNA precursor molecules (pre-tRNAs) are found
in all three kingdoms of life72. To obtain a mature tRNA these intron sequences have to be
removed by splicing from the pre-tRNAs73. After cleavage by the tRNA-splicing
endonuclease complex, the intron-cleaved tRNA are composed of a free 5’-hydoxyl group
at the 3’-exon and a 2’,3’-cyclic phosphate group at the 5’-exon71. To become rejoined into
a mature tRNA molecule, the 5’- and 2’,3’-termini have to be processed by a
multi-functional tRNA ligase, named Trl174 that consists of three enzymatic activities in
functionally independent domains, a cyclic phosphodiesterase domain, a PNK domain, and
an RNA ligase domain75,76. Trl1 is able to first open the 2’-3’-cyclic phosphate74, than to
phosphorylate the 5’-hydroxyl terminus74, and finally to rejoin both exon halves to provide
a mature tRNA molecule74. Interestingly, no homolog Trl1 tRNA ligase was identified in
higher eukaryotes, suggesting that Trl1 is a distinctive enzyme of yeast tRNA maturation
pathway (Figure 1.1.2 C).
However, the signature 5’-phosphate is not only important for sealing reactions by
ligases. RNA turnover is also controlled by the phosphorylation state of nucleotides at the
5’-terminus8-16. In contrast to prokaryotes, the mRNA of eukaryotes is modified at the
5’-terminus, where methylated GMP is incorporated to form the m7GpppN cap5. Despite
these differences in 5’-capping, the processes in mRNA decay are remarkably similar
between eukaryotes and prokaryotes. In both cases, the 5′-phosphate represents the
preferential substrate for nucleases in RNA turnover by exonucleases10-16. In eukaryotes,
prior to the RNA decay the 5’-cap is hydrolyzed to a 5’-phosphate77,78 Once the 3′-poly(A)
tail is removed, the mRNA becomes degraded from the 5′-terminus by the 5′-3′
exonuclease Xrn179,80 (Figure 1.1.2 C). Furthermore, the signature 5’-phosphate couples
the specificity of Xrn1 to its processivity7. In prokaryotes, the 5’-triphosphate of the
mRNA is hydrolyzed to 5’-phoshate8 and the mRNA is subsequently committed to
degradation by RNase E, an enzyme that recognized the signature 5’-phosphate6 (Figure
1.1.2 C). In addition, nucleases involved in RNA interference81 (RNAi) also depend on the
signature 5’-phosphate. Two prominent examples of enzymes that recognize the
5’-phosphate are the RNA induced silencing complex (RISC)82 (Figure 1.1.2 C) and the
DICER83 nuclease. Strinkingly, a signature 5’-phosphate is also generated by enzymatic
5
1 Introdduction
cleavage of RISC84. Thus, thee molecularr basis of th
he signaturee 5’-phosphhate has a ceentral
role in tthe nucleotide metaboliism that enssures contro
olled decay of RNA moolecules.
6
1 Introduction
Figure 1.1.2. PNK involved in the nucleotide metabolism. A, mPNKs are a recurrent linker for
repair of different lesion types. B, T4 PNK together with T4 ligase heals lesions of tRNA to
overcome the host defense mechanism. PNKp-Hen1 complex is involved in a special RNA repair
pathway that methylates the RNA after ligation to confer protective immunity. C, Trl1 is a
tri-functional RNA ligase associated with tRNA maturation. The 5’-kinase activity of Trl1 is
important to re-ligation of the 5’- and 3’-exons to produce a mature tRNA. Efficient mRNA decay
relies on the presence of a signature 5’-phosphate. The exonucleases Xrn1 and RNase E only
show efficient enzymatic activity on mRNA substrates with 5’-phosphate group. mRNAs are
committed to degradation by decapping or removing the triphosphate group (RppH). In addition,
the signature 5’-phosphate group is also required for siRNAs and miRNAs to become incorporated
into RISC. Adapted from17,71.
1.2
Clp1,anoveleukaryoticRPNK
Since eukaryotic RNA-maturation and RNA-degradation processes strongly rely on the
presence of 5’-phosphorylated nucleotides, the existence of RPNKs in eukaryotic
organisms was conceivable. Although this particular enzymatic reaction was already
described in 1979 in HeLa cell extracts85, characterization of the responsible enzymes
started only in 200786-88. Human Clp1 (hsClp1) is the first identified eukaryotic RPNK and
is involved in different RNA maturation pathways; (i) in cleavage/polyadenylation of RNA
polymerase II transcripts89,90, (ii) in tRNA maturation52,90-92, and (iii) in 5’-phosphorylation
of synthetic siRNA molecules88,93.
Eukaryotic Clp1 orthologs display a high degree of evolutionary conservation94.
Interestingly, Clp1-like PNKs were also identified in plants, archaea and are predicted in
prokaryotes86,95,96, thereby covering all three kingdoms of life. Based on their sequence
conservation, these homologous proteins can be grouped into a novel subfamily of PNKs,
the Clp1 protein family. Despite their structural and sequence similarities, not all
eukaryotic Clp1 orthologs show 5’-kinase activity. Paradoxically, in contrast to hsClp1,
Clp1 from Saccharomyces cerevisiae (scClp1) appears to be enzymatically inactive94 and
functionally different from hsClp1. Consistent with this observation is the fact that hsClp1
is unable to compensate for the loss of scClp197. Although enzymatically inactive, scClp1
is still able to bind ATP molecules94,98-100. Thus, it was suggested that ATP-binding might
be required for structural stability of the protein98-100.
Since archaeal Clp1 from Pyrococcus horikoshii (phClp1) is enzymatically active95,
it is conceivable that 5’-kinase activity is ancestral to the Clp1 protein family. scClp1,
therefore, represents a variant that has lost its enzymatic activity due to distinctive features
of the tRNA maturation pathway in yeast (see 1.1.3). Furthermore, it is anticipated that
7
1 Introdduction
scClp1 solely funcctions in pree-mRNA 3′--terminus processing
p
as
a an integrral subunit of
o the
988-100
3’-cleavvage polyaddenylation machinery
m
, whereas hsClp1 represents
r
aan RPNK th
hat is
involvedd in multiplle RNA matturation patthways as a recurrent liinker88,90,92,997.
1.2.1 h
hsClp1’sin
nvolvemen
ntinRNAm
metabolism
(i) mRN
NA processsing. Clp1 was origiinally identtified as a member off the eukarryotic
mRNA 3’-cleavagee and polyaadenylation machinery
y89,101,102 (Figure 1.2.1 A
A). Compo
onents
of this m
machinery are
a structurrally conserrved in eukaaryotes103. scClp1
s
is paart of the CF
C IA
complexx, and has been show
wn to interaact Cleavag
ge–Polyaden
nylation Faactor (CPF)98. In
mammaals, Clp1 caan be foun
nd as a subbunit of thee CF IIm, suggesting a function
n as a
bridgingg factor beetween CF Im and CP
PSF89. The interaction
n between Clp1 and Pcf11
P
another subunit off the cleavage factor, is evolutio
onary conserved betw
ween scClp1
1 and
hsClp1 orthologs103.
8
1 Introduction
Figure 1.2.1: Proposed function of hsClp1 in mRNA 3’-end processing. A, hsClp1 is part of the
3’-end processing machinery and required for cleavage and polyadenylation of pre-mRNA. There it
acts as a bridging factor and interactions Pcf11. B, hsClp1 might has a function in transcription
termination or mRNA decay by maintaining the signature 5'-phosphate group on the RNA. The
5‘-nascent transcript already contains a terminal 5’-phosphate after cleavage. This phosphate is
required for dissociation of the Pol II from the DNA by Xrn2 “torpedo model”104. hsClp1 is
suggested to maintain this phosphate. Furthermore, hsClp1 might have an additional function in
maintenance of the 5'-phosphate of mRNAs committed to degradation by Xrn1. Adapted from92.
After identification of hsClp1’s RNA-specific 5’-kinase activity, Martinez and
Weitzer suggested the involvement of hsClp1 in mRNA degradation88,92 (Figure 1.2.1 B).
Although the 5’-phosphate is important for efficient degradation by the exonucleases
Xrn17 and Xrn215, both mRNA and its nascent transcript show no free 5’-hydroxyl
termini5,105,106. Therefore, it was assumed that Clp1 is involved in maintaining the
phosphorylation state of the 5’-terminus to counteract putative phosphatases92. This
hypothesis, however, is in contrast to recent results showing that the 5’-kinase activity of
hsClp1 is dispensable for mRNA 3′-end cleavage90. Similar to scClp1, hsClp1 seems to
function only as a scaffold protein in the mRNA 3’- cleavage and polyadenylation
machinery. Consistent with this assumption is the oberservation that the generation of a
Clp1-knockout produced a lethal phenotype in mice, but a kinase deficient mutant
(K127A) was viable90.
(ii) tRNA processing. The endonuclease complex for tRNA-splicing is associated with
hsClp191. Since hsClp1 was additionally shown to 5’-phosphorylate intron cleaved
tRNAs88, it is likely that this enzyme is involved in tRNA-splicing (Figure 1.2.2).
Currently, there are several postulated tRNA-splicing pathways71. However, all of them are
based on the same underlying principle that splicing of pre-tRNA occurs in two steps:
cleavage and rejoining of exon halves71. Characteristic ends for this cleavage reaction are a
5’-hydroxyl terminus and a 2’,3’-cyclic phosphate, respectively71. Apart from the yeast
ligation pathway, mammals harbor a distinctive pathway that relies on a 2’,3’-cyclic
phosphate ligase named HSPC117107. This enzyme rejoins the splice site junction by
ligation of the 2’,3’-cyclic phosphate with a free 5’-hydroxyl group and consequently a
5’-kinase activity is obsolete for this mechanism71. However, in addition to the
HSPC117-dependent pathway, a second “yeast-like” pathway is hypothesized, involving a
PNK (hsClp188,92), a cyclic phosphodesterase108, and phosphotransferease109. However,
5’,3’-RNA ligase is still missing110. The yeast-like ligation pathway also explains the
9
1 Introdduction
8
functionnal differennce between
n scClp194 aand hsClp188
since yeaast Trl1 alreeady contaiins an
additionnal RNA-sppecific 5’-kinase activitty76.
10
1 Intro
oduction
Figurre 1.2.2: Pro
oposed func
ction of hsC
Clp1 in tRNA
A-splicing. In
ntron sequennces of pre-tR
RNAs are
remo
oved by enzyymatic activitty of the tRN
NA-splicing endonuclease
e
e complex. hhsClp1 was shown to
be asssociated to this complex91. The clea
avage reaction leads to a 2',3' cyclicc phosphate group at
the 5’-exon and a 5'-hydroxyl group at the
e 3’-exon. Siimilar to the yeast ligatioon pathway (T
Trl1-like),
hsClp
p1 phosphorrylates the 5’-terminus
5
o
of intron-cleaved 3’-exons. In combbination with
h a cyclic
phosp
phodiesterasse and RNA ligase activi ty, the 3’- an
nd 5’-exon ha
alves can bee rejoined to obtained
a m
mature tRNA
A. The rem
maining 2’-p
phosphate can be re
emoved by the activity of a
2′-pho
osphotransfe
erase. Adaptted from92.
Figurre 1.2.3: Prroposed fun
nction of hs
sClp1 in RN
NAi. hsClp1 was shownn to 5’-phosphorylate
synth
hetic siRNAss. This signature 5’-phossphate group
p on the RNA
A is requiredd for RISC assembly
a
and subsequentt gene sile
encing. Alth
hough Dicer-cleaved siRNAs
s
are generated with a
osphate, hsC
Clp1 may have a function
n in the main
ntenance the phosphate tto counterac
ct putative
5'-pho
RNA phosphatases. The protection of the
e signature 5’-phosphate
5
ensures effiicient RISC assembly
a
and sstability of an
n Ago2-siRNA
A complex. A
Adapted from
m92.
m Remirez and colleaagues show
wing that
Conssistent withh this hypothesis are rresults from
expreession of hssClp1 in budding yeastt can compllement cond
ditional andd lethal mutations in
9
the kkinase moduule of yeast or plant tRN
NA ligases97
.
In contraast, trans-complementtation was impossible
i
with overexxpressed sccClp1 or
kinasse-defectivee hsClp1977. These rresults sup
pport a putative
p
roole of hsC
Clp1 in
11
1 Introduction
5'-terminus-healing similar to the yeast-like tRNA-splicing pathway92. Furthermore, recent
experiments with transgenic kinase-defective Clp1 mice identified a novel tyrosine pretRNA as a substrate of Clp1’s 5’-kinase activity90. Thus, Clp1 from higher eukaryotes
seems to be involved in the phosphorylation of tRNA 3'-exon halves.
(iii) RNAi. Besides its function in the 3’-terminus processing machinery89 and its role in
tRNA processing88,90,92, hsClp1 was shown to 5’-phosphorylate synthetic siRNAs in
vitro88,93 (Figure 1.2.3). The signature 5’-phosphate is required for the incorporation of
siRNA and miRNAs into RISC111. Although siRNAs and microRNAs already contain a
5’-phosphate group111,112, Weitzer and Martinez speculated whether hsClp1 or other PNKs
(Nol9) are involved in the maintenance of this phosphate groups and thereby counteract
putative phosphatases92. Recent results showed, however that Clp1’s 5’-kinase activity is
not essential for gene silencing by miRNAs90.
In conclusion, hsClp1 participates in multiple RNA maturation pathways88,90,92. However,
the 5’-kinase activity seems to be negligible for the function in the 3’-end processing of
pre-mRNA90. Both scClp1 and hsClp1 are rather functioning as a structural scaffold for
mRNA maturation. But in contrast to scClp1, hsClp1 is additionally part of the
tRNA-splicing machinery91 and involved in tRNA maturation88,90,92.
1.2.2 StructuralcharacterizationofscClp1
The three-dimensional structure of the enzymatically inactive scClp1 in association with a
fragment of its CF IA interaction partner Pcf11 was determined by X-ray crystallography
at medium resolution (Figure 1.2.4 A and B)94. This analysis revealed a multi-domain
protein composed of a central PNK domain, sandwiched between additional N- and
C-terminal domains (NtD and CtD). The NtD (amino-acids 1 to 100) is composed of a
β-sheet sandwich, whereas the PNK domain (amino-acids 101 to 341) contains the active
site region of the protein. The PNK domain adopts a classical , fold that is similar to the
previously described T4 PNK51,52 and the DNA-specific mammalian PNK42 (Figure 1.2.4
C). Characteristic for the PNK domain are two evolutionarily conserved motifs, the P-loop
and the Walker B motif113,114. The P-loop is typically found in adenine or guanine
triphosphate binding proteins19 and it corroborates scClp1’s ability to bind ATP molecules.
In contrast to T4 PNK and mPNK, the ATP binding site of scClp1 is obstructed by the
12
1 Introduction
additional NtD94 (Figure 1.2.4 C). Furthermore, the amino acid sequences as well as the
structural architecture of the proteins suggest that scClp1 and T4 PNK or mPNK have not
evolved from a common ancestor. The CtD (amino-acids 342 to 446) is folded in a mixture
of α-helices, -strands and random coils.
1.2.3 TheRNA‐specificClp1proteinfamily
Based on recent bioinformatical characterizations of hsClp1, the Clp1 protein family can
be extended by two newly identified homologs, Nol9 and Grc386,87. These closely related
proteins also contain a highly conserved P-loop motif, a Walker B motif and were shown
to be RPNKs86,87. Nol9 and Grc3 are widely conserved eukaryotic proteins and display an
extensive sequence similarity to the PNK domain of other members of the Clp1 protein
family, whereas the additional NtD and CtD are less conserved. Interestingly, in the
archaeal Clp1 homolog the NtD is even missing95. All biochemically characterized
members (hsClp1, phClp1, Nol9 and Grc3) of the Clp1 protein family were shown to
phosphorylate a variety of different polynucleotide substrates in vitro, having preference
for single-stranded RNA as well as double-stranded RNA and DNA substrates86-88,95.
Therefore, the structural features of the PNK domain seem to define the in vitro specificity
of the polynucleotide substrate. Despite of this promiscuous RNA-phosphorylation,
members of the Clp1 protein family contribute specifically to different RNA maturation
pathways. Exemplarily, Nol9 and Grc3 were shown to be involved in rRNA
maturation86,87,115,116, whereas hsClp1 is described as a recurrent linker in mRNA and
tRNA processing88,90,92. This observation might be explained by regulatory functions of
NtD and CtD. As known for mPNK, the N-terminal FHA domain serves as a
protein-protein interaction platform that recruits mPNK to its molecular target42. This
domain-dependent recruitment would explain the distinguishable in vivo substrate
specificities of hsClp1 compared to Nol9 and Grc3. Furthermore, scClp1 was shown to
interact with a variety of proteins involved in 3’-end processing, specifically with its
central domain and the additional NtD and CtD98. However, it remains unclear if these
additional domains also play a role in enzyme specificity for the enzymatically active
members of the Clp1 protein family. Therefore, the function of additional domains and the
identification of putative interaction partners will be of future interest to understand the in
vivo function of RPNKs in more detail.
13
1 Introdduction
Figure 1
1.2.4. Structture of scClp1 in comp
parison with
h the structurally homo
ologous pro
oteins
T4 PNK
K and mPNK
K. A, A ribbo
on representa
ation of scClp1 showing the N-termi nal domain (NtD),
ATP-bind
ding domain (AbD) and the
t C-termin al domain (C
CtD). Structural motifs im
mportant for liigandbinding a
are highlighte
ed in purple (P-loop) and
d green (Wallker B). scClp
p1 was crysttallized in complex
The orientattion is
with scP
Pcf11 (PDB 2NPI). ATP and scPcf1 1 are shown as a stick
k model. B, T
rotated b
by 180 degrree. C, The homologouss structures of the kinas
se domains of T4 PNK (PDB
1RRC) a
and mPNK (PDB
(
3ZVN)) were superrimposed to scClp1 based on their P-loop motiff. The
bound A
ADP molecule
es of T4 PNK
K and mPNK
K are shown as a sick mo
odel (yellow).. The ATP-binding
site of scClp1 is covvered by the NtD, where
eas the T4 PNK
P
and the mPNK show
w an “open” ATPbinding ssite.
1.3
Thephosphorylttransferreaction
nmechanism
Since hhsClp1 beloongs to the P-loop kinnase, the reeaction mecchanism off the phosp
phoryl
transferr is evolutioonary conseerved and sshould reseemble the one
o of previ
viously desccribed
P-loop kkinases117. Currently, our knowleedge about this reactio
on mechanissm mainly relies
on expeeriments obtained from
m mononuclleotide kinaases (NMP-kinases)118 and GTPasses119.
14
1 Introduction
However, to fully understand enzyme catalysis of PNKs the knowledge about the nature of
the transition state is mandatory. So far, structural details of the PNK-catalyzed phosphoryl
transfer reaction mechanism are missing and need to be further investigated. The transition
state theory distinguishes between two extreme cases that are characterized by the
sequence of bond formation and bond breakage of the leaving group120. A dissociative
transition state is described as a metaphosphate, the bond to the leaving group is fully
broken and the bond to the nucleophile absent120 (Figure 1.3.1). In contrast to the
dissociative transition state, the associative transition state shows a penta-coordinated
leaving group120 (Figure 1.3.1). Depending on the nature of the transition state, charges are
differently distributed and need to be specifically compensated (Figure 1.3.1).
Theoretically, the transition state is inaccessible for structural characterization, due to its
marginal lifetime. Therefore, a special experimental setup is necessary to obtain structural
information. Instead of capturing the real transition state, compounds are used that mimic
the transition state geometry and charge distribution121. In several crystal structures
aluminum fluoride, beryllium fluoride, and vanadate were shown to function as transition
state analogs for the transferred phosphoryl group122,123. Currently, no structural data of a
transition state analog in PNKs are available. This information, however, would greatly
facilitate our understanding of the reaction mechanism and amongst other questions is
addressed in this work.
A characteristic feature shared by kinases is an evolutionary conserved P-loop
motif with the consensus sequence GxxxxGK[S/T]19. The phosphates of the nucleoside
triphosphate are bound by a hydrogen-bonding network formed by the main-chain amides
of the P-loop113. The Walker A lysine of the P-loop interacts electrostatically with the
β- and γ-phosphates in a bifurcated manner113. The metal cofactor (Mg2+) is octahedrally
coordinated by the conserved Ser/Thr residue, the oxygen atoms of the β- and
γ-phosphates, and water molecules19. The Walker B is classically composed of a catalytic
aspartate that activates the acceptor oxygen by deprotonation, a prerequisite for the
nucleophylic attack on the γ-phosphate group of the ATP molecule52.
15
1 Introdduction
Figure 1
1.3.1. Nature
e of the tra
ansition statte for phosp
phoryl trans
sfer reactio ns. The tran
nsition
state the
eory describe
es the phosphoryl transffer reaction either with an
a associativve (SN2-like
e) or a
dissociattive (SN1-likke) mechan
nism. The d
differences between bo
oth transitionn states arre the
sequencce of bond formation and bond brreakage betw
ween the attacking nuccleophile an
nd the
bridging oxygen atom of the β- and γ-phos phate group
p. In case of an associattive mechan
nism a
pentagonal bipyrimid
dal structure is formed, w
whereas diss
sociative tran
nsition state involves a planar,
p
trigonal, metaphosph
hate (PO3–) structure.
s
Be
esides their different
d
geom
metry, the tw
wo transitions
s state
structure
es also differr in charge distribution. In
n contrast to an associattive transitionns state with
h a net
charge o
of –3, the dissociative cas
se has a net charge of –1
1. Adapted frrom124.
1.4
Scopeoffthisthessis
Even thhough the newly
n
identiified RPNK
Ks are functtionally well characterrized86,87,95,1115,116,
our struuctural know
wledge abo
out the prottein family
y is limited. hsClp1 w
was shown to be
involvedd in mRNA
A as well as tRNA pprocessing and RNAii as a recuurrent linkeer88,92.
Howeveer, the moleecular targeet of its 5’-kkinase activ
vity remainss elusive. T
To date, only
y few
crystal sstructures of
o PNKs witth bound nuucleotides1255,126 are avaiilable and nno structure of an
enzymaatically activve variant of Clp1 has bbeen provid
ded.
16
1 Introduction
Therefore, the first aim of this PhD thesis is to gain structural insights into Clp1’s
substrate specificity by X-ray crystallography. The structural data presented in this
thesis together with the comprehensive biochemical characterization provide an
experimental and conceptual framework to understand Clp1’s mode of RNA-recognition
and RNA-specificity.
The second aim of this thesis addresses Clp1’s mechanism of the phosphoryl transfer
reaction. Our current knowledge of the phosphoryl transfer reaction and the transition
state architecture is mainly based on observations made for NMP-kinases. The
characterization of crystal structures in catalytically important states of Clp1 provides a
model for the phosphoryl transfer reaction mechanism of PNKs. This model can be applied
in general since their active sites are highly conserved. Furthermore, based on the crystal
structure of an active Clp1 ortholog it was possible to identify enzymatically important
residues that are missing in scClp1.
The third aim was to decipher the function of NtD and CtD in an enzymatically active
Clp1. As indicated by the crystal structure of the inactive scClp1 ortholog, the central
domain containing the P-loop motif is sandwiched between additional NtD and CtD with
unknown functions94. In the case of scClp1, it was shown that these domains are involved
in protein-protein interactions98. To elucidate the function of these additional domains in
catalysis, truncated variants of Clp1 were designed and biochemically characterized. Based
on these experiments, it was possible to show that the NtD is essential for ATP binding,
whereas the CtD seems to be important in the correct positioning of enzymatically relevant
residues.
17
1 Introduction
18
2 Material and Methods
2 MaterialsandMethods
2.1 Materials
2.1.1 Chemicals
If not otherwise stated, all chemicals were purchased from Carl Roth (Karlsruhe, Germany),
Fluka (Deisenhofen, Germany), GERBU Biotechnik (Gaiberg, Germany), Honeywell/Riedel
de Haën (Seelze, Germany), Invitrogen (Paisley, UK), Merck (Darmstadt, Germany), NEB
(Schwalbach, Germany), Serva (Heidelberg, Germany), and Sigma-Aldrich (Deisenhofen,
Germany) in analytical grade purity.
2.1.2 Crystallizationscreens
Crystallization screen
Company
Classics Suite
Qiagen, Hilden, Germany
JCSG Core Suites
Qiagen, Hilden, Germany
PEGs and PEGs II Suites
Qiagen, Hilden, Germany
Ammoniumsulfate Suite
Qiagen, Hilden, Germany
Wizard I and II
Emerald, Bainbridge Island, WA, USA
Additive Screen
Hampton Research, Alison Viejo, CA, USA
19
2 Material and Methods
2.1.3 Buffers
Buffer name
Buffer composition
Lysis buffer 1
20 mM HEPES-NaOH pH 7.0
100 mM NaCl
10 mM β-mercaptoethanol
Lysis buffer 2
20 mM HEPES-NaOH pH 7.0
100 mM NaCl
5 mM β-mercaptoethanol
Anion-exchange buffer
20 mM HEPES-NaOH pH 7.0
1 mM MgCl2
2 mM dithioerythritol (DTE)
Size exclusion buffer
10 mM HEPES-NaOH pH 7.0
100 mM NaCl
1 mM MgCl2
2 mM DTE
Crystallization buffer
10 mM HEPES-NaOH pH 7.0
100 mM NaCl
1 mM MgCl2
4 mM tris(2-carboxyethyl)phosphine (TCEP)
Dialysis buffer
20 mM HEPES-NaOH pH 7.3
100 mM NaCl
2 mM MgCl2
4 mM CaCl2
2 mM DTE
2x Phosphorylation buffer
100 mM HEPES-NaOH pH 8.0
200 mM NaCl
30 mM MgCl2
20
2 Material and Methods
2x ATPase buffer
20 mM HEPES-NaOH pH 7.0
200 mM NaCl
28 mM KCl
20 mM MgCl2
1 mM ethylenediaminetetraacetic acid (EDTA)
0.5 mM
(NADH)
nicotinamide
adenine
dinucleotide
0.4 mM phosphoenol pyruvat (PEP)
SDS-PAGE running buffer
25 mM Tris-HCl pH 8.0
200 mM glycine
0.1 % (w/v) sodium dodecyl sulfate (SDS)
SDS-PAGE separation gel buffer
1.5 M Tris-HCl pH 8.8
0.4 % (w/v) SDS
SDS-PAGE stacking gel buffer
1.5 M Tris-HCl pH 6.8
0.8 % (w/v) SDS
SDS-PAGE stacking gel (5 %)
125 mM Tris-HCl pH 6.8
15 % (v/v) Rotiphorese® Gel 30 (37.5:1)
0.1 % (w/v) ammonium persulfate (APS)
0.1 % (v/v) tetramethylethylenediamine (TEMED)
SDS-PAGE separation gel (15 %)
375 mM Tris-HCl pH 8.8
0.1 % (w/v) SDS
50 % (v/v) Rotiphorese® Gel 30 (37.5:1)
0.08 % (w/v) APS
0.08 % (v/v) TEMED
10x TBE buffer
890 mM Tris-HCl pH 8.0
890 mM boric acid
20 mM EDTA
21
2 Material and Methods
RNA-urea gel (9 %)
38 % (v/v) Rotiphorese®
sequencing gel concentrate (19:1)
52 % (v/v) Rotiphorese® sequencing gel diluent
1x TBE buffer
0.08 % (w/v) APS
0.08 % (v/v) TEMED
6x DNA loading dye
30 % (v/v) glycerol
0.25 % (w/v) bromphenol blue
0.25 % (w/v) xylene cyanole FF
5x SDS loading dye
50 mM Tris-HCl pH 7.0
10 % (v/v) glycerol
10 % (w/v) SDS
0.1 % (w/v) bromophenol blue
0.9 M DTE
0.1 M β-mercaptoethanol
RNA-urea loading dye
95 % (v/v) formamide
5 mM EDTA
0.025 % (w/v) SDS
0.25 % (w/v) bromophenol blue
0.25 % (w/v) xylene cyanol FF
Methylene blue staining solution
0.4 M sodium acetate
2.28 % (v/v) acetic acid
0.01 % (w/v) methylene blue
TFB-1
30 mM potassium acetate pH 5.8
50 mM MnCl2
100 mM RbCl
10 mM CaCl2
15% (v/v) glycerol
TFB-2
10 mM MOPS-NaOH pH 7.0
10 mM RbCl
75 mM CaCl2
15% (v/v) glycerol
22
2 Material and Methods
2.1.4 Growthmedia
All bacterial cultures were grown in lysogeny broth (LB) medium. LB-agar plates for single
colony growth were solidified by addition of bacto-agar. Both LB medium and LB-agar plates
were routinely provided by the media kitchen of the Max Planck Institute for Medical
Research (Heidelberg, Germany). Depending on the used E. coli strain and the plasmid
constructs, different antibiotics were used as selection markers. For growth of E. coli strain
BL21-CodonPlus(DE3)-RIL, the media was supplemented with 34 μg/ml of chloramphenicol.
In case of expression experiments, the media additionally contained either 50 μg/ml of
kanamycin when cells were transformed with pET28b vector or 100 μg/ml of ampicillin after
transformation with pET21a vector.
Media name
Components
LB
10 g/l bacto tryptone
5 g/l bacto-yeast extract
10 g/l NaCl
Adjusted to pH 7.0 with NaOH
LB-agar plates
25 g/l Standard-I-nutrient-broth
15 g/l bacto agar
For
selenomethionine
labeling
BL21-CodonPlus(DE3)-RIL
were
of
protein,
grown
in
cell
cultures
minimal
of
media
E.
coli
supplemented
strain
with
selenomethionine instead of methionine. If not elsewhere stated, cells were treated similar to
native expression experiments. The media were prepared according to127.
23
2 Material and Methods
2.1.5 Bacterialstrains
Cells of E. coli strain DH5α were used for plasmid DNA propagation, whereas cells of strain
BL21-CodonPlus(DE3)-RIL were used for protein overexpression.
E. coli strains
Genotype
Reference
DH5α
F- φ80lacZΔM15 Δ(lacZYA-argF)U169
128
recA1 endA1 hsdR17(rk-, mk +)
phoA supE44 thi-1 gyrA96 relA1 λE. coli B F– ompT hsdS(rB– mB–)
Novagen,
dcm+ Tetr gal λ(DE3) endA The
USA
Wisconsin,
[argU ileY leuW Camr]
2.1.6 Plasmids
Protein expression of the different construct variants was performed with two plasmids of the
pET expression vector series (Novagene, Madison, USA). The full-length gene and truncation
variants thereof were cloned into the multiple cloning sites with the indicated restriction sites.
The pET28b vector carries both an N-terminal hexahistidine-tag fused to a thrombin
cleavage site and an optional C-terminal hexahistidine-tag sequence. The expression construct
was designed with an N-terminal hexahistidine-tag fusion. To prevent an additional Cterminal hexahistidine-tag, a stop-codon was introduced. In contrast to pET28b, the pET21a
vector carries only a C-terminal hexahistidine-tag. Both plasmids are isopropyl-β-Dthiogalactopyranosid (IPTG) inducible expression systems.
24
2 Material and Methods
Name
Description
Reference
pET28b
T7-based expression system
coding for a recombinant
protein of N-terminal
hexahistidine-tag, a thrombin
protease cleavage site fused
to the protein of interest;
KanR
Novagene
pET21a
T7-based expression system
coding for a recombinant
protein with C-terminal
hexahistidine-tag; AmpR
Novagene
pET28b-ceClp1
Cloned via NcoI/NotI
This thesis
pET28b-ceClp1-K127A
Cloned via NcoI/NotI
This thesis
pET28b-ceClp1-K127R
Cloned via NcoI/NotI
This thesis
pET28b-ceClp1-W233A
Cloned via NcoI/NotI
This thesis
pET28b-ceClp1ΔC104
Cloned via NcoI/NotI
This thesis
pET21a-ceClp1ΔN107
Cloned via NdeI/NotI
This thesis
pET28b-ceClp1ΔC310
Cloned via NcoI/NotI
This thesis
pET21a-ceClp1ΔNC315
Cloned via NdeI/NotI
This thesis
2.2 Methods
2.2.1 MolecularBiology
2.2.1.1 CompetentE.colicellsandtransformation
Chemically competent E. coli DH5α and E. coli BL21-CodonPlus(DE3)-RIL cells were used
for cloning and expression experiments, respectively. Chemically competent cells were
provided by Maike Gebhard (Max Planck Institute for Medical Research) and produced using
a modified protocol of Hanahan129. Cells from a glycerol stock were transferred onto an
LB-agar plate and incubated overnight. A single colony was used to inoculate an overnight
pre-culture with antibiotics if required. This pre-culture was diluted 1:100 into 0.5 l of LB
medium. The cell suspension was incubated at 37 °C until an OD600 of 0.5 was reached.
Afterwards, cells were harvested by centrifugation (6000 x g, 4 °C, 10 min). The supernatant
25
2 Material and Methods
was discarded and the cell pellet was washed with 50 ml of TFB-1. A final centrifugation step
was performed (6000 x g, 4 °C, 10 min) and the cell pellet was resuspended in 4 ml of TFB-2.
Aliquots of 50 μl were immediately flash-frozen in liquid nitrogen and stored at -80 °C until
further use.
For transformation of E. coli cells with plasmid-DNA, aliquots of chemically
competent cells were thawed on ice. After thawing, 1-2 µl of plasmid DNA (20-200 ng/μl)
was added to the aliquot and incubated on ice for additional 15-30 minutes. Cells were heat
shocked at 42 °C for 45 sec and immediately transferred onto ice for 2-3 minutes.
Subsequently, 450 μl LB medium was added to the aliquot and incubated at 37 °C for
60 minutes under constant shaking at 750 rpm in a thermomixer comfort (Eppendorf,
Hamburg, Germany). Prior to plating, the transformed cells were harvested by centrifugation
at 4500 rpm in a tabletop centrifuge and resuspended in approximately 50 μl of LB medium.
This cell suspension was plated on a pre-warmed LB-agar plate supplied with appropriate
antibiotics. Plates were incubated at 37 °C overnight for single colony growth.
2.2.1.2 Polymerasechainreaction(PCR)
PCR was used to amplify genetically modified DNA-fragments, taking advantage of build in
restriction sites within the PCR-primer sequences. A standard PCR protocol was used in a
TPersonal thermocycler (Biometra, Göttingen, Germany). The protocol was adapted from
Saiki and colleagues using a thermostable DNA-dependent DNA polymerase130. Primers were
synthesized by Eurofins MWG Operon (Eurofins, Ebersberg, Germany) and the components
for the PCR reaction were purchased at New England Biolabs (NEB, Schwalbach, Germany).
Annealing temperature and elongation time were adapted according to primer characteristics
and to the particular prduct size, respectively. PCR products were separated by agarose gel
electrophoresis. These separated DNA fragments were excised and purified by the use of the
gel extraction kit according to the manufacturer’s recommendations (Qiagen, Hilden,
Germany).
26
2 Material and Methods
Component
PCR template
Primer
Formamide
Volume
[µl]
Concentration
50 ng/µl
0.5
100 pmol/µl
0.5
> 99.5 %
3.5
MgSO4
100 mM
2-4
dNTPsa
10 mM
5
5
10 x Thermopol
bufferb
Vent polymerase
2 U/µl
1 µl
ddH2O
ad 50 µl
a
Mixture of dNTPs containing dCTP, dATP, dGTP and dTTP
b
200 mM Tris-HCl pH 8.8, 100 mM (NH4)2SO4, 100 mM KCl, 20 mM MgSO4, 1 %
Triton X-100
Standard PCR thermocyler protocol
Program
Temperature [°C]
Time [min]
Initial denaturation
95
5
Denaturation
95
0.5
Annealing
55
1
Elongation
68
1 min per 1000 bp
Final elongation
68
10
Storage
4
Until further use
Standard PCR experiments were performed with 30 cycles of denaturation, annealing, and
elongation.
27
2 Material and Methods
2.2.1.3 Site‐directedmutagenesis
The mutagenesis procedure was carried out according to the QuikChange protocol
(Strategene, La Jolla, USA). For primer design, an online browser software tool of Agilent
Technologies (Agilent, Santa Clara, USA) was used and resulting primers were ordered from
Eurofins MWG Operon (Eurofins Ebersberg, Germany). The QuikChange PCR reaction
(50 µl) contained 1x Pfu reaction buffer (Stratagene La Jolla, USA), 2.5 μM of each sense and
antisense primer, 0.2 mM dNTP mix, 50 ng template DNA and 2.5 U Pfu Turbo polymerase
(Stratagene La Jolla, USA).
In all QuikChange reactions, the following PCR program was used on a TPersonal
thermocycler (Biometra, Göttingen, Germany):
Site-directed mutagenesis PCR reaction
Reaction
Temperature [°C]
Time [min]
Initial Denaturation
95
5
Denaturation
95
0.5
Annealing
55
1
Elongation
68
1 min per 1000 bp
Final Elongation
68
10
Storage
4
Until further use
Denaturation, annealing and elongation were repeated in 16 cycles.
Annealing temperature and elongation time were adapted according to primer melting
temperature and the particular plasmid size, respectively. Finally, the reaction sample was
incubated with DpnI (10 U) at 37°C for 1-2 hours to remove methylated template DNA.
Subsequently, the nicked PCR product was transformed into chemically competent E. coli
DH5α cells. Success of site directed mutagenesis was verified by sequencing of the respective
plasmid DNA from single colonies.
28
2 Material and Methods
2.2.1.4 AmplificationofplasmidDNA
A single colony of E. coli strain DH5α was picked after transformation and used to inoculate
a 4 ml LB overnight culture at 37 °C. Appropriate antibiotics were supplemented according to
the selection markers of the used plasmids. Plasmid DNA was isolated according to the
manufacturer’s instruction using a QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany).
Concentration and purity of DNA was determined spectroscopically measuring absorbance at
260 nm and purity was verified by measuring the 260/280 nm ratio using a NanoDrop
ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, USA).
2.2.1.5 RestrictiondigestandligationofPCRproducts
DNA from PCR products or from plasmid DNA was incubated with restriction enzymes
(NEB, Schwalbach, Germany) for at least 3 hours at 37 °C. To prevent religation, plasmid
DNA was additionally incubated with 10 U of calf intestine phosphatase (NEB, Schwalbach,
Germany) at 37 °C for 1 hour.
Standard digestion protocol
Component
DNA
Concentration
~50 ng/µl
Volume [µl]
xxb
Restriction enzyme 1
10-20 U/µl
1
Restriction enzyme 2
10-20 U/µl
1
10 mg/µl
0.5
BSAb
10 x NEB buffer
5
ddH2O
ad 50 µl
a
If necessary, bovine serum albumine (BSA) was added
b
The volume was adjusted to the concentration of the stock solution
29
2 Material and Methods
After restriction digest, samples were either purified with a PCR purification kit (Qiagen,
Hilden, Germany) or by agarose gel electrophoresis (see. 2.2.1.8) Separated DNA-fragments
were then purified with a gel extraction kit (see 2.2.1.6; Qiagen, Hilden, Germany).
2.2.1.6 DNAextractionfromagarosegels
If necessary, DNA bands of interest were excised from the gel with a sterile razor blade and
the DNA was subsequently extracted using a QIAquick Gel Extraction Kit (Qiagen, Hilden,
Germany) according to the manufacturer’s instructions.
2.2.1.7 Ligationreaction
Ligation reactions were performed according to the manufacturer’s instruction using a
Rapid DNA ligation kit (Fermentas, St. Leon-Rot, Germany).
Standard ligation protocol
Component
Concentration
Amount
Vector
50-100 ng
Insert
100-300 ng
5 x Ligation buffer
T4 DNA ligase
4 µl
5 U/µl
ddH2O
1 µl
Ad 20 µl
2.2.1.8 Agarosegelelectrophoresis
Agarose gel electrophoresis was used for analysis of DNA molecules that were obtained after
restriction digest experiments or PCR reactions. Desired DNA molecules were purified in a
gel containing 1-2 % agarose (w/v). For this, agarose was heated in 1x TBE buffer and cast
before polymerization. To visualize the DNA under UV light irradiation, ethidium bromide
30
2 Material and Methods
was added prior to solidification to a final concentration of 0.5 μg/ml. Electrophoresis was
performed in 1x TBE buffer at 70-100 V for 1-2 hours. Samples were supplemented with
DNA loading dye. As size markers, 1-10 μl of a NEB ruler 1 kb DNA Ladder were used.
2.2.1.9 Sequencing
To verify successful cloning of the different plasmid constructs, each plasmid DNA was
sequenced at Eurofins/MWG (Ebersberg, Germany). Sequencing samples contained
20-200 ng/μl template DNA mixed with sequencing primer in a concentration of 10 μM. The
sequencing primers used for the respective constructs are listed in the appendix.
2.2.1.10 Cloningoffull‐lengthClp1
For generation of the expression construct pET28b-ceClp1, the coding sequence of ceClp1
was amplified by PCR from Caenorhabditis elegans cDNA (see. 2.2.1.2) According to the
descriptions for a standard PCR reaction, the coding sequence was amplified using the
primers ceClp1-pET28b-For and ceClp1-pET28b-Rev. The sequencing primers used for the
respective constructs are listed in the appendix (see 6.1.1). The sequence of the forward
primer contained an NcoI restriction site, whereas the reverse primer was designed with a
NotI restriction site (see 2.2.1.5). The PCR product was purified with a QIAquick PCR
Purification Kit and subsequently digested with the two restriction enzymes NcoI and NotI.
The pET28b vector was digested with identical restriction enzymes. After restriction digest,
both the linearized plasmid DNA and the PCR product were analyzed by agarose gel
electrophoresis (see 2.2.1.8). The separated DNA-fragments were extracted from the excised
gel slices using a QIAquick Gel Extraction Kit (see 2.2.1.6; Qiagen, Hilden, Germany). The
PCR product was cloned into the linearized pET28b vector using a Rapid DNA Ligation Kit
(Fermentas, St. Leon-Rot, Germany). The ligation leads to an open reading frame (ORF)
coding for an N-terminal hexa-histidine fusion. For plasmid propagation, plasmid constructs
were transformed into E. coli DH5α cells (see 2.2.1.1 and 2.2.1.4).
31
2 Material and Methods
2.2.1.11 CloningofC‐terminaltruncations
The expression construct pET28b-ceClp1ΔC104 was designed to express a C-terminally
truncated variant of ceClp1. A stop codon leading to a truncated ORF was introduced by
site-directed mutagenesis into the pET28b-ceClp1 construct using the QuikChange protocol
(see 2.2.1.3; Strategene, La Jolla, California, USA). Thereby, the encoding codon triplet of
residue 105 was mutated to a stop codon using the primer pair ceClp1ΔN104-pET28b-For and
ceClp1ΔN104-pET28b_Rev. The sequencing primers used for the respective constructs are
listed in the appendix (see 6.1.1).
The expression construct pET28-ceClp1ΔC310 was designed as a C-terminal
truncation of the full-length gene obtained from pET28b-ceClp1. A stop codon was
introduced into the coding sequence of ceClp1 by site directed mutagenesis using the
QuikChange protocol (Strategene, La Jolla, California, USA). QuikChange primers were
designed to replace the codon of residue 310 with a stop codon (ceClp1ΔC310-pET28b-For
and ceClp1ΔC310-pET28b-Rev). The sequencing primers used for the respective constructs
are listed in the appendix (see 6.1.1).
2.2.1.12 CloningofN‐terminaltruncations
The ORF of an N-terminally truncated ceClp1 variant was amplified from the vector
pET28b_ceClp1 using PCR primers designed with an NdeI and a NotI restriction site. Similar
to the description for the cloning of the pET28b-ceClp1-Clp1 vector, the truncated coding
sequence spanning from residue 107 to residue 428 was cloned into a pre-linearized pET21a
vector (pET21a-ceClp1ΔN107). This ligation leads to an ORF encoding a C-terminally hexahistidine fused protein tagged and was kingly provided from Bernhard Dichtl (Zürich,
Switzerland).
The expression construct pET21-ceClp1NC314 was designed as an N-terminal
truncation. Primers for amplification contained an NdeI (ceClp1ΔN314-pET28b-For) and a
NotI (ceClp1ΔN314-pET28b-Rev) restriction site. This PCR product was digested and cloned
into a pre-linearized pET21a vector. The ORF encodes for a recombinant protein of
ceClp1ΔN314 fused to C-terminal hexahistidine tag. The sequencing primers used for the
respective constructs are listed in the appendix (see 6.1.1).
32
2 Material and Methods
2.2.1.13 Cloningofsingleaminoacidmutationsbysitedirectedmutagenesis
The
expression
constructs
pET28b_Clp1_K127A,
pET28b_Clp1_K127R,
and
pET28b_Clp1_W233A were obtained by site directed mutagenesis based on full-length
ceClp1. The codon-triplet of residue 127 was replaced with a codon coding either for alanine
or arginine using the QuikChange protocol (Strategene, La Jolla, California, USA).
Furthermore, the codon-triplet of residue 233 was also replaced with a codon coding for
alanine. Primers for site directed mutagenesis were applied on the full-length expression
construct
pET28b_ceClp1
(pET28b_Clp1_K127A_for
and
pET28b_Clp1_K127Arev;
pET28b_Clp1_K127R_for and pET28b_Clp1_K127R_rev; pET28b_Clp1_W233A_for and
pET28b_Clp1_W233A_rev). The sequencing primers used for the respective constructs are
listed in the appendix (see 6.1.1).
2.2.2 Biochemistry
2.2.2.1 SDS‐polyacrylamidegelelectrophoresis
Proteins were analyzed according to their molecular weight by discontinuous sodium dodecyl
sulfate polyacrylamide gel electrophoresis (SDS-PAGE), initially described by Lämmli131.
For gel casting and electrophoresis, a Protean 4 system (Bio-Rad, Munich, Germany) and a
PowerPack 200 (BioRad, Munich, Germany) power source were used, respectively. Mini-gels
were casted in parallel, starting with the separation gel. For the separation gel, 7.3 ml ddH2O
was mixed with 7.5 ml separation gel buffer and 15 ml aqueous 30 % acrylamide-,
bisacrylamide (37.5:1, Roth, Karlsruhe, Germany) to obtain a 15 % gel. The polymerization
reaction was started by addition of 250 μL of 10 % (w/v) APS (Merck, Darmstadt, Germany)
and 25 μL TEMED (Serva, Heidelberg, Germany). To remove air bubbles, the gel was
overlaid with isopropanol. To prepare the stacking gel, which was added on top of the
separation gel, 8.9 ml ddH2O were mixed with 3.8 ml stacking gel buffer and 2.3 ml 30 %
acrylamide-, bisacrylamide (37.5:1). Polymerization was initiated by addition of 150 μl 10 %
(w/v) APS and 15 μl TEMED. Gels were stored for up to two weeks at 4 °C until further use.
Prior to loading, the protein samples were supplemented with 5x SDS PAGE sample buffer
and boiled at 95 °C for 5-10 minutes. Electrophoresis was performed in SDS PAGE running
buffer at 35 mA per gel. Gels were stained using Page Blue protein staining solution
(Fermentas St. Leon-Rot, Germany) and destained with ddH2O.
33
2 Material and Methods
2.2.2.2 Expressionofrecombinantproteins
Large scale protein expression was performed in E. coli BL21-CodonPlus(DE3)-RIL cells
harboring the respective expression plasmid constructs. LB medium (2 l) was supplemented
with appropriate antibiotics and inoculated with a 1:100 dilution of an overnight culture. Cells
were grown until they reached an OD600 of 1.0. Once cells were in their exponential growth
phase, cultures were cooled down to 18 °C and protein expression was induced by addition of
1 mM IPTG. Cells were then incubated at 18 °C overnight. After cell harvesting, the pellets
were resuspended in the appropriate buffers (see 2.1.3) and used either directly for protein
purification or shock frozen with liquid nitrogen and stored at -80 °C. The bacterial cell wall
was broken by pulsed-sonication for 20 minutes on ice and supernatants were cleared by
centrifugation.
2.2.2.3 Purificationoffull‐lengthceClp1
For purification of wild-type ceClp1 and variants obtained by site directed mutagenesis,
pellets were resuspended in lysis buffer 1. The full-length protein was expressed using
pET28b vector (pET28b_ceClp1, pET28b_ceClp1_K127A pET28b_ceClp1_K127R, and
pET28b_ceClp1W233A) and initially purified by metal ion chromatography using Ni2+-NTA
material (Qiagen, Hilden, Germany). The column (~ 1 ml column volume (CV)) was
equilibrated with lysis buffer 1 prior to applying the supernatant and subsequently washed
with 3 CV of lysis buffer 1. Bound proteins were eluted with 6 ml of lysis buffer 1 containing
additional 300 mM imidazole. To remove the N-terminally fused hexahistidine-tag, the
protein was cleaved with thrombin (Sigma-Aldrich, Deisenhofen, Germany). For this, eluted
fractions were dialysed against dialysis buffer overnight and incubated with 80 U of thrombin.
Cleaved ceClp1 protein was diluted with anion-exchange buffer until a final conductivity of
5 mS/cm was reached and loaded onto a MonoQ column (GE Healthcare, Freiburg,
Germany), equilibrated with anion-exchange buffer. Bound proteins were eluted over a linear
gradient of 30 CV to the anion-exchange buffer containing additional 600 mM NaCl.
Fractions containing pure ceClp1 were pooled and concentrated using Amicon-15 centrifugal
filter units (Millipore) with a molecular weight cut-off of 10 kDa. Finally, the protein solution
was applied onto a Superdex 75 (GE Healthcare, Freiburg, Germany) column equilibrated in
size exclusion buffer and eluted protein fractions were concentrated as described above. All
variants generated by site directed mutagenesis were purified accordingly. Protein batches
34
2 Material and Methods
used for crystallization experiments were purified similarly, except that DTE was used in a
concentration of 4 mM instead of 2 mM, and that it was replaced with 4 mM TCEP in the
final purification step.
2.2.2.4 PurificationoftruncationvariantsofceClp1
For purification of the N- and C-terminal truncation variants, pellets were resuspended in lysis
buffer
2.
The
truncation
variants
were
expressed
using
pET28b-ceClp1ΔC104,
pET21a-ceClp1ΔN107, pET28b-ceClp1ΔC310, pET21a-ceClp1ΔN314, and initially purified
by metal ion chromatography using TALON (Clontech).
The untagged N-terminal truncated pET21a-ceClp1ΔN107 variant was co-expressed
with pET28b-ceClp1ΔC104 to improve solubility. Furthermore, pET28b-ceClp1ΔC310 also
was co-expressed with pET21a-ceClp1ΔN314. The supernatant containing the co-expressed
truncation variants ceClp1ΔN107 and ceClp1ΔC104 as well as ceClp1ΔC310 and
ceClp1ΔN314 was loaded on to a TALON Clontech (Saint-Germain-en-Laye, France) metal
affinity column (1 ml cv) equilibrated with buffer 2.
After a washing step with buffer 2 the untagged ceClp1ΔN107 was eluted from the
immobilized ceClp1ΔC104 in a high salt wash of buffer 2 containing additional 1 M NaCl.
The pooled wash fractions were concentrated using Amicon-15 centrifugal filter units with a
MWCO of 10 kDa and in a re-chromatography step using TALON Clontech (Saint-Germainen-Laye, France) resin, minor traces of ceClp1ΔC104 were removed. A buffer exchange to
buffer 2 was performed by concentration of the flow-through and re-dilution with buffer 2 in a
concentration step as described above. The N-terminal domain of ceClp1, the ceClp1ΔC104
variant was expressed alone and purified over a TALON Clontech (Saint-Germain-en-Laye,
France) metal affinity column (1 ml cv) equilibrated with buffer 2. After washing with 3 cv of
buffer 1, the bound protein was eluted with 6 cv of buffer 2 containing additional 150 mM
imidazole.
The supernatant of the co-expressed truncation variants ceClp1ΔC310 and
ceClp1ΔN314 was loaded on to a TALON Clontech (Saint-Germain-en-Laye, France) metal
affinity column (1 ml cv) equilibrated with buffer 2. After washing with 3 cv of buffer 2, the
bound proteins were eluted with 6 cv of buffer 2 containing additional 150 mM imidazole.
35
2 Material and Methods
2.2.2.5 Phosphorylationassay
5’-kinase activity of ceClp1 towards different RNA and DNA substrates was analyzed
using a phosphorylation assay. ceClp1 was tested either with a dinucleotide RNA substrate
guanylyl(3′→5′)cytidine (G1C2) or with RNA and DNA oligonucleotides (20 nt, see
Appendix 6.1.2). Each reaction was performed under limiting enzyme concentrations (1 µM)
in 1x phosphorylation buffer. In case of the G1C2, ceClp1 was incubated with 250 µM ATP
and 250 µM G1C2 for 2 hours and subsequently the reaction was analyzed by
chromatographic separation of the products ADP and 5’P- G1C2. In addition, G1C2 (500 µM)
was also used in a two fold excess relative to ATP to prove that ceClp1 catalysis the
phosphorylation of G1C2 in a one-to-one molar ratio. Phosphorylation was determined by
chromatographic purification (see 2.2.2.6).
For the phosphorylation assay with oligonucleotides, ceClp1 was incubated with
1 mM ATP and three different RNA and DNA substrates. Oligonucleotides were provided
either as single-stranded, double-stranded (blunt-end), or double-stranded (3’-overhang)
substrates. To produce a double-stranded substrate, two complementary strands were mixed in
equimolar amounts (0.1 mM). For annealing, nucleotide solution was heated to 95°C for
5 minutes and gradually cooled to room temperature. The resulting double-stranded substrate
was stored at -20°C until further use. Phosphorylation reactions were incubated at 30 ºC and
quenched upon addition of 8 M urea. Control experiments were performed with T4 PNK
(20 U) in an identical experimental setup. Phosphorylation efficiency was determined by
denaturing PAGE (see 2.2.2.9).
2.2.2.6 Chromatographicpurificationofthephosphorylationproducts
This approach is designed for baseline separation of reaction products based on changes of the
ionic net charge after phosphoryl transfer reaction. After the incubation time of 2 hours, each
reaction (200 µl) was diluted 1:2 with ddH2O, filtered, and finally applied to a MonoQ
column (5/50, GE Healthcare, Freiburg, Germany) equilibrated with ddH2O at 4 °C. The
baseline separation was achieved using a linear gradient to 80 mM NaCl. Under these
conditions, G1C2 eluted at 5.8 ml, 5’P-G1C2 at 8.7 ml, ADP at 9.2 ml, and ATP at 10.8 ml.
During chromatographic separation, absorbance at 260 nm and 280 nm and conductivity
(mS/cm) were monitored.
36
2 Material and Methods
2.2.2.7 Quantitativeproductionof5’P‐G1C2
5’P-G1C2, which is used for a single turnover reverse reaction, was produced enzymatically
using ceClp1 (see 2.2.2.8). The reaction sample containing 2 μM of recombinant ceClp1
supplemented with 500 µM of G1C2 and 500 µM of ATP in phosphorylation buffer was
incubated at 25 °C for 8 hours. After incubation the reaction mixture was diluted 1:2 with
ddH2O and loaded onto a MonoQ anion exchange column (10/100, GE Healthcare, Freiburg,
Germany) equilibrated with ddH2O. G1C2 was baseline separated from the phosphorylated
5’P-G1C2 molecule using a linear gradient of 80 mM NaCl. To prevent contamination from
residual traces of ADP, an ATP recovery system was used containing pyruvat kinase (PK)
and 500 µM PEP. Eluted fractions containing 5’P-G1C2 were collected and concentrated using
a vacuum concentrator. 5’P-G1C2 was resuspended in ddH2O and residual salt was removed
by gel filtration chromatography. The sample (500 µl) was applied onto a Superdex 75
column (5/50, GE Healthcare, Freiburg, Germany) and equilibrated with ddH2O. Desalted
fractions of 5’P-G1C2 were concentrated as described above by a vacuum concentrator and
stored at -20 °C.
2.2.2.8 Reversereaction
The reverse reaction was analyzed chromatographically under single turnover conditions
(see 2.2.2.6). The reaction sample contained 60 µM of recombinant ceClp1 supplemented
with 50 µM of 5’P-G1C2 and 50 µM of ADP in the phosphorylation buffer incubated for 20
and 40 minutes. Identical to the analysis of the forward reaction, the sample was applied to a
MonoQ column (GE Healthcare, Freiburg, Germany). To inhibit enzyme activity of possible
adenylate kinase contaminations, all samples of single turnover reactions were treated with
100 µM of diadenosine pentaphosphate (Ap5A). Ap5A is known as a potent inhibitor of
adenylate kinases functioning as a bisubstrate inhibitor132. Based on results obtained from the
forward reaction, it was shown that ceClp1 5’-kinase activity is not inhibited by Ap5A.
Additional control experiments were performed to show that ceClp1 protein batches do not
posses any basal adenylate kinase activity on their own. ceClp1 was incubated with 50 µM of
ADP. The forward reaction was also performed under single turnover conditions. The reaction
consisted of 60 µM of ceClp1 supplement with 100 µM of Ap5A, 50 µM of G1C2 and 50 µM
of ATP incubated for 20 minutes.
37
2 Material and Methods
2.2.2.9 Denaturingpolyacrylamidegelelectrophoresis(Urea‐PAGE)
Phosphorylation efficiency of ceClp1 towards single-stranded and double-stranded RNA and
DNA oligonucleotides (20 nt) was analyzed by denaturing Urea-PAGE. For this, 9 % (v/v)
polyacrylamide gels (19:1 acrylamide to bisacrylamide ratio) containing 8.3 M urea were
used. Prior to electrophoresis, RNA samples were denatured with urea (4 M) and RNA and
supplemented with sample buffer. Gels were cast in glass plates (40 cm x 33 cm) with 1 mm
spacers and were run in 0.5x TBE buffer. Polymerization of the polyacrylamide gel solution
was initiated by addition of 0.08 % (w/v) APS and 0.08 % (v/v) TEMED. To remove excess of
APS and TEMED after polymerization, a pre-run step was applied at 25 W for 1 hour. After
loading, the gel was run at 25 W until G1C2 and 5’P-G1C2 were separated, based on their
different electrophoretic mobility. RNA and DNA oligonucleotides were visualized by
methylene blue staining.
2.2.2.10 Steady‐statekinaseassay
A coupled colorimetric assay was used to determine consumption of ATP during RNA
phosphorylation by ceClp1. This assay enables quantification of ceClp1’s kinase activity.
ATP consumption of Clp1 is coupled to the oxidation of NADH using a recovery system. The
recovery system consists of two components. The first is PK that converts PEP and ADP to
ATP and pyruvate, thus responsible for ATP-regeneration. The second is lactate
dehydrogenase (LDH) that uses pyruvate to produce lactate by using NADH as reducing
agent. This decrease in the amount of NADH was followed by measuring absorbance change
at 340 nm and allowed spectroscopical quantification of Clp1’s ATP hydrolysis rates. All
experiments were performed at constant ATP concentrations of 1 mM. To achieve this, 0.2
µM of ceClp1 was incubated in ATPase buffer supplemented with 7 U/ml pyruvate kinase
(Sigma-Adrich, Deisenhofen, Germany) and 10 U/ml lactate dehydrogenase (Sigma-Adrich,
Deisenhofen, Germany). Prior to the kinetic measurements, the reaction mixture was
incubated at room temperature for 20 minutes to obtain a stable baseline and subsequently the
reaction was started by addition of ceClp1. To determine the Michaelis-Menten kinetics for
the RNA dinucleotide G1C2 and urylyl-(3’-5’)guanosine (U1G2), varying RNA concentration
were used.
38
2 Material and Methods
Changes in absorbance at 340 nm were monitored in a 1 cm quarz cuvette (Hellma,
Mühlheim, Germany) using a JASCO V-650 spectrophotometer. The initial rates of the
reaction were calculated using equation 1.1.
v
Abs340 / t
N NADH  d
[equation 1.1]
ΔAbs340:
Difference of absorption at 340 nm
Δt:
Time-difference (min or s)
εNADH:
Absorption coefficient of NADH (6220 M-1 cm-1)
cp :
Protein concentration (M)
d:
Path length of cuvette (cm)
Initial rates were plotted against the substrate concentration and fitted using the standard
Michaelis-Menten using equation 1.2:
v
vmax  S
S  Km
[equation 1.2]
S:
Substrate concentration
vmax:
Maximal apparent reaction velocity
Km:
Apparent Michaelis-Menten constant
Turnover numbers were calculated according to the following equation using
equation 1.13:
K cat 
cp:
v
cp
[equation 1.3]
Protein concentration (M)
39
2 Material and Methods
2.2.3 BiophysicalMethods
2.2.3.1 CDspectroscopyofClp1mutantvariants
Protein stability of wild type ceClp1 and mutant variants was compared using CD
spectroscopy. Melting curves were determined using change in ellipticity at 222 nm from
10°C to 60°C with a heat-rate of 0.5 °C/min. Protein samples (15 µM) in size exclusion buffer
were analyzed using a Jasco J-810 spectropolarimeter (Jasco, Groß-Umstadt, Germany) with
1 mm CD cuvettes (Hellma, Müllheim, Germany).
2.2.4 CrystallographicMethods
To identify suitable crystallization conditions, protein solution was pre-mixed with nucleotide
ligands and applied to a variety of crystallization screens in a 96-well plate sitting drop vapor
diffusion setup with the Mosquito nanolitre pipetter (TTP LabTech LTD, Melbourn, UK).
Manual optimization in Linbro plates using the hanging drop vapor diffusion method was
unsuccessful. For this, crystallization conditions were optimized in 96-well plates. Prior to
flash cooling in liquid nitrogen, crystals were transferred with a cryo-loop into a
cryo-protectant solution.
2.2.4.1 Crystallizationandcryo‐protection
Initial crystallization trails were performed with full-length ceClp1 (450 μM) crystallized in
the presence of 1.5 mM adenyl-5’-(,-imido)triphosphate (AppNHp, Jena Bioscience, Jena,
Germany) and 1.5 mM G1C2 or ADP and 1.5 mM G1C2 or 1.5 mM of ATP in a three-fold
excess compared to the protein concentration. Crystals were obtained at 20 °C using the
sitting drop vapor diffusion method with a reservoir solution composed of 100 mM
Na2HPO4/KH2PO4 between pH 5.0 and 6.0, 200 mM NaCl, 15 mM MgCl2, 90 mM sarcosine
and 25 % (w/v) polyethylene glycol (PEG) 1000. All crystals that were obtained under the
described condition were suitable for soaking experiments. To ensure full occupancy of the
soaked ligands, ligand concentration was increased to 10 mM in the cryo-protectant. An
artificial apo structure of ceClp1 was obtained by soaking of a ternary complex
(ceClp1•ATP•Mg2+) with mother-liquor containing alkaline phosphatase solution. The
40
2 Material and Methods
inhibited substrate bound state was obtained after incubation of ceClp1 with 10 mM G1C2 or
U1G2 and 10 mM AppNHp (ceClp1-AppNHp-G1C2 and ceClp1-AppNHp-U1G2). The
transition state analog (ceClp1-ADP-AlF4--G1C2) was obtained after soaking with 10 mM
ADP and G1C2 dinucleotide and aluminum fluoride. AlFX was generated in situ by addition of
20 mM NaF and 5 mM AlCl3. In order to obtain crystals bound to ATP (ceClp1-ATP),
crystals grown in the presence of ADP and subsequently incubated with 10 mM ATP.
Residual traces of ADP were removed upon addition of a recycling system that converts ADP
to ATP (PK and PEP). Crystals bound to the G1A2A3A4 tetra-nucleotide (ceClp1 RNA 4mer)
were obtained by soaking with 10 mM ADP and 5 mM tetra-nucleotide. Finally, crystals
bound to their educts were obtained by incubation with 10 mM ATP and G1C2 dinucleotide
(ceClp1-ADP). The crystallized protein was still enzymatically active. However, those
crystals appeared to have slightly suffered from incubation (max. resolution of 2.6 Å) but
were virtually identical to crystals soaked with 10 mM ADP and 100 mM MgCl2 (max.
resolution of 2.1 Å). For flash cooling in liquid nitrogen, crystals were cryo-protected by
addition of 5 % (v/v) PEG 600 and 2.5 % (w/v) sucrose to their respective soaking solutions.
In order to identify the position of the catalytic divalent ion by, Mg2+ was replaced by Mn2+ in
the ATP bound structure.
2.2.4.2 Datacollection
Diffraction data were collected at 100 K either at the synchrotron beamline X10SA at the
Swiss Light Source (SLS) in Villingen, Switzerland or in-house at the Max Planck Institute
for Medical Research. All data sets were processed with the XDS package
133
. Diffraction
data were indexed and integrated with XDS and scaled and merged with XSCALE.
XDSCONVERT was used to convert files into a format compatible with the programs used
for phasing, model building, and refinement. For calculation of a free R-factor, 5 % of
reflections from diffraction data sets were randomly assigned and excluded from refinement.
2.2.4.3 Structuredetermination
Experimental phases for ceClp1 were obtained from a single anomalous diffraction
experiment on selenomethionine-labelled protein crystals and experimental phases were
obtained using SHELX134. An initial model for ceClp1 was built manually using COOT135
and refined with CNS136. Phase extension to a high-resolution native ceClp1 data set was
41
2 Material and Methods
performed following a rigid-body refinement protocol in CNS136. In later stages of ceClp1
structure refinement, iterative cycles of refinement using REFMAC137 including TLS
refinement138 were performed, followed by manual model improvement using COOT135.
Water molecules were assigned manually. Phases for different nucleotide-bound forms of
ceClp1 were obtained by rigid-body refinement137 with the refined structure of
selenomethionine-labelled ceClp1, followed by cycles of manual model improvement and
refinement using REFMAC137 including TLS refinement138. Intermediate and final structures
were evaluated with MOLPROBITY139 and PROCHECK140. All figures were drawn using
PYMOL141. Surface-potentials have been calculated using the APBS plug-in application142.
2.2.5 Bioinformaticand‐computationalmethods
2.2.5.1 Sequencealignment
Multiple
sequence
alignments
were
generated
with
ClustalW
software
(http://www.ebi.ac.uk/Tools/msa/clustalw2) and secondary protein structure predictions were
obtained by the PSIPRED tool (http://bioinf.cs.ucl.ac.uk/psipred/). Sequence alignments were
illustrated using the program ALSCRIPT143 and the sequence conservation was determined by
the AMAS-server144.
2.2.5.2 StructuralmodelingofthePcf11interactioninterface
The interaction interface between ceClp1 and cePcf11 was deduced by superposition of the
previous provided crystal structure of scClp1 bound to scPcf1194 (PDB 2NPI) and the crystal
structure of ceClp1 (ceClp1-AppNHp-G1C2). Structures of both orthologs were superimposed
by secondary-structure matching and the peptide representing scPcf11 was mutated according
to a sequence alignment between scPcf11 and cePcf11. Side-chain geometry of cePcf11 was
optimized to avoid steric clashes with the structure model of ceClp1. All modeling steps were
performed by COOT software135. The superposition of individual polypeptide chains was
performed using LSQKAB145.
42
3 Results
3 Results
3.1
StructuralcharacterizationofceClp1
This chapter presents crystallization and subsequent structure determination of Clp1 by X-ray
crystallography, providing the first three-dimensional structures of a eukaryotic, RPNK.
ceClp1 was crystallized in complex with various RNA substrates that differ in length and
sequence. The molecular basis of RNA-recognition and RNA-specificity were investigated
using these RNA bound crystal structures. In addition, Clp1 was crystallized at important
states of its catalytic cycle providing new insights into the phosphoryl transfer reaction
mechanism of PNKs.
3.1.1
Bioinformatical characterization of the Clp1 protein family
In contrast to enzymatically inactive scClp194, the human ortholog functions as a 5’-PNK88.
Furthermore, hsClp1 is the first identified eukaryotic PNK that is specific for single stranded
RNA rather than DNA substrates85,92. Unfortunately, hsClp1 is insoluble (Bernhard Loll,
personal communication) and thus unsuitable for crystallization experiments. This
necessitated the search for other enzymatically active orthologs that could be used for
crystallization experiments. The amino acid sequences of Clp1 orthologs from a
representative selection of eukaryotic organisms were analyzed by multiple structure-based
sequence alignment (Figure 3.1.1). The aim of this bioinformatical characterization was to
identify conserved structural features that might be specific to enzymatically active variants of
Clp1.
Characteristic for the entire eukaryotic Clp1 protein family is the common P-loop
motif (phosphate-binding loop) and the Walker B motif113,114. The P-loop motif consists of a
flexible loop between strand β10 and helix α2 and shows the consensus sequence
GxxxxGK[S/T] (Figure 3.1.1). The distal Walker B motif consists of a conserved aspartate
residue located after strand β11. The P-loop and the Walker B are involved in substrate
binding and enzyme catalysis19,113. Inspection of multiple sequence alignment within higher
eukaryotes revealed two additional conserved motifs.
The first constitutes the LID module, a mostly helical structure located within α7,
while the second is a clasp-like binding loop (referred to as clasp) located after strand β14.
43
3 Results
The LID module contains the conserved RxxxxR sequence and was suggested to stabilize the
transition state by neutralizing emerging negative charges of the transferred phosphoryl group
as known for other P-loop kinases118. A comparison of orthologs from higher eukaryotes with
Clp1 from fungi revealed that both or at least one of the conserved arginine residues is
replaced by hydrophobic amino acids (Fig. 3.1.1). The second motif, the clasp, has the
consensus sequence TxGW and plays a major role in RNA-binding. In contrast to higher
eukaryotes, yeast orthologs showed no conservation of the clasp motif.
The low sequence conservation of Clp1 from fungi compared to higher eukaryotes
suggests that scClp1 has lost its enzymatic activity during evolution by distinct mutations.
ATP-binding by the P-loop motif, however, remained unaffected94,98-100. Orthologs from
higher eukaryotes, in contrast, show high sequence conserved within the four critical motifs
(Figure 3.1.1). Thus, it is most likely that orthologs from higher eukaryotes function in a
similar way as shown for hsClp1. Interestingly, ceClp1 carries a unique insertion in the loop
region connecting the NtD with the PNK domain. Since this insertion might have a positive
affect on Clp1’s protein solubility, the protein was purified and subsequently applied to
structure determination experiments. ceClp1 has the highest homology to Clp1 from Xenopus
laevis with a sequence identity of 78.7 %. However, it also shows a significant homology to
both, hsClp1 and Clp1 from Drosophila melanogaster, with a sequence identity of 48.0 % and
42.0 %, respectively.
44
3 Resultss
45
3 Results
Figure 3.1.1 Sequence
e compariso
on of Clp1 o
orthologs. Sequence
S
alig
gnment of C
Clp1 in both higher
h
eukaryotes (Caenorhab
bditis elegan
ns: NP_001
1040858; Ho
omo sapien
ns: NP_0011136069; Xen
nopus
laevis: NP_
_001084787, and Drosop
phila melanog
gaster: NP_610876) as well
w as fung i (Saccharom
myces
cerevisiae: NP_014893, Schizosaccharomycces pombe: NP_59374
41, and C
Candida alb
bicans:
XP_712695
5). The seco
ondary struc
cture elemen
nts are sho
own above the
t
alignmennt, with α-h
helices
depicted ass cylinders an
nd β-strands as arrows, a
assigned acc
cording to the
e ceClp1 cryystal structure
e. The
coloring of the seconda
ary structure correspond
ds to the dom
main color frrom Figure 33.1.4 (NtD: green,
g
PNK domain: cyan, and CtD: red). The
T P-loop m
motif, the Wallker B motif, the LID moddule, and the clasp
are highligh
hted by red boxes.
b
Residues involved
d in the switc
ch on/off of Lys127
L
are m
marked with a gray
triangle, wh
hereas residu
ues, which in
nteract with tthe RNA, are
e marked witth a yellow ccircle. Cataly
ytically
important rresidues are
e indicated by
b black tria
angles. The
e Walker A threonine iss involved in
i the
coordination
n of the mettal cofactor (purple circle
e). Residues subjected to
o site-directeed mutagene
esis in
ceClp1 are
e marked with
w
a black circle below
ow the align
nment. Cons
served residdues are co
olored,
according to
o the degree
e of conserva
ation decreassing from darrk green to yellow.
3.1.2 Purrification of
o ceClp1
To expresss and puriffy recombin
nant ceClp 1, the cDN
NA sequencce was clonned via stan
ndard
methods innto the baacterial exp
pression vecctor pET28
8b. The seequence-verrified expreession
plasmid w
was transforrmed into E. coli BL
L21-CodonP
Plus(DE3)-R
RIL cells aallowing fo
or an
IPTG-induucible overeexpression. The proteinn constructt was designed to encoode for a fusion
f
variant of cceClp1 withh an N-term
minal hexahiistidine tag. The resulting recombiinant protein
n was
overexpresssed, cells were
w
lysed using
u
soniccation, and the protein was then ppurified from
m the
supernatannt. Overexppressed ceC
Clp1 was fo
found in the soluble fraction
f
of E. coli ly
ysates.
2
Further puurification included th
hree steps using Ni2+
affinity, anion excchange and
d size
exclusion chromatogrraphy. The protein sollution was applied
a
to Ni
N 2+-resin. After a waashing
e
with imidazole. The N-term
minal hexahiistidine tag was removed by
step, the prrotein was eluted
a proteolyttic digest ussing thromb
bin and the wild type ceClp1
c
prottein was fur
urther purified by
anion excchange (MoonoQ) and
d size excllusion chro
omatograph
hy. A typiical purificcation
procedure yielded thrree milligraams of prootein per litter of cell culture. Hoomogeneity
y was
46
3 Resultss
veriffied by SDS-PAGE off peak fracctions, show
wing the prresence of a single prrotein bandd
migrrating to a molecular
m
weight
w
of aroound 54 kD
Da, which iss in agreemeent with thee calculatedd
masss (Figure 3.22.1).
Figure 3.1.2 Ex
xpression and
a
purificattion of ceCllp1 in E. co
oli. Aliquot off
2
ceC
Clp1 obtaine
ed after Ni2+
-NTA, anio
on exchangge, and siz
ze exclusion
n
chro
omatographyy were analy
yzed by SDS
S-PAGE (15 %). The gel was stained
d
with
h Coomassie
e Blue confirrming purifica
ation of ceCllp1 to high homogeneity..
M: low molecula
ar weight ma
arker.
3.1.33
Co‐crysstallization
nofceClp1
1
Durinng the lastt years, sig
gnificant prrogress hass been mad
de in analyyzing the biochemical
b
l
propeerties of hsClp188,92. Furthermore
F
e, the crystaal structure of inactive scClp1 was publishedd
in 22001 at medium
m
ressolution94. Nevertheless, the molecular m
mechanism of Clp1’ss
RNA
A-specificityy and the understandin
u
ng of the phosphoryl
p
transfer reeaction on a structurall
levell remained elusive. This thesis, therefore, aimed for crystallizattion of ceC
Clp1 in thee
ligannd-free form
m and in com
mplex with its ligands. ceClp1 was purified (ssee 3.1.2) and
a used forr
initiaal crystallizzation experriments at a concentraation of 650 µM. Cryystallization
n trials of a
ligannd-free ceCllp1 remaineed unsuccesssful, possib
bly due to domain flexxibility of apo-ceClp1
a
preveenting crysttal formatio
on. Howeveer, co-crystaallization off 450 µM cceClp1 togeether with a
non-hhydrolysablle ATP an
nalog (ApppNHp) and an RNA dinucleotidde (G1C2), both in a
threee-fold excesss (~1.4 µM
M) relative tto protein, finally
f
led to
t spontaneeous crystall formation..
Screeening was performed
p
in a sittingg drop vapo
or diffusion setup in a 96-well pllate format..
Initiaal crystals of ceClp1, co-crystalllized with
h AppNHp and G1C2, were obttained afterr
12-48 hours in a buffer con
ntaining 20 % (w/v) PE
EG 1000, 10
00 µM Na/K
K phosphatee buffer pH
H
6.2 aand 200 µM
M NaCl (Figu
ure 3.1.3 A)). These con
nditions weere further ooptimized by
y screeningg
with varying PEG
P
1000 concentratio
c
ons (20-25 % (w/v)) in
i a pH raange from 5.5 to 7.5..
Crystallization was
w further improved uusing an ad
dditive screeen. The addditive screen
n identifiedd
two components that supported crystaal growth to
owards sing
gle crystal m
morphology
y. The firstt
was MgCl2 at a concentraation of 15 mM, the second
s
wass sarcosine at a conceentration off
mM. The opttimization procedure
p
fi
finally led to
o needle-shaped singlee crystals off a maximall
90 m
size oof about 1000-200 µm growing
g
in a precipitan
nt solution containing
c
222 % (w/v) PEG 1000,,
47
3 Results
100 mM Na/K phosphate buffer pH 6-7, 200 mM NaCl, 15 mM MgCl2, and 90 mM sarcosine
(Figure 3.1.3 B). Crystal growth was achieved only with a 96-well plate, while up-scaling into
a 24-well plate format (Linbro plates) remained unsuccessful. During crystal harvesting, a
cryo-protection solution was used preventing ice formation during flash-cooling in liquid
nitrogen. This cryo-protectant consisted of mother-liquor with additional 5 % (v/v) PEG 600
and 2.5 % (w/v) sucrose. The cryo-protectant solution was supplemented with AppNHp and
G1C2 at the same concentration as used for co-crystallization in order to avoid substrate
diffusion from the active site of the protein crystals. Based on diffraction data obtained from
the crystals, an electron density map was generated showing both a bound AppNHp molecule
and a Mg2+ ion in the active site (ceClp1-AppNHp-Mg2+), whereas no electron density for the
RNA molecule was observed.
The crystal structure of ceClp1-AppNHp-Mg2+ revealed that crystallization of ceClp1
depends on the presence of a nucleoside triphosphate, whereas presence of G1C2 during
crystallization experiments was not required. Thus, ceClp1 was also co-crystallized with a
three-fold excess of ATP (1.3 µM) or ADP (1.3 µM) using the same precipitant conditions as
described for growth of ceClp1-AppNHp-Mg2+ crystals. In both cases crystals started to grow
after one day. However, no electron density for the Mg2+ ion was observable for the ceClp1ADP complex. Increasing the MgCl2 concentration in the cryo-protectant up to 100 mM
finally enabled Mg2+ incorporation, leading to a ternary complex of ceClp1 (ceClp1-ADP).
Since hsClp1 was shown to use different divalent metal cofactors85, Mg2+ and Mn2+ were
tested for crystallization experiments. Crystals co-crystallized with ATP were transferred to a
cryo-protectant solution in which MgCl2 was replaced with 15 mM of MnCl2 (ceClp1-ATP).
Furthermore, in these experiments a ATP recycling system was added to the cryo-protectant
to remove residual traces of ADP arising from spontaneous ATP hydrolysis. This recycling
system consisted of PK that converts ADP to ATP using PEP as substrate. The crystals were
transferred to the cryo-protectant solution and incubated for 30 minutes. In contrast to soaking
experiments, in which no recycling-system was added, a mixture of ADP/ATP was observed
in the electron density map. To identify ceClp1’s in vivo substrate, attempts to crystallize
ceClp1 with double stranded DNA duplexes or with an RNA anti-codon loop analog were
made, but co-crystallization was unsuccessful.
48
3 Resultss
Figurre 3.1.3 Cry
ystals of ceC
Clp1. A, Inte
ergrown crys
stals appear after one daay by using a precipitantt
solutiion of 20 % (w/v) PEG
G 1000, 100
0 mM Na/K phosphate buffer pH 66.2, 200 mM NaCl. B,,
Crysttallization conditions were
e optimized to 22 % (w/v
v) PEG 1000, 100 mM Naa/K phospha
ate buffer pH
H
6-7, 2
200 mM NaC
Cl, 15 mM MgCl
M 2 and 90
0 mM sarcos
sine, resulting in growth of single needle-shaped
d
crysta
als.
To obtain a crystal structure
s
off apo-ceClp
p1, crystals of
o the ceClpp1-ADP com
mplex weree
soakeed in mother-liquor supplement
nted with alkaline
a
ph
hosphatase. After incu
ubation forr
2 houurs, the crysstals were harvested
h
annd transferreed into cryo
o-protected ssolution (ap
po-ceClp1).
To charaacterize ceC
Clp1’s RNA
A-binding site,
s
the cry
ystal structuure of an RNA
R
boundd
form
m was required. Whereeas co-crysttallization was
w unsuccessful, soakking experiments withh
ceClpp1 crystals resulted in crystals witth bound RN
NA substrates. Crystalls were transferred intoo
a cryyo-protectant containiing a ~25--fold excess of AppN
NHp (10 m
mM) and of
o an RNA
A
dinuccleotides (G
G1C2 and U1G2; 10 mM
M) relative to the prottein concenntration, leaading to thee
crysttal structurre of ceCllp1-AppNH
Hp-G1C2 an
nd ceClp1-AppNHp-U
U1G2. To ensure
e
fulll
occuupancy of the
t
ligand-binding sitte, these crystals werre incubateed for 2 ho
ours in thee
cryo--protectant solution and
d subsequenntly harvestted for flash
h-cooling.
Based on soaking experimentts, it was possible
p
to capture the
he first tran
nsition statee
analoog of a PNK
K. To obtain this trans ition state analog,
a
crysstals of ceC
Clp1-ADP were
w
soakedd
in a cryo-protecctant solution containiing a ~25-ffold excess of ADP ((10 mM) an
nd of G1C2
(10 m
mM) relativve to the protein
p
conccentration, 100 mM of MgCl2 annd aluminu
um fluoridee
(AlFx) as transition state analog. Thhis compou
und is geneerally usedd as a transition statee
analoog121 and was
w generatted in situ by additio
on of 20 mM NaF wiith 5 mM AlCl
A 3. Thiss
resullted in a teetrafluoroaluminate (A
AlF-4) transiition state analog. Sinnce it was previouslyy
show
wn that the coordination
c
n number oof aluminum
m depends on
o the pH att, which thee enzyme iss
crysttallized (pH
H 4.5 leads to AlF-4 andd pH 8.5 lead
ds to trifluo
oroaluminatte (AlF3))1466, the bufferr
systeem of the cryo-protectant was chaanged from
m Na/K phosphate bufffer to a HE
EPES bufferr
systeem (22 % (w
w/v) PEG 1000, 50 mM
M HEPES-N
NaOH pH 7,
7 200 mM NaCl, 15 mM
m MgCl2,
49
3 Results
90 mM sarcosine). Furthermore, soaking experiments were performed with pH ranging from
8 to 9, in order to induce formation of an AlF3 transition state analog. However, at high pH it
was impossible to trap a mimic and no AlF3 was bound.
Based on the electron density map, it is not possible to distinguish between an AlF-4
and a tetrafluoromagnesate (MgF42-) transition state analog147. Since both analogs would have
been possible, two control soaking experiments were performed without AlCl3. Both
experiments differed in the type of the divalent metal cofactor. The 100 mM MgCl2 of the
cryo-protectant solution of the first experiment were replaced with 100 mM of MnCl2 for the
second experiment. Capturing of the transition state analog without AlCl3 would indicate the
existence of an MgF42- or tetrafluoromanganate (MnF4) transition state analog in the active
site of ceClp1. Furthermore, a transition state analog formed by MnF4 could easily be
identified by an anomalous scattering signal of the manganese atom. However, no electron
density for a transition state analog for MgF42- and MnF4 was obtained, suggests that indeed
AlF4- was bound. Additionally, another trial was performed aiming for capturing the transition
state analog using G1A2A3A4. Previous observations with the transition state mimic suggested
full occupancy of the RNA-binding site at much lower concentrations during a transition state
as compared to a substrate bound state. The cryo-protectant was, therefore, supplemented with
5 mM of G1A2A3A4. Analysis of crystals soaked under these conditions showed no transition
state analog, instead those crystals resemble a product bound state (ceClp1-4mer).
Additionally, soaking experiments showed that ceClp1 remains enzymatically active
within the crystal. To obtain crystals captured with their educts, cryo-protectant was
supplemented with a ~25 fold excess of ATP and G1C2. Analysis of the diffraction data from
those crystals showed reaction product formation (ADP) in the electron density map, whereas
phosphorylated RNA was missing. These structures were virtually identical to the
ceClp1-ADP crystal. Since turnover crystals appeared to have slightly suffered from
incubation (max. resolution of 2.6 Å), the ceClp1-ADP structure (max. resolution of 2.1 Å) is
referred to as RNA released product bound state.
All diffraction data were collected either at the PXII beamline at the SLS in Villigen,
Switzerland, or in-house using a Rigaku MicroMax 007 HF microfocus X-ray generator.
Space group, unit cell parameters, and the overall Wilson B-factor are given in Table 3.1.1.
All crystals belonged to a trigonal crystal symmetry system and the space group was P31.
50
3 Results
3.1.4 De novo phasing of ceClp1 by single-wavelength anomalous diffraction
experiments
Experimental phases of ceClp1 were obtained by de novo phasing using a single-wavelength
anomalous diffraction (SAD) experiment148. To obtain selenomethionine labeled protein,
Clp1 was overexpressed similar to native protein, except that cells were grown in minimal
medium, in which the amino acid methionine was substituted with its seleno-derivate127.
Protein with incorporated selenomethionine was purified (see 2.2.2.3) and crystallized similar
to the native protein. Precipitant conditions were similar to the description for native protein
(22 % (w/v) PEG 1000, 100 mM Na/K phosphate buffer pH 6-7, 200 mM NaCl, 15 mM
MgCl2, and 90 mM sarcosine). Flash-cooling was achieved with a cryo-protectant containing
the mother liquor supplemented with 5 % (v/v) PEG 600, 2.5 % (w/v) sucrose, 1.3 mM
AppNHp, and 1.3 mM G1C2 (ceClp1-SeMet). A highly redundant data set from
selenomethionine crystals that diffracted up to 2.3 Å resolution was collected at the
absorption edge of selenium. Collected diffraction images were indexed, integrated, and
scaled using the XDS package133.
The positions of the selenium atoms were localized and native phases were determined
using SHELXD149 and SHELXE134, respectively. The SHELX programs are embedded in the
graphical user interface HKL2MAP that facilitates usage150. SHELXD identified 11 selenium
atom positions with a correlation coefficient of 31.2 % (all) and 17.9 % (weak) as the best
solution. The pseudo-free correlation coefficient, which is calculated by randomly leaving out
10 % of the reflections, was 54 %.
Similar to the observations made for native crystals, the SeMet-Clp1 crystals belong to
the trigonal crystal system and have the space group P31. The crystals contain one molecule
per asymmetric unit. Seven selenomethionine positions could unambiguously be identified
and were used to calculate an initial electron density map. An initial model for ceClp1 was
manually built using COOT135 and refined with CNS136. In later stages of ceClp1 structure
refinement, iterative cycles of refinement were performed using REFMAC137 including TLS
refinement138 with manual model improvement using COOT135. They led to a final Rwork and
an Rfree of 19.2 % and 24.0 %, respectively, at a resolution of 2.3 Å.
51
3 Results
3.1.5
Phasing and refinement of the different ceClp1 complexes
Diffraction data of the different nucleotide free-/bound-forms of ceClp1 were processed with
the XDS package133 and phases were obtained by rigid-body refinement137 using the refined
model of selenomethionine-labelled ceClp1 as a starting model. Once phases were calculated,
models were manually improved by COOT135 and subsequently refined using REFMAC137
with restrained maximum likelihood refinement including TLS refinement138. Water
molecules were assigned manually. Data collection and refinement statistics are summarized
in Table 3.1.1.
3.1.6 Crystal structure of apo-ceClp1
The overall architecture of ceClp1 can be subdivided into three distinct domains, a central
PNK domain flanked by additional N- and C-terminal domains (NtD and CtD, Figure 3.1.4 A
and B). Diffraction data of the apo-structure had a maximum resolution of 2.1 Å. The
structure model was refined to an R-factor of 18.8 % with an Rfree of 23.2 % (Table 3.1.1;
ceClp1-apo). Based on the electron density map, a continuous polypeptide chain starting at
residue 4 and ending at residue 425 could be modeled. Only two unassigned loop regions
from residue 105 to residue 111 and from residue 331 to residue 349 could not be modeled
due to ambiguous electron density. Furthermore, the electron density map revealed the
presence of two inorganic phosphate molecules (Pi). One Pi is bound to the P-loop motif,
positioned at the expected location of β-phosphate of an ATP molecule113. The other Pi is
bound to an arginine residue (Arg406) at the molecular surface of the protein. There was also
additional density observed for a PEG molecule, which was at a van der Waals distance to a
hydrophobic patch on the surface of ceClp1 (composed of residues Trp62, Leu88, His91,
Try137, and Arg140).
The PNK domain is the central domain of the ceClp1 structure, spanning from residue 110 to
residue 310. It contains the active site region, and the fold of the PNK domain is structurally
related to T4 PNK (r.m.s.d.: 3.1 Å)51,52. The PNK domain of ceClp1 is composed of an α/β
fold also called Rossmann fold151. Characteristic for the Rossmann fold is a central β-sheet
flanked by α-helices. The central β-sheet has a topological arrangement of four parallel
strands in a 4-1-3-2 order, similar to the previously described T4 PNK51,52. A diagram
describing the topological arrangement of secondary structures within ceClp1 is presented in
Figure 3.1.4 C. However, in contrast to T4 PNK, ceClp1 forms three additional strands
52
3 Results
leading to a seven-stranded central β-sheet. Whereas the PNK domain is structurally
conserved and related to members of the kinase protein family19, the additional NtD and CtD
constitute novel structural features. A structural comparison of this domain from ceClp1 with
the PBD database using the DALI server152 revealed the greatest structural similarities
between ceClp1 and the yeast ortholog94.
Table 3.1.1 X-ray diffraction data collection and refinement statistics of ceClp1.
Data collection
Data set
ceClp1-SeMet
ceClp1-ATP
ceClp1-ADP
ceClp1-AppNHp-G1C2
Space group
P31
P31
P31
P31
Unit cell (a,b,c [Å])
100.7, 100.7, 40.3
99.3, 99.3, 40.8
100.0, 100.0, 40.4
100.4, 100.4, 40.3
Wavelength [Å]
0.9786
0.9793
0.9792
0.9786
Resolution [Å] a
50-2.3 (2.4-2.3)
50-2.3 (2.4-2.3)
50-2.0 (2.1-2.0)
50-2.0 (2.1-2.0)
No. reflections a
40589 (3026)
19707 (2323)
30056 (4086)
30691 (7581)
Completeness [%] a
99.7 (99.5)
98.5 (97.0)
98.6 (99.1)
99.9 (99.8)
<I/(I)> a
13.2 (2.9)
12.4 (4.8)
14.8 (2.4)
17.2 (3.8)
Rmeas a, b
9.6 (57.6)
8.5 (33.4)
5.8 (59.9)
6.2 (46.7)
Redundancy
4.8 (4.7)
3.6 (3.6)
3.5 (3.5)
4.4 (4.3)
19.2 (32.9)
18.2 (22.3)
19.1 (43.6)
18.2 (28.8)
24.0 (42.4)
23.5 (35.4)
24.2 (47.6)
22.8 (31.1)
Overall B factor [Å2]
31.5
39.8
38.1
36.2
Wilson [Å2]
43.8
46.1
45.8
42.1
0.012
0.011
0.009
0.011
1.312
1.275
1.135
1.290
allowed [%]
92.3
92.1
91.8
92.1
additionally allowed
[%]
7.7
7.9
8.2
7.9
PDB entry
Refinement
Rwork a, c
Rfree
a, d
r.m.s.d. e from ideal
geometry:
bond length [Å]
bond angles [°]
Ramachandran
statistics:
53
3 Results
Table 3.1.1 X-ray diffraction data collection and refinement statistics of ceClp1.
Continue.
Data collection
ceClp1
ceClp1
ceClp1
RNA 4mer
ADP-AlF4--G1C2
AppNHp-U1G2
Space group
P31
P31
P31
P31
Unit cell (a,b,c [Å])
99.7, 99.7, 40.8
100.7, 100.7, 40.4
100.7, 100.7, 40.3
99.42, 99.42,
40.40
Wavelength [Å]
0.9786
1.5418
0.9782
0.9792
Resolution [Å] a
43-2.4 (2.5-2.4)
50-2.2 (2.4-2.2)
43.6-2.1 (2.2-2.1)
49.7-2.3 (2.4-2.3)
No. reflections a
17640 (2001)
22553 (2442)
26620 (3434)
19795 (2374)
Completeness [%] a
99.6 (99.1)
96.9 (83.7)
99.9 (99.8)
99.7 (99.7)
<I/(I)> a
17.9 (2.7)
18.9 (7.0)
12.62 (4.44)
17.34 (3.46)
Rmeas a , b
6.8 (56.4)
5.9 (22.7)
14.0 (48.2)
5.5 (41.8)
Redundancy
3.9 (3.7)
3.1 (3.0)
6.0 (5.3)
3.51 (3.59)
Rwork a, c
18.5 (34.7)
22.7 (56.2)
20.2 (41.3)
19.0
Rfree a, d
21.9 (35.9)
28.0 (57.7)
25.0 (40.7)
23.3
Overall B factor [Å2]
34.1
26.5
21.9
35.1
Wilson [Å2]
58.1
36.8
42.0
40.0
0.013
0.009
0.010
0.010
1.355
1.155
1.330
1.090
Allowed
92.4
91.7
92.4
96.3
additionally allowed
7.6
8.3
7.6
3.7
Data set
ceClp1-apo
PDB entry
Refinement
r.m.s.d. e from ideal geometry:
bond length [Å]
bond angles [°]
Ramachandran statistics:
a
values in parentheses refer to the highest resolution shell
zb
Rmeas = h [n/(n-1)]1/2 i  Ih - Ih,i/ hi Ih,i.where Ih is the mean intensity of symmetry-equivalent
reflections and n is the redundancy.
c
Rwork= h Fo – Fc/  Fo (working set, no  cut-off applied)
d
Rfree is the same as Rwork, but calculated on 5% of the data excluded from refinement.
e
root-mean-square deviation (r.m.s.d.) from target geometries.
54
3 Results
Besides this structural conservation, the PNK domain of ceClp1 also showed a significant
structural homology to a signal recognition particle GTPase153 (Z-score 10.8, r.m.s.d.: 2.6 Å,
and a sequence identity of 16 %; PDB accession code: 2J7P, chain: B).
The NtD covers residues 4 to 105 and is composed of a sandwich of two β-sheets. The
electron density map did not allow modeling of the first three amino acids, suggesting that
they are disordered. In contrast to the PNK domain, a DALI-server search152 in the PDB
database revealed no significant similarity to any solved structure. The most similar fold
identified was Mif2p, a conserved DNA-binding kinetochore protein154 (Z-score 8.4, r.m.s.d.:
1.8 Å, and 12 % sequence identity, PDB; accession code: 2VPV, chain: B).
The CtD is slightly larger than the NtD and encompasses residues 311 to 425, folded
in a mixture of β-strands and random coils. For the CtD, a PDB database search152 revealed
low structural homology to the α-subunit of an ATP synthase155 (Z-score: 4.5, r.m.s.d.: 2.5 Å,
and a sequence identity of 9 %; PDB accession code: 3OEH, chain: T). Based on previously
presented structural data of yeast Clp1 (scClp1), it was shown that the interface between the
central PNK domain and CtD is involved in binding of Pcf1194. Pcf11 together with scClp1
are subunits of the cleavage factor CF IA involved in the 3’-end processing machinery of
eukaryotic Pol II transcripts102.
Based on a sequence alignment, the active site region of ceClp1 at the PNK domain
can be divided into an ATP- and an RNA-binding site that contains four recognizable
structural motifs associated with substrate binding and enzyme catalysis (Figure 3.1.1). The
most prominent motif for kinases is the P-loop sequence
121
GxxxxGKT128. Besides the
P-loop, ceClp1 contains a Walker B motif 151DxxQ154, and two motifs that are only conserved
among Clp1 family members of higher eukaryotes. These motifs are a LID module
288
RxxxxR293 and a clasp
230
TxCGW233. To structurally characterize the substrate binding
properties of ceClp1, crystals of the complex were generated by soaking with AppNHp, Mg2+
and G1C2 (inhibited substrate bound state) or by co-crystallization with ATP-Mn2+ (substrate
bound state), and then analyzed.
55
3 Results
Figure 3.1.4 Structure
e of RNPK ceClp1.
c
A, A ribbon rep
presentation of ceClp1 w
with the N-terminal
ase domain (PNK dom
main) and th
he C-terminaal domain (CtD).
domain (NttD), polynuccleotide kina
Structural m
motifs importa
ant for ligand
d-binding or ccatalysis are
e highlighted in color; P-looop (purple), clasp
(orange), LIID module (b
blue), catalyttic base (bla
ack). B, Close-up view off the active ssite in an ide
entical
orientation a
and color co
ode. Residue
es important for substrate
e binding and
d catalysis aare shown as
s stick
representations. C, Diag
gram showin
ng the topolo
ogical arrang
gement of se
econdary struucture eleme
ents in
ceClp1.
3.1.7 RN
NA‐binding
gsite
This thesiss provides the first crystal struccture of a eukaryotic RPNK bouund to an RNA
substrate. The crystall structure of ceClp1 bound to an
a RNA diinucleotide was solved
d and
refined to an R-factoor of 18.2 % and an R free of 22..8 % at a resolution
r
oof 2.0 Å (F
Figure
3.1.5 A).
Cryystals werre soaked with ApppNHp and
d G1C2 representing
r
g an inhiibited
substrate-bbound state. The RNA
A-binding ssite could be
b localized
d in a clefft and show
wed a
positively charged electrostatic surface pottential (Fig
gure 3.1.5 B).
B Both nuucleotides of
o the
56
3 Results
RNA were clearly visible and were found to bind to the conserved clasp comprising the
consensus sequence TxGW. On the one hand, the indol function of the tryptophan residue
(Trp233) mimics a base stacking interaction to the ultimate base (G1) and provides an initial
platform for RNA-binding (Figure 3.1.5 A). On the other hand, the penultimate base (C2) is
stacked against G1, thereby perfectly sandwiching G1 between the side chain of Trp233 and
C2 in a clasp-like binding mechanism. Apparently, G1 gets locked by Trp233 and C2. This
observation suggests that ceClp1 requires two bases as minimal substrate length for efficient
binding (see 3.2.2). G1 gets locked to the Trp233 binding platform only in the presence of a
second base.
Another interaction between clasp and RNA substrate is mediated by the main-chain
amide of a glycine residue (Gly232) (Figure 3.1.5 A). Gly232 forms a hydrogen bond to the
5’-hydroxyl group of the ultimate ribose. The 5’-hydroxyl group is also recognized by
catalytic aspartate of the Walker B motif (Asp151), forming an additional hydrogen bond.
These interactions most likely are involved in the correct positioning of the RNA substrate.
The free 5′-hydroxyl group of RNA points towards the γ-phosphate of the ATP molecule, as
expected for an in-line mechanism52.
ceClp1 also interacts with the oxygen atoms of the bridging phosphate backbone. A
glutamine residue Gln154 forms a hydrogen bond to an oxygen atom of the phosphate group.
Furthermore, guanidinium groups of Arg293 and Arg297 contribute to RNA-binding by
salt-bridge formation to the phosphate groups of the RNA. Interestingly, ceClp1 was shown to
phosphorylate G1C2 with the same efficiency as longer single-stranded RNA oligonucleotides
(Bernhard Loll, personal communication). To explain this biochemical observation, the
structure of ceClp1 bound to a longer RNA oligonucleotide (G1A2A3A4) was solved (Figure
3.1.5 C). The structure model was refined to an R-factor of 18.5 % and an Rfree of 22 % at a
resolution of 2.4 Å. An unambiguous electron density was visible for all nucleotides. Binding
modes between the dinucleotide and the oligonucleotide were found to be virtually identical
(r.m.s.d.: of 1.2 Å, Figure 3.1.5 D). After the first two nucleotides, no additional interaction
contacts between RNA and ceClp1 were observed. Interestingly, a water molecule (W41)
visible in the dinucleotide structure interacts with the 3’-hydroxyl group of C2 and thereby
indicates the putative position of the 3’-5’ bridging phosphate group of a longer substrate
(Figure 3.1.5 A). Indeed, the position of the bridging phosphate group found in the
oligonucleotide crystal structure is identical to the position of W41. The second phosphate
group forms a salt bridge to Arg297. The Watson-crick base pairing interfaces of the
nucleotides are solvent exposed. In both structures, RNA models perfectly match an ideal
57
3 Results
RNA in thhe A-form conformatio
c
on and basee moieties are
a in anti-cconformatioon. Furtherm
more,
the sugar m
moieties werre in the fav
voured 3’-enndo conform
mation.
Figure 3.1..5 Structura
al basis of RNA-binding
R
g by ceClp1
1. A ribbon representatioon of ceClp1 with
polynucleotide kinase domain (PNK
K domain) an d C-terminal domain (CtD). Structuraal motifs imp
portant
for ligand-b
in color; P-loop (purple)), clasp (oraange), LID module
binding or ca
atalysis are highlighting
h
m
(blue), catalytic base (b
black). Dinucleotide (yello
ow), and watter molecules (red) are sshown in stic
ck and
sphere reprresentations.. The ultimatte base (G1) is sandwich
hed in the clasp betweeen Trp233 an
nd the
penultimate
e base (C2). Trp233 is pa
art of the cla
asp. B, ceClp1 with bou
und G1C2 is shown as su
urface
representation colored according
a
to the electrosstatic surface
e potential (co
ontouring froom +5 kT/e in
n blue
to −5 kT/e in red). C, stick
s
represe
entation of th
he RNA-bind
ding site bou
und to G1A2A 3A4. The ulttimate
base (G1) iss sandwiched
d in the clasp
p between T
Trp233 and penultimate
p
base
b
(A2). Trpp233 is part of the
clasp. D, Su
uperposition of G1C2 and G1A2A3A4 sh
howing a virttually identica
al conformattion.
58
3 Results
The 2‘-hydroxyl group of the ultimate base is recognized by a hydrogen bond formed
with the main chain carbonyl of a threonine residue (Thr191, RNA-sensor). This hydrogen
bond could be an explanation for hsClp1’s ability to discriminate single stranded RNA against
single stranded DNA88. Interestingly, based on the oligonucleotide conformation, one could
model an ideal double stranded A-form RNA bound to ceClp1 without any steric hindrance
by superposition with the single stranded RNA (Figure 3.1.6 A). The modeled double-strand
also displays no steric clashes for the model of longer double stranded RNA as a linear
substrate. In contrast to the 5’-acceptor strand, the modeled complementary strand showed no
contact with the protein except when substrates with a 3’-overhang were used (Figure 3.1.6
B). In conclusion, these two crystal structures explain Clp1’s ability to phosphorylate both
single stranded and double stranded RNA substrates88 (see 3.2.2).
Although the interactions of ceClp1 with the RNA backbone as well as the
base-stacking interaction imposed no sequence specificity for the RNA substrate, the effect of
sequence variation on RNA-binding was verified. For this, the dinucleotide U1G2 was soaked
into crystals of ceClp1 and the third structural model of ceClp1 bound to RNA was obtained.
This crystal structure was refined to an R-factor of 20.2 % and an Rfree of 25 % at a resolution
of 2.1 Å (Figure 3.1.7). The guanosine and the uracile nucleotides of both dinucleotide bound
structures were found to be virtually identical positioned (Figure 3.1.7 A and B), mostly
through the clasp-like binding mechanism and interactions with the sugar phosphate
backbone. Furthermore, both crystal structures showed unspecific Watson Crick-like
hydrogen bonds. However, the penultimate base of the U1G2 structure was flipped away into a
syn-conformation and thus did not perfectly stack with the ultimate base (Figure 3.1.7 B). The
observation of a syn- instead of an anti-conformation of the penultimate base of U1G2 is most
likely an artifact due to the use of dinucleotides. For longer nucleotides, the penultimate base
will most likely be in an anti-conformation due to stacking interactions with neighboring
nucleotides.
59
3 Results
Figure 3.1.6 Structurral features of RNA s
specificity and
a
RNA recognition.
r
. A, Ribbon
n-stick
representation of the RN
NA-binding site
s bound to
o an RNA olig
gonucleotide
e (G1A2A3A4).. An ideal do
oubled
stranded (blunt end) A-ffrom RNA oligomer was superimpose
ed onto the single
s
strandded G1A2A3A4. The
individual do
an ideal douubled strande
omains are color
c
coded similar to Fig
g. 3.1.4. B, Furthermore
F
ed (3’overhang) A
A-from RNA oligomer was modeled.
A
B
Figure 3.1.7 Compariison of strructural dettails for RN
NA-binding to ceClp1.. A, Ribbon
n-stick
representation of the RNA-binding
R
site bound
d to the dinu
ucleotide G1C2. B, as w
well as the U1G2.
Structural m
motifs importtant for RNA
A-binding hig
ghlighting in color; clasp (orange), LLID module (blue),
(
catalytic ba
ase (black). The
T
dinucleo
otide (yellow
w), and wate
er molecules (red) are shhown in stic
ck and
sphere reprresentation.
60
3 Results
3.1.8 ATPbindingsite
The ATP binding site was characterized using the ternary complex structure of ceClp1 bound
to an ATP molecule and Mn2+ ion. This structure was refined to an R-factor of 18.2 % and an
Rfree of 23.5 % at a resolution of 2.3 Å (Figure 3.1.8). Electron density for the ATP ligand was
unambiguously identified within the active site region. In contrast to the RNA-binding site,
the ATP molecule binds to the conserved P-loop motif (Figure 3.1.8) that is formed at the
interface between NtD and PNK domain. Interestingly, ceClp1 showed substantial differences
in the ATP-binding site when compared to all other previously described PNKs51,52. Whereas,
in all other PNKs42,51,52 the P-loop is part of a solvent-exposed, open-ended channel, the ATP
binding site of ceClp1 is shielded by NtD. More precisely, the ATP binding pocket is not
completely obstructed by NtD. Instead, a channel is created at the interdomain boundary of
NtD and the PNK (Fig. 3.1.8; ATP/ADP exchange channel). This cleft might enable
ATP/ADP exchange during catalysis. The triphosphate moiety of the ATP molecule is bound
in a narrow channel that is traversing the core of the PNK domain. The wall of this mainly
positively charged channel is formed by residues of the classical P-loop motif, the LID
module that contains conserved arginine residues (Arg288 and Arg293) and a catalytic
divalent metal cofactor that is octahedrally coordinated. Although the ATP-Mn2+ structure
showed a distorted octahedral coordination sphere of the divalent metal cofactor, it was
complete in crystal structures of ceClp1 in complex with an Mg2+ ion. The octahedral
coordination sphere of the Mg2+ ion is formed by the hydroxyl group of Thr128, which is the
only direct interaction with protein and metal cofactor. The remaining sites of the octahedron
were occupied by two non-bridging - and -phosphate oxygen atoms and three water
molecules (W1, W2, and W3). The metal-coordinating water molecules are held in place by
hydrogen bonds between W1 and the oxygen atom of side-chain Asn229, and W2, which
forms a hydrogen bond to the oxygen atom of Asp151 (Figure 3.1.9).
The nucleobase and ribose of the ATP molecule are bound in the domain interface of
NtD and PNK domain. Within this pocket, N6 of the adenine ring forms a hydrogen bond to
the side-chain oxygen atom of Glu16. Furthermore, the nucleobase is at a perfect π-stacking
orientation with the side chain of a phenylalanine (Phe39), which is part of NtD (Figure
3.1.8). The 3’-hydroxyl group of the ribose moiety forms a hydrogen bond to the main-chain
carbonyl of Arg56. The tri-phosphate moiety of the ATP molecule is held in place by
hydrogen bonds of the β- and γ- phosphate with the main chain amides of the P-loop motif as
61
3 Results
well as cooordination-bbonds to thee metal cofaactor. The α-phosphate
α
group is onnly bound by
b the
main chainn amide of the
t P-loop.
Inteerestingly, the
t Walker A lysine oof the P-loo
op (Lys127
7) shows neeither interaaction
with both ATP nor with AppN
NHp (Figurre Figure 3.1.8
3
and 3.1.10).
3
Insstead, Lys127 is
captured ouutside the active
a
site and
a forms hyydrogen bo
onds betweeen its ε aminno group an
nd the
mainchain of Gly121 and Thr230
0, residues w
which are part
p of the P-loop
P
and thhe Clasp. This
T is
in contrastt to the prrevailing sttructural annd biochem
mical dogmaa, which suuggests thaat the
Walker A lysine is innvolved in the
t coordinaation of β- and γ-phossphate group
ups of nucleeoside
o this non-ccanonical co
onformation
n of the “arrrested” Ly
ys127,
triphosphatte113,114. In contrast to
the two arrginine residues (Arg2
288 and Arrg293) of th
he LID mo
odule interaact in a classical
manner wiith β- and γ-phosphatte groups. T
These resid
dues togeth
her with thee divalent metal
cofactor prrovide a poositively chaarged envirronment thaat compensaates the neggatively charged
phosphoryl groups. The
T oppositee side of thee active sitee channel iss closed byy the 5’-hyd
droxyl
group of thhe bound RN
NA substratte.
Figure 3.1.8
8 Characterrization of th
he ATP bind
ding site of ceClp1.
c
A ribbon represeentation of ceClp1
c
with the N--terminal dom
main (NtD), polynucleotiide kinase domain
d
(PNK
K domain) aand the C-terminal
domain (CtD
D). Structura
al motifs impo
ortant for liga
and-binding or
o catalysis as
a highlightinng in color (p
purple)
clasp (orang
ge), and LID
D module (blu
ue). The AT P ligand (yellow, and the
e Mn2+ ion (ppurple) are shown
s
as stick and
d sphere rep
presentation. Residues im
mportant for substrate bin
nding and caatalysis are shown
s
as stick rep
presentation. The adenos
sine moiety of the ATP molecule intteracts with the NtD, wh
hereas
the triphosp
phate group is bound to
o the P-loop
p motif within the active
e site. Furthhermore, the
e ATP
binding site is obstructed by the NtD
D, which form
ms a channel between the
e inter-doma in boundary of the
PNK domain and the NttD leading ov
ver to the trip
phosphate binding site.
62
3 Resultss
ordination o
of the Mg2+ ion.
i
Ribbon representatioon of ceClp1
1 with bound
d
Figurre 3.1.9. Octtahedral coo
2+
Mg . Structural motifs
m
importtant for ligan d-binding or catalysis as highlightingg in color; P-loop (purple))
clasp
p (orange), LID
L
module (blue), cata
alytic base (black).
(
The ATP, G1C22 (yellow), th
he Mg2+ ion
n
(purp
ple), and water molecules
s (red) are sh
hown in stick
k and sphere
e representat
ation. Residues importantt
for the coordinatio
on of the diva
alent metal ccofactor are shown
s
as stick representtation..
3.1.9
9 Phosphooryl transfeer reaction
n mechanism
m of ceClp1
1
Althoough extennsive stud
dies are aavailable on
o the ph
hosphoryl transfer reaction off
monoonucleotidee kinases118, relativelyy little prog
gress has been made ttowards thee structurall
undeerstanding of
o the enzzymatic meechanism of
o PNK. Th
his thesis provides catalytically
c
y
impoortant crysttal structurees of the reaction traajectory off ceClp1. T
The crystall structuress
repreesent an inhhibited substtrate bound state, a tran
nsition statee analog andd a product bound statee
with released RN
NA.
9.1 Inhibitted substratte bound staate 3.1.9
Baseed on the strructure mod
del of the innhibited sub
bstrate boun
nd state it w
was possible to describee
the R
RNA-bindinng site and
d the bindiing of App
pNHp to th
he P-loop m
motif (Figu
ure 3.1.10)..
Presuumably, thee γ-phosphate group off the ATP molecule
m
is oriented
o
in-lline with respect to thee
seconnd substratte (G1C2), creating tthe correct geometry to enablee the transsfer of thee
γ-phoosphoryl grroup to thee 5’-hydroxxyl group. Both the structural ddata and siite directedd
mutaagenesis stuudies suggeested Asp1551 to be th
he catalyticc general bbase that acctivates thee
5’-hyydroxyl grooup. During
g enzyme caatalysis, Assp151 initiates the reacction by dep
protonationn
of thhe 5’-hydroxxyl group of the ultimaate base. Ass mentioned
d before, the
he lysine ressidue pointss
63
3 Results
outwards from the active site prior to the phosphoryl transfer reaction. The LID arginines are
involed in ATP-binding by interactions with their guanidinuim group and the oxygen atoms
of the phosphate β- and γ-phosphate groups.
3.1.9.2 Transitionstatemimicwithaluminumtetrafluoride
Our current understanding of phosphoryl transfer reactions is strongly supported by structural
information on transition state analogs. It has been shown that aluminum, beryllium fluoride,
and vanadate can be used as transition state analogs of the phosphoryl transfer reaction122,123.
These analogs of a transferred phosphate group resemble transition state geometry and charge
distribution121. This thesis provides the first crystal structure of a transition state analog
obtained from a eukaryotic PNK (Figure 3.1.11). The crystal structure of a ternary complex
(ADP-AlF4--Mg2+-G1C2) was refined to an R-factor of 22 % and an Rfree of 28 % at a
resolution of 2.1 Å. Although bound AlF4- does not represent the trigonal bipyramidal
transition state geometry, it is a good approximation and thus broadly accepted to represent a
mimic of the transition state121. The octahedrally-coordinated aluminum atom interacts with
four equatorial fluorine atoms, while the two axial positions are occupied by an oxygen atom
of ADP’s β-phosphate and an oxygen atom of the attacking nucleophile, respectively. The
aluminum atom is positioned between the donor oxygen atom of ADP and the acceptor
oxygen atom of the polynucleotide with a distance of 2.1 Å and 2.2 Å, respectively. This
apical coordination of donor and acceptor oxygen atoms suggests an in-line associative
reaction mechanism similar to that for UMP kinases118. Negative charges of the penta
coordinated transition state seem to be neutralized by electrostatic interactions with positively
charged side chains of Lys127, Arg288, Arg293 and the Mg2+ ion. The observation of a
charge-neutralized transition state is consistent with an associative reaction mechanism.
Moreover, RNA and ADP move towards each other under the concerted actions of Lys127,
Arg288, and Arg293. Importantly, the Walker A lysine no longer interacts with Gly121 and
Thr230 but instead becomes activated. After the transfer of the phosphoryl group, RNA is
released from the active site, which leads to the product-bound state with bound ADP.
64
3 Resultss
Figurre 3.1.10. In
nhibited sub
bstrate boun
nd state. A, stick representation of a close-up viiew from the
e
2+
active
e site of ceClp1 crystalliz
zed in the inh
hibited substrrate bound state
s
AppNHpp-Mg -G1C2, prior to the
e
transition state co
omplex, the Walker A lyssine (Lys127
7) is “arreste
ed” outside tthe active sitte. B, Active
e
electron denssity of ATP-M
Mg2+-G1C2. T
The grey me
esh depicts Fo
F - Fc differrence density
y (contoured
d
site e
at 3.0
0 σ) calculatted prior to inclusion of tthe nucleotid
des or a watter-metal com
mplex in the model. The
e
nucle
eotides are depicted
d
as stick model s. The orien
ntation is the
e same as iin A. The co
olor code off
structtural motifs important for ligand-bindi ng or catalys
sis is similar to Figure. 3. 1.4.
65
3 Results
on state ana
alog. stick re
epresentation
n of a close-up view from
m the active site
s of
Figure 3.1.11 Transitio
sition state a
analog ADP
P-AlF4--Mg2+-G1C2. The Lys127 bec
comes
ceClp1 crystallized witth the trans
B, Active site
e electron density of ADP
P-AlF4--Mg2+-G
- 1C2.
activated “sswitched-on” in the transiition state. B
The grey mesh depicts Fo - Fc diffe
erence densitty (contoured
d at 3.0 σ) calculated priior to the inc
clusion
of the nucle
eotides or a metal
m
comple
ex in the mod
del. The nuc
cleotides are depicted as stick models
s. The
orientation is the same
e as in A. The color cod
de of structu
ural motifs im
mportant forr ligand-binding or
gure. 3.1.4.
catalysis is similar to Fig
66
3 Results
3.1.9.3 Product bound state with released RNA
Since ceClp1 is active in the crystalline state, it was possible to obtain a product bound
state by soaking crystals of ceClp1 with its substrates. After incubation of ceClp1 crystals
with ATP and G1C2, crystal structures revealed electron density for an ADP molecule, while
phosphorylated G1C2 was not present. This crystal structure represents the RNA released
product bound state and was refined to an R-factor of 19 % and an Rfree of 24 % at a
resolution of 2.0 Å (Figure 3.1.12). It is likely that the lack of electron density for
phosphorylated RNA results from electrostatic repulsion between β-phosphate of ADP and
5’-phosphorylated RNA. Interestingly, this crystal structure is characterized by a “switched
off” Lys127 that points out of the active site, similar to the substrate-bound state. Lys127
switches into an “arrested” state prior to and after the transition state. Furthermore, the latch
arginines, Arg288 and Arg293, are reoriented. In a latch-catch mechanism, Arg293 of the LID
module forms a salt bridge with Asp124 of the P-loop motif. In conclusion, ceClp1 seems to
use a molecular gating mechanism characterized by a reorganization of the active site such
that Lys127, Arg288, and Arg293 stabilize negative charges of the transferred phosphoryl
group in the transition state, while preventing ATP hydrolysis in the absence of an RNA
substrate.
Interestingly, the arginines of the LID module seem to function in a latch-catch
mechanism (Figure 3.1.11). These arginine latches are completely activated only in the
transition state and form, together with Lys127, an environment that enables correct
positioning and charge stabilization of the transition state (Figure 3.1.11).
3.1.10 Summaryofthestructuralresults
This thesis provides the first crystal structure of a eukaryotic PNK with a bound RNA
oligonucleotide. Furthermore, high-resolution structures of enzymatically relevant states of
ceClp1 were determined (AppNHp-Mg2+-G1C2, ADP-Mg2+-AlF4--G1C2, and ADP-Mg2+).
Clp1’s RNA-specificity and RNA recognition are described based on these structural data.
Additionally, it was possible to elucidate the phosphoryl transfer reaction mechanism.
Interestingly, crystal structures showed a non-canonical Walker A lysine (Lys127) that seems
to act as a molecular switch. Activation of this switch is achieved in the transition state
complex while before and after the Walker A lysine is in an arrested conformation. In contrast
to other nucleotide kinases, Clp1 seems to use a substrate-gating mechanism.
67
3 Results
Figure 3.1..12 RNA rele
eased produ
uct bound sttate. stick re
epresentation
n of a close--up view from the
active site o
of ceClp1 cryystallized in the RNA rele
eased produc
ct bound state ADP-Mg2++. Prior to and
d after
the transitio
on state com
mplex, the Walker
W
A lys ine (Lys127) is “arrested” outside tthe active site. B,
Active site e
electron denssity of ADP-M
Mg2+. The grrey mesh depicts Fo - Fc
c difference ddensity (conttoured
at 3.0 σ) ca
alculated priior to the inc
clusion of th
he nucleotide
es or a meta
al complex iin the model. The
nucleotides are depicte
ed as stick models.
m
The
e orientation is the same
e as in A. T
The color co
ode of
structural m
motifs importa
ant for ligand-binding or ccatalysis is similar to Figu
ure. 3.1.4.
68
3 Results
3.2
BiochemicalcharacterizationofClp1fromC.elegans
This section describes the biochemical characterization of ceClp1, which was shown to be a
bona fide RPNK enzyme similar to previously described human orthologs88. In addition to
high sequence conservation between these orthologs, ceClp1 and hsClp1 have identical RNA
and DNA substrate specificities. The enzymatic characterization of ceClp1 showed that the
dinucleotide is the minimal substrate for an efficient phosphoryl transfer. Intriguingly, in case
of short RNA substrates (dinucleotides), ceClp1 shows a clear preference for purine against
pyrimidine bases at the ultimate position, which is in accordance with the crystal structure.
Furthermore, no basal ATPase activity or product inhibition could be measured for ceClp1,
which suggested a highly regulated 5’-kinase activity. Moreover, the 5’-kinase reaction is
irreversible under single turnover conditions.
3.2.1 SubstratespecificityoftheClp1proteinfamily
An RPNK activity of ceClp1 was demonstrated using an in vitro phosphorylation assay with
different RNA and DNA substrates. A comprehensive biochemical characterization of hsClp1
revealed a strong preference for RNA substrates as the 5’-phosphate acceptor88. To compare
the specificity of ceClp1 with hsClp1, the enzyme (2 µM) was incubated with different single
stranded and double stranded oligonucleotides (0.1 mM), followed by a separation using
denaturing PAGE. The control reactions with T4 PNK verified the differential electrophoretic
mobility of phosphorylated oligonucleotide substrates (Fig. 3.2.1). In accordance with
observations made for hsClp1, ceClp1 also discriminates single stranded RNA against single
stranded DNA (Fig. 3.2.1 A). However, ceClp1 is able to phosphorylate double-strand
nucleotide RNA and DNA substrates with blunt ends. But again, the efficiency is lower for
double stranded DNA (Fig. 3.2.1 B). Similar results were obtained for double stranded RNA
and double stranded DNA molecules with a 3’-overhang (Fig. 3.2.1 C). Interestingly, blunt
ended double stranded RNA substrates showed higher phosphorylation efficiency when
compared to double stranded RNA substrate with a 3’-overhang. Even though ceClp1 showed
a clear preference for single stranded RNA and double stranded RNA, it also displayed a
residual activity towards single stranded DNA during a longer incubation time. Complete
69
3 Results
turnover of single straanded DNA
A substratess was observed after 2 hours of inncubation. These
T
results reveealed conseerved substrate specificcity between
n ceClp1 and hsClp1.
Figure 3.2..1 Substrate
e specificity
y of ceClp1 on RNA and DNA substrates. ceeClp1 (1 µM
M) was
incubated w
with various RNA and DN
NA oligonuclleotide substtrates (20 ntt of length), aat a concenttration
of 100 µM. Substrate molecules were eithe
er single-strranded oligonucleotide, double-stra
anded
oligonucleottides contain
ning a blunt end or dou
uble-stranded
d oligonucleo
otides with a 3’-overhan
ng. T4
PNK reactio
on samples are
a shown as
s positive co
ontrols (Ctrl). A, ceClp1 has
h robust kinnase activity
y upon
incubation w
with single stranded
s
RNA
A, whereas ssingle strand
ded DNA wa
as inefficientlly phosphory
ylated.
Notably, aftter 2 h incubation, basal phosphoryla
ation of single stranded DNA
D
was obsserved. B, ceClp1
c
displays en
nzymatic acttivity on dou
ubled-strande
ed oligonucleotides with
h blunt endss, however, RNAmolecules a
are more effiiciently phosphorylated. C
C, double stranded RNA
A oligonucleootides contain
ning a
3’-overhang
g are less effficiently phos
sphorylated b
by ceClp1 an
nd DNA mole
ecules are noot phosphory
ylated.
3.2.2 Minimal subsstrate requiirements off ceClp1
The substrrate specificcity assay (see 3.2.1) aas well as recent
r
in vittro experim
ments by Weitzer
and Martinnez88 indicated that Clp1
C
is ablee to phosp
phorylate a number off different RNA
polynucleootides. How
wever, inform
mation abouut ceClp1’s requiremen
nts on the m
minimal sub
bstrate
length of R
RNA is missing. To deefine these m
minimal sub
bstrate requ
uirements, 55’-kinase acctivity
of ceClp1 towards suubstrates with varyingg lengths was determin
ned. Activitty was anaalyzed
using Micchaelis-Mennten kineticcs. All meaasurements were perfformed undder steady state
conditions. Two dinuucleotides, G1C2 and U 2G2, identified in the crystal struucture were used
K activity asssays. ceClp
p1 was usedd at limiting
g enzyme concentratio
c
on (0.2 µM)). The
for 5’-PNK
70
3 Results
reaction mixture containing 1 mM ATP and 0.5 mM of either G1C2 or U1G2 and pre-incubated
for 15 minutes at 20°C. After this pre-incubation step, the reaction was started by addition of
ceClp1 and then analyzed using a coupled calorimetric ATPase assay (Figure 3.2.2). Although
ceClp1 showed a significant ATP turnover with both dinucleotides, phosphorylation of G1C2
was much more efficient compared to U1G2 (Figure 3.2.2 A). After an initial lag phase, the
ATP turnover of G1C2 and U1G2 reached its maximum velocity. The same experiment was
performed with 0.5 mM of adenosine- and cytosine 3’-monophosphate (3’-AMP and
3’-CMP). Neither with 3’-AMP nor with 3’-CMP, was an ATP turnover measured. Results of
these measurements, together with structural data (see 3.1.7), suggest that correct binding to
the clasp motif of ceClp1 is impaired for 3’-nucleoside monophosphate substrates. These
conclusions are further supported by results obtained for hsClp1 and the mammalian PNK,
which also showed no 5’ -kinase activity on 3’ -nucleoside monophosphates85,88. Currently,
only T4 PNK, has been shown to phosphorylate 3’-nucleoside monophosphates46. Therefore,
a control reaction was performed with T4 PNK under identical conditions. As expected, in the
presence of 3’-AMP and 3’-CMP, enzymatic activity could be detected for T4 PNK (Figure
3.2.2 B).
Michaelis-Menten kinetics of ceClp1 were measured at varying dinucleotide (G1C2
and U1G2) concentrations. The Vmax and Km values were obtained by fitting to a
Michaelis-Menten equation (Equation 1.2). At saturating ATP concentration of 1 mM, ceClp1
(0.2 µM) showed an apparent kcat of 2.5 s-1 and Km of 103.0 μM for G1C2, and a kcat of 2.3 s-1
and Km of 1122.0 μM for U1G2 (Figure 3.2.3). These kinetic data suggest that under saturating
dinucleotide concentrations, ceClp1 displays similar turnover rates for G1C2 and U1G2.
However, a tenfold higher Km value indicated a much lower substrate affinity for U1G2. This
can be explained by the crystal structure of ceClp1 bound to the U1G2, which shows an
imperfect base-stacking interaction mimic of the ultimate base (U1) to Trp233. The
penultimate base (G2) is flipped into a syn-conformation. However, the position of the 5’hydroxyl was in ideal hydrogen bonding distance to Gly232 and Asp151. These structural
details are reflected in the increased Km value and identical turnover rates at saturation.
Interestingly, experiments with longer nucleotide substrate revealed similar Michaelis-Menten
constants for G1C2. Michaelis-Menten values of a penta nucleotide G1A2A3A4A5 determined
under identical conditions as in this thesis exhibited a kcat = 2.2 s-1 and Km = 99.0 μM (Bernard
Loll). As suggested by the protein-RNA interaction profile seen in the crystal structure of
ceClp1 (Figure 3.1.5), substrate affinities for G1C2 and G1A2A3A4A5 are almost identical.
71
3 Results
Althhough ceClp1 showed
d a preferennce for purrine bases at the ultim
mate positio
on of
dinucleotidde substratees, it is qu
uestionable whether this
t
is the case with oligonucleeotide
substrates as well. It
I is more likely thaat for long
ger substrattes, all basses are in antib
ng, which may
m stabilizze the impeerfect interaaction
conformatiion with coontinuous base-stackin
between a pyrimidine base and
d Trp233 att the ultim
mate position
n. To test whether ceeClp1
discriminattes purine and pyrim
midine basess even for longer sub
bstrate, an experimentt was
performed comparingg two com
mplementaryy oligonucleotides (20mer) thatt differ at their
ultimate annd penultim
mate positio
ons (G1C2-1 8mer and C1C2-18merr). Since it was shown
n that
ceClp1 haas similar turnover
t
rattes for lonnger oligonu
ucleotide (18mer) com
mpared to G1C2
(Bernard L
Loll), RNA
A concentrattions (0.1 m
mM) were chosen
c
in which
w
the eenzyme is under
u
unsaturatedd RNA connditions. Th
hese measurrements ind
dicated that both singlle stranded RNA
substrates aare phosphoorylated witth the same efficiency (Figure 3.2.5).
In conclusion, ceClp1 iss unable too phosphorrylate 3’-nu
ucleoside m
monophosph
hates,
whereas RN
RNA dinucleeotides reprresent a suittable substraate for 5’-k
kinase activiity. Furtherm
more,
ceClp1 dissplayed no further preeference wiith respect to length of
o the oligoonucleotide. The
dinucleotidde G1C2 waas identified
d as the minnimal substrrate requireements of ceeClp1. Based on
the structuural data annd the biocchemical exxperiments,, ceClp1 seeems to haave no sequ
uence
specificitiees for longerr RNA subsstrates.
Figure 3.2.2
2 Nucleotide
e sequence length dep endence of the 5’-kinas
se activity. T
The kinase activity
a
of ceClp1 a
and T4 PNK
K was monittored spectrroscopically based on a continuouss coupled-AT
TPase
activity assa
ay. A, ceClp
p1 (0.2 µM) was incuba ted with 1 mM
m ATP, and
d 0.5 mM off 3’-AMP, 3’’-CMP
G1C2, and.U
U1G2. Once a dinucleotid
de was used
d as a substrrate, ceClp1 showed a ssignificant turrnover
of ATP, in contrast to mononucleot
m
tides. The d inucleotide appears
a
to be
b ceClp1’s minimal sub
bstrate
requirementt. B, A contrrol experimen
nt was condu
ucted with th
he T4 PNK (5
5 U) supplem
mented with 1 mM
ATP and 0.5 mM of tw
wo different mononucleot
m
tide substrate
es (3’-AMP and 3’-CMP
P). T4 PNK shows
s
clear kinase activity to
owards both substrates,, whereas no
n basal ac
ctivity was ddetectable without
w
acceptor.
72
3 Resultss
Figurre 3.2.3 Min
nimal substrrate require
ements of ceClp1.
c
A, The
T Michaeliss-Menten kin
netics of the
e
initial enzyme velocities were
e plotted aga
ainst varying
g dinucleotide concentrattions. All me
easurementss
were performed at
a constant ATP
A
concentrrations (1 mM
M). The black line depictss the fit of the data using
g
the sstandard Micchaelis-Mentten equation
n. B, Calculated consta
ants resultinng from the fit with the
e
stand
dard Michaelis-Menten model.
m
C, and
d D, Residuals of the Michaelis-Menteen plot.
quence specificity of c
ceClp1 towa
ards oligonu
ucleotides. S
e kinetics off
Steady state
Figurre 3.2.4 Seq
ceClp
p1 measured
d with two different singlle stranded RNA
R
oligonu
ucleotides. B
Both oligonuc
cleotides are
e
comp
plementary to
o each other.
73
3 Results
3.2.3
Characterization of ceClp1’s 5’-kinase activity
Using an enzymatically coupled spectroscopic ATPase assay, it was possible to characterize
the substrate specificity of ceClp1. However, these experiments have the drawback that the
quantification of enzymatic activity only relies on conversion of ATP to ADP. Therefore, a
second experimental setup for an in vitro phosphorylation assay was used to support kinetic
data obtained with the coupled-enzyme assay (Figure 3.2.5). The second assay is based on a
MonoQ anion exchange chromatography protocol designed for binding and baseline
separation of ceClp1’s 5’-kinase reaction substrates from products.
Steady-state kinetics were measured under limiting enzyme concentrations (0.2 µM) in
a reaction mixture containing saturating ATP (1 mM) and 0.5 mM G1C2 concentrations. For
the second assay, ceClp1 (2µM) was incubated with ATP and G1C2, each in a concentration
of 250 µM, at 25 °C for 2 hours. Due to a change in ionic net charge after the phosphoryl
transfer reaction, substrates and products can be separated by anion exchange
chromatography. To assess possible basal ATPase activity of ceClp1, both assays were
performed in absence of G1C2, which showed no basal ATPase activity (Figure 3.2.5 A and
B). However, after the addition of RNA a significant ATP turnover and an accumulation of
the reaction products ADP and 5’P-G1C2 were detected. Using the phosphorylation assay, it
was also tested if the net charge of the RNA is modified in an ATP-independent manner,
which was not the case (Figure 3.2.6 C). Furthermore, it was possible to show that the kinase
reaction proceeds in a one-to-one stoichiometry. Therefore, ceClp1 was incubated with ATP
(500 µM) in two-fold excess compared to the G1C2 concentration. Observed ATP turnover
was proportional to the conversion of G1C2 to 5’P-G1C2. ceClp1 shows neither unstimulated
basal ATPase activity nor RNA-stimulated basal ATPase activity (Figure 3.2.5 D).
In addition, ceClp1 was tested for product inhibition using the coupled calorimetric
ATPase assay at varying G1C2 concentrations. ATP turnover rates were followed by the
decrease in absorbance at 340 nm due to oxidation of NADH (Figure 3.2.6). The reaction rate
of each measurement obtained under varying G1C2 concentration was plotted and analyzed.
Maximum velocity of the reaction was reached after an initial lag phase. Although a slight
decrease of ATP turnover rates at the end of the reaction was measured, it is possible that this
decrease is due to dilution effects. Consistent with this hypothesis the weaker decrease of
ATP turnover rates at higher G1C2 concentration. Thus, ceClp1 seems to show no product
inhibition by accumulation of 5’P-G1C2.
74
3 Resultss
Figurre 3.2.5 Biochemical ch
haracterizattion of ceClp
p1. A possib
ble basal AT Pase activity
y was tested
d
by a spectroscop
pically based
d coupled-AT
TPase activitty assay as well as a phhosphorylatio
on assay. A,,
ceClp
p1 (0.2 µM) incubated with
w 1 mM AT
TP, and with
h or without 0.5 mM of G 1C2, showiing no basall
activity. B, 5’-sub
btrate phosph
horylation wa
as analyzed on a MonoQ
Q anion exchhange colum
mn. Elution off
the n
nucleotide wa
as achieved with an incrreasing linea
ar NaCl gradient that dete
termines the conductivityy
(brow
wn). Dependiing on the numbers of p hosphoryl grroups, nucleotides bind w
with a differe
ent affinity to
o
the co
olumn matrixx (RNA-5’OH
H < RNA-5’P < ADP < AT
TP). The abso
orbance at 2260 nm (red) and 280 nm
m
(blue) is indicate
ed in the plo
ots. Note tha
at the RNA-dinucleotide
e (G1C2) cann be distinguished from
m
adenine molecule
es (ATP/ADP
P) based on their different A260/A280 ratio of ~2 aand ~7, respectively. Forr
the n
negative con
ntrol, ceClp1
1 was incub
bated with 250
2
µM of ATP
A
but witthout G1C2, showing no
o
NA was exc
conve
ersion of AT
TP to ADP. C, An ATP
P-indepented
d modificatio
on of the RN
cluded by a
phosp
phorylation assay
a
withou
ut ATP, show
wing no net charge
c
chang
ge of G1C2. D
D, The reaction was also
o
condu
ucted with an excess of 500µM of A
ATP. The obs
served turno
over of ATP pproceeds un
ntil the entire
e
RNA substrate is consumed.
75
3 Results
Figure 3.2..6 Characterization of production inhibition of
o ceClp1. Based on a spectroscop
pically
enzyme-cou
upled ATPasse activity as
ssay, data w
were recorded
d to test for product inhibbition. ceClp1 (0.2
µM) incuba
ated at saturated ATP co
oncentrationss (1 mM) un
nder varying G1C2 conceentrations (50, 75,
100, 150 µM
M).
3.2.4 Revverse reacttion of ceCllp1
The exactt function of Clp1’s RNA 5’-kkinase activ
vity has rem
mained eluusive until now.
However, tthe identificcation of a potential
p
reeversal of th
he phosphorrylation reacction would
d give
new insighhts into RNA
A maturatio
on pathwayss. Experimeents on the human
h
ortho
holog suggessted a
quasi-irrevversible 5’-kkinase reacttion85. This is in contraast to measu
urements w
with thermosstable
archaeal Clp1 homoloog95, which exhibited a reversible reaction.
r
Fu
urthermore, a reversal of
o the
kinase reacction is alsoo demonstraated for otheer PNKs lik
ke T4 PNK156,157. ceClpp1 was testeed for
the putativve reverse reaction under
u
singlle-turnover condition. An ideal single-turn
nover
experimentt is charactterized by th
he fact thatt every substrate molecule is initiially bound
d by a
single enzyme. If ceC
Clp1 indeed is a quassi-unidirecttional kinasse, no equillibrium bettween
n be reachedd. Since the reaction prroceeds intoo the directiion of
forward annd reverse reaction can
the energettically favoorable forwaard reactionn, it is impossible to deetect the connversion off ADP
to ATP. The reverse reaction waas analyzedd by anion exchange
e
chromatograaphy, based
d on a
purificationn protocol that
t
enabless baseline seeparation off ADP and ATP (Figurre 3.2.8). ceeClp1
(60 µM) w
was incubatted with a phosphoryla
p
ated dinucleeotide (5’P--G1C2) andd ADP, each
h at a
concentratiion of 50 µM
M. After an
n incubationn time of 20
0-40 minutees at 20 °C, the reaction
n was
quenched aand conversion of AD
DP to ATP w
was measurred. The tw
wo substratees ADP and
d 5’PG1C2 weree eluting froom the anio
on exchangee column as
a a single peak,
p
but coould be basseline
separated from the potential
p
prroducts (AT
TP and G1C2). The reeverse reacction experiiment
revealed thhat ceClp1 is
i functioning as a quaasi-irreversiible polynuccleotide kinnase, showin
ng no
76
3 Resultss
form
mation of AT
TP or G1C2 after 20 orr even 40 minutes
m
of in
ncubation (FFigure 3.2.8
8 A and B)..
A coontrol expeeriment un
nder identiccal conditio
on was performed inn the absen
nce of thee
phosphate-donoor (5’P-G1C2) to excludde any basall adenylate kinase
k
activvity (Figure 3.2.8 C). Iff
o ATP andd
a pottential adennylate kinasse contaminnation was present, a conversionn of ADP to
AMP
P would havve been detected. How
wever, no forrmation of ATP or AM
MP was meaasured afterr
20 m
minutes incuubation timee. Furtherm
more, to excclude that high
h
proteinn concentraation in thee
singlle turnover experiment
e
affects Clpp1’s enzymaatic activity, the enzym
me was also tested in itss
forw
ward directioon. Therefore, ceClp1 (60 µM) was
w incubateed with AT
TP and G1C2 each at a
concentration of 50 µM. The
T analyzeed reaction mixture revealed a coomplete con
nversion off
ATP and G1C2 to
t ADP and
d 5’P-G1C2 (Figure 3.2.8 D) after an incubatioon time of 20
2 minutes..
m between forward annd reverse reaction iss
This result impplies that, again, no eequilibrium
reachhed and thatt the 5’-kinaase reactionn is the enerrgetically favorable direection.
Figurre 3.2.8 Rev
versal of the
e polynucleo
otide kinase
e reaction. A,
A and B, reeaction mixtu
ures (200 µl))
conta
aining 60 µM of ADP, 50 µM of 5’P-G
G1C2, and 50 µM of ceClp
p1 were incubbated for 20 and 40 min,,
respe
ectively. The
e products were
w
analyzzed on a MonoQ
M
anion
n exchange column. Elution of the
e
nucle
eotide was achieved
a
with
h an increassing linear NaCl
N
gradien
nt similar to the descripttion in figure
e
3.2.1. Even after 40 min, ceC
Clp1 showed no reverse reaction which would leaad to produc
ction of G1C2
ATP in a sin
ngle turnoverr experimentt. C, to exclu
ude any bas
sal adenylatee kinase activity, ceClp1
and A
was iincubated on
nly with 50 µM
µ of ADP ffor 40 min. Consistently
C
, no accumuulation of AT
TP and AMP
P
were detectable. D, shows th
he forward re
eaction in a single
s
turnov
ver experimeent. The reac
ction mixture
e
ng 50 µM off ATP and 5
50 µM of G1C2 was incubated for 200 min. In contrast to the
e
(200 µl) containin
reverrse reaction, the PNK forw
ward reactio n led to the accumulation
a
n of ADP andd 5’P-G1C2.
77
3 Results
3.2.5
Summary of the enzymology studies
Altogether, the biochemical characterization demonstrates that ceClp1 is an RPNK with
identical substrate specificity as hsClp1. Although ceClp1 is able to phosphorylate RNA and
DNA polynucleotides, it shows a much higher affinity for RNA substrates. Furthermore,
steady-state kinetics revealed that ceClp1 requires at least a dinucleotide for an efficient
phosphoryl transfer reaction. Intriguingly, the dinucleotide G1C2 represents the minimal
substrate for ceClp1, showing similar Michaelis-Menten kinetics as measured for an
oligonucleotide substrate. However, in case of the dinucleotide substrate, ceClp1 displays a
preference for purine against pyrimidine bases at the ultimate position as indicated by the
tenfold increased Km of U1G2 compared to G1C2. Interestingly, the characterization of ceClp1
reveals no basal ATPase activity, and moreover, ceClp1 seems to show no product inhibition.
In contrast to other PNKs42,51,52, ceClp1 functions as a quasi-unidirectional enzyme with no
reversal of the 5’-kinase reaction.
3.3
MutagenesisstudiesonceClp1
3.3.1 SitedirectedmutagenesisofLys127andTrp233
Previous mutagenesis studies in our lab allowed the identification of a number of catalytically
important residues for ceClp1 enzyme activity (Table 3.3.1). Conservative and nonconservative mutations were introduced based on the sequence alignment of Clp1 from higher
eukaryotes (Figure 3.1.1). As expected, hot spots for mutation sites were found in highly
conserved regions like the P-loop motif, the Walker B motif, the LID module and the clasp.
This characterization focused in particular on conserved residues of the PNK domain (Pro122,
Thr123, Asp124, Asp151, Glu154, Asp235, Arg288, and Arg293) to show their potential
involvement in enzyme catalysis. In addition, in this thesis structure-guided site-directed
mutagenesis and biochemical methods were employed to supplement the previous mutational
characterization of ceClp1 (Table 3.3.1). The wild-type protein and mutated protein-variants
(K127A, K127R, and W233A) were expressed and purified as described for crystallization
experiments (see 3.1.2, Figure 3.3.1 A). Subsequently, purified protein variants were
characterized for their 5’-kinase activity using an enzymatically coupled spectroscopic
ATPase assay (Figure 3.3.1 B). Steady state kinetics were measured under limiting protein
concentration (2 µM) with excess of ATP (1 mM) and G1C2 (500 µM). To exclude any
78
3 Results
destabilizing effects by the mutation, the thermodynamic stability of all mutant variants was
verified by determining the apparent melting point (Tm) using circular dichroism
spectroscopy. All mutant variants revealed no significant differences compared to the wild
type (Figure 3.3.1 C and D).
Since the Walker A lysine (Lys127) of the P-loop showed a non-canonical “switchedoff” conformation prior to and after the transition state, this residue was considered to be part
of a molecular gating mechanism. Therefore, the structure-function relation was examined by
a conservative substitution. Lys127 was replaced with an arginine residue and a
non-conservative alanine substitution. Based on ATPase activity, Lys127 proved to be
essential for enzyme catalysis since K127R was abolished in enzymatic activity. It is possible
that K127R mutation might interfere with the bifurcated binding of the β- and γ- phosphates
groups by Lys127 at the transition state structure (Figure 3.1.10). Non-conservative mutations
of the Walker A lysine have previously been shown to abolish enzyme activity of P-loop
kinases and therefore serve as a negative control88. As predicted, the 5’-kinase activity of the
K127A variant was abolished (Figure 3.3.1 B).
The crystal structure of ceClp1 bound to an RNA substrate (Figure 3.1.5) showed a
base stacking interaction mimic between the indol group of Trp233 and the ultimate base of
the substrate. Alanine substitution of Trp233 (W233A) also abolished the 5′-kinase activity of
ceClp1. Therefore, the π-stacking interaction between Trp233 and the ultimate nucleobase
seems to be critical for RNA-binding of ceClp1.
79
CD [mdeg]
3 Results
Figure 3.3..1 Site direc
cted mutag
genesis of k
key residue
es of ceClp1 involved in catalysis
s and
RNA-bindin
ng interaction. A, SDS--PAGE of ce
eClp1 mutan
nt variants. B,
B Kinase acctivity of wild
d type
Clp1 and different muta
ant variants monitored sspectroscopic
cally based on a coupleed ATPase activity
a
assay. The experiment was
w conductted with 2 μM
M protein, 1 mM
m ATP, and 500 µM G11C2. Wild typ
pe (wt)
protein sho
ows 5’-kinase
e activity. The
T
mutated variants K1
127R and W233A
W
show
wed no enzy
ymatic
activity. A W
Walker A lyssine mutant (K127A) serrves as a co
ontrol for an inactive enzzyme. C, Th
hermal
melting currve exempliffied for nativ
ve ceClp1. T
Thermal unffolding is an
n irreversiblee process due
d
to
precipitation
n of the prote
ein. D, Deterrmined apparrent melting temperature
es (Tm) of ceC
Clp1 wild typ
pe and
variants. Alll protein sam
mples show comparable
c
g
global stabilitty.
3.3.2 Deletionofth
heN‐terminalandC
C‐terminaldomain
ceClp1 is a multi-dom
main protein
n with a cenntral PNK domain containing the aactive site region
(Figure 3.11.4). The ceentral domaain is flankeed by an add
ditional NtD
D and CtD, but the fun
nction
of these addditional doomains hass remained elusive so far. To ad
ddress the qquestion to what
extend the NtD and CtD
C contribu
ute to 5’-kinnase activity
y, various NN and C-terrminal trunccation
mutant varriants of ceC
Clp1 were designed
d
andd tested for enzymatic activity.
80
3 Results
Table 3.3.1 Michaelis-Menten steady-state constants of wild type ceClp1 and mutated variants
using G1C2 as substrate. Titration experiment was performed at a constant concentration of 1 mM
ATP and 0.2 μM ceClp1.
Enzyme
KM [μM]
kcat [s-1]
ceClp1wt
109
2.5
ceClp1P122A
187
1.8
ceClp1P122S
691
3.1
ceClp1T123A
residual activity
ceClp1T123S
185
0.9
ceClp1D124A
1764
1.3
ceClp1D124Q
ceClp1D124N
no activity
612
ceClp1K127A
no activity
ceClp1K127R
no activity
ceClp1D151A
Function
Reference
This thesis
A P-loop residue involved in main chain
interactions with the nucleoside triphosphate.
*
A P-loop residue involved in main chain
interactions with the nucleoside triphosphate.
*
A P-loop residue that interacts via its side chain
with both residue Arg288 and the ATP molecule.
*
0.3
*
*
*
*
The Walker A lysine is usually involved in
coordination of the β- and γ-phosphate groups.
This thesis
no activity
The catalytic base is involved in the deprotonation
of the 5’-OH group.
*
ceClp1Q154A
no activity
The side chain of this residue interacts with 5’→3’
phosphate group of the ultimate base.
*
ceClp1W233A
no activity
This residue is involved in a base stacking
interaction mimic with the ultimate base.
This thesis
The side chain of this residue forms a salt bridge
interaction with Arg261.
*
The side chain of this residue is involved in charge
neutralization of phosphate groups.
*
ceClp1D235A
ceClp1R288A
ceClp1R288K
244
3.2
no activity
276
2.2
ceClp1R293L
no activity
ceClp1R293K
residual activity
The side chain of this residue is involved in charge
neutralization of phosphate groups.
This thesis
*
*
*
* Mutant variants were measured from Bernhard Loll.
3.3.2.1 PurificationoftruncatedceClp1proteinvariants
To functionally and structurally characterize truncation variants of ceClp1, different
constructs were designed and cloned via a standard cloning procedure into bacterial pET21a
and pET28b expression vectors. Resulting sequence-verified clones were transformed into E.
coli BL21-CodonPlus(DE3)-RIL cells allowing for an IPTG-inducible overexpression. The
recombinant protein variants were overexpressed, followed by cell lysis via sonication, and
81
3 Results
then further purified. In contrast to full-length protein, truncated variants appeared to be rather
insoluble.
For NtD truncation variants, two constructs were designed in which the polypeptide
chain was truncated in or after the N-terminal helix α1 (ceClp1ΔN85 and ceClp1ΔN107,
Figure 3.1.1). Whereas purification of ceClp1ΔN85 was unsuccessful due to solubility
problems, purification of ceClp1ΔN107 could be optimized by co-expression with a protein
variant coding for NtD (ceClp1ΔC104). Both domains bound to each other and formed a
complex that could be separated after immobilization on a Talon affinity column. Although
the purification procedure included a single Talon affinity chromatography step, samples were
highly homogeneous as confirmed by SDS-PAGE. Separated proteins were subsequently
tested for 5’-kinase activity, which was measured by an enzymatically coupled spectroscopic
ATPase assay.
Similar to the truncation experiments characterizing the functional importance of the
NtD, three protein variants with a C-terminal deletion were designed. The truncations were
located in helix α7 (ceClp1ΔC288, ceClp1ΔC304, and ceClp1ΔC310). Since only
ceClp1ΔC310 was soluble, this mutated protein variant was purified and 5’-kinase activity
was measured. Furthermore, ceClp1ΔC310 was also co-expressed with a protein variant
coding for the CtD (ceClp1NC315) and the complex was tested for 5’-kinase activity.
3.3.2.2 CharacterizationofthetruncatedceClp1proteinvariants
Truncation of the NtD as well as CtD were tested enzymatically using two different IPTG
inducible truncation variants ceClp1ΔN107 and ceClp1ΔC310 (Figure 3.3.2 A and B). Based
on the ATP-bound crystal structure of ceClp1, specific interactions between the nucleobase
and conserved residues of NtD were identified (Figure 3.1.8). The ATP molecule interacts,
via hydrogen bonds, with the side chain of Glu16 and the main chain carbonyl of Arg56.
Furthermore, the nucleobase forms a stacking interaction with Phe36 of the NtD. In addition,
there is evidence for domain flexibility of the NtD. Sequence alignment of the eukaryotic
Clp1 protein family shows a highly conserved putative hinge region within the loop
connecting NtD with PNK domain Gly114 (Figure 3.1.1). This loop appears to be flexible,
since a poorly defined electron density was observed for the crystal structure. Therefore, it
seemed plausible that deletion of NtD should affect ceClp1 5’-kinase activity. Indeed, as a
consequence of truncation, the ceClp1ΔN107 variant showed an abolished enzymatic activity
82
3 Results
(Figure 3.3.2 C). Moreover, the mutated protein variant ceClp1ΔN107 was prone to
aggregation, probably due to solvent-exposed hydrophobic patches. Interestingly, 5’-kinase
activity of the ceClp1ΔN107 variant was restored by -complementation with the NtD (Figure
3.3.2 A).
In contrast to NtD, residues of CtD do not contribute to the active site of ceClp1
(Figure 3.1.4). However, the CtD and the PNK domain are linked downstream after helix α7.
Helix α7 contains the LID module and a conserved arginine residue (Arg297, Figure 3.1.1).
The LID module has an important role in charge stabilization of the transition state.
Furthermore, Arg293 and Arg297 form hydrogen bonds to the backbone phosphates of the
RNA substrate (Figure 3.1.5). Therefore, a truncation within helix α7 (ceClp1ΔC288) was
expected to affect ceClp1 enzymatic activity, since three highly conserved arginine residues
(Arg288, Arg293, and Arg297) are removed. Indeed, truncation at the end of helix α7
(ceClp1ΔC310) diminished ceClp1 5’-kinase activity (Figure 3.3.2 D). It is likely that the LID
module becomes disordered upon removal of CtD. To restore 5’-kinase activity of
ceClp1ΔC310, the CtD (ceClp1ΔN315) was trans-complemented. However, no kinase
activity could be detected and thus trans-complementation is not sufficient to overcome a
possible distortion of helix α7.
In conclusion, both additional domains affect ceClp1 5’-kinase activity. Whereas NtD
is involved in ATP binding and shows domain flexibility, CtD could have a function in the
correct positioning of the three conserved Arg288, Arg293, and Arg297 that are part of the
LID module.
3.3.3
Summary
In conclusion, the site-directed mutagenesis experiments demonstrate the importance of the
Walker A lysine for enzyme catalysis. Furthermore, the clasp-motif was identified, which is a
novel feature of members of the Clp1 protein family.
To complement these mutagenesis experiments, ceClp1 was further characterized by
deletion of the additional domains NtD and CtD. Based on these experiments, NtD seems
highly flexible and this flexibility might be involved in a regulatory function. So far, the
function of the CtD remained largely elusive. However, initial results suggest an important
role in the positioning of the LID module.
83
3 Results
Figure 3.3.2
2 Characterrization of N-terminal
N
a nd C-termin
nal truncatio
on variants of ceClp1. A and
B, Ribbon d
diagram of th
he different truncation
t
va
ariants used for enzymattic activity. T
The kinase activity
a
he descriptiion for the coupled AT
was monito
ored spectro
oscopically similar to th
TPase assa
ay. All
experimentss were perfo
ormed with 2 μM of prote
ein, 1 mM ATP,
A
and 500
0 µM G1C2. R
Reaction mix
xtures
were equilib
brated at 20 °C for 10 min
m prior to in
ndividual me
easurements. C. A truncaation of ceC
Clp1 to
ceClp1∆N107 abrogates enzymatic activity. How
wever, after long-term incubation of cceClp1∆N10
07 and
ceClp1∆C104 a nucle
eotide kinase
e activity w
was recoverred. The co
ontrol experiiments without a
nucleotide ssubstrate (G1C2) showed
d no a basal ATPase acttivity for the complex of cceClp1∆N10
07 and
ceClp1∆C104. D, Both ceClp1∆C3
310 alone orr co-express
sed with ceC
Clp1∆N314 w
were abolish
hed in
enzymatic a
activity.
84
4 Discussion
4 Discussion
4.1 TheClp1proteinfamily
In the course of this thesis it could be shown that Clp1 orthologs from higher eukaryotes
represent novel types of RPNKs. Based on their structural conservation, Clp1 homologs
define a subfamily of RPNKs, the Clp1 protein family. These RPNPs show broad
phylogenetic distribution and can be found in all three kingdoms of life86-88,95,96. Although
structurally conserved, members of the Clp1 protein family were shown to be involved in
different RNA maturation pathways86-88,90. This functional difference might be due to
additional NtD and CtD flanking the PNK domain. The NtD and CtD are novel structural
elements that are unique to the Clp1 protein family. Truncation experiments with the NtD and
CtD revealed their importance for ceClp1’s enzymatic activity. Based on these observations,
it is conceivable that either the in vivo function or the regulation of enzymatic activity in Clp1
depends on the structural dynamics of the additional domains, which will be discussed in
more detail in the following sections.
4.1.1 SequenceconservationoftheClp1proteinfamily
ceClp1 is a multidomain protein composed of a central PNK domain flanked by additional
NtD and CtD (Figure 3.1.1 and Figure 3.1.4). While hsClp1 contributes to mRNA processing,
tRNA splicing and RNAi88,90,92, the homologous proteins Nol9 and Grc3 were shown to
participate in rRNA maturation86,87,115,116. Based on a sequence alignment, these enzymes
define the Clp1 protein family, owing the highest degree of conservation within the PNK
domain. The PNK domain is subdivided into an ATP- and an RNA-binding site. The
sequence alignment also revealed four conserved structural motifs associated with substrate
binding and enzyme catalysis (Figure 3.1.1 and Figure 4.1.1). The most prominent motifs for
PNK are the P-loop motif GxxxxGK[T/S] and the Walker B motif DxxQ113,114. However,
Clp1 revealed two additional motifs that are only conserved among members of the Clp1
protein family. These motifs are the LID module (RxxxxR) and the clasp motif (TxGW). The
sequence of the central PNK domain seems to be conserved throughout all kingdoms of
life86,88,95,96 (Figure 4.1.1). Jain and Shuman95 identified a Clp1 archaeal homolog from
Pyrococcus horikoshii that displays similar RNA-specificity in vitro as shown for ceClp1
(Figure 3.2.1). Moreover, sequence alignments also suggested a bacterial candidate Clp1
85
4 Discussion
homolog in Nitrosococcus halophilus (nhClp1) Figure (4.1.1). This “GTPase or GTP-binding
protein-like protein” (CP001798.1) shows a high sequence homology to the PNK and the CtD
domain. Interestingly, in both the bacterial (nhClp1) as well as the archaeal Clp1 (phClp1)
homolog, no evidences for an NtD are found. Furthermore, Nol9 and Grc3 display sequence
alterations in their NtD and CtD. On the basis of these observations it can be assumed that the
additional NtD and CtD have a regulatory or a recruiting function important for the different
in vivo functions of the enzymes. Since only Clp1 homologs in eukaryots consist of an NtD, it
is likely that an NtD with a putative regulatory function was acquired during evolution to add
an additional layer of regulation to its 5’-kinase activity.
4.1.2 PutativeregulatoryfunctionoftheadditionalNtDandCtD
With respect to the additional domains, ceClp1 is structurally different from all other
previously described PNKs51,52,69. In contrast to the T4 PNK as well as the mPNK with an
“open” ATP-binding site, the NtD of ceClp1 was covering its active site (Figure 4.1.2). The
crystal structure of ceClp1 revealed interactions between the nucleobase and conserved
residues of the NtD (Glu16, Phe38, and Arg56). Interestingly, the mPNK125 also shows these
kinds of interactions, however, in case of ceClp1, the ATP binding site is completely
obstructed from the NtD (Figure 3.1.4). Truncation experiments of the NtD and the CtD
showed the importance of both domains for Clp1’s 5’-kinase activity (Figure 3.3.2).
Truncation of both domains abolished the enzymatic activity, but only through
trans-complementation of the NtD ceClp1’s enzyme activity could be restored (Figure
3.3.2 C). These experiments suggest a function of the NtD in ATP-binding. This hypothesis is
consistent with the conservation of the ATP-binding site within higher eukaryotes showing
parts of the NtD that interact with ATP to be highly conserved (Figure 4.1.2). The CtD is
distal to the α-helical LID module and truncation of the CtD might affect the alignment of the
catalytically relevant arginine residues of the LID module. Thus, the CtD appeared to be
important for the correct positioning of the LID module. A structural comparison of the NtD
and CtD with structures of the PDB database revealed structural homology of the NtD to the
FHA domain (r.m.s.d: 2 Å) of the mPNK. The FHA domain is involved in factor
recruitment42 by recognition of the phosphorylated forms of the two scaffold proteins XRCC1
and XRCC4. Although both the FHA domain and the NtD were composed of a two β-sheet
sandwich, the topological arrangement of the strands is different (Figure 3.1.4). Moreover, a
conserved loop of the FHA that recognizes the phosphorylated forms of XRCC1 and XRCC4
86
4 Discussionn
is miissing in thee NtD of ceeClp1. This fact suggessts that in contrast to thhe FHA of the mPNK,,
ceClpp1 might be unable to
o recognize phosphoryllated scaffo
old proteins for factor recruitment
r
t
simillar to the observations
o
s of the DN
NA repair433,44. However, interactition studies on scClp1
identtified speciffic protein-p
protein inteeraction parrtners that are
a recognizzed by sing
gle domainss
or at the inter-domain boun
ndary of twoo domains94,98-100. Strik
kingly, som
me of these interactions
i
s
11 are distuurbed by mutations
m
in
n the P-looop motif su
upporting a
such as the onee to scPcf1
strucctural role of
o the P-lo
oop in Clp1198-100. Sim
milar interaction studie s were perrformed forr
hsClpp1 but not in a comp
parable quaality89,91. Th
hese interacction studiess showed that
t
hsClp1
interaacts with Pcf11,
P
a factor
f
of thhe mRNA 3’-end pro
ocessing89. Moreover, it is alsoo
assocciated with a protein complex innvolved in tRNA-spliccing, the T
TSEN comp
plex91. Thiss
obserrvation leaad to the assumption
a
p1 is a reccurrent linkker of diffe
ferent RNA
A
that hsClp
matuuration pathways92. Sin
nce the interraction interrface of Pcff11 from S. cerevisiea and scClp1
8
is knnown and both
b
proteiins are connserved in eukaryotes89
, it was ppossible to model thee
interaaction of ceeClp1 and Pcf11
P
from C
C. elegans (cePcf11)
(
(F
Figure 4.1.33).
87
4 Discussioon
Figure 4.1.1. Sequenc
ce conserv
vation and domain arc
chitecture of
o the Clp11 protein fa
amily.
Members o
of the Clp1 protein
p
family
y are presen
nt in all thre
ee kingdoms of life and show the highest
homology in
n the PNK domain.
d
Seq
quence align ment revealed 4 structu
ural motifs beeing importa
ant for
enzyme catalysis and/o
or substrate binding. Th
he P-loop and Walker B motif show
ow high sequence
conservatio
on, whereas the LID mo
odule and th e clasp are less conserrved throughhout all mem
mbers.
Clp1 homologs are fro
om Caenorhabditis elega
gans (ceClp1
1): NP_0010
040858; from
m Saccharom
myces
cerevisiae (scClp1): NP_014893,
N
from Pyrrococcus ho
orikoshii (phClp1): NP
P_142196.1; from
Nitrosococccus halophilu
us (nhClp1): CP001798.1
1; from homo
o sapiens (N
Nol9): NP_0778930.3, and
d from
Saccharomyyces cerevissiae (Grc3): NP_013065.
N
1.
P
Previously presented
p
structural
s
ddata of thee inactive scClp194 sshowed thaat the
interface bbetween the central PNK
P
domaain and thee CtD is in
nvolved in the bindin
ng of
scPcf1194. scPcf11 toggether with scClp1 are subunits off the cleavage factor CFF IA involv
ved in
y of eukaryyotic Pol II transcriptts102. Basedd on a strucctural
the 3’-endd processingg machinery
model of thhe interactioon interfacee between cceClp1 and cePcf11,
c
it was shownn that the pro
oteinprotein intteractions arre highly conserved inn eukaryotees (Figure 4.1.3).
4
This model pro
ovides
the basis too discuss a CtD-depend
dent regulattion of the enzymatic
e
activity
a
in C
Clp1 by affeecting
the correctt position off the LID module.
m
Figure 4.1.2
2 Conservattion of the active
a
site o
of ceClp1. An
n “Open-Boo
ok view” of ceeClp1 showin
ng the
conserved A
ATP- and RN
NA-binding site.
s
The am ino acid seq
quence conse
ervation (Figgure 3.1.1) of Clp1
from higherr eukaryotes is mapped onto the mo
olecular surface representation and the molecule
e was
cut into tw
wo halves shown
s
side by side. T
The bound AppNHp
A
and RNA aree shown as stick
representation and were
e mapped on
nto the active
e site of ceCllp1 on each half of the m
molecule.
88
4 Discussionn
Figurre 4.1.3 Protein-protein
n interaction
n interface between
b
ceC
Clp1 and ceePcf11. A, Based
B
on the
e
crysta
al structure of
o scClp1 associated with
h scPcf11, cePcf11
c
was superimposeed and modeled into the
e
structture of ceClp
p1. cePcf11 is presented
d as stick mo
odel and the
e amino acidd sequence conservation
c
n
(Figure 3.1.1) of ceClp1
c
and seClp1
s
is ma
apped onto th
he molecularr surface reppresentation. The surface
e
is colored accord
ding to the degree
d
of co
onservation, decreasing from dark ggreen to yelllow. B, The
e
individual domain
ns are color coded simillar to Fig. 3.1.4. C, Amino acid seqquence align
nment of the
e
regions of cePcf11 and scPcf11 interactin g with Clp1.
4.1.3
3 Structuraldiffere
encebetweeenscClp1
1andceClp
p1
Strucctural compparison betw
ween scClpp1 and ceCllp1 revealed
d catalyticaally importaant residuess
that are requireed for ceCllp1 5’-kinasse activity. Whereas hsClp1
h
seeems to be involved
i
inn
NA and tRN
NA processsing88-90,92, scClp1 waas shown to
o be inactivve94. Even though thee
mRN
enzyyme is inacttive, the bin
nding of AT
TP seems to
o be crucial for factor rrecruitmentt during thee
cleavvage and polyadenylattion reactioon of RNA polymerasse II transcr
cripts98-100. The
T crystall
struccture of scC
Clp1 (PDB 2NPI) revveals Gln13
33 (the correspondingg residue in
n ceClp1 iss
Asp1124) to tighhtly interactt with the pphosphate groups
g
of ATP
A
and to distort thee nucleotidee
bindiing site (Fiigure 4.1.4). Site direected mutag
genesis cou
uld show thhat a mutaation of thee
Asp1124 in ceClp1 to Gln abolishes enzymatic activity, whereas
w
muttations to Asn
A or Alaa
show
wed much leess dramaticc effects (Fiigure 3.2.1)). Furthermo
ore, scClp1 lacks a ressidue that iss
impoortant to neeutralize th
he negative charges off the transiition state as shown for ceClp1
(Figuure 4.1.4 A and B). Th
he arginine rresidues of the LID mo
odule (the ccorresponding residuess
in cceClp1 are Arg288 and
a
Arg2993) are no
on-conservatively muttated to Val316
V
andd
conseervatively to
t Lys321. Therefore,, it is sugg
gested that scClp1 hass lost its caapability off
89
4 Discussioon
enzymatic activity duuring evolu
ution, whereeas ATP-biinding by the
t P-loop motif rem
mained
9
important ffor structuraal integrity98-100
.
Figure 4.1.4
4. Comparis
son of the active
a
site re
egion of acttive and inactive Clp1 o
orthologs. A,
A The
inhibited substrate boun
nd state of ceClp1
c
B, scC
Clp1 in complex with AT
TP-Mg2+. Ressidues that render
r
scClp1 inca
apable of AT
TP hydrolysis
s are shown in stick representation. For
F better coomparison, scClp1
s
has an iden
ntical orienta
ation as ceC
Clp1. The co
orresponding
g amino acid
ds in ceClp11 are indicated in
brackets. In
n contrast to ceClp1, scC
Clp1 shows n
no clasp sequence motif. Furthermore
re, the LID module
m
has lost its function in scClp1 due to an excha
ange of Arg2
288 to Val31
16. Gln133 fforms a hyd
drogen
bond to the ATP, which interferes with a correct positioning of
o the ATP.
4.2
RN
NA‐recogn
nitionand
dRNA‐speecificityin
nPNKs
PNKs are kknown to be
b importantt in DNA aand RNA reepair, RNA maturationn as well as RNA
degradationn processees. Based on
o their sppecific funcction, thesee enzymes show diffferent
6
substrates specificity60,86-88,158
. Crystals of cceClp1 boun
nd to RNA provide firsst structural data
R
rate. A com
mparison off ceClp1 w
with DNA bound
b
of a PNK in complexx with an RNA-substr
K (PDB 1R
RRC) and the mPNK (3ZVN) rrevealed a novel
crystal struucture of the T4 PNK
“RNA-sensor” and a clasp-like
c
binding
b
mecchanism. Th
hese differen
nces and thee implicatio
ons of
s
the structurral data for its substratee specificityy are discusssed in this section.
Since hsC
Clp1 was shhown to phosphorylatte in vivo siRNAs bu
ut also tRN
NAs, a detailed
characterizzation of itts substratee specificitty was pro
ovided64,88,900. Howeverr, the strucctural
features invvolved in thhe RNA-speecificity as w
well as in RNA
R
recogn
nition have rremained ellusive
so far. For crystallizattion experim
ments, the un
uncharacterizzed Clp1 orrtholog from
m C. elegans was
used. ceClp1 was iddentified as
a a bona fide RPNK
K showing
g similar inn vitro sub
bstrate
88
t the recen
ntly identifieed RPNKs (hsClp1
(
, Nol9
N 87, Grc3386, and arcchaeal
specificity compared to
Clp195; Figgure 3.2.1).. Interesting
gly, in addiition to theiir similar in
n vitro subsstrate specifficity,
90
4 Discussion
these homologs also show a high sequence homology of the PNK domain (Figure 4.1).
Therefore, it is likely that structural features of the PNK domain are responsible for the
specificity of the 5’-nucleotide substrate in vitro. These structural features mainly depend on
the clasp motif and on the “RNA-sensor” recognizing the 2’-hydroxyl group at the ultimate
position of the RNA.
Similar to previously described PNKs42,51,52, ceClp1 interacts with its RNA substrate
via a combination of several interactions including the backbone phosphates, sugars, and the
bases themselves (Figure 3.1.6). In ceClp1, the ultimate nucleobase stacks on the hydrophobic
side chains of Trp233. This residue is part of the clasp motif. Mutation of Trp233 to alanine
effectively impaired enzyme catalysis (Figure 3.3.1). Even though the clasp motif represents a
novel consensus sequence (TxGW) for RNA-binding, it is highly conserved throughout
higher eukaryotes (Figure 3.1.1). This binding mechanism is similar to other PNKs42,51,52. The
T4 PNK and the mPNK have similar hydrophobic interaction pockets. In contrast to Trp233,
nucleobases are stacked to smaller hydrophobic residues (Val135 and Val477, T4 PNK and
mPNK, respectively42,51,52). Another unique feature of Clp1’s clasp motif is a hydrogen bond
formed by Gly232 to the 5’-hydroxyl group of the RNA substrate. Thereby, the clasp ensures
that the RNA is positioned in a favored 3'-endo form and in anti-conformation (Figure 3.1.5).
Interestingly, despite their similarities in binding of the ultimate nucleobase, ceClp1
shows a major difference in the orientation of the nucleotide, which is inverted compared to
the other PNKs. This observation suggests convergent evolution of Clp1 in the protein family
of PNKs. Despite these differences, the binding mechanism of the 3’-5’ bridging phosphate
groups seems to be conserved within the protein family of PNKs42,51,52. Again, Clp1, T4 PNK,
and mPNK show a similar binding mechanism. In each case the first two 3’-5’ bridging
phosphate groups are recognized by conserved arginine residues (Arg34, Arg38 and Arg395,
Arg432; T4 PNK and mPNK, respectively). The structure described in this thesis shows a
hydrogen bond as well as salt bridges between three conserved residues (Gln154, Arg293, and
Arg297) and the first two 3’-5’ bridging phosphate groups being important for binding
(Figure 3.1.5). The importance was also shown by comprehensive mutational analysis
(Table 3.3.1). In this respect, ceClp1 is similar to the majority of RNA-binding enzymes,
which are interacting more with the bound 3’-5’ bridging phosphate groups and less with the
nucleobase or the nucleosugar159. However, in contrast to the T4 PNK, which was previously
described to phosphorylate 3’-nucleoside monophosphates46, ceClp1 was unable to use this
substrate (Figure 3.2.2). Although structurally similar, ceClp1 imposes different requirements
on minimal substrate length compared to T4 PNK. Comparison of the active site of all three
91
4 Discussion
PNKs reveals slight differences in binding of the first bridging phosphate. T4 PNK126 shows
the tightest interaction with its substrate compared to ceClp1 and mPNK125. Thus, ceClp1 and
mPNK require dinucleotides or oligonucleotides for efficient binding42,85.
As shown by experiments aiming at identification of the preferred substrate
(Figure 3.2.1), Clp1 can discriminate single-stranded RNA against single-stranded DNA88.
The possible explanation for this selectivity of ceClp1 is the main chain carbonyl of Thr191
that forms a hydrogen bond to the ribose 2'-hydroxyl group of the ultimate nucleotide (Figure
3.1.5). In contrast to published crystal structures of the other PNKs42,51,52, Clp1 is the only
PNK that recognizes the 2’-hydroxyl group. Interestingly, Clp1’s RNA-sensor only
recognizes the 2’-hydroxyl group at the ultimate position of the substrate. Phosphorylation
assays showed that replacing the RNA molecule at the ultimate position by a
desoxyribonucleotide abolishes hsClp1’s ability to phosphorylate synthetic substrate in
hsClp188. Strikingly, it was possible to model an ideal double-stranded A-form RNA bound to
ceClp1 without any steric hindrance by superposition with the single-stranded RNA observed
in the crystal structure. This model explains why Clp1 is able to phosphorylate both, single as
well as double-stranded RNAs88. The modeled double-stranded RNA has no steric clashes
when modeled for longer double-stranded RNA substrates (Figure 3.1.6 A). In contrast to the
5’-acceptor strand, the complementary strand shows no contact with the protein, except for
modeled substrates with a 3’-overhang (Figure 3.1.6 B). The continuous base stacking
interaction of the 3’overhang of RNA has to be interrupted to become accommodated in the
RNA-binding site of ceClp1. This observation might explain why double-stranded
3’-overhang RNA and DNA substrates were phosphorylated less efficiently compared to blunt
end substrates. In conclusion, the crystal structures of Clp1 bound to RNA explain the
selectivity for both single-stranded and double-stranded RNA substrates.
4.3
Phosphoryltransferreaction
Even though extensive biochemical and structural studies are available on the phosphoryl
transfer reactions of P-loop kinases, relatively little is known about the functional mechanism
in PNKs. This thesis provides structural data about the first reaction trajectory of PNKs.
Crystal structures of ceClp1 were trapped during enzyme catalysis such as an inhibited
substrate bound state, a transition state analog and a product bound state. By comparing
ceClp1’s active site geometry between ground state and transition state, a general model for
enzyme catalysis in PNKs is derived and will be discussed in detail in this section.
92
4 Discussion
4.3.1 AmolecularmodelofthephosphoryltransferreactioninPNKs
Similar to other P-loop kinases, the active site of ceClp1 is formed by residues of the classical
P-loop motif, a LID module, and a catalytic Mg2+ ion19. Interestingly, although the Walker A
lysine of the P-loop is classically involved to interact in a bifurcated manner with the β- and
γ-phosphate groups of ATP113, ceClp1’s Lys127 was captured outside the active site in both
the inhibited substrate bound state (Figure 3.1.10) and the product bound state (Figure 3.1.12).
In the ground state, Lys127 was “switched-off”, whereas in the crystal structure of the
transition state analog, Lys127 was switched into an active conformation (Figure 3.1.11). The
“switched-on” Lys127 assists in enzyme catalysis by stabilizing the developing negative
charges on the leaving group (3.1.11). The observation of the “arrested” Walker A lysine is in
contrast to the prevailing structural and biochemical dogma underlying the Walker A
lysine113. To exclude the possibility that ceClp1 belongs to the special group of Walker A
lysine independent P-loop kinases160-163, the functional role of Lys127 for enzyme catalysis
was tested by non-conservative and conservative point mutations. Both mutant variants,
K127A and K127R (Table 3.3.1), were abolished in their enzymatic activity similar to the
results of Walker A lysine mutants in hsClp188. This suggests that Clp1 uses a substrate gating
mechanism composed of the catalytically important Lys127, which is switched on and off
during enzyme catalysis. A highly regulated RPNK would have been in accordance with the
biochemical characterization on ceClp1 revealing a quasi-unidirectional 5’-kinase (Figure
3.2.8) that suppresses futile side reactions (Figure 3.2.5 and 3.2.6).
Despite this exclusive substrate-gating mechanism, the overall architecture of
ceClp1’s active site resembles those of previously described P-loop kinases19. The proposed
model of the phosphoryl transfer reaction follows a mechanism of a classical associative
transition state124 (Figure 3.1.11). The enzymatic reaction starts with the nucleophilic attack of
the deprotonated 5’-hydroxyl group. Based on the crystal structure of ceClp1, the 5’-hydroxyl
group of the bound RNA was shown to be in an ideal hydrogen bonding distance to the
catalytic base (Asp151). Asp151 is proposed to abstract the proton of the attacking
5’-hydroxyl group prior to the nucleophilic attack on the -phosphate group of ATP. In fact,
mutation of Asp151 to alanine abrogates the phosphoryl transfer reaction in ceClp1 (Table
3.3.1), supporting the function of Asp151 as general base. This result could also be shown for
the equivalent residue in hsClp197. The function of the general base is highly conserved in
PNKs42,51,52. A comparison of ceClp1 with a selection of other PNKs42,51,52 (Asp35 and
93
4 Discussion
Asp396 in T4 PNK and mPNK, respectively) reveals a highly conserved aspartate residue in
ideal hydrogen bonding distance to the respective position. Additionally, site-directed
mutagenesis showed the catalytical importance of those general bases164. Also, the
5’-hydroxyl group of the incoming nucleophile is further recognized by a hydrogen bond from
the main-chain amide of Gly232 (see 3.1.5), which is part of the RNA-binding clasp. Gly232
might contribute to the correct position of the attacking nucleophile.
On the opposite site of the attacking nucleophile, an ATP molecule is positioned
enabling an efficient in-line phosphoryl transfer reaction. The phosphate groups of the ATP
interact with the main-chain amide of the residues of the P-loop motif. In addition to the
P-loop motif, an α-helical LID module together with a catalytic Mg2+ ion assists in charge
neutralization of the ATP phosphate groups19,113. In the inhibited substrate bound state, two
catalytically important arginine residues Arg288 and Arg293 form salt-bridges between their
guanidinium groups and the oxygen atoms of the β- and γ-phosphate groups. Both arginine
residues are essential for enzyme catalysis. Non-conservative mutations (R288A, R293L) led
to an effectively abolished enzymatic activity, to an extent comparable with the Walker A
lysine mutant K127A (Table 3.3.1). Interestingly, conservative mutation to lysine (R288K
and R293K) revealed functional differences between Arg288 and Arg293. Whereas R288K
appeared to be slightly impaired in kinase activity, R293K was inactivity (Table 3.3.1). Thus,
Arg288 seems to be important for neutralizing the developing negative charges during the
transition state. However, Arg293 additionally contributes to the RNA-binding site by its
guanidinium group interacting in a bidendate mode forming salt-bridges to the phosphate
groups of the ATP as well as the 5’-3’ bridging phosphate of the RNA (Figure 3.1.5).
A LID module with conserved arginine residues is a common feature of the reaction
mechanism in PNKs and P-loop kinases in general19. A structural comparison with other
described PNKs identified similar arginine residues in T4 PNK51,164 (Arg122 Arg126).
Mutagenesis studies on T4 PNK showed the enzymatic relevance of those arginine residues as
it was shown for ceClp1165. Based on the contribution of Arg293 in RNA-binding it is
suggested that this residue also plays a role in RNA release after the phosphoryl transfer
reaction. In the inhibited substrate bound state Arg293 interacts with the phosphate backbone
of the RNA, but in the crystal structures of ceClp1 bound to ADP and G1A2A3A4, Arg293 is
reoriented that breaks up the salt bridge to the RNA. Arg293 is captured by Asp124 as shown
for the product bound state. Both residues function in a latch-catch mechanism166.
Mutagenesis studies on Asp124 showed additionally the enzymatic relevance of the
latch-catch by mutations of Asp124 and Arg293 (Table 3.3.1).
94
4 Discussionn
In the traansition staate, AlF4- is perfectly orientated
o
ap
pical to bothh the oxygeen atoms off
the A
ADP and thhe RNA neccessary for the “in-linee” transfer mechanism
m. Superposition of thee
substtrate-boundd and the traansition staate analog revealed
r
thaat the donorr and accep
ptor oxygenn
atom
ms approachhes each oth
her by 0.9 aand 1.2 Å, respectively (Figure 44.3.1). Similarly to thee
“on-ooff switcheed” Lys127
7, also the LID modu
ule Arg288 and Arg2293 residuees are fullyy
activvated to neuutralize negative charg es of the trransfer reaction. Whilee the lysine switch hass
not bbeen describbed before, arginine sw
witches are reminiscent
r
of the situaation observ
ved in NMP
P
kinasses118. Conssidering the positively charged viccinity that iss necessary to stabilizee charges off
the trransition sttate analog, ceClp1’s pphosphoryl transfer reaaction mechhanism is su
uggested too
follow an assoociative reaactions meechanism124. In concllusion, bassed on thee structurall
or enzyme catalysis is generallyy
conseervation off the PNK domain, ouur proposed model fo
appliicable for PN
NKs from all
a three kinngdoms of liife.
Figurre 4.3.1: ce
eClp1 show
ws local co
onformationa
al changes during thee phosphoryl transferr
reacttion. Cross-eyed stereo-view of the
e inhibited su
ubstrate bou
und state (bllue), the transition state
e
analo
og (red), and
d the RNA re
eleased prod uct bound sttate (yellow). In order to visualize conformationall
chang
ges, crystal structures were
w
superim
mposed and are shown as a stick m
models. A co
omparison off
these
e crystal structures revea
als only mino
or conformational change
es, besides thhe movemen
nt of Lys127,,
Arg28
88, and Arg2
293. Furtherm
more, ATP a nd RNA mov
ve in closer proximity.
p
95
4 Discussion
Table 4.3.1. Table of proteins owing a non-conventional Walker A lysine
residue.
Nr.
PDB
ID
Resolution
[Å]
Protein name
P-loop
Sequence
P-loop
residue
number
and chain
Ligand
Walker A
Lys
interactions
Non-conventional Walker A lysine structures ceClp1 polynucleotide
kinase
2.10
GPTDVGKT
121-128 (A)
ANP
no nucleotide
interactions
ceClp1 polynucleotide
kinase
2.10
GPTDVGKT
121-128 (A)
ADP
no nucleotide
interactions
FeoB iron iransporter
1.50
23-30 (A)
GDP
no nucleotide
interactions
ClpB
2.35
595-602 (A)
ADP
no nucleotide
interactions
Gimap2
1.90
29-36 (A)
GTP
no nucleotide
interactions
Gimap2
2.80
29-36 (A)
GDP
no nucleotide
interactions
Ribosome biogenesis
GTP-binding protein
EngB
1.90
30-37 (A)
GDP
no nucleotide
interactions
3QKT
Rad50 ABC-ATPase
1.90
30-37 (A)
ANP
γ-phosphate
3PXN
Kinesin family member
Kin10/NOD
2.60
87-94 (A)
ADP
no nucleotide
interactions
4AYT
ABC transporter
ABCB10
2.85
527-534 (A)
ACP
β-phosphate
11
3NH9
ABCB6
2.10
GPSGAGKS
623-630 (A)
ATP
β-phosphate
12
2JDI
F1-ATPase
1.90
GDRQTGKT
169-176 (A)
ANP
β-phosphate
2VHJ
P4 protein from
bacteriophage
1.80
130-137 (A)
ADP
no nucleotide
interactions
EngA
2.60
182-189 (A)
GCP
no nucleotide
interactions
3K0S
DNA mismatch repair
protein MutS
2.20
614-621 (A)
ADP
no nucleotide
interactions
3RC3
Helicase Suv3
10-17 (A)
ANP
no nucleotide
interactions
1NKS
Adenylate kinase
8-15 (F)
ADP
no nucleotide
interactions
495-502 (X)
ADP
no nucleotide
interactions
169-176 (A)
ANP
β-phosphate
1
2
3
4
5
6
XXXX
XXXX
3A1S
4FCW
2XTN
2XTO
3PQC
7
8
9
10
13
14
15
16
17
18
19
4DCV
1R6B
2CK3
GCPNVGKT
GPTGVGKT
GKTGTGKS
GKTGTGKS
GRSNVGKS
2.08
2.57
ClpA
2.25
F1-ATPase
1.95
GQNGSGKS
GQTGTGKS
GPSGSGKS
GKGNSGKT
GRPNVGKS
GPNMGGKS
GPTNSGKT
GIPGVGKS
GPTGVGKT
GDRQTGKT
96
4 Discussionn
A
B
Figurre 4.3.2: Crystal
C
stru
uctures of non-conven
ntional and
d conventio
onal Walke
er A lysine
e
resid
dues. Close up view of the active site of diffe
erent P-loop kinases shoowing eitherr A, a non-conve
entional or B,
B a conventtional confor mation of the Walker A-lysine. The W
Walker A lys
sine and the
e
nucle
eotide are represented as
s stick-mode
el, and the P-loop
P
motif as
a ribbon diaagram. The conventional
c
l
Walker A lysine variants
v
are obtained fro
om PDB entrries from 6.2
2.1 (substratee bound statte), whereass
the non-conventio
onal Walker A lysine resid
dues belong to table 4.3.1.
4.3.2
2 Putativefunction
nofthenon
n‐canonica
alWalkerAlysineLyys127
Confformational gating is already
a
know
wn since 19
981167, but a compreheensive charaacterizationn
of gaating mechhanisms in general is still missin
ng. The welll known W
Walker A lysine is ann
unexxpected residdue for such
h a gating m
mechanism.. Apparently
y, prior to th
the phospho
oryl transferr
reacttion, Lys1277 is in a “g
gated” confoormation (F
Figure 3.1.10). The resiidue is arreested by thee
P-looop and more importanttly by the cclasp, which
h is involved in RNA-bbinding. ceC
Clp1 seemss
to im
mpose novell constraintss on ligand binding, wh
hereas other PNKs42,511,164 seem to
o follow thee
classsical reactioon mechaniism of phoosphoryl traansferases. In the trannsition statee complex,,
howeever, the Walker
W
A lysine has to re-orientatte to compeensate emerrging negatiive chargess
durinng the phosphoryl traansfer reactiion. Indeed
d, in the sttructure witth the tran
nsition statee
analoog (ADP-AlF4--G1C2), Lys127 hass moved intto the activee site and innteracts in a bifurcatedd
mannner with ann oxygen attom of the -phosphaate group an
nd a fluorinne atom off the planarr
AlF4-, representiing a transittion state geeometry of a phosphory
yl group beiing transferrred (Figuree
3.1.111). Crystallization arteefacts by thhe use of th
he non-hydrrolysable A
ATP analog (AppNHp))
weree excluded since the ATP structture also showed
s
a Walker
W
A llysine in an
a identicall
“arreested” confoormation (Figure 3.1.8)).
97
4 Discussion
According to the Protein Data Bank, there are a significant number of other P-loop
kinases with similar Walker-A lysine residues in an “arrested” conformation (Table 4.3.1).
However, the relevance of this peculiar conformation of the Walker A lysine remained
undiscussed in the literature until now. A superposition of these crystal structures shows that
the triphosphate groups do not align and that their position is rather randomly distributed
(Figure 4.3.2 A). In contrast, triphosphate groups of cases in which the Walker A lysine
adopts the classical conformation, these phosphate groups are virtually identically oriented
(Figure 4.3.2 B). The eukaryotic Clp1 seems to have evolved a sophisticated mechanism to
repress basal site activities. One of these regulatory mechanisms is the newly identified
Walker A lysine functioning as a molecular door step. Also the latch-catach (Arg293-Asp124)
seems to play an important role in enzyme catalysis. These two active site regulation elements
are in addition to the function of the NtD and CtD that were also schon to affect enzymatic
activity of ceClp1.
4.4 ConclusionsandOutlook
4.4.1
Clp1anoveleukaryotic,RPNK
This work provides new structural, functional and mechanistic insights into the Clp1 protein
family. RPNK represents a novel type of PNK that shows high sequence conservation within
the PNK domain. Based on a sequence alignment, four different consensus sequences were
identified that are characteristic for this protein family. The clasp motif together with the
“RNA-sensor” enables Clp1 to contribute to different RNA maturation pathways as a
recurrent linker. Both single-stranded RNA as well as double stranded DNA and RNA is
accommodated in the RNA-binding site of this protein family86-88,95. Whereas hsClp1 is part
of the mRNA, tRNA and RNAi maturation pathways88,92, Nol987 and Grc386,115,116 are also
involved in rRNA maturation. It could therefore be suggested that the different RNA
maturation pathways are interconnected to a global RNA metabolism controlled by the Clp1
protein family. Currently, knowledge about in vivo substrates of Clp1 is limited90. However,
the structural data of this thesis could provide the framework to understand RNA-recognition
and RNA-specificity in more detail. In combination with a detailed biochemical
characterization, future experiments will unveil novel in vivo substrates of the
RNA-metabolism.
98
4 Discussion
In contrast to previously described PNKs, Clp1 also consists of additional NtD and
CtD which are flanking the PNK domain. The initial characterization of truncation variants
presented in this work suggests that the enzymatic activity of ceClp1 is directly affected by
these additional domains. Future experiments aiming towards a better understanding of the
function of these additional domains will be reasonable. The additional NtD and CtD possibly
serve as a protein-protein interaction interface. This interface may regulate Clp1’s enzymatic
activity by structural dynamics in domain flexibility. Possible strategies to test this
protein-induced regulation include a phosphorylation assay measured in the presence of
putative interaction partners. A possible candidate for initial screening experiments could be
Pcf11, the interaction partner of the CFIIm89.
4.4.2
FunctionaldiversitywithintheClp1proteinfamily
Although ceClp1 and scClp1 show a high sequence similarity, both enzymes are functionally
different. scClp1 is enzymatically inactive94, however, still able to bind ATP and ADP
molecules94,98-100. ATP binding seems to be required for factor recruitment during the
cleavage and polyadenylation reaction of RNA polymerase II transcripts98-100. The crystal
structure of ceClp1 revealed structural features involved in enzyme catalysis. A comparison
between
scClp1
and
ceClp1
shows
that
catalytically
important
residues
are
non-conservatively mutated, the P-loop motif, however, remains unaffected. Apparently,
scClp1 has lost its PNK activity during evolution, whereas structural integrity of scClp1 still
requires ATP-binding98-100. The bioinformatical and structural data suggest catalytically
important residues that are required for active Clp1 homologs. These sequence profiles can be
used to search for other active variants in all three kingdoms of life. A functional comparison
of Clp1 homologs from all kingdoms of life may allow new insights into RNA metabolism
such as underestimated cross-connection between mRNA, tRNA, rRNA, and RNAi.
4.4.3
The central dogma of the Walker A lysine has to be reconsidered
The presented structural work of ceClp1 provides a detailed model of the phosphoryl transfer
reaction mechanism for a RPNK. Based on structural conservation of the PNK domain, this
proposed model is generally applicable for PNKs from all three kingdoms of life.
Interestingly, in ceClp1 the enzymatic activity is regulated by a substrate-gating
99
4 Discussion
mechanism168. The Walker A lysine of ceClp1 functions as a door step that prevents futile
ATP hydrolysis. This observation is consistent with experiments showing that ceClp1 has no
basal ATPase activity and only functions in the forward reaction as a quasi unidirectional
enzyme. Eukaryotic Clp1 seems to have evolved a sophisticated mechanism to repress basal
site activities. Interestingly, a detailed search in the PDB database identified several other
proteins showing a putative molecular switch for the Walker A lysine. In all cases, the Walker
A lysine is in an arrested conformation, however, a corresponding activated structure is
missing. Therefore, it will be an exciting task to further characterize a representative selection
of these proteins. The identification of other Walker A lysine switches would argue for a
novel function of the P-loop motif.
100
4 Discussion
5 Acknowledgement
This work would have been impossible without the continuous support of so many people
during the last years. These people contributed in various ways and I am very grateful for all
the help. I would like to thank…
… my direct supervisor Dr. Anton Meinhart for giving me the opportunity to work on a
challenging and exciting project in his laboratory, for his scientific guidance and for his help
trying to bring me on the right track.
... Prof. Dr. Ilme Schlichting for her willingness of being my first appraiser and for the
continuous support.
… Dr. Bernhard Loll for his advice and guidance in the beginning of the project.
… Dr Tatiana Domratcheva for her chemical expertise and her excellent suggestions.
… Dr. Thomas Barends and Dr. Andreas Winkler for help with crystallographic problems.
... Present and former members of the Meinhart lab for the scientific support and the social
and enjoyable lab atmosphere throughout the years. Thanks to Andrea, Christina, Brad,
Hannes, Iris, Julian, Juliane and Stefano. In particular, I would like to thank Maike Gebhardt
for her great assistance and her kindness.
... Chris Roome for excellent IT service and for spreading good mood very morning.
… Present and former members of the BMM department for the great working atmosphere.
… Stephan and Anikó for their contiuous support with delicious and exotic smoothies.
… Important people outside the RNA-world. Verena, Stefan, Jan, Lena, Tine, and Jannik thank you for providing a buffer to survive the rainy days of research.
At last, I would like to express my deepest gratitude towards my parents, my brothers
and my two nieces Lara and Sueda for their continuous support and their love.
My apologies to all others who are not mentioned by name but contributed in many ways.
101
References
References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Steitz, T. A. DNA polymerases: structural diversity and common mechanisms. The
Journal of biological chemistry 274, 17395-17398 (1999).
Brueckner, F., Ortiz, J. and Cramer, P. A movie of the RNA polymerase nucleotide
addition cycle. Current opinion in structural biology 19, 294-299,
doi:10.1016/j.sbi.2009.04.005 (2009).
Van Rompay, A. R., Johansson, M. and Karlsson, A. Phosphorylation of nucleosides
and nucleoside analogs by mammalian nucleoside monophosphate kinases.
Pharmacology & therapeutics 87, 189-198 (2000).
Richards, J., Liu, Q., Pellegrini, O., Celesnik, H., Yao, S., Bechhofer, D. H., Condon,
C. and Belasco, J. G. An RNA pyrophosphohydrolase triggers 5'-exonucleolytic
degradation of mRNA in Bacillus subtilis. Molecular cell 43, 940-949,
doi:10.1016/j.molcel.2011.07.023 (2011).
Shuman, S. Structure, mechanism, and evolution of the mRNA capping apparatus.
Progress in nucleic acid research and molecular biology 66, 1-40 (2001).
Dorleans, A., Li de la Sierra-Gallay, I., Piton, J., Zig, L., Gilet, L., Putzer, H. and
Condon, C. Molecular basis for the recognition and cleavage of RNA by the
bifunctional 5'-3' exo/endoribonuclease RNase J. Structure 19, 1252-1261,
doi:10.1016/j.str.2011.06.018 (2011).
Jinek, M., Coyle, S. M. and Doudna, J. A. Coupled 5' nucleotide recognition and
processivity in Xrn1-mediated mRNA decay. Molecular cell 41, 600-608,
doi:10.1016/j.molcel.2011.02.004 (2011).
Deana, A., Celesnik, H. and Belasco, J. G. The bacterial enzyme RppH triggers
messenger RNA degradation by 5' pyrophosphate removal. Nature 451, 355-358,
doi:10.1038/nature06475 (2008).
Caponigro, G. and Parker, R. Mechanisms and control of mRNA turnover in
Saccharomyces cerevisiae. Microbiological reviews 60, 233-249 (1996).
Mackie, G. A. Ribonuclease E is a 5'-end-dependent endonuclease. Nature 395, 720723, doi:10.1038/27246 (1998).
Tock, M. R., Walsh, A. P., Carroll, G. and McDowall, K. J. The CafA protein required
for the 5'-maturation of 16 S rRNA is a 5'-end-dependent ribonuclease that has
context-dependent broad sequence specificity. The Journal of biological chemistry
275, 8726-8732 (2000).
Walsh, A. P., Tock, M. R., Mallen, M. H., Kaberdin, V. R., von Gabain, A. and
McDowall, K. J. Cleavage of poly(A) tails on the 3'-end of RNA by ribonuclease E of
Escherichia coli. Nucleic acids research 29, 1864-1871 (2001).
Jiang, X. and Belasco, J. G. Catalytic activation of multimeric RNase E and RNase G
by 5'-monophosphorylated RNA. Proceedings of the National Academy of Sciences of
the United States of America 101, 9211-9216, doi:10.1073/pnas.0401382101 (2004).
Pellegrini, O., Mathy, N., Condon, C. and Benard, L. In vitro assays of 5' to 3'exoribonuclease activity. Methods in enzymology 448, 167-183, doi:10.1016/S00766879(08)02609-8 (2008).
Stevens, A. and Poole, T. L. 5'-exonuclease-2 of Saccharomyces cerevisiae.
Purification and features of ribonuclease activity with comparison to 5'-exonuclease-1.
The Journal of biological chemistry 270, 16063-16069 (1995).
Stevens, A. 5'-exoribonuclease 1: Xrn1. Methods in enzymology 342, 251-259 (2001).
102
References
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Weinfeld, M., Mani, R. S., Abdou, I., Aceytuno, R. D. and Glover, J. N. Tidying up
loose ends: the role of polynucleotide kinase/phosphatase in DNA strand break repair.
Trends in biochemical sciences 36, 262-271, doi:10.1016/j.tibs.2011.01.006 (2011).
Amitsur, M., Levitz, R. and Kaufmann, G. Bacteriophage T4 anticodon nuclease,
polynucleotide kinase and RNA ligase reprocess the host lysine tRNA. The EMBO
journal 6, 2499-2503 (1987).
Leipe, D. D., Koonin, E. V. and Aravind, L. Evolution and classification of P-loop
kinases and related proteins. Journal of molecular biology 333, 781-815 (2003).
Morrison, J. F. and Heyde, E. Enzymic phosphoryl group transfer. Annual review of
biochemistry 41, 29-54, doi:10.1146/annurev.bi.41.070172.000333 (1972).
Wolters, S. and Schumacher, B. Genome maintenance and transcription integrity in
aging and disease. Frontiers in genetics 4, 19, doi:10.3389/fgene.2013.00019 (2013).
Loeb, L. A. and Harris, C. C. Advances in chemical carcinogenesis: a historical review
and prospective. Cancer research 68, 6863-6872, doi:10.1158/0008-5472.CAN-082852 (2008).
Paz-Elizur, T., Sevilya, Z., Leitner-Dagan, Y., Elinger, D., Roisman, L. C. and Livneh,
Z. DNA repair of oxidative DNA damage in human carcinogenesis: potential
application for cancer risk assessment and prevention. Cancer letters 266, 60-72,
doi:10.1016/j.canlet.2008.02.032 (2008).
Madhusudan, S. and Middleton, M. R. The emerging role of DNA repair proteins as
predictive, prognostic and therapeutic targets in cancer. Cancer treatment reviews 31,
603-617, doi:10.1016/j.ctrv.2005.09.006 (2005).
McKinnon, P. J. DNA repair deficiency and neurological disease. Nature reviews.
Neuroscience 10, 100-112, doi:10.1038/nrn2559 (2009).
Hoeijmakers, J. H. DNA damage, aging, and cancer. The New England journal of
medicine 361, 1475-1485, doi:10.1056/NEJMra0804615 (2009).
Burma, S., Chen, B. P. and Chen, D. J. Role of non-homologous end joining (NHEJ)
in maintaining genomic integrity. DNA repair 5, 1042-1048,
doi:10.1016/j.dnarep.2006.05.026 (2006).
Valko, M., Izakovic, M., Mazur, M., Rhodes, C. J. and Telser, J. Role of oxygen
radicals in DNA damage and cancer incidence. Molecular and cellular biochemistry
266, 37-56 (2004).
Ciccia, A. and Elledge, S. J. The DNA damage response: making it safe to play with
knives. Molecular cell 40, 179-204, doi:10.1016/j.molcel.2010.09.019 (2010).
Sclafani, R. A. and Holzen, T. M. Cell cycle regulation of DNA replication. Annual
review of genetics 41, 237-280, doi:10.1146/annurev.genet.41.110306.130308 (2007).
Sherman, M. H., Bassing, C. H. and Teitell, M. A. Regulation of cell differentiation by
the DNA damage response. Trends in cell biology 21, 312-319,
doi:10.1016/j.tcb.2011.01.004 (2011).
Buchko, G. W. and Weinfeld, M. Influence of nitrogen, oxygen, and nitroimidazole
radiosensitizers on DNA damage induced by ionizing radiation. Biochemistry 32,
2186-2193 (1993).
Mills, K. D., Ferguson, D. O. and Alt, F. W. The role of DNA breaks in genomic
instability and tumorigenesis. Immunological reviews 194, 77-95 (2003).
Alagoz, M., Gilbert, D. C., El-Khamisy, S. and Chalmers, A. J. DNA repair and
resistance to topoisomerase I inhibitors: mechanisms, biomarkers and therapeutic
targets. Current medicinal chemistry 19, 3874-3885 (2012).
Shiokawa, D. and Tanuma, S. DLAD, a novel mammalian divalent cation-independent
endonuclease with homology to DNase II. Nucleic acids research 27, 4083-4089
(1999).
103
References
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Harosh, I., Binninger, D. M., Harris, P. V., Mezzina, M. and Boyd, J. B. Mechanism
of action of deoxyribonuclease II from human lymphoblasts. European journal of
biochemistry / FEBS 202, 479-484 (1991).
Pheiffer, B. H. and Zimmerman, S. B. 3'-Phosphatase activity of the DNA kinase from
rat liver. Biochemical and biophysical research communications 109, 1297-1302
(1982).
Habraken, Y. and Verly, W. G. The DNA 3'-phosphatase and 5'-hydroxyl kinase of rat
liver chromatin. FEBS letters 160, 46-50 (1983).
Wiederhold, L., Leppard, J. B., Kedar, P., Karimi-Busheri, F., Rasouli-Nia, A.,
Weinfeld, M., Tomkinson, A. E., Izumi, T., Prasad, R., Wilson, S. H., Mitra, S. and
Hazra, T. K. AP endonuclease-independent DNA base excision repair in human cells.
Molecular cell 15, 209-220, doi:10.1016/j.molcel.2004.06.003 (2004).
Whitehouse, C. J., Taylor, R. M., Thistlethwaite, A., Zhang, H., Karimi-Busheri, F.,
Lasko, D. D., Weinfeld, M. and Caldecott, K. W. XRCC1 stimulates human
polynucleotide kinase activity at damaged DNA termini and accelerates DNA singlestrand break repair. Cell 104, 107-117 (2001).
Chappell, C., Hanakahi, L. A., Karimi-Busheri, F., Weinfeld, M. and West, S. C.
Involvement of human polynucleotide kinase in double-strand break repair by nonhomologous end joining. The EMBO journal 21, 2827-2832,
doi:10.1093/emboj/21.11.2827 (2002).
Bernstein, N. K., Williams, R. S., Rakovszky, M. L., Cui, D., Green, R., KarimiBusheri, F., Mani, R. S., Galicia, S., Koch, C. A., Cass, C. E., Durocher, D., Weinfeld,
M. and Glover, J. N. The molecular architecture of the mammalian DNA repair
enzyme, polynucleotide kinase. Molecular cell 17, 657-670,
doi:10.1016/j.molcel.2005.02.012 (2005).
Loizou, J. I., El-Khamisy, S. F., Zlatanou, A., Moore, D. J., Chan, D. W., Qin, J.,
Sarno, S., Meggio, F., Pinna, L. A. and Caldecott, K. W. The protein kinase CK2
facilitates repair of chromosomal DNA single-strand breaks. Cell 117, 17-28 (2004).
Koch, C. A., Agyei, R., Galicia, S., Metalnikov, P., O'Donnell, P., Starostine, A.,
Weinfeld, M. and Durocher, D. Xrcc4 physically links DNA end processing by
polynucleotide kinase to DNA ligation by DNA ligase IV. The EMBO journal 23,
3874-3885, doi:10.1038/sj.emboj.7600375 (2004).
Ali, A. A., Jukes, R. M., Pearl, L. H. and Oliver, A. W. Specific recognition of a
multiply phosphorylated motif in the DNA repair scaffold XRCC1 by the FHA
domain of human PNK. Nucleic acids research 37, 1701-1712,
doi:10.1093/nar/gkn1086 (2009).
Richardson, C. C. Phosphorylation of nucleic acid by an enzyme from T4
bacteriophage-infected Escherichia coli. Proceedings of the National Academy of
Sciences of the United States of America 54, 158-165 (1965).
Novogrodsky, A., Tal, M., Traub, A. and Hurwitz, J. The enzymatic phosphorylation
of ribonucleic acid and deoxyribonucleic acid. II. Further properties of the 5'-hydroxyl
polynucleotide kinase. The Journal of biological chemistry 241, 2933-2943 (1966).
Novogrodsky, A. and Hurwitz, J. The enzymatic phosphorylation of ribonucleic acid
and deoxyribonucleic acid. I. Phosphorylation at 5'-hydroxyl termini. The Journal of
biological chemistry 241, 2923-2932 (1966).
Berkner, K. L. and Folk, W. R. Polynucleotide kinase exchange as an assay for class II
restriction endonucleases. Methods in enzymology 65, 28-36 (1980).
Chaconas, G. and van de Sande, J. H. 5'-32P labeling of RNA and DNA restriction
fragments. Methods in enzymology 65, 75-85 (1980).
104
References
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
Galburt, E. A., Pelletier, J., Wilson, G. and Stoddard, B. L. Structure of a tRNA repair
enzyme and molecular biology workhorse: T4 polynucleotide kinase. Structure 10,
1249-1260 (2002).
Wang, L. K., Lima, C. D. and Shuman, S. Structure and mechanism of T4
polynucleotide kinase: an RNA repair enzyme. The EMBO journal 21, 3873-3880,
doi:10.1093/emboj/cdf397 (2002).
Silber, R., Malathi, V. G. and Hurwitz, J. Purification and properties of bacteriophage
T4-induced RNA ligase. Proceedings of the National Academy of Sciences of the
United States of America 69, 3009-3013 (1972).
Tyndall, C., Meister, J. and Bickle, T. A. The Escherichia coli prr region encodes a
functional type IC DNA restriction system closely integrated with an anticodon
nuclease gene. Journal of molecular biology 237, 266-274 (1994).
Levitz, R., Chapman, D., Amitsur, M., Green, R., Snyder, L. and Kaufmann, G. The
optional E. coli prr locus encodes a latent form of phage T4-induced anticodon
nuclease. The EMBO journal 9, 1383-1389 (1990).
Penner, M., Morad, I., Snyder, L. and Kaufmann, G. Phage T4-coded Stp: doubleedged effector of coupled DNA and tRNA-restriction systems. Journal of molecular
biology 249, 857-868, doi:10.1006/jmbi.1995.0343 (1995).
Cameron, V. and Uhlenbeck, O. C. 3'-Phosphatase activity in T4 polynucleotide
kinase. Biochemistry 16, 5120-5126 (1977).
Sirotkin, K., Cooley, W., Runnels, J. and Snyder, L. R. A role in true-late gene
expression for the T4 bacteriophage 5' polynucleotide kinase 3' phosphatase. Journal
of molecular biology 123, 221-233 (1978).
David, M., Borasio, G. D. and Kaufmann, G. T4 bacteriophage-coded polynucleotide
kinase and RNA ligase are involved in host tRNA alteration or repair. Virology 123,
480-483 (1982).
Lillehaug, J. R., Kleppe, R. K. and Kleppe, K. Phosphorylation of double-stranded
DNAs by T4 polynucleotide kinase. Biochemistry 15, 1858-1865 (1976).
Lillehaug, J. R. and Kleppe, K. Kinetics and specificity of T4 polynucleotide kinase.
Biochemistry 14, 1221-1225 (1975).
van de Sande, J. H. and Bilsker, M. Phosphorylation of N-protected
deoxyoligonucleotides by T4 polynucleotide kinase. Biochemistry 12, 5056-5062
(1973).
Fontanel, M. L., Bazin, H. and Teoule, R. Sterical recognition by T4 polynucleotide
kinase of non-nucleosidic moieties 5'-attached to oligonucleotides. Nucleic acids
research 22, 2022-2027 (1994).
Chan, C. M., Zhou, C. and Huang, R. H. Reconstituting bacterial RNA repair and
modification in vitro. Science 326, 247, doi:10.1126/science.1179480 (2009).
Pedersen, K., Zavialov, A. V., Pavlov, M. Y., Elf, J., Gerdes, K. and Ehrenberg, M.
The bacterial toxin RelE displays codon-specific cleavage of mRNAs in the ribosomal
A site. Cell 112, 131-140 (2003).
Neubauer, C., Gao, Y. G., Andersen, K. R., Dunham, C. M., Kelley, A. C., Hentschel,
J., Gerdes, K., Ramakrishnan, V. and Brodersen, D. E. The structural basis for mRNA
recognition and cleavage by the ribosome-dependent endonuclease RelE. Cell 139,
1084-1095, doi:10.1016/j.cell.2009.11.015 (2009).
Winther, K. S. and Gerdes, K. Enteric virulence associated protein VapC inhibits
translation by cleavage of initiator tRNA. Proceedings of the National Academy of
Sciences of the United States of America 108, 7403-7407,
doi:10.1073/pnas.1019587108 (2011).
105
References
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
Wang, P., Chan, C. M., Christensen, D., Zhang, C., Selvadurai, K. and Huang, R. H.
Molecular basis of bacterial protein Hen1 activating the ligase activity of bacterial
protein Pnkp for RNA repair. Proceedings of the National Academy of Sciences of the
United States of America 109, 13248-13253, doi:10.1073/pnas.1209805109 (2012).
Wang, L. K., Das, U., Smith, P. and Shuman, S. Structure and mechanism of the
polynucleotide kinase component of the bacterial Pnkp-Hen1 RNA repair system.
RNA 18, 2277-2286, doi:10.1261/rna.036061.112 (2012).
Martins, A. and Shuman, S. An end-healing enzyme from Clostridium thermocellum
with 5' kinase, 2',3' phosphatase, and adenylyltransferase activities. RNA 11, 12711280, doi:10.1261/rna.2690505 (2005).
Popow, J., Schleiffer, A. and Martinez, J. Diversity and roles of (t)RNA ligases.
Cellular and molecular life sciences : CMLS 69, 2657-2670, doi:10.1007/s00018-0120944-2 (2012).
Chan, P. P. and Lowe, T. M. GtRNAdb: a database of transfer RNA genes detected in
genomic sequence. Nucleic acids research 37, D93-97, doi:10.1093/nar/gkn787
(2009).
Abelson, J., Trotta, C. R. and Li, H. tRNA splicing. The Journal of biological
chemistry 273, 12685-12688 (1998).
Greer, C. L., Peebles, C. L., Gegenheimer, P. and Abelson, J. Mechanism of action of
a yeast RNA ligase in tRNA splicing. Cell 32, 537-546 (1983).
Sawaya, R., Schwer, B. and Shuman, S. Genetic and biochemical analysis of the
functional domains of yeast tRNA ligase. The Journal of biological chemistry 278,
43928-43938, doi:10.1074/jbc.M307839200 (2003).
Wang, L. K., Schwer, B., Englert, M., Beier, H. and Shuman, S. Structure-function
analysis of the kinase-CPD domain of yeast tRNA ligase (Trl1) and requirements for
complementation of tRNA splicing by a plant Trl1 homolog. Nucleic acids research
34, 517-527, doi:10.1093/nar/gkj441 (2006).
Coller, J. and Parker, R. Eukaryotic mRNA decapping. Annual review of biochemistry
73, 861-890, doi:10.1146/annurev.biochem.73.011303.074032 (2004).
Franks, T. M. and Lykke-Andersen, J. The control of mRNA decapping and P-body
formation. Molecular cell 32, 605-615, doi:10.1016/j.molcel.2008.11.001 (2008).
Houseley, J. and Tollervey, D. The many pathways of RNA degradation. Cell 136,
763-776, doi:10.1016/j.cell.2009.01.019 (2009).
Parker, R. and Song, H. The enzymes and control of eukaryotic mRNA turnover.
Nature structural & molecular biology 11, 121-127, doi:10.1038/nsmb724 (2004).
Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E. and Mello, C. C.
Potent and specific genetic interference by double-stranded RNA in Caenorhabditis
elegans. Nature 391, 806-811, doi:10.1038/35888 (1998).
Ma, J. B., Yuan, Y. R., Meister, G., Pei, Y., Tuschl, T. and Patel, D. J. Structural basis
for 5'-end-specific recognition of guide RNA by the A. fulgidus Piwi protein. Nature
434, 666-670, doi:10.1038/nature03514 (2005).
Park, J. E., Heo, I., Tian, Y., Simanshu, D. K., Chang, H., Jee, D., Patel, D. J. and
Kim, V. N. Dicer recognizes the 5' end of RNA for efficient and accurate processing.
Nature 475, 201-205, doi:10.1038/nature10198 (2011).
Martinez, J. and Tuschl, T. RISC is a 5' phosphomonoester-producing RNA
endonuclease. Genes & development 18, 975-980, doi:10.1101/gad.1187904 (2004).
Shuman, S. and Hurwitz, J. 5'-Hydroxyl polyribonucleotide kinase from HeLa cell
nuclei. Purification and properties. The Journal of biological chemistry 254, 1039610404 (1979).
106
References
86
87
88
89
90
91
92
93
94
95
96
97
98
99
Braglia, P., Heindl, K., Schleiffer, A., Martinez, J. and Proudfoot, N. J. Role of the
RNA/DNA kinase Grc3 in transcription termination by RNA polymerase I. EMBO
reports 11, 758-764, doi:10.1038/embor.2010.130 (2010).
Heindl, K. and Martinez, J. Nol9 is a novel polynucleotide 5'-kinase involved in
ribosomal RNA processing. The EMBO journal 29, 4161-4171,
doi:10.1038/emboj.2010.275 (2010).
Weitzer, S. and Martinez, J. The human RNA kinase hClp1 is active on 3' transfer
RNA exons and short interfering RNAs. Nature 447, 222-226,
doi:10.1038/nature05777 (2007).
de Vries, H., Ruegsegger, U., Hubner, W., Friedlein, A., Langen, H. and Keller, W.
Human pre-mRNA cleavage factor II(m) contains homologs of yeast proteins and
bridges two other cleavage factors. The EMBO journal 19, 5895-5904,
doi:10.1093/emboj/19.21.5895 (2000).
Hanada, T., Weitzer, S., Mair, B., Bernreuther, C., Wainger, B. J., Ichida, J., Hanada,
R., Orthofer, M., Cronin, S. J., Komnenovic, V., Minis, A., Sato, F., Mimata, H.,
Yoshimura, A., Tamir, I., Rainer, J., Kofler, R., Yaron, A., Eggan, K. C., Woolf, C. J.,
Glatzel, M., Herbst, R., Martinez, J. and Penninger, J. M. CLP1 links tRNA
metabolism to progressive motor-neuron loss. Nature 495, 474-480,
doi:10.1038/nature11923 (2013).
Paushkin, S. V., Patel, M., Furia, B. S., Peltz, S. W. and Trotta, C. R. Identification of
a human endonuclease complex reveals a link between tRNA splicing and pre-mRNA
3' end formation. Cell 117, 311-321 (2004).
Weitzer, S. and Martinez, J. hClp1: a novel kinase revitalizes RNA metabolism. Cell
Cycle 6, 2133-2137 (2007).
Kenski, D. M., Cooper, A. J., Li, J. J., Willingham, A. T., Haringsma, H. J., Young, T.
A., Kuklin, N. A., Jones, J. J., Cancilla, M. T., McMasters, D. R., Mathur, M., Sachs,
A. B. and Flanagan, W. M. Analysis of acyclic nucleoside modifications in siRNAs
finds sensitivity at position 1 that is restored by 5'-terminal phosphorylation both in
vitro and in vivo. Nucleic acids research 38, 660-671, doi:10.1093/nar/gkp913 (2010).
Noble, C. G., Beuth, B. and Taylor, I. A. Structure of a nucleotide-bound Clp1-Pcf11
polyadenylation factor. Nucleic acids research 35, 87-99, doi:10.1093/nar/gkl1010
(2007).
Jain, R. and Shuman, S. Characterization of a thermostable archaeal polynucleotide
kinase homologous to human Clp1. RNA 15, 923-931, doi:10.1261/rna.1492809
(2009).
Xing, D., Zhao, H. and Li, Q. Q. Arabidopsis CLP1-SIMILAR PROTEIN3, an
ortholog of human polyadenylation factor CLP1, functions in gametophyte, embryo,
and postembryonic development. Plant physiology 148, 2059-2069,
doi:10.1104/pp.108.129817 (2008).
Ramirez, A., Shuman, S. and Schwer, B. Human RNA 5'-kinase (hClp1) can function
as a tRNA splicing enzyme in vivo. RNA 14, 1737-1745, doi:10.1261/rna.1142908
(2008).
Ghazy, M. A., Gordon, J. M., Lee, S. D., Singh, B. N., Bohm, A., Hampsey, M. and
Moore, C. The interaction of Pcf11 and Clp1 is needed for mRNA 3'-end formation
and is modulated by amino acids in the ATP-binding site. Nucleic acids research 40,
1214-1225, doi:10.1093/nar/gkr801 (2012).
Haddad, R., Maurice, F., Viphakone, N., Voisinet-Hakil, F., Fribourg, S. and
Minvielle-Sebastia, L. An essential role for Clp1 in assembly of polyadenylation
complex CF IA and Pol II transcription termination. Nucleic acids research 40, 12261239, doi:10.1093/nar/gkr800 (2012).
107
References
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
Holbein, S., Scola, S., Loll, B., Dichtl, B. S., Hubner, W., Meinhart, A. and Dichtl, B.
The P-loop domain of yeast Clp1 mediates interactions between CF IA and CPF
factors in pre-mRNA 3' end formation. PloS one 6, e29139,
doi:10.1371/journal.pone.0029139 (2011).
Gross, S. and Moore, C. Five subunits are required for reconstitution of the cleavage
and polyadenylation activities of Saccharomyces cerevisiae cleavage factor I.
Proceedings of the National Academy of Sciences of the United States of America 98,
6080-6085, doi:10.1073/pnas.101046598 (2001).
Minvielle-Sebastia, L., Preker, P. J., Wiederkehr, T., Strahm, Y. and Keller, W. The
major yeast poly(A)-binding protein is associated with cleavage factor IA and
functions in premessenger RNA 3'-end formation. Proceedings of the National
Academy of Sciences of the United States of America 94, 7897-7902 (1997).
Proudfoot, N. New perspectives on connecting messenger RNA 3' end formation to
transcription. Current opinion in cell biology 16, 272-278,
doi:10.1016/j.ceb.2004.03.007 (2004).
Kim, M., Krogan, N. J., Vasiljeva, L., Rando, O. J., Nedea, E., Greenblatt, J. F. and
Buratowski, S. The yeast Rat1 exonuclease promotes transcription termination by
RNA polymerase II. Nature 432, 517-522, doi:10.1038/nature03041 (2004).
Moore, C. L., Skolnik-David, H. and Sharp, P. A. Analysis of RNA cleavage at the
adenovirus-2 L3 polyadenylation site. The EMBO journal 5, 1929-1938 (1986).
Sheets, M. D., Stephenson, P. and Wickens, M. P. Products of in vitro cleavage and
polyadenylation of simian virus 40 late pre-mRNAs. Molecular and cellular biology
7, 1518-1529 (1987).
Popow, J., Englert, M., Weitzer, S., Schleiffer, A., Mierzwa, B., Mechtler, K.,
Trowitzsch, S., Will, C. L., Luhrmann, R., Soll, D. and Martinez, J. HSPC117 is the
essential subunit of a human tRNA splicing ligase complex. Science 331, 760-764,
doi:10.1126/science.1197847 (2011).
Kurihara, T., Fowler, A. V. and Takahashi, Y. cDNA cloning and amino acid
sequence of bovine brain 2',3'-cyclic-nucleotide 3'-phosphodiesterase. The Journal of
biological chemistry 262, 3256-3261 (1987).
Spinelli, S. L., Malik, H. S., Consaul, S. A. and Phizicky, E. M. A functional homolog
of a yeast tRNA splicing enzyme is conserved in higher eukaryotes and in Escherichia
coli. Proceedings of the National Academy of Sciences of the United States of America
95, 14136-14141 (1998).
Englert, M., Sheppard, K., Gundllapalli, S., Beier, H. and Soll, D. Branchiostoma
floridae has separate healing and sealing enzymes for 5'-phosphate RNA ligation.
Proceedings of the National Academy of Sciences of the United States of America 107,
16834-16839, doi:10.1073/pnas.1011703107 (2010).
Hammond, S. M., Bernstein, E., Beach, D. and Hannon, G. J. An RNA-directed
nuclease mediates post-transcriptional gene silencing in Drosophila cells. Nature 404,
293-296, doi:10.1038/35005107 (2000).
Nykanen, A., Haley, B. and Zamore, P. D. ATP requirements and small interfering
RNA structure in the RNA interference pathway. Cell 107, 309-321 (2001).
Saraste, M., Sibbald, P. R. and Wittinghofer, A. The P-loop--a common motif in ATPand GTP-binding proteins. Trends in biochemical sciences 15, 430-434 (1990).
Walker, J. E., Saraste, M., Runswick, M. J. and Gay, N. J. Distantly related sequences
in the alpha- and beta-subunits of ATP synthase, myosin, kinases and other ATPrequiring enzymes and a common nucleotide binding fold. The EMBO journal 1, 945951 (1982).
108
References
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Castle, C. D., Sardana, R., Dandekar, V., Borgianini, V., Johnson, A. W. and
Denicourt, C. Las1 interacts with Grc3 polynucleotide kinase and is required for
ribosome synthesis in Saccharomyces cerevisiae. Nucleic acids research 41, 11351150, doi:10.1093/nar/gks1086 (2013).
Kitano, E., Hayashi, A., Kanai, D., Shinmyozu, K. and Nakayama, J. Roles of fission
yeast Grc3 protein in ribosomal RNA processing and heterochromatic gene silencing.
The Journal of biological chemistry 286, 15391-15402, doi:10.1074/jbc.M110.201343
(2011).
Knowles, J. R. Enzyme-catalyzed phosphoryl transfer reactions. Annual review of
biochemistry 49, 877-919, doi:10.1146/annurev.bi.49.070180.004305 (1980).
Schlichting, I. and Reinstein, J. Structures of active conformations of UMP kinase
from Dictyostelium discoideum suggest phosphoryl transfer is associative.
Biochemistry 36, 9290-9296, doi:10.1021/bi970974c (1997).
Wittinghofer, A. Phosphoryl transfer in Ras proteins, conclusive or elusive? Trends in
biochemical sciences 31, 20-23, doi:10.1016/j.tibs.2005.11.012 (2006).
Lassila, J. K., Zalatan, J. G. and Herschlag, D. Biological phosphoryl-transfer
reactions: understanding mechanism and catalysis. Annual review of biochemistry 80,
669-702, doi:10.1146/annurev-biochem-060409-092741 (2011).
Wittinghofer, A. Signaling mechanistics: aluminum fluoride for molecule of the year.
Current biology : CB 7, R682-685 (1997).
Antonny, B. and Chabre, M. Characterization of the aluminum and beryllium fluoride
species which activate transducin. Analysis of the binding and dissociation kinetics.
The Journal of biological chemistry 267, 6710-6718 (1992).
Goodno, C. C. and Taylor, E. W. Inhibition of actomyosin ATPase by vanadate.
Proceedings of the National Academy of Sciences of the United States of America 79,
21-25 (1982).
Matte, A., Tari, L. W. and Delbaere, L. T. How do kinases transfer phosphoryl
groups? Structure 6, 413-419 (1998).
Garces, F., Pearl, L. H. and Oliver, A. W. The structural basis for substrate recognition
by mammalian polynucleotide kinase 3' phosphatase. Molecular cell 44, 385-396,
doi:10.1016/j.molcel.2011.08.036 (2011).
Eastberg, J. H., Pelletier, J. and Stoddard, B. L. Recognition of DNA substrates by T4
bacteriophage polynucleotide kinase. Nucleic acids research 32, 653-660,
doi:10.1093/nar/gkh212 (2004).
Van Duyne, G. D., Standaert, R. F., Karplus, P. A., Schreiber, S. L. and Clardy, J.
Atomic structure of FKBP-FK506, an immunophilin-immunosuppressant complex.
Science 252, 839-842 (1991).
Taylor, R. G., Walker, D. C. and McInnes, R. R. E. coli host strains significantly
affect the quality of small scale plasmid DNA preparations used for sequencing.
Nucleic acids research 21, 1677-1678 (1993).
Hanahan, D. Studies on transformation of Escherichia coli with plasmids. Journal of
molecular biology 166, 557-580 (1983).
Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis,
K. B. and Erlich, H. A. Primer-directed enzymatic amplification of DNA with a
thermostable DNA polymerase. Science 239, 487-491 (1988).
Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature 227, 680-685 (1970).
Lienhard, G. E. and Secemski, II. P 1 ,P 5 -Di(adenosine-5')pentaphosphate, a potent
multisubstrate inhibitor of adenylate kinase. The Journal of biological chemistry 248,
1121-1123 (1973).
109
References
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
Kabsch, W. Xds. Acta crystallographica. Section D, Biological crystallography 66,
125-132, doi:10.1107/S0907444909047337 (2010).
Sheldrick, G. M. A short history of SHELX. Acta crystallographica. Section A,
Foundations of crystallography 64, 112-122, doi:10.1107/S0108767307043930
(2008).
Emsley, P. and Cowtan, K. Coot: model-building tools for molecular graphics. Acta
crystallographica. Section D, Biological crystallography 60, 2126-2132,
doi:10.1107/S0907444904019158 (2004).
Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., GrosseKunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J.,
Rice, L. M., Simonson, T. and Warren, G. L. Crystallography & NMR system: A new
software suite for macromolecular structure determination. Acta crystallographica.
Section D, Biological crystallography 54, 905-921 (1998).
Murshudov, G. N., Vagin, A. A. and Dodson, E. J. Refinement of macromolecular
structures by the maximum-likelihood method. Acta crystallographica. Section D,
Biological crystallography 53, 240-255, doi:10.1107/S0907444996012255 (1997).
Winn, M. D., Isupov, M. N. and Murshudov, G. N. Use of TLS parameters to model
anisotropic displacements in macromolecular refinement. Acta crystallographica.
Section D, Biological crystallography 57, 122-133 (2001).
Lovell, S. C., Davis, I. W., Arendall, W. B., 3rd, de Bakker, P. I., Word, J. M., Prisant,
M. G., Richardson, J. S. and Richardson, D. C. Structure validation by Calpha
geometry: phi,psi and Cbeta deviation. Proteins 50, 437-450, doi:10.1002/prot.10286
(2003).
Laskowski, R. A., MacArthur, M. W., Moss, D. S. and Thornton, J. M. PROCHECK:
a program to check the stereochemical quality of protein structures. J. Appl. Cryst.,
283-291 (1993).
DeLano, W. L. The PyMOL Molecular Graphics System. Palo Alto, CA, USA. (2002).
Baker, N. A., Sept, D., Joseph, S., Holst, M. J. and McCammon, J. A. Electrostatics of
nanosystems: application to microtubules and the ribosome. Proceedings of the
National Academy of Sciences of the United States of America 98, 10037-10041,
doi:10.1073/pnas.181342398 (2001).
Barton, G. J. ALSCRIPT: a tool to format multiple sequence alignments. Protein
engineering 6, 37-40 (1993).
Livingstone, C. D. and Barton, G. J. Protein sequence alignments: a strategy for the
hierarchical analysis of residue conservation. Computer applications in the
biosciences : CABIOS 9, 745-756 (1993).
Kabsch, W. A solution for the best rotation to relate two sets of vectors. Acta
crystallographica. Section D, Biological crystallography A32, 922–923 (1997).
Schlichting, I. and Reinstein, J. pH influences fluoride coordination number of the
AlFx phosphoryl transfer transition state analog. Nature structural biology 6, 721-723,
doi:10.1038/11485 (1999).
Baxter, N. J., Blackburn, G. M., Marston, J. P., Hounslow, A. M., Cliff, M. J., Bermel,
W., Williams, N. H., Hollfelder, F., Wemmer, D. E. and Waltho, J. P. Anionic charge
is prioritized over geometry in aluminum and magnesium fluoride transition state
analogs of phosphoryl transfer enzymes. Journal of the American Chemical Society
130, 3952-3958, doi:10.1021/ja078000n (2008).
Hendrickson, W. A., Smith, J. L. and Sheriff, S. Direct phase determination based on
anomalous scattering. Methods in enzymology 115, 41-55 (1985).
Schneider, T. R. and Sheldrick, G. M. Substructure solution with SHELXD. Acta
crystallographica. Section D, Biological crystallography 58, 1772-1779 (2002).
110
References
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
Pape, T. and Schneider, T. R. HKL2MAP: a graphical user interface for phasing with
SHELX programs. J. Appl. Cryst. 37, 843-844 (2004).
Rao, S. T. and Rossmann, M. G. Comparison of super-secondary structures in
proteins. Journal of molecular biology 76, 241-256 (1973).
Holm, L. and Sander, C. Dali/FSSP classification of three-dimensional protein folds.
Nucleic acids research 25, 231-234 (1997).
Gawronski-Salerno, J. and Freymann, D. M. Structure of the GMPPNP-stabilized NG
domain complex of the SRP GTPases Ffh and FtsY. Journal of structural biology 158,
122-128, doi:10.1016/j.jsb.2006.10.025 (2007).
Cohen, R. L., Espelin, C. W., De Wulf, P., Sorger, P. K., Harrison, S. C. and Simons,
K. T. Structural and functional dissection of Mif2p, a conserved DNA-binding
kinetochore protein. Molecular biology of the cell 19, 4480-4491,
doi:10.1091/mbc.E08-03-0297 (2008).
Arsenieva, D., Symersky, J., Wang, Y., Pagadala, V. and Mueller, D. M. Crystal
structures of mutant forms of the yeast F1 ATPase reveal two modes of uncoupling.
The Journal of biological chemistry 285, 36561-36569, doi:10.1074/jbc.M110.174383
(2010).
Berkner, K. L. and Folk, W. R. Polynucleotide kinase exchange reaction: quantitave
assay for restriction endonuclease-generated 5'-phosphoroyl termini in DNA. The
Journal of biological chemistry 252, 3176-3184 (1977).
van de Sande, J. H., Kleppe, K. and Khorana, H. G. Reversal of bacteriophage T4
induced polynucleotide kinase action. Biochemistry 12, 5050-5055 (1973).
Karimi-Busheri, F., Lee, J., Tomkinson, A. E. and Weinfeld, M. Repair of DNA strand
gaps and nicks containing 3'-phosphate and 5'-hydroxyl termini by purified
mammalian enzymes. Nucleic acids research 26, 4395-4400 (1998).
Bahadur, R. P., Zacharias, M. and Janin, J. Dissecting protein-RNA recognition sites.
Nucleic acids research 36, 2705-2716, doi:10.1093/nar/gkn102 (2008).
Rusnak, P., Haney, P. and Konisky, J. The adenylate kinases from a mesophilic and
three thermophilic methanogenic members of the Archaea. Journal of bacteriology
177, 2977-2981 (1995).
Kath, T., Schmid, R. and Schafer, G. Identification, cloning, and expression of the
gene for adenylate kinase from the thermoacidophilic archaebacterium Sulfolobus
acidocaldarius. Archives of biochemistry and biophysics 307, 405-410,
doi:10.1006/abbi.1993.1607 (1993).
Lacher, K. and Schafer, G. Archaebacterial adenylate kinase from the
thermoacidophile Sulfolobus acidocaldarius: purification, characterization, and partial
sequence. Archives of biochemistry and biophysics 302, 391-397,
doi:10.1006/abbi.1993.1229 (1993).
Jha, S., Karnani, N., Dhar, S. K., Mukhopadhayay, K., Shukla, S., Saini, P.,
Mukhopadhayay, G. and Prasad, R. Purification and characterization of the N-terminal
nucleotide binding domain of an ABC drug transporter of Candida albicans:
uncommon cysteine 193 of Walker A is critical for ATP hydrolysis. Biochemistry 42,
10822-10832, doi:10.1021/bi0345900 (2003).
Wang, L. K. and Shuman, S. Domain structure and mutational analysis of T4
polynucleotide kinase. The Journal of biological chemistry 276, 26868-26874,
doi:10.1074/jbc.M103663200 (2001).
Wang, L. K. and Shuman, S. Mutational analysis defines the 5'-kinase and 3'phosphatase active sites of T4 polynucleotide kinase. Nucleic acids research 30, 10731080 (2002).
111
References
166
167
168
Vale, R. D. Switches, latches, and amplifiers: common themes of G proteins and
molecular motors. The Journal of cell biology 135, 291-302 (1996).
McCammon, J. A. and Northrup, S. H. Gated binding of ligands to proteins. Nature
293, 316-317 (1981).
Gora, A., Brezovsky, J. and Damborsky, J. Gates of Enzymes. Chemical reviews,
doi:10.1021/cr300384w (2013).
112
Appendix
6 Appendix
6.1 AppendixoftheMaterialsandMethodssection
6.1.1 Listofprimers
Cloning Primers
Primer name
Sequence [5’→3’]
ceClp1-pET21a-For
CGA AGC ATC ATA TGA GCG AGG AGA ATG
TTC
ceClp1-pET21a-Rev
GAG TGC GGC CGC TCG TTT TAT TTG ATC AT
ceClp1-pET28b-For
CG AAG CAT CAT ATG AGC GAG GAG AAT GTT
C
ceClp1-pET28b-Rev
TGC GGC CGC TCA TCG TTT TAT TTG ATC ATC
ceClp1ΔN314-pET21a_For
GCA TTA GCC ATA TGC CAT TCA CAT TTG ACG
Quikchange Primers
Primer name
Sequence [5’→3’]
ceClp1ΔN107-pET21a(ΔHis)-For
ATC TTG ATG ATC AAA TAA AAC GAT AGC TTG
CGG CCG C
ceClp1ΔN107-pET21a(ΔHis)-Rev GCG GCC GCA AGC TAT CGT TTT ATT TGA TCA
TCA AGA T
ceClp1ΔC104-pET28b-For
GAA GAA GAG AGA GGA ACA GTA GGC TGG
AAA CTC GAA TAA G
ceClp1ΔC104-pET28b-Rev
CTT ATT CGA GTT TCC AGC CTA CTG TTC CTC
TCT CTT CTT C
ceClp1ΔC310-pET28b-For
TCT ACG GAA CCC GTG CCT AGA ATC TCT ACC
CAT TCA C
ceClp1ΔC310-pET28b-Rev
GTG AAT GGG TAG AGA TTC TAG GCA CGG GTT
CCG TAG A
ceClp1K127A-pET21a-For
ACC AAC GGA CGT CGG AGC AAC CAC AGT
CTC C
113
Appendix
ceClp1K127A-pET21a-Rev
GGA GAC TGT GGT TGC TCC GAC GTC CGT TGG
T
ceClp1K127R-pET28b-For
ACC AAC GGA CGT CGG AAG AAC CAC AGT
CTC C
ceClp1K127R-pET28b-Rev
GGA GAC TGT GGT TCT TCC GAC GTC CGT TGG
T
114
Appendix
6.1.2 Oligonucleotidesequences.
oligonucleotide
sequence
Single stranded RNA
5’-GCG AGA CAG UGU GAC UUU GG-3’
Double stranded RNA with
5’-GCG AGA CAG UGU GAC UUU GG-3’
blunt ends
Double stranded RNA with a
single stranded 3’-overhang
5’-CCA AAG UCA CAC UGU CUC GC-3’
5’-GAG ACA GUG UGA CUU UGG AC-3’
5’-CCA AAG UCA CAC UGU CUC GC-3’
Single stranded DNA
5’-GCG AGA CAG TGT GAC TTT GG-3’
Double stranded DNA with
5’-GCG AGA CAG TGT GAC TTT GG-3’
blunt ends
Double stranded DNA with a
single stranded 3’-overhang
5’-CCA AAG TCA CAC TGT CTC GC-3’
5’-GAG ACA GTG TGA CTT TGG AC-3’
5’-CCA AAG TCA CAC TGT CTC GC-3’
115
Appendix
6.2 AdditionalTables
6.2.1 ProteinsshowingaconventionalWalkerAlysineresidue.
Nr.
PDB
ID
Protein name
Resolution
[Å]
P-loop
Sequence
P-loop
residue
number
and chain
Ligand
Walker A
Lys
interactions
Conventional Walker A lysine: Substrate bound state structures
1
3RWM
Ypt32
2.00
2
4GP7
Pnkp-Hen1 RNA repair
system
2.00
3
3QF7
Mre11:Rad50 complex
1.90
4
3A4L
O-phosphoseryltRNA(Sec) kinase
1.80
5
2IYW
Shikimate Kinase
1.85
2D7C
Rab11 in complex with
FIP3 Rab-binding
domain
1.75
7
2CBZ
Multidrug resistance
protein 1 nucleotide
binding domain 1
1.50
8
1YZL
Rab9 GTPase
1.85
9
3RLF
Maltose-binding
protein/maltose
transporter complex
2.20
10
3GPL
RecD2
2.50
11
4A6X
RadA C-terminal
ATPase domain
1.55
12
2GCO
RhoC
1.40
13
3MYK
Myosin
1.84
14
3FVQ
Nucleotide binding
domain of FbpC
1.90
15
2QT0
Nicotinamide riboside
kinase 1
1.92
16
2OLR
Phosphoenolpyruvate
carboxykinase
1.60
6
GDSGVGKS
GSSGSGKS
GPNGAGKS
GLPGVGKS
GLPGSGKS
20-27 (A)
GNP
β- and γphosphate
15-22 (A)
ATP
β- and γphosphate
30-37 (A)
ANP
β- and γphosphate
11-18 (A)
ANP
β- and γphosphate
9-16 (A)
ATP
β- and γphosphate
18-25 (A)
GTP
β- and γphosphate
678-685 (A)
ATP
β- and γphosphate
14-21 (A)
GNP
β- and γphosphate
36-43 (A)
ANP
β- and γphosphate
360-367 (A)
ANP
β- and γphosphate
138-145 (A)
ATP
β- and γphosphate
12-129 (A)
GNP
β- and γphosphate
179-186 (X)
ANP
β- and γphosphate
37-44 (A)
ATP
β- and γphosphate
10-17 (A)
ANP
β- and γphosphate
248-255 (A)
ATP
β- and γphosphate
GDSGVGKS
GQVGCGKS
GDGGVGKS
GPSGCGKS
GGPGTGKS
GEFGSGKT
GDGACGKT
GESGAGKT
GASGCGKT
GVTNSGKT
GLSGTGKT
116
Appendix
17
1OXV
GlcV, the ABC-ATPase
of the glucose ABC
transporter
18
3Q9L
MinD
2.34
19
3HQD
Kinesin Eg5 motor
domain
2.00
20
3EW9
RADA recombinase
2.40
GPSGAGKT
1.95
GKGGVGKT
GQTGTGKT
GMFGSGKT
38-45 (A)
ANP
β- and γphosphate
10-17 (A)
ATP
β- and γphosphate
105-112 (A)
ANP
β- and γphosphate
105-112 (A)
ANP
β- and γphosphate
7-14 (A)
AlF4
β- and γphosphate
13-20 (A)
Al F3
β- and γphosphate
13-20 (A)
Al F3
β- and γphosphate
10-17 (R)
Al F3
β- and γphosphate
41-48 (A)
AlF4
36-43 (A)
AlF4
223-230 (A)
Conventional Walker A lysine: Transition state analogue structures
1
3SR0
Adenylate kinase
1.57
2
5UKD
UMP/CMP kinase
1.90
3
1E2E
Thymidylate kinase
2.00
4
1WQ1
RAS-RASGAP
complex
2.50
5
3T34
Dynamin-related
protein 1A (AtDRP1A)
2.40
6
3PUW
MBP-Maltose
transporter
2.30
7
2GJ9
MnmE G-domain
2.00
8
3MSX
RhoA with GAP
domain of ArhGAP20
1.65
9
2WOJ
GET3
2.00
10
3T12
MglA in complex with
MglB
2.20
11
2XZO
UPF1 helicase
2.39
12
1IHU
Aresenite translocating
ATPase
2.15
13
4ANJ
Myosin VI
2.60
14
3BH7
RP2-Arl3 complex
1.90
15
1W9I
Myosin II
1.75
2G77
Gyp1 TBC domain in
complex with Rab33
GTPase
2.26
16
GPPGAGKG
GGPGSGKG
GVDRAGKS
GAGGVGKS
GGQSSGKS
GPSGCGKS
GRPNAGKS
GDGACGKT
GKGGVGKT
GPGLSGKT
GPPGTGKT
GKGGVGKT
GESGAGKT
GLDNAGKT
GESGAGKT
-
-
β- and γphosphate
-
β- and γphosphate
AlF4
-
β- and γphosphate
12-19 (A)
Al F3
β- and γphosphate
25-32 (A)
AlF4
19-26 (A)
AlF4
492-499 (A)
-
β- and γphosphate
-
β- and γphosphate
AlF4
-
β- and γphosphate
324-341 (A)
Al F3
β- and γphosphate
151-158 (A)
AlF4
24-31 (A)
-
β- and γphosphate
AlF4
-
β- and γphosphate
179-186 (A)
Be F3
β- and γphosphate
40-47 (B)
Al F3
β- and γphosphate
GDSNVGKT
117
Appendix
17
1W0J
F1-ATPASE
2.20
18
3SS8
NFeoB
2.51
19
2GTP
RGS1 in complex with
the activated Gi alpha
1
2.55
20
1GRN
CDC42-GAP complex
2.10
GGAGVGKT
GNPNSGKT
155-162 (D)
Be F3
-
β- and γphosphate
AlF4
-
β- and γphosphate
10-17 (A)
Al F3
β- and γphosphate
15-22 (A)
ADP
β- and γphosphate
371-378 (A)
APD
β- and γphosphate
10-17 (A)
APD
β- and γphosphate
11-18 (A)
ADP
β- and γphosphate
87-94 (A)
ADP
β- and γphosphate
96-130 (A)
ADP
β- and γphosphate
9-16 (A)
ADP
β- and γphosphate
98-105 (A)
GDP
β- and γphosphate
528-535 (A)
ADP
β- and γphosphate
27-34 (A)
ADP
β- and γphosphate
371-378 (A)
ADP
β- and γphosphate
138-145 (A)
ADP
β- and γphosphate
91-98 (A)
ADP
β- and γphosphate
27-34 (A)
GDP
β- and γphosphate
105-112 (A)
ADP
β- and γphosphate
10-17 (A)
ADP
β- and γphosphate
8-15 (A)
AlF4
40-47 (A)
GAGESGKS
GDGAVGK
β- and γphosphate
Conventional Walker A lysine: Product bound state structures
1
4GP6
Pnkp-Hen1 RNA repair
system
2.1
2
3ZVN
Polynucleotide kinase
3'-phosphatase
2.15
3
4EDH
4
3A4M
O-phosphoseryltRNA(Sec) kinase
1.8
5
3DC4
Kinesin family member
NOD
1.9
6
2ZFI
7
2IYQ
8
2DPX
9
2H58
10
1UJ2
11
1YQT
12
4A7O
RadA C-terminal
ATPase domain
13
4A14
KIF7
14
3REF
EhRho1
15
3FYH
Recombinase
16
2QSY
Nicotinamide riboside
kinase 1
Thymidylate kinase
Kif1A Motor Domain
Shikimate kinase
Rad GTPase
KIFC3 motor domain
Uridine-cytidine kinase
2
RNase-L Inhibitor
1.3
1.55
1.8
1.8
1.85
1.8
1.9
1.88
1.6
1.95
1.9
1.95
GSSGSGKS
GFPGAGKS
GPEGAGKS
GLPGVGKS
GQTGTGKS
GQTGAGKS
GLPGSGKS
GAPGVGKS
GSERVGKS
GGTASGKS
GPNGIGKT
GEFGSGKT
GQTGSGKT
GDGAVGKT
GVFGSGKT
GVTNSGKT
118
Appendix
17
1XMV
RecA
1.9
18
1SVL
SV40 large T antigen
helicase domain
19
1F6B
SAR1
1.7
20
2UKD
UMP/CMP kinase
2.2
1.95
GPESSGKT
GPIDSGKT
GLDNAGKT
GGPGSGKG
119
66-73 (A)
ADP
β- and γphosphate
426-433 (A)
ADP
β- and γphosphate
32-39 (A)
GDP
β- and γphosphate
12-19 (A)
ADP
β- and γphosphate
Appendix
6.3 Abbreviations
ADP
adenosine-5’-diphosphate
ATP
adenosine-5’-triphosphate
A.U.
arbitrary units
AppNHp
adenyl-5’-(,-imido)triphosphate
BSA
bovine serum albumin
CD
circular dichroism
C-terminus
carboxy terminus
CtD
C-terminal domain
DNA
deoxyribonucleic acid
DTE
dithioerythritol
EDTA
ethylenediaminetetraacetic acid
G1C2
guanylyl(3′→5′)cytidine
GTP
guanosine-5’-triphosphate
HEPES
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
IPTG
isopropyl β-D-1-thiogalactopyranoside
LB
lysogeny broth medium
mRNA
messenger ribonucleic acid
NADH
nicotinamide adenine dinucleotide
N-terminus
amino terminus
NtD
N-terminal domain
ORF
open reading frame
120
Appendix
PCR
polymerase chain reaction
PDB
protein database
PEP
phosphoenolpyruvate
PNK
Polynucleotide kinase
RNA
ribonucleic acid
RPNK
RNA-specific PNK
rRNA
ribosomal RNA
TCEP
(tris(2-carboxyethyl)phosphine)
TEMED
tetramethylethylenediamine
Tris
tris(hydroxymethyl)aminomethane
tRNA
transfer RNA
SDS
sodium dodecyl sulfate
SDS-PAGE
sodium dodecyl sulfate polyacrylamide gel electrophoresis
U1G2
urylyl-(3’→5’)guanosine
UV
ultra violet light
% w/v
mass/volume percentage
% v/v
volume/volume percentage
121
Appendix
122
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement