Journal of Chromatography A, 1216 (2009) 5296–5301
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Aminoglycoside antibiotics may interfere with microbial amino sugar analysis Chao Liang a,b,∗ , Joel A. Pedersen a,c , Teri C. Balser a a
Department of Soil Science, University of Wisconsin, Madison, WI 53706, USA Great Lakes Bioenergy Research Center (GLBRC), University of Wisconsin, Madison, WI 53706, USA c Molecular and Environmental Toxicology Center, University of Wisconsin, Madison, WI 53706, USA b
a r t i c l e
i n f o
Article history: Received 13 November 2008 Received in revised form 27 April 2009 Accepted 6 May 2009 Available online 15 May 2009 Keywords: Amino sugar Aminoglycoside Biomarker Microbe
a b s t r a c t The amino sugars (e.g., glucosamine, galactosamine, mannosamine, muramic acid) in soils are frequently employed as biomarkers of microbial residues. The analysis of amino sugars in environmental matrices, however, is expected to be more complicated than their determination in isolated microbial cells. In this study, we employed a widely used protocol for amino sugar analysis, and found that some aminoglycoside antibiotics interfere with amino sugar quantification in vitro. The method converts the aminoglycosides to compounds that coelute with the aldononitrile acetate derivatives of the amino sugars. Specifically, streptomycin significantly interferes with muramic acid analysis, and kanamycin, tobramycin and amikacin hamper glucosamine measurement. Mass spectrometry confirmed that the interfering compounds from aminoglycosides are not actually genuine microbial amino sugar monomers (bacterial muramic acid or fungal glucosamine), and are most likely to be N-methyl glucosamine or 3-amino-3-deoxy-glucopyranose. In contrast to their effects on muramic acid and glucosamine analyses, aminoglycosides do not interfere with galactosamine and mannosamine quantification. The few data that exist on the environmental occurrence of aminoglycoside antibiotics suggest they occur at only trace levels. Our findings may have implications for the qualitative and quantitative validity of results from amino sugar assays in some context. Application of the aldononitrile acetate derivatization method to samples (especially in selective microbial cultures using aminoglycosides as inhibitors) requires that potential interference be evaluated. Published by Elsevier B.V.
1. Introduction Analytical approaches to studying microbial biomass and communities are crucial for a broad understanding of the status and function of microorganism within ecosystems. Strategies for microbial analysis often rely on traditional laboratory culturedependent techniques as well as those based on direct extraction and analysis of specific biomarkers (e.g., lipids, amino sugars, ergosterol) [1–3]. Biomarker analyses circumvent the biases inherent in culture-based methods, and are therefore generally considered more reliable. Biomarkers are typically organic molecules with measurable cellular abundances and known biosynthetic origins. In environmental microbiology, potential biomarkers identified in laboratory cultures have been applied to field samples to infer the composition of microbial communities in nature. The presence of biomarker
∗ Corresponding author at: Department of Soil Science, University of Wisconsin, 1525 Observatory Drive, Madison, WI 53706-1299, USA. Tel.: +1 608 265 4850; fax: +1 608 265 2595. E-mail address:
[email protected] (C. Liang). 0021-9673/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.chroma.2009.05.010
compounds reflects a specific biological source as well as its degradates (e.g., muramic acid traces peptidoglycan parent compounds and residues) in the environment. The ability for a biomarker to reveal useful, accurate information depends on the taxonomic specificity of that marker, as well as the robustness of methods used for its analysis. Biomarker analysis in complex environmental matrices may be complicated by the presence of interfering compounds not encountered in analysis of isolated cells [4,5], and therefore requires extensive validation. Substantial effort has been focused on determining the taxonomic specificity of biomarkers; however, commonly employed analytical methods still lack comprehensive validation and thus warrant further investigation. The amino sugars glucosamine (GluN), galactosamine (GalN), mannosamine (ManN) and muramic acid (MurA) are associated with microbial cell walls and are commonly present in soils; plants and arthropods contribute negligible amounts to the total soil amino sugar pool [6,7]. Muramic acid occurs as N-acetylmuramic acid in the bacterial cell wall peptidoglycan and is not produced by eukaryotic cells [8]. Galactosamine is considered to be primarily of bacterial origin [9]. Bacterial peptidoglycan contains equal amounts of MurA and GluN, but in soils, GluN derives predominantly from fungal chitin rather than bacterial peptidoglycan [7]. The origin of
C. Liang et al. / J. Chromatogr. A 1216 (2009) 5296–5301
ManN is the subject of controversy [10], and ManN is best used only in combination with other amino sugars to estimate the total microbial amino sugar pool [7,11]. Derivatization of amino sugars to aldononitrile acetates followed by gas chromatography (GC) with flame ionization detection (FID) has emerged as a convenient method to simultaneously analyze multiple amino sugars [7,12,13]. Although the aldononitrile acetate method for amino sugar analysis has gained popularity in the earth sciences, the introduction of artifacts from other compounds present in environmental matrices has not been adequately investigated. We recently reported that the presence of the widely used aminoglycoside antibiotic streptomycin significantly interferes with MurA analysis by a commonly used aldononitrile acetate derivatization GC-FID method, potentially compromising the utility of MurA as a biomarker under certain conditions[5]. Aminoglycoside antibiotics are composed of amino sugars attached to an aminocyclitol, and occur in soil environments due to anthropogenic inputs and in situ production by actinobacteria. Most commercially produced aminoglycoside antibiotics are used in human therapy and animal husbandry [14,15]; a small fraction is used to control plant pathogens [16]. Aminoglycosides are poorly absorbed in the gastrointestinal tract of dosed organisms ( 95) for the MurA derivative were m/z = 115, 181, 236, 296, while those for the GluN derivative were m/z = 98, 127, 187, 289. The mass spectrum of the streptomycin derivative contained major ions (>95) at m/z = 112, 123, 187, 289, while those of kanamycin, tobramycin and amikacin derivatives exhibited a very similar pattern with major ions at m/z = 126, 168, 186, 288. Although the mass spectra of these three aminoglycoside derivatives were largely similar to that of the GluN derivative, several major ions differed by one mass unit.
Fig. 2. Mass spectra of aldononitrile acetate derivatives. Std 1: muramic acid, Std 2: glucosamine, A: streptomycin, C: kanamycin, E: tobramycin, F: amikacin.
5300
C. Liang et al. / J. Chromatogr. A 1216 (2009) 5296–5301
4. Discussion Our first objective was to determine whether aminoglycoside antibiotics would interfere with amino sugar identification. In the amino sugar method investigated, several pairs of amino sugar derivatives and aminoglycoside derivatives had similar retention times: the MurA derivative eluted with the streptomycin derivative, and the GluN derivative eluted at a retention time similar to those of the kanamycin, tobramycin and amikacin derivatives (Fig. 1). These data suggested that streptomycin would interfere with MurA analysis by this method, and that kanamycin, tobramycin and amikacin would interfere with GluN analysis. Our findings reveal that the presence of aminoglycoside antibiotics in samples may interfere with identification of amino sugars if confirmatory analyses are not conducted, potentially compromising interpretation of the bacterial and fungal residue mass. To address our second objective of quantifying the potential errors by the presence of aminoglycosides, we calculated molar response factors for the derivatives of aminoglycoside antibiotics relative to those of the amino sugars with which they may coelute (Table 1). We found that with the exception of netilmicin and spectinomycin, the aminoglycosides tested interfere with GluN or MurA quantification in the widely employed protocol of acid hydrolysis, aldononitrile acetate derivatization and GC-FID analysis. For quantitative estimation of GluN in environmental samples, the peaks induced by streptomycin, neomycin and gentamicin derivatives are not expected to significantly interfere because the FID signals from these aminoglycosides are low in our study (Fig. 1); furthermore, the molar response factors for streptomycin, neomycin and gentamicin derivatives were ∼50 times lower relative to that of the GluN derivative (e.g., if streptomycin was present at a 10-fold molar excess, the peak of the streptomycin derivative would be 20% of that for the GluN) indicating that, unless present in extremely large concentrations, they are unlikely to substantially interfere with the quantification of GluN (Table 1). To determine the nature of the interference in amino sugar analysis caused by the presence of aminoglycosides, we examined the structures of the amino sugar and aminoglycoside derivatives. Our purpose was to evaluate whether the interference by aminoglycosides was specific to the method employed or expected to be experienced in other methods for amino sugar analysis. If the processing of samples containing aminoglycoside antibiotics generate GluN or MurA, the interference by these antibiotics may be unavoidable even if other methods are adopted. Conversely, if the
interference by aminoglycosides is attributable to coelution under the chromatographic conditions employed, the problem may be addressed by altering separation conditions (e.g., changing the stationary phase, altering temperature ramp), using a confirmatory detector, or employing a different analytical method. To achieve the objective of determining the nature of the interference (whether by actual microbial amino sugars or by other compounds), we examined the mass spectra of the amino sugar and aminoglycoside derivatives. The mass spectra of organic molecules are frequently distinct enough to allow their discrimination from one another. We compared the mass spectra of the MurA derivative with that of the streptomycin derivative, and that of the GluN derivative with those of the kanamycin, tobramycin and amikacin derivatives (Fig. 2). We previously detailed the interference of streptomycin with MurA analysis, and proposed that the interference was due to N-methyl glucosamine formed from streptomycin after glycosidic cleavage [5]. In the present study, we found that the aldononitrile acetate derivatives from kanamycin, tobramycin and amikacin produce very similar fragmentation patterns, suggesting that the interfering compounds from these antibiotics likely correspond to a single compound or to stereoisomers yielding very similar mass spectra that do not allow differentiation. Comparison of the structures of these antibiotics (Fig. 3) reveals that kanamycin, tobramycin and amikacin do not possess a GluN moiety (2-amino-2-deoxy-glucopyranose); rather, they contain the GluN isomer (3-amino-3-deoxy-glucopyranose). The hypothesized reaction pathway for the derivatization of this GluN isomer is presented in Fig. 4. The identity of this interfering GluN isomer cannot be verified by GC–MS because an authentic standard is not commercially available at this time. In this GluN isomer, the amino group is located on carbon 3 rather than carbon 2 as in GluN. This difference does not appear to influence chromatographic retention under the conditions employed. The presence of the spatial isomers (stereoisomers) among kanamycin, tobramycin and amikacin provides an explanation for the largely similar mass spectra produced by the aldononitrile acetate derivatives from these compounds. The lack of precise correspondence between the mass spectra of the derivatives from these aminoglycoside antibiotics (kanamycin, tobramycin and amikacin) and that of the GluN derivative may be due to the positional isomerism in the structures of these aminoglycosides relative to GluN. The contrasting mass spectra of the aminoglycoside and amino sugar derivatives indicate that the interference is due to coelution of the derivatives from these aminoglycosides that differed from those of MurA or GluN. Accordingly, the interference of
Fig. 3. Chemical structure of kanamycin, tobramycin and amikacin. Grey region shows the stereoisomers that are expected to interfere with glucosamine analysis.
C. Liang et al. / J. Chromatogr. A 1216 (2009) 5296–5301
5301
Fig. 4. Hypothesized structural isomers of aldononitrile acetate derivatives of (A) glucosamine (2-amino-2-deoxy-glucopyranose) and (B) 3-amino-3-deoxy-glucopyranose from the aminoglycoside antibiotics kanamycin, tobramycin and amikacin.
aminoglycosides with amino sugar analysis is due to methodological artifacts that are distinct from microbial amino sugars, and in principle, may be addressed by protocol modifications. We have demonstrated that aminoglycoside antibiotics may interfere with microbial amino sugar analysis. Many investigators have determined amino sugars in environmental samples by acid extraction, derivatization to aldononitriles, and separation of derivatives by GC [12]; most previous studies relied solely on the non-selective detector FID [21–28]. At present, the few data that exist on the environmental occurrence of aminoglycoside antibiotics indicate that they are present at trace concentrations (e.g., g L−1 in water [19]). We expect that the levels of aminoglycoside antibiotics from natural and anthropogenic sources are too low to significantly interfere with amino sugar quantification. On the other hand, some previous amino sugar determinations were based on microorganisms cultured on selective media containing aminoglycoside antibiotics [2,29]. Conclusions from such studies may warrant re-examination in light of the present findings. Based on this study, our key concern is to alert researchers to consider the possibility of artifacts when using amino sugars as the microbial residue biomarkers, especially in the selective microbial culture using aminoglycoside antibiotics as inhibitors. Furthermore, the potential for aminoglycoside antibiotics to interfere with amino sugar analysis should be considered in design of future studies. Existing procedures may need to be refined or new methods developed. Depending on the study objectives, some possible solutions present themselves: unambiguous identification of trace MurA and GluN in complex matrices requires sophisticated instrumentation, for example, gas chromatography–tandem mass spectrometry with multiple reaction monitoring; precise quantification of MurA and GluN in the presence of interferents may be addressed by the calibration on some dominant mass ions unique to the biomarker derivatives. In summary, we demonstrate that specific aminoglycosides can interfere with the analysis of key amino sugars (viz. MurA and GluN) by a widely used GC-FID technique that relies on aldononitrile acetate derivatization method. Such interference has implications for evaluating the role played by microorganisms in ecological processes. Observing a chromatographic peak at the correct retention time does not constitute definitive identification of an amino sugar. This contribution serves as a cautionary note regarding potential difficulties that may be encountered when microbial
biomarkers identified in laboratory cultures are applied to the environment. Acknowledgements This study was funded by grants from USDA-CSREES McIntireStennis Act and DOE Great Lakes Bioenergy Research Center. The authors gratefully acknowledge the indispensable assistance of Drs. Harry W. Read, Kennedy F. Rubert IV and Xin Wei. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]
[18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]
J.R. Vestal, D.C. White, Bioscience 39 (1989) 535. B. Glaser, M.B. Turrion, K. Alef, Soil Biol. Biochem. 36 (2004) 399. X.R. Zhao, Q. Lin, P.C. Brookes, Soil Biol. Biochem. 37 (2005) 311. M.A. Hoijer, M.J. Melief, C.G. Vanheldenmeeuwsen, F. Eulderink, M.P. Hazenberg, Infect. Immun. 63 (1995) 1652. C. Liang, H.W. Read, T.C. Balser, Microb. Ecol. 57 (2009) 494. J.W. Parsons, in: E.A. Paul, J.N. Ladd (Eds.), Soil Biochemistry, Marcel Dekker, New York, 1981. W. Amelung, in: R. Lal, J.M. Kimble, R.F. Follett, B.A. Stewart (Eds.), Assessment Methods for Soil Carbon, CRC/Lewis Publishers, Boca Raton, FL, 2001, p. 233. K.H. Schleife, O. Kandler, Bacteriol. Rev. 36 (1972) 407. F.J. Sowden, K.C. Ivarson, Can. J. Soil Sci. 54 (1974) 387. R.R.R. Coelho, D.R. Sacramento, L.F. Linhares, Eur. J. Soil Sci. 48 (1997) 425. C. Liang, X. Zhang, K.F. Rubert, T.C. Balser, Biol. Fertil. Soils 43 (2007) 631. X. Zhang, W. Amelung, Soil Biol. Biochem. 28 (1996) 1201. G.O. Guerrant, C.W. Moss, Anal. Chem. 56 (1984) 633. L.S. Gonzalez, J.P. Spencer, Am. Fam. Physician 58 (1998) 1811. A.K. Sarmah, M.T. Meyer, A.B.A. Boxall, Chemosphere 65 (2006) 725. P.S. McManus, V.O. Stockwell, G.W. Sundin, A.L. Jones, Annu. Rev. Phytopathol. 40 (2002) 443. H.F. Chambers, in: J.G.L. Hardman, L.E. Limbird, A.G. Gilman (Eds.), Goodman and Gilman’s the Pharmacological Basis of Therapeutics, McGraw-Hill Companies, Inc, New York, NY, 2001, p. 1219. S. Thiele-Bruhn, J. Plant Nutr. Soil Sci. 166 (2003) 145. D. Löffler, T.A. Ternes, J. Chromatogr. A 1000 (2003) 583. J.M. Serrano, M. Silva, J. Chromatogr. A 1117 (2006) 176. D. Said-Pullicino, K. Kaiser, G. Guggenberger, G. Gigliotti, Chemosphere 66 (2007) 2166. J. Niggemann, C.J. Schubert, Geochim. Cosmochim. Acta 70 (2006) 2229. C. Liang, R. Fujinuma, L. Wei, T.C. Balser, Forestry 80 (2007) 65. G. Guggenberger, S.D. Frey, J. Six, K. Paustian, E.T. Elliott, Soil Sci. Soc. Am. J. 63 (1999) 1188. R.T. Simpson, S.D. Frey, J. Six, R.K. Thiet, Soil Sci. Soc. Am. J. 68 (2004) 1249. W. Amelung, J.M. Kimble, S. Samson-Liebig, R.F. Follett, Soil Sci. Soc. Am. J. 65 (2001) 1704. X. Zhang, W. Amelung, Y. Yuan, S. Samson-Liebig, L. Brown, W. Zech, Appl. Soil Ecol. 11 (1999) 271. D. Solomon, J. Lehmann, W. Zech, Biol. Fertil. Soils 33 (2001) 33. G. Wang, H. He, X. Zhang, J. Li, L. Han, J. Wang, Chin. J. Appl. Ecol. 17 (2006) 827.
