J. Vet. Sci. (2002), 3(2), 103-108 J O U R N A L O F Veterinary Science

Simultaneous Determination of Various by Liquid Chromatography/Mass Spectrometry Youn-Hwan Hwang, Jong-Hwan Lim, Byung-Kwon Park and Hyo-In Yun*

Division of Veterinary Pharmacology and Toxicology, College of Veterinary Medicine, Chungnam National University Received J a n . 4, 2002 / Accept ed Ap r . 29, 2002

ABSTRACT7) poultry, those are 0.125 g/kg for and 0.1 g/kg for tylosin. In order to monitor residues, simple, Macrolides are frequently use d in veterinary confirmatory and simultaneous analytical methods are m edicine as therape utic and preve ntive agents for required. various diseases. It is difficult to determine m acrolides Microbiological assays were widely used for determination sim ultaneously w ith conventional m ethods due to of macrolide antibiotics [3, 4]. Unfortunately, these methods their sim ilar structures. A sim ultaneous analysis for could not be used for simultaneous analysis due to lacks of erythrom ycin, roxithrom ycin, tiam ulin and tylosin their specificities. Gas chromatography-mass spectrometry w ith LC/MS has bee n develope d. Se paration w as (GC-MS) supplies good sensitivity and selectivity [5], but pe rform ed on C18 reversed phase colum n. Mobile direct analysis for macrolides antibiotics is difficult because phase w as gradiently flow e d w ith 10 m M am m onium of their thermal labile property and low volatility. acetate and m ethanol. The m ass spectrome ter w as Liquid chromatographic methods have been reported for run in the positive m ode and sele ctive ion monitoring the determination of macrolide antibiotics: UV absorption m ode. The m olecular ions w ere [M+H]+ form at m /z [6-11], fluorimeteric [12-14], ch em ilu m in escen ce [15] a n d 837.5 for erythromycin, at m/z 859.5 for roxithrom ycin, elect r och em ica l det ection [16, 17] m et h ods h a ve been u sed at m /z 494.2 for tiam ulin and at m/z 916.7 for tylosin. for det erm in a tion , bu t t h ese m et h ods h a ve sh ow n h igh Lim its of detection w ere in the range from 0.001 to lim it s of d et ect ion . 0.01 ㎍/g low er than their MRLs. Recently, several simultaneous determination methods of macrolide antibiotics have been developed by mass Keyw ards : simultaneous determination; liquid chromato- spectrometry coupled with HPLC [18-20]. The determination graphy/mass spectrometry; macrolides antibiotics methods of macrolides by LC/MS have advantages such as high specificity and selectivity due to each molecular mass. 1. Introduction The aim of this study is to develop a more simple, rapid and effective method for the simultaneous determination of Macrolide antibiotics have 12-, 14-, 16- or 17-membered three macrolide antibiotics (erythromycin, macrocyclic lactone ring, which is bound to several amino and tylosin) and a pleuromultlin antibiotic (tiamulin) by and/or neutral sugars (fig.. 1). Because of their effective LC/MS with electrospray interface. Although tiamulin does antimicrobial activity against Gr am-posit ive bacteria, my- not belong to a group of macrolide antibiotics, we coplasma, chlamydia, they are frequently used in industrial determined this drug due to its similar structure to tylosin. animals to treat and prevent diseases or as growth promotants [1]. 2. Materials and methods Incorrect use of these antibiotics may leave residues in edible tissues causing toxic effects on consumers, e.g., 2-1 Che micals and reagents allergic reactions in hypersensitive individuals, or indirectly, Erythromycin, roxithromycin and tylosin were supplied problems through the induction of resistant strains of by S in il Ch em ica ls (Seou l, Kor ea ). Tia m u lin w a s su pplied by bacteria [2]. Therefore, the Sourth Korea has set maximum Da esu n g M icr obia ls (Seou l, K or ea ). H P LC gr a d e w a t er a n d residue limits (MRLs) for macrolide antibiotics in edible m et h a n ol w er e pu rch a sed fr om J .T. & Ba k er (N ew J er sey, tissues of food-producing animals. The MRLs of erythromycin U SA). Rea gen t gr a de a m m on iu m a ceta te w a s pu r ch a sed and tylosin are 0.1 g/kg in bovine and porcine. In case of fr om S igm a (Missou ri, U SA). Th e in dividu a l stock sta n d a r d solu t ion s w er e p repa r ed a s ＊Corresponding author: Hyo-In Yun 1 m g/㎖ in methanol and working standard solutions were Tel: +82-42-821-6759 Fax: +82-42-822-5780 prepared weekly by dilution of stock standard solutions with E-mail: [email protected] 104 Youn-Hwan Hwang, Jong-Hwan Lim, Byung-Kwon Park and Hyo-In Yun

Fig. 1. The structure of erythromycin (A), roxithromycin (B), tylosin (C) and tiamulin (D). m eth a n ol. All st a n da r d solu t ion s w er e st ored a t 4℃ and range of 0.001 ㎍/㎖ ~ 5 ㎍/㎖. were stable for at least 1 month under this condition. Limit of detection (LOD) and limit of quantitation (LOQ) Deionized or distilled water of 18.2 ㏁ cm-1 resistivity was were based on the signal-to-noise ratio based on their areas. used throughout the experiment. The signal-to-noise ratio of 3 was accepted for the LOD and that of 10 for the LOQ. 2-2 Instrumentation and chromatographic conditions Samples were analyzed by a Hewlett-Packard 1100 series 3. Results LC/MSD system. It consisted of a G1322A degasser, a G1312A binary pump, a G1315A photo-diode-array detector, 3-1. Chrom atographic separation a 59987A electrospray interface and a 5989B mass spectrometer. All drugs used for the experiment were separated under The separation was performed on Nova-Pak C18 reverse the adopted conditions within 18 min (Fig. 2). Each phase column (4 ㎛, 3.9 mm x 150 mm I.D., Waters, USA). separation of erythromycin (15 min), roxithromycin (16 Analytical system was operated with a gradient elution at min), tylosin (12 min) and tiamulin (14 min) was achieved flow rate of 0.5 ㎖/min. The mobile phase consisted of 10 successfully, on the same chromatogram. mM ammonium acetate (A) and methanol (B). Gradient runs were programmed as follows: 100% B for 3 min, 3-2. Mass spectra decrease from 100% to 90% B for 6 min, decrease from 90% For each molecule, the produced ions on mass spectra to 5% B for 6 min, 5% B for 5 min, re-equilibration with were the molecular related ion [M+H]+, two adduct ions 100% B for 5 min, post-run with 100% B for 10 min, until [M+Na]+ and [M+K]+, and several fragmentation ions (Fig. the next sample injection. 3). The molecular ions, [M+H]+, at m / z 734.5, 837.5, 494.4 The nebulizer gas was flowed at 45 p.s.i., 350℃ and 9.0 and 916.5 for erythromycin, roxithromycin, tiamulin and l/min and quadrupole was heated to 100℃. Mass spectrometer tylosin were represented dominantly. Except tiamulin, other was run in the positive mode and scan mode from m/z 100 drugs produced two adduct ions, [M+Na]+ and [M+K]+, at to 1000. Fragmentation voltage was 100 V. Analysis was m/ z 756.5 and 772.5 for erythromycin, at m/ z 859.5 and carried out at the room temperature. 875.5 for roxithromycin, at m/ z 938.5 and 954.5 for tyloisin (Fig. 3). The appearance of fragmentation ions was due to 2-3 Calibration curve s and the limit of de tection the dissociation of amino or sugar moieties on the structure Calibration curves have been constructed by plotting area of drugs. The m / z 576.5 and 679.5 in mass spectra of against the standard concentrations of macrolides in the erythromycin and roxithromycin were corresponding with Simultaneous Determination of Various Macrolides by Liquid Chromatography/Mass Spectrometry 105

F ig . 2. Tot a l ion ch rom a togra ph y (TIC, A) of m a crolide a n d t ia m u lin . E xt ra ct ion s ch r om a t ogr a ph y (E IC) of tia m u lin (B), t ylosin (C), er yth r om ycin (D) a n d r oxit h rom ycin (E ). a [M-desosamine+H]+. The m / z 158.1 of erythromycin and wit h pr e-colu m n deriva tiza tion pr ocedu r es r equ ires lon g roxithromycin was corresponding with a [desosamine + H]+, sep a r a t ion t im es a n d is less sen sit ive t h a n LC/MS [12-14]. and the m/ z 115.1 of roxithromycin was corresponding with In a ddition, fluorim etric detection is lim ited for sim u lt a n eou s a [cladinose-OCH3 + H]+. The fragment ions of tiamulin, det er m in t ion beca u se of t h e differen t der iva t iza t ion m et h od m/ z 192.1 was a moiety of [2-(diethylamino)-ethyl, thio] of ea ch dr u g. K ees et al. [16] a n d Dr ea ssi et a l. [17] h a ve acetic acid dissociating from m olecu la r ion . Th e fr a gm en t repor t ed for t h e determ in a tion m et h ods of er yth rom ycin a n d ion s of t ylosin , m / z 742.5 a n d 772.5 w er e corr espon din g roxith rom ycin u sin g H P LC w ith elect roch em ica l det ector , wit h [M -m ycin ose+H ]+ a n d [M -m yca rose+H ]+, r espect ively. wh ich is m or e sen sit ive t h a n U V detector. Bu t , th ese Th ese r esu lt s were su m m a r ized in Ta ble 1. m et h ods a re difficu lt to set u p a n a lyt ic con dit ion beca u se th e d et er m in a t ion m et h ods by elect r ochemical detection are 3-3 . Lin e a rity a n d th e lim it o f d e te c tio n very sensitive to environmental condition. All experim en ted dr u gs in t h e r a n ge of 0.001 ㎍/g ~ 5㎍/g The determination method by gas chromatography-mass showed good linearity, with correlation coefficient of spectrometry (GC-MS) has been reported [5]. This method 0.99(Table 2). Th e lim it of detection a n d lim it of qu a n t it a t ion needs the derivatization procedures for each macrolide, thus r a n ged from 0.001 t o 0.01 ㎍/g and from 0.005 to 0.05 ㎍/g taking a long time for the determination of macrolides. LC/ (Ta ble 3), r espect ively. Th ese figu r es w er e m u ch lower t h a n MS which omits the derivatization procedures was successfully t h e M RLs set u p by t h e S ou th K or ea . applied to determine several macrolides. Simultaneous de- termination methods by HPLC with UV detector [6-11] have 4. D iscussion been developed, but these methods are difficult to detect LC/M S w a s h igh ly sen sit ive a n d select ive for th e macrolides such as erythromycin and roxithormycin due to sim u lt a n eou s d et er m in a t ion of m a cr olid es com pa r in g w ith their weak UV absorbance. The fluorimetric detection with ot h er pu blish ed m et h ods. Sever a l m et h ods w er e r epor t ed for pre-column derivatization procedures requires long separation sim u lt a n eou s det er m in a t ion of m a crolides. Sim u lt a n eou s times and is less sensitive than LC/MS [12-14]. In addition, determ in a tion m et h ods by H P LC w ith U V d et ect or [6-11] fluorimetric detection is limited for simultaneous determintion h a ve been developed, bu t th ese m eth ods a r e difficu lt to because of the different derivatization method of each drug. detect m a crolides su ch a s eryt h r om ycin a n d r oxit h or m ycin Kees et al. [16] and Dreassi et al. [17] have reported for the du e to t h eir w ea k U V a bsor ba n ce. Th e flu orim et ric det ect ion determination methods of erythromycin and roxithromycin 106 Youn-Hwan Hwang, Jong-Hwan Lim, Byung-Kwon Park and Hyo-In Yun

(A)

(B)

(C)

(D )

ig. 3. Th e m a ss spect ra of erythromycin (A), roxithromycin (B), tylosin (C) and tiamulin (D). Simultaneous Determination of Various Macrolides by Liquid Chromatography/Mass Spectrometry 107

Table 1. Summarized mass spectra of the drugs used experiments

Molecular ion Adduct ions D rugs Molecular m ass Fragm entation ions [M + Na]+ [M + N a]+ ; [M + K]+ Erythromycin 733.5 734.5 756.5; 772.5 158.1; 558.3; 576.5 Roxithrom ycin 836.5 837.5 859.5; 875.5 115.1; 158.1; 679.5 Tiamulin 493.4 494.4 － 192.1 Tylosin 915.5 916.5 938.5; 945.5 742.5; 772.5; 794.5

Table 2. The linearity of the drugs

Equationa D rugs RSDb Linearity(r) Slope(106) Intercept(104) Erythrom ycin 7.02 -1.93 1.45 0.99 Roxithrom ycin 4.90 -73.49 8.43 0.99 Tiam ulin 43.3 -65.12 4.2 0.99 Tylosin 3.9 -32.2 4.2 0.99

a is used for the back-calculation of the drug concentration in the sample, y=mx + b (m, the slope; b, the intercept; x, the amount; y, the area) b is the residual standard deviation of calibration curve in the regression analysis and is represented as 105

Table 3. LOD, LOQ and reproducibility of four drugs

D rugs LOD(㎍/㎖) LOQ(㎍/㎖) Reproducibility(r) Erythromycin 0.005 0.02 0.99 Roxithrom ycin 0.01 0.05 0.99 Tiamulin 0.001 0.005 0.99 Tylosin 0.001 0.01 0.99 using HPLC with electrochemical detector, which is more that of major component. Even changing the mobile phase, sensitive than UV detector. But, these methods ar e difficult the minor component was not separated and was moved to set up analytic condition because the determination together with the major component. The chromatographic methods by electrochemical detection are very sensitive to property of partially overlapped peak in erythromycin is environmental condition. similar to that of the major component. As this overlapped As described in above as to the simultaneous determination peak resulted from the addition of the minor component, of macrolides, there are several problems such as weak UV LC/MS based on their molecular weight could identify the absorption, long separation time and difficult derivatization minor component. procedure. Our method has solved previous problems by application of liquid chromatorygraphy/mass spectrometry 5. Conclusion (LC/MS). LC/MS minimizes chromatographic separation and method development time in confirming the molecular LC/MS with electrospray is a simple, rapid and effective identities of the target substance. technique for the simultaneous determination of macrolides. The partially overlapped peak in erythromycin observed The fragmentation patterns provide the confirmatory in Fig. 3 needs some discussion in this study. Macrolides information of macrolides. The relevance of these studies for were generally composed of more than one structural the determination of macrolide in biomatrices remains component. In determination of macrolides, major components further investigated. were generally used as indicators to evaluate the residue levels [7]. However, minor components could be also Reference remained in edible tissues. This peak in erythromycin indicates a major component combined with a minor 1. P rescott, J .F. and Baggot, J .D. Antimicorbial component. Its mass spectrum pattern was different from Therapy in Veterinary Medicine (2nd ed.), Iowa State 108 Youn-Hwan Hwang, Jong-Hwan Lim, Byung-Kwon Park and Hyo-In Yun

University Press, Ames, Iowa, 1993. chloride and fluorescence detection. J Chromatogr B Biomed 2. Moats, W.A. and Medina, M.B. Veterinary Drug Sci Appl, J Chromatogr B Biomed Sci Appl, 1998, Residues, ACS Symposium Series 636, American Chemical 720(1-2), 89-97. Society, Washington, DC, 1996, p. 5. 13. Tsuji, K. Fluorimetric determination of erythromycin 3. Ministry of Health and We lfare. Official Methods for and erythromycin ethylsuccinate in serum by a high- Residual Substances in Livestock Products, Ministry of performance liquid chromatographic post-column, on-stream Health and Welfare, J apan, 1994. derivatization and extraction method. J Chromatogr, 4. AOAC International. Official Methods of Analysis of 1978, 1 58, 337-48. AOAC International (16th ed.), AOAC International, 14. To d , M., B ia re z, O., N icila s, P . a n d P e tije a n O. Arlington, VA, 1995. Sen sitive deter m in a t ion of josa m ycin a n d r ok it a m ycin in 5. Takatsuki, K., Ushizaw a, I. and Shoji, T. Gas plasm a by h igh -perform an ce liquid ch rom a tography w it h chromatographic-mass spectrometric determination of fluorescence detection. J Chrom a togr, 1992, 575(1), 171-6. macrolide antibiotics in beef and pork using single ion 15. H is a k o , M., Ma sa sh i, Y., Is a m u , A. a n d H a y a o , U . monitoring. J Chromatogr, , 1987, 391(1), 207-217 H ighly sen sitive h igh -perform an ce liquid ch r om a t ogr a ph ic 6. Carme n, L., Rosa, C., Ram on, C., Mercè, G. and determ in ation m ethod for a new eryth rom ycin der iva t ive, D olors P rat M. Determination of macrolide antibiotics E M523, a n d it s m a jor m eta bolit es in h u m a n pla sm a a n d by liquid chromatography. J Chromatogr A, 2001, u rin e usin g postcolu m n tris(2,2'-bipyridine) r u t h en iu m (Ⅲ) 910(2), 285-290. ch em ilu m in escen ce detect ion ., J Ch rom a togr B Biom ed 7. Horie, M., Saito, K., Ishii, R., Yoshida, T., Haram aki, Sci Appl, 1997, 690 (1-2), 305-313. Y. and Nakazaw a, H. Simultaneous determination of 16. Ke e s , F ., S p a n g le r, S . a n d We lle n h o fe r, M. Deter- five macrolide antibiotics in meat by high-performance m in a t ion of m a cr olides in biologica l m a t r ices by h igh - liquid chromatography. J Chromatogr A, 1998, 812(1-2), perform an ce liquid chrom a togra ph y with electrochem ica l 295-302. detection. J Ch rom a togr A, 1998, 812(1-2), 287-293. 8. Moats, W.A. Chromatographic methods for determination 17. D re a ss i, E ., Co rti, P ., B e zzin i, F . a n d F u rla n e tto , S . of macrolide antibiotic residues in tissues and milk of H igh -p er for m a n ce liqu id ch r om a t ogr a p h ic a ssa y of food-producing animals. J Assoc Off Anal Chem, 1985, erythrom ycin from biologica l m a trix usin g electr och em ica l 68(5), 980-4 or u ltr a violet det ect ion . An a lyst , 2000, 125(6), 1077-81. 9. Kanfer, I., Skinner, M. F. and Walker, R. B. Analysis 18. N ie ss e n , W. M. A. An a lysis of a n tibiot ics by liqu id of macrolide antibiotics. J Chromatogr. A, 1998, 821(1-2), ch r om a t ogr a p h y¯m a ss spect rom etr y. J Ch r om a t ogr A, 255-286. 1998, 8 12(1-2 ), 53-75. 10. How ard, D. and Cow en, T. Determination of tiamulin 19. D ra is ci, R., P a lle s ch i, L., F e rre tti, E., Ac h e n e , L. hydrogen fumarate in animal feeds using high-performance a n d Ce cilia , A. Con firm a t or y m eth od for m a cr olide liquid chromatography. Analyst, 1982, 107(1272), 319-23. residues in bovine tissues by micro-liquid ch r om a t ogr a ph y¯ 11. Markus, J . R. and Sherm a, J ., Method. III. Liquid t a n dem m a ss spectr om et r y. J Ch r om a t ogr A, 2001, chromatographic determination of tiamulin hydrogen 926 (1), 97-104. fumarate in complete swine meal feeds. J AOAC Int, 20. D u b o is , M., F lu c h a rd , D ., S io r, E. a n d D e la h a u t, P . 1993, 76(2), 449-50. Id en tifica t ion a n d qu a n t ifica t ion of five m a crolide 12. Sastre Toraño, J . and Guche laar, H. J . Quantitative a n t ibiotics in severa l t issu es, eggs a n d m ilk by liqu id determination of the macrolide antibiotics erythromycin, ch rom a tograph y¯electrospray ta ndem m a ss spect rom etr y. roxithromycin, and in human J Ch rom a togr B Biom ed Sci Appl, 2001, 753(2), serum by high-performance liquid chromatography using 189-202. pre-column derivatization with 9-fluorenylmethyloxycarbonyl