3-Ketoacyl-ACP synthase III FabH1 is essential for branched-chain DSF family signals in Xanthomonas oryzae pv. oryzae

Abstract

The 3-ketoacyl-ACP synthase III (FabH), a key enzyme for bacteria growth, catalyses the last step of the initiation of bacterial fatty acid synthesis. Rice bacterial blight is caused by the pathogen Xanthomonas oryzae pv. oryzae (Xoo), which is widely studied as a model bacterium. Bioinformatics analysis showed that the X. oryzae pv. oryzae PXO99A genome encodes three FabH homologous proteins with unknown functions. In this study, we found that only PXO_02706 (fabH1) encodes a functional FabH, the key enzyme in the production of branched-chain fatty acid, which is essential for the branched-chain diffusible signal factor family signals in Xoo. Interestingly, we found that FabH1 is not essential for fatty acid biosynthesis in Xoo. Pathogenicity analysis showed that loss of fabH1 caused a significant decrease in virulence of Xoo. Genetic and phenotypic analyses revealed that fabH1 plays a key role in multiple Xoo virulence-related activities, including exopolysaccharide (EPS) production, biofilm formation, motility, and resistance to environmental stresses.

Background

Xanthomonas oryzae pv. oryzae (Xoo) is among the top ten most devastating plant pathogenic bacteria, and is the causal agent of rice bacterial leaf blight disease (Nino-Liu et al. 2006; Mansfield et al. 2012). During blight infection, Xoo activates the diffusible signal factor (DSF) family quorum sensing (QS) signals, which globally regulates pathogenicity (He and Zhang 2008; He et al. 2010). Previous studies have shown that DSF-family signal components, such as DSF (cis-11-methyl-2-dodecenoic acid), CDSF (cis, cis-11-methyldodeca-2,5-dienoic acid), IDSF (cis-10-methyl-2-dodecenoic acid), and BDSF (cis-2-dodecenoic acid), regulate virulence factor production in Xanthomonas spp. by the rpf gene cluster, including rpfF, rpfB, rpfC, and rpfG that encode DSF synthase, acyl-CoA ligase, sensor, and response regulator, respectively (Slater et al. 2000; Andrade et al. 2006; He et al. 2010; An et al. 2014; Bi et al. 2014; Wang et al. 2016). 3-Hydroxyacyl-ACPs are the precursors of DSF-family signal components (Zhou et al. 2015, 2017b). Previous studies showed that 3-hydroxyacyl-ACPs are generated by the fatty acid synthesis (FAS) pathway (Zhou et al. 2015).

FAS is important for all organisms (White et al. 2005; Zhang and Rock 2008). Fatty acids (FAs) are not only required for biosynthesis of cell membrane phospholipids (Zhang and Rock 2008), but also are the intermediates of the biosynthesis of many bioactive substances, such as biotin (Lin and Cronan 2011), rhamnolipids (Du et al. 2017), lipoic acid (Cronan et al. 2005), lipid A (Zhang and Rock 2008), and QS signalling molecules including DSFs and N-acyl-homoserine lactones (AHLs) (Zhou et al. 2015; Huang et al. 2016). De novo FA synthesis in bacteria is catalysed by the type II fatty acid synthesis system (FAS II), which is composed of a series of small independent enzymes (Cronan and Rock 2008). Intermediates in FAS II are shuttled from between biosynthesis enzymes by a small highly acidic acyl carrier protein (ACP) (Campbell and Cronan 2001; Cronan and Rock 2008). Fatty acid biosynthesis begins with the conversation of acetyl-CoA into malonyl-CoA by acetyl-CoA carboxylase AccABCD (Davis et al. 2000). Malonyl-CoA:ACP transacylase (FabD) then transfers malonyl-CoA to malonyl-ACP (Magnuson et al. 1992). Malonyl-ACP further combines with an acetyl-CoA group in the first condensation reaction, catalysed by 3-ketoacyl-ACP synthetase III (KAS III), to form 3-ketobutyryl-ACP (Cronan and Rock 2008). KAS III is a key enzyme for bacterial fatty acid biosynthesis that determines the type and amount of fatty acids in different bacteria (Marrakchi et al. 2002; White et al. 2005). In Escherichia coli, FabH functions as a KAS III that preferentially utilises acetyl-CoA and malonyl-ACP to synthesize the straight-chain fatty acids (SCFAs) (Heath and Rock 1995), whereas its counterparts in many Gram-positive bacteria, such as Bacillus subtilis, selectively catalyse branched-chain acyl-CoA and malonyl-ACP to generate branched-chain fatty acids (BCFAs) (Choi et al. 2000).

Xoo has a complex fatty acid profile which is comprised of saturated SCFAs, unsaturated fatty acids (UFAs), and BCFAs (Vauterin et al. 1996), but the FAS pathway in Xoo is less known. In the present study, we found that PXO_02706 (fabH1) is the only functional KAS III. It plays a key role in the biosynthesis of BCFAs and branched-chain DSF family signal components in Xoo. Pathogenicity analysis showed that the fabH1 gene was required for full virulence of Xoo on rice. Knockout of fabH1 resulted in a significant decrease in the production of several pathogenicity-related virulence activities, including exopolysaccharide (EPS) production, biofilm formation, motility, and resistance to environmental stresses.

Results

Only fabH1 can complement the growth of Ralstonia solanacearum fabH deletion mutant

According to the genome annotation, the Xanthomonas oryzae pv. oryzae PXO099A genome contains three putative fabH genes (Fig. 1a), the fabH1 (PXO_02706), fabH2 (PXO_00990), and fabH3 (PXO_03957). All three putative FabHs contain a Cys-His-Asn catalytic triad (Fig. 1b), implying that they may function in initiation of fatty acid synthesis. The fabH1 gene is located within a cluster of fatty acid synthesis genes (plsX-fabH1-fabD-fabG-acpP-fabF). About 60.1% of the residues of FabH1 protein are identical to the orthologous of E. coli. fabH2 and fabH3, encoding 3-ketoacyl-ACP synthase III (KAS III) candidates, are located in a separate operon that is far away from the above-mentioned genes. FabH2 and FabH3 share only 27.1% and 28.6% identity with E. coli orthologous, respectively.

Fig. 1

FabH homologue coding genes and the complementation of R. solanacearum fabH knockout strain ΔRsfabH. a Loci of XoofabHs. The fabH1 locus is similar to the E. coli fabH locus. b Structure-based amino acid alignments of FabH homologues from Xoo. Residues that constitute the His-Asn-Cys catalytic triad are highlighted (Cys112, His244, and Asn274 in EcFabH). c R. solanacearum fabH knockout strain ΔRsfabH carrying the plasmid of pMF1(fabH1) was grown on BG medium in the absence of octanoic acid

To explore whether these proteins function in fatty acid synthesis initiation, the plasmids pMF1 (fabH1), pMF2 (fabH2), and pMF3 (fabH3) were transformed into the R. solanacearum fabH deletion mutant ΔRsfabH, which does not grow on BG medium without octanoic acid supplementation (Mao et al. 2015). The ΔRsfabH harbouring plasmid pMF1 grew well in the absence of octanoic acid, whereas pMF2 and pMF3 could not rescue the growth of ΔRsfabH under the same conditions (Fig. 1c). To investigate the functions of FabH1 in fatty acid synthesis, the fatty acids in the resulting strains were determined by gas chromatography–mass spectrometry (GC–MS). R. solanacearum GMI1000 does not produce branched-chain fatty acids (BCFAs) (Table 1). However, the ΔRsfabH mutant harbouring pMF1 contained ~ 48.8% BCFAs, including iso-C_15:0, iso-C_17:0, and antesio-C_17:0 (Table 1). These results indicate that only fabH1 encodes a functional KAS III, enabling the initiation of fatty acid biosynthesis in vivo.

Table 1 Fatty acid composition of ΔRsfabH complementary strains and Xoo mutant strains

Analysis of Xoo FabH activities in vitro

In order to perform a direct in vitro assay of Xoo FabHs’ activities, the recombinant proteins were expressed and purified from E. coli (Fig. 2a). Various E. coli fatty acid biosynthetic proteins, including FabD, FabB, FabG, FabA, FabI, holo-ACP, and Vibrio harveyi acyl-ACP synthetase (AasS), were also purified using the same method.

Fig. 2

XooFabH1 encodes a 3-ketoacyl-ACP synthase III to initiate FASII. a SDS-PAGE analysis of purified XooFabHs. b Condensation of malonyl-ACP with acetyl-CoA by purified EcFabH (lane 3), XooFabH1 (lane 4), XooFabH2 (lane 5), or XooFabH3 (lane 6). The migration positions of octanoyl-ACP (C_8:0-ACP, lane 7) on gel are shown. c Condensation of malonyl-ACP with isobutyryl-CoA (lane 1) or isovaleryl-CoA (lane 3) by purified FabH1. The migration positions of octanoyl-ACP (C_8:0-ACP, lane 4) on gel are shown. d Condensation of malonyl-ACP with acyl-CoAs (C_4:0-CoA to C_12:0-CoA) by purified XooFabH1 (lane 3 to lane 7). The migration positions of hexanoyl-ACP (C_6:0-ACP, lane 8) and octanoyl-ACP (C_8:0-ACP, lane 9) on gel are shown. The initiation fatty acid synthesis was reconstructed using a combination of E. coli FabZ, FabG, FabI, and Xoo FabH1 with NADH, and NADPH as cofactors, and malonyl-ACP plus acetyl-CoA or isoacetyl-CoA as substrates

To investigate the functions of Xoo FabHs, the initial reaction was reconstructed by adding of acetyl-CoA, malonyl-ACP, FabHs, and E. coli fatty acid synthesis enzymes mentioned above, and the products were analysed by conformationally sensitive gel electrophoresis. E. coli FabH could condense acetyl-CoA and malonyl-ACP to generate 3-oxobutyryl-ACP, and finally produce butyryl-ACP (Fig. 2b, lane 3). Interestingly, FabH1 could produce longer chain products, namely hexanoyl-ACP and octanoyl-ACP (Fig. 2b, lane 4). These results suggest that FabH1 not only catalyses the initiation reaction, but also functions in the elongation cycle, where it utilizes acyl-ACP and malonyl-ACP to generate C_6:0-ACP or C_8:0-ACP. However, adding of FabH2 or FabH3 in the reaction system failed to produce any new bands (Fig. 2b, lanes 5 and 6), indicating that neither fabH2 nor fabH3 encodes KAS III.

To investigate the substrate specificity of FabH1, branched-chain acyl-CoAs (isobutyryl-CoA or isovaleryl-CoA) were tested. The results showed that FabH1 used branched-chain acyl-CoAs, isobutyryl-CoA, or isovaleryl-CoA as a primer to initiate fatty acid synthesis (Fig. 2c, lanes 1 and 3). The activities of FabH1 to utilise straight-chain acyl-CoAs (butyryl-CoA, hexanoyl-CoA, octanoyl-CoA, decanoyl-CoA, or dodecanoyl-CoA) as substrates were also tested, where each of the acyl-CoAs could be converted to the longer acyl-ACPs by FabH1 (Fig. 2d). We also found that FabH2 and FabH3 failed to elongate any of the straight-chain or branched-chain acyl-CoAs (Additional file 1: Figure S1). Thus, only FabH1 functions in fatty acid biosynthesis.

FabH1 is not essential for growth, but is essential for BCFA biosynthesis in Xoo

The physiological functions of FabHs were further explored via gene deletion mutants. Using suicide plasmids carrying in-frame gene deletions to disrupt fabH1, fabH2, or fabH3, single-gene knockout of each of the three putative fabH genes in the PXO99A strain was achieved. Deletion of any of the fabH genes did not affect their growth on NA medium, suggesting that none of the genes are essential for growth of Xoo (Additional file 1: Figure S2).

To test whether FabH1 has 3-ketoacyl-ACP synthetase III activity in vivo, we analysed the fatty acid composition of ΔfabH1, ΔfabH2, and ΔfabH3 by GC–MS. The FA profile of the wild type (WT) strain Xoo PXO99A included iso-BCFAs, anteiso-BCFAs, saturated straight-chain fatty acids (SCFAs), and unsaturated SCFAs (Table 1). In Xoo PXO99A, anteiso-BCFAs were only observed with trace amounts, and iso-BCFAs were the predominant BCFAs, including iso-C_15:0 (7.72 ± 0.05%) and iso-C_17:0 (8.72 ± 0.84%). The predominant FAs in Xoo were SCFAs, mainly the n-C_16:0 (41.64 ± 2.2%) and n-C_16:1 (30.9 ± 2.27%). Deletion of fabH1 resulted in the loss of BCFAs (< 1%), while the composition of n-C_16:0 and n-C_16:1 increased up to 48.79 ± 0.92% and 37.37 ± 2.53%, respectively (Table 1). Biosynthesis of BCFAs could be restored to WT levels by introducing a single copy of fabH1 (Table 1). We also analysed the FA composition in ΔfabH2 and ΔfabH3, and no significant changes were found in either ΔfabH2 or ΔfabH3. Thus, FabH1 is likely responsible for BCFA biosynthesis, and FabH2 and FabH3 might not be involved in fatty acid biosynthesis.

FabH1 is required for biosynthesis of branched-chain DSF family signal components in Xoo

In this study, we confirmed that fabH1 deletion disrupted BCFA biosynthesis. Therefore, we believe that deletion of fabH1 would result in the defective branched-chain DSF family signal component biosynthesis. To test this hypothesis, DSF family signal production was examined in DSF high-yielding ΔrpfBC-derived strains using high-performance liquid chromatography (HPLC) (Zhou et al. 2017a). Both ΔrpfBCfabH1 and ΔrpfBCfabH1/fabH1 complementary strains were constructed. The results showed that deletion of fabH1 significantly decreased IDSF production. BDSF production was not affected, while CDSF and DSF signal components were not detectable in the mutant strain (Fig. 3). As expected, the introduction of a single copy of fabH1 restored CDSF, IDSF, and DSF production to WT levels (Fig. 3). These results indicate that precursors of DSF family signals are derived from intermediate metabolites of the FAS pathway, and FabH1 is the key enzyme in the production of branched-chain DSF family signals in Xoo.

Fig. 3

Diffusible signal factor (DSF)-family signal components produced by fabH1 deletion mutant. a DSF signal components produced by fabH1 deletion mutant that was grown in 50 mL NA medium at 30°C for 24 h were collected and detected by HPLC. b The relative DSF-family signal component levels at 24 h after inoculation. The amounts of signal molecules were calculated based on peak areas. Values are the means ± SDs from three independent experiments. The asterisks above the error bars indicate significant differences compared with the wild-type strain (***P < 0.001). All experiments were repeated three times with similar results

fabH1 is required for full virulence of Xoo

Previous studies showed that DSF promotes functions required for attachment and biofilm formation, and suppresses motility in Xoo (Torres et al. 2007; He and Zhang 2008; Zhu et al. 2011; Rai et al. 2012). Our current results showed that ΔfabH1 was defective in branched-chain DSF family signal component biosynthesis. To evaluate whether fabH1 contributes to Xoo virulence, a leaf-clipping virulence assay using susceptible rice plants (Oryza sativa L. japonica cv. Nipponbare) was conducted. The WT PXO99A strain caused an average lesion length of 10.7 ± 1.4 cm at 2 weeks post inoculation (Fig. 4). The ΔfabH1 mutant infection resulted in a significantly reduced lesion length in all infected leaves (5.6 ± 1.0 cm), and the complementary strain ΔfabH1/fabH1 maintained similar pathogenicity to the WT strain (Fig. 4). These results suggest that deletion of fabH1 resulted in the significantly decreased pathogenicity of Xoo on rice leaves.

Fig. 4

fabH1 is required for the full virulence of Xoo. a Lesion lengths on 5 weeks old rice leaves inoculated with the Xoo strains after 2 weeks. b Calculated lesion lengths were recorded from the Xoo strains inoculated rice leaves in a growth chamber. Values are the means ± SDs from three independent experiments. The asterisks above the error bars indicate significant differences compared with the wild-type strain (^**P < 0.01). All experiments were repeated three times with similar results

To understand why ΔfabH1 showed reduced pathogenicity on rice plants, we evaluated several pathogenicity-related virulence factors produced by Xoo strains. Previous reports have demonstrated that DSF positively regulates EPS production and biofilm formation (Slater et al. 2000; Torres et al. 2007). Production of EPS was therefore investigated, and the amounts of EPS produced by WT PXO99A, ΔfabH1, and the complementary strain ΔfabH1/fabH1 were 8.1 ± 0.28, 2.0 ± 0.1, and 7.9 ± 0.1 mg/mL, respectively (Fig. 5a). The amount of EPS produced by ΔfabH1 was significantly lower than that of the WT strain, but the amount of EPS produced by ΔfabH1/fabH1 was not statistically different from that produced by the WT strain (Fig. 5a).

Fig. 5

Phenotypic changes of EPS production, biofilm formation, motility, resistance to environmental stress in fabH1 deletion mutant. a Production of extracellular polysaccharide (EPS) in Xoo strains. b Biofilm production and its quantification of the Xoo strains on polystyrene tubes. c, d Typical bacterial halos, formed on semisolid 0.3% and 0.5% agar plates after incubation at 30°C for 2 days. e Relative activities of extracellular enzymes (cellulase and amylase) produced by Xoo strains on NA plates. f The Xoo strains were spotted onto NA plates containing increasing levels of NaCl, SDS, pH or H₂O₂. Values are the means ± SDs from three independent experiments. The asterisks above the error bars indicate significant differences compared with the wild-type strain (^**P < 0.01). All experiments were repeated three times with similar results

Next, the crystal violet assay was performed to monitor the formation of biofilm by the ΔfabH1 mutant. The results showed that deletion of fabH1 decreased biofilm production to 66.5% in the WT strain (Fig. 5b). DSF-mediated QS systems not only control the production of many virulence factors in Xoo, but are also involved in the regulation of Xoo motility (Rai et al. 2012). Thus, we evaluated the motility ability of Xoo strains in semisolid motility agar. The results showed that the diameter of the swimming zone of the WT strain was 2.06 ± 0.3 cm, while that of ΔfabH1 was only 0.81 ± 0.18 cm. In addition, the diameter of the swarming zone of the WT strain was 1.99 ± 0.11 cm, while the swarming zone of ΔfabH1 was only 0.88 ± 0.05 cm. These results indicated that deletion of fabH1 reduced Xoo swimming and swarming, and the complementary strain restored their motility ability to WT levels (Fig. 5c, d). The amylase and cellulase production assays showed that the ΔfabH1 mutant produced WT levels of these three extracellular enzymes, suggesting that it is not involved in regulating the production of extracellular enzymes in Xoo (Fig. 5e).

The growth of Xoo-derived strains on NA medium was investigated under various environmental challenges, including H₂O₂, the anionic surfactant sodium dodecyl sulphate (SDS), low pH, and high salt concentration. The results showed that 1.5% NaCl, 0.055 mM H₂O₂, 0.01% SDS, pH 5.5, and pH 8.0 differentially affected ΔfabH1 strain growth, indicating that ΔfabH1 was sensitive to many environmental stresses mentioned above. The fabH1 complementary strain grew well on medium containing 1.5% NaCl, 0.055 mM H₂O₂, or 0.01% SDS, but ΔfabH1/fabH1 only partially restored growth at pH 5.5 or pH 8.0 (Fig. 5f). Taken together, these findings indicate that fabH1 positively regulated EPS production, biofilm formation, motility, and resistance to environmental stresses.

Discussion

Xoo, the pathogen causing bacterial leaf blight, affects the quality and yield of rice worldwide (Nino-Liu et al. 2006; He et al. 2010). Pathogenicity of Xoo is extensively regulated by the DSF-based quorum sensing system (Zhou et al. 2017a), while DSF signals are initiated by the intermediates of the FAS system (Zhou et al. 2015), which is considered as an attractive target for antibacterial agent development (Heath et al. 2001). However, the fatty acid synthesis pathway in Xoo is poorly understood. FabH is a key enzyme in fatty acid synthesis, hence our study focused on elucidating the functions of FabH in Xoo.

Although three genes are annotated as FabH-coding candidates in the genome of Xoo, only fabH1 could rescue the growth of R. solanacearum ΔRsfabH (Fig. 1c), and enables the mutant to produce large quantities of BCFAs (Table 1). Moreover, FabH1 showed catalytic activity to various substrates in vitro, but the activities of FabH2 and FabH3 were not detected (Fig. 2). The ΔfabH1 mutant showed a similar growth phenotype as the WT strain (Additional file 1: Figure S2), but BCFA production was abolished (Table 1). Moreover, DSF was almost undetectable, but similar amounts of BDSFs were produced by deleting of fabH1 (Fig. 3). FabH1 is the key enzyme in the production of BCFAs and branched-chain DSF family signal components in Xoo (Fig. 6). Deletion of fabH1 also caused a significant decrease in pathogenicity on host plants, and the reduced production of virulence-related factors and resistance to stresses may collectively contribute to the attenuated pathogenicity (Figs. 4, 5).

Fig. 6

Schematic biosynthetic pathway of fatty acid and DSF-family signal components in Xoo. DSF, cis-11-methyl-2-dodecenoic acid; CDSF, cis, cis-11-methyldodeca-2,5-dienoic acid; IDSF, cis-10-methyl-2-dodecenoic acid; BDSF, cis-2-dodecenoic acid; FabH1, 3-ketoacyl-ACP synthase III; and RpfF, acyl-ACP thioesterase/dehydratase

Protein sequence alignment analysis showed that FabH2 and FabH3 shared only 27.1% and 28.6% identity with E. coli FabH, respectively, implying that FabH2 and FabH3 may not possess 3-ketoacyl ACP synthesase III activities. The hypothesis was confirmed in the complementary experiments and in vitro enzymatic analysis. The mutant ΔfabH2 and ΔfabH3 displayed similar fatty acid profiles, and showed a same growth pattern with wild-type strain, indicating that neither fabH2 nor fabH3 is essential for fatty acid synthesis and bacterial growth. However, further study is necessary to explore the functions of FabH2 and FabH3.

Compared with SCFAs, BCFAs contain an additional methyl group, and are essential in maintaining membrane fluidity in many Gram-positive bacteria (Choi et al. 2000; Li et al. 2005; Singh et al. 2009). However, the fabH1 deletion mutant did not produce BCFAs (Table 1), but its growth curve was similar to that of the WT strain (Additional file 1: Figure S2), implying that BCFAs were not essential for growth under the tested conditions. Similar results were also found in our previous study with the other phytopathogen Xanthomonas campestris pv. campestris (Xcc) (Yu et al. 2016). Only trace amounts of BCFAs (< 1%) were detected when XccfabH was replaced by E. coli fabH (Yu et al. 2016). We thus speculate that BCFAs are not important for the growth in the Xanthomonas genus. One plausible explanation is that both Xoo and Xcc produce a complex composition of FAs, including BCFAs, SCFAs, and UFAs, in which UFAs can replace the functions of BCFAs. The increased percentage of UFAs in the fabH1 deletion mutant is consistent with this hypothesis. However, whether UFAs in Xoo are essential for growth is not clear, and our group is attempting to answer this question by exploring the biosynthetic mechanism of UFAs in Xoo.

Many studies have showed that KAS III is the sole enzyme that catalyses the condensation step in the initiation of the FAS II system (Lin and Cronan 2011; Mao et al. 2015), hence the fabH gene is essential for bacterial growth (Lai and Cronan 2003; Parsons and Rock 2013), and is a promising target for antibacterial drug discovery (Marrakchi et al. 2002). However, our study indicated that the ΔfabH1 mutant grew well on nutrition-rich media (Additional file 1: Figure S2), indicating that fabH1 is not essential for bacterial growth in Xoo. The opposite result was found in studies on Xcc, in which no fabH deletion mutant was isolated, but the E. coli fabH replacement strain Xcc fabH::EcfabH was constructed, suggesting that Xcc fabH is an essential gene for Xcc growth (Yu et al. 2016). One possible reason for this observation is that another 3-ketoacyl-ACP synthase (FabB or FabF) may partially compensate for the activity of FabH in the deletion mutant. Indeed, several reports showed that overproduction of FabB or FabF can bypass the initiation step in FAS II (Morgan-Kiss and Cronan 2008; Meng et al. 2018). However, this was not the case in Xoo, because the ΔfabH1 strain grew comparably well to the WT strain, and overexpression of Xoo fabB or fabF could not rescue the growth of R. solanacearum ΔfabH (Additional file 1: Figure S3). Based on these results, we speculate that a novel mechanism initiates FAS in Xoo, and this is currently being explored in our group.

Conclusions

In this study, we determined that FabH1 is the key enzyme in the production of BCFAs and branched-chain DSF family signal components in Xoo, and provided strong evidence that DSF family signal molecules are synthesised via the FAS pathway in Xanthomonas species. Pathogenicity analysis showed that loss of fabH1 caused a significant decrease in the virulence of Xoo on rice leaves. Genetic and phenotypic evidence showed that fabH1 plays a key role in multiple Xoo pathogenicity-related functions, including EPS production, biofilm formation, motility, and resistance to environmental stresses.

Methods

Materials

Malonyl-CoA, acetyl-CoA, fatty acids, NADPH, NADH, and antibiotics were purchased from Sigma-Aldrich. Takara Biotechnology Co., Ltd. provided molecular biology reagents. Novagen provided the pET-28(b) vectors. Ni–NTA agarose columns were purchased from Invitrogen. Bio-Rad Co. provided the Quick Start Bradford dye reagent. The oligonucleotide primers were synthesized by Sangon Biotech (Shanghai) Co., Ltd.

Bacterial strains, plasmids, and bacterial growth conditions

The bacterial strains and plasmids are listed in Additional file 2: Table S1. E. coli were grown at 37°C in Luria–Bertani medium (yeast extract, 5 g/L; peptone, 10 g/L; NaCl, 10 g/L; pH 7.0) (Hanahan 1983). Ralstonia solanacearum strains were grown at 30°C in BG medium (bacto peptone, 10 g/L; yeast extract, 1 g/L; casamino acids, 1 g/L; glucose, 5 g/L; pH 7.0) (Cheng et al. 2012). Xoo strains were grown at 30°C in nutrient agar (NA) medium (beef extract, 3 g/L; yeast extract, 1 g/L; peptone, 5 g/L; sucrose, 10 g/L; pH 7.0) (Zhang et al. 2022). The composition of XOLN medium: 0.7 g/L K₂HPO₄, 0.2 g/L KH₂PO₄, 1 g/L (NH₄)₂SO₄, 0.1 g/L MgCl₂, 0.01 g/L FeSO₄, 0.001 g/L MnCl₂, 0.0625% tryptone, 0.0625% yeast extract, and 2 g/L sucrose (He et al. 2010). The antibiotics and inducers were added at the concentrations of 30 μg/mL kanamycin sulfate, 30 μg/mL gentamicin, or 30 μg/mL chloramphenicol. Isopropyl-β-D-thiogalactoside (IPTG) was used at a final concentration of 1 mM. Octanoic acid was used at a final concentration of 1 mM.

Complementation of the R. solanacearum fabH deletion strain

The Xoo fabHs genes were amplified from genomic DNA of X. oryzae pv. oryzae PXO99A using primers listed in Additional file 2: Table S2. Then the amplified fragment was purified, and inserted into the multicopy plasmid pSRK-Gm (Khan et al. 2008) to generate pMF1 (fabH1), pMF2 (fabH2), and pMF3 (fabH3), respectively. These plasmids were confirmed by nucleotide sequencing performed at Sangon Biotech (Shanghai) Co., Ltd. The derivatives of E. coli strain S17-1 carrying these plasmids or empty vector plasmids was mated with R. solanacearum fabH deletion strain ΔRsfabH (Mao et al. 2015) on BG plates with octanoic acid for 48 h at 30°C. The cells were suspended in BG medium and appropriate dilutions were spread onto BG plates (with octanoic acid) containing chloramphenicol (to select against the donor strain) and gentamicin. The transformed strains were inoculated onto the BG plates with or without octanoic acid, and growth was determined after 2 days incubation at 30°C.

Expression and purification of proteins

pET-28(b) carrying Xoo fabHs genes were transformed into BL21 (DE3) cells. The recombinant proteins were then purified using nickel-ion affinity column (Qiagen) following the standard protocol. The purified proteins were examined with SDS-PAGE electrophoresis. The E. coli proteins FabD, FabH, FabB, FabG, FabZ, FabI, holo-ACP, and Vibrio harveyi AasS proteins were purified as described previously (Zhu et al. 2010).

Assays of 3-ketoacyl-ACP synthase III activity in vitro

The activities of KAS III in the initiation of fatty acid synthesis were assessed with reaction mixtures containing 0.1 M sodium phosphate (pH 7.0), 1 mM β-mercaptoethanol, 0.1 μg each of E. coli FabD, FabG, FabA, and FabI, 50 μM holo-ACP, 50 μM NADPH, 50 μM NADH, 100 μM malonyl-CoA, and 100 μM acetyl-CoA (C_2:0-CoA to C_12:0-CoA, isobutyryl-CoA, and isovaleryl-CoA) in a final volume of 40 μL. The reactions were initiated by the addition of FabHs to the mixture, followed by incubation for 1 h at 37°C. The resulting products were distinguished with conformationally sensitive gel electrophoresis on 17.5% polyacrylamide gels containing 0.5 M urea optimized for separation (Mao et al. 2015). The gel was stained with InstantBlue Coomassie Protein Stain (Abcam).

Construction of in-frame deletion mutants and complementation

X. oryzae pv. oryzae PXO99A was used as the parental strain for the generation of deletion mutants, as described previously, and the primers used were listed in Additional file 2: Table S2. For single-copy complementation of ΔfabH1, the coding region of fabH1 along with its promoter region was amplified by PCR and was cloned in a versatile mini-Tn7T delivery vector mini pUC18-Tn7T-Gm plasmid (Choi and Schweizer 2006). The resultant construction was transferred into ΔfabH1 mutant by electroporation with the helper plasmid pTNS2 and the complementation strain was selected as described previously (Choi and Schweizer 2006).

Analysis of fatty acid profile

To determine fatty acid composition, the strains were grown in BG medium or NA medium to an OD₆₀₀ of 0.6. Cultures were collected by centrifugation and were washed three times with water. Fatty acid methyl esters were generated and extracted and the profile was analyzed by GC–MS as previously described (Yu et al. 2016; Dong et al. 2021).

Pathogenicity assays

The pathogenicity of Xoo strains was tested on rice cultivars (Oryza sativa L. japonica cv. Nipponbare) by leaf-clipping, as previously described (Ray et al. 2000). Briefly, the bacterial were grown in NA medium at 30°C for 2 days, and the cells were washed and resuspended in sterile PBS buffer to an OD₆₀₀ of 0.1. The cells were inoculated on the leaves by leaf-clipping. The lesion length was measured at 2 weeks post inoculation. Twenty leaves for each tested strain were inoculated. Each strain was tested with at least three separated experiments.

Extraction and purification of DSF-family signal components from Xoo culture

The method for extraction and purification was described previously (Zhou et al. 2015; Zhou et al. 2017a). To quantify DSF-family signal components production in the culture of the Xoo strains by High-performance liquid chromatography (HPLC), Xoo strains were cultured in liquid medium for 24 h and 50 mL of the supernatant was collected. The crude ethyl acetate extract was filtered through a 0.45-µm Minisart filter and then concentrated to 0.1 mL for HPLC assays. The extracted matters (20 µL) were examined with a C18 reverse-phase HPLC column (4.6 mm × 250 mm, Agilent Technologies), and were eluted with water: methanol (23:77 v/v, 0.1% formic acid) at a flow rate of 1 mL/min in a Shimadzu prominence LC-20AT system (Shimadzu International Trading Co. Limited) with a UV220 detector. All experiments were repeated three times with similar results.

Motility and biofilm formation assays

The swimming and swarming motility assays were performed on a semi-solid plate containing 0.3% agarose and 0.5% agarose, respectively (Antar et al. 2020). The bacteria of each strain were cultured in NA medium in a shaking incubator at 30°C for 2 days, and the cells were washed and resuspended in water to a final OD₆₀₀ of 2.0. Aliquots (2 μL) of bacteria were spotted onto the swarming plates, and were incubated at 30°C for 3 days, respectively. The assays were repeated at least three times. Biofilm formation assay was described previously (Li and Wang 2012), with slight modifications. The assay was performed by incubating the diluted bacteria inoculated into the borosilicate glass for 7 days at 30°C without shaking. Non-adherent bacteria and medium were removed and washed three times. The attached biofilm was incubated with 0.1% (w/v) crystal violet solution for 30 min at room temperature, and thoroughly washed twice with water and then air-dried. The crystal violet bound cells was solubilized in 90% ethanol and quantified by measuring absorption at 590 nm. The assays were repeated at least three times.

Measurement of the extracellular enzyme activity and EPS production

The activities of extracellular cellulase and amylase were determined on NA plates containing 0.5% (w/v) carboxymethylcellulose (CMC) (for cellulase), 0.2% xylan (for xylanase), and 0.1% (w/v) starch (for amylase), respectively (Wei et al. 2007). Two μL of each Xoo strain culture (OD₆₀₀≈ 2.0) was incubated at 30°C for 3 days. The CMC plates were stained with 1% Congo Red solution. The xylanase plates and starch plates were stained with I₂-KI (I₂ at 0.08 mol/L and KI at 3.2 mol/L) solution. The zones of clearance around the spots due to the degradation of the substrate were photographed. The relative activities of the enzymes were indicated by the diameters of the clear zones. Three plates were inoculated in each experiment, and each experiment was repeated three times.

To measure EPS production, each Xoo strain culture (1 mL, OD₆₀₀ ≈ 2.0) was used to inoculate 50 mL of NA liquid medium containing 4% glucose at 30°C for 4 days, with shaking at 180 rpm (Long et al. 2018). EPS was precipitated from the culture supernatant by adding of three volumes of 100% cold ethanol. The precipitated EPS were pelleted by centrifugation at 12,000 g, 4°C for 15 min, then air dried and weighed. Three flasks were inoculated in each experiment and were repeated three times.

Effects of environmental factors on the growth of Xoo

The bacteria of each strain were cultured in NA medium in a shaking incubator at 30°C for 2 days, and the cells were washed and resuspended in water to a final OD₆₀₀ of 2.0. Aliquots (5 μL) of bacteria were spotted onto NA plates containing increasing levels of NaCl, SDS, pH, or H₂O₂. All plates were incubated at 30°C for 3 days.

Statistics and reproducibility assays

The experimental datasets were subjected to analyses of variance using GraphPad Prism 8.0. The significance of the treatment effects was determined by the F value (P = 0.05). If a significant F value was obtained, separation of means was accomplished by Fisher’s protected least significant difference at P ≤ 0.05.