Highly Selective Separation of Aminoglycoside Antibiotics on A

J. Sep. Sci. 2014, 37, 1781–1787 1781

Jie Wei1 Research Article Aijin Shen2 Huihui Wan3 Jingyu Yan2 Highly selective separation of Bingcheng Yang1 Zhimou Guo2 ∗ aminoglycoside antibiotics on a zwitterionic Feifang Zhang1 Click TE-Cys column Xinmiao Liang1,2

1School of Pharmacy, East China Hydrophilic interaction liquid chromatography has emerged as a valuable alternative ap- University of Science and proach to ion-pair chromatography for the separation of aminoglycoside antibiotics in recent Technology, Shanghai, China years. However, the resolution of structurally related aminoglycosides is a great challenge 2Key Laboratory of Separation Science for Analytical owing to the limited selectivity. In this work, a cysteine-based zwitterionic stationary phase Chemistry, Dalian Institute of (named Click TE-Cys) was utilized and compared with ﬁve commonly used hydrophilic Chemical Physics, Chinese interaction liquid chromatography columns. Click TE-Cys displayed much better selectivity Academy of Sciences, Dalian, for structurally similar aminoglycosides. The retention behaviors of aminoglycosides were China ࣙ 3Chemistry Analysis and investigated in detail, revealing that low pH (2.7 or 3.0) and high buffer concentration ( 50 Research Center, Faculty of mM) were preferable for achieving good peak shape and selectivity. Effective resolution Chemical, Environmental and of ten aminoglycosides including spectinomycin, dihydrostreptomycin, streptomycin, gen- Biological Science and tamicin C1, gentamicin C2/C2a, gentamicin C1a, kanamycin, paromonycin, tobramycin, Technology, Dalian University of Technology, Dalian, China and neomycin was realized at optimized conditions. Additionally, spectinomycin and its related impurities were successfully resolved. The results indicated the great potential of the Click TE-Cys column in the separation of aminoglycoside mixtures and related impurities. Received January 23, 2014 Revised April 22, 2014 Keywords: Aminoglycosides / Hydrophilic interaction chromatography / Separa- Accepted April 23, 2014 tion selectivity / Zwitterionic stationary phases DOI 10.1002/jssc.201400080 Additional supporting information may be found in the online version of this article at the publisher’s web-site

1 Introduction In the past few years, LC–MS/MS has become the most popular detection technique for the detection of aminoglyco- Aminoglycosides (Fig. 1) are broad-spectrum antibiotics side residues in foods. The strongly polar aminoglycosides (-mycin from Streptomyces spp. and -micin from Mi- exhibited limited retention in traditional RPLC. Therefore, cromonospora spp.) that have been widely used for treating ion-pair chromatography (IPC) has been employed for the Gram-negative and some Gram-positive bacterial infections. efﬁcient separation of aminoglycosides [5–13]. However, IPC The antibacterial mechanism of aminoglycosides is based is subject to long equilibration time, unstable retention time, on its irreversible binding to the ribosomal 30S subunit, and short column life, together with contamination of the thereby inhibiting protein synthesis [1, 2]. Aminoglycosides instrument and inhibition of MS detection sensitivity. HILIC are extensively implemented in veterinary practice for the has emerged as a valuable alternative in separating polar an- treatment of a wide range of infections, or administered tibiotics [14]. In recent years, the application of HILIC for as feed additives for prophylaxis and growth promoters [3]. resolving aminoglycosides has received increasing popular- The abuse of aminoglycosides would accelerate the emer- ity [15–19]. Without the use of ion-pair reagents, HILIC exhib- gence of aminoglycoside-resistant pathogens and/or cause ited much higher sensitivity than IPC [20]. But the separation aminoglycoside-residues in animal-originated foods, which selectivity of aminoglycosides in HILIC is limited. Kumar potentially threat human health. The European Union has and coworkers [21] investigated the resolution (Rs)ofamino- laid down strict maximum residue limits (MRLs) for aminoglycosides on the bare silica, amino, amide, and zwitterionic glycosides in a variety of food matrices [4]. stationary phases. The results demonstrated that the zwitterionic ZICR -HILIC column possessed the best selectivity. Nonetheless, effective separation of some structurally similar Correspondence: Dr. Feifang Zhang, School of Pharmacy, East aminoglycosides such as streptomycin/dihydrostreptomycin China University of Science and Technology, 130 Meilong Road, and gentamicin C1/C1a/C2/C2a was not achieved. Shanghai, 200237, China E-mail: [email protected] Fax: +86-21-64250627

∗ Abbreviations: IPC, ion-pair chromatography; Rs, resolution; Additional corresponding author: Dr. Zhimou Guo, Tf, tailing factor E-mail: [email protected]

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com 1782 J. Wei et al. J. Sep. Sci. 2014, 37, 1781–1787

Figure 1. Chemical structures of aminoglycoside antibiotics.

Consequently, the selection of a HILIC stationary phase with tained from National Institutes for Food and Drug Control superior selectivity is of great significance. (Beijing, China). Spectinomycin hydrochloride, actinamine Recently, we have developed a silica-based cysteine- (impurity A), actinospectinoic acid (impurity B), and (4S)- bonded zwitterionic HILIC stationary phase named Click TE- dihydrospectinomycin (impurity C) were kindly gifted from Cys [22, 23]. Click TE-Cys showed better hydrophilicity than National Institutes for Food and Drug Control. Sample so- commercially available HILIC columns (e.g. ZICR -HILIC, lutions were prepared in ammonium formate (100 mM, pH TSK Amide-80). It has been successfully applied in the sep- 3.2)/acetonitrile/water (5:40:55, v/v/v). aration of oligosaccharides, alkaloids, and protein tryptic di- Five commercially available HILIC columns were utilized gests with high selectivity and efficiency. In view of its good for comparison. Unitary Diol and XAmide were purchased hydrophilicity and selectivity, a Click TE-Cys column was from Acchrom (Beijing, China). TSKgel Amide-80 was from employed for the separation of aminoglycosides, aiming at Tosoh (Tokyo, Japan). Both zwitterionic HILIC columns from improving the selectivity of structurally related aminoglyco- Merck (Darmstadt, Germany) were ZICR -HILIC and ZICR - sides. The comparison of Click TE-Cys with five commercial cHILIC. A homemade Click TE-Cys column was used. The HILIC columns was performed. The versatility and capabil- characteristics of the columns are listed in Table 1. ity of Click TE-Cys column was validated by the Rs of ten Acetonitrile and methanol of HPLC grade were obtained commonly used aminoglycosides. Besides, the separation of from Merck (Darmstadt, Germany). Water was purified using spectinomycin with its related impurities was successfully a Milli-Q purification system (Billerica, MA, USA). Ammo- realized. nium formate (97%) and formic acid (98%) were obtained from J&K Scientific (Beijing, China).

2 Materials and methods 2.2 LC with evaporative light scattering detection 2.1 Materials and reagents A Waters ACQUITY UPLC H-Class system (Milford, MA, Streptomycin sulfate (90%) and dihydrostreptomycin USA) consisting of a quaternary solvent manager, a FTN sesquisulfate (95%) were purchased from Dr. Ehrenstorfer sample manager, and an evaporative light scattering detector (Augsburg, Germany). Gentamicin sulfate (gentamicin was (ELSD) was used for chromatographic separation. The mo- a mixture of four components, gentamicin C1: 27.1%, C1a: bile phase was composed of ammonium formate aqueous 25.3%, C2: 30.1%, and C2a: 17.4%, among which gentamicin solution (mobile phase A), acetonitrile (B), and water (C). C2 and C2a were stereoisomers), kanamycin sulfate, paro- Throughout the experiment, the ratio of mobile phase A in momycin sulfate, neomycin sulfate, and tobramycin were ob- the mobile phase was kept constant to obtain a constant buffer

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com J. Sep. Sci. 2014, 37, 1781–1787 Liquid Chromatography 1783

Ta b l e 1 . The physical and chemical characteristics of the applied columns

Column name Bonded phase Chemical structure Dimensions Particle size Pore size

Unitary Diol Diol 4.6 × 150 mm 5 ␮m 100 A˚

XAmide Amide 4.6 × 150 mm 5 ␮m 100 A˚

TSKgel Amide-80 Amide 2.0 × 150 mm 3 ␮m80A˚

ZIC-HILIC Sulfobetaine 4.6 × 150 mm 5 ␮m 200 A˚

ZIC-cHILIC Phosphorylcholine 4.6 × 150 mm 3 ␮m 100 A˚

Click TE-Cys Cysteine 4.6 × 150 mm 5 ␮m 100 A˚

concentration. Flow rates of 1 and 0.2 mL/min were set for centration of 100 mM although gentamicin C1, C1a, and columns with size of 4.6 and 2 mm id, respectively. Column C2/C2a were not resolved. When the buffer concentration temperature was maintained at 30ЊC. The ELSD parameters was decreased to 10 mM, the Rs became worse. TSK Amide- were as follows: gain, 50; gas (N2)pressure,30psi;nebu- 80 exhibited good separation selectivity for streptomycin and lizer in heated mode at 70% (42ЊC); drift tube temperature, dihydrostreptomycin at both 100 mM (Rs = 1.90) and 10 mM 75ЊC. (Rs = 1.81) ammonium formate concentrations. However, the gentamicin components could not be resolved under ei- ther condition. With ammonium formate of 100 mM, the 3 Results and discussion isolation of dihydrostreptomycin and streptomycin was realized with Rs of 1.37 and 1.86 on ZIC-HILIC (sulfobetaine) 3.1 Comparison of separation selectivity between and ZIC-cHILIC (phosphorylcholine) [24, 25], respectively. Click TE-Cys and commercially available HILIC Gentamicin C1, C2/C2a, and C1a were partially resolved columns into two peaks on ZIC-HILIC, while three peaks were dis-

played on the ZIC-cHILIC column with Rs of 3.31 and 1.31. To investigate the separation selectivity of Click TE-Cys for ZIC-cHILIC displayed better selectivity than ZIC-HILIC for aminoglycosides, two sets of structurally similar components, aminoglycosides probably due to the existence of positively i.e. streptomycin and dihydrostreptomycin, gentamicin C1, charged quaternary ammonium group at the distal end of C1a, and C2/C2a (as shown in Fig. 1) were chosen as probes. the stationary phase. At 10 mM ammonium formate, the re- Five HILIC columns including diol-, amide-, and zwitterion- tention times of dihydrostreptomycin and streptomycin were based stationary phases (Table 1) were utilized for compari- increased obviously and no gentamicin peak was observed son. Two buffer concentration levels of 100 and 10 mM were within 20 min on both zwitterionic columns. The electro- employed for the separation. The acetonitrile gradient was static attraction between aminoglycosides and the zwitteri- optimized according to the hydrophilicity of each column to onic stationary phases was enhanced with decreasing buffer ensure retention times of 8–10 min at 100 mM ammonium concentration. The Click TE-Cys column, with positively and formate. negatively charged groups parallel to the surface of silica As shown in Fig. 2, the Unitary Diol column showed no gel [23] presented good selectivity for aminoglycosides at

Rs for streptomycin and dihydrostreptomycin, while only two 100 mM ammonium formate (Fig. 2). The Rs between di- peaks were observed for gentamicin C1, C1a, and C2/C2a hydrostreptomycin and streptomycin was 1.86, while the Rs components at both buffer concentration levels. Baseline between gentamicin C1 and C2/C2a, C2/C2a, and C1a were separation of streptomycin and dihydrostreptomycin was 3.30 and 2.18. In comparison, Click TE-Cys exhibited much achieved with a Rs of 1.49 on XAmide column at buffer con- better separation selectivity for aminoglycosides with similar

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com 1784 J. Wei et al. J. Sep. Sci. 2014, 37, 1781–1787

Figure 2. Separation of dihydrostreptomycin/streptomycin and gentamicin C1/C1a/C2/C2a on different columns. (A) streptomycin and dihydrostreptomycin; (B) gentamicin C1,C1a, and C2/C2a. Conditions: mobile phase A, 500 mM (or 100 mM) ammonium formate, pH 3.0, mobile phase B, acetonitrile, mobile phase C, water; gradient, 20% of 500 mM (100 mM in constant) or 10% of 100 mM (10 mM in constant) ammonium formate, the acetonitrile gradient was 70–50% (Unitary Diol, XAmide), 65–30% (TSK Amide-80), 60–30% (ZIC-HILIC, ZIC-cHILIC), and 60–10% (Click TE-Cys) in 20 min. structures than other HILIC columns at high level of buffer stronger. The results were further supported by the change concentration. of peak shapes in terms of tailing factors (Tf). At low pH values, e.g., pH 2.7 or 3.0, a Tf of 1.0–1.2 was obtained, indicating rather good peak symmetry. When the pH value was 3.2 Retention behaviors of aminoglycosides on the increased to 3.8 or 4.6, peak shape deteriorated signiﬁcantly. Click TE-Cys column Even though ammonium formate of 100 mM was utilized in the mobile phase, electrostatic attraction could not be com- In order to understand the retention characteristics of amino- pletely shielded and severe tailing peaks were still observed. glycosides on the Click TE-Cys column, the inﬂuence of Thus, pH 2.7 or 3.0 was preferable for resolving aminoglyco- buffer pH and concentration on the retention and peak shape sides. of dihydrostreptomycin, streptomycin, gentamicin C1, C1a, Meanwhile, the effect of buffer concentration was stud- and C2/C2a were investigated. The pH value of ammonium ied at pH of 2.7 (Fig. 4). With increasing buffer concentration formate was changed from 2.7 to 4.6, and the buffer concen- from 25 to 100 mM, the retention times of aminoglycosides tration in the mobile phase was kept constant at 100 mM. decreased and the peak shapes were effectively improved. The As shown in Fig. 3, the retention of aminoglycosides in- results indicated that electrostatic attraction involved in the creased with the increase of pH. As the ionization of carboxyl separation of aminoglycosides, i.e. the interaction between (and silanol) groups on Click TE-Cys material was enhanced aminoglycosides and carboxyl groups on Click TE-Cys still ex- with increasing pH value, the electrostatic attraction between isted even at pH value of 2.7. Perhaps the oppositely charged aminoglycosides and Click TE-Cys stationary phase became groups distributed parallel to the surface of the stationary

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com J. Sep. Sci. 2014, 37, 1781–1787 Liquid Chromatography 1785

Figure 4. Effect of buffer concentration on the retention and peak shape of aminoglycosides. Conditions: mobile phase, (A) 250 mM Figure 3. Effect of buffer pH on the retention and peak shape of ammonium formate, pH 2.7, (B) acetonitrile, C, water; gradient, aminoglycosides. Conditions: mobile phase, A, 250 mM ammo- 10/20/30/40% A in constant (corresponding to 25, 50, 75, and nium formate, pH 2.7/3.0/3.8/4.6, B, acetonitrile, C, water; gradi- 100 mM of buffer in the mobile phase), 60–20% B in 30 min. Note: ent, 40% A in constant, 60–20% B in 30 min. Note: DHS, dihy- DHS, dihydrostreptomycin; STR, streptomycin; GNT, gentamicin. drostreptomycin; STR, streptomycin; GNT, gentamicin. The tailing factor of GNT C1 at 50 mM was not acquired due to 2− incomplete separation with SO4 .

phase were convenient for the contact between aminoglycosides and carboxyl groups. High concentration of ammonium icin C1a and kanamycin occurred at a pH value of 3.0 and ࣙ formate ( 50 mM) was advantageous for improving the peak ammonium formate of 50 mM. The Rs capability of Click shape. In addition, the investigation of the chromatographic TE-Cys was much better than the traditional zwitterionic sta- behavior of aminoglycosides on the Click TE-Cys column and tionary phase [21]. An interesting phenomenon should be − 2− other commercial HILIC columns under different water con- noted that the anions Cl and SO4 could be well separated tents is discussed in the Supporting Information. (Fig. 5) on the Click TE-Cys column, indicating its feasibil- ity for simultaneous analysis of polar pharmaceuticals and counterions [26]. 3.3 Application in the separation of aminoglycoside On the other hand, the separation of aminoglycoside mixtures and impurities with its impurities, which are always structurally related substances, is important for the quality control of pharmaceu- To further demonstrate the applicability and versatility, ticals. In the product of spectinomycin hydrochloride (for the isolation of ten aminoglycoside mixtures including injection), impurities and related substances such as acti- spectinomycin, dihydrostreptomycin, streptomycin, gentam- namine (impurity A), actinospectinoic acid (impurity B) and icin C1, gentamicin C2/C2a, gentamicin C1a, kanamycin, (4S)-dihydrospectinomycin (impurity C) always exist. Quality paromonycin, tobramycin, and neomycin was carried out on monitoring was generally accomplished by IPC [27, 28]. In the Click TE-Cys column. According to the above results, low this work, the separation of spectinomycin and its impuri- pH (2.7 or 3.0) and high buffer concentration (ࣙ50 mM) were ties was successfully performed on the Click TE-Cys column

beneﬁcial for improving the peak shape. Thus, the Rs of all the in HILIC mode, revealing good selectivity (Fig. 6). In con- 10 aminoglycosides was implemented at the optimized con- sequence, Click TE-Cys stationary phase has great potential

ditions (Fig. 5). Good separation selectivity and peak shape for the Rs of aminoglycosides and impurities with similar were achieved except for one condition. Co-elution of gentam- structures.

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com 1786 J. Wei et al. J. Sep. Sci. 2014, 37, 1781–1787

4 Conclusion

The improvement of separation selectivity for aminoglycoside antibiotics in HILIC mode is challenging, espe- cially for structurally similar aminoglycosides such as streptomycin/dihydrostreptomycin, gentamicin C1/C1a/C2/C2a, etc. This work introduced a novel zwitterionic HILIC column (Click TE-Cys) for the highly selective separation of aminoglycosides. Compared with commercially available HILIC columns, Click TE-Cys exhibited superior separation selectivity for structurally similar aminoglycosides. Systematic op- timization of the separation conditions indicated that low pH value and high buffer concentration in the mobile phase was advantageous for improving selectivity and peak shape. The versatility of the method was further validated by the

efﬁcient Rs of ten aminoglycoside mixtures. Besides, the isolation of spectinomycin with its impurities was successfully realized. We believe that this method will be useful for the de- termination of trace aminoglycoside residues in food or other ﬁelds.

This work was ﬁnancially supported by National Natural Science Foundation of China (no. 21177040, 21005077). We also thank the Open Foundation of Key Laboratory of Industrial Ecology and Environmental Engineering for ﬁnancial support.

The authors have declared no conﬂict of interest.

Figure 5. Separation of aminoglycosides at (a) 50 mM, (b) 75 Mm, and (c) 100 mM buffer concentration with pH 2.7 and 3.0. Conditions: Mobile phase: A, 250 mM ammonium formate 5 References with the corresponding pH values; B, acetonitrile; C, water. Gra- dient: 20/30/40% A in constant, 0–30 min, 60–20% B. Compounds: [1] Magnet, S., Blanchard, J. S., Chem. Rev. 2005, 105, (1) spectinomycin, (2) dihydrostreptomycin, (3) streptomycin, (4) 477–497. gentamicin C1, (5) gentamicin C2/C2a, (6) gentamicin C1a, (7) kanamycin, (8) paromomycin, (9) tobramycin, (10) neomycin, (11) [2] Kohanski, M. A., Dwyer, D. J., Collins, J. J., Nat. Rev. − 2− Microbiol. 2010, 8, 423–435. Cl , (12) SO4 . [3] McGlinchey, T. A., Rafter, P. A., Regan, F., McMahon, G. P. , Anal. Chim. Acta 2008, 624, 1–15. [4] Ofﬁcial Journal of the European Union, Commission Regulation (EU) No 37/2010 of 22 December 2009 on pharmacologically active substances and their classiﬁ- cation regarding maximum residue limits in foodstuffs of animal origin. [5] Clarot, I., Regazzeti, A., Auzeil, N., Laadani, F., Citton, M.,Netter,P.,Nicolas,A.,J. Chromatogr. A 2005, 1087, 236–244. [6] Clarot, I., Chaimbault, P.,Hasdenteufel, F., Netter, P., Nico- las, A., J. Chromatogr. A 2004, 1031, 281–287. [7] van Bruijnsvoort, M., Ottink, S. J. M., Jonker, K. M., de Boer, E., J. Chromatogr. A 2004, 1058, 137–142. [8] Cherlet, M., De Baere, S., De Backer, P., J. Mass Spec- Figure 6. Separation of spectinomycin and its related impurities. trom. 2007, 42, 647–656. Conditions: mobile phase, A, 100 mM ammonium formate, pH 3.2, B, acetonitrile, C, water; gradient, keep 5% of mobile phase [9] Kaufmann, A., Butcher, P., Maden, K., Anal. Chim. Acta A in constant, and mobile phase B changed from 7535% in 2012, 711, 46–53. 20 min. [10] Almeida, M. P., Rezende, C. P., Souza, L. F., Brito, R. B., Food Addit. Contam. A 2012, 29, 517–525.

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com J. Sep. Sci. 2014, 37, 1781–1787 Liquid Chromatography 1787

[11] van Holthoon, F. L., Essers, M. L., Mulder, P. J., Stead, S. [19] Oertel, R., Neumeister, V., Kirch, W., J. Chromatogr. A L.,Caldow,M.,Ashwin,H.M.,Sharman,M.,Anal. Chim. 2004, 1058, 197–201. Acta 2009, 637, 135–143. [20] Gremilogianni, A. M., Megoulas, N. C., Koupparis, M. A., [12] Granja, R., Nino, A. M. M., Zucchetti, R. A. M., Nino, R. E. J. Chromatogr. A 2010, 1217, 6646–6651. M., Patel, R., Salerno, A. G., Anal. Chim. Acta 2009, 637, [21] Kumar, P.,Rubies, A., Companyo, R., Centrich, F., J. Sep. 64–67. Sci. 2012, 35, 498–504. [13] Zhu, W. X., Yang, J. Z., Wei, W., Liu, Y. F., [22] Shen, A. J., Guo, Z. M., Cai, X. M., Xue, X. Y., Liang, X. Zhang, S. S., J. Chromatogr. A 2008, 1207, M., J. Chromatogr. A 2012, 1228, 175–182. 29–37. [23] Shen, A. J., Guo, Z. M., Yu, L., Cao, L. W., Liang, X. M., [14] Kahsay, G., Song, H. Y., Van Schepdael, A., Cabooter, Chem. Commun. 2011, 47, 4550–4552. D.,Adams,E.,J. Pharm. Biomed. Anal. 2014, 87, 142–154. [24] Jiang, W., Fischer, G., Girmay, Y., Irgum, K., J. Chro- matogr. A 2006, 1127, 82–91. [15] Kumar, P.,Rubies, A., Companyo, R., Centrich, F., J. Sep. Sci. 2012, 35, 2710–2717. [25] Jiang, W., Irgum, K., Anal. Chem. 2002, 74, 4682–4687. [16] Bohm, D. A., Stachel, C. S., Gowik, P., Food Addit. Con- [26] Liu, X. D., Pohl, C. A., J. Chromatogr. A 2012, 1232, tam. A 2012, 29, 189–196. 190–195. [17] Ishii, R., Horie, M., Chan, W., MacNeil, J., Food Addit. [27] Zhou, J. Y., Zhang, L., Wang, Y., Yan, C., J. Sep. Sci. 2011, Contam. A 2008, 25, 1509–1519. 34, 1811–1819. [18] Oertel, R., Renner, U., Kirch, W., J. Pharm. Biomed. Anal. [28] Wang, J., Hu, X. J., Tu, Y., Ni, K. Y., J. Chromatogr. B 2006, 2004, 35, 633–638. 834, 178–182.

C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com