The disease riemerella anatipestifersis, caused by the Gram-negative bacterium Riemerella anatipestifer, occurs primarily in 1–8-week-old ducks but is most common in more susceptible 2–3-week-old ducklings. It is currently the most economically damaging bacterial infection affecting the global duck industry. Symptoms are characterized by fibrinous pericarditis, glissonitis, airbag inflammation, and meningitis (Segers et al., 1993). A total of 21 R. anatipestifer serotypes have been identified, with no significant cross-protection reported, making it difficult to control the disease through vaccination (Pathanasophon et al., 1995, 2002). R. anatipestifer serotypes 1, 2, and 10 are responsible for most major outbreaks in China. In recent years, many R. anatipestifer strains with differing virulence have been isolated from duck farms in China (Yuan et al., 2011; Wang et al., 2015; Zhang et al., 2015; Song et al., 2016). Additionally, several virulence factors have been identified that associate with disease severity, including VapD (Chang et al., 1998), CAMP cohemolysin (Crasta et al., 2002), outer membrane protein A (OmpA; Hu et al., 2011), Nicotinamidase PncA (Wang et al., 2016), and putative genes associated with lipopolysaccharide (LPS) synthesis (Yu et al., 2016).

Bacteria sense and adapt to their environment via two-component signaling systems (TCSs). A paired TCS typically has a sensing histidine kinase (HK) coupled to a response regulator (RR). The sensing of a signal by the HK leads to autophosphorylation on a histidine residue. Subsequent transfer of the phosphate to an aspartate residue on the cognate RR facilitates the binding of the RR to its specific DNA. Each phosphorylated RR regulates specific genes that enable individual bacteria to sense environmental factors and respond to stresses (Stock et al., 2003; Mike et al., 2014). TCSs are present in nearly all sequenced bacterial genomes, as well as in some fungal, archaeal, and plant species (Skerker et al., 2008).

TCSs are involved in the regulation of a variety of important biological functions in bacteria. In particular, PhoP/PhoQ TCSs control the transcription of key virulent genes in several bacterial pathogens, including Escherichia coli, Shigella spp. (Tobe, 2008), Yersinia pestis (O'Loughlin et al., 2010), and Salmonella typhimurium (Tran et al., 2016). TCSs also control the expression of Pneumococcal surface antigen A (PsaA), subsequently regulating virulence and resistance to oxidative stress in Streptococcus pneumonia (McCluskey et al., 2004). Related two-component system also regulate genes that are essential to virulence and complex lipid biosynthesis. For example, SenX3, and RegX3 form a TCS that is involved in Mycobacterium tuberculosis virulence (Ryndak et al., 2008). The TCS is expressed during phosphate starvation and is required for phosphate uptake and aerobic respiration (Parish et al., 2003; Haydel et al., 2012). Additionally, in TB, the TCS PrrAB is required during early intracellular infection (Haydel et al., 2012) and the TCS MprAB responds to envelope stress by regulating stress-responding and virulence-associated genes (He et al., 2006; Pang et al., 2007).

The bacterium R. anatipestifer YM was isolated in Yunmeng, Hubei province, China, and is a highly virulent strain classified as serotype 1 (Zhou et al., 2011). We have previously used in vivo-induced antigen technology (IVIAT) to identify in vivo-induced protein antigens from R. anatipestifer. This predicted the involvement of a putative TCS. To research the function of this TCS in R. anatipestifer, we constructed a mutant with the putative TCS genes, RAYM_RS09735 and RAYM_RS09740, deleted to investigate their biological characteristics. We found that RAYM_RS09735/RAYM_RS09740 form a PhoP/PhoR TCS, the first reported such TCS in Gram-negative bacteria. We confirmed that the RAYM_RS09735/RAYM_RS09740 TCS is an important global transcription regulator and regulates the expression of virulence-associated genes in R. anatipestifer. This may provide the theoretical basis for further study into the molecular pathogenesis of R. anatipestifer and facilitate the design of genetically engineered vaccines against R. anatipestifer.

Bacterial two-component regulatory systems (TCSs), consisting of a sensing histidine kinase and a response regulator, mediate gene expression in response to environmental stimuli (Stock et al., 2003). To achieve this, histidine kinases sense environmental signals and autophosphorylate. This additional phosphate is then transferred to an aspartic acid residue on the corresponding response regulator (Hoch, 2000; Alm et al., 2006). The phosphorylated response regulator elicits a diverse range of downstream responses, including an enhancement of its DNA binding ability. This allows the TCS to modulate target gene expression (West and Stock, 2001; Lin et al., 2015). Using an in vivo antigen technology (IVIAT) to examine ducks infected with R. anatipestifer, we have identified a putative TCS involving the genes RAYM_RS09740 and RAYM_RS09735. Bioinformatics analysis of these two genes, and their proteins, revealed that RAYM_RS09740 and RAYM_RS09735 encode a histidine kinase and response regulator, respectively. We also found that, like most TCSs, RAYM_RS09735 and RAYM_RS09740 have the same promoter and co-transcribe as an operon (Aggarwal et al., 2016). RAYM_RS09735 and RAYM_RS09740 possess all of the characteristic TCS domains essential to their biochemical activities and responses. RAYM_RS09735 and RAYM_RS09740 were predicted to belong to the OmpR and BaeS families, respectively, and RAYM_RS09735 was further predicted to be a phosphate regulon sensor protein (PhoR).

Our study aimed to investigate the RAYM_RS09735/RAYM_RS09740 TCS and explore its potential functions in R. anatipestifer. To better evaluate the role of RAYM_RS09735 and RAYM_RS09740 in R. anatipestifer, we constructed the mutant strain YMΔRS09735/RS09740 in which the RAYM_RS09735 and RAYM_RS09740 genes were mutated by deleting a 1400-bp fragment from the wild-type RA-YM strain. RT-PCR and qRT-PCR were used to validate the mutant strain (Figures 2A,B). We found that upstream and downstream genes were significantly down-regulated in the YMΔRS09735/RS09740 mutant, although the strain showed the same growth characteristics as the wild-type in TSB (Figure 2C).

LD₅₀ analysis of the two strains demonstrated the RAYM_RS09735/RAYM_RS09740 mutant had significantly reduced virulence. Indeed, the strain was almost avirulent to ducklings (Table 3). A normal R. anatipestifer infection is characterized by septicemia and the ability to colonize and develop in the tissues of their host. We found that the bacterial loads in the liver, spleen, heart, brain, and blood of ducks infected with the mutant YMΔRS09735/RS09740 strain were significantly lower than that of wild-type RA-YM infected ducklings (Figure 3). In addition, pathological histological analysis using hematoxylin and eosin staining showed that the mutant strain also induced significantly less damage to the heart, brain, and spleen, compared to the wild-type strain. There was still slight damage to the liver (Figure 4). Our results indicate that the RAYM_RS09735 and RAYM_RS09740 genes are important mediators of R. anatipestifer virulence during infection in ducklings. Therefore, RAYM_RS09740 was determined to be a response regulator, with a combined promoter affecting downstream genes. The protein also altered the expression of target genes. Our study therefore indicated that virulence factors in R. anatipestifer could be regulated by RAYM_RS09740.

RAYM_RS09735 and RAYM_RS09740 share up to 70% sequence identity with Flavobacteriaceae species, such as Cloacibacterium, Epilithonimonas, and Chryseobacterium. These species have been previously reported to exhibit drug resistance. Molecular docking analysis identified inhibitors of various RRs in OmpR family to affect this observed drug resistance. Therefore targeting TCSs has become an attractive option for the development of new drugs, particularly to combat M. tuberculosis (Banerjee et al., 2016). These studies also demonstrate that TCSs can be used as the theoretical basis for the study of drug resistance mechanisms in Flavobacteriaceae.

To investigate how RAYM_RS09735 and RAYM_RS09740 regulate gene expression, RNAseq analysis was performed to identify differentially expressed genes in the mutant YMΔRS09735/RS09740 strain. This revealed 112 genes that were upregulated (13.9%) and 693 genes that were down-regulated (86.1%) relative to the wild-type strain (approximately one-third of genes had differential expression). To validate the RNAseq results, we randomly assessed the differential expression of 10 genes using qRT-PCR, confirming the accuracy of the RNAseq data (Figure 7). Gene ontology (GO) enrichment analysis showed that differentially expressed genes were involved in a wide-variety of biological processes, cellular components, and molecular functions, including 21 distinct pathways (Figure 6). Therefore, we hypothesize that the RAYM_RS09735/RAYM_RS09740 TCS is a global expression regulator that controls the expression of virulence genes, thus affecting the pathogenicity of R. anatipestifer.

KEGG pathway analysis showed that RAYM_RS09740 is a PhoP protein, indicating that RAYM_RS09735 and RAYM_RS09740 form a PhoP/PhoR TCS. PhoP/PhoR TCSs have been implicated in several biological processes in Gram-positive bacteria, such as B. subtilis (Hulett et al., 1994), Streptomyces lividans (Sola-Landa et al., 2003), and M. tuberculosis (Perez et al., 2001). Specifically, PhoP/PhoR has been shown to control respiration(Birkey et al., 1998), cell wall metabolism (Minnig et al., 2005), culture metabolism (Thomas et al., 2012), and biofilm formation (Bluskadosh et al., 2013). Most importantly, in many pathogens PhoP/PhoR TCSs have also been shown to regulate pathogenesis (Perez et al., 2001; Ryndak et al., 2008). PhoP and PhoQ TCS components have been reported in many Gram-negative bacteria, including E. coli (Eguchi et al., 2007), Edwardsvilla (Lv et al., 2012), and Salmonella (Tran et al., 2016). PhoP/PhoR is also an important TCS in M. tuberculosis and is a key player in virulence (Ryndak et al., 2008). However, PhoP/PhoR TCSs have not been previously observed in Gram-negative bacteria. Our data shows that a putative PhoP/PhoR TCS is involved in R. anatipestifer virulence, although further phosphotransfer profiling will be required to validate RAYM_RS09735/RAYM_RS09740 as a PhoP/PhoR TCS.

Our data also suggest that RAYM_RS09740 can act as a transcription factor and plays a role in regulating genes expression, although there are likely other TCSs present in R. anatipestifer. This is based on our observation that, after the deletion of RAYM_RS09735 and RAYM_RS09740, other TCSs regulated gene expression. We hypothesize that cross-talk and -regulation may occur between a multitude of two-component systems in R. anatipestifer, similarly to other organisms. For example, the transcription factor NemR is a redox-regulated transcriptional repressor in E. coli (Gray et al., 2013) and a carbohydrate-responsive regulatory protein (BadR) is a transcriptional repressor of rpoS in Borrelia burgdorferi (Miller et al., 2013).

The PhoP/PhoR TCS we identified was initially predicted to be part of the phosphate (Pi) stress-response in R. anatipestifer. PhoR acts to sense environmental phosphate (Pi) levels in bacteria through the ABC-type phosphate-specific transport (Pst) system and the protein PhoU. PstS is a periplasmic protein that binds Pi with high affinity and PhoP induces the promoter activity of pstS. We found with RNAseq and qRT-PCR that the mRNA expression of pstS was 8-fold higher (Figure 8A) in the mutant strain and different from other Gram-negative bacteria. Therefore, RAYM_RS09735 and RAYM_RS09740 may act more like the PhoP/PhoR system of M. tuberculosis. Additionally, M. tuberculosis PhoP/PhoR does not respond to phosphate starvation and is similar to the PhoP/PhoQ TCS in Salmonella. In Salmonella PhoP/PhoQ TCS, PhoQ senses low Mg²⁺ levels, and PhoP activates expression of genes encoding high-affinity Mg²⁺ transport systems (Walters, 2006). We aim to investigate the environmental factors sensed by R. anatipestifer PhoP/PhoR in future work.

Further, DEG analysis showed that the expression of several other signal transduction system genes was altered in the YMΔRS09735/RS09740 strain. It is likely that PhoP can influence the expression of transcription factors in other signal transduction system and there is cross-talk or cross regulation between PhoP/PhoR and other TCSs in R. anatipestifer. As more than one third of genes demonstrated differential expression, it is unlikely that the PhoP/PhoR TCS directly regulates all of these DEGs. However, the TCS we identified may affect wider gene expression by influencing the expression of other transcription factors in a variety of signal transduction systems.

Additionally, the CRISPR-Cas bacterial immune system has been shown to be present in R. anatipestifer (Chamnongpol et al., 2003; Zhu et al., 2016) and we found that the expression of cas9, cas1, and cas2 were all different in the YMΔRS09735/RS09740 mutant, compared to the wild-type. Recent studies have reported that quorum sensing has a role in controlling the Pseudomonas aeruginosa CRISPR-Cas adaptive immune system (Høyland-Kroghsbo et al., 2016) and the TCS BfiSR affects the production of proteins involved in virulence, post-translational modification, and quorum sensing (Petrova and Sauer, 2010). We therefore reasoned that TCS might also affect the CRISPR-Cas immune system of R. anatipestifer.

We found 13 genes encoding hypothetical proteins that were downregulated more than 4-fold. At present, M949_RS01915, M949_1360, AS87_01735, AS87_03730, M949_1556, and AS87_04050 are all hypothetical proteins but have been reported to associate with virulence in R. anatipestifer. Bacterial lipoprotein (BLP) and the hydrolase Nlp/P60 were both found to be significant down-regulated in the mutant YMΔRS09735/RS09740 strain (Figure 8B). Nlp/P60 is involved in cell growth and division, autolysis, and invasion (Xu et al., 2015). Bacterial lipoproteins on the cell surface include an important class of virulence factor that is expressed by many pathogens. BLP can promote TLR2 expression, TLR2-induced NF-κB activation, and IL-6 production, leading to immune inflammation (Buddelmeijer, 2015). The clinical symptoms of R. anatipestifersis are characterized by fibrinous pericarditis, glissonitis, airbag inflammation, and meningitis. Our pathological results demonstrated that the damage induced by the mutant YMΔRS09735/RS09740 strain to the heart, brain, and spleen was significantly less severe than the wild-type strain (Figure 4). We therefore hypothesize that R. anatipestifer reduced expression of BLP and the hydrolase Nlp/P60, subsequently affecting inflammation in ducklings.

Other downregulated genes we identified included transcription factors, metabolic genes, multiple molecular chaperones, and TonB-dependent receptors. The primary role of DNA binding proteins and transcription factors is to regulate the expression of the genes and RAYM_RS09740 may regulate transcription by modifying the expression of these transcription factors. Molecular chaperone proteins assist in the folding of nascent polypeptide chains and prevent the aggregation of denatured proteins (Hartl and Hayerhartl, 2002). We found that transcript for the molecular chaperones dnaJ, dnaK, grpE, clpP, tir, and skp were all downregulated in the mutant YMΔRS09735/RS09740 strain, suggesting that the RAYM_RS09735/RAYM_RS09740 TCS may regulate the expression of the chaperone genes to influence the severity of R. anatipestifersis. The DnaK molecular chaperone system includes DnaK, DnaJ, and GrpE (Mayer et al., 2000; Ben-Zvi and Goloubinoff, 2001; Genevaux et al., 2007) and is related to stress tolerance and virulence in Salmonella. DnaK can also regulate the expression of virulence islands and a DnaK mutation strain was found to be largely avirulent (Takaya et al., 2004). ClpP was found to be a virulence-related factor in Actinobacillus pleuropneumoniae, with a clpP gene deletion strain demonstrating decreased biofilm production (Xie et al., 2013). We also found that TonB-dependent receptors were downregulated. These receptors have been reported to associate with hemin iron acquisition and R. anatipestifer virulence.

In this study, RAYM_RS09735 and RAYM_RS09740 were predicted to be components of a PhoP/PhoR TCS in R. anatipestifer using bioinformatics analysis. Furthermore, deletion of the RAYM_RS09735 and RAYM_RS09740 genes significantly decreased R. anatipestifer virulence. RNAseq analysis showed that RAYM_RS09740 is a global expression regulator in R. anatipestifer. We also demonstrated that the RAYM_RS09735/RAYM_RS09740 TCS contributes to virulence, signal translation, and the CRISPR-Cas adaptive immune system of R. anatipestifer. BLP and the hydrolase Nlp/P60 were predicted to be two virulence factors involved. Although common in Gram-positive bacteria, the predicted PhoP/PhoR TCS we have putatively characterized is the first such TCS identified in Gram-negative bacteria. Our study is also the first to report this TCS in R. anatipestifer, providing new insight into the role of TCSs in regulating R. anatipestifer virulence.

All the animal experiments were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals from Research Ethics Committee, Huazhong Agricultural University, Hubei, China. All procedures performed in studies involving animals were in accordance with the ethical standards of the institution or practice at which the studies were conducted.

Funding acquisition: ZL. Investigation: YW. Methodology: YW, TL, and XY. Project administration: SL, ML, SH, and DB. Resources: ZZ. Supervision: ZL. Validation: ZL. Visualization: YW. Writing—original draft: YW. Writing—review's editing: ZL.

This work was supported by grants from the Wuhan Science and Technology Bureau (No. 2015020101010070 to ZL) and Natural Science Foundation of Hubei Province (No. 2015CFB268 to ZL).