|Home | About | Journals | Submit | Contact Us | Français|
The lincomycin biosynthetic gene lmbX was deleted in Streptomyces lincolnensis ATCC 25466, and deletion of this gene led to abolition of lincomycin production. The results of complementation experiments proved the blockage in the biosynthesis of lincomycin precursor 4-propyl-l-proline. Feeding this mutant strain with precursor derivatives resulted in production of 4′-butyl-4′-depropyllincomycin and 4′-pentyl-4′-depropyllincomycin in high titers and without lincomycin contamination. Moreover, 4′-pentyl-4′-depropyllincomycin was found to be more active than lincomycin against clinical Staphylococcus isolates with genes determining low-level lincosamide resistance.
Lincosamides form a small yet clinically important group of antibiotics. One of the naturally occurring members of this group, lincomycin A (LIN), is active against many Gram-positive bacteria, such as staphylococci and streptococci. Its semisynthetic derivative clindamycin (CLI) is prescribed for the treatment of some infections caused by anaerobic bacteria and is also applied against the causative agent of malaria, Plasmodium falciparum (9, 22).
Biosynthesis of lincomycin proceeds via two separate branches from tyrosine and d-glucose to the aglycone 4-propyl-l-proline (PPL) and methylthiolincosamide (MTL), respectively. Condensation of these two precursors via an amide bond by a multimeric synthetase yields N-demethyllincomycin (NDL), which is subsequently methylated to form LIN (3) (Fig. (Fig.1A).1A). Various LIN derivatives have been prepared chemically, and in particular, the 4′-alkyl-4′-depropyllincomycin set has been determined to be more active and have a broader antimicrobial spectrum than LIN has. Of this set, 4′-butyl, pentyl, and hexyl analogs have been shown to be particularly effective (12, 13). Moreover, demethylated and chlorinated 4′-alkyl-1′-demethyl-4′-depropylclindamycins had higher activities against Plasmodium spp. than clindamycin did (14). These more potent 4′-alkyl derivatives of lincomycin can be prepared by a multistep and costly chemical synthesis (13). Alternatively, feeding or genetic modifications of the natural biosynthetic pathway could be used. For example, the addition of PPL derivatives with extended alkyl residues to fermentation broths of a producer strain in a process termed precursor-directed biosynthesis has been described previously (25). This resulted in a mixture of both LIN and its more biologically active derivative being produced, which is not desirable because of the need to separate and purify the product of interest.
In this study we present a new mutasynthetic approach to the preparation of the two known 4′-alkyl-4′-depropyllincomycins (Fig. (Fig.1B)1B) by feeding a mutant strain defective in PPL biosynthesis with PPL derivatives as a practical alternative to total chemical synthesis. We tested the biological activity of these LIN derivatives on a collection of Staphylococcus strains with defined resistance profiles that have been described previously (15-17).
The early reactions involved in biosynthesis of proline derivatives are the same for lincomycin and several pyrrolobenzodiazepine antibiotics (5, 10, 11). Presumably, the genes shared by lincomycin and benzodiazepine gene clusters could code for enzymes of PPL biosynthesis. To confirm the hypothesis and obtain a mutant defective in PPL production, we deleted one of the shared genes, lmbX, in the LIN-producing Streptomyces lincolnensis ATCC 25466 type strain. The inactivation of lmbX was achieved by the Redirect targeting system (4) using the LK6 cosmid, which bears the whole LIN biosynthetic gene cluster (7), inactivation primers Xf (5′-CGCGCCCATCCTGCACAGCGCACCGGAGGAAGCATGATCATTCCGGGGATCCGTCGACC-3′) and Xr (5′-GAGAAAAGAGCCGCTGACGCAAGGGGCCCTCGGCGACTATGTAGGCTGGAGCTGCTTC-3′) (nucleotide extensions with sequence identity to regions upstream and downstream of lmbX, respectively, are underlined) and checking primers chXf (5′-CCGGCATCAACGACT-3′) and chXr (5′-CCAGATGGAACGAATTCA-3′). The lmbX deletion strain is hereafter referred to as the ΔlmbX strain. For detection of LIN in fermentation broth, we cultivated the type and mutant strains and performed ultra-performance liquid chromatography (UPLC) analysis of the respective broths by the method of Olsovska et al. (18). We revealed that LIN production was under the limit of detection in the ΔlmbX strain (Fig. (Fig.2).2). The minor peak that eluted at 2.95 min had a UV spectrum different from that of LIN.
Next, we tested the ability of the LIN precursors PPL and MTL to complement the mutation in feeding experiments, in which the ΔlmbX strain was cultivated on GYM agar plugs (glucose [4 g/liter], yeast extract [4 g/liter], malt extract [10 g/liter], CaCO3 [2 g/liter], agar [12 g/liter] [pH 7.2]) with PPL or MTL (6) added at a final concentration of 200 mg/liter for 10 days at 28°C. Agar plugs were subsequently placed on B1 agar (beef extract [10 g/liter], peptone [10 g/liter], NaCl [5 g/liter], agar [20 g/liter] [pH 7.2]) overlaid with the indicator strain Kocuria rhizophila CCM 552. Growth inhibition zones, indicating the production of antimicrobial compound, were detected after 24 h of incubation at 30°C. The production of the compound was detected only in the case of feeding with the precursor PPL. UPLC analysis of the fermentation broth confirmed the antimicrobial compound to be LIN (based on a comparison of retention times and UV spectra with those of the LIN standard) (Fig. (Fig.2).2). The restoration of LIN production after PPL addition proved the participation of LmbX in the PPL biosynthetic branch. Nevertheless, the precise enzymatic role of LmbX remains unknown and is the aim of further studies.
In order to feed the ΔlmbX strain with PPL derivatives, we prepared 4-butyl-l-proline and 4-pentyl-l-proline (BUPL and PEPL, respectively) based on the aldol condensation of protected l-pyroglutamic acid with corresponding aldehyde. The resulting aldols were dehydrated using MsCl-Et3N to yield 4-alkylidenepyroglutamates, which gave cis-4-substituted pyroglutamates after hydrogenation of the double bond. Inversion of the configuration at C-4 by 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) followed by a two-step reduction of the resulting amides led, after deprotection, to BUPL and PEPL. For details of synthetic procedures and analytical data, see supplemental material.
In mutasynthetic experiments, BUPL and PEPL were added at the start of fermentation in AVM medium (18) to a final concentration of 100 mg/liter. We found that addition of either BUPL or PEPL to cultivation broth of the ΔlmbX mutant defective in PPL production led to the formation of new compounds (Fig. (Fig.3).3). The compounds were isolated by UPLC or high-performance liquid chromatography (HPLC) in order to perform mass spectrometry and nuclear magnetic resonance experiments (for details, see supplemental material and Fig. S2 and S3 in the supplemental material). The results of these experiments confirmed synthesis of either 4′-butyl-4′-depropyllincomycin (BULIN) or 4′-pentyl-4′-depropyllincomycin (PELIN) and, simultaneously, indicated the broad substrate flexibility of the LmbC enzyme, which is responsible for the recognition and activation of PPL, prior to the synthesis of NDL (S. Kadlčík, unpublished data). This phenomenon has been observed for various enzymes involved in the biosyntheses of secondary metabolites (8, 19, 21) and enables a great structural diversity of products.
The antimicrobial activities of LIN and CLI derivatives with extended alkyl chains against a collection of Staphylococcus aureus, Enterococcus faecalis, Escherichia coli, and Salmonella schottmuelleri strains have been determined previously (12). We used a set of 13 clinical macrolide- and/or lincosamide-resistant Staphylococcus isolates (15-17) for a series of experiments evaluating the antimicrobial activity of BULIN and PELIN, depending on the presence of certain resistance genes. S. aureus ATCC 29213 and S. aureus CIP 107907 were used as a lincosamide-sensitive indicator strain and a strain carrying the vgaA gene, respectively. The agar diffusion method (24) showed PELIN to be more biologically active than BULIN; therefore, further experiments were conducted only with PELIN. We determined the MIC of PELIN by the agar dilution method in microtitration plates. A 2-μl inoculum of a 0.5 McFarland suspension was spotted onto 100-μl Mueller-Hinton agar plugs with the following antibiotics and concentrations: PELIN, 0.125 to 128 μg/ml; LIN, 0.125 to 256 μg/ml; and CLI, 0.125 to 16 μg/ml. Microtitration plates were incubated at 37°C for 24 h. As summarized in Table Table1,1, strains expressing the constitutive ermC gene, coding for rRNA adenine N-6-methyltransferase and determining macrolide-lincosamide-streptogramin B (MLSB) resistance (23), were resistant to PELIN to the same level as CLI and LIN. Furthermore, high susceptibility of strains with an inducible ermC gene showed that PELIN, just as LIN and CLI, did not induce resistance. The recently described vgaALC gene, coding for an ABC transporter, confers resistance not only to streptogramins, as its evolutional variant vgaA does (2), but also to lincosamides (16), including PELIN (indicated by the LC subscript). Strains carrying the msrA gene, coding for another ABC transporter and conferring resistance only to macrolide antibiotics (20) were, as expected, susceptible to all lincosamides tested. Altogether, PELIN, LIN, and CLI had the similar antimicrobial activity against strains with ermC, vgaALC, and msrA genes (Table (Table1).1). In contrast, PELIN was found to have almost the same activity as CLI against isolates with a resistance given by lnuA gene, which codes for a lincosamide nucleotidyltransferase (1). Strains carrying this gene, as well as a strain with both lnuA and msrA, showed higher susceptibility to CLI and PELIN than to LIN. On the other hand, the combination of lnuA and msrA genes with vgaALC mimicked the MLS resistance phenotype (Table (Table11).
In conclusion, we have reported the first example of highly effective mutasynthesis of LIN derivatives. Our more feasible way of synthesis could be used for developing chlorinated alternatives, which have a greater antimalarial activity than currently used clindamycin (14). Nevertheless, we should take into consideration the need for chemical synthesis of the precursors, which is less but still costly and time-consuming. Therefore, preparation of a strain producing hybrid antibiotics on lincosamide without or with minimal chemical modifications is more favorable and is the aim of our future studies.
We thank Gabriela Novotná and Terry Evans for comments and critical reading of the manuscript. The Redirect targeting system was obtained from Plant Bioscience Limited (Norwich, United Kingdom).
This work was supported by the Grant Agency of the Academy of Sciences of the Czech Republic (grant IAA500200810), the Ministry of Education, Youth and Sports of the Czech Republic (grant 2B08064), and Institutional research concept (AV 0Z 50200510).
Published ahead of print on 16 November 2009.
†Supplemental material for this article may be found at http://aac.asm.org/.