A transferred regulator that contributes to Xanthomonas oryzae pv.oryzicola oxidative stress adaptation and virulence by regulating the expression of cytochrome bd oxidase genes

WANG Pei-hong ,WANG Sai ,NlE Wen-han ,WU Yan ,lftikhar AHMAD, ,Ayizekeranmu YlMlNG,HUANG Jin,CHEN Gong-you,ZHU Bo

1 Shanghai Yangtze River Delta Eco-Environmental Change and Management Observation and Research Station/Shanghai Cooperative Innovation Center of Agri-Seeds,School of Agriculture and Biology,Shanghai Jiao Tong University,Shanghai 200240,P.R.China

2 Department of Environmental Sciences,COMSATS University Islamabad,Vehari-Campus,Vehari 61100,Pakistan

AbstractHorizontal gene transfer (HGT) has been well documented as a driving force in the evolution of bacteria. It has been shown that a horizontally acquired gene,xoc_2868,involved in the global response against oxidative stress and pathogenicity of Xanthomonas oryzae pv.oryzicola strain BLS256. However,as a transcriptional factor (TF),the regulatory mechanism of XOC_2868 has not yet been revealed. Here,evolutionary analysis suggested XOC_2868 might be co-transferred with its physically proximate downstream genes from a Burkholderiaceae ancestor. Interestingly,RNA-seq data of wild-type (BLS256) and Δxoc_2868 strains under oxidative stress showed that XOC_2868 did not regulate the expression of its adjacent genes,but remarkably influenced the expression of several genes involved in the extracellular polysaccharide (EPS) production and xanthan biosynthesis. Chromatin immunoprecipitation-sequence(ChIP-seq) combined with transcriptome analysis revealed that XOC_2868 directly regulates a cydAB operon,encoding two subunits of cytochrome bd oxidase and involved in redox balance. Consistent with Δxoc_2868 strain,cydA-and cydAB-knockout mutants also showed a higher sensitivity to H2O2 along with a reduced bacterial virulence compared with the wild-type strain. Overall,our findings raise the possibility of regulatory circuit evolution shaped by HGT and driven by selection and reveal a novel regulatory pathway that regulates the expression of cytochrome bd oxidase and thus contributes to the virulence of BLS256.

Keywords:HGT,transcriptional factor,Xanthomonas oryzae pv.oryzicola,oxidative stress adaptation and virulence

1.lntroduction

Horizontal gene transfer (HGT),the acquisition of genetic materials across species boundaries,is a prevalent phenomenon across the biosphere. Numerous examples showed that HGT can serve as a major driver of the evolution in bacteria to provide new capabilities and traits in various ecological niches (Ochmanet al.2000;Bacciuet al.2004;Goyal 2019). HGT is also the main mechanism for spreading of genes associated with virulence and antibiotic resistance,contributing to pathogenic bacteria to successfully colonize hosts and express virulence-related genes (Lindsay 2010;Liuet al.2014). As the causative agent of bacterial leaf streak,Xanthomonas oryzaepv.oryzicola(Xoc) would be exposed to a variety of harmful conditions in the host environment and immune system during the invasion of rice. In the study of the molecular mechanisms ofXocresponding to oxidative stress during invasion,over 30%of the coding genes in theXocrepresentative strain BLS256 genome were identified as hypothetical genes(Fanget al.2019). Among them,xoc_2868gene was proven to participate in the oxidative stress response and pathogenesis (Fanget al.2019). Interestingly,xoc_2868was a horizontally acquired gene encoding a transcription factor and located adjacent to many genes encoding transposase,highlighting the importance of considering its contribution to the oxidative stress adaption and the evolution of bacterial regulatory networks.

From an evolutionary point of view,HGT optimize existing abilitiesviashaping bacterial genomes.Specifically,the adaptation to a dynamic environment demands proper regulation of gene expression in bacteria(Ishiiet al.2007;Priceet al.2008;Perez and Groisman 2009),which is realized by transcription factors (TFs)that sense the cellular environments and bind to specific DNA sequences or motifs (Priceet al.2008;Ali and Seshasayee 2020). Therefore,to unveil its regulatory potential,it is important to locate TFs binding sites or motifs where TFs interact with DNA.

Due to the rich genome sequence data and the wide application of computational analyses,the regulatory features of horizontally acquired DNA has attracted an increasing attention (Babu and Teichmann 2003;Gelfand 2006;Lozada-Chávezet al.2006;Perez and Groisman 2009;Ali and Seshasayee 2020). In the previous literature,it has been revealed that most TFs (including putative regulators) regulate adjacent operons inEscherichia coli(Hershberget al.2005;Priceet al.2008).The proximity of these regulators to their target genes were interpreted as co-transfer through HTG mechanism and maintained by repeated transfers (Priceet al.2008). However,there is also an evidence that TFs do not usually co-evolve with their target genes in bacteria,because TFs evolve much faster than the target genes(TGs) (Babuet al.2004,2006;Lozada-Chávezet al.2006) and show less conservation than their regulated genes (Ali and Seshasayee 2020).

The genes are functionless without expression,therefore,it is interesting to unveil the proper response of the foreign XOC_2868 under oxidative stress. In this study,we address the role of XOC_2868 in adaptation evolution ofXocviaunderlying its regulatory interaction with its TGs,including figuring out its binding sites and functional analysis of the target genes.

2.Materials and methods

2.1.Phylogenetic analysis

For homologs searching of interested genes,their nucleotide sequences and protein sequences were compared against GenBank non-redundant nucleotide(NT) database and RefSeq non-redundant protein sequence (NR) database,by using the BLASTN and BLASTP with the E value cutoff of 10-4. The real homologue genes were selected based on the query coverage and identity values over 80%. The multiple sequences were aligned online with MAFFT (http://mafft.cbrc.jp/alignment/server/). Maximum likelihood were generated in the IQTree online tool (http://iqtree.cibiv.univie.ac.at/?user=guest&jobid=191204085022),and all phylogenetic trees were visualized with iTOL (http://itol.embl.de/itol.cgi). We performed 1 000 bootstraps to gain branch support values.

2.2.Bacterial strains,culture conditions,plasmids and primers

The bacterial strains and plasmids used in this study are listed in Appendix A. Primers are listed in Appendix B.Escherichia colistrains were cultivated at 37°C on Luria-Bertani (LB) agar plates or LB medium.Xocstrains were cultured at 28°C on nutrient agar (NA) plates,NA without sucrose (NAN) plates,NA with 10% sucrose (NAS) plates and in nutrient broth (NB) medium. When required,kanamycin (Km) was added in growth media with the final concentration of 25 μg mL-1. DH5aand BL21 (DE3)in this work were used for plasmid construction and XOC_2868 purification,respectively.

2.3.Construction of in-frame deletion mutants

Homologous double-crossover recombination was used to construct in-frame deletion mutations in this study. In brief,two fragments flanking the left and right of target genes were amplified,sequenced,and ligated into the pKMS1 suicide vectors atBamHI andSalI sites (Guoet al.2012). These recombinant plasmids were electroporated into BLS256 and screened on NAN plates supplemented with Km to obtain single homologous crossover colonies.NAS plates were then used for selecting the second double-crossover recombinants. Sucrose resistant and Km-sensitive colonies were further verified by PCR amplification and sequencing.

For chromatin immunoprecipitation sequencing (ChIPseq),the His6tag coding sequence was fused at the C-terminal ofxoc_2868 viaPCR. Then the chromosomalxoc_2868copy in BLS256 was substituted with the above DNA insertion by double crossover through the suicide vector pKMS1. All the primers used to construct the above strains are listed in Appendix B.

2.4.Oxidative stress treatments and total RNA preparation

H2O2resistance assays were performed as described previously with some modifications (Upadhyaet al.2013).Xocstrains (WT and Δxoc_2868) were grown overnight to an OD600of 1.0 in fresh NB medium,treated with 0.1 mmol L-1H2O2the next morning in a 28°C shaking incubator for 15 min. After collection by centrifugation at 4°C,each cell pellet was washed with cold PBS and centrifugated,repeated twice,and then immediately subjected to RNA purification using the RNeasy Protect Bacteria Mini Kit(Qiagen,Germany). Two biological replicates were harvested for each strain.

2.5.RNA-seq and data analysis

A total of 10 μg total RNA per sample was treated with the MICROBExpress Bacterial mRNA Enrichment Kit (Ambion,USA) and RiboMinus? Transcriptome Isolation Kit(Bacteria) (Invitrogen,USA) following the manufacturer’s instructions. cDNA was generated according to instructions given in SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen,USA). The Illumina Paired End Sample Prep Kit was used for RNA-seq library creation according to the manufacturer’s instructions as follows:Fragmented cDNA was end-repaired,ligated to Illumina adaptors,and amplified by 18 cycles of PCR. Pairedend 250-bp reads were generated by high-throughput sequencing with the Illumina Hiseq2500 Genome Analyzer Instrument (USA).

After removing the low quality reads and adaptors,RNA-seq reads were aligned to the corresponding BLS256 genome using Tophat 2.0.7 (Trapnellet al.2009),allowing for a maximum of two mismatches. If reads mapped to more than one location,only the one showing the highest score was kept. Reads mapping to rRNA and tRNA regions were removed from further analysis. After getting the reads number from every sample,edgeR with TMM normalization method was used to determine the differentially expressed genes (DEGs). Significantly differentially expressed genes (FDR value＜0.01) were selected for further analysis. Cluster 3.0 and Treeview 1.1.6 were used to generate the heatmap cluster based on the RPKM values (de Hoonet al.2004;Saldanha 2004).

2.6.ChlP-seq analysis

Stationary-phase cultures ofXocstrains were grown overnight to an OD600of 1.0 in fresh NB medium,treated with 0.1 mmol L-1H2O2the next morning in a 28°C shaking incubator for 15 min. Subsequently,treated samples were cross-linked (1% formaldehyde,20 min,RT),quenched (2 mol L-1glycine,10 min),washed three times with ice cold PBS,and collected by centrifugation(4°C,5 min). Two biological replicates were flash-frozen with liquid nitrogen,stored at -80°C and transported on dry ice to Wuhan IGENEBOOK Biotechnology Co.,Ltd.,China,for ChIP-seq libraries generation and sequencing.

According to ENCODE guidelines (Landtet al.2012),the DNA was sonicated to generate fragments,which were size-selected by SPRI between 250-350 bp. After end repair and adaptor ligation,the enriched DNA was amplified with 15 PCR cycles prior to sequencing. Illumina ChIP-seq libraries were conducted following validation on the Bioanalyzer 2100 (Agilent,USA) and Qubit fluorometer(Invitrogen,Carlsbad,CA,USA). Primary analysis of the ChIP-seq raw reads was carried out as described previously (Landtet al.2012). Clean reads were aligned against theXocstrain BLS256 genome. Peak calling was conducted by MACS2 (version 2.1.1.20160309) with default parameters (bandwidth,300 bp;model fold,5-50;q-value,0.05). The motif identification was done using the MEME Program on the set of sequences defined by the XOC_2868-binding regions. To reveal potential roles of genes,the EasyGO Gene Ontology (GO) enrichment analysis tool was employed to perform GO enrichment analysis and cluster Profiler in R package was used for KEGG enrichment analysis.

2.7.Electrophoretic mobility shift assay (EMSA)

For EMSA,XOC_2868-His6recombinant protein was expressed by constructing the corresponding expression vectorviacloningxoc_2868into pET30aatEcoRI andXhoI sites and purified fromE.coliBL21following the protocalof the BeyoGoldTMHis-tag purification Resin(Beyotime,China). DNA probes needed in this assay were gained by PCR amplication with the primer pair (Site F/R,listed in Appendix B). The DNA probe was labelled using EMSA Probe Biotin Labeling Kit (Beyotime). The protein-binding reactions were performed according to the manufacture’s instruction (Chemiluminescent EMSA Kit,Beyotime). Finally,the signals were detected using X-ray film in a dark room.

2.8.Quantitative real-time PCR (RT-qPCR)

The bacteria samples used for RT-qPCR were the same as those used for RNA-seq. Total RNA was purified using EasyPure RNA Kit (Transgen Biotech,Beijing,China). cDNA was synthesized from 1 μg total RNA with TransScript II One-Step gDNA Removal and cDNA Synthesis SuperMix (Transgen Biotech,Beijing,China).RT-qPCR assay was conducted with ABI 7500 Quantitative PCR System (Applied Biosystems,Foster City,CA) using transStart Tip Green qPCR SuperMix (TransGen Biotech,Beijing,China). The gene-specific primers are listed in Appendix B and the relative expression of target genes were normalized togyrBusing the 2?ΔΔCtmethod.

2.9.Bacterial growth analysis under H2O2 stress

In this assay,all strains were grown overnight to an OD600of 1.0 in fresh NB medium,diluted to OD600of～0.1 the next morning and then inoculated with a fresh NB medium without or with the addition of 0.075 mmol L-1H2O2. Bacterial growth was automatically determined on Bioscreen C (Labsystem,Helsinki,Finland) at 420-580 nm,at regular intervals of 15 min for 40 h with continuous shaking at 28°C.

2.10.Pathogenicity assays

Oryza sativaL.cultivar ‘Yuanfengzao’ plants,susceptible toXoc,were grown in our field stations (Shanghai Jiao Tong University) for 1 month. To prepare bacterial inoculum,Xanthomonascells were grown overnight in NB broth,harvested at the exponential phase of growth by centrifugation,and resuspended in sterile water to reach OD600of 0.8 after washing twice. The pathogenicity ofXocstrains was examined by using leaf piercing and evaluating water-soaked symptoms based on lesion length measurement at 14 days after inoculation (DAI).Ten fully expanded leaves were inoculated for each independent experiment.

3.Results

3.1.Physically proximate genes of xoc_2868 were also transferred from a Burkholderiaceae ancestor

As shown in Fig.1-A,the other two hypothetical genes,xoc_2866andxoc_2867,present betweenxoc_2868and many transposase encoding genes. Notably,a comprehensive BLAST sequence analysis showed that both of them were only present inXocand absent in otherXanthomonas. Based on the phylogenetic analysis with high bootstrapping support (Fig.1-B and C),it can be further inferred thatxoc_2866andxoc_2867inXocmight transfer from aBurkholderiaceaeancestor over the course of evolutionary history,which is consistent with the analysis ofxoc_2868in the previous study (Fanget al.2019).

Fig.1 Evolution of the physically proximate genes of XOC_2868. A,genetic structure of the co-linearly arranged genes adjacent to XOC_2868. Genes are represented by boxes,black boxes corresponding to genes coding for unknown proteins and grey boxes indicating genes coding for transposase. B and C,phylogeny of the xoc_2867 and xoc_2866. The phylogenetic tree shown was calculated using the maximum likelihood (ML) program in the IQTree online tool and visualized with iTOL. Only the ML values≥50% were shown. Bar=0.1 substitution per site.

3.2.Transcriptome analysis of Δxoc_2868 strain and wild-type strain under oxidative stress

To explore the regulatory interaction between XOC_2868 and the possible TGs,we next carried out RNA-seq analysis of Δxoc_2868strain and wild-type strain(BLS256) under the treatment of H2O2,each with two samples. The comparative analysis indicated thatxoc_2866andxoc_2867were not directly or indirectly regulated by XOC_2868 under oxidative stress (data shown in Appendix C). In addition,five genes associated with fructose and glucose metabolism were downregulated at least 4-fold in mutant strain (Appendix C),These genes includedoprB(xoc_1972),xoc_1973(PTS fructose transporter subunit IIBC),xoc_1974(1-phosphofructokinase),mtp(xoc_1975) andxoc_1976(LacI family transcriptional regulator). In addition,their expression profile was examined through RT-qPCR with the specific primers (Appendix B). Consistent with the RNA-seq results,all of them were indeed significantly downregulated (Fig.2-A).

Fig.2 Integrated analysis and validation of the results of RNA-seq and ChIP-seq. A,verification of RNA-seq results by RT-qPCR.Fold change values are shown and normalized to gyrB. Error bars represent the standard deviations (SD) of three technical replicates. B,integrated analysis of differentially expressed genes (DEGs) in RNA-seq and potential target genes (TGs) in ChIP-seq. C,potential XOC_2868 binding motifs were identified by MEME from the ChIP-seq peak regions and present within protein-coding sequences of xoc_1426. D,EMSA experiments were conducted in the presence of purified XOC_2868-His6 and the corresponding probe listed in Appendix C. Labeled cydAB_BS,free probe labeled with biotin;lane 1,probe only;lanes 2 and 3,probe with purified XOC_2868-His6.

3.3.Genome-wide transcriptional and binding profiles of XOC_2868

To identify genes directly regulated by the XOC_2868,we performed ChIP-seq experiment using a recombinant strain,a C-terminal His6taggedxoc_2868in BLS256.

The ChIP-seq results were shown in Appendix D,including the DNA binding sites of XOC_2868 throughout the genome,loci and enrichment fold. The integrated analysis of RNA-seq data and ChIP-seq data was conducted to identify direct regulatory targets of XOC_2868. As shown in the Fig.2-B,a total of 229 and 479 genes were differentially regulated or specially bound by XOC_2868 from the results of RNA-seq and ChIP-seq,respectively. In addition,16 potential TGs of XOC_2868 shared by them and three of those genes are annotated genes (shown in Appendix E). Of note,no significant enrichment of DNA binding sites was found for all the above down-regulated genes except foroprB. TheoprBmentioned earlierencodes an outer membrane protein B,which functions as a carbohydrate-selective porin inXoo(Baeet al.2018),Xcc(Ficarraet al.2017) andPseudomonasspecies(Wylie and Worobec 1995;Chevalieret al.2017). The other two genes arecydABandmutM,which also have been demonstrated to be essential for bacterial cellular survival (Endleyet al.2001;Landova and Silhan 2020).ThemutMgene encodes a bacterial DNA glycosylase that is important in repairing the oxidized DNA (Landova and Silhan 2020).

ThecydABoperon encodes two subunits (CydA and CydB) of cytochromebdubiquinol oxidase that is one kind of the terminal oxidases and is widespread in bacteria and involved in redox balance (Degli Espostiet al.2015;Safarianet al.2016). RT-qPCR was firstly utilized out to examine the regulatory role of XOC_2868 forxoc_1426andxoc_1427during infection. Compared with the wild type strain BLS256,obvious decreased expression ofxoc_1426andxoc_1427were observed in the Δxoc_2868mutant strain under H2O2treatment (Fig.2-A),which was consistent with the RNA-seq data and confirmed the reliability of sequencing results again.

The MEME analysis on the ChIP-seq data revealed the potential XOC_2868 binding motif (5′-TTCGACAT-3′;Fig.2-C),which was within the 3′ untranslated region(UTR) ofoprB(Appendix E) and within the coding sequences ofcydAB(Fig.2-C). EMSA was thus only carried out to verify whether XOC_2868 binds tocydAB.Band shifts were observed with using the purified XOC_2868-His6and the probes containing the potential binding motif (Fig.2-D). Taken together,the above results revealed that XOC_2868 directly controls the expression ofxoc_1426andxoc_1427.

3.4.Roles of XOC_2868 regulated genes under oxidative stress and in the bacterial virulence

The mutant strains ΔcydA(with the deletion ofxoc_1426)and ΔcydAB(with the deletion ofxoc_1426andxoc_1427) were generatedviadouble crossover using the suicide vector pKMS1 to evaluate the contributions of the XOC_2868 regulated genes under oxidative stress. The growth curves of WT and three mutant strains (Δxoc_2868,ΔcydAand ΔcydAB) under oxidative stresses were measured in four repeats and representative curves were shown in Fig.3. Obviously,all the strains showed similar growth patterns in the absence of H2O2in nutrient broth medium. When cultures were added with 0.075 mmol L-1H2O2,these mutant strains displayed different lags compared to the wild type,curves corresponding to the minimum lag time of 2 h (ΔcydA),6 h (Δxoc_2868),and 18 h (ΔcydAB),respectively. This indicated that the lack ofcydAorcydABalso increased the sensitivity of BLS256 to H2O2. It led us speculated that they might also play a role in pathogenesis of BLS256. Not surprisingly,these three knockout mutants in this study showed decreased pathogenicity when compared with the wild-type strain(Fig.4).

Fig.3 Comparison of the H2O2 resistance of wild type (Xanthomonas oryzae pv.oryzicola,Xoc) and knockout mutant strains in nutrient broth (NB) medium. All strains were grown overnight to an OD600 of 1.0 in fresh NB medium,diluted to OD600 of～0.1 the next morning and then inoculated with a fresh NB medium without or with the addition of 0.075 mmol L-1 H2O2. Growth was monitored at 420-580 nm during 0-40 h using a Bioscreen C (growth curves) maintained at 28°C with continuous shaking. The data shown in the figure are the mean OD values (quadruplicate). Shaded areas show the standard error of the means.

Fig.4 Pathogenicity assays for BLS256,Δxoc_2868,ΔcydA,and ΔcydAB. A,water-soaked symptoms of inoculated leaves(Yuanfengzao,one-month-old) at 14 days after inoculation (DAI). B,box plots display the respective distributions of lesion lengths on the adult inoculated with the Xanthomonas oryzae pv.oryzicola (Xoc) strains. Error bars represent the SD calculated from 10 replicate measurements. Statistical analysis was done by a GraphPad Prism 8.0,significant values are marked with asterisks (＊＊＊,P＜0.001).

4.Discussion

In light of the above-mentioned fact that most bacterial regulators are adjacent to operons that they regulate,we combined a comprehensive BLAST sequence analysis and the phylogenetic analysis to characterize genesxoc_2866andxoc_2867that lie downstream ofxoc_2868.Based on their physically-linked character and the same evolutionary source,we hypothesized that the three genes were likely acquired as an unit and XOC_2868 might be a neighbor regulator. However,RNA-seq analysis suggested thatxoc_2866andxoc_2867were not the regulatory genes ofxoc_2868under H2O2treatment.One possible explanation is that these regulatory proteins can act at distant sites,and the close proximity of TFs and their TGs is not essential for the regulatory mechanism(Lawrence and Roth 1996). However,an alternative explanation is that bacterial TFs have a tendency to evolve faster than their regulated genes (Babuet al.2004,2006;Lozada-Chávezet al.2006). Thus,we reasoned thatxoc_2868,xoc_2866andxoc_2867might diverge inXocafter the co-transfer eventviaHGT,enabling it to adapt some specific niches. The absence ofxoc_2868,xoc_2866andxoc_2867in its closely related strains,Xoo,can also be a circumstantial evidence to support our inference.

Notably,the regulatory genes involved in fructose and glucose metabolism have previously been reported in response to the host environment stimuli (Andréet al.2005;Kimet al.2016). For example,TheXanthomonas citrissp.citri(Xcc) lacking OprB andXanthomonas oryzaepv.oryzae(Xoo) lacking OprB are both also highly susceptible to H2O2(Ficarraet al.2017;Baeet al.2018).Further,the OprB family is also known to be involved in the biosynthesis of the extracellular polysaccharide(EPS) xanthan (Slateret al.2000;Vojnovet al.2001),probably by modulating the rate of glucose transport,which has knock-on effects on biofilm formation and virulence withinXoo(Baeet al.2018) andXcc(Ficarraet al.2017). Previous studies have shown that BLS256 lacking XOC_2868 showed enhanced sensitivity to H2O2(Fanget al.2019),which is also verified in this study. Accordingly,we hypothesized thatoprBin BLS256 is also involved in the tolerance to H2O2. In addition,the XOC_2868 mutant also showed obviously reduced virulence. Thus,it can be postulated that the downregulation of these genes involved in fructose and glucose metabolism caused by the loss of XOC_2868 in axoc_2868mutant,which,in return,influenced xanthan production and impaired H2O2tolerance and pathogenicity,although this needs to be investigated further. At this point,it was clear that XOC_2868 was indeed functional in cellular survival. Collectively,these findings indicate that XOC_2868 suggest the existence of novel and potential binding sites for XOC_2868 after transferring under oxidative stress.

The genes directly regulated by the XOC_2868 were identified by the ChIP-seq assay and their interaction were substantiated by EMSA. In addition,we examined the roles of the XOC_2868 regulated genes,cydAB,in the oxidative stress response and pathogenicity. In this study,the deletion ofcydAorcydABincreased the sensitivity of BLS256 to H2O2along with decreased pathogenicity.Consistently,mutants defective in cytochromebdalso display a hypersensitivity to H2O2inE.coli(Lindqvistet al.2000),Azotobacter vinelandii(Edwardset al.2000) andPorphyromonas gingivalis(Leclercet al.2015). Morover,disruption ofMycobacterium tuberculosiscytochromecmaturation (CCM) would induce the overexpression of cytochromebdand the hyper-resistance to H2O2(Smallet al.2013). Along the same line,exposure to external H2O2could increase the expression of genes encoding cytochromebdinE.coli(Lindqvistet al.2000). Those observations point to a crucial role for cytochromebdin protecting bacteria cells from oxidative stress. This possibility may be due to the extraordinarily high oxygen affinity of CydAB (Borisovet al.2011). This oxidase is employed for reducing molecular oxygen or dioxygen to water (Safarianet al.2016) and thus for respiratory protection (Poole and Hill 1997) and even the colonization of O2-limited niches by pathogenic bacteria (Endleyet al.2001;Baughn and Malamy 2004;Shiet al.2005). This hypersensitivity of mutants may be due to the inability to remove superoxide anion and H2O2in the periplasm(Borisovet al.2010).

As outlined above,several lines of evidence point to the positive correlation between the cytochromebdexpression and virulence in bacteria. For example,M.tuberculosiscytochromebdwas observed to be upregulatedin vivoduring the transition from acute to chronic infection of mouse lungs,inversely,reduced virulence was observed for a mutant strain defective in expressing cytochromebd(Shiet al.2005). The mutation of cytochromebdinBrucella abortusalso resulted in a attenuated bacterial virulence in a murine infection model(Endleyet al.2001). All these evidences strongly indicate thatbd-type oxidases play a critical role in facilitating bacterial survival during successful pathogen infection and pathogenesis.

Our result confirmed that although the high correlation between co-transfer and regulatory relationships between TFs and their TGs was detected in the previous study(Priceet al.2008),however,it is not a 100% reliable indicator to predict the function of an acquired or uncharacterized regulator. Based on our results,we have summarized the acquisition,evolution and function of XOC_2868 in a model (Fig.5). As shown in Fig.5-A,xoc_2868as well as the two physically-linked genes were likely transferred as a unit,however,evolved separately and maintained by selection to adapt the oxidative stress and other possible competitive microenvironments.During infection,the production of reactive oxygen species(ROS) activates the transcription of stress response genes in BLS256. XOC_2868 binds tocydABoperon to protect cells from oxidative stressviaremoving H2O2and other ROS and thus contribute to bacterial virulence. TheoprBand Fru-PTS related genes have been shown to be associated with EPS production and xanthan biosynthesis(Ficarraet al.2017;Baeet al.2018),which are crucial for the biofilm formation of theXanthomonadaceaefamily (Katzenet al.1998). The indirect activation of the expression ofoprBand Fru-PTS related genes thus assist in colonization and dissemination within the host.

Fig.5 Evolutionary hypothesis for the early diversification and functional mechanism of XOC_2868. A,hypothetical model of acquisition and evolution of XOC_2868. B,functional mechanism of XOC_2868 mediating the survival determinants under oxidative stress. ROS means reactive oxygen species and the dotted lines indicate indirect activation. More details of the model were discussed in the text.

5.Conclusion

This study investigated the function of a horizontally acquired gene,xoc_2868,inXocBLS256. Results showed thatxoc_2868might be co-transferred with its physically proximate downstream genes,but they evolved separately inXocBLS256 and were retained under selection pressure. As a transcriptional factor,XOC_2868 directly regulatescydABby binding to a novel regulatory site,protecting cells from oxidative stress by scavenging H2O2and other ROS. In addition,the indirect activation of genes related to EPS production and xanthan biosynthesis promote its colonization and dissemination in the host,thus participating in the pathogenicity of BLS256.These results highlight the possibility of the influence of the HGT on the virulence and adaptability of BLS256 during its evolution.

Acknowledgements

This work was supported by the National Key R&D Program of China (2018YFD0201202 and 2017YFD0201108),the Agri-X Interdisciplinary Fund of Shanghai Jiao Tong University,China (Agri-X2017010),the Shanghai Committee of Science and Technology,China(19390743300),and the National Natural Science Foundation of China (31200003).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm