Xanthomonas albilineans HtpG is required for biosynthesis of the antibiotic and phytotoxin albicidin

Abstract

Xanthomonas albilineans, the causal agent of leaf scald disease of sugarcane, produces a highly potent polyketide‐peptide antibiotic and phytotoxin called albicidin. Previous studies established the involvement of a large cluster of genes in the biosynthesis of this toxin. We report here the sub‐cloning and sequencing of an additional gene outside of the main cluster and essential for albicidin biosynthesis. This gene encodes a 634‐amino‐acid protein that shows high identity with the Escherichia coli heat shock protein HtpG. Complementation studies of X. albilineans Tox⁻ mutants confirmed the requirement of htpG for albicidin biosynthesis and revealed functional interchangeability between E. coli and X. albilineans htpG genes. HtpG was co‐localised with albicidin in the cellular membrane, i.e., the cellular fraction where the toxin is most probably biosynthesised. Here we show the requirement of an HtpG protein for the biosynthesis of a polyketide‐peptide antibiotic.

Leaf scald, Nonribosomal peptide synthetases, Polyketide synthases, Sugarcane

1 Introduction

Xanthomonas albilineans is a systemic, xylem‐invading, slow growing, Gram‐negative bacterium that causes leaf scald disease of sugarcane (interspecific hybrids of Saccharum species) [1]. Leaf scald symptoms include chlorosis, necrosis, rapid wilting, and plant death. Chlorosis‐inducing strains of the pathogen produce several toxic compounds. The major toxic component, named albicidin, plays a key role in pathogenesis [2]; it inhibits chloroplast DNA replication, resulting in blocked chloroplast differentiation and appearance of chlorotic leaf streaks that are characteristic of the disease [3]. Albicidin also inhibits prokaryotic DNA replication and is bactericidal to a range of Gram‐positive and Gram‐negative bacteria [4]; this compound is, therefore, of interest as a potential clinical antibiotic. Although the chemical structure of albicidin remains unknown, it has been partially characterised as a polyketide‐peptide antibiotic with a molecular weight of about 842 Da [4].

Transposon mutagenesis study revealed that at least two separate genomic regions are involved in albicidin production by X. albilineans [5]. Among a total of 50 generated albicidin deficient mutants, 45 were localized in region XALB1. This region contains 20 ORFs including three major genes with a modular architecture characteristic of polyketide synthases (PKSs) and nonribosomal peptide synthetases (NRPSs), as well as several putative resistance, regulatory and modifying genes [6,7]. NRPSs synthesise peptides by condensation of individual amino‐acid or amino‐acid analogs in a biosynthetic pathway that does not involve ribosomes [8]. Based on in silico analyses of the albicidin PKSs and NRPSs, an albicidin biosynthesis pathway has been proposed [7]. Two other mutants were located in XALB2, a region containing only one gene coding for a phosphopantheteinyl transferase. This enzyme is required for post‐translational activation of albicidin PKS and NRPS enzymes [7,9]. Three additional mutants, XaAM12, XaAM13 and XaAM36, were complemented by a cosmid, pALB639 (Table 1) that did not cross hybridise with cosmids containing XALB1 or XALB2 [5].

1

Bacterial strains, plasmids and oligonucleotides used in this study

Strains, plasmids or oligonucleotides	Relevant characteristicsa	Reference or source
Strains
Escherichia coli
DH5α	F‐f 80dlac ZΔM15 Δ(lac ZYA‐arg FU169 deo R recA 1 endA 1 hsdR17 (rk− k+ supE44 thi‐1 gyrA 96 relA 1	Gibco‐BRL
Xanthomonas albilineans
Xa23	Wild type from sugarcane (Florida)	[5]
Xa23R1	Spontaneous Rifr derivative of Xa23	[5]
XaAM10	Xa23R1::Tn5‐gusA in XALB2, Kmr, Rifr, Tox−	[5]
XaAM12	Xa23R1::Tn5‐gusA in XALB3, Kmr, Rifr, Tox−	[5]
XaAM13	Xa23R1::Tn5‐gusA in XALB3, Kmr, Rifr, Tox−	[5]
XaAM15	Xa23R1::Tn5‐gusA in XALB1, Kmr, Rifr, Tox−	[5]
XaAM36	Xa23R1::Tn5‐gusA in XALB3 Kmr, Rifr, Tox−	[5]
Plasmids
pBR325	Tcr, Apr, Cmr	Gibco‐BRL
pGEMT	Apr	Promega
pRK2073	pRK2013 derivative, Kms (npt::Tn7), Spr, Tra+, helper plasmid	Cited in [5]
pUFR043	IncW, Mob+, LacZα, Gmr, Kmr, Cos	Cited in [5]
pALB639	36‐kb insert from Xa23R1 in pUFR043, Gmr, Kmr	[5]
pALB 540	47‐kb insert from Xa23R1 in pUFR043, Gmr, Kmr	[5]
pAM12.1	11kb Eco RI fragment carrying Tn5 and flanking sequences of mutant XaAM12 in pBR325, Kmr, Tcr, Apr, Cmr	[5]
pAM36.2	8‐kb Eco RI fragment carrying Tn5 and flanking sequences of mutant XaAM36 in pBR325, Kmr, Tcr, Apr, Cmr	[5]
pALB639A	9.7‐kb Eco RI‐Sau 3AI insert from pALB639 in pUFR043, Gmr, Kmr	This study
pEV639	2.6‐kb Sal I insert from pALB639A in pUFR043, Gmr, Kmr	This study
pGEMTalbHtpG	1948 bp PCR insert from Xa23R1 in pGEMT, Apr	This study
pGEMTalbHtpG bis	X. albilineans htpG ORF plus 111bp downstream from stop codon in pGEMT, Apr	This study
pGEMTcolHtpG	2343 bp PCR insert from DH5α in pGEMT, Apr	This study
pAlbH	X. albilineans htpG in fusion with lacZ in pUFR043, Gmr, Kmr	This study
pHisAlbH	Hexa histidine tag in fusion with X. albilineans htpG in pAlbH, Gmr, Kmr	This study
pColH	E. coli htpG in fusion with lacZ in pUFR043, Gmr, Kmr	This study
Oligonucleotides
GUSN	5′tgcccacaggccgtcgagt3′	Genome express
alb22F	5′tttgaattcgcacctaccgatgagcgtgg3′	Genome express
alb22R	5′tttggatccgtgcgtcactgcttacgccg3′	Genome express
RACE22a	5′gtgcgaatccttcttctgatc3′	Genome express
RACE22b	5′ctcatgccgatgccgttgtcg3′	Genome express
RACE22c	5′tccagaagttccggcttgacc3′	Genome express
ColHTF	5′tttgaattccatgaaaggacaagaaactcgtgg3′	Genome express
ColHTR	5′gcctgcggaatggtacgcgggaagccgtcc3′	Genome express
HistagalbhtpgF	5′aattgcgtcgaccaccatcaccatcaccat3′	Genome express
HistagalbhtpgR	5′aattatggtgatggtgatggtggtcgacg3′	Genome express

Tox⁻: deficient in albicidin production.

^aAp^r, Cm^r, Gm^r, Km^r, Rif^r, Sp^r, Tc^r: resistant to ampicillin, chloramphenicol, gentamicin, kanamycin, rifampicin, spectinomycin, tetracycline, respectively.

We report here the subcloning, sequencing and homology analysis of a third region involved in albicidin biosynthesis, XALB3, present in cosmid pALB639.

2 Materials and methods

2.1 Bacterial strains, plasmids and culture conditions

The source of bacterial strains and their relevant characteristics are described in Table 1. X. albilineans and E. coli culture conditions and bacterial conjugation have been described elsewhere [5,10].

2.2 Albicidin detection assay

Albicidin was detected by a microbiological assay as described previously [5]. Alternatively, liquid culture media, lysates or cell fractions were spotted on LB plates overlaid with a top agar containing a suspension of E. coli susceptible to albicidin. Albicidin was quantified with a bioassay, using the following formula [11]: free albicidin (ng/ml) = 4.576 e^{(0.315 × Inhibition zone diameter in mm)}.

2.3 Nucleic acid manipulations, oligonucleotide synthesis and DNA sequencing

Standard molecular techniques were used to manipulate DNA [10] except that X. albilineans total genomic DNA was prepared for Southern blot hybridisation as described by Gabriel and De Feyter [12]. Automated DNA sequencing was performed by Genome Express (Grenoble or Montreuil, France). Sequence analyses were carried out using Sequence Navigator™ (Applied Biosystems, Inc.) and SeqMan (DNASTAR Inc.) programs.

2.4 Sequencing of the Tn5 insertional site of Tox⁻ mutants located in XALB3 and sequencing of the genomic region XALB3

The complete double‐strand nucleotide sequence of the 9673 bp Eco RI–Sau 3AI genomic fragment, subcloned into pALB639A, was determined by sequencing pAM12.1, pAM36.2 and pALB639A (Fig. 1; GenBank Accession No. AM039979). The Tn5 insertional site of mutants XaAM12 and XaAM36 was determined by sequencing pAM12.1 and pAM36.2, respectively (Table 1 and [5]), using primer GUSN (Table 1) that annealed 135 bp downstream from the insertional sequence IS50L of Tn5‐gus A. Tn5 transposon insertion site could not be determined in mutant XaAM13 because, for unknown reasons, subcloning and PCR amplification of the Tn5 flanking region were unsuccessful.

1

Physical map and genetic organisation of DNA fragments containing the htpG gene in different xanthomonads. (A) Physical map and genetic organisation of the DNA fragment containing the genomic region XALB3 involved in albicidin production. The shaded box at the top of the figure represents the XALB3 genomic region. E, B, S and Sa are restriction sites used for subcloning of XALB3 and Xanthomonas albilineans htpG; they are for Eco RI, Bgl II, Sal I and Sau 3AI, respectively. The DNA inserts carried by plasmids pALB639, pALB639A, pEV639 and pAlbH are represented by thick bars at the top of the figure. Position of the Tn5 insertional sites of mutants XaAM12 and XaAM36 are indicated by vertical arrows. The DNA regions corresponding to Tn5 flanking sequences in pAM12.1 and pAM36.2 are represented by the bars at the bottom of the figure. The location and direction of htpG and of the surrounding ORFs are indicated by the thick arrows. The transcription initiation site of htpG is indicated by a broken arrow on pEV639. Proposed functions from left to right: Hyp4 and Hyp3: Hypothetical proteins related to Caulobacter crescentus Hypothetical proteins AAK22542 and AAK22541, respectively, ggt: Gamma‐glutamyltranspeptidase, Fox*: interrupted ferredoxin ORF, Hyp2: Hypothetical protein, kdtB: lipopolysaccharide synthesis enzyme, Hyp1: Hypothetical protein, and btuB: TonB dependent receptor. (B) Physical map and genetic organisation of the DNA fragment containing the htpG gene in X. campestris pv. campestris ATCC 33913. From left to right: egl2: Cellulase (AAM41665), ggt: Gamma‐glutamyltranspeptidase (AAM41666), Fox: Ferredoxin (AAM41667), Hyp2: Hypothetical protein XCC2390 (AAM41668), kdtB: lipopolysaccharide synthesis enzyme (AAM41669), Hyp1: Hypothetical protein XCC2392 (AAM41670), htpG: Heat shock protein G (AAM41671), Hyp0: Hypothetical protein XCC2394 (AAM41672) and btuB: TonB dependent receptor (AAM41673). (C) Physical map and genetic organisation of the DNA fragments containing the htpG gene in X. axonopodis pv. citri ATCC 306. From left to right: egl2: Cellulase (AAM37373), ggt: Gamma‐glutamyltranspeptidase (AAM37374), Fox: Ferredoxin (AAM37375), Hyp2: Hypothetical protein XAC2525 (AAM37376), kdtB: lipopolysaccharide synthesis enzyme (AAM37377), Hyp1: Hypothetical protein XAC2527 (AAM37378), htpG: Heat shock protein G (AAM37379), RhsD: RhsD protein (AAM37380), Hyp0: Hypothetical protein XAC2530 (AAM37381) and btuB: TonB dependent receptor (AAM37382). (D) Southern blot analysis of Eco RI digest of pALB639 (lane 1), and of total genomic DNA from Tox⁻ mutants XaAM10 (lane 2), XaAM12 (lane 3), XaAM13 (lane 4), XaAM15 (lane 5), XaAM36 (lane 6) and Xa23R1 (lane 7) using pALB639A as a probe. The approximate size (in kb) of the different bands is indicated on the left of the lanes. The double band of lane 1 corresponds to the pUFR043 vector derived band and the 10 kb XALB3 Eco RI–Sau 3AI subgenomic fragment of pALB639 (the Eco RI restriction site located downstream from Sau 3AI in pALB639 originates from the pUFR043 vector cassette). The 11 kb band in Xa23R1 (lane7) corresponds to the XALB3 Eco RI subgenomic fragment. As expected, the same Eco RI 11 kb band was found in XALB1 mutant (lane 5) and in XALB2 mutant (lane 2). The two bands of 8 and 12 kb, that hybridised in XALB3 mutants (lanes 3, 4 and 6), resulted from the digestion of the Eco RI restriction site present in the 9 kb Tn5 transposon.

2.5 Sequences analysis

Nucleotide sequences were translated in all six reading frames using EditSeq (DNASTAR Inc.). Potential products of ORFs longer than 100 bases were compared to protein databases using the PSI‐BLAST program [13] on the NCBI site (http://www.ncbi.nlm.nih.gov/). Sequences were aligned using the Megalign Software (DNASTAR Inc.). The TERMINATOR program [14], run with the Wisconsin Package™ GCG software (Accelrys, San Diego, CA) was used to identify putative Rho‐independent transcription terminators. In silico cellular localisation was predicted using the PSORT software [15] on the PSORT web server (http://psort.nibb.ac.jp).

2.6 Mapping of the X. albilineans htpG transcription initiation

Total RNA was extracted from 2‐day‐old cultures of X. albilineans Xa23R1 using Master complete RNA purification kit (Epicentre, Madison, WI, USA) according to the manufacturer's recommendations, except that cells were initially resuspended in RNA protect Bacteria reagent (Qiagen, Courtaboeuf, France). 5′ rapid amplification of cDNA end (5′ RACE) was subsequently performed with the 5′/3′ RACE kit (Roche, Meylan, France) according to the manufacturer's recommendations and with oligonucleotides RACE22a, RACE22b and RACE22c, that anneal 363, 266 and 176 bases downstream from the htpG ATG, respectively (Table 1). The resulting PCR product was sequenced using oligonucleotide RACE22c.

2.7 Subcloning of the X. albilineans htpG gene

The Sal I DNA fragment corresponding to the XALB3 DNA region from nucleotide 5510 to nucleotide 8124, was subcloned into pUFR043 in the opposite direction to the promoter of the lacZ operon, yielding plasmid pEV639 (Table 1 and Fig. 1). Simultaneously, a 1948 bp fragment corresponding only to the 1905 bp ORF of htpG was PCR‐amplified from cosmid pALB639A with primers alb22F and alb22R (Table 1). The PCR fragment was cloned into pGEMT (Promega) and several clones of the resulting plasmid pGEMT/albHtpG were sequenced. Plasmid pGEMT/albHtpGbis contained an insert corresponding to the intact htpG ORF plus 111 bp downstream from the stop codon. This insert was then directionally subcloned as an Eco RI–Sal I fragment into pUFR043 to drive transcription of htpG from the promoter of the lacZ operon of the vector. The resulting construct, pAlbH, together with pEV639, were used for complementation studies.

2.8 Cloning of the E. coli htpG gene

A 2343 bp fragment containing htpG from E. coli was PCR amplified from purified DH5α genomic DNA with primers ColHTF and ColHTR (Table 1). PCR fragment was cloned into pGEMT, sequenced, and subcloned as an Eco RI–Sal I fragment into pUFR043. The resulting clone, pColH, carrying htpG driven by the promoter of the lacZ operon, was used for complementation studies (Table 2).

2

Complementation studies of insertion mutants of Xanthomonas albilineans Xa23R1

Donor	Recipient or Tox− mutant (Genomic region)
	XaAM12 (XALB3)	XaAM13 (XALB3)	XaAM36 (XALB3)	XaAM10 (XALB2)	XaAM15 (XALB1)
pEV639	+	+	+	−	−
pAlbH	+	+	+	−	−
pColH	+	+	+	−	−
pALB639A	+	+	+	−	−
pUFR043	−	−	−	−	−
none	−	−	−	−	−

+: restoration of albicidin production, −: no restoration of albicidin production. All experiments were performed at least in duplicate with 15–20 exconjugants obtained from two independent triparental conjugations (with the exception of XaAM36 × pColH for which only two exconjugants were obtained by triparental conjugation).

2.9 Expression, detection and purification of histidine‐tagged HtpG derivatives

An histidine affinity tag coding sequence, generated from the annealing of oligonucleotides HistagalbhtpgF and HistagalbhtpgR (Table 1), was inserted as an in frame Eco RI linker upstream from htpG in pAlbH, generating pHisAlbH. This construct was used for complementation studies and immunodetection of HtpG. His‐tagged HtpG purification was performed with Ni‐NTA affinity beads (Qiagen SA, Courtaboeuf, France) according to the manufacturer's recommendations (QIAexpressionist™) except that cells were lysed by serial passage through a French press at 14,000 psi. Western blot immunodetection of His‐tagged HtpG was performed with penta‐His antibody (Qiagen SA, Courtaboeuf, France) according to the manufacturer's recommendations (QIAexpress™ Handbook).

2.10 Subcellular fractionation

The cells from 250 ml of four day‐old cultures of X. albilineans were pelleted by centrifugation (30 min at 10,000g). The pellet was resuspended in one tenth volume of lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 10 mM imidazole). Each whole cell suspension was then processed three times in a French press at 14,000 psi. The resulting lysate was centrifuged for 30 min at 12,000g to pellet the remaining intact cells. The resulting clarified lysate was ultracentrifuged for two hours at 200,000g to pellet the cell membranes. The supernatant, containing the cytosolic fraction was used to purify His‐tagged HtpG. The cell membrane pellet was briefly washed once with lysis buffer, resuspended in 200 μl PBS and was further processed for His‐tagged HtpG and albicidin detection.

3 Results and discussion

3.1 Cloning and sequencing of the XALB3 region

Plasmid pALB639A (Table 1) harbours a 10 kb fragment from pALB639 that hybridized with plasmid pAM12.1 carrying the Tn5 insertion site and flanking region of mutant XaAM12. This plasmid restored albicidin production in Tox⁻ mutants XaAM12, XaAM13 and XaAM36 and not in XALB1 and XALB2 Tox⁻ mutants, demonstrating that a third region involved in albicidin production, XALB3, was subcloned. Plasmid pALB639A was used to probe Eco RI restriction digests of pALB639 and genomic DNA from Xa23R1 and Tox⁻ mutants, and resulting banding patterns confirmed that XALB3 region differed from XALB1 and XALB2 (Fig. 1). The DNA sequence of the XALB3 region corresponding to the 9673 bp Eco RI–Sau 3AI insert of pALB639A was determined (GenBank accession N° AM039979). The Tn5 insertion sites of XaAM12 and XaAM36 were located at positions 6500 and 7232, respectively (Fig. 1). An Eco RI restriction site was located 1 kb downstream from the cloning Sau 3AI restriction site of pALB639 on the basis of the restriction analysis of pAM36.2 and Southern blot analysis of Xa23R1 (Fig. 1).

3.2 Genetic organisation of XALB3

Open reading frame analyses of the 9673 bp Eco RI–Sau 3AI fragment revealed that only one ORF was disrupted in mutants XaAM12 and XaAM36. This ORF, designated albXXII, was located between nucleotides 6105 (ATG) and 8009 (TAA). The ATG is preceeded at position −9 by the putative ribosome binding site sequence GGAG. A putative rho‐independent terminator was identified at position 8082, 73 bases downstream from the ORF. The transcription initiation site was mapped by RACE, 53 bases upstream from the ATG, at position 6052. No close matches to E. coli σ⁷⁰ promoter −10 (TATAAT) and −35 (TTGACA) consensus sequence were found upstream from the transcription initiation site. Alignment of the potential products of the flanking ORFs revealed a total of six putative ORFs (Fig. 1). With the exception of the first two ORFs (Hyp4 and Hyp3), whose predicted products are, respectively, similar to Caulobacter crescentus conserved hypothetical proteins (GenBank Accession No. AAK22542 and AAK22541), the genetic organisation of the XALB3 region is closely related to the corresponding region in X. axonopodis pv. citri strain 306 (Genbank Accession No. NC003919) and X. campestris pv. campestris strain ATCC33913 (Genbank Accession No. NC003902) with a degree of identity/similarity of potential protein products of at least 69% and 76%, respectively.

3.3 Homology analysis

The product of the albXXII ORF is a protein of 634 aa with a predicted size of 71.5 kDa. This protein is very similar to heat shock proteins G [85% identity with X. campestris pv. campestris HtpG (Genbank Accession No. AAM41671) or with X. axonopodis pv. citri HtpG (Genbank Accession No. AAM37379) and 60% identity with the E. coli HtpG (Genbank Accession No. AAA23460)]. Accordingly, albXXII was renamed htpG. Bacterial HtpGs are homologues of the yeast Hsp82 and of the human Hsp90α molecular chaperones (Genbank Accession No. P02829 and P07900 which show 41% and 37% similarity, respectively, with E.coli HtpG). All these proteins share a common structural plan ( [16] and Fig. 2). The htpG gene can be deleted in E. coli with no effect on the viability of the strain, but with a decreased growth rate at high temperatures [17]. Similarly, besides abolishing albicidin production, interruption of htpG did not affect the viability of X. albilineans at temperatures 28 and 30 °C (data not shown). Although its in vivo role remains unknown, the E. coli HtpG would facilitate de novo protein folding in stressed E. coli cells, presumably by interacting with the DnaK–DnaJ–GrpE molecular chaperone system [18].

2

Amino acid alignment of yeast Hsp82, human Hsp90α, and E. coli and X. albilineans HtpGs using the clustal method. The black lines below the sequences represent the ATP binding domain as described by Prodromou and Pearl [16]. Arrowheads represent the location of conserved Glu 33 (open arrowhead) and Asp 79 (black arrowhead) residues in the Hsp82 nucleotide binding site. Location of the insertion of transposon Tn5 in mutants XaAM12 and XaAM36 is indicated by a star below the sequence. Vertical bars represent the borderlines between the N‐terminal ATP binding region, the intermediate region and the carboxyterminal dimerisation domain as described by Nemoto et al. [23]. Horizontal box below the sequence alignment corresponds to the carboxy‐terminal dimerisation region. Residues in bold are identical in all four organisms. Shaded residues are identical in at least two organisms.

3.4 Complementation analysis of XALB3 mutants

Transfer of pAlbH (containing X. albilineans htpG fused with the promoter of the lacZ operon) into XaAM12, XaAM13 and XaAM36 restored albicidin production, thus confirming that this ORF was involved in albicidin biosynthesis (Table 2). Transfer of pEV639 (containing htpG, together with both the native promoter within the 542 bp upstream from the transcription initiation site and the putative rho‐independent terminator, in opposite direction to the promoter of the lacZ operon) into XaAM12, XaAM13 and XaAM36 also restored albicidin production (Fig. 1 and Table 2). Transposon mutagenesis classically generates polar mutations that affect all downstream cistrons of an operon. However, the complementation results with pEV639 indicated that the X. albilineans htpG ORF was not transcribed as a part of a larger operon (as could also be assessed from the genetic organisation of the XALB3 region), and confirmed that the promoter of this ORF is present in the 542 bp upstream from the transcription initiation site.

Additionally, the E. coli htpG gene, expressed from the promoter of the lacZ operon in pColH, also complemented the three XALB3 mutations, and fully restored albicidin production in XaAM12, XaAM13 and XaAM36. This heterologous complementation revealed the substantial functional interchangeability of the two prokaryotic htpG genes, a finding consistent with similar studies using eukaryotic Hsp90 [19].

3.5 HtpG inhibition assay

Because Hsp90 is antagonised by the antibiotic novobiocin [20], we hypothesised that this coumarinic antibiotic, that does not affect X. albilineans growth, may affect HtpG and therefore albicidin biosynthesis. To test the antagonistic activity of novobiocin on albicidin biosynthesis, X. albilineans Xa23R1 was grown for four days in liquid MW medium containing increasing amounts of novobiocin, from 30 μg/ml to 1.5 mg/l. Growth of Xa23R1, regularly monitored by absorbance at OD = 600, was not affected up to 150 μg/ml. In contrast, albicidin biosynthesis was reduced at the first tested concentration of novobiocin (30 μg/ml), and was no longer detectable at 45 μg/ml of novobiocin.

Overall, the high degree of identity between X. albilineans and E. coli HtpGs, the functional complementation of XaAM12, XaAM13 and XaAM36 Tox⁻ mutants with the X. albilineans and the E. coli htpG ORFs under the promoter of the lacZ operon, and the inhibition of albicidin biosynthesis by a Hsp90 antagonist, confirmed that X. albilineans HtpG is required for albicidin biosynthesis.

3.6 Subcellular localisation of HtpG and albicidin

In silico cellular localisation prediction with software PSORT was performed to predict the cellular localisation of the three PKS and NRPS enzymes involved in albicidin biosynthesis: AlbI (6879 aa, Genbank Accession No. CAE52339), AlbIV (941 aa, Genbank Accession No. CAE52342) and AlbIX (1959 aa, Genbank Accession No. CAE52334). AlbI and, to a lesser extent, AlbIV and AlbIX, were predicted to be membrane localised, a result consistent with the membrane localisation of polyketide synthase in Myxococcus xanthus [21]. We therefore hypothesised that albicidin was synthesised at the level of the cytoplasmic membrane and we analysed the subcellular localisation of HtpG and albicidin.

In preliminary western blot experiments, a polyclonal antibody raised against human Hsp90 (Stressgen Biotechnologies) did not specifically react with X. albilineans HtpG (data not shown). A histidine tag coding sequence was therefore translationally fused to the amino terminal end of the protein, allowing both purification and specific detection. This construct (pHisAlbH) restored albicidin biosynthesis after transfer by triparental conjugation into Tox⁻ mutant XaAM12 (XALB3 region) but not after transfer into Tox⁻ mutant XaAM15 (XALB1 region), indicating that the His tag did not interfere with the in vivo function of HtpG in albicidin biosynthesis. A His tagged‐HtpG specific 68kDa band was detected in whole cell protein extracts, regardless the cell type (Tox⁺ of XaAM12‐ pHisAlbH and Tox⁻ of XaAM15‐ pHisAlbH), demonstrating the utility of the His tag in HtpG detection (Fig. 3).

3

Localisation of His tagged‐HtpG in subcellular fractions of X. albilineans mutants. Tox⁺ and Tox⁻X. albilineans strains (Albicidin defective mutant XaAM12 complemented with pHisAlbH (H) or transformed with pUFR043 (U) and mutant XaAM15 transformed with pHisAlbH (H) or pUFR043 (U), respectively) were fractionated as indicated in Materials and Methods section. The fractions (left to right) included whole cells (Cells), washed ultracentrifuge pelleted membrane fraction of cleared lysate (Memb), ultracentrifuge supernantant of cleared lysate (Cyt) and eluate of Ni‐NTA His affinity chromatography column used for purification of His tagged HtpG from ultracentrifuge supernantant of cleared lysate (Cyt*). Nylon (Nitrocellulose) membranes were probed with mouse anti penta‐His antibody. All lanes were loaded with the same amount of sample (20 μl). Arrowhead on the right of the figure corresponds to a molecular mass of ∼ 68 kDa.

Protein extracts from Tox⁻ and Tox⁺ strains were fractionated into membrane and cytosolic fractions. His‐tagged HtpG was found in the membrane fraction in albicidin producing cell type (XaAM12 mutant complemented by pHisAlbH) as well as in Tox⁻ mutant (XaAM15 mutant transformed but not complemented by pHisAlbH). Tagged HtpG was also detected in the cytosolic fraction, but only after a ∼15‐fold concentration of the protein on Ni‐NTA affinity beads. The presence of HtpG in the membrane is consistent with previous findings on localisation of the Porphyromonas gingivalis HtpG [22] (that shows 28% similarity with X. albilineans HtpG). Most interestingly, albicidin was also found in the membrane fraction (∼ 250 ng/ml), but not in the cytoplasmic fraction (even after concentration). Co‐localisation of albicidin and X. albilineans HtpG in the membrane fraction and predicted membrane localisation of the albicidin structural genes AlbI, AlbIV and AlbIX, although not a proof by itself, are in agreement with the involvement of X. albilineans HtpG in the NRPS/PKS biosynthesis machinery. Despite of its undoubtful involvement in the biosynthesis of this hybrid non ribosomal peptide‐polyketide antibiotic, the precise role of HtpG in albicidin biosynthesis remains unclear. In preliminary double hybrid experiments (Stratagene bacteriomatch II system) using X. albilineans HtpG as the bait, no client protein was identified for HtpG when a X. albilineans Sau 3AI genomic library and a sub‐library, made with the previously identified albicidin biosynthesis genes, were screened (data not shown).

4 Conclusion

This study represents a milestone in our effort to help clarify the role of albicidin in leaf scald disease progress since the availability of all three genomic regions known to be involved in albicidin production (XALB1, XALB2 and XALB3) offers the possibility to characterise all enzymes of the albicidin biosynthesis pathway, including structural, resistance, secretory and regulatory elements. In addition, cloning all of the biosynthetic genes involved in albicidin production may help to overcome limitations in albicidin studies due to low yields of toxin production by wild‐type X. albilineans [3]. Proof that all genes involved in albicidin biosynthesis are now identified will require expression in a heterologous host. Since all mutants isolated in the earlier, extensive transposon mutagenesis study of X. albilineans [5], have now been complemented, it is possible, if not likely, that all of the albicidin biosynthetic genes have been cloned. If so, this effort offers the possibility to engineer high level co‐expression of the genes to obtain higher albicidin production. It also provides the potential to manipulate the albicidin biosynthetic machinery in order to obtain structural variants of this potent novel class of antibiotics.

Acknowledgements

E. Vivien was supported by a fellowship from the Direction Scientifique du CIRAD. We thank Dr J.B. Morel for critical reading of the manuscript and Sandrine Duplan for her expertise and skillfull assistance. We also thank Dr J.F. Dubremet for the French press facilities at Montpellier University ‐ UMII.