Analytical Profiling of Biosynthetic Intermediates Involved in the

The method of creating DOS-containing aminoglycoside profiles (a list of ..... The order of elution of the four components was C1a, C2 isomers (C2 and...
1 downloads 0 Views 253KB Size
Download PDF
Anal. Chem. 2007, 79, 4860-4869

Analytical Profiling of Biosynthetic Intermediates Involved in the Gentamicin Pathway of Micromonospora echinospora by High-Performance Liquid Chromatography Using Electrospray Ionization Mass Spectrometric Detection Je Won Park,† Jay Sung Joong Hong,‡ Niranjan Parajuli,† Hwa Soo Koh,† Sung Ryeol Park,† Mi-Ok Lee,§ Si-Kyu Lim,§ and Yeo Joon Yoon*,†

Center for Intelligent Nano-Bio Materials, Division of Nano Sciences and Department of Chemistry, Ewha Womans University, Seoul 120-750, Korea, Interdisciplinary Program of Biochemical Engineering and Biotechnology, Seoul National University, Seoul 151-742, Korea, and GenoTech Corporation, Daejeon 305-343, Republic of Korea

In the present study, we developed a sensitive and highly selective method of detecting the biosynthetic intermediates involved in the gentamicin pathway from a cell culture of Micromonospora echinospora. A novel extraction method utilizing a dual solid-phase extraction (SPE) technique was employed to purify and recover all of the gentamicin-related components from the cell culture broth, and high-performance liquid chromatography (HPLC) coupled with electrospray ionization mass spectrometry (ESI-MS/MS) was used to analyze the extractant for gentamicin intermediates. The pH of the culture broth was adjusted to an acidic condition of pH 2 prior to the extraction. The samples were first cleaned with a reversedphase AccuBOND C18 cartridge, and then the aminoglycosidic components were purified using a cationic exchanger OASIS MCX cartridge. The detection limit of a gentamicin standard spiked in blank medium processed by this method was found to be approximately 5 ng for each component of the gentamicin C complex, and the mean recovery for each component of standard gentamicin was above 91% when analyzed by HPLC-ESI-MS/MS. We further demonstrated that this method enables the analytical profiling of the gentamicin-related compounds produced by wild-type M. echinospora ATCC 15835, which mainly produces the gentamicin C complex, and the UV-induced mutant strain KCTC 10506BP, which produces gentamicin B as the major product. Seven intermediates (paromamine, gentamicin A2, B, X2, A, JI20A, and JI-20B) besides the gentamicin C complex were detected in the culture broth of both M. echinospora strains when analyzed by MS/MS for the distinct fragmentation patterns of each gentamicin component. This report displays the first example of the HPLC profiling in * To whom correspondence should be addressed. E-mail: [email protected]. Phone: +82 2 3277-4082. Fax: +82 2 3277-3419. † Ewha Womans University. ‡ Seoul National University. § GenoTech Corporation.

4860 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

a wide range of structurally related biosynthetic intermediates involved in the gentamicin pathway. Aminoglycosides are a large and diverse class of clinically important antibiotics that are used in the treatment of a wide variety of susceptible Gram-negative bacterial infections.1 Many of them occur naturally as metabolites of actinomycetes, particularly in the genera Streptomyces and Micromonospora. Aminoglycosides contain two or more aminosugars linked by glycosidic bonds to an aminocyclitol component, which in the majority of cases is 2-deoxystreptamine (DOS). The antibiotic gentamicin, along with kanamycin and tobramycin, is a typical example of such DOS-containing aminoglycosides. Gentamicin was discovered in 1963 in the fermentation product of Micromonospora purpurea or M. echinospora and has been found to be effective against severe Gram-negative bacterial infections.2 Recently, the potential antiviral properties of some gentamicin adducts have also been demonstrated.3 Aminoglycosides are often synthesized as a mixture of structurally related compounds by the above genera, and therefore, the therapeutic product contains a mixture of these aminoglycosides. The gentamicin drug complex used in clinical practice is made up of four closely related components designated as C1, C1a, C2a, and C2 (gentamicin C series), which differ from each other in their degree of methylation on the auxiliary aminosugar purpurosamine moiety (Figure 1). From the proposed gentamicin biosynthetic pathway,4,5 M. echinospora produces a number of aminoglycoside intermediates in addition to the gentamicin C series. All of these share DOS, to which one (purpurosamine) or two (garosamine and purpurosamine) sugar moieties are added, (1) Umezawa, H.; Hooper I. R. Aminoglycoside Antibiotics; Springer-Verlag: Berlin, 1982. (2) Weinstein, M. J.; Luedemann, G. M.; Oden, E. M.; Wagman, G. H.; Rosselet, J. P.; Marquez, J. A.; Coniglio, C. T.; Charney, W.; Herzog, H. L.; Black, J. J. Med. Chem. 1963, 122, 463-464. (3) Litovchick, A.; Evdokimov, A. G.; Lapidot, A. Biochemistry 2000, 39, 28382852. (4) Testa, R. T.; Tilley, B. C. Jpn. J. Antibiot. 1979, 32 (Suppl.), S47-S59. (5) Unwin, J.; Standage, S.; Alexander, D.; Hosted, T., Jr.; Horan, A. C.; Wellington, E. M. J. Antibiot. 2004, 57, 436-445. 10.1021/ac070028u CCC: $37.00

© 2007 American Chemical Society Published on Web 05/24/2007

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

4861

Figure 1. Molecular structures of the aminoglycoside intermediates involved in the proposed gentamicin biosynthetic pathway in M. echinospora.

generating DOS-containing aminoglycosides. The gentamicin intermediates could include paromamine and gentamicin A and C series on the basis of both the degree of sugar attachment and the type of functional group attached to the sugar. The molecular structures of these intermediates and their proposed biosynthetic pathway are illustrated in Figure 1. The resistant strains began to appear after decades of widespread use of natural aminoglycosides. The emergence of drugresistant pathogens that produce aminoglycosides-modifying enzymes, which inactivate the antibiotic activities of aminoglycosides, has highlighted the need to continue searching for novel aminoglycosides and semisynthetic derivatives that retain activity against the resistant organisms.6 In particular, the chemical modification of the existing aminoglycosides was pursued during the 1970s. This effort led to the discovery of some available precursors for semisynthetic derivatives; amikacin and isepamicin were derived from kanamycin A and gentamicin B, respectively.7 However, no synthetic aminoglycoside derivatives were found whose therapeutic properties were markedly superior to those of amikacin.8 Over the past decade, the advances made in the biotechnology used to characterize and manipulate the particular biosynthetic gene clusters found in microbes has led to the development of appealing strategies, such as combinatorial biosynthesis, not only for improving the production of antibiotics but also for the generation of novel antibiotics with greater potency. These genetic studies have encouraged the exploration of diverse aminoglycoside antibiotic biosynthetic pathways, and several biosynthetic gene clusters have been reported.5,9-11 The M. echinospora gene cluster for the gentamicin biosynthetic pathway has recently been identified, and the function of some genes encoding the very early steps of gentamicin biosynthesis were characterized in vitro.5,10 Especially, Unwin et al.5 found the gentamicin gene cluster in M. echinospora for the first time by using molecular biology tools such as cosmid library and gene disruption approaches. Furthermore, both the role of open reading frames within the cluster and the synthetic route to the gentamicin have been proposed based on nucleotide database homology searching. On the other hand, Kharel et al.10 identified the function of a specific gene within the gentamicin gene cluster corresponding to the biosynthesis of a precursor of DOS, 2-deoxy-scylloinosose, from glucose-6-phosphate. To date, the biosynthetic steps from DOS to other gentamicin-related aminoglycosides have not been established yet. For functional identification and manipulation of some putative or unknown genes involved in the gentamicin pathway, a more efficient detection method of DOS-containing aminoglycoside intermediates which were formed during the biosynthesis of the gentamicin complex should be prerequisite. The development of an analytical method for the reliable detection and identification of the DOS-containing aminoglycosides intermediates found in the culture of M. echinospora is essential (6) Thornsberry, C.; Barry, A. L.; Jones, R. N.; Baker, C. N.; Badal, R. E.; Packer, R. R. Antimicrob. Agents Chemother. 1980, 18, 338-345. (7) Kondo, S.; Hotta, K. J. Infect. Chemother. 1999, 5, 1-9. (8) Price, K. E. Antimicrob. Agents Chemother. 1986, 29, 543-548. (9) Kharel, M. K.; Subba, B.; Basnet, D. B.; Woo, J. S.; Lee, H. C.; Liou, K.; Sohng, J. K. Arch. Biochem. Biophys. 2004, 429, 204-214. (10) Kharel, M. K.; Basnet, D. B.; Lee, H. C.; Liou, K.; Moon, Y. H.; Kim, J. J.; Woo, J. S.; Sohng, J. K. Mol. Cells 2004, 18, 71-78. (11) Huang, F.; Haydock, S. F.; Mironenko, T.; Spiteller, D.; Li, Y.; Spencer, J. B. Org. Biomol. Chem. 2005, 3, 1410-1418.

4862

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

for further research. In the decade following the discovery of gentamicin,2 there have been a few reports concerned with the separation and isolation of the gentamicin-related aminoglycosides generated from the fermentation of M. echinospora.12,13 However, to the best of our knowledge, there has been no study on the high-performance liquid chromatography (HPLC) profiling of the DOS-containing aminoglycoside derivatives implicated in the gentamicin biosynthetic pathway from M. echinospora. Analytical chemists have focused on the monitoring of the residual gentamicin complex in a variety of biological fluids14-16 and environments (such as hospital wastewater),17 since the overdosage of gentamicin is associated with serious side effects (nervous and renal damage). Furthermore, the environmental pollution of gentamicin residues may also cause the appearance of antibiotic-resistant strains in pathogens.17 Many techniques have been developed to detect or measure the level of the gentamicin complex in different environments, and they are based on microbiological assays,18 immunoassays,19 gas chromatography (GC),20 or HPLC.13,16,17,21 Neither microbiological nor immunoassays are suitable for the profiling of the metabolites, because they lack specificity and are not able to distinguish between the individual gentamicin intermediates. The following chemical features of aminoglycosides obscure their chromatographic determination: (a) their very hydrophilic and nonvolatile nature not only restricts the routine usage of solvent extraction and reversedphase chromatographic columns for separation but also requires derivatization procedures for GC to be applicable; (b) they do not possess UV- or fluorescence-absorbing chromophores and therefore cannot be detected by common spectrometric techniques such as UV and fluorescence detection, making their quantification difficult. Furthermore, the presence of coextractable salts, sugars, and proteins from M. echinospora culture affects the HPLC separation of aminoglycosides and causes low recovery and poor uniformity between analyses. An efficient extraction procedure is required to remove the above interferences prior to detection using the spectrometric method. Solid-phase extraction (SPE) can be a potential tool to detach the interferences and concentrate the sample. On the basis of the basic (primary and secondary amine) properties present in the aminoglycoside structures, most reported SPE procedures for analyzing biological fluid samples rely on either cationic exchange at low pH or on C18 at high pH,15,22 although cationic exchange resin coupled with flash chromatog(12) Wagman, G. H.; Marquez, J. A.; Weinstein, M. J. J. Chromatogr. 1968, 34, 210-215. (13) Lee, B. K.; Condon, R. G.; Wagman, G. H.; Katz, E. Antimicrob. Agents Chemother. 1976, 9, 151-159. (14) Al-Amoud, A. I.; Clark, B. J.; Chrystyn, H. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2002, 769, 89-95. (15) Kaale, E.; Long, Y.; Fonge, H. A.; Govaerts, C.; Desmet, K.; Van Schepdael, A.; Hoogmartens, J. Electrophoresis 2005, 26, 640-647. (16) Lecaroz, C.; Campanero, M. A.; Gamazo, C.; Blanco-Prieto, M. J. J. Antimicrob. Chemother. 2006, 58, 557-563. (17) Lo ¨ffler, D.; Ternes, T. A. J. Chromatogr., A 2003, 1000, 583-588. (18) Arcelloni, C.; Vaiani, R.; Paroni, R. J. Chromatogr., A 1998, 812, 111-116. (19) Wei, T. Q.; Chu, V. P.; Craig, A. R.; Duffy, J. E.; Obzansky, D. M.; Kilgore, D.; Masulli, I. S.; Sanders, C. M.; Thompson, J. C. Clin. Chem. 1999, 45, 388-393. (20) Preu, M.; Guyot, D.; Petz, M. J. Chromatogr., A 1998, 818, 95-108. (21) Heller, D. N.; Peggins, J. O.; Nochetto, C. B.; Smith, M. L.; Chiesa, O. A.; Moulton, K. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005, 821, 22-30. (22) Heller, D. N.; Clark, S. B.; Righter, H. F. J. Mass Spectrom. 2000, 35, 3949.

raphy has also been used to isolate aminoglycosides from the fermentation broth.23 Aminoglycosides are not retained on conventional reversedphase C18 columns unless ion-pair agents are used. Although these agents are known to generate both a small shift of the retention time and memory effects, HPLC with fluorinated ion-pairing agents has been widely used for the analysis of aminoglycosides since the first demonstration of this technique in 1984.24 Inventive separation techniques using hydrophilic interaction chromatography25 or capillary electrophoresis15 have been viewed as promising alternatives to HPLC in drug-monitoring studies. However, HPLC separation with ion-pairing agents remains the most highly selective and sensitive technique for identifying aminoglycosides. Several HPLC methods have been proposed for the determination of aminoglycosides, but due to their low UV absorption and absence of fluorescence, analysis by HPLC is only feasible with pre- or postcolumn derivatization. Among pre- or postcolumn reagents for the detection of aminoglycosides, 1-fluoro-2,4-dinitrobenzene was most often used as the reagent permitting fluorescence or UV detection.26,27 Other reagents that could also have been used include phenylisocyanate28 for UV detection and 9-fluorenylmethyl chloroformate for fluorescence detection.29 Even if derivatization methods coupled with UV or fluorescence detectors are sensitive and specific enough to be employed for the determination of aminoglycosides, there exist a number of drawbacks to their use: the requirement of an additional HPLC system such as a reaction coil, the generation of degradation products and impurities which interfere with the detection of the aminoglycoside derivatives, and the uncertainty and incompleteness of the derivatizing reaction. To avoid the above problems in the sample derivatization process, refractive index detection and evaporative light scattering were used to detect nonchromophores.24,30 The ideal method of detecting aminoglycosides is mass spectrometry (MS), as it gives definite chemical evidence which can be diagnostic for molecular masses and their distinct fragmentation patterns. Therefore, HPLC coupled with MS becomes the most desirable analytical method for the determination of aminoglycosides with low detection limits (of the order of nanograms). A variety of ionization methods such as thermospray, fast atom bombardment, and electrospray ionization (ESI) have been reported for aminoglycoside analyses.22,31,32 ESI in combination with tandem MS (ESI-MS/MS) was shown to be more sensitive than other techniques for aminoglycoside analysis, and recent studies have demonstrated the usage of HPLC-ESI(23) Maehr, H.; Schaffner, C. P. J. Chromatogr. 1967, 30, 572-578. (24) Inchauspe´, G.; Samain, D. J. Chromatogr. 1984, 303, 277-282. (25) Oertel, R.; Neumeister, V.; Kirch, W. J. Chromatogr., A 2004, 1058, 197201. (26) Lu, J.; Cwik, M.; Kanyok, T. J. Chromatogr., B: Biomed. Sci. Appl. 1997, 695, 329-335. (27) Peng, G. W.; Gadalla, M. A.; Peng, A.; Smith, V.; Chiou, W. L. Clin. Chem. 1977, 23, 1838-1844. (28) Kim, B. H.; Lee, S. C.; Lee, H. J.; Ok, J. H. Biomed. Chromatogr. 2003, 17, 396-403. (29) Stead, D. A.; Richards, R. M. J. Chromatogr., B: Biomed. Appl. 1996, 675, 295-302. (30) Clarot, I.; Chaimbault, P.; Hasdenteufel, F.; Netter, P.; Nicolas, A. J. Chromatogr., A 2004, 1031, 281-287. (31) Getek, T. A.; Vestal, M. L.; Alexander, T. G. J. Chromatogr., A 1991, 554, 191-203. (32) Kotretsou, S. I.; Constantinou-Kokotou, V. Carbohydr. Res. 1998, 310, 121127.

MS/MS to determine the level of residual aminoglycoside antibiotics in biological fluids or animal tissues.21,22,33 The objectives of this study were to (a) develop an efficient SPE procedure for the sample preparation of the DOS-containing aminoglycoside intermediates generated from the gentamicin biosynthetic pathway in the M. echinospora culture, (b) establish an MS-compatible HPLC method capable of resolving the individual gentamicin intermediates, (c) set up an ESI-MS/MS system for detecting and identifying the different DOS-containing aminoglycoside components based on their m/z values and characteristic fragmentation spectra, and (d) evaluate the HPLC-ESIMS/MS method developed herein by comparing the DOScontaining aminoglycoside profile produced from wild-type M. echinospora ATCC (American Type Culture Collection) 158352 with that from the mutant strain KCTC 10506BP (Korean Collection for Type Cultures, Korean Patent No. 10-0519142-0000), which was derived by UV-induced mutagenesis. The method of creating DOScontaining aminoglycoside profiles (a list of aminoglycosides shown in Figure 1 except G418) from various gentamicinproducing strains developed in this study can be applied to other mutant or engineered strains which produce aminoglycosides, and the profiles generated can be useful in discerning the biosynthetic pathways of gentamicin or other aminoglycoside antibiotics in the future. EXPERIMENTAL SECTION Chemicals and Materials. Gentamicin sulfate (composed of mainly C1, C1a, C2, and C2a), paromomycin, and N-Z-Amine type A were obtained from Sigma, heptafluorobutyric acid (HFBA) was from Fluka Chemie GmbH, LC-grade acetonitrile, methanol, and water were from J. T. Baker, and culture medium yeast extract was from BD. Concentrated hydrochloric acid (HCl), ammonium hydroxide (NH4OH), glucose, soluble starch, and calcium carbonate were of reagent grade. Paromamine was synthesized as described previously34 using the acidic hydrolysis of paromomycin, which is commercially available. Because the aminoglycosides are known to be highly absorptive to polar surfaces such as glassware,21 the conical tubes, microtubes, and pipet tips used throughout the procedures were made of polypropylene (Becton Dickinson Labware or Axygen Scientific). The C18 SPE cartridge (AccuBOND II ODS-C18 3 cm3/200 mg) was supplied by Agilent Technologies, and the cationic exchanger SPE (OASIS MCX 3 cm3/60 mg) and vacuum manifold were obtained from Waters. A stock gentamicin solution (100 µg/mL, corrected for the purity of the free base) was prepared in water, and stored in an amber microtube at -20 °C, and was serially diluted to make standard solutions with concentrations of between 10 and 1000 ng/mL. These working solutions were stable for up to a month at 4 °C. Bacterial Strains and Culture Conditions. M. echinospora ATCC 15835 (wild-type strain for gentamicin C complex production) and the derivative KCTC 10506BP (UV-induced mutant of the wild-type strain) have been previously described.2,10 Both strains were grown for a week at 28 °C in 50 mL of N-Z amine medium (1% glucose, 2% soluble starch, 0.5% yeast extract, 0.5% N-Z-Amine, and 0.2% calcium carbonate) in 500 mL baffled (33) Cherlet, M.; Baere, S. D.; Backer, P. D. J. Mass Spectrom. 2000, 35, 13421350. (34) Haskell, T. H.; French, J. C.; Bartz, Q. R. J. Am. Chem. Soc. 1959, 81, 34803481.

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

4863

Erlenmeyer flasks on a rotary shaker. The whole culture broths were subjected to further study for the analysis of the aminoglycosidic intermediates. All experiments were done in triplicate. Aminoglycoside Extraction and Cleanup. The extraction and cleanup procedures for the DOS-containing aminoglycosides in the M. echinospora cultures included acid extraction by adjusting the pH value of the broths to pH 2 and dual SPE cleanup using both the reversed-phase C18 and the cationic exchange cartridges, prior to the HPLC-ESI-MS/MS analysis. Acid extraction was performed by a slightly modified version of the procedure described by Reiblein et al.,35 whereas the dual SPE cleanup procedure was developed to make clearer analytes from the crude extracts. This was done by adding a C18 SPE cleanup step to the minor modification of the method described by Heller et al.,22 which employs a weak cationic exchanger (WCX) SPE cartridge. Whole cultures were transferred to a polypropylene conical tube, and the pH was adjusted to ∼2 with concentrated HCl (approximately 120 µL), as verified by pH paper. The cultures were agitated for 1 h on a laboratory rocker prior to centrifugation at 20 000g for 5 min at 4 °C. The supernatants (approximately 45 mL) were passed through an AccuBOND C18 SPE column, preconditioned with 3 mL of methanol and then 3 mL of water, and the eluates were subsequently loaded onto an OASIS MCX column, previously conditioned with 3 mL of methanol followed by 3 mL of 10 mM HCl solution at pH 2. The column was washed with 6 mL of 10 mM HCl solution and then air-dried for about 30 s. The attached aminoglycosides were eluted two times with 0.75 mL of 5% methanolic NH4OH (concentrated NH4OH/methanol ) 5:95, v/v), evaporated to dryness at room temperature by vacuum centrifugation, and kept in a freezer until analysis. The aminoglycosides were reconstituted to 200 µL with water, and a portion of this solution was subjected to HPLC-ESI-MS/MS analysis. The recoveries of the gentamicin C complex standard (1 µg/ mL) spiked into a blank N-Z amine medium were determined to examine both the performance of the dual SPE protocol and the matrix effect of the media on the isolation of the aminoglycosides during the extraction and cleanup steps. They were extracted as described above and further analyzed by HPLC-ESI-MS/MS. From the comparison of the chromatographic peak areas obtained from both spiked blank samples with those from the standards, the percentage recovery was calculated. All of the experiments were done in triplicate. Aminoglycoside HPLC-ESI-MS/MS Analysis. Analyses of the DOS-containing aminoglycosides obtained from the cultures of the two M. echinospora strains were performed using a Waters/ Micromass Quattro micro/MS interface consisting of a Waters 2695 separation module connected directly to a Micromass Quattro micro MS. Separation was performed on a 50 mm × 2.1 mm XTerra MS C18 (3.5 µm, Waters) reversed-phase column. The analytes were eluted at a flow rate of 180 µL/min with a gradient of 10 mM (v/v) aqueous HFBA (A) and 50% (v/v) acetonitrile with the same HFBA concentration (B) at 10% B from 0 to 5 min, to 90% B for 30 min, maintained at 90% B for 2 min, and then to 10% B for another 8 min for column re-equilibration. The column effluent was directed to the ESI-MS, which was operated in the (35) Reiblein, W. J.; Watkins, P. D.; Wagman, G. H. Antimicrob. Agents Chemother. 1973, 4, 602-606.

4864

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

positive ion mode, without splitting. The instrument was tuned by the direct infusion of a solution of gentamicin C complex (10 µg/mL) in the ion source at 50 µL/min. The optimization parameters of the ESI-MS/MS system were based on the maximum generation, first of the protonated molecular ions (parents), and then of the corresponding fragment (product) ions. The following tune parameters were retained for the optimum ESIMS detection of the aminoglycosides: the capillary voltage and cone voltage were 4 kV and 30 V, respectively; the source and desolvation temperatures were 130 and 250 °C, respectively; the desolvation gas and cone gas flow rates were 500 and 50 L/h, respectively. The scan mass range was m/z 100-500 Da when run in the scan mode. The collision energy in the MS/MS mode, concurring with full argon-induced fragmentation of the parent molecules, was found to be 0.4 V (pressure reading 7.49 × 10-6 mbar). The molecular structures and fragmentation patterns for each DOS-containing aminoglycoside present in the M. echinospora cultures, which were obtained under these conditions, are shown in Figures 1, 2, and 4. Most of the observed patterns are in good agreement with those estimated from the ESI-MS fragmentation data of the gentamicin C complex published by others.17,22 To evaluate the sensitivity and facility of this analytical profiling technique for the detection of the aminoglycoside antibiotics compared to those of other drug-monitoring methods, the quantification of a gentamicin C complex standard spiked into a blank N-Z amine medium was conducted using MS/MS in multiple reactions monitoring (MRM) mode. This was done by choosing the two mass ions (listed in Figure 2) set to detect a transition of the parent ion to the product ion specific to the selected analytes (C1, 478 > 157; C2 isomers, 464 > 143; C1a, 450 > 129). RESULTS AND DISCUSSION Prior to establishing a complete method, the following steps were taken: the recoveries of the gentamicin C complex standards were checked at each step, dual SPE cleanup was used to obtain cleaner extracts more quickly, the HFBA ion-paring agent was used to resolve all of the gentamicin intermediates in a single chromatogram, and ESI-MS/MS was used not only to detect the aminoglycoside without derivatization procedures but to uncover the molecular information of each separated DOS-containing aminoglycoside intermediate. Extraction of Aminoglycoside Using Dual SPE Cleanup. Since the hydrophilic character of gentamicin makes its extraction by common organic solvents impossible, the adjustment of the pH of the culture broth coupled with an OASIS MCX SPE cleanup procedure for extracting aminoglycosides in blank N-Z amine medium was investigated. On the basis of the recovery of the gentamicin spiked into a blank medium from the OASIS MCX cartridge, modifying the pH to ∼2 with concentrated HCl solution led to the greatest amount of recovered gentamicin (>95%) among the wide range of pHs tested, ranging from 2 to 10 (Supporting Information Figure S-1). The similar usage of an extremely acidic extraction procedure has been previously reported.35 Reiblein et al. found that various aminoglycoside antibiotics bound to the mycelium of actinomycetes were able to be extracted at extreme pH values. The aminoglycosides have free NH2 functionalities with high pKa values which should be protonated at the strongly acidic pH (∼2), so they must be adsorbed, in these acidic conditions,

Figure 2. Main fragmentation pathway of gentamicin C components in ESI-MS/MS detection.

onto a cationic exchanger. OASIS MCX cationic exchange cartridges are also known to be stable at extreme pHs. Therefore, acidic extraction coupled with the OASIS MCX cleanup technique appears to be a suitable extraction procedure for the gentamicin analyses. However, it was observed that the extracts obtained from the single use of an OASIS MCX SPE cartridge brought out a very deep color, and the chromatogram of the extracts showed a large number of messy peaks that could seriously interfere with the aminoglycoside analysis. The presence of these endogenous interferences would also lead to the suppression of the ionization of the aminoglycosides in the ESI-MS/MS detection procedure. The resins charged in the OASIS SPE column are known to possess a slightly hydrophobic character, and it is likely that the eluents from the OASIS MCX cartridge contain a portion of the hydrophobic interferences. To avoid the noisy baseline in the chromatogram caused by these matrixes, several different brands of reversed-phase C18 cleanup cartridges (Agilent, Waters, Supelco, and Varian) were tested as a precleanup step before using the OASIS MCX cartridge. All of the extracts passing through the additional C18 SPE columns showed no apparent difference in the recovery rate (>90%) of gentamicin obtained from the single use of the OASIS MCX cartridge, although the extracts from the AccuBOND C18 SPE column (Agilent) was visually much cleaner than those from the other C18 SPE columns. This is the first demonstration on the extraction of aminoglycosides using the dual

SPE cleanup, and it can be concluded that the combination of acid extraction and the dual SPE cleanup procedure employing both AccuBOND C18 and OASIS MCX cartridges can be used for the clean and rapid extraction of gentamicin intermediates in M. echinospora culture. These results also suggest that the above extraction method may be relevant for the analysis of other important aminoglycoside antibiotics utilizing a modified HPLC gradient system. Separation of Aminoglycoside Using HFBA and ESI-MS/ MS Detection. Since fluorinated ion-pairing reagents with moderate volatility were successfully applied to the reversed-phase chromatographic separation of aminoglycoside antibiotics in 1984,24 various reagents such as trifluoroacetic acid (TFA), pentafluoropropionic acid (PFPA), and HFBA have been investigated.22,36 The chromatographic retention behavior of the gentamicin standard, which was spiked into a blank N-Z amine medium followed by the established cleanup procedure, using the above three ion-pairing reagents added to the mobile phases (water-acetonitrile mixture) was investigated to determine the best-suited ion-pairing reagent for the following research purposes: there should be a reasonable difference in the retention times in order to resolve the diverse DOS-containing aminoglycoside peaks, the suppression effects of the ion-pairing reagent (36) McLaughlin, L. G.; Henion, J. D.; Kijak, P. J. Biol. Mass Spectrom. 1994, 23, 417-429.

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

4865

Figure 3. Representative MRM mass chromatograms and spectra obtained for spiked gentamicin C components: (a-c) MRM mass chromatograms of gentamicin C1, C2 isomers, and C1a, respectively; (d-f) fragmentation ion mass spectra of gentamicin C1, C2 isomers, and C1a, respectively.

on the ionization of the analytes should be minimized, and the matrix effects (shift of the retention time and nonsymmetric peak) and memory effects (reappearance of aminoglycoside in subsequent analyses causing false positive results) must be avoided. These adverse effects were observed by other researchers in the determination of aminoglycoside antibiotics in milk and wastewater samples.17,22 From the three reagents that were previously investigated, HFBA was accepted for the ESI-MS/MS detection of nonderivatized gentamicin, because it allowed for the strong retention of the gentamicin peaks on a reversed-phase column and the usage of this reagent minimized the suppression effects 4866

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

on ionization, as compared to the other ion-pairing reagents (data not shown). These results are in accordance with those of a previous study reported by Heller et al.,22 which demonstrated that the length of the carbon chain in the fluorinated ion-pairing reagent was positively correlated with the chromatographic retention of aminoglycoside antibiotics on a reversed-phased column. This group also observed that the suppression of aminoglycoside ionization caused by the addition of HFBA to the mobile phase was less serious than that caused by TFA. To lessen the matrix and memory effects known to commonly occur in aminoglycoside analysis, the level of HFBA added to the mobile

Figure 4. Proposed fragmentation pathways of the DOS-containing aminoglycoside intermediates involved in the gentamicin biosynthesis in M. echinospora strains.

phases was varied. Excessive levels of HFBA (>30 mM) in the mobile phase were found to cause very severe matrix effects, the distortion of the peak shapes, and unstable retention times (Supporting Information Figure S-2). The employment of a lower concentration of HFBA ( m/z product ion) for the main components of the gentamicin C complex in the positive mode of LC-ESI-MS/MS were monitored as a series of chromatograms: C1, 478 > 157 (Figure 3a); C2 isomers, 464 > 143 (Figure 3b); C1a, 450 > 129 (Figure 3c). The order of elution of the four components was C1a, C2 isomers (C2 and C2a), and C1, with the peaks being separated by ∼1 min each. The protonated molecular (parent) ions of the four gentamicin C series [M + H]+ were scanned at m/z 450 (C1a), 464 (C2 and C2a), and 478 Da (C1), respectively. After fragmentation by MS/ MS, these ions generated tandem mass spectra with diagnostic fragmentation (product) ions (Figure 3d-f). The gentamicin isomers, C2 and C2a, can hardly be separated on a reversed-phase column, and further differentiation by MS and MS/MS was also impracticable, since they have the same molecular mass and yield the same MS/MS fragmentation patterns. There were three typical product ions observed for the gentamicin C complex standard at m/z 322, 205, and 160 Da, which correspond to (a) the loss of the purpurosamine ring (the left aminosugar unit in the gentamicin Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

4867

Table 1. Aminoglycoside Profiles Found in Wild-Type M. echinospora ATCC 15835 and the UV-Induced Mutant KCTC 10506BPa protonated ion

M. echinospora strain

DOS-containing aminoglycosideb

parent (m/z)

production (m/z)

ATCC 15835 (µg/mL)c

KCTC 10506BP (µg/mL)c

paromamine gentamicin A2 gentamicin B gentamicin X2 gentamicin A JI-20A JI-20B gentamicin C1a gentamicin C2 isomers gentamicin C1

324 456 483 483 469 482 496 450 464 478

307, 289, 163 295, 205, 133 322, 205, 162, 160 322, 205, 162, 160 308, 205, 146 322, 205, 161, 160 322, 205, 175, 160 322, 205, 163, 160, 129 322, 205, 163, 160, 143 322, 205, 163, 160, 157

4.8 ( 0.2d 7.3 ( 0.3e

4.1 ( 0.3d 3.4 ( 0.3e 7.5 ( 0.4e

1.2 ( 0.2e 0.7 ( 0.1e 1.0 ( 0.1e 3.1 ( 0.2f 5.0 ( 0.1f 8.7 ( 0.3f

0.5 ( 0.1e 0.4 ( 0.1e 6.4 ( 0.1f 1.9 ( 0.2f 1.7 ( 0.2f

a Both strains were cultivated in N-Z amine media at 28 °C for a week. b DOS-containing aminoglycosides identified are listed in the order of their elution from the HPLC column. c All the values are represented as µg/mL culture ( standard deviation (n ) 3). d Values are semiquantified based on the relative intensity of a specific product ion at m/z 163 ([intensity of relevant mass peak]/[intensity of the reference mass peak at m/z 163 produced from the standard gentamicin Cs]) compared to that of standard gentamicin Cs. e Values are semiquantified based on the relative intensity of a specific product ion at m/z 205 ([intensity of relevant mass peak]/[intensity of the reference mass peak at m/z 205 produced from the standard gentamicin Cs]) compared to that of standard gentamicin Cs. f Values are quantified using MS/MS in MRM mode (C1, 478 > 157; C2 isomers, 464 > 143; C1a, 450 > 129).

C series), (b) the successive cleavage of the garosamine moiety (the right aminosugar unit) followed by dehydration and cyclization with the DOS moiety (the middle aminosugar unit), and (c) the associated loss of two sugar (DOS and purpurosamine moiety) units, respectively (Figure 2). However, the characteristic product ions for the gentamicin C1a, C2 isomers, and C1 component were produced at m/z 129, 143, and 157 Da, respectively. These product ions account for the remaining purpurosamine moieties on the individual gentamicin standards. The appearance of peaks in a selective ion chromatogram at m/z 160, 205, or 322 Da must be considered to be diagnostic of the presence of the gentamicin C series component. The specific product ions can be useful in determining the identity of each compound and can also be chosen as distinctive fragmentation ions for quantifying each gentamicin standard using MS/MS in MRM mode. Calibration curves were established separately for each gentamicin component, and there was a linear correlation (gentamicin C1a, r2 > 0.98; C2a/C2, r2 > 0.93; C1, r2 > 0.99) between the amount of gentamicin spiked (20-1000 ng/mL) and the MS/MS detection response. The detection limit obtained by this HPLCESI-MS/MS analysis at a signal-to-noise ratio of 7:1 was found to be approximately 5 ng for each component of the gentamicin C complex, corresponding to about 1.2 ng/mL in the culture broths. In comparison to other studies based on HPLC-ESI-MS/MS analysis, this detection limit is more sensitive than the 25 ng/g limit indicated for gentamicin monitoring in animal tissues31 and is similar to that of other analyses21 that were recently applied to detect gentamicin residues in biological fluids such as milk, plasma, and urine. A greater increase in sensitivity (as little as 0.2 ng/mL) was reported17 in the case of a sensitive method using SPE cleanup for the determination of gentamicin residues in hospital wastewater. The accuracy of the method developed herein was evaluated by recovery experiments. When compared to the gentamicin standard injected directly into the LC-ESI-MS/MS, mean recoveries ranging from 91% to 94% for each component of the gentamicin standard spiked into the blank N-Z amine medium were obtained when the combination of acid extraction and dual 4868 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

SPE was employed prior to the analysis with HPLC-ESI-MS/MS detection. Analytical Profiling of the Gentamicin Intermediates Produced in the M. echinospora Strains. The proposed method was applied to the detection of the gentamicin intermediates produced in the two M. echinospora strains. Supporting Information Figure S-3a shows a typical total ion chromatogram of the M. echinospora ATCC 15835 cultures obtained using the established procedures, and Supporting Information Figure S-4a shows a total ion chromatogram of the M. echinospora KCTC 10506BP cultures obtained by the same analytical procedures. In the chromatogram obtained from the wild-type ATCC 15835 culture extract, diverse peaks were detected on the basis of the protonated molecular ions of intermediates whose formation is predictable in the proposed gentamicin biosynthetic pathway.5 Further analysis by MS/MS was then performed on the suspected peaks (Supporting Information Figure S-3b-i). As only the gentamicin C complex was commercially available as an authentic standard, the fragmentation patterns of the various gentamicin intermediates (except the C series) were inferred from those of the gentamicin C complex obtained by the ESI-MS/MS analysis (Figure 4). The identity of each DOS-containing aminoglycoside was verified through the comparison of the product ions produced in the mass spectra shown in Supporting Information Figures S-3 and S-4 with the predicted fragmentation patterns shown in Figure 4. A total of eight different gentamicin intermediates were verified in the ATCC 15835 strain, and they were eluted in the following order: paromamine, gentamicin A2, gentamicin X2, gentamicin A, JI-20B, gentamicin C1a, gentamicin C2 isomers, and gentamicin C1 (see the Supporting Information Figures S-3 and S-4). Their fragment patterns analyzed in this study by ESI-MS/MS were in good agreement with the proposed patterns (Figures 2 and 4). In particular, they contain the typical product ions at m/z 205, corresponding to the cleavage of the garosamine moiety followed by dehydration and cyclization with the DOS moiety. This seems to be a conserved product ion for all DOS-containing aminoglycoside intermediates implicated in the gentamicin biosynthetic

pathway, because there is no substitution found on the DOS moiety in the aminoglycoside structure. Therefore, the selection of the ion chromatogram at m/z 205 can provide useful information about the presence of DOS-containing aminoglycosides in the M. echinospora culture. In addition, all other intermediates besides the gentamicin A series (A2 and A) yield the common product ions at m/z 160 and 322 together with m/z 205. Supporting Information Figure S-4b-i shows the mass spectra of the DOScontaining intermediates produced in the UV-induced mutant KCTC 10506BP. The DOS-containing aminoglycoside profile of this M. echinospora mutant strain was different in its composition and quantity from that of the wild-type strain. The main intermediates detected in the mutant culture are found to be gentamicin B and C1a on the basis of the peak height of the product ion at m/z 205, whereas those found in the wild type are of the gentamicin C1 and C2 isomers; an intermediate JI-20A appears, but the other intermediate, JI-20B, observed in the wild type is not detectable in the mutant, KCTC 10506BP. The lack of authentic standards for various aminoglycosides may cast doubt on our deduction that other DOS-containing aminoglycosides follow the same fragmentation pathway as that obtained from the gentamicin C complex standard. To confirm the above inference regarding the fragment patterns, a limited study on the synthesis of paromamine34 was carried out. It was determined that both the retention time in the selected ion chromatogram scanned at m/z 324 and the mass spectra of the authentic paromamine were consistent with those of the paromamine detected in both M. echinospora cultures (see the Supporting Information Figure S-5). This result demonstrates that the method proposed in this study for the determination of the gentamicin intermediates in the bacterial strain by HPLC-ESI-MS/ MS is reliable for the profiling of the DOS-containing aminoglycoside metabolites present. There are 11 kinds of DOS-containing aminoglycoside intermediates which are known to exist in the gentamicin biosynthetic pathway, and the putative roles of the biosynthetic genes were proposed in a previous study.5 The presence of all of the DOS-related components except G418 (shown in Figure 1) was noticed in the two different M. echinospora strains (shown in Table 1). As proposed by Unwin et al.,5 the biosynthetic pathway for the production of the gentamicin C complex in M. echinospora is thought to be branched into two routes, as follows: the first serial path from gentamicin X2 to JI-20A, gentamicin C1a and C2a, and the other serial path from gentamicin X2 to JI-20B, gentamicin C2 and C1 (Figure 1). They also suggested that these routes could be separated from each other as a function of a particular gene, which encodes the methyltransferase acting at the 6′ position of the purpurosamine moiety on the aminoglycoside. A qualitative comparison (based on the relative peak height of a specific product ion at m/z 205 in the mass spectrum of the analyte) of the abundance of the DOS-containing aminoglycosides between the different strains of M. echinospora indicates that the KCTC 10506BP strain seems to be mutated on the above-mentioned methyltransferase gene, in that the main component gentamicin C1a obtained in the mutant culture belongs among the gentamicin intermediates involved in the first path, and moreover, JI-20A

appeared while JI-20B disappeared (shown in Table 1). In particular, gentamicin B, which may be a shunt metabolite derived from JI-20A in gentamicin biosynthesis and is known to be a valuable aminoglycoside for the synthesis of isepamicin, is generated as a major product in the mutant strain. Detailed studies on the structural elucidation of these aminoglycosides using nuclear magnetic resonance spectroscopy are needed. Moreover, finding the specific genes, which are mutated by UV mutagenesis, would be helpful for elucidating their functions in the gentamicin biosynthetic pathway in the M. echinospora strains. Recently, the biosynthesis of aminoglycoside antibiotics has begun to be explored intensively.37

(37) Llewellyn, N. M.; Spencer, J. B. Nat. Prod. Rep. 2006, 23, 864-874.

AC070028U

CONCLUSIONS This study describes a rapid, sensitive, and highly selective technique by which the DOS-containing aminoglycoside intermediates involved in the gentamicin biosynthetic pathway in M. echinospora can be profiled. Acid extraction coupled with dual SPE cleanup using AccuBOND C18 and OASIS MCX cartridges was found to provide adequate cleanup of the extract. Our HPLC-ESIMS/MS technique can provide the detailed mass-related information that is now needed for the differentiation of the gentamicin intermediates. The detection limit of the gentamicin standard obtained in this HPLC-ESI-MS/MS analysis was found to be about 5 ng for each component of the gentamicin C complex, and the mean recoveries for each component of the gentamicin standard spiked into blank N-Z amine medium were above 91%. This indicates that the proposed method is sensitive and reproducible enough to profile the gentamicin intermediates which are present in low concentrations in M. echinospora. This method was successfully applied to the analytical profiling of the gentamicin intermediates produced in two M. echinospora strains (wild type and UV-induced mutant), where seven intermediates besides the gentamicin C complex were detected, including paromamine, gentamicin A2, B, X2, A, JI-20A, and JI-20B. HPLC-ESI-MS/MS analysis of the above components from the culture extracts resulted in the expected fragmentation patterns. This analytical approach to DOS-containing aminoglycoside profiling can be coupled with genetic strategies to provide a more detailed understanding of the gentamicin biosynthetic pathway. ACKNOWLEDGMENT This work was supported by MOCIE (Ministry of Commerce, Industry and Energy), the SRC program of the Korea Science and Engineering Foundation (KOSEF) through the Center for Intelligent NanoBio Materials at Ewha Womans University (Grant R112005-008-00000-0), and the Seoul R&BD program (10816). The authors thank the Ministry of Education for the Brain Korea 21 fellowship. SUPPORTING INFORMATION AVAILABLE Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org Received for review January 5, 2007. Accepted April 18, 2007.

Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

4869