Analysis of Phloxine B and Uranine in Coffee by High

J. Agric. Food Chem. 1998, 46, 1005−1011
1005
Analysis of Phloxine B and Uranine in Coffee by High-Performance
Liquid Chromatography and Capillary Zone Electrophoresis after
Solid Phase Extraction Cleanup
Jocelyn P. Alcantara-Licudine, Ngoc Lan Bui, Michael K. Kawate, and Qing X. Li*
Department of Environmental Biochemistry, University of Hawaii at Manoa, Honolulu, Hawaii 96822
A method was developed for the analysis of phloxine B and uranine, photoactive dyes being evaluated
as fruit fly toxicants, in coffee cherries and green and roasted beans. The analytes were measured
by high-performance liquid chromatography (HPLC) and capillary zone electrophoresis (CZE) using
visible and fluorescence detectors after cleanup with disposable amino cartridges. A mixture of
methanol (MeOH)/acetonitrile/n-butylamine (n-BA) (1/1/0.05) effectively extracted phloxine B and
uranine from coffee cherries and green beans. The method yielded good recoveries of phloxine B
(66-89%) and uranine (75-100%) at spike levels of 0.05-1.0 μg/g from coffee cherries. Good
recoveries of phloxine B (82-95%) and uranine (95-110%) were obtained from green beans at spike
levels of 0.25-1.0 μg/g. Addition of sodium hexametaphosphate in roasted beans prior to extraction
with MeOH/acetone/n-BA (1/1/0.05) yielded good recoveries of phloxine B (72-77%) and uranine
(79%) at spike levels of 0.5-1.0 μg/g. HPLC and CZE are adequate for determining these analytes.
The major advantages of CZE are short analysis time and use of inexpensive columns and aqueous
buffer.
Keywords: CZE; HPLC; solid-phase extraction; phloxine B; uranine; coffee
INTRODUCTION
Coffee is one of the fastest growing agricultural
industries in Hawaii, and the coffee producers compete
in world trade on the basis of quality (Goto and
Fukunaga, 1991). The high quality of Kona coffee has
been known worldwide for over a century, and this is
attributed to the coffee cultural practices of the Hawaiian farmers. However, coffee serves as a major host to
several Tephritid fruit fly species (i.e., Malaysian,
Mediterranean, Melon, and Oriental flies) that were
introduced in the islands (Vargas et al., 1995, 1997).
Exportation of host plant produce (e.g., papaya and
mango) of these flies from Hawaii is subject to quarantine and therefore causes trade limitations of agricultural produce. Tephritid fruit flies also significantly
damage crops. Migration of these flies threatens agriculture in fruit fly free areas. Suppression of the fly
population can increase crop yield, favor produce trade,
and limit migration to a noninfested area.
Currently, malathion is the primary insecticide being
used to control these flies, but it has received strong
public opposition because of perceived public health and
environmental concerns (Service, 1995). SureDye, containing phloxine B and uranine (Figure 1), has shown
to effectively control fruit flies and could replace
malathion (Liquido et al., 1995a,b; Mangan and Moreno,
1995). When fruit flies consume phloxine B and are
exposed to light, phloxine B presumably generates
singlet oxygen which is toxic to fruit flies. Uranine is
assumed to act as a synergist, although the exact mode
of action is yet unknown (Heitz, 1995).
* Author to whom correspondence should be addressed
[telephone (808) 956-2011; fax (808) 956-5037; e-mail
gingl@hawaii.edu].
Figure 1. Chemical structures of phloxine B (where X ) Br,
Y ) Cl) and uranine (where X ) Y ) H).
Phloxine B and uranine are xanthene dyes widely
used for drugs and cosmetics in the U. S. Phloxine B
and uranine have low mammalian toxicity [phloxine B
acute oral LD50 (mg/kg) rat 8400, mouse 310, dog >4600;
uranine oral LD50 (mg/kg) rat 6700, mouse 4700)]
(Klaassen, 1973; Lutty, 1978; McDonald et al., 1974;
Yankell and Loux, 1977). An allowable intake of
phloxine B is 1.25 mg/kg of body weight for human
consumption (Food and Drug Administration, 1982).
However, little is known on their toxic effects in wildlife
and other nontarget organisms in the ecosystems.
Related studies on the environmental effects of these
dyes are currently being undertaken.
Field pilot tests were conducted in Hawaii, California,
and Guatemala on the use of the xanthene dyes for
suppression and eradication of fruit flies (Li et al.,
1997a; Liquido et al., 1997). SureDye was aerially
sprayed on 126 acres of coffee in Kauai, HI, in 1996.
Part of the study was to develop analytical methods to
assess the fate of these dyes in the environment. We
previously reported an analytical method for phloxine
B, uranine, and related xanthene dyes in soil using
supercritical fluid extraction (Alcantara-Licudine et al.,
S0021-8561(97)00750-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/24/1998
1006 J. Agric. Food Chem., Vol. 46, No. 3, 1998
1997). No method was available for the analysis of
phloxine B and uranine in coffee. This paper reports
an analytical procedure developed for these dyes in
coffee cherries and green and roasted beans.
Phloxine B and uranine are photosensitive, nonvolatile, water soluble, and stable in basic solutions. Thus,
extraction of these polar aromatic dyes from plant
matrices was difficult. A mixture of solvents and salts
were found to effectively extract phloxine B and uranine
from various coffee matrices. Solid phase extraction
(SPE) was explored as a cleanup procedure for coffee
extracts. SPE was widely used for cleaning up plant
extracts, food, and environmental samples (Hiemstra et
al., 1995; Redondo et al., 1996; Schenck et al., 1994).
SPE was used to clean up coffee extracts for the analysis
of carbohydrates (Prodolliet et al., 1995, 1996) and
chlorogenic acids (Bicchi et al., 1995). Availability of
commercial SPE columns made this procedure simple
and convenient for routine analysis.
Capillary zone electrophoresis (CZE) is increasingly
used for determining pesticide residues in grains
(Krynitsky and Swineford, 1995) and environmental
samples (Brumley, 1995; Wu et al., 1995). A CZE
procedure was recently developed for determining phloxine B and uranine in water samples (Li et al., 1997b).
In this study, CZE was used and compared with HPLC
for the analysis of these analytes in coffee extracts.
MATERIALS AND METHODS
Coffee Samples. Coffee cherries were collected from a
coffee field in Kauai, HI. The cherries were processed and
roasted following a standard procedure approved for USDA
Interregional Research Project No. 4. Fresh coffee cherries
were pulped and then fermented for 20-24 h. The fermented
cherries were rinsed and dried in an oven at 53 °C for 16 h to
yield the green coffee beans. The green beans were roasted
at 200 °C until the beans turned dark to produce the roasted
beans. Moisture contents were 16.8, 10.7, and 4.8% for coffee
cherries and green and roasted beans, respectively.
Fresh coffee cherries and dry ice [1/1 (w/w)] were chopped
by a food processor. The sample was transferred to a mason
jar, loosely capped, and stored in the freezer overnight to allow
complete volatilization of dry ice. The jar was tightly sealed
the following day and stored at -25 °C for subsequent analysis.
Green or roasted beans were ground in a coffee grinder to pass
a 20 mesh sieve. Ground-up samples were transferred to
mason jars and stored at 5 °C.
Samples were fortified by addition of phloxine B or uranine
solution in MeOH to 25-50 g of fresh coffee cherries or green
beans and 12.5 g of roasted beans. The samples sat about 30
min prior to extraction.
Chemicals and Supplies. Phloxine B and uranine were
obtained from ICN Biochemicals (Cleveland, OH). Phloxine
B was purified as previously reported (Alcantara-Licudine et
al., 1997). Sodium borate decahydrate (Na2B4O7‚10H2O), boric
acid (H3BO3), n-butylamine (n-BA), and optima grade methanol (MeOH), acetonitrile (ACN), acetone, ethyl acetate (EtOAc),
and methylene chloride (CH2Cl2) were obtained from Fisher
Scientific (Pittsburgh, PA). Other reagents used in this study
were of analytical or HPLC grade. SPE columns (amino, SAX,
cyano, diol, octadecyl, and phenyl) were from either J. T. Baker
(Phillipsburg, NJ) or Varian (Walnut Creek, CA).
Optimization of Extraction Procedure for Coffee.
Cherries. Various solvents (acetone, ACN, H2O, and MeOH)
alone or in mixtures with n-BA and salts [NaCl, (NaPO3)6, Na4EDTA, Na2SO4, NH4OAc, and Na2CO3] were tested to extract
phloxine B and uranine from coffee cherries. Phloxine B and
uranine were spiked at 3 μg/g on whole cherries (50 g) in a
500 mL Erlenmeyer flask or on ground-up cherries in a 1 quart
mason jar covered with aluminum foil. The spiked samples
sat about 30 min and were swirled every 10 min. The whole
Alcantara-Licudine et al.
cherries were extracted with 100 mL of solvent three times
by shaking the flask on a Burrel shaker for 10 min. The
ground-up cherries were homogenized with 100 mL of solvent
with a Sorvall Omni mixer at high speed (4 on a scale of 10)
and repeated twice. The extracts were combined and decanted
through glasswool to a 1 L round-bottom flask. After the
extract was concentrated by rotary evaporation to about 100
mL, it was defatted twice with 50 mL of hexane. The defatted
extract was then concentrated to 50 mL. An aliquot of the
extract was filtered through a 0.45 μm acrodisc filter for HPLC
analysis.
Roasted Beans. Phloxine B and uranine were spiked in
roasted beans (12.5 g) at 5 μg/g and extracted three times with
100 mL of MeOH, ACN, acetone, or a mixture of these solvents
and 5 g of (NaPO3)6 or Na4EDTA. Clarifying agents, e.g., MgO
(Dick, 1995) or lead acetate [Pb(OAc)2] (Tsumura et al., 1994)
were evaluated for cleaning up extracts prior to SPE. Liquidliquid partitioning was also tried to eliminate interferences.
Water (20 mL) was added in extracts (100 mL) which were
then partitioned twice with hexane, EtOAc, CH2Cl2, or methyl
tert-butyl ether (MTBE) (2 × 50 mL).
SPE Cleanup for Coffee Extracts. Several types of SPE
columns (amino, SAX, cyano, diol, octadecyl, and phenyl) were
preliminarily tested to clean up the extracts. Amino and SAX
columns were further investigated. Phloxine B and uranine
(10 μg each) in 10 mL of MeOH/ACN/n-BA [1/1/0.05 (v/v/v)]
were passed through a SPE column (100 mg sorbent) after
conditioning with aqueous HCl (1 mL, 0.05 M). After the
column was washed with 5 mL of distilled water, the dyes were
eluted with NaOH (0.5 mL, 0.5 M)/MeOH (5 mL) from the
amino column and with NaOH (0.5 mL, 1 M)/MeOH (5 mL)
from the SAX column. Various amounts (5-50 μg) of analytes
in 10 mL of MeOH/ACN/n-BA (1/1/0.05) were applied on amino
and SAX columns (100 mg) to examine column capacity. The
efficiency of cleanup and elution was evaluated as recoveries
of the dyes from the columns.
Phloxine B and uranine were dissolved in different solvent
systems (CH2Cl2, ACN, acetone, MeOH, H2O, or mixtures of
these solvents and n-BA) at 1 μg/mL, and a 10 mL aliquot of
the spike solution was applied on the amino column (100 mg).
Column conditioning and elution steps were the same as those
described above. Analytes were recovered to evaluate the
effect of different solvent systems on the trapping efficiency
of the amino column. In a separate experiment, hexane, CH2Cl2, EtOAc, ACN, acetone, MeOH, and H2O (10 mL each) were
successively passed through the amino column (200 mg) to
examine selective removal of interferences after application
of the analytes.
The matrix load was optimized for SPE cleanup of coffee
cherry and green and roasted bean extracts. Phloxine B and
uranine were spiked at levels of 0.5 and 1 μg/mL in blank coffee
extracts (MeOH/ACN/n-BA, 1/1/0.05) containing 0.75-10 g
equiv coffee matrix. A 10 mL aliquot and a 20 mL aliquot of
the 1 μg/mL spike extract were passed through 100 and 200
mg amino columns, respectively. A 40 mL aliquot of the 0.5
μg/mL spike extract was passed through a 500 mg amino
column. The dyes were eluted from the column with aqueous
NaOH (0.5 mL, 1.0 M)/MeOH (5 mL) followed by an additional
5 mL of MeOH.
Extraction and Cleanup Procedures. Cherries and
Green Beans. A 25 g sample of coffee cherries or green beans
was placed in a 1 pint mason jar and extracted with 100 mL
of MeOH/ACN/n-BA (1/1/0.05) for 5 min using a Sorvall Omni
mixer at low speed (2.5 on a scale of 10). The mixture was
filtered through a Buchner funnel, fitted with a no. 7 glass
fiber filter, into a 500 mL suction flask. The sample was
reextracted twice, and the extracts were combined. The
extract was then cleaned up using SPE.
Roasted Beans. The samples (12.5 g) were mixed with
(NaPO3)6 (2.5 g), moistened with distilled H2O (20 mL), and
extracted following the same procedure used for coffee cherries
and green beans except for MeOH/acetone/n-BA (1/1/0.05)
instead of MeOH/ACN/n-BA (1/1/0.05).
SPE Cleanup of Coffee Extracts. An optimized SPE cleanup
procedure is illustrated in Scheme 1. Sample extracts (30-
Phloxine B and Uranine in Coffee after SPE Cleanup
Scheme 1. SPE Cleanup Procedure for Phloxine B
and Uranine in Coffee Extracts
J. Agric. Food Chem., Vol. 46, No. 3, 1998 1007
Table 1. Screening Extraction Methods for Phloxine B
and Uranine in Coffee Cherries
recoveryb (%)
solvent system
H2O (80 °C)
aqueous NH4OAc (1.3 M)
MeOH
MeOH
ACN
MeOH/ACN (1/1)
MeOH/acetone (1/1)
MeOH/acetone/n-BA (1/1/0.05)
MeOH/ACN/n-BA (1/1/0.05)
MeOH/ACN/n-BA (1/1/0.05)
MeOH/ACN/n-BA (1/1/0.05)
MeOH/ACN/n-BA (1/1/0.05)
MeOH/ACN/n-BA (1/1/0.05)
salta
Na2CO3
NaCl
Na2SO4
(NaPO3)6
Na4EDTA
phloxine B uranine
NDc(ND)
ND (ND)
ND (59)
38 (19)
49 (65)
79 (70)
42 (67)
59 (58)
83 (74)
75 (55)
73 (68)
49 (55)
47 (60)
ND (bpd)
ND (bp)
ND (135)
80 (81)
63 (50)
81 (72)
83 (60)
63 (60)
96 (83)
86 (60)
71 (77)
ND (ND)
56 (50)
a 5 g of salt were added. b Recoveries were from ground-up or
whole (data in parentheses) cherries. Data were the means of two
replicates. c ND ) not detected. d bp ) broad peak.
40 mL) were applied on the amino column (500 mg) after it
was washed with MeOH and activated with HCl/MeOH.
When the solution was about 5 mm in the column, hexane,
acetone, and MeOH were passed successively through the
column to remove interferences. Finally, the dyes were eluted
with aqueous NaOH/MeOH followed by an additional 5 mL of
MeOH. The flow was controlled with a Supelco vacuum
manifold (set at 20 psi) connected to a water-suction pump.
The elution flow was approximately 5-6 mL/min. After SPE
cleanup, the extracts were concentrated to the appropriate
volume for HPLC or CZE analysis.
HPLC Analysis. HPLC analysis was done according to the
method previously described (Alcantara-Licudine et al., 1997)
but with the following modifications: column, Alltima C18 LL
(Alltech, Deerfield, IL); mobile phase, ACN-0.5 M NH4OAc
buffer with linear gradient increase of ACN 20-80% within
15 min, followed by 100% ACN isocratic elution for 10 min
and equilibrated back to 20% ACN for 20 min.
CZE Analysis. The analysis was performed according to
an established procedure for phloxine B and uranine (Li et
al., 1997b) on a Dionex capillary electrophoresis system (CES1) (Dionex Corp., Sunnyvale, CA) with ultraviolet-visible
(UV-vis) and fluorescence (Fl) detectors. The running buffer
was 10 mM Na2B4O7 and 50 mM H3BO3 (pH 8.5). Buffers and
sample solutions were filtered through a 0.45 μm Gelman
membrane filter. The fused silica capillary used for separation
has an effective length of 65 cm with 75 μm i.d. A new
capillary was pretreated by flushing with a 1 N NaOH solution
for at least 5 min. Prior to daily analysis, the capillary was
conditioned by flushing with distilled deionized water for 3
min, a 0.1 N NaOH solution for 3 min, and finally with the
running buffer for 5 min. A constant potential was 20 kV,
and the maximum current was 300 μA. Absorptions at 546
and 493 nm were monitored for phloxine B and uranine,
respectively. Fluorescent detection of uranine was carried out
at 493 (excitation) and 515 nm (emission). The injection mode
was by gravity with the sampler head height set at 100 mm
for 10 s.
RESULTS AND DISCUSSION
Screening Extraction Solvents (Table 1). Screening various solvent systems suggested that a mixture
of MeOH/ACN/n-BA (1/1/0.05) is suitable to extract
phloxine B and uranine from coffee cherries (Table 1).
This solvent system extracted 74 and 83% of the
phloxine B and uranine applied, respectively, from
whole cherries. The recoveries increased approximately
Figure 2. Amino and SAX columns (100 mg sorbent) for
trapping phloxine B and uranine.
10% when cherries were homogenized. Addition of salt
[i.e., NaCl, Na2SO4, (NaPO3)6, or Na4EDTA] to the
samples did not improve recoveries. Analytes were not
extracted by hot water alone or 1.3 M aqueous NH4OAc even though the dyes are very water soluble. ACN
or MeOH alone gave variable recoveries of phloxine B
and uranine. Poor recoveries of phloxine B and uranine
by MeOH or H2O may be due partly to sample loss
because a thick emulsion was formed during extraction
and viscosity of the extract increased during solvent
evaporation. Failure to partition phloxine B and uranine in nonpolar solvents such as CH2Cl2, diethyl ether,
EtOAc, and hexane limited the elimination of interferences by liquid-liquid partitioning. HPLC peak tailing
and broadening were encountered after approximately
15-20 injections. The contaminated HPLC column was
reconditioned following the method of Dolan and Snyder
(1989).
Selection of SPE Columns for Trapping Phloxine B and Uranine (Figure 2). SPE was evaluated
for sample cleanup because of its selectivity for a wide
range of compounds including polar and anionic analytes. Phloxine B and uranine were quantitatively
extracted from water samples using phenyl SPE cartridges (Li et al., 1997b). Among the SPE columns
screened (amino, SAX, cyano, diol, octadecyl, and phenyl), the amino and SAX columns effectively retained the
analytes when the dyes were dissolved in MeOH/ACN/
n-BA (1/1/0.05) and passed through the column. The
other columns did not effectively retain the dyes. The
amino column quantitatively recovered phloxine B (96%)
1008 J. Agric. Food Chem., Vol. 46, No. 3, 1998
Figure 3. Effect of different solvent systems on the recovery
of phloxine B and uranine from the amino column (100 mg
sorbent). Ratio of MeOH/ACN or acetone/n-BA was 1/1/0.05
(v/v/v).
and uranine (100%). The SAX column recovered 77%
of the phloxine B and 97% of the uranine applied. The
amino and SAX columns effectively trapped anionic
phloxine B and uranine and released neutral and
cationic interferences such as carbohydrates and chlorophyll.
Subsequent trapping and eluting of the analytes were
tested on the amino and SAX columns (Figure 2). The
amino column had approximately equal efficiency, illustrated as recovery for phloxine B and uranine. Over
80% of the dyes were recovered when up to 25 μg of
uranine and 20 μg of phloxine B were applied per
column (100 mg); beyond these loads, recoveries declined. However, the recoveries of phloxine B were
lower than uranine on the SAX column. The SAX
column is a stronger anion exchanger than the amino
column; therefore, a higher concentration of base is
required to elute the dyes. The amino column was
preferred in this study.
Optimization of Solvent Systems for Trapping
Analytes and Effective On-Column Cleanup (Figure 3). In general, the adsorption of the analytes on
the amino column varied when phloxine B and uranine
were dissolved in different solvents. Both analytes were
quantitatively adsorbed (>96%) on the column when the
dyes were dissolved in a mixture of MeOH/ACN/n-BA
(1/1/0.05) or MeOH/H2O (1/1). MeOH/ACN/n-BA (1/1/
0.05) also was effective in extracting these dyes from
coffee cherries. Therefore, no solvent transfer step was
needed for the sample cleanup when the amino column
was used. After the dyes were adsorbed on the column,
hexane, CH2Cl2, EtOAc, ACN, and MeOH (10 mL each)
were successively passed through the column. These
solvents cleaned up interferences in the columns without removing the target analytes (data not shown).
Subsequently, the analytes were eluted with aqueous
NaOH/MeOH.
SPE Cleanup for Phloxine B and Uranine Spiked
in Coffee Extracts (Figure 4 and Table 2). Phloxine
B and uranine were spiked in coffee extracts (MeOH/
ACN/n-BA, 1/1/0.05) which were equivalent to various
amounts of cherries. The amounts of phloxine B and
uranine adsorbed decreased as the matrix load increased (Figure 4). The 100 mg column retained over
83% of the phloxine B applied up to 1.25 g equiv matrix
load. However, only 40% of the uranine applied, which
was first eluted from the column, was recovered with a
0.75 g equiv matrix load. The 200 mg column quantitatively recovered uranine (>92%) when the matrix load
was equivalent to 5 g of coffee cherries. Phloxine B
Alcantara-Licudine et al.
Figure 4. Effect of matrix load on the recovery of phloxine B
and uranine from the amino column.
recoveries (g83%) were good up to a 2.5 g equiv matrix
load but decreased to 74% at a 5 g equiv matrix load.
MeOH insoluble interferences were coeluted with the
analytes. The extract formed a two phase solution
which was readily separated by centrifugation prior to
analysis.
The sorbent/matrix ratio was examined on a larger
amino column (500 mg sorbent) (Table 2). A matrix load
of up to a 5 g equiv of coffee cherries gave good
recoveries (73-103%). The cleanup step yielded a clear,
golden yellow extract. Cleanup with this column was
also examined for analytes spiked in green and roasted
bean extracts. Recoveries were 82-101% up to 2.5 g
equiv of green beans but decreased to 46% at a 5.0 g
equiv matrix load. When the dyes were spiked in
roasted bean extracts, up to a 1.25 g equiv load can be
applied without sacrificing good recoveries. However,
the roasted bean extract left dark brown colored coextractives in the column which suggested the need of
cleanup prior to SPE.
Roasted Bean Extraction (Table 3). When the
dyes were spiked in roasted beans, recoveries of phloxine
B (32%) and uranine (29%) were low following the same
extraction (MeOH/ACN/n-BA, 1/1/0.05) and cleanup
procedures for coffee cherries or green beans. MeOH/
acetone/n-BA (1/1/0.05) recovered 35% of phloxine B and
37% of uranine in roasted beans. Roasted coffee differs
in chemical composition from raw cherries or green
beans by the presence of humic acids and mineral oxides
formed during the roasting process (Clarke and Macrae,
1985; Sivetz and Desrosier, 1979). These compounds
may strongly interact with the dyes and consequently
result in poor recoveries. MgO was commonly used as
a clarifying agent in tea analysis (Dick, 1995). MgO
turned the dark brown extract to a light yellow brown
solution. However, the recoveries of phloxine B and
uranine were very low (0-46%) using different combinations of solvents and MgO (Table 3). Addition of lead
acetate to clarify the extract as reported by Tsumura
et al. (1994) resulted in a clear golden yellow solution;
however, the recoveries of phloxine B and uranine were
only 57 and 30%, respectively. We previously found that
MeOH/n-BA/Na4EDTA were effective modifiers for supercritical fluid extraction and that MeOH/n-BA/
(NaPO3)6 was effective for conventional solvent extraction of phloxine B and uranine in soil (AlcantaraLicudine et al., 1997). However, use of MeOH/ACN/nBA/Na4EDTA yielded low recoveries of phloxine B (9%)
and uranine (25%) from roasted beans. Use of MeOH/
ACN/n-BA/(NaPO3)6 gave good recovery of uranine
Phloxine B and Uranine in Coffee after SPE Cleanup
J. Agric. Food Chem., Vol. 46, No. 3, 1998 1009
Table 2. Amino Columna Cleanup for Phloxine B and Uranine Spiked in Coffee Extracts
recovery ((SD)b (%)
equiv sample
load (g)
0
0.75
1.25
2.5
5.0
7.5
a
coffee cherry
green bean
phloxine B
uranine
96 ( 4
84 ( 6
94 ( 6
89 ( 3
73 ( 11
92 ( 7
95 ( 10
103 ( 11
97 ( 3
88 ( 14
roasted bean
phloxine B
uranine
phloxine B
uranine
96 ( 4
82 ( 9
92 ( 5
46 ( 10
45 ( 24
101 ( 6
90 ( 5
92 ( 14
47 ( 19
17 ( 7
87 ( 12
88 ( 5
68 ( 3
73 ( 4
90 ( 6
95 ( 3
82 ( 13
69 ( 6
500 mg sorbent. b Data were the means of three to five replicates. SD ) standard deviation.
Table 3. Screening Solvent Extraction Method and
Cleanup Prior to SPE for Phloxine B and Uranine in
Roasted Beans
recoverya (%)
solvents
salts
MeOH/ACN/n-BA (1/1/0.05)
MeOH/acetone/n-BA (1/1/0.05)
MeOH
ACN
MeOH/ACN (1/1)
MeOH/acetone (1/1)
MeOH/acetone (1/1)
MeOH/ACN/n-BA (1/1/0.05)
MeOH/ACN/n-BA (1/1/0.05)
MeOH/acetone/n-BA (1/1/0.05)
a
MgO
MgO
MgO
MgO
Pb(OAc)2
Na4EDTA
(NaPO3)6
(NaPO3)6
phloxine B uranine
31.9
34.9
37.5
ND
19.8
22.6
56.7
8.9
49.6
70.1
28.5
36.7
46.0
ND
39.1
17.1
29.6
24.8
79.3
70.4
Data were the means of two to four replicates.
Table 4. Recoveries of Phloxine B and Uranine from
Coffee Cherries and Green and Roasted Beans
recovery ((SD)a (%)
spike concentration
(μg/g)
1.00
0.50
0.25
0.125
0.05
phloxine B
Coffee Cherries
86 ( 3
88 ( 7
84 ( 7
81 ( 8
67 ( 2
uranine
97 ( 10
100 ( 2
95 ( 3
87 ( 2
77 ( 3
1.0
0.5
0.25
Green Beans
95 ( 5
92 ( 10
82 ( 11
110 ( 8
105 ( 6
95 ( 14
1.0
0.5
0.25
Roasted Beans
72 ( 8
77 ( 11
bpb
79 ( 16
79 ( 24
bp
a Data measured by HPLC were the means of three to six
replicates. SD ) standard deviation. b bp ) broad peak.
(79%) but not phloxine B (50%). Surprisingly, MeOH/
acetone/n-BA/(NaPO3)6 recovered about 70% of phloxine
B and uranine from roasted beans. MeOH/acetone/nBA/(NaPO3)6 also extracted endogenous materials from
coffee producing a dark brown solution even after SPE
cleanup. Liquid-liquid partition was tested prior to
SPE to remove interferences. However, recoveries
decreased to 7-47% for phloxine B and 45-69% for
uranine after partitioning (without acidification) with
CH2Cl2, EtOAc, diethyl ether, hexane, or MTBE.
Recoveries of Phloxine B and Uranine Spiked
in Coffee Samples (Scheme 1 and Table 4). The
analytes in coffee cherries were extracted with MeOH/
ACN/n-BA (1/1/0.05) and cleaned up with the amino
column (Scheme 1). The recoveries of phloxine B and
uranine ranged from 67 to 88% and from 77 to 100%,
respectively, at 0.05-1.00 μg/g spike levels (Table 4).
Figure 5. Correlation between HPLC and CZE analyses of
uranine and phloxine B in coffee samples.
The extraction and cleanup procedures developed for
coffee cherries were applicable to green beans. Good
recoveries were obtained for phloxine B (82-95%) and
uranine (95-110%) in green beans spiked at 0.25-1.0
μg/g levels. Analyte peak heights slightly decreased and
extra peaks appeared on the CZE electropherograms
after the extracts were refrigerated overnight at 5 °C.
Also, the color of the extracts changed from clear golden
yellow to green. This indicated that the analytes in
extracts may be unstable in storage. Therefore, quantitation should be conducted immediately after extraction. Extraction using MeOH/acetone/n-BA (1/1/0.05)
and (NaPO3)6 gave recoveries of 72-77% for phloxine
B and 79% for uranine in roasted beans spiked at 0.51.0 μg/g. Recoveries at the 0.25 μg/g spike level could
not be determined due to peak broadening.
Comparison of HPLC and CZE Determinations
(Figures 5 and 6 and Table 5). Phloxine B and
uranine in coffee extracts were determined by CZE and
HPLC. The results by CZE correlated closely with those
by HPLC (Figure 5). The correlation coefficients (r)
were about 0.99 for phloxine B (slope, 0.91) and uranine
(slope, 1.08).
1010 J. Agric. Food Chem., Vol. 46, No. 3, 1998
Alcantara-Licudine et al.
visible detector. Fluorescence detection is more sensitive and selective, and eliminated interferences. Some
difficulties, e.g. peak broadening, changing a guard
column, and loss of sensitivity, often occurred during
HPLC analysis of coffee extracts, particularly from
roasted beans. The CZE method avoided these difficulties. CZE analysis of phloxine B and uranine also had
advantages of using aqueous buffer and short analysis
time (12 min per run by CZE vs 45 min by HPLC). CZE
used a simple and inexpensive column compared to an
expensive HPLC column. HPLC and CZE have different separation mechanisms and thus complement each
other for chemical identification and characterization.
Partitioning of analytes between the stationary and
mobile phases and the eluate flow driven by pressure
are the main separation mechanism of HPLC. CZE
separation is due to differential electrophoretic migration of the analytes and electroosmotic flow of the bulk
solution in an electric field.
CONCLUSION
Figure 6. CZE electropherograms of phloxine B [visible (vis),
λ ) 546 nm] and uranine [vis; λ ) 493 nm, fluorescence (Fl);
λexc ) 493 nm, λem ) 515 nm] in coffee cherry extracts (A, top)
and in 1.0 μg/mL standard solutions in MeOH (B, bottom).
Table 5. Comparison of HPLC and CZE Analysis of
Phloxine B and Uranine
compound
comparison
parameter
phloxine B
uranine
Mean Recovery (%) and Coefficient of Variationa
HPLC-vis
81.2 (7.4)
91.4 (5.4)
CZE-vis
81.0 (6.8)
90.6 (8.8)
CZE-Fl
92.4 (5.2)
Limit of Detection (μg/mL)b and Coefficient of Variation
HPLC-vis
0.041 (2.4)
0.071 (2.8)
CZE-vis
0.191 (1.1)
0.125 (2.4)
CZE-Fl
0.075 (2.7)
a Data were the means of all recoveries of the analyte spiked
at levels of 0.05-1.00 μg/g in ground coffee cherries, and these of
all coefficients of variation (CV%, in parentheses). b Data were the
means of three to five replicates. Signal to noise ratio was
approximately 3.
HPLC and CZE gave very comparable average recoveries of phloxine B (81%) and uranine (91-92%) in coffee
samples (Table 5). The results by HPLC fluctuated
similarly with those by CZE. The mean coefficients of
variation (CV%) ranged from 5 to 9% for both HPLC
and CZE measurements. The minimum detectable
concentrations of phloxine B were 0.04 and 0.19 μg/mL
by HPLC-vis and CZE-vis, respectively. The minimum detectable concentrations of uranine were 0.07,
0.13, and 0.08 μg/mL by HPLC-vis, CZE-vis, and
CZE-Fl, respectively. The detection limit of phloxine
B by CZE-vis was about five times higher than that
by HPLC-vis. The detection limit of uranine by CZEvis was almost doubled compared to HPLC-vis. However, the detection limit of uranine by CZE-Fl was
similar to that by HPLC-vis. The high detection limit
by CZE was attributed to the small injection volume and
short detection path length (i.e., 75 μm i.d. capillary).
Figure 6 shows the CZE electropherograms of phloxine B and uranine in coffee extracts and standard
solutions. When an analyte concentration in coffee
samples was lower than 0.1 ppm, extracts needed to be
concentrated for CZE analysis. Consequently, uranine
comigrated with an interference(s) as determined by
A mixture of MeOH/ACN/n-BA (1/1/0.05) adequately
extracted phloxine B and uranine from coffee cherries
and green beans. MeOH/acetone/n-BA (1/1/0.05) and
(NaPO3)6 were used to extract phloxine B and uranine
from roasted beans. Amino SPE columns were used for
matrix cleanup for the analysis of phloxine B and
uranine in all coffee matrices tested. Good recoveries
of the dyes were obtained from the coffee samples.
HPLC and CZE were comparable for the detection of
these dyes in coffee samples. This method will be useful
in monitoring phloxine B and uranine residues.
ACKNOWLEDGMENT
We gratefully acknowledge Ms. J. Coughlin in this
department for processing green and roasted coffee
beans and Drs. R. Cunningham, N. Liquido, and G.
McQuate at USDA-ARS in Hilo, HI, for providing the
fresh coffee cherries. We thank Drs. M. David and R.
Lodevico in this laboratory for their review of the
manuscript.
LITERATURE CITED
Alcantara-Licudine, J. P.; Kawate, M. K.; Li, Q. X. Method
for the analysis of phloxine B, uranine, and related xanthene
dyes in soil using supercritical fluid extraction and high
performance liquid chromatography. J. Agric. Food Chem.
1997, 45, 766-773.
Bicchi, C. P.; Binello, A. E.; Pellegrino, G. M.; Vanni, A. C.
Characterization of green and roasted coffees through the
chlorogenic acid fraction by HPLC-UV and principal component analysis. J. Agric. Food Chem. 1995, 43, 1549-1555.
Brumley, W. C. Techniques with potential for handling
environmental samples in capillary electrophoresis. J. Chromatogr. Sci. 1995, 33, 670-685.
Clarke, R. J., Macrae, R. Eds. Coffee; Elsevier Applied Science
Publishers: New York, 1985; Vol. 1: Chemistry.
Dick, R. S., Chapter Ed. Coffee and tea. In Official Methods
of Analysis, 16th ed.; Cunnif, P., Ed.; AOAC International:
Gaithersburg, MD, 1995; Chapter 30.
Dolan, J. W.; Snyder, L. R. Separation problems: Band tailing
and peak distortion. In Troubleshooting LC systems; Humana Press: Totowa, NJ, 1989.
Food and Drug Administration. D&C red no. 27 and D&C red
no. 28. Fed. Regist. 1982, 47, 42566-42569.
Goto, Y. B.; Fukunaga, E. T. Coffee: Harvesting and processing
for top quality coffee; Extension Circular 359; Hawaii
Phloxine B and Uranine in Coffee after SPE Cleanup
Institute of Tropical Agriculture and Human Resources,
University of Hawaii at Manoa: Honolulu, HI, 1991.
Heitz, J. R. Pesticidal applications of photoactivated molecules.
In Light-Activated Pest Control; Heitz, J. R., Downum, K.
R., Eds.; ACS Symposium Series 616; American Chemical
Society: Washington, DC, 1995.
Hiemstra, M.; Joosten, J. A.; Kok, A. D. Fully automated solidphase extraction cleanup and on-line liquid chromatographic
determination of benzimidazole fungicides in fruit and
vegetables. J. AOAC Int. 1995, 78 (5), 1267-1274.
Klaassen, C. D. Comparison of the toxicity of chemicals in
newborne rats to bile duct-ligated and sham-operated rats
and mice. Toxicol. Appl. Pharmacol. 1973, 24, 37-44.
Krynitsky, A. J.; Swineford, D. M. Determination of sulfonylurea herbicides in grains by capillary electrophoresis. J.
AOAC Int. 1995, 78 (4), 1091-1096.
Li, Q. X.; Alcantara-Licudine, J. P.; Wang, L. Environmental
fate of phloxine B and uranine. J. Photobiochem. Photobiol.
1997a, 65 (special issue), 66S.
Li, Q. X.; Alcantara-Licudine, J. P.; Li, L. P. Determination of
phloxine B and uranine in water by capillary zone electrophoresis. J. Chromatogr. Sci. 1997b, 35 (12), 573-577.
Liquido, N. J.; McQuate, G. T.; Cunningham, R. T. Lightactivated toxicity of phloxine B and uranine to Mediterranean fruit fly, Ceratitis capitata (Wiedemann) (Diptera:
Tephritidate), adults. In Light-Activated Pest Control; Heitz,
J. R., Downum, K. R., Eds.; ACS Symposium Series 616;
American Chemical Society: Washington, DC, 1995a.
Liquido, N. J.; McQuate, G. T.; Cunningham, R. T. Lightactivated toxicity of phloxine B and fluorescein in methyleugenol to Oriental fruit fly, Bactrocera dorsalis (Hendel)
(Diptera: Tephritidate) males. In Light-Activated Pest
Control; Heitz, J. R., Downum, K. R., Eds.; ACS Symposium
Series 616; American Chemical Society: Washington, DC,
1995b.
Liquido, N. J.; Cunningham, R. T.; McQuate, G. T. Photoactive
dyes mixed in baits and lures for fruit fly suppression and
eradication. J. Photobiochem. Photobiol. 1997, 65 (special
issue), 65S.
Lutty, G. A. The acute intravenous toxicity of biological stains,
dyes, and other fluorescent substances. Toxicol. Appl. Pharmacol. 1978, 44, 225-249.
Mangan, R. L.; Moreno, D. S. Development of phloxine B and
uranine bait for control of Mexican fruit fly. In LightActivated Pest Control; Heitz, J. R., Downum, K. R., Eds.;
ACS Symposium Series 616; American Chemical Society:
Washington, DC, 1995.
McDonald, T.; Kasten, K.; Hervey, R.; Gregg, S.; Robb, C. A.;
Borgmann, A. R. Acute and subacute toxicity evaluation of
intravenous sodium fluorescein in mice, rats and dogs.
Toxicol. Appl. Pharmacol. 1974, 29, 97-98.
J. Agric. Food Chem., Vol. 46, No. 3, 1998 1011
Prodolliet, J.; Bruelhart, M.; Blanc, M. B.; Leloup, V.; Cherix,
G.; Donnelly, C. M.; Viani, R. Adulteration of soluble coffee
with coffee husks and parchments. J. AOAC Int. 1995, 78
(3), 761-767.
Prodolliet, J.; Bugner, E.; Feinberg, M. Determination of
carbohydrates in soluble (instant) coffee by anion-exchange
chromatography with pulsed amperometric detection: summary of collaborative study. J. AOAC Int. 1996, 79 (6),
1400-1407.
Redondo, M. J.; Ruiz, M. J.; Boluda, R.; Font, G. Optimization
of a solid-phase extraction technique for the extraction of
pesticides from soil samples. J. Chromatogr. A 1996, 719,
69-76.
Schenck, F. J.; Wagner, R.; Hennessy, M. K.; Okrasinski, J.
L. Screening procedure for organochlorine and organophosphorous pesticide residues in eggs using a solid-phase
extraction cleanup and gas chromatographic detection. J.
AOAC Int. 1994, 77 (4), 1036-1040.
Service, R. F. New compounds make light of pests. Science
1995, 268, 806.
Sivetz, M.; Desrosier, N. W. Physical and chemical aspects of
coffee. In Coffee Technology; AVI Publishing: Westport, CT,
1979.
Tsumura, Y.; Wada, I.; Fujiwara, Y.; Nakamura, Y.; Tonogai,
Y.; Ito, Y. Simultaneous determination of 13 synthetic
pyrethroid metabolite, 3-phenoxybenzoic acid, in tea by gas
chromatography. J. Agric. Food Chem. 1994, 42, 2922-2925.
Vargas, R. I.; Walsh, W. A.; Nishida, T. Colonization of newly
planted coffee fields: Dominance of Mediterranean fruit fly
over Oriental fruit fly (Diptera: Tephritidae). J. Econ.
Entomol. 1995, 88 (3), 620-627.
Vargas, R. I.; Prokopy, R. J.; Duan, J. J.; Albrecht, C.; Li, Q.
X. Capture of wild Mediterranean and Oriental fruit flies
(Diptera: Tephritidae) in Jackson and McPhail traps baited
with coffee juice. J. Econ. Entomol. 1997, 90 (1), 165-169.
Wu, J.; Wong, M. K.; Lee, H. K. Capillary zone electrophoretic
determination of heterocyclic aromatic amines in rain. J.
Chromatogr. Sci. 1995, 33, 712-716.
Yankell, S.; Loux, J. Acute toxicity testing of erythrosine and
sodium fluorescein in mice and rats. J. Peridontol. 1977,
48, 228-231.
Received for review September 2, 1997. Revised manuscript
received January 5, 1998. Accepted January 5, 1998. This
study was supported by USDA-ARS award number 58-53204-549 and University of Hawaii Research Council award.
N.L.B. was a NSF-REU scholar for the summer 1996-1997
programs.
JF9707501
Download PDF

advertising