Journal of Pharmaceutical and Biomedical Analysis 75 (2013) 1– 6
Contents lists available at SciVerse ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis
jou rn al h om epage: www.elsev ier .com/ locate / jpba
ssay at low ppm level of dimethyl sulfate in starting materials for API synthesissing derivatization in ionic liquid media and LC–MS
elu Grinberga, Florin Albub, Keith Fandricka, Elena Iorgulescuc, Andrei Medvedovicib,c,∗
Chemical Development Department, Boehringer Ingelheim Pharmaceuticals Inc., 900 Ridgebury Rd., Ridgefield, CT 06877-0368, USABioanalytical Laboratory, SC Labormed Pharma SA, #44B Th. Pallady Blvd., Bucharest 032266, RomaniaUniversity of Bucharest, Faculty of Chemistry, Department of Analytical Chemistry, #90 Panduri Av., Bucharest 050663, Romania
r t i c l e i n f o
rticle history:eceived 12 October 2012eceived in revised form 7 November 2012ccepted 9 November 2012vailable online 19 November 2012
eywords:imethyl sulfateerivatization with dibenzazepine oryridinePLC/(+)ESI-MS/MS
a b s t r a c t
Dimethyl sulfate (DMS) is frequently used in pharmaceutical manufacturing processes as an alkylatingagent. Trace levels of DMS in drug substances should be carefully monitored since the compound canbecome an impurity which is genotoxic in nature. Derivatization of DMS with dibenzazepine leads to for-mation of the N-methyl derivative, which can be retained on a reversed phase column and subsequentlyseparated from other potential impurities. Such derivatization occurs relatively slowly. However, it canbe substantially speed up if ionic liquids are used as reaction media. In this paper we report the use of 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide (IL1) and 1-butyl-4-methylpyridiniumtetrafluoroborate (IL2) as reaction media for the derivatization of DMS with dibenzazepine. It was deter-mined that the stoichiometry between the substrate and DMS may be 1:1 or 2:1, in relation with thenature of the reaction media. An (+)ESI-MS/MS approach was used for quantitation of the derivatized
ILIC/(+)ESI-MS (SIM)onic liquid as derivatization media
product. Alternatively, DMS derivatization may be carried out with pyridine in acetonitrile (ACN). The N-methylpyridinium derivative was separated by hydrophilic interaction liquid chromatography (HILIC)and detected through (+)ESI-MS (in the SIM mode). In both cases, a limit of quantitation (LOQ) of0.05 �g/ml DMS was achievable, with a linearity range up to 10 �g/ml. Both analytical alternatives wereapplied to assay DMS in 4-(2-methoxyethyl)phenol, which is used as a starting material in the synthesisof metoprolol.
. Introduction
Dimethyl sulfate (DMS) is commonly used as an alkylatinggent in organic synthesis. In vivo and in vitro tests carried outn DMS have demonstrated the carcinogenic potential of the com-ound [1,2]. The issue of genotoxic impurities has rapidly acquired
ncreased attention from the pharmaceutical industry [3,4] and theorresponding regulatory bodies [5,6]. Quantitation of trace levelsf DMS in complex matrices is challenging due to their reactiv-ty and increased polar characteristics. Experimental tools used forontrol and analysis of alkyl esters of alkyl and aryl sulfonic acidsn active pharmaceutical ingredients have been recently reviewed7]. Earlier approaches dealt with gas chromatographic (GC) analy-
is by direct sample injection [8,9] with a FID or MS detection. Theajor drawback of such an analytical solution relates to the fre-
uent need for cleaning the inlet port to avoid formation of ghost
∗ Corresponding author at: University of Bucharest, Faculty of Chemistry, Depart-ent of Analytical Chemistry, #90 Panduri Av., Bucharest 050663, Romania.
peaks produced by thermal degradation of the matrix deposited onthe internal surface of the liner. Isolation of DMS from the matrixthrough different techniques prior to injection in the GC can beconsidered as a practical solution to the above mentioned prob-lem. Liquid–liquid extraction of DMS with methyl t-butyl ether hasbeen used for aqueous soluble active pharmaceutical ingredient(API) intermediates with satisfactory results [10].
Another important approach for the determination of DMSis through derivatization. For example, the reaction of DMS inaqueous media with sodium thiosulfate leads to formation ofmethylthiocyanate, which can be directly analyzed by head-spaceGC–MS [11]. Formation of the methylisothiocyanate byproduct wasobserved only in minor amounts. Pentafluorothiophenol was alsoused to act as a methylation substrate for DMS, allowing analysisof the derivative by head-space GC–MS [12]. 2-Mercaptopyridinewas successfully used for derivatization of DMS. The fluorescenceproduct was determined by RPLC and detected by fluorescence[13]. Alternatively, trialkylamines (more precisely triethylamine,
in the case of DMS) were used as derivatization reagents for alkyl-ating compounds [14]. A quantitative derivatization was obtainedafter heating at 50–60 ◦C for 1 h. The resulting quaternary ammo-nium ion was subsequently separated from the reaction mixture
y means of a hydrophilic interaction mechanism in liquid chro-atography (HILIC). The permanent charge of the target compoundakes it readily detectable through (+)ESI/MS detection carried out
n the selective ion monitoring (SIM) mode.HILIC separation mechanism has recently emerged as a potent
ool for separation of polar compounds [15–18]. Although HILICs considered a promising separation technique for pharmaceut-cals, the reversed phase separation mechanism still remains arst choice in most quality control laboratories. This was theajor reason for considering an alternative reagent for derivati-
ation of DMS, by making the reaction product separable throughPLC with increased retention. We focused on apolar compoundsontaining a single secondary amino moiety acting as a methyla-ion substrate for DMS. Dibenzazepine was consequently used aserivatization agent. Kinetic studies revealed that a highly ionicnvironment such as ionic liquids [19] along with temperaturencreased the speed of the derivatization process. Pyridine has alsoeen considered as a reagent in order to illustrate the formation ofuaternary ammonium derivatization compounds. The resulting N-ethyl derivative was separated from the reaction matrix in HILICode. The two alternatives were validated and used to assay DMS
n 4-(2-methoxyethyl)phenol, a starting material used in the syn-hesis of metoprolol tartrate (a selective �1 receptor blocker usedn the treatment of several diseases of the cardiovascular system,specially hypertension).
. Experimental
.1. Reagents
Acetonitrile (ACN) gradient grade for HPLC and formiccid extra pure grade were obtained from Merck (Darmstadt,ermany). Water for chromatography (minimum resistivity of8.2 M� and maximum TOC content of 30 ng/ml) was producedsing a TKA Lab HP 6UV/UF instrument (TKA Water Purifi-ation Systems GmbH, Niederelbert, Germany). DMS (cat. no.186309, purity ≥ 99.8%) was purchased from Sigma–Aldrich. 4-(2-ethoxyethyl)phenol (cat. no. B23548, purity 98%) was obtained
rom Alfa Aesar. Ionic liquids, 1-butyl-1-methylpyrrolidiniumis(trifluoromethylsulfonyl)imide (cat. no. 38894, purity ≥ 98%)urther noted (IL1), and 1-butyl-4-methylpyridinium tetrafluo-oborate (cat. no. 73261, purity ≥ 97%), further noted (IL2) wereurchased from Fluka Analytical. Derivatization reagents, namelyH-dibenz[b,f]azepine (cat. no. 143650, batch 1452630, purity9.8%) and pyridine (cat. no. 270407, HPLC grade, purity ≥ 99.9%)ere purchased from Sigma–Aldrich.
.2. Instrumentation
Experiments were performed with an Agilent 1200 SL seriesC/MSD (Agilent Technologies) system consisting of the followingodules: degasser (G1379B), binary pump (G1312B), automated
njector (G1367C) and thermostat (G1330B), column thermostatG1316B), ESI standard source (G1948B) and triple quadrupole
ass spectrometric detector (G2571A). System control, data acqui-ition and interpretation were made with the Agilent Mass Hunteroftware version B 04.01 incorporating both qualitative and quan-itative packages. A thermoreactor Spectroquant®, model TR420rom Merck was used as a heating device for controlling derivati-ation reaction temperature.
.3. Derivatization procedure and sample preparation
The derivatization reaction is based on the methylation actionf DMS at the secondary or tertiary amino moiety of the reagentsee Fig. 1). Basically, a volume of 0.25 ml from a solution in ACN or
Fig. 1. Methylation pathways of different substrates with DMS.
ionic liquids containing DMS was mixed with 25 �l of a solution inthe same solvent containing 10 mg/ml of the derivatization reagentin a 1.1 ml extreme recovery HPLC glass vial (conical shaped bot-tom). After tightly capping the vial, the reaction mixture was heatedat a given temperature for a precise time period. Optimizationof the preparation procedure was made by studying the kineticsof the derivatization reactions in different solvents and at differ-ent temperatures. Upon completion of the reaction, the vials weretransferred to the thermostated autosampler at 25 ◦C and kept forcooling until injection was made.
To study the kinetics of the reaction, 5 �g/ml DMS solutionsin the appropriate solvent were subjected to derivatization. Forlinearity studies, the following stock solutions were made in theappropriate solvent: 10 �g/ml and 100 �g/ml DMS; 200 mg/ml4-(2-methoxyethyl)phenol containing no residual DMS. Standardsolutions of 0.05, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, and 10 �g/mlDMS were obtained by serial dilution of the 10 �g/ml DMS solu-tion. Equivalent amounts of 4-(2-methoxyethyl)phenol were addedto the previous solutions. These standard solutions correspond to0.5, 1, 2.5, 5, 7.5, 10, 25, 50, 75 and 100 �g/ml of DMS in 4-(2-methoxyethyl)phenol. The tested samples were typically preparedat 100 mg/ml of 4-(2-methoxyethyl)phenol in the appropriate sol-vent.
2.4. Chromatographic methods
The RPLC separations were carried out on a ZorbaxEclipse XDB-C18 Narrow Bore Rapid Resolution column,100 mm × 2.1 mm × 3.5 �m d.p. (Agilent Technologies, cat. no.961753-902) fitted with a Phenomenex Guard Cartridge C18,2 mm × 4 mm (prod. No. AJO-4286, Phenomenex Inc.). The columnwas thermostated at 25 ◦C. The elution was isocratic, and themobile phase consisted of a mixture of aqueous 0.1% formicacid solution and ACN at a ratio of 4/6 (v/v), with a flow rate of0.4 ml/min; 1 �l was injected into the chromatographic column.
The HILIC separations were carried out on a Luna HILIC column,150 mm × 2 mm × 3 �m d.p. (cat. no. 00F-4449-B0, PhenomenexInc.), thermostated at 25 ◦C. The column was operated in the iso-cratic mode, using a mobile phase consisting of an aqueous 50 mMammonium formate buffer (pH 3.2) and ACN 3/97 (v/v) at a flow rateof 0.8 ml/min; 1 �l was injected into the chromatographic column.
When the solvent being used was ACN, the drawing/dispensingspeed of the metering pump of the autosampler was set to500 �l/min. When ionic liquids were used as solvating media, thisparameter was reduced at 100 �l/min.
2.5. MS detection parameters
The operational parameters controlling the ESI source were asfollows: drying gas (N2) temperature: 350 ◦C; drying gas flow:13 l/min; pressure of the nebulizing gas: 60 psi; capillary voltage:
ceutical and Biomedical Analysis 75 (2013) 1– 6 3
4cai2t
3
3
lstMbhift
Eeqasait
aorHptatn
ealfmrisssM
aohaposrt
trp
N. Grinberg et al. / Journal of Pharma
000 V. The fragmentor potential was set at 135 V. For MS/MS appli-ations, the potential of the collision cell was set to 15 V (N2 islso used as collision gas). The mass transition used for monitor-ng the N-methyl dibenzazepine derivatization product was m/z08.3–193.3. When single MS detection in the SIM mode was used,he m/z of N-methylpyridinium molecular ion is 94.2.
. Results and discussion
.1. Derivatization
The classical approach to the industrial synthesis of metopro-ol uses 4-(2-methoxyethyl)phenol as a starting material, whichuccessively reacts with epichlorohydrin and isopropyl amineo yield the active ingredient in its free base form [20]. 4-(2-
ethoxyethyl)phenol is obtained from p-(�-hydroxyethyl)phenyl-enzyl ether through methylation with DMS, followed by catalyticydrogenation. Since DMS is a genotoxic compound it is very
mportant to determine the residual amount of DMS remainingrom the starting material in the final active ingredient, andhroughout the synthetic pathway of the process.
Both DMS and 4-(2-methoxyethyl)phenol are not ionized in theSI sources. Consequently, a direct injection alternative, conductedither in the RPLC or HILIC separation modes, fails to produce ade-uate results. On the other hand, DMS exhibits increased reactivity,nd is readily decomposed during the different stages of the synthe-is. Development of a selective derivatization procedure targeting
simultaneous chemical stabilization and enhanced electrosprayonization efficiency may be consequently considered as an attrac-ive experimental tool.
The nucleophilic substrate represented by a tertiary aliphaticmine has been successfully used for reacting DMS. The ionic naturef the resulting quaternary ammonium derivative in solution rep-esents a solid basis for increased sensitivity through MS detection.ydrophilicity mediated interaction between the derivatizationroduct and the stationary phase is thus required to produce reten-ion within the chromatographic column. Although pharmaceuticalpplications based on HILIC gained popularity over the last decade,he RP mode still represents the most popular separation mecha-ism.
Consequently, our attention focused on molecular structuresxhibiting increased hydrophobic characteristics and a secondarymmonium moiety acting as a potential substrate for the methy-ating action of DMS. The hydrophobic character was targeted toacilitate increased chromatographic retention under RP separation
echanism conditions. The resulting tertiary amino derivative iseadily protonated in acidic solutions and produces enhanced ion-zation within the ESI source, with a direct positive impact on theensitivity of the MS detection. As the molecular structure of theubstrate is suitable for fragmentation in a collision cell, additionalelectivity may be added in the detection stage by using a tandemS approach, by monitoring a specific single mass transition.Dibenzazepine is an attractive candidate for this approach,
s the structure contains the required functional group (a sec-ndary amino moiety), which can give the compound an increasedydrophobic character (log Pcalc. = 4.1); it is also commercially avail-ble as a highly pure compound. Additionally, we consideredyridine as an alternative substrate for the electrophilic attackf DMS. Obviously, the chromatographic separation mechanismhould be again switched to HILIC, in order to produce increasedetention of the resulting product against the excess of the deriva-ization agent and 4-(2-methoxyethyl)phenol.
The possibility of using the above mentioned substrates forhe assay of DMS as a residual genotoxic impurity in raw mate-ials mainly depends upon the fast kinetics of the derivatizationrocess, the chemical stability of the resulting derivative, and its
Fig. 2. Pseudo-first-order kinetic plots obtained for DMS derivatization with diben-zazepine in ACN and IL2 at 80 ◦C, and with pyridine in ACN at 60 ◦C.
MS detectability (in terms of ionization yields and indifference toeventual matrix effects).
3.2. Kinetics of the derivatization reaction
Ionic liquids exhibit significant electrorestrictive behavior,showing a radial and angular distribution of the solvent ions aroundthe solute which become more structured and permit the chargeseparation to begin to occur [21]. These features affect the possibil-ity of ionic liquids to solvate long-lived dissolved species as well astransient species [22]. As such the rate constant of a certain reactioncan increase in an ionic liquid compared to a molecular solvent.
If we consider that c0 is the initial concentration of DMS and c isits concentration at time t, than a plot of ln[c]/[c0] versus t shouldbe linear, and thus consistent with pseudo-first-order kinetics [23](k is the reaction rate constant).
ln(
[c][c0]
)= −k × t (1)
Indeed, plots performing ln[c/c0] versus time were linear andcharacterized by correlation coefficients higher than 0.9. The low-est correlation coefficient (0.9644) was obtained in IL2 at 120 ◦C.This may be explained by the increased speed of the derivatiza-tion reaction, limiting the number of determinations and addinguncertainty to the experimental results. A few examples of thekinetic plots of different derivatization substrates in different sol-vents are presented in Fig. 2. The results for the kinetic studies alongwith the half-life of the pseudo first order reaction (t1/2 = ln 2/k) arepresented in Table 1.
From the data of Table 1 it is possible to ascertain that the rateconstant in both ionic liquids is several orders of magnitude higherthan in ACN. An increase of temperature naturally increases therate constant. The rate constant in IL1 at 80 ◦C is higher that in IL2,probably due to a difference in solvation and structure of the IL1relative to IL2. At the same time t1/2 decreases significantly in thetwo ionic liquids relative to ACN.
Thus, the use of ionic liquids IL1 and IL2 as reaction mediaappeared as an alternative to ACN, as they reduce the reaction time.It should be noted that no derivatization arises if dimethylsulfox-
ide (DMSO) is used as reaction media. As illustrated in Fig. 3, theuse of IL2 as a reaction solvent at 80 ◦C instead of ACN leads toa reduction of the derivatization process by a factor of 19. Whencomparing IL1 to ACN under the same temperature conditions, a
4 N. Grinberg et al. / Journal of Pharmaceutical and Biomedical Analysis 75 (2013) 1– 6
Table 1Kinetic parameters of the derivatization reactions (DMS derivatized with dibenzazepine and pyridine, respectively) carried out in different solvents.
Derivatization agent Solvent Temperature (◦C) k (h−1) t1/2 (h−1) rxy Sampled timeinterval (h)
long as the effluent is directed toward the ESI source by means of a
Pyridine Acetonitrile60 2.5365
80 7.3651
eduction by a factor of 34 is obtained. Comparison between dura-ions of a complete derivatization in IL1 compared to IL2 leads to aeduction by a factor of 1.8 at 80 ◦C and by 2.5 at 120 ◦C. However,f comparing the absolute peak areas corresponding to N-methylibenzazepine produced in IL1 and IL2 after complete derivatiza-ion at 80 ◦C and 120 ◦C, one can observe that values are doubledhen using IL2. The observed behavior should not be explained
hrough matrix effects (signal enhancement) arising in the source,otentially induced by the presence of ionic liquids, as long as, dur-
ng their elution from the chromatographic column, the effluentow is diverted from entering the ESI source (by means of a sixort rotative valve). The only pertinent explanation for this exper-
mental observation is that, in IL2, methyl hydrogensulfate (MHS)roduced through methylation of dibenzazepine by DMS is furtherethylating the substrate to yield sulfuric acid (see process II in
ig. 1). It appears that in IL2 the stoichiometry of the interactionubstrate: DMS is 2:1 instead of 1:1. In IL1, only the first stage ofhe reaction is achieved, as is the case for ACN at 60 ◦C and 80 ◦C.n IL2, even at 60 ◦C, both methylation stages are completed. Thetudy of samples corresponding to derivatization periods up to 60 hndicates that the N-methyl substituted dibenzazepine is stable inhe reaction media and the second methylation stage did not arisen IL1, even at increased temperatures.
A temperature of 120 ◦C was applied when using IL1 and IL2s reaction media (but not in ACN, as it boils at 82 ◦C). Such aemperature still maintains DMS and 4-(2-methoxyethyl)phenoln the liquid state, as it is far below their corresponding boil-
ng points (188 ◦C with decomposition for DMS and 234 ◦C for-(2-methoxyethyl)phenol). A faster derivatization was achieved
ig. 3. Kinetics of the derivatization reaction of DMS with dibenzazepine at differentemperatures and reaction media.
0.27 0.9966 0–1.33 100.09 0.9854 0–0.5 6
at 120 ◦C compared to 80 ◦C (7.8 fold in IL1, 5.5 folds in IL2). Theseexperimental facts are illustrated by data contained in Fig. 3.
Derivatization with pyridine was made in ACN at 60 ◦C and 80 ◦C.The reactivity of the substrate is enhanced, quantitative trans-formations being obtained after 70 and 25 min, respectively, asillustrated in Fig. 4. No insights could be obtained about the deriva-tization reaction carried out in ionic liquids. The injection of theionic liquids as sample diluents in the chromatographic columnoperated under the HILIC mode seriously affects separation effi-ciency and peak shape (distorted peaks). Consequently, the assayof the derivatization product obtained in ionic liquids was not pos-sible.
3.3. Chromatographic separations
Derivatization of DMS with dibenzazepine was monitored bymeans of the isocratic RPLC method described in Section 2.Although the concentration of ACN in the mobile phase wasimportant (60%), the retention of the derivatization product wasconsistent. More precisely, assuming the injector’s valve switchingsignal as indicator of the column’s void time, the following reten-tion factors were calculated: the organic cations of IL1 and IL2 – 0;4-(2-methoxyethyl)phenol – 0.47; dibenzazepine – 3.67; impurityin the derivatization reagent exhibiting the same mass transition asthe derivatization product – 6.19; derivatization product – 7.36. As
rotative valve only 3 min after injection (the moment correspondsto a retention factor of 4.17), the reaction solvent, the raw material,and the excess of derivatization reagent are not reaching the ion
Fig. 4. Kinetics of the derivatization reaction of DMS with pyridine at differenttemperatures.
N. Grinberg et al. / Journal of Pharmaceutical and Biomedical Analysis 75 (2013) 1– 6 5
Fig. 5. Chromatograms resulting upon separation of the derivatization productsaoM(
srs
biTtmtT3rclmAaii
3
tzdf
te
sult
s
for
the
assa
y
of
DM
S
afte
r
der
ivat
izat
ion
wit
h
dib
enza
zep
ine
or
pyr
idin
e.
Der
ivat
izat
ion
med
ium
Tem
per
atu
re(◦
C)
Du
rati
on
of
the
der
ivat
izat
ion
pro
cess
(min
)
Test
ed
lin
ear
dom
ain
(�g/
ml)
Nu
mbe
r
ofco
nce
ntr
atio
nle
vels
Para
met
ers
of
the
regr
essi
on
mod
el
LOQ
((10
×
s A−
A)/
B)(�
g/m
l)
Prec
isio
n
inte
rval
b:
RSD
%(c
once
ntr
atio
nle
vel)
Acc
ura
cy
inte
rval
:re
cove
ry
(%)
(con
cen
trat
ion
leve
l)
Prec
isio
n(R
SD%
at0.
5
�g/
ml,
n
=
10)
Prec
isio
n(R
SD%
at5
�g/
ml,
n
=
10)
Slop
e
(B)
s BIn
terc
ept
(A)
s AC
orre
lati
onco
effi
cien
t(r
xy)
e
AC
N
60
1440
0.1–
10
6
27,0
31
1629
−235
3
190.
4
0.99
28
0.16
2.8
(L6)
–20.
0
(L1)
85.4
(L6)
–100
.9
(L1)
14.4
8.6
AC
N
80
1440
0.1–
10
6
47,7
67
2208
−358
3
258
0.99
58
0.13
3.3
(L6)
–13.
9
(L1)
99.1
(L1)
–114
.5
(L6)
12.3
8.4
IL1
120
30
0.1–
10
8
320,
920
12,2
75
1234
3153
0.99
56
0.09
0.9
(L5)
–8.8
(L1)
86.8
(L6)
–100
.3
(L4)
8.1
2.6
IL2
120
30
0.1–
10
8
639,
821
8850
41,8
24
2273
0.99
94
0.04
4.0
(L6)
–6.6
(L1)
100.
6
(L2)
–103
.8
(L6)
9.3
8.8
IL2a
120
30
0.1–
10
8
638,
394
18,3
52
140,
111
4717
0.99
75
0.05
5.2
(L5)
–9.3
(L2)
96.9
(L5)
–108
.4
(L2)
7.7
3.2
AC
Na
80
30
0.05
–10
10
2,21
2,97
5
65,7
52
−48,
552
9092
0.99
65
0.06
2.8
(L4)
–20.
0
(L10
)
101.
0
(L5)
–113
.8
(L3)
1.3
0.8
on
was
mad
e
on
sam
ple
s
spik
ed
wit
h
DM
S
free
4-(2
-met
hox
yeth
ylp
hen
ol)
at
100
mg/
ml l
evel
.en
den
t
rep
lica
tes
on
each
con
cen
trat
ion
leve
l (n
=
6).
gainst matrix components (API starting material, excess of derivatization reagent,ther impurities). (A) Separation of N-methyl dibenzazepine by RPLC with (+)ESI-S/MS detention; (B) separation of N-methyl pyridinium by HILIC with (+)ESI-MS
SIM) detection. Conditions are given in Section 2.
ource. Consequently, matrix effects on ionization are consistentlyeduced. Typical chromatograms illustrating the results of the RPeparation method are provided in Fig. 5A.
Derivatization of DMS with pyridine was monitored by a HILICased chromatographic separation mechanism, also carried out
n isocratic elution conditions (details are given in Section 2).he injection valve switching signal indicates a void time ofhe column of 0.52 min. For the experimental conditions, 4-(2-
ethoxyethyl)phenol and pyridine are eluted at or close to the voidime of the column (retention factors of 0.17 and 0.21, respectively).he N-methyl pyridinium derivative exhibits a retention factor of.38. Through ionization, an impurity in pyridine characterized by aetention factor of 1.21 produced a fragment at m/z = 94.2, which isonsequently detected on the same mass channel as the target ana-yte. However, by diverting the effluent out of the ion source until
inute 1.8, it is possible to avoid the detection of the compound.s expected for an ionic compound, N-methyl pyridinium elutess a tailing peak (see Fig. 5B). Although the tailing phenomenons not severe, some quantitation errors may appear if the manualntegration feature is used.
.4. Validation results
The linearity, limit of quantitation, precision and recovery ofhe methods based on the derivatization procedures with diben-azepine and pyridine are summarized in Table 2. In the case of theerivatization using dibenzazepine substrate, data are given for dif-
erent derivatization conditions (temperature and reaction media).
From data contained in Table 2, one can first observe thathe presence of the matrix (4-(2-methoxyethyl)phenol) atxtremely high concentration levels does not adversely affect the Ta
ble
2V
alid
atio
n
re
Der
ivat
izat
ion
reag
ent
Dib
enza
zep
in
Pyri
din
e
aC
alib
rati
bSi
x
ind
ep
6 ceutica
dezsdtoidt2ttcb
p(fdboo(trcwADttw(tfs
4
(azsprstmmf
A
eP
[
[
[
[
[
[
[
[
[
[
[
[
4755–4762.[22] C. Chappe, M. Malvaldi, C.S. Pomelli, Ionic liquids: solvation ability and polarity,
Pure Appl. Chem. 81 (2009) 767–776.
N. Grinberg et al. / Journal of Pharma
erivatization process and the produced quantitative results. Asxpected, less sensitivity was obtained when making derivati-ation in ACN for 24 h, although precision and accuracy are notubstantially affected (the reaction rate is slow; consequently,ifferences between periods used for completion of the methyla-ion process are not inducing major experimental errors). One canbserve that different substrate: DMS stoichiometry is obtainedn IL1 and IL2. As the slope of the linear regression doubles whenerivatization is produced in IL2 compared to IL1, it seems obvioushat in IL2 the stoichiometry between the substrate and DMS is:1 instead of 1:1. The high sensitivity obtained when assayinghe N-methyl pyridinium derivative is due to the existence ofhe target compound being in an ionic state in solution. In suchonditions, the MS tandem detection may be avoided, and replacedy the single MS-SIM operating mode.
Three different production batches of 4-(2-methoxyethyl)henol were analyzed through the two alternative methodsderivatization with dibenzazepine in IL2 for 30 min at 120 ◦Collowed by the RPLC separation and MS/MS detection, anderivatization with pyridine in ACN for 30 min at 80 ◦C followedy the HILIC separation and MS-SIM detection). Both meth-ds found no detectable levels of DMS in one batch. For thether two batches, the RPLC method determined 0.52 ± 0.05 �g/mlmean ± standard deviation) and 1.2 ± 0.13 �g/ml DMS concentra-ion levels in 4-(2-methoxyethyl)phenol. The HILIC approach yieldsesults of 0.61 ± 0.05 �g/ml and 0.98 ± 0.12 �g/ml (number of repli-ate determinations was 5). Determined values fall reciprocallyithin the ±20% limits, which can be considered as acceptable.dditionally, the same methods were used to assay the residualMS in metoprolol tartrate resulting from the synthesis process of
he former 4-(2-methoxyethyl)phenol batches. It should be notedhat metoprolol tartrate is not soluble in IL2 and consequently DMSas extracted from the solid material in IL2 through sonication
15 min). In all cases DMS was not detectable. It clearly appearshat the intrinsic reactivity of the analyte does not leave it untrans-ormed upon the chemical and physical production stages of theynthesis process.
. Conclusions
Assay of DMS at low �g/ml levels in 4-(2-methoxyethyl)phenolstarting material for metoprolol synthesis) was possible throughpplying two analytical alternatives: (a) derivatization with diben-azepine in ionic liquids at 120 ◦C for 30 min, followed by a RPLCeparation and (+)ESI-MS/MS detection; (b) derivatization withyridine in ACN at 80 ◦C for 30 min, followed by a HILIC sepa-ation and (+)ESI/MS (SIM) detection. Both alternatives behaveimilarly. Depending on the nature of the ionic liquid used as a reac-ion medium, the stoichiometry between the substrate and DMS
ay be 1:1 or 2:1. Validation of the data obtained through thisethod recommends both analytical approaches as reliable tools
or assaying DMS in different chemical environments.
cknowledgements
Andrei Medvedovici, Florin Albu and Elena Iorgulescu acknowl-dge the financial support given by the Romanian ProjectNII ID PCE 2011 3 0152.
[
l and Biomedical Analysis 75 (2013) 1– 6
References
[1] G.R. Hoffmann, D.J. Crowley, P.J. Theophiles, Comparative potencies of induc-tion of point mutations and genetic duplications by the methylating agentsmethylazoxymethanol and dimethyl sulfate in bacteria, Mutagenesis 17 (2002)439–444.
[2] Z. Sobol, M.E. Engel, E. Rubitski, W.W. Ku, J. Aubrecht, R.H. Schiestl, Genotoxicityprofiles of common alkyl halides and esters with alkylating activity, Mutat. Res.633 (2007) 80–94.
[3] C.D. Humfrey, Recent developments in the risk assessment of potentially geno-toxic impurities in pharmaceutical drug substances, Toxicol. Sci. 100 (2007)24–28.
[4] D.A. Pierson, B.A. Olsen, D.K. Robbins, K.M. DeVries, D.L. Varie, Approachesto assessment, testing decisions, and analytical determination of geno-toxic impurities in drug substances, Org. Process Res. Dev. 13 (2009)285–291.
[5] Committee for Medicinal Products for Human Use (CHMP), European MedicinesAgency (EMEA), Guideline on the Limits of Genotoxic Impurities, London, UK,June 2006 (EMEA/CHMP/QWP/251344/2006).
[6] U.S. Department of Health and Human Services, Food and Drug Administra-tion, Center for Drug Evaluation and Research (CDER), Guidance for Industry– Genotoxic and Carcinogenic Impurities in Drug Substances and Products:Recommended Approaches, Silver Spring, MD, USA, December 2008.
[7] D.P. Elder, A. Teasdale, A.M. Lipczynski, Control and analysis of alkyl esters ofalkyl and aryl sulfonic acids in novel active pharmaceutical ingredients (APIs),J. Pharm. Biomed. Anal. 46 (2008) 1–8.
[8] M. Seymour, Determination of residual dimethyl sulfate in a lipophilic bulkdrug by wide bore capillary gas chromatography, J. Chromatogr. 463 (1989)216–221.
[9] N.V.V.S.S. Raman, K.R. Reddy, A.V.S.S. Prasad, K. Ramakrishna, Validated chro-matographic methods for the determination of process related toxic impuritiesin pantoprazole sodium, Chromatographia 68 (2008) 481–484.
10] J. Zheng, W.A. Pritts, S. Zhang, S. Wittenberger, Determination of low ppmlevels of dimethyl sulfate in an aqueous soluble API intermediate usingliquid–liquid extraction and GC–MS, J. Pharm. Biomed. Anal. 50 (2009)1054–1059.
11] C.R. Lee, F. Guivarch, C.N.V. Dau, D. Tessier, A.M. Krstulovic, Determination ofpolar alkylating agents as thiocyanate/isothiocyanate derivatives by reactionheadspace gas chromatography, Analyst 128 (2003) 857–863.
12] R. Alzaga, R.W. Ryan, K. Taylor-Worth, A.M. Lipczynski, R. Szucs, P. Sandra,A generic approach for the determination of residues of alkylating agentsin active pharmaceutical ingredients by in-situ derivatization–headspace–gaschromatography–mass spectrometry, J. Pharm. Biomed. Anal. 45 (2007)472–479.
13] J.G. Hoogerheide, R.A. Scott, Use of 2-mercaptopyridine for the determination ofalkylating agents in complex matrices: application to dimethyl sulfate, Talanta65 (2005) 453–460.
14] J. An, M. Sun, L. Bai, T. Chen, D.Q. Liu, A. Kord, A practical derivatization LC/MSapproach for determination of trace level alkyl sulphonates and dialkyl sulfatesgenotoxic impurities in drug substances, J. Pharm. Biomed. Anal. 48 (2008)1006–1010.
15] D.Q. Liu, M. Sun, A.S. Kord, Recent advances in trace analysis of pharmaceuticalgenotoxic impurities, J. Pharm. Biomed. Anal. 51 (2010) 999–1014.
16] Y. Guo, S. Gaiki, Retention and selectivity of stationary phases forhydrophilic interaction liquid chromatography, J. Chromatogr. A 1218 (2011)5920–5938.
17] T. Ikegami, K. Tomomatsu, H. Takubo, K. Horie, N. Tanaka, Separation efficien-cies in hydrophilic interaction liquid chromatography, J. Chromatogr. A 1184(2008) 474–503.
18] A.J. Alpert, HILIC at 21: reflections and perspective, J. Chromatogr. A 1218 (2011)5879–5880.
19] A. Berthod, M.J. Ruiz-Angel, S. Carda-Brosch, Ionic liquids in separation tech-niques, J. Chromatogr. A 1184 (2008) 6–18.
21] M.N. Kobrak, Electrostatic interactions of neutral dipolar solute with a fusedsalt: a new model for solvation in ionic liquids, J. Phys. Chem. B 111 (2007)
23] J.H. Espenson, Chemical Kinetics and Reaction Mechanism, McGraw-Hill BookCompany, New York, 1981, p. 13–17.
