Heterologous Production of Albicidin: a Promising Approach to Overproducing and Characterizing This Potent Inhibitor of DNA Gyrase

Albicidin is produced by Xanthomonas albilineans which is a slow-growing bacterium and the causal agent of sugarcane leaf scald (13). Albicidin is involved in the pathogenicity of X. albilineans and inhibits the replication of chloroplastic DNA (6, 7). Albicidin also inhibits DNA replication in Escherichia coli at nanomolar concentrations (4), whereas mammalian cells are unaffected at 8 μg/ml (3). A recent study showed that albicidin targets DNA gyrase with features of inhibition that differ from those of other known antibiotics (11). Albicidin is synthesized by a unique hybrid polyketide synthase-nonribosomal peptide synthetase (PKS-NRPS) pathway that does not resemble any other pathway described to date. Three genomic regions (XALB1, XALB2, and XALB3) involved in albicidin biosynthesis were cloned and sequenced (12, 14, 15, 17). XALB1 contains the majority of albicidin biosynthetic genes in a cluster of 20 open reading frames (ORFs) (albI to albXX). ORFs albI, albIV, and albIX encode three large PKSs and NRPSs, whereas the other ORFs are resistance, modifying, and regulatory genes. XALB2 and XALB3 each contain a single biosynthetic gene, albXXI and albXXII, whose products are, respectively, a phosphopantetheinyl transferase and the heat shock protein HtpG. The antibiotic activity of albicidin against a wide range of gram-positive and gram-negative pathogenic bacteria (Enterobacter aerogenes, E. coli, Haemophilus influenzae, Klebsiella pneumoniae, Shigella sonnei, and Staphylococcus aureus) is of interest for the development of new antibacterial drugs (4, 5). Low yields of albicidin production in slow-growing X. albilineans have slowed studies of its chemical structure and potential therapeutic applications. A heterologous system for albicidin overproduction is therefore highly desirable (i) to increase the levels of expression of the biosynthetic enzymes, (ii) to obtain large quantities of albicidin, and (iii) to characterize albicidin through nuclear magnetic resonance and mass spectrometry analyses. We report here the construction of a two-plasmid expression system harboring the complete albicidin biosynthetic gene set; its transfer into a fast-growing heterologous host, X. axonopodis pv. vesicatoria; and the subsequent production of albicidin.

Two wide-host-range plasmids derived from IncW vector pUFR043 (10) and IncP vector pLAFR3 (16) were used for the transfer of the complete albicidin biosynthetic machinery. The previously described IncW plasmid pALB571 (12, 14) was used to transfer albI to albIX (the insert corresponds to nucleotides 19001 to 55839 of accession no. AJ586576 ). Plasmid pLAFRK7 was constructed by ligating together into pLAFR3 the three following genomic DNA fragments: (i) nucleotides 5896 to 19216 of accession no. AJ586576 (harboring albX to albXX), (ii) nucleotides 1 to 2986 of accession no. AJ586577 (harboring albXXI), and (iii) nucleotides 5510 to 8119 of accession no. AM039979 (harboring albXXII), respectively. Both plasmids were transferred by triparental mating into X. axonopodis pv. vesicatoria strain Xcv 91-118 (1), and bioassays for albicidin production were performed with exconjugants. Triparental matings were performed as described previously (12) except that exconjugants were selected on SPA medium (2% sucrose, 0.5% peptone, 1.5% agar) supplemented with kanamycin at 50 μg/ml (for the transfer of pALB571) and tetracycline at 12 μg/ml (for the transfer of pLAFRK7). Bioassays for albicidin production were performed as described previously (12) except that (i) exconjugants were spotted onto SPA medium supplemented with gentamicin at 3 μg/ml, tetracycline at 12 μg/ml, and morpholinepropanesulfonic acid (MOPS) buffer and (ii) the plates were overlaid with a suspension of E. coli DH5α harboring pUFR043 and pLAFR3 empty plasmids that confer resistance to gentamicin and tetracycline, respectively. All exconjugants harboring plasmids pALB571 and pLAFRK7 produced an antibiotic that inhibited the growth of E. coli DH5α. No antibiotic production was observed with control exconjugants harboring either pALB571 or pLAFRK7 and one empty vector (and therefore missing some albicidin biosynthetic genes). These data demonstrated that transfer of the complete albicidin biosynthetic machinery into X. axonopodis pv. vesicatoria led to the heterologous production of an antibiotic that was named albicidin^ves. The exconjugant yielding the largest DH5α growth inhibition zone was designated Xves-alb and was used for further analyses.

Cross-resistance between albicidin and albicidin^ves was characterized by bioassays with agar cultures of Xves-alb using E. coli strains expressing a wide range of albicidin resistance determinants. Strain DH5αAlb^r (a spontaneous albicidin-resistant DH5α derivative [12]) and strain DH5αAlbD (expressing the albicidin-detoxifying gene albD [19]) were completely resistant to albicidin^ves. Strains harboring albXIV (an albicidin efflux pump gene conferring albicidin resistance in E. coli [8]), albXIX (an albicidin qnr gene conferring resistance in E. coli [11]), or sbmC (a gyrase inhibitor resistance gene [2]) were also resistant to albicidin^ves but to a lesser extent. This resistance pattern was identical to the one observed with X. albilineans cultures or with albicidin semipurified from X. albilineans. In addition, several E. coli DH5α clones that spontaneously grew within the growth inhibition zone of Xves-alb were resistant to albicidin, confirming the cross-resistance between the two antibiotics. Albicidin and albicidin^ves appeared, therefore, to display the same biological activities.

The growth of Xves-alb and the production of albicidin^ves were monitored in several rich liquid media supplemented with gentamicin at 3 μg/ml and tetracycline at 12 μg/ml: NYG (0.3% yeast extract, 0.5% peptone, 2% glycerol), TY (0.3% yeast extract, 0.5% tryptone, 5 mM CaCl₂), and YM (0.3% yeast extract, 0.3% malt extract, 1% peptone). Maximum albicidin^ves activity was obtained in early stationary phase from shaken cultures grown at 28°C, and the best results were obtained in NYG with an albicidin^ves titer of 970 μg/liter. Albicidin was quantified with a bioassay, using the following formula developed by Zhang et al. (20): free albicidin (ng/ml) = 4.576 e^{(0.315 × inhibition zone diameter in mm)}. This titer was six times higher than the titer of albicidin produced by X. albilineans in the optimized SP8 medium (0.5% sucrose, 0.23% peptone, 0.1% yeast extract, 3 mM K₂HPO₄, 1 mM MgSO₄; pH 7). In order to minimize the presence of catabolites in the supernatant and to facilitate the subsequent purification of albicidin^ves, we investigated growth of Xves-alb and the production of albicidin^ves in minimal media without gentamicin and tetracycline. We tested the XVM2 medium [20 mM NaCl, 10 mM (NH₄)₂SO₄, 5 mM MgSO₄, 1 mM CaCl₂, 0.16 mM KH₂PO₄, 0.32 mM K₂HPO₄, 0.01 mM FeSO₄, 0.03% Casamino Acids; pH 6.7] (18) supplemented with a single carbon source (mannitol, glucose, fructose, sucrose, or glycerol). The best albicidin^ves titer (200 μg/liter) was obtained in XVM2 supplemented with 2% glycerol. A decrease of albicidin production in minimum media was also observed in X. albilineans (4).

The chromatographic behaviors of albicidin and albicidin^ves were then compared by thin-layer chromatography (TLC). Albicidin^ves was analyzed from cultures of Xves-alb grown to early stationary phase (48 h) in 5 liters of XVM2 medium supplemented with 2% glycerol and without any antibiotic, and albicidin was analyzed from cultures of X. albilineans strain Xa23R1 to stationary phase (96 h) in 5 liters of SP8 medium. Supernatants from bacterial cultures were loaded onto Amberlite XAD-7 (Sigma) resin column, washed, and eluted with methanol. The fractions exhibiting an antibiotic activity at a 1,000-fold dilution were pooled and vacuum dried to a final volume of 0.5 ml. Then, 2-μl samples were separated by TLC chromatography with silica gel 60F₂₅₄ TLC aluminum sheets (20 by 20 cm; Merck, Darmstadt, Germany) using methanol as the mobile phase. Subsequently, strips of the TLC plate were placed for 2 h at room temperature on the surface of a top agar layer containing an E. coli indicator strain. Inhibition zones, observed after overnight incubation at 37°C, documented the positions of antibiotics separated by TLC. Two indicator strains sensitive to albicidin and two indicator strains resistant to albicidin were tested (Fig. 1). All indicator strains harbored pUFR043 and pLAFR3 empty plasmids that confer resistance to gentamicin and tetracycline, respectively (antibiotics that were present in the inoculum of Xves-alb). A single growth inhibition zone was observed at the same position with the two DH5α derivative sensitive indicator strains for the samples prepared from X. axonopodis pv. vesicatoria strain Xves-alb and from X. albilineans strain Xa23R1. The absence and a significant intensity reduction of inhibition zones with resistant strains DH5αAlb^r and DH5αAlbD, respectively, confirmed that the antibiotics separated by TLC chromatography were albicidin.

In conclusion, we describe here the heterologous production of albicidin in the fast-growing bacterium X. axonopodis pv. vesicatoria, with production yields sixfold higher than those observed with X. albilineans. In addition to growing faster than X. albilineans, this heterologous production system has the substantial advantage of being more easily modifiable by genetic engineering because albicidin biosynthetic genes are cloned in plasmids. The following modifications are promising for improving the production yield of albicidin in this heterologous system: (i) the replacement of native promoters by constitutive promoters in order to suppress negative transcriptional control of albicidin production; (ii) the replacement of TTG codons for the initiation of translation by ATG codons in order to suppress any negative posttranscriptional control due to the presence of these TTG codons; (iii) the overexpression of the albicidin efflux pump (and/or the Qnr albicidin resistance protein) in order to facilitate the excretion of the antibiotic into the culture medium and potentially increase resistance; and (iv) the overexpression of the biosynthetic enzymes functioning in trans in order to improve the efficacy of the PKS-NRPS system. This latter approach was successfully used to improve the production of leinamycin by Streptomyces atroolivaceus (9). The present study is a first significant step in the overproduction of albicidin that will be necessary to purify large quantities of albicidin and to characterize this potent inhibitor of DNA gyrase that differs from other known antibiotics.

• Open in new tab
• Download powerpoint

FIG. 1.

Analysis of antibiotics separated by TLC from supernatants of cultures of strains X. albilineans Xa23R1 and X. axonopodis pv. vesicatoria Xves-alb. Samples prepared from bacterial supernatants were subjected to TLC, and TLC strips were laid on top agar layers containing different E. coli indicator strains: DH5α (lane a); DH5αAlb^r (lane b), a spontaneous albicidin-resistant DH5α derivative described previously (12); DH5α transformed with the pGEX-4T-3 empty vector (Pharmacia) (lane c); and DH5α transformed with pGEXAlbD (DH5αAlbD) and expressing albD, an albicidin-detoxifying gene conferring albicidin resistance in E. coli (19) (lane d). All E. coli indicator strains were transformed with pUFR043 and pLAFR3 empty vectors to confer resistance to gentamicin and tetracycline, respectively. The initial positions of the samples on the chromatograms are indicated by “x”.

• Copyright © 2007 American Society for Microbiology