Journal of Chromatography A, 1202 (2008) 118–123

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Streamlining methodology for the multiresidue analysis of ␤-lactam antibiotics

in bovine kidney using liquid chromatography–tandem mass spectrometry夽
Katerina Mastovska ∗ , Alan R. Lightfield
US Department of Agriculture, Agricultural Research Service, Eastern Regional Research Center, 600 East Mermaid Lane, Wyndmoor, PA 19038, USA

a r t i c l e i n f o a b s t r a c t

Article history: A previously reported multiresidue method for the analysis of 11 important ␤-lactams (amoxicillin, ampi-
Received 4 February 2008 cillin, cefazolin, cephalexin, cloxacillin, desfuroylceftiofur cysteine disulfide (DCCD), deacetylcephapirin,
Received in revised form 3 July 2008 dicloxacillin, nafcillin, oxacillin, and penicillin G) in bovine kidney has been further streamlined. The
Accepted 8 July 2008
method is based on a simple extraction using acetonitrile–water (4:1, v/v), followed by dispersive solid-
Available online 11 July 2008
phase extraction clean-up with C18 sorbent, concentration of an extract aliquot, and filtration of the
final extracts using syringeless filter vials, which are used for the sample introduction in the liquid
Keywords:
chromatographic–tandem mass spectrometric (LC–MS/MS) analysis. The recoveries have been improved
Antibiotics
␤-Lactams
by adding the internal standard [13 C6 ]sulfamethazine to the homogenized sample before the extraction
Penicillins step, which enabled a proper control of the volume changes during the sample preparation. Average recov-
Cephalosporins eries of fortified samples were 87–103% for all ␤-lactams, except for DCCD, which had an average recovery
Stability of 60%. Based on the results of the stability study and LC mobile phase tests, methanol has been elimi-
Degradation nated from the entire method, including the LC–MS/MS analysis. The best overall LC–MS/MS (electrospray
Liquid chromatography–mass spectrometry positive ionization) performance was achieved by using 0.1% formic acid as an additive in both parts of
Sample preparation the mobile phase, in water and in acetonitrile. To prevent carry-over in the LC–MS/MS analysis, the LC
method was divided into two parts: one serving as an analytical method for injection of the sample and
elution of the analytes and the other one, starting at a highly organic mobile phase composition, being
dedicated for injection of a solvent, washing of the system, and equilibration of the column to the initial
conditions of the analytical method. In this way, a blank solvent is injected after each sample, but these
in-between injections contribute minimally to the overall sample throughput.
Published by Elsevier B.V.

1. Introduction identification/confirmation purposes. Our laboratory efforts aim at

development of quantitative multiresidue method(s) for veterinary
␤-Lactam antibiotics are one of the most widely applied antimi- drug residues based on LC–MS/MS analysis and transfer of these
crobial drugs in current veterinary practice. The class of ␤-lactams methods to the FSIS and other laboratories.
includes penicillins and cephalosporins that both have a ␤-lactam Several LC–MS-based methods for the analysis of one or more
ring in their structures, but this ring is fused to a five-membered ␤-lactams in bovine kidney have been previously reported [1–7].
thiazolidine or a six-membered dihydrothiazine ring, respectively. Two of the multiresidue methods were developed in our labora-
In the USA, bovine kidney is the target tissue for monitoring tory by Fagerquist and Lightfield at first for confirmatory analysis
of ␤-lactam (and other antibiotics) levels in cattle within enforce- using an ion trap MS instrument [3], which was later replaced by a
ment and surveillance programs of the Food Safety and Inspection triple quadrupole MS instrument for a more quantitative analysis
Service (FSIS) of the US Department of Agriculture. Currently, the [4]. Furthermore, the latter method employed dispersive solid-
FSIS uses a semi-quantitative microbial assay to determine violative phase extraction (SPE) clean-up instead of the previously used
levels of ␤-lactams (and some other antibiotics) and employs liquid cartridge-based SPE, which simplified and speeded up the sample
chromatography–tandem mass spectrometry (LC–MS/MS) only for preparation.
As opposed to microbial assays, LC–MS-based methods usually
employ organic solvents (mainly methanol or acetonitrile) for sam-
夽 Mention of brand or firm name does not constitute an endorsement by the U.S. ple preparation and analyte elution, which can pose problems in
Department of Agriculture above others of a similar nature not mentioned. terms of analyte stability during the analytical process and standard
∗ Corresponding author. Tel.: +1 215 233 6645; fax: +1 215 233 6642. solution/extract storage. Tyczkowska et al. [8] previously reported
E-mail address: katerina.mastovska@ars.usda.gov (K. Mastovska). relatively rapid degradation of cloxacillin in methanol (MeOH) and

0021-9673/$ – see front matter. Published by Elsevier B.V.

doi:10.1016/j.chroma.2008.07.009
 K. Mastovska, A.R. Lightfield / J. Chromatogr. A 1202 (2008) 118–123 119

MeOH–water solutions, presumably due to the methyl ester for- system (Dubuque, IA, USA). Liquid-headspace supplied nitrogen
mation [9]. MeOH–water (50:50, v/v) was used for the preparation serving as nebulizer, curtain, and collision gas in LC–MS/MS was
of ␤-lactam stock solutions in the above-mentioned methods [3,4]. obtained from Air Products (Allentown, PA, USA). Bovine kidneys
Thus, a question arose, which was related to the use of MeOH in the (checked to be ␤-lactam free) were obtained from a local grocery
␤-lactam method, including not only standard preparation but also store. Sorbents tested for dispersive SPE included C18 from J.T. Baker
its use as a part of the mobile phase in the LC–MS/MS analysis, since (Phillipsburg, NJ, USA), primary secondary amine (PSA) from Var-
MeOH is a frequently employed mobile phase solvent in ␤-lactam ian (Harbor City, CA, USA) and graphitized carbon black (GCB) from
LC–MS methods [4,5,7,10–14]. Supelco (Bellefonte, PA, USA).
In this study, we evaluated stability of ␤-lactams at differ-
ent conditions in solutions prepared in solvents typically used in 2.2. Stability study experiments
LC–MS-based methods for analysis of antibiotics. We also investi-
gated the impact of the LC mobile phase composition (% addition For the stability study experiments, test solutions of the 10 ␤-
of formic acid, presence of MeOH, etc.) on LC–MS/MS analysis sen- lactams were prepared in replicates at 100 ng/mL in water, MeCN,
sitivity, selectivity, ruggedness, and speed for 14 ␤-lactams. Finally, MeCN–water (50:50, v/v), MeOH, or MeOH–water (50:50, v/v). The
we further improved the previously reported multiresidue method same set of solutions was also prepared with addition of 0.1%
for the analysis of ␤-lactams in bovine kidney [4] by streamlining FA. Propranolol was added as an internal standard to all solu-
the procedure and by introducing a proper control of the volume tions at 200 ng/mL. The test solutions were analyzed by LC–MS/MS
changes during the sample preparation. immediately after their preparation (t = 0 h) and stored in different
conditions: in dark and clear vials at room temperature and light,
in dark vials in the refrigerator (+4 ◦ C) and in polypropylene tubes
2. Experimental in the freezer (−20 ◦ C). The solutions stored in dark vials at room
temperature were analyzed in 6 h intervals for 5 days. All solutions
2.1. Chemicals and materials were analyzed every week for 8 weeks. Responses of analytes were
compared to freshly prepared solutions.
Reference standards, all 95% or higher purity, were obtained
from US Pharmacopeia (Rockville, MD, USA); except for cefadroxil 2.3. LC–MS/MS conditions
and propranolol (used as an internal standard), which were
obtained from Sigma (St. Louis, MO, USA). Deacetylcephapirin was LC–MS/MS analysis was performed using an Agilent 1100 LC
provided by the Center of Veterinary Medicine of the US Food system with a binary pump, autosampler, column heater (kept at
and Drug Administration (Laurel, MD, USA) and by the European 30 ◦ C), and degasser (Agilent Technologies, Palo Alto, CA, USA) inter-
Union Community Reference Laboratory for Antimicrobial Residues faced to an API 3000 triple quadrupole mass spectrometer (Applied
in Food (Fougères, France). A metabolite of ceftiofur, desfuroyl- Biosystems; Toronto, ON, Canada). Injection volume was 5 ␮L in
ceftiofur cysteine disulfide (DCCD), was provided by Pharmacia the stability study to enable direct injection of analyte solutions
(Kalamazoo, MI, USA) and Pfizer (New York, NY, USA). The isotopi- in all tested solvents without affecting peak focusing of the early
cally labeled standard of phenyl-[13 C6 ]sulfamethazine (90%) was eluting analytes. Injection volume of 10 ␮L was used in all other
obtained from Cambridge Isotope Laboratories (Andover, MA, USA). instances when analytes where introduced in aqueous solutions
Individual stock solutions (at about 2000 ␮g/mL) were pre- (in the mobile phase evaluation studies and the analysis of kidney
pared in water, except for cefazolin, in which case MeCN–water extracts). A Phenomenex Prodigy ODS3 column (150 mm × 3 mm;
(50:50, v/v) was used. For the stability study, a composite stock 5 ␮m particle size, 100 Å pore size), coupled to a C18 4 mm × 3 mm
standard solution containing 10 ␤-lactams (amoxicillin, ampicillin, guard column (both from Phenomenex; Torrance, CA, USA), was
cefazolin, cephapirin, cloxacillin, dicloxacillin, nafcillin, oxacillin, employed for the LC separation. The flow rate of the mobile phase
penicillin G, and penicillin V) was prepared at 100 ␮g/mL in water. was 300 ␮L/min. A Valco (Houston, TX, USA) divert valve was placed
For the evaluation of the LC mobile phase composition, a com- between the column outlet and MS source to eliminate the intro-
posite stock standard solution containing 13 ␤-lactams (the above duction of co-extracted matrix components into the MS instrument
10 analytes and 3 additional compounds: cefadroxil, cephalexin, prior and after elution of ␤-lactams.
and deacetylcephapirin) was prepared at 100 ␮g/mL in water. In the stability study, the binary mobile phase was composed of
This solution was diluted 1000-fold in water (to 100 ng/mL) and A, 0.1% FA in water and B, 0.1% FA in MeCN. A linear gradient, starting
DCCD was added at 1000 ng/mL to prepare a test solution of 14 from 2% B and going to 100% B within 15 min followed by a hold at
␤-lactams for the mobile phase evaluation study. For the recov- 100% B till 19.5 min, was used in the analytical method. Afterwards,
ery studies validating the method for multiresidue analysis of a wash method was performed for the injection of MeCN to prevent
␤-lactams in bovine kidney, one composite stock standard solution carry-over between samples. The wash method started at 100% B
(at 50 ␮g/mL in water) contained 11 ␤-lactams (amoxicillin, ampi- (held for 8 min), followed by a fast ramp (within 0.5 min) to 2% B,
cillin, cefazolin, cephalexin, cloxacillin, DCCD, deacetylcephapirin, which was held until 15.5 min to equilibrate the column for the
dicloxacillin, nafcillin, oxacillin, and penicillin G). Another compos- next run. Thus, the total cycle time was adjusted to take 36 min
ite stock standard solution (also at 50 ␮g/mL in water) contained (including two injections taking 0.5 min each) to analyze the 10
[13 C6 ]sulfamethazine (an internal standard) and two ␤-lactams different test solutions evaluated in the stability study in 6 h. For
(cefadroxil and penicillin V) used for quality control purposes the analysis of kidney extracts, a faster gradient in the analytical
(cefadroxil and penicillin V are generally not used in veterinary method was used, going from 2% B to 100% B in 10 min and holding
medicine). All stock standard solutions were stored in 1 mL portions at 100% B until 14.5 min. The wash method was also shorter: a 5-
in polypropylene tubes at −20 ◦ C. min hold at 100% B, followed by a fast ramp to 2% B, which was held
MeOH, MeCN, and hexane were high-purity grade solvents for until 10.5 min, resulting in a cycle time of 26 min.
residue analysis from J.T. Baker (Phillipsburg, NJ, USA) and Burdick In the mobile phase evaluation study, the percentage of FA in
& Jackson (Muskegon, MI, USA), respectively. Formic acid (FA) was both mobile phase components A and B was varied (x = 0.4, 0.2, 0.1,
obtained as a 98% solution for MS from Fluka (Buchs, Switzerland). 0.05, 0.025, 0.01, 0.005, and 0%, v/v). The mobile phase A was x%
Ultrapure water was obtained from a Barnstead water purification FA in water. The mobile phase B was x% of FA in (1) MeOH; (2)
120 K. Mastovska, A.R. Lightfield / J. Chromatogr. A 1202 (2008) 118–123

Table 1
Compound-specific LC–ESI(+)-MS/MS conditions for the tested ␤-lactams

Analyte tR (min) DP (V) FP (V) Precursor ion Product ion CE (V) CPX (V)

Deacetylcephapirin 7.8 47 143 382 152 37 12

226 27 16

Amoxicillin 8.0 25 129 366 349 13 10

208 19 14

Cefadroxil 8.1 24 110 364 114 29 10

208 15 16

Cephapirin 8.4 42 182 424 292 21 8

152 31 14

DCCD 8.4 66 240 549 183 43 18

241 29 18

Ampicillin 8.5 36 142 350 106 29 8

192 23 14

Cephalexin 8.6 27 117 348 158 13 12

174 19 14

Cefazolin 10.3 26 140 455 323 17 10

156 23 12

Penicillin G 11.9 36 180 335 160 17 14

176 17 12

Penicillin V 12.3 21 110 351 160 17 14

114 45 8

Oxacillin 12.5 31 160 402 160 19 14

243 17 8

Cloxacillin 12.8 31 160 436 277 19 8

160 19 12

Nafcillin 12.9 36 160 415 199 19 14

171 49 16

Dicloxacillin 13.4 31 150 470 160 19 12

311 19 10

tR , retention time; DP, declustering potential; FP, focusing potential; CE, collision energy; CXP, collision cell exit potential.

MeOH–MeCN (50:50, v/v); (3) MeCN. A linear gradient, starting kidney sample into a 50-mL disposable polypropylene centrifuge
from 2% B and going to 100% B within 10 min followed by a hold tube (Corning, Lowell, MA, USA); (2) add 100 ␮L of 1 ␮g/mL com-
at 100% B until the elution of the last analyte (dicloxacillin), was posite standard solution of [13 C6 ]sulfamethazine (serving as an
used for each A and B combination. Then, the column was well internal standard to compensate for volume changes), penicillin V
equilibrated to the same initial conditions (for repetitive analyses, and cefadroxil (a penicillin and a cephalosporin, respectively, serv-
n = 4) or to the initial conditions for the next tested mobile phase ing as quality control standards indicating potential problems in
composition. real analysis of unknown samples) in water; (3) add 2 mL water
The MS determination was performed in electrospray (ESI) posi- and 8 mL acetonitrile; (4) vortex briefly, shake vigorously for 5 min;
tive mode combined with monitoring of the most abundant MS/MS (5) centrifuge at 3450 rcf (5000 rpm using a RT6000B centrifuge
(precursor → product) ion transitions (dwell time of 75 ms for from Sorvall, Newtown, CT, USA) for 5 min; (6) decant the super-
each transition). The following MS conditions were used: entrance natant into a disposable polypropylene 15 mL tube (Corning, Lowell,
potential of 10 V, ionspray voltage of 4500 V, and ion source tem- MA, USA) with 500 mg of C18 sorbent; (7) vortex briefly, shake for
perature of 525 ◦ C. The curtain gas regulator was set at 40 psi with 30 s; (8) centrifuge at 3450 rcf for 1 min; (9) place 5 mL aliquot
optimum relative setting (in the Analyst software) of 11. The nebu- of the supernatant into a graduated tube; (10) evaporate extract
lizer and collision gas regulators were set at 90 psi with optimum to <1 mL; (11) make up the volume to 1 mL with water; and (12)
relative settings of 14 and 12, respectively. Table 1 gives compound- transfer the extract into the chamber of a Mini-UniPrep Syringe-
specific MS/MS conditions and retention times obtained using the less Filter vial (Whatman, Florham Park, NJ, USA) and compress the
mobile phase composition and gradient employed in the analysis filter (poly(vinylidene difluoride), PVDF, 0.45 ␮m) plunger to filter
of kidney extracts. For the internal standard propranolol, MS/MS the extracts, which are then ready for LC–MS/MS analysis that is
transitions m/z 260 → 116 and m/z 260 → 183 were monitored in performed on the same day as the sample extracts were prepared.
the stability study. For the internal standard [13 C6 ]sulfamethazine,
MS/MS transitions m/z 285 → 186, 124, and 114 were used in the
analysis of kidney extracts. 3. Results and discussion

3.1. Evaluation of the mobile phase composition for LC–MS

2.4. Sample preparation method for the analysis of ˇ-lactams in analysis of ˇ-lactams
bovine kidney
As mentioned in Section 1, MeOH is a very popular mobile
The optimized sample preparation procedure entailed the fol- phase solvent in LC–MS analysis of ␤-lactams in spite the poten-
lowing steps: (1) weigh 1 g of thoroughly homogenized bovine tial for analyte degradation. In our stability study experiments
 K. Mastovska, A.R. Lightfield / J. Chromatogr. A 1202 (2008) 118–123 121

performed at room temperature in dark vials, very rapid degrada-

tion of all tested ␤-lactams (somewhat slower for cephalosporins)
was observed in MeOH and a slower degradation rate in the 50:50
MeOH–water mixture (still about 40–60% of the tested analytes left
in the 50:50 MeOH–water solutions even after 5 days of storage). On
the other hand, very good stability was observed in water, MeCN,
and MeCN–water solutions, which were also found to be suitable
solvents for long-term storage of standard solutions at the lower
tested temperatures.
Similarly to MeOH, FA is a frequently employed mobile phase
modifier in LC–MS analysis of antibiotics, including ␤-lactams
[3–6,10–13,15–17]. Wiese and Martin [18] previously reported
degradation of penicillin G at lower pH conditions. In our sta-
bility study, the 0.1% addition of FA to water and other aqueous
solutions caused rapid degradation of monobasic penicillins, espe-
cially penicillin G and nafcillin. Degradation of other tested
penicillins (amoxicillin and ampicillin) and cephalosporins was
less pronounced (still about 40% of these analytes left in the
solutions even after 5 days in water with 0.1% FA at room
temperature).
To test the effect of MeOH and FA presence in the mobile phase
on analyte responses and method ruggedness, we designed experi-
ments involving the use of MeOH, MeCN, and MeOH–MeCN (50:50,
v/v) as the organic part of the LC mobile phase (part B). At the same
time, we varied the percentage of FA (0–0.4%, v/v) in both parts of
the mobile phase (mobile phase A was water with x% of FA).
Independent of the amount of FA in the mobile phase, MeCN pro-
vided overall better sensitivity for the tested ␤-lactams than MeOH.
The early eluting analytes (from deacetylcephapirin to cephalexin)
Fig. 1. Effect of the LC mobile phase composition on responses (peak heights) of (A)
were less affected by the organic mobile phase component, but deacetylcephapirin and (B) cloxacillin using different amounts of FA (0.005–0.4%)
2.2–4.1-fold higher maximum responses were achieved with MeCN in the mobile phase A (water) and B, which was MeOH or MeCN or MeCN–MeOH
versus MeOH for the later eluting analytes (from cefazolin to (50:50, v/v).
dicloxacillin). In the case of MeCN, highest overall sensitivity was
obtained with 0.1% FA (0.2% FA was the optimum for MeOH), which
also corresponded with the maximum sensitivity of the late eluting throughput. We divided the LC method into two parts. The first
analytes. Early eluters were less affected by the different amounts part served as an analytical method for injection of the sample and
of FA in the mobile phase. They overall gave the best responses at elution of the analytes. The LC gradient of the analytical method
0.01% FA addition, but signals of the late eluters were very low at ends at 100% B and holds there until the last analyte is recorded
these conditions. Thus, we found 0.1% FA in MeCN to provide the by the MS instrument. The second method was a wash method
best overall sensitivity for LC–ESI(+)-MS/MS multiresidue analysis that started at 100% B and served for injection of a solvent (MeCN),
of ␤-lactams as compared to other tested mobile phase composi- washing of the system, and equilibration of the column to the initial
tions. Fig. 1 demonstrates the above discussed results, comparing conditions of the analytical method. In this way, the in-between
responses (peak heights) of an early eluting analyte deacetyl- samples injections of blanks contribute minimally to the overall
cephapirin and a late eluting analyte cloxacillin obtained with sample throughput.
different % FA in the aqueous and organic (MeCN, MeCN–MeOH,
or MeOH) mobile phase components.
In terms of the speed of the analysis, MeCN provided the over-
all fastest analyte elution as compared to MeOH and MeOH–MeCN.
Fig. 2 shows retention times obtained for the last eluting analyte
(dicloxacillin) using different mobile phase compositions. A similar
trend can be observed for all late eluting analytes. For FA concen-
trations in MeCN ≤ 0.05%, the retention times shifted dramatically
to higher values. Thus, 0.1% FA addition also ensured good robust-
ness because only very slight changes in analyte retention times
(and also responses) would occur with a less accurate addition of
0.1% FA. Furthermore, lower FA concentrations (≤0.05%) also led
to deterioration of peak shapes (significant peak broadening) for
cefazolin and DCCD, which shifted to longer retention times more
readily than the other ␤-lactams, thus resulting in different elution
orders (e.g., DCCD eluting after ampicillin and cephalexin) as the %
FA decreased in the mobile phase.
To prevent carry-over in LC–MS analysis, analysts often inject
a blank solvent after each (or a high-level) sample using the same
Fig. 2. Effect of the LC mobile phase composition on dicloxacillin retention time
LC method to wash the system (mainly the needle, injection port, using different amounts of FA (0.005–0.4%) in the mobile phase A (water) and B,
and valve). This procedure, however, significantly reduces sample which was MeOH or MeCN or MeCN–MeOH (50:50, v/v).
122 K. Mastovska, A.R. Lightfield / J. Chromatogr. A 1202 (2008) 118–123

3.2. Streamlined method for multiresidue analysis of ˇ-lactams almost the entire extract after the dispersive-SPE clean-up, and a
in bovine kidney filtration of the concentrated extract using a syringe filter). In our
modification, the internal standard [13 C6 ]sulfamethazine is added
The previously reported sample preparation method [4] is based to the homogenized sample before the extraction step together
on the extraction of homogenized tissue with MeCN–water (4:1, with two ␤-lactams (penicillin V and cefadroxil), which serve for
v/v), followed by a dispersive SPE clean-up using octadecyl silica method performance controls. These two compounds are not gen-
(C18 ) as a sorbent. After the clean-up step, the extract is concen- erally used in veterinary medicine, thus can serve as quality control
trated by evaporation, resulting in the final extract in water. The 4:1 standards indicating potential problems in the ␤-lactam method
MeCN–water mixture used as 10 mL per 1 g sample was found to be performance. The use of a suitable internal standard allows working
optimal for both ␤-lactam extraction and sample deprotenization with sample aliquots rather than the entire extract, which reduces
[19]. the time required for the extract concentration step. Also, the use
Dispersive-SPE technique involves a simple mixing of the extract of an aliquot in the case of the sample transfer after the disper-
with a sorbent that removes matrix interferences but does not sive SPE clean-up helps to avoid the presence of C18 particles in the
retain the analytes [20]. C18 sorbent removes highly lipophilic com- final extract. The final extract (after the evaporation step) contains
pounds [21] (we found 500 mg per 10 mL of the kidney extract mainly water (a weak solvent), thus the presence of C18 particles
to be an optimum in terms of clean-up and cost-effectiveness). would potentially lead to lower analyte recoveries due to their
Other effective sorbents applied in the dispersive format include retention/partition on the sorbent.
primary secondary amine and graphitized carbon black that are To further simplify the procedure, we used Mini-UniPrep
employed mainly for removal of fatty acids and pigments (and some syringeless filters for the filtration of the final extracts instead of
other matrix-coextractives), respectively. Unfortunately, PSA and syringe filters. The syringeless filters consist of two parts: a cham-
GCB cannot be used in the analysis of ␤-lactams because they con- ber and a filter plunger that together form an autosampler vial that
tain a carboxylic group that is being retained by PSA, resulting in can be used for sample storage and for sample introduction using
lower analyte recoveries. GCB has a high affinity towards planar common autosamplers. The extract is transferred into the chamber.
molecules, therefore also highly retains nafcillin and DCCD, which Then, the filter plunger is compressed to filter the extract and to seal
recoveries were only about 20–30% when 50 mg GCB was added to the vial. Various filtration media can be used depending on the sam-
1 mL of 100 ng/mL ␤-lactam solution in 4:1 MeCN–water. ple type. We tested two media suitable for aqueous extracts: nylon
In our effort to further streamline the method and make it more and PVDF; and found nylon to be problematic for certain analytes
cost-effective, we also tested the use of hexane as an alternative to (lower recoveries obtained with nylon especially for amoxicillin,
the dispersive C18 clean-up. Hexane is immiscible with MeCN and penicillin G, and nafcillin). Thus, PVDF (0.45 ␮m) was employed as
water, thus can be used for removal of lipophilic compounds from a filtration medium in the syringless filters for filtration of bovine
MeCN–water extracts [22]. In our evaluation, we added 5 mL of hex- kidney extracts.
ane to 10 mL of kidney extract, shaked well, and removed the upper Table 2 gives ␤-lactam recoveries obtained at three spiking lev-
hexane layer after a centrifugation step. Using the same extract, els (10, 50, and 250 ng/g) in bovine kidney. Average recoveries
we also performed the dispersive-C18 clean-up and the hexane were 87–103%, except for DCCD, which had an average recovery
clean-up followed by the dispersive-C18 procedure. In each case, of 60%. The previous method gave average recoveries of 70–75%
the amount of matrix co-extractives left as a residue after extract (58% for DCCD), thus the streamlined procedure (see Section 2.4
evaporation was determined gravimetrically and compared versus for the details) improved the analyte recoveries, mainly by account-
the situation without any clean-up. The dispersive-C18 clean-up ing for volume changes and losses during the sample preparation.
removed about threefold more matrix co-extractives than hexane. DCCD is the only free form of ceftiofur that can be monitored in
Also, no additional clean-up was achieved using the combination a multiresidue method with other ␤-lactams. Upon intramuscular
of hexane and C18 as compared to C18 alone. injection, ceftiofur metabolizes rapidly to desfuroylceftiofur, which
A quantitative limitation of the previously introduced method has a free thiol group that binds via disulfide bonds to cysteine
[4] was an insufficient control over volume changes and losses residues in peptides and proteins [23–25]. The lower DCCD recovery
during the procedure because the internal standard (penicillin V) can be explained by the disulfide bond between desfuroylceftiofur
was added to the final extract just before the LC–MS/MS analysis. and cysteine undergoing exchange with protein thiols or disulfide
Thus, the internal standard could not account for variable water bonds in the kidney tissue, resulting in losses during the extrac-
content in the kidney samples and losses during the three sample tion/deproteinization step and/or nondetection in the LC–MS/MS
transfer steps (a decantation step after the extraction, a transfer of method [4,26]. To determine total residues of ceftiofur in a single-

Table 2
␤-Lactam recoveries and relative standard deviations (RSDs) obtained in bovine kidney samples fortified at 10, 50, and 250 ng/g (DCCD at 100, 500, and 2500 ng/g)

Analyte Recovery (RSD, %)

10 ng/g (n = 6) 50 ng/g (n = 6) 250 ng/g (n = 6) Overall (n = 18)

Deacetylcephapirin 100 (11) 96 (11) 98 (7) 98 (9)

Amoxicillin 88 (6) 85 (12) 87 (8) 87 (9)
DCCD 54 (12) 66 (21) 59 (5) 60 (16)
Ampicillin 91 (9) 87 (16) 89 (9) 89 (11)
Cephalexin 89 (10) 88 (14) 84 (12) 87 (11)
Cefazolin 100 (5) 103 (5) 104 (7) 102 (6)
Penicillin G 95 (7) 110 (9) 104 (10) 103 (10)
Oxacillin 98 (4) 97 (6) 102 (7) 99 (6)
Cloxacillin 97 (7) 94 (6) 102 (7) 98 (7)
Nafcillin 97 (5) 101 (7) 107 (8) 101 (8)
Dicloxacillin 96 (3) 92 (7) 102 (8) 97 (8)
 K. Mastovska, A.R. Lightfield / J. Chromatogr. A 1202 (2008) 118–123 123

residue method, desfuroylceftiofur has to be released from the References

disulfide bonds by their reduction, followed by stabilization of the
thiol group by derivatization (e.g., acetylation) [26]. [1] V. Hormazabal, M. Yndestad, J. Liq. Chromatogr. Relat. Technol. 21 (1998) 3099.
[2] Y. Ito, Y. Ikai, H. Oka, H. Matsumoto, Y. Miyazaki, K. Tabeka, H. Nagase, J. Chro-
matogr. A 911 (2001) 217.
4. Conclusions [3] C.K. Fagerquist, A.R. Lightfield, Rapid Commun. Mass Spectrom. 17 (2003) 660.
[4] C.K. Fagerquist, A.R. Lightfield, S.J. Lehotay, Anal. Chem. 77 (2005) 1473.
[5] M. Becker, E. Zittlau, M. Petz, Anal. Chim. Acta 520 (2004) 19.
A previously reported multiresidue method for the analysis of [6] K. Granelli, C. Branzell, Anal. Chim. Acta 586 (2007) 289.
important ␤-lactams in bovine kidney has been further stream- [7] T.A.M. Msagati, M.M. Nindi, Food Chem. 100 (2007) 836.
[8] K.L. Tyczkowska, R.D. Voyksner, A.L. Aronson, J. Chromatogr. 594 (1992) 195.
lined by using an aliquot rather the entire extract and by employing [9] M.I. Page, Adv. Phys. Org. Chem. 23 (1987) 165.
syringeless filter vials for final extract filtration and storage. The [10] T. Goto, Y. Ito, S. Yamada, H. Matsumoto, H. Oka, J. Chromatogr. A 1100 (2005)
recoveries have been improved by a proper control of the volume 193.
[11] D.N. Heller, M.A. Ngoh, Rapid Commun. Mass Spectrom. 12 (1998) 2031.
changes during the sample preparation. Based on the results of the
[12] F. Bruno, R. Curini, A. Di Corcia, M. Nazzari, R. Samperi, J. Agric. Food Chem. 49
stability study and LC mobile phase tests, MeOH has been elimi- (2001) 3463.
nated from the entire method, including the LC–MS/MS analysis. [13] D. Hurtaud, B. Delepine, P. Sander, Analyst 119 (1994) 2731.
Water, MeCN, and their mixtures can be used in the standard solu- [14] D.M. Holstege, B. Puschner, G. Whitehead, F.D. Galey, J. Agric. Food Chem. 50
(2002) 406.
tion and sample preparation. The best overall LC–ESI(+)-MS/MS [15] R. Lindberg, P.-Å. Jarnheimer, B. Olsen, M. Johansson, M. Tysklind, Chemosphere
performance was achieved by using 0.1% FA as an additive in the 57 (2004) 1479.
mobile phase (in water and in MeCN). However due to the sta- [16] S. Riediker, A. Rytz, R.H. Stadler, J. Chromatogr. A 1054 (2004) 359.
[17] E. Daeseleire, H. de Ruyck, R. van Renterghem, Rapid Commun. Mass Spectrom.
bility issues, FA should not be added to the final extracts prior to 14 (2000) 1404.
the LC–MS/MS analysis to match the mobile phase composition. [18] B. Wiese, K. Martin, J. Pharm. Biomed. Anal. 7 (1989) 67.
To prevent carry-over in the LC–MS/MS analysis, the LC method [19] W.A. Moats, R.D. Romanowski, J. Agric. Food Chem. 46 (1998) 1410.
[20] M. Anastassiades, S.J. Lehotay, D. Štajnbaher, F.J. Schenck, J. AOAC Int. 86 (2003)
was divided into two parts: one serving as an analytical method 412.
for injection of the sample and elution of the analytes and the [21] S.J. Lehotay, K. Mastovska, S.J. Yun, J. AOAC Int. 88 (2005) 630.
other one, starting at a highly organic mobile phase composition [22] K. Mastovska, S.J. Lehotay, J. Agric. Food Chem. 54 (2006) 7001.
[23] P.S. Jaglan, B.L. Cox, T.S. Arnold, M.F. Kubrick, D.J. Stuart, T.J. Gilbertson, J. Assoc.
(100% MeCN with 0.1% FA in our case), being dedicated for injec- Off. Anal. Chem. 73 (1990) 26.
tion of a solvent (MeCN), washing of the system, and equilibration [24] M.G. Beconi-Barker, R.D. Roof, L. Millerioux, F.M. Kausche, T.J. Vidmar, E.B.
of the column to the initial conditions of the analytical method. In Smith, J.K. Callahan, V.L. Hubbard, G.A. Smith, T.J. Gilbertson, J. Chromatogr.
B 673 (1995) 231.
this way, a blank solvent is injected after each sample, but these
[25] C.K. Fagerquist, R.R. Hudgins, M.R. Emmett, K. Hakansson, A.G. Marshall, J. Am.
in-between injections contribute minimally to the overall sample Soc. Mass Spectrom. 14 (2003) 302.
throughput. [26] R.E. Hornish, P. Hamlow, S. Brown, J. AOAC Int. 86 (2003) 30.