JOURNAL OF MASS SPECTROMETRY, VOL.

31, 418-426 (1996)

Hyphenation of Coupled-column Liquid

Chromatography and Thermospray Tandem Mass
Spectrometry for the Rapid Determination of
P2-Agonist Residues in Bovine Urine Using
Direct Large-volume Sample Injection.
Set-up of Single-residue Methods for Clenbuterol
and Salbutamol

Elbert A. Hogendoorn and Piet van Zoonen

National Institute of Public Health and Environmental Protection (RIVM), P.O. Box 1, 3720 BA, Bilthoven, The Netherlands

Aldo Polettinit and M a r i a Montagna

Institute of Legal Medicine, Laboratory of Forensic Toxicology, University of Pavia, Via Forlanini 12, 27100 Pavia, Italy

In a previous work, coupled-column L C (LC/LC) with direct large-volume sample injection for the analysis of
fl,-agonists in bovine urine was investigated, in view of its combination with thermospray (TSP) tandem mass
spectrometric (MS/MS) detection. In this work, the potential of both conventional L C and LC/LC coupled to
TSP-MS/MS for the rapid determination of P,-agonists in bovine urine was evaluated. I t was first established that,
in order to avoid disturbances in TSP-MS/MS detection due to changes in pressure and flow rate of the mobile
phase, LC/LC is preferable to conventional LC. Then, selecting clenbuterol and salbutamol as model compounds
for the class of P,-agonists, two single-residue LC/LC/TSP-MS/MS methods were developed, permitting the deter-
mination of the analytes in bovine urine samples with a limit of quantification of at least 0.1 ng, ml-', with a
sample throughput of 4 h-I. The methods were validated by analysing both spiked urine samples and samples
containing clenbuterol or salbutamol residues. Mean recoveries from urine samples spiked with clenbuterol a t levels
of 5 and 0.22 ng ml-' were 107% (n = 9, RSD = 7.4%) and 102% ( n = 6, RSD = 14%), respectively. A bovine
urine sample containing 1.28 ng ml-' clenbuterol residue (previously quantified by GC/MS) was re-analysed by
LC/LC/TSP-MS/MS on four different days (n = 3, each day) and yielded a mean concentration of 1.31 ng ml-',
with a reproducibility of 8.4%. Bovine urine samples spiked with salbutamol at levels of 10.8 and 0.55 ng ml-'
yielded mean recoveries of 101% (n = 3, RSD = 1%) and 89% (n = 3, RSD = 13%), respectively. In bovine urine
samples containing salbutamol residues, the concentrations of unconjugated salbutamol were found to be in the
range 8-57 ng ml-'.

KEYWORDS: coupled-column liquid chromatography; thermospray tandem mass spectrometry; P,-agonists;

urine; direct injection

biological matrices, followed by confirmation of posi-

INTRODUCTION
fi,-Agonists are widely used in zootechnics to enhance
meat production, and active residues of the drugs may
be present in meat, milk and processed d e r i v a t i v e ~ l - ~
Owing to their 'anabolic' and stimulating properties, b2-
agonists are also used by athletes for doping pur-
p o s e ~ . This
~ , ~ raises public health concerns and high-
lights the need for effective control strategies.
The diagnosis of illicit administration of p,:agonists
is mainly based at present on screenings by immuno-
chemical methods7-'' which can detect j?,-agonists resi-
dues in urine, vitreous humor, meat extracts and other
7 Author to whom correspondence should be addressed.
tives by gas chromatography/mass spectrometry (GC/
MS).11-26 In addition to the time-consuming and
laborious sample preparation (involving hydrolysis,
extraction, purification and derivatization) necessary
prior to GC/MS analysis, this analytical strategy may,
in some cases, lack the required selectivity and sensi-
tivity. Hence the availability of more rapid, sensitive
and reliable methods for the assay of fi,-agonists in
biological samples would significantly increase the eft%
cacy of control procedures.
Liquid chromatography (LC) provides significant
advantages over GC for the analysis of j?,-agonists,
including compatibility of the biological matrix with the
chromatographic system and the lack of any need for
derivatization. However, its a p p l i ~ a t i o n ~ . ' ~ ha
- ~ s' been
limited by the inadequate selectivity and sensitivity of

CCC 1076-5 174/96/040418-09 Received 9 November 1995

0 1996 by John Wiley & Sons, Ltd. Accepted 3 January 1996
 LC/LC/TSP-MS/MS ANALYSIS OF B,-AGONISTS IN URINE
common LC detectors on the one hand and by the diffi-
culties of interfacing LC with selective detection
methods such as mass spectrometry (MS) on the other.
The development of effective LC/MS interfaces, such
as thermospray (TSP) and electrospray (ESI), has
opened up new possibilities in the application of
LC/MS to the analysis of j?,-agonists, and some
methods have been proposed for the determination of
residues of these drugs in bovine urine.32p34Blanchflo-
wer and Kennedy32 proposed a method based on LC/
TSP-MS for the assay of clenbuterol, while Debrauwer
and B o r i e ~have
~ ~ applied ESI-MS, after off-line LC
separation, to the detection of clenbuterol residues.
Flow injection analysis (FIA) followed by TSP-tandem
mass spectrometry (MS/MS) has been used by van
Rhijn et ~ 1 to .screen
~ j?,-agonists.
~ Recently, Doerge et
al. 3 s have proposed an LC/MS/MS method using an
atmospheric pressure chemical ionization (APCI) inter-
face for the screening of j?,-agonists in plasma extracts.
Tandem mass spectrometry yields very high detection
selectivity. By selecting appropriate parent ions in the
first mass analyser and by monitoring specific fragments
(product ions) produced by collision of parent ions with
an inert gas (e.g. argon) with the second mass analyser,
the signal-to-noise ratio is increased considerably, thus
increasing the detectability of the analytes. Although
MS/MS can be used for analytical purposes (with a
direct probe or with FJA,34 for example), a purification
step prior to analysis is unavoidable, particularly in the
case of bovine urine samples containing non-volatile
compounds (e.g. inorganic salts). Indeed: almost all the
LCjMS and MS/MS procedures available in the liter-
ature adopt off-line sample preparation procedures
based on solid-phase extraction (SPE)33or liquid-liquid
partitioning or without purification by back-
e~traction.~~
Column-switching LC can easily provide automated
clean-up procedures. However, recent reviews36337
reveal that the number of applications of this technique
on-line with TSP-MS detection are limited. Precolumn
switching has been successfully applied as a phase
switching system (PSS), providing both purification and
solvent compatibility with the TSP i n t e r f a ~ e . ~ ~ . ~ ~
We have recently investigated method development
strategies as applied in pesticides residue analysis using
coupled-column LC (LC/LC)40-44 in automated single-
residue (SRM) and multi-residue analysis (MRM) of P2-
agonists in urine samples.45 The major advantage of
LC/LC is the utilization of the separation power of the
first separation column. This feature is particularly
favourable in the case of polar analytes with little C,,
retention, such as terbutaline or salbutamol. Compared
with automated sample processing procedures, such as
precolumn switching (PC/LC) and SPE/LC using
hydrophobic materials, LC/LC offers the advantages of
efficient clean-up and chromatographic analysis in com-
bination with large-volume injection of aqueous
samples.
In this study, we investigated the potential of the
combination of both conventional LC and of LC/LC
with TSP-MS/MS in the analysis of two B,-agonists
selected as model compounds : clenbuterol and salbuta-
mol. Using the results of our previous LC work4' and
earlier mass spectrometric experimental the
419
potential of these hyphenated techniques in the rapid,
on-line determination o f B,-agonists in bovine urine
was evaluated.
EXPERIMENTAL
Chemicals and reagents
-
Clenbuterol HCl, salbutamol- H,SO,, and terbutaline
-H,SO, (internal standard for salbutamol) were
obtained from Biomedica Foscama (Rome, Italy), Glaxo
Wellcome (London, UK) and Pharmacia (Milan, Italy),
respectively. A 1.00 mg ml-' stock solution of each
analyte was prepared in methanol and stored at
- 20 "C. Stock solutions were diluted in HPLC-grade
water and kept in a refrigerator at about 4°C. A 5 pg
-
ml- dilution of terbutaline H,S04 was prepared for
-
direct addition to urine samples. Clenbuterol-d, HCl,
obtained from Rijksinstituut voor Volksgezondheid en
Milieu (Bilthoven, The Netherlands) was used as an
internal standard for clenbuterol. Clenbuterol-d, was
dissolved in HPLC-gradc water at a concentration of 4
pg m l '.~
Methanol (UV-grade), water (HPLC-grade) and tri-
ethylamine (p.a.) (99.5%) were obtained from Carlo
Erba (Milan, Italy) and ammonium acetate (p.a.) (99%)
and formic acid (p.a.) (96%) from Aldrich (Milan, Italy).
Mobile phases containing 0.1 .M ammonium acetate,
0.01 M triethylamine and 0.17 M formic acid consisting
of methanol-water (30: 70, v/v) and methanol-water
(18:82, v/v) were used for clenbuterol and salbutamol
analysis, respectively. Before use, a freshly prepared
mobile phase was filtered through a type HA 0.45 pm
filter (Millipore-Waters, Milford, MA, USA).
Equipment
The LC/MS instrumentation, depicted schematically in
Fig. 1, consisted of the following components: a
Rheodyne (Cotati, CA, USA) model 7125 manual injec-
tor, IV; a Valco (VICI, Schenkon, Switzerland) EC4W
security switching valve controlled by the MS system,
SV; two Rheodyne model 7010 switching valves, HV-A
and HV-B, installed and controlled by a MUST device
from Spark Holland (Emmen, The Netherlands); a
Waters Model 510 isocratic LC pump, IP; a Waters
Model 600 MS gradient LC pump equipped with a
flow-rate stabilization option and a helium degassing
system, GP; a Waters Model 441 UV detector set at
280 nm, UV; a Perkin-Elmer (Norwalk, CT, USA)
Model 561 recorder; a Finnigan MAT (San Jose, CA,
USA) TSQ-700 tandem mass spectrometer, MS/MS,
equipped with a Finnigan MAT TSP-2 thermospray
interface ; and a cold trap containing liquid nitrogen,
CT.
Two 50 x 5.6 mm I.D. columns packed with 3 pm
Microspher C,, particles from Chrompack (Bergen op
Zoom, The Netherlands) were used as the first (C-1) and
second (C-2) columns. A Chrompack 10 x 3 mm I.D.
420 E. A. HOGENDOORN ET AL.

M
U

Figure 1 . Overview of the LC/MS system. I V = injection valve with 5 0 0 pl loop; SV =security valve; M U S T = two-valve switching device:
-
I P = isocratic LC pump; G P =gradient LC pump; UV UV detector; TSP-MS/MS -thermospray tandem mass spectrometer; CT -solvent
cold trap; C-1, C - 2 =first and second separation columns; M = mobile phase used on C-1 and C-2; W =waste.

RP-guard column was inserted before each separation
column. A flow rate of 1.3 ml/min of the appropriate
mobile phase (see above) was used with both C-1 and
c-2.
The TSP interface was operated in the buffer ioniza-
tion mode (filament and discharge off). The temperature
of the source block was 200°C and the vaporizer tem-
perature was 105 and 103 "C for clenbuterol and salbu-
tamol analysis, respectively. The repeller was set to 60 V
in both cases.
MS/MS analysis was performed in the selected reac-
tion monitoring, precursor ion mode. The reactions
monitored were m/z 277 203 (clenbuterol), m/z
-+
Quantification of clen buterol or salbutamol in bovine
urine samples was performed by comparing the
clenbuterol/clenbuterol-d, or salbutamol/terbutaline
TSP-MS/MS response peak-area ratio obtained for the
sample with that obtained for the injection of standards.
In the case of clenbuterol, a calibration graph was pre-
pared by randomly injecting standards of clenbuterol at
concentrations of 0.2, 1, 5, 10 and 20 ng m l - ' (each
point in triplicate). After each day of analysis both
columns were washed with 20 ml of methanol, while the
TSP interface was washed with methanol (5 ml) fol-
lowed by water (5 ml).

283 + 203 (clenbuterol-d,), m/z 240 -+ 166 (salbutamol)

and m / z 226 + 152 (terbutaline). The collision gas
(argon) pressure and collision energy were 1.7 mTorr (1
Torr = 133.3 Pa) and -25 eV, respectively, for clenbu-
terol analysis and 1.2 mTorr and - 15 eV, respectively,
for salbutamol analysis, while the multiplier was set to
2000 and 1750 eV, respectively.
The instrumental conditions adopted for clenbuterol
and salbutamol analysis are summarized in Table 1.
Urine sample pretreatment
Bovine urine samples (about 5 ml) were filtered through
0.22 pm Millex-GS filters (Millipore). To a volume of
4.5 ml of sample (filtered bovine urine or an aqueous
solution of the analytes), 30 pl of the appropriate inter-
nal standard solution (see above) were added prior to
injection.

Table 1. Experimental LC/LC/TSP-MS/MS conditions for the determination

of clenbuterol and salbutamol
Stage Parameter Clenbuterol Salbutamol

LC/LC Injection volume (PI) 500 500

Clean-up time (min) 2.8 1.2
Transfer time (min) 0.75 0.50

TS P Repeller (V) 60 60
Vaporizer temperature ("C) 105 103
Source block temperature ("C) 200 200
MS/MS Collision energy (eV) -25 15
Collision gas (argon) pressure (mTorr) 1.7 1.2
Reactions monitored (rn/z) 277 --t 203 240 + 1 6 6
283 -+ 203 226 --t 1 5 2
 LC/LC/TSP-MS/MS ANALYSIS OF P,-AGONISTS IN URINE
RESULTS AND DISCUSSION
General aspects
Our previous work4' was mainly focused on finding
suitable (reversed phase) LC/LC conditions for the effi-
cient separation of #I,-agonists and on the on-line pro-
cessing of urine samples including large-volume
injection prior to the adoption of TSP-MS/MS as used
in the present study. Basically, two approaches were
employed : a single-residue (SRM) and a multi-residue
methodology (MRM). UV detection was used to
monitor the analytes during the optimization pro-
cedures, As regards MRM, the optimization procedure
provided adequate conditions to allow the separation of
eight b,-agonists (isoprenaline, cimaterol, terbutaline,
salbutamol, fenoterol, ractopamine, clenbuterol and
mabuterol) with a large difference in C,, retention. An
efficient separation was attained in less than 15 min by
means of either linear, one-step or two-step gradient
elution. It was also established that the LC/LC system
could easily perform the on-line analyses of numerous
bovine urine samples (200 pl), by applying a clean-up
volume of at least three times the void volume (V,) of
the first column and a transfer volume of about 13 ml to
transport all analytes from the first to the second
column. It was clearly demonstrated that under these
conditions UV detection is insufficiently specific to
detect all target analytes in urine samples.
However, in the case of SRM focused on the determi-
nation of one analyte, the selectivity and sensitivity
attainable by LC/LC could be enhanced considerably.
By applying (i) large-volume sample injection (1.5 ml),
(ii) a very accurately adjusted clean-up volume and a
small transfer volume (0.15 ml), and (iii) a pH-based step
gradient involving an acidic (pH 3.5) and a slightly basic
(pH 7.5) mobile phase on the first and second columns,
respectively, clenbuterol could be determined at the pg
I-' level in bovine urine samples in about 10 min with
UV detection (245 nm).
As a first step in investigating the potential of LC/
LC/TSP-MS/MS for the automated determination of
b,-agonists in bovine urine, this study was focused on
SRM. Clenbuterol and salbutamol were selected as
model compounds to represent the class of #I,-agonists
because of their different polarities. Furthermore, the
availability of matrices containing residues of these two
analytes allowed the validation of methods on real
samples .
Selection of LC-switching mode
Considering the extreme selectivity of MS/MS detec-
tion, the use of two high-efficiency separation columns
as a first approach may appear to be excessive. Hence,
to avoid unnecessary complexity in the LC system, the
option of performing on-line analysis using a single
column was first investigated.
The following configuration was adopted, using the
UV detector to simulate the TSP interface: the gradient
LC pump (GP) was used to deliver the mobile phase to
the TSP interface, while the isocratic LC pump (IP)
42 1
delivered the mobile phase to the column to perform
the clean-up. Just before the elution of the analyte, the
gradient pump (GP) was temporarily switched on-line
with the column to introduce the LC effluent into the
TSP. However, under these conditions, the pressure and
the baseline monitored with the UV detector were
highly unstable. The flow-rate stabilization option avail-
able on the GP proved to be ineffective, requiring too
long to stabilize the abrupt and large change in pressure
occurring when the GP was switched on-line with the
column. This has deleterious implications for interfacing
LC with TSP-MS, as fluctuations in pressure and flow
rate may have a dramatic influence on the ionization
processes in the TSP interface, and consequently on the
stability of ion current.47 We attempted to improve this
situation by placing a 'dummy' column, similar to the
analytical column, between the GP and TSP to over-
come the large change in pressure when the GP is
switched on-line with the column. With this configu-
ration, the differences in flow rate and pressure were
considerably reduced, but nonetheless a significant
baseline drift was still observed on the UV detector.
Although not experimentally verified on the TSP inter-
face, it is highly probable that this would adversely
affect the ion current stability, making the single-
column switching approach unsuitable for use with
TSP-MS/MS.
Disturbances in mobile phase delivery during intro-
duction of the analyte into the TSP interface can be
avoided by using LC/LC (see Fig. 1). There is a tempo-
rary increase in pressure when the two columns (C-1
and C-2) are switched on-line during the transfer of the
analyte from C-1 to C-2. However, as verified by moni-
toring the read-out of the pressure and the stability of
the baseline by UV detection, the retention of the
analyte on C-2 is sufficient to provide time to attain
stable conditions during TSP-MS/MS detection. It can
therefore be concluded that, in addition to permitting
large-volume injection and on-line clean-up, an impor-
tant feature of LC/LC is to provide stable elution condi-
tions for TSP analysis.
Set-up of TSP-MS conditions
Previous and literature data33,34indicated that
electrolyte-mediated ionization (filament and discharge
off) is a suitable ionization mode for the MS/MS
analysis of #I,-agonists. Under electrolyte-mediated ion-
ization, in fact, most p,-agonists give the protonated
molecular ion, [MH]', as the base peak in the spec-
trum and little fragmentation is observed. In terms of
sensitivity this is a favourable condition for MS/MS
analysis, as almost all the signal in the TSP mass spec-
trum is carried by the precursor ion. In addition to
[MH]', an intense peak due to the loss of water is
present in the TSP mass spectrum of the more polar
#I,-agonists, such as salbutamol and r i t ~ d r i n e . ~ ~
It is known that vaporizer temperature (VT) is a criti-
cal parameter for successful TSP/MS. It is dependent on
mobile phase composition, and particularly on the per-
centage of organic modifier, and also on flow rate and
the analyte, and should be optimized to yield maximum
sensitivity for the target compound. We selected VT for
422 E. A. HOGENDOORN ET AL.

+%v?*! -b +Ssngatdd [MH - 741' ion (m/z 166), and a corresponding

I \ I
I I
I \ increase in the abundance of a fragment at m/z 148.
A B C A B C A

SRM for clenbuterol

0.05 AU (280 nrn)
On the basis of earlier a mobile phase of
methanol-buffer (30: 70, v/v) was chosen, yielding a
capacity factor ( k ) of about 5 for clenbuterol. It must be

-
emphasized that columns packed with 3 pm C,, par-
ticles are particularly appropriate in this type of
analysis. With a flow rate of 1.3 ml min-' (selected for
adequate TSP interfacing), this type of column behaves
similarly to a high-speed column, reducing both the
elution volume and retention time of the anayte.
TSP-MS/MS conditions were determined by FIA (20
p1 loop injections) as mentioned above. Under these
0 4 8 12 min conditions, the limit of quantification (signal-to-noise
Figure 2. UV monitoring of the mobile phase ( M ) flowing on ratio (S/N) = 10) was about 10 pg of clenbuterol,
C - I , illustrating the sample sequence of the clenbuterol analysis Taking into account the increase in peak volume after
using LC/LC/TSP-MS/MS. A =injection of sample on C-1; A-
6 =clean-up time on C - I ; B-C = transfer time (C-I on-line with
chromatographic analysis, we calculated that a sample
C-2); C-A = TSP-MS/MS analysis of clenbuterol (C-1 off-line injection volume of 0.2 ml would be sufficient to detect
with C-2). clenbuterol at the required level of 0.1 ng ml-'.
However, to increase the robustness of the procedure, a
sample injection volume of 0.5 ml was selected.
both SRMs by FIA of clenbuterol and salbutamol in The column switching conditions, that is, the clean-
the respective mobile phases, increasing VT from 90 to up time on C-1 and the transfer time on the analyte-
120°C (in 5°C steps) and measuring absolute peak containing fraction from C-1 to C-2, are easily
areas. At the beginning of each session, the optimum determined by injecting 0.5 ml of a standard solution
VT was checked by FIA and adjusted if needed. An without column switching and monitoring the signal by
adjustment of 2 "C with respect to the previous value UV detection (see Fig. 1). It was our intention to
was sometimes required. perform a cleaning of C-1 with methanol after the
analysis of urine as in our previous As shown in
Fig. 1, the valve switching system was connected so that
Set-up of MS/MS conditions the gradient pump could perform this step after the
analysis of a sample. However, because of the high
Combined loss of water and methylpropene from the selectivity of the detection mode, we investigated the
protonated molecular ion to yield an ion of mass possibility of omitting this step. Experiments indicated
[MH -~741 is the most important fragmentation
+
that, when analysing a standard solution after each
pathway in the collisionaIly activated dissociation urine sample, a C-1 washing step was superfluous,
spectra of B,-agonists containing a tert-butylamino allowing a reduction in analysis time. In addition, accu-
group (e.g. clenbuterol, salbutamol and rate settings of the clean-up and transfer time were also
t e r b ~ t a l i n e ) . We
~ ~ ,therefore
~~ optimized the collision unnecessary, making the system very robust to the
gas (argon) pressure and collision energy to yield the changes in retention time expected because of contami-
maximum abundance of the [MH - 741 product ion +
nation of C-1 after repeated injections in the absence of
for each analyte. Selecting a collision gas pressure in the a washing step. The analytical strategy that was finally
range 1.2-1.7 mTorr, the optimum collision energy was adopted, consisting of analysis of a sample followed by
found to be -25 eV for clenbuterol and - 15 eV for analysis of standard, is illustrated in Fig. 2, which shows
salbutamol. In the case of salbutamol, higher collision the UV signal monitored on C-1. Under these condi-
energies resulted in a reduction in the abundance of the tions the total time of analysis is only 7 min for the
sample plus additional 7 min for the standard.
The linearity of the final procedure (see Experimental)
_ _ _ ~ ___
was tested by randomly analysing standards of clenbu-
Table 2. Clenbuterol recoveries from bovine urine terol with concentrations in the range 0.2-20 ng m1-I.
Spiked level No of Mean recovery RSD
The calibration plot obtained (y = 0 . 1 6 2 ~- 0.0189)
Day (ngim1) experiments We) ("/I showed good linearity over a two-decades range
1 0.22 3 106 9.5 (r2 = 0.9993). However, further experimental work
3 0.22 3 97 16.8 revealed that the precision of analysis of urine samples
1 5.3 3 107 7.0 containing less than 1.0 ng ml-' was improved by
3 5.3 6 108 7.8 applying a one-point calibration, using a standard con-
taining a concentration of clenbuterol close to that in
Overall recovery (YO) 105.5
10.4
the sample.
Repeatability (%)
Reproducibility (YO) 10.2 The method was validated by analysing blank
samples of bovine urine spiked with clenbuterol at
 LC/LC/TSP-MS/MS ANALYSIS OF b,-AGONISTS IN URINE 423

levels of 0.22 and 5.3 ng ml-*, on two different days. tification limit of 0.1 ng ml-' (S/N = 10) can be easily
The results, given in Table 2, show that both the repea- achieved.
tability and reproducibility were satisfactory. The availability of a bovine urine sample in which a
Figure 3 shows the chromatograms obtained for the concentration of clenbuterol residue of 1.28 ng ml- '
LC/LC/TSP-MS/MS analysis of a blank bovine urine had been previously measured by GC/MS4 gave the
sample (A), and for the same urine sample spiked at a opportunity to investigate the performance of the
level of 0.22 ng ml- of clenbuterol (B). As can be seen, method on a real sample. In order to obtain informa-
the developed procedure is highly selective, and a quan- tion about both repeatability and reproducibility, the

El04
p.000

4001
zoo

A A
m/z 277+203
5:38
296114
100 -
3 588,lO1 - Et05
2.961

50-

mh 283-203
I " . - . -

5:40
3165'17
...__
301 1,613
100-

i
3.165

50-

=+05
mh 283+203

100-
4.758

50-

mh 277-203
C
5:40
281544
3 2 1 9,10 4
100- -6t05
2.845

50-

mh 283-203
, . . . ' I . . . I . . . . , ...,.... . .,.... , , . . .

Figure 3. Ion chromatograms obtained for the LC/LC/TSP-MS/MS analysis of clenbuterol in bovine urine samples. (A) Blank urine
sample; (B) blank urine sample spiked with clenbuterol at a level of 0.22 ng ml-' ; (C) bovine urine sample containing a clenbuterol residue
of 1.31 ng ml-'. The MS/MS transitions monitored for each analyte are indicated. Peaks: 1 -clenbuterol; 2 =internal standard.
 424 E. A. HOGENDOORN ET AL.

sample was analysed at four different days. On each day

Table 3. Results of Clenbuterol determination in a bovine
the LC/LC/TSP-MS/MS system was initialized by pre-
urine sample which had been previously found to
contain a residue of 1.28 ng ml-' of clenbuterol paring a new mobile phase and checking the LC/LC
using a GC/MS method (see text for explanation) and TSP-MS/MS conditions as described above. O n
two days some optimization by FIA was necessary, but
No. of Mean concentration RSD could be accomplished in less than 1 h. Information
Day experiments ( n s m l ') 1%)
from this validation procedure is given in Table 3,
1 1.30 5.4 which shows good agreement with results obtained by
2 1.27 16.0 GC/MS and acceptable repeatability and repro-
3 1.35 9.5
4
ducibility. The ion chromatograms obtained for the LC/
1.29 5.8
LC/TSP-MS/MS analysis of this sample are displayed
Overall concentration (ng mi-') 1.31 in Fig. 3(C).
Repeatability (YO) 9.9
Reproducibility (%) 8.4
SRM for salbutamol

Because of the presence of two further hydroxy groups

in the molecule, the polarity of salbutamol is higher
Table 4. Recoveries of salbutamol from bovine than that of clenbuterol. Conseauentlv. a reduction in
urine
the percentage of organic modiger in ;he mobile phase
Spiked level No. of Mean recovery RSD is necessary. In our previous 15% methanol
(neiml) experiments (%.) ("/) was used to provide some separation between the
0.55 3 89 12.7 analyte and terbutaline. As we intended to luse terbuta-
10.8 3 101 1 .o line as an internal standard for salbutamol, and because
Overall recovery (%) 95.2 the two analytes were easily separated by MS/MS
Repeatability (%) 9.0 owing to their different molecular masses, a mobile
phase containing 18% methanol was chosen (see
Experimental) to allow co-elution of the two analytes
(the capacity factors were 3.1 and 3.0 for salbutamol
and terbutaline, respectively). This has important impli-
Table 5. Determination of unconjugated salbu- cations for the column switching procedure, allowing
tamol in bovine urine samples which minimization of the transfer time of the analytes from
had been previously found to be posi- C-1 to C-2. In the case of clenbuterol, this aspect was
tive for fi,-agonists by an enzyme-
linked immunosorbent assay (see text
not considered as a deuterated internal standard was
for explanation) available.
The same strategy as developed for clenbuterol was
Sample Concentration Mean c~ncenlration ASD applied to LC/LC/TSP-MS/MS sample processing :
code (ng mi-') (ng ml-') ("1
analysis of urine samples with interpolated analysis of
A-5 56.5 56.7 7.7 standards rendering a total time of analysis
52.5 (sample + standard) of less than 15 min. As can be
61.2 expected when using a mobile phase with a low eluotro-
A-10 7.9 8.4
9.0
pic strength, a high UV background signal at 280 nm
A-8 30.9 - - was observed on C-1 (range 0.2-1.0 AUFS; data not
shown) during analysis of urine samples. However,

1 1
2:36
955623
6341,501
9 .5 5 6

2:oo 2:20 2:40 3:OO mh 240+166

3:20

Figure 4. Ion chromatograms obtained for the LC/LC/TSP-MS/MS analysis of salbutamol in a bovine urine sample containing a residue Of
7.9 ng ml-' of unconjugated salbutamol. The MS/MS transitions monitored for each analyte are indicated. Peaks: 1 -salbutamol:
2 = internal standard.
 LC/LC/TSP-MS/MS ANALYSIS OF B2-AGONISTS IN URINE
neither interferences in analyte detection nor deleterious
chromatographic effects were observed during the pro-
cessing of about 40 samples per day.
The method was validated by the analysis of blank
bovine urine spiked with salbutamol at two different
levels, 0.55 and 10.8 ng rn1-I. The results, given in Table
4, show acceptable recovery and repeatability at both
levels. The limit of quantification (S/N = 10) was esti-
mated to be 0.1 ng ml-'.
The LC/LC/TSP-MS/MS method was also used for
the analysis of three bovine urine samples which were
found to be positive for b,-agonists by enzyme-linked
immunosorbent assay.' The presence of salbutamol was
confirmed by GC/MS. The results of this experiment,
given in Table 5, show that the current methodology is
able to determine quantitatively unconjugated salbuta-
mol in bovine urine samples. Figure 4 shows the ion
chromatograms obtained for the LC/LC/TSP-MS/MS
analysis of a bovine urine sample (code A-10; see Table
5) in which a concentration of 7.9 ng ml-' of uncon-
jugated salbutamol was measured. The potential of the
described instrumental configuration in the direct
analysis of salbutamol conjugates was not investigated.
CONCLUSIONS
Our results show that the LC/LC/TSP-MS/MS
hyphenated system is a very efficient approach for the
rapid, selective and sensitive analysis of p,-agonists in
REFERENCES
1 . J. P. Hanrahan, Befa-Agonisrs and Their €ffecfs on Animal
Growfh and Carcass Qualify. Elsevier Applied Science,
Barking (1987).
2. J. F. Martinez-Navarro, Lancet 336,1311 (1991).
3. C. Puke, D. Lamaison, G. Keck, C. Bostvironnols, J. Nicolas
and J. Descotes, Vet. Hum. Toxicol. 33,480(1 988).
4. A. Polettini. A. Groppi. M. C. Ricossa and M. Montagna, Biol.
Mass Spectrom. 22.457 (1993).
5. Muscling in on clenbuterol (editorial), Lancet 340, 403
(1992).
6. Infernal Report of Acfivifies. Games of the XXV Olympiad of
Barcelona '92. Report IMIM/FT/DOPING/JJ00/1/1.2, lnstitut
Municipal d'lnvestigacio Medica. Barcelona (1992).
7. I. Yamamoto and K. Iwala, J. lmrnunoassay 3,155 (1982).
8. M . Paleologo Oriundi, G. Giacomini, F. Ballaben, F. Berti, F.
Benedetti, R . Bagnati and E. Bastiani, Food Agric. lmmunol. 4,
73 (1992).
9. M . J. Sauer, R . J . H. Pickett and A. L. MacKenzie, Anal. Chim.
Acta 275, 195 (1 993).
10. C. T. Elliott, J. D. G. McEvoy, W. J. McCaughey, D.H. Shortt
and S. R. H. Crooks,Analyst 118,447(1993).
1 1 . J. G. Leferink, I. Wagernaker-Engels, R. A. A. Maes, H.
Lamont, R . Pauwels and M. van der Straeten, J . Chromatogr.
143,299 (1977).
12. R . A. Clare, 0 . S. Davies and T. A. Baillie. Biomed. Environ.
Mass Spectrom. 6,31 (1979).
13. H. J. Forster. K. L. Rominger, E. Ecker, H. Peil and A. Wittrock,
Biomed. Environ. Mass Spectrom. 17.417 (1988).
14. J. Schmid and A. Bucheler, Biorned. Environ. Mass Spectrom.
17,415(1988).
15. J. Girault, P. Gobin and J. B. Fourtillan, Biomed. Environ.
Mass Spectrom. 19,80 (1990).
16. J. Girault and J. B. Fourtillan, J . Chromafogr. 518,41 (1990).
17. A. Polettini. M. C. Ricossa, A. Groppi and M . Montagna, J.
Chromatogr. 564,529 (1 991 ).
425
samples of bovine urine. The two SRMs developed for
clenbuterol and salbutamol demonstrate that the
approach can be easily adapted to different p,-agonists
and other analytes amenable to TSP-MS/MS. In addi-
tion, the high sample throughput yields a favourable
cost-benefit ratio for the expensive equipment used.
Owing to the high selectivity of the detection mode,
the ability of LC/LC to yield an efficient clean-up is
hardly utilized. In the configuration used here, the main
features of LC/LC are to provide (after a simple on-line
clean-up step) (i) a very stable flow during the intro-
duction of the analyte into the TSP-MS/MS system and
(ii) the ability to perform large-volume sample injection
and efficient chromatographic elution of polar analytes.
All these features are favourable to sensitivity.
Useful applications may be expected in the on-line
analysis of serum samples by applying an internal
surface reversed-phase (ISRP) column as the first
column in the proposed methodology. This interesting
possibility and the potential of LC/LC/TSP-MS/MS in
multi-residue analysis of b,-agonists will be a subject
for future research.
Acknowledgement
This work was carried out with financial support from the European
Commission (DG XII. Science and Technology, Division XII-H-l), in
the framework of the Human Capital and Mobility Programme, as
part of the network project Hyphenation Analytical Chemistry for
Environmental and Public Health Research in the European Union
(Grant No. ERBCHRXCT930274).
18. M. C. Dumasia and E. Houghton, J . Chromafogr. 564, 503
(1991).
19. L. Leyssens, C Driessen, A. Jacobs, J. Czech and J. Raus, J.
Chromatogr. 564,51 5 (1991 ).
20. L. A. van Ginkel, R . W. Stephany and H. J. van Rossum, J.
Assoc. Off.Anal. Chem. 75,554 (1992).
21. W. J . Blanchflower, S. A. Hewitt, A. Cannavan, C. T. Elliott
and D. G. Kennedy, Biol. Mass Specfrom. 22.326 (1993).
22. T. Visser, M. J. Vredenbregt. A. P. J. M. de Jong, L. A. van
Ginkel, H. J. van Rossum and R . W. Stephany, Anal. Chim.
Acfa 275,205 (1993).
23. M.-P. Montrade, B. Le Bizec, F. Monteau, B. Siliart and F.
Andre, Anal. Chim.Acta 275,253 (1993).
24. J. A. van Rhijn. W. A. Traag and H. H. Heskamp, J . Chromafo-
gr. 619,243 (1993).
25. L. A. van Ginke1.J. Chromatogr. 564,363 (1991).
26. A. Polettini, M. Montagna, J. Segura and X de la Torre, J .
Mass Specfrom. 31,47 (1995).
27. B. Diquet. L. Doare and P. Simon, J . Chromafogr. 336, 415
(1 984).
28. C. Eddins, J. Harnann and K. Johnson, J. Chromafogr.Sci. 23,
308 (1985).
29. H. H. D. Meyer, L. Rinke and I. Diirsch. J . Chrometogr. 564,
551 (1991).
30. H. Hooijerink. R . Schilt, W. Haasnoot and D. Courtheyn, J.
Pharm. Biomed. Anal. 9,485 (1991).
31. D. Courtheyn, C. Desaever and R . Verhe, J. Chromafogr. 564.
537 (1991).
32. W. J. Blanchflower and D. G. Kennedy, Biomed. Environ.
Mass Specfrom. 18,935 (1989).
33. L. Debrauwer and G. Bories. Anal. Chim. Acre 275, 331
(1 993).
34. J. A. van Rhijn, M. O'Keeffe. H. H. Heskamp and S.Collins, J.
Chromafogr.A 712,67(1995).
35. D. R. Doerge, S. Bajic, L. R. Blamkenship, S . W. Preece and
426 E. A. HOGENDOORN ET AL.
M .I. Churchwell, J. Mass Specfrom. 30, 91 1 (1995).
36. P. Arpino, Mass Spectrom. Rev. 9, 631 (1 990).
37. P. Arpino, Mass Spectrom. Rev. 11, 3 (1992).
38. E. R . Verhey, H. J. E. M. Reeuwijk, G. F. LaVos, W. M. A.
Niessen, U. R . Tjaden and J. van der Greef, Biomed. Fnviron.
Mass Specfrom. 16, 393 (1988).
39. A. Walhagen, L.-E. Edholm, E. E. M . Heeremans, R . A. M. van
der Hoeven, W. M. A. Niessen, U. R. Tjaden and J. van der
Greef. J . Chromatogr. 474, 257 (1 989).
40. E. A. Hogendoorn, U. A. Th. Brinkman and P. van Zoonen, J.
Chromatogr. 644, 307 (1 993).
41. E. A. Hogendoorn, R. Hoogerbrugge, P. van Zoonen, C. E.
Goewie and P. J. Schoenmakers, J . Chromafogr. 552, 113
(1 991 ).
42. E. A. Hogendoorn, R. Hoogerbrugge, E. Dijkman, S. M. Gort,
U. A. Th. Brinkman and P. van Zoonen, J. Chromatogr. Sci.
31,433 (1993).
43. S.M. Gort, E. A. Hogendoorn, E. Dijkman, P. van Zoonen and
R. Hoogerbrugge, Chromarographia 42, 17 (1996).
44. E. A. Hogendoorn and P. van Zoonen, J. Chromatogr. A 703,
149 (1995).
45. A. Polettini, M . Montagna, E. A. Hogendoorn, E. Dijkman, P.
van Zoonen and L. van Ginkel, J . Chromatogr. A 695, 19
(1995).
46. A. Polettini, in Doping e Scienze Medico-Forensi, Proceedings
of r'Vl Convegno Nazionale Gruppo Tossicologi Forensi
S.1.M.LA.". Pavia, September 1993, edited by M. Montagna,
0. Crippa and P. Maternini, p. 226. EDIMES, Pavia (1993).
47. W. Genuit and H. van Binsbergen, J . Chromatogr. 474, 145
(1989).