RcsAB is a major repressor of Yersinia biofilm development through directly acting on hmsCDE, hmsT, and hmsHFRS

Biofilm formation in flea gut is important for flea-borne transmission of Yersinia pestis. There are enhancing factors (HmsHFRS, HmsCDE, and HmsT) and inhibiting one (HmsP) for Yersinia pestis biofilm formation. The RcsAB regulatory complex acts as a repressor of Yesinia biofilm formation, and adaptive pseudogenization of rcsA promotes Y. pestis to evolve the ability of biofilm formation in fleas. In this study, we constructed a set of isogenic strains of Y. pestis biovar Microtus, namely WT (RscB+ and RcsA-), c-rcsA (RscB+ and RcsA+), ΔrcsB (RscB- and RcsA-), and ΔrcsB/c-rcsA (RscB- and RcsA+). The phenotypic assays confirmed that RcsB alone (but not RcsA alone) had an inhibiting effect on biofilm/c-di-GMP production whereas assistance of RcsA to RcsB greatly enhanced this inhibiting effect. Further gene regulation experiments showed that RcsB in assistance of RcsA tightly bound to corresponding promoter-proximal regions to achieve transcriptional repression of hmsCDE, hmsT and hmsHFRS and, meanwhile, RcsAB positively regulated hmsP most likely in an indirect manner. Data presented here disclose that pseudogenization of rcsA leads to dramatic remodeling of RcsAB-dependent hms gene expression between Y. pestis and its progenitor Y. pseudotuberculosis, enabling potent production of Y. pestis biofilms in fleas.

The hmsHFRS orthologs can be found in several bacterial species 15 , including the genetically close pgaABCD operon in Escherichia coli 16 . c-di-GMP binds to PgaC and PgaD (homologues of HmsR and HmsS, respectively), which stabilizes the PgaCD enzymatic complex and thereby activates its glycosyltransferase activity to produce exopolysaccharide 17 . Without c-di-GMP binding, PgaD fails to interact with PgaC and both of them are subject to proteolysis 17 . Y. pestis might employ the conserved c-di-GMP-HmsRS association mechanism to control exopolysaccharide production.
The Enterobacteriaceae Rcs phosphorelay system is an atypical two-component regulatory system composed of three proteins, RcsB, RcsC and RcsD 18 . RcsC and RcsD are membrane-bound proteins, while RcsB is a cytoplasmic one. RcsC acts as a sensor kinase catalyzing autophosphorylation of RcsD and RcsB, and the resulting phosphate group is then transferred to RcsD and finally to RcsB. Phosphorylated RcsB (RcsB-p) acts as a transcriptional regulator alone or upon binding with an auxiliary protein RcsA. The RcsAB complex recognizes a consensus box sequence TAAGAAT-ATTCTTA, which is a 7-7 invert repeat, within the promoter-proximal regions of its target genes mainly including those responsible for exopolysaccharide biosynthesis, flagellar mobility, and Rcs autoregulation (Table S1, and Fig S1).
The biofilm formation of Y. pestis and its genetically very closed progenitor Y. pseudotuberculosis is negatively regulated by the Rcs phosphorelay system 19,21 . The rcsA gene is inactivated in Y. pestis due to a 30 bp duplication insertion in its coding region, and replacing the rcsA pseudogene with functional rcsA allele of Y. pseudotuberculosis strongly represses Y. pestis biofilm formation and essentially abolished flea blockage 19,21 . The conversion of rcsA to a pseudogene during evolution from Y. pseudotuberculosis to Y. pestis is most likely a case of positive Darwinian selection 19,21 .
The present work discloses that the RcsAB complex acts as a major repressor of Y. pestis biofilm formation through directly repressing transcription of hmsCDE, hmsT and hmsHFRS meanwhile positively regulating hmsP in an indirect manner. The above results denote dramatic remodeling of biofilm-related hms gene expression between and Y. pestis and its progenitor Y. pseudotuberculosis due to adaptive pseudogenization of a regulatory gene rcsA.
Taken together, RcsB alone (but not RcsA alone) has an inhibiting effect on biofilm/c-di-GMP production, whereas assistance of RcsA to RcsB greatly enhances this inhibiting effect.
Regulation of hmsCDE, hmsT, hmsHFRS and hmsP by RcsAB. RcsAB box-like sequences could be found within the promoter-proximal regions of hmsCDE, hmsT and hmsHFRS, indicating that they might serve as direct RcsAB targets (Table S3), which promoted us to elucidate RcsAB-dependent expression of these candidate genes. hmsP was also included in the following gene regulation analyses.
The primer extension assays indicated the relative mRNA levels of each of hmsC (Fig. 2a), hmsT (Fig. 3a) and hmsH (Fig. 4a) in the below four strains showed the following tendency: c-rcsA ,WT , DrcsB < DrcsB/c-rcsA. This observation was further confirmed by determination of the promoter activities of the above four genes by LacZ fusion (Fig. 2b, Fig. 3b   MBP-RcsA protected a single upstream region of each of hmsC (Fig. 2d), hmsT (Fig. 3d) and hmsH (Fig. 4d). The above observations indicated that RcsB-p in assistance of RcsA tightly bound to the corresponding promoter-proximal regions to achieve transcriptional repression of hmsCDE, hmsT and hmsHFRS.
By contrast, the relative mRNA levels (determined by primer extension, Fig. 5a) of hmsP showed the following tendency: c-rcsA. WT. DrcsB < DrcsB/c-rcsA, which was further validated by quantitative RT-PCR (data not shown). However, LacZ fusion assay (Fig. 5b) indicated that RcsAB had no regulatory effect on  promoter activity of hmsP. In addition, both EMSA (Fig. 5c) and DNase I footprinting (data not shown) indicated no association between RcsAB and hmsP upstream DNA. Therefore, RcsAB positively regulated hmsP most likely in an indirect manner.
Organization of RcsAB-dependent promoters. Transcription starts determined by primer extension were considered as transcribed promoters for indicated genes and, accordingly, core promoter 210 and 235 elements could be predicted. Each of hmsCD (Fig. 2e), hmsT (Fig. 3e), hmsP (Fig. 4e) and hmsHFRS (Fig. 5d) had a single transcribed promoter. It should be noted that all the above data were consistent with our previous report on regulation of hms genes by Y. pestis ferric uptake regulator Fur 20 .

Discussion
Transcriptional repression of genes for biofilm exopolysaccharide biosynthesis by RcsB with assistance of its auxiliary protein RcsA has been characterized in several bacterial species (Table S1). The present work confirms RcsAB-mediated tight inhibition of Y. pestis c-d-GMP/exopolysaccharide/biofilm production by using a set of isogenic strains of Y. pestis biovar Microtus, namely WT (RscB1 and RcsA-), c-rcsA (RscB1 and RcsA1), DrcsB (RscB-and RcsA-), and DrcsB/c-rcsA (RscB-and RcsA1). RcsAB acts as a major repressor of Y. pestis biofilm formation through directly repressing transcription of biofilm-enhancing genes hmsCDE, hmsT and hmsHFRS and meanwhile positively regulating biofilm-enhancing one hmsP in an indirect manner. RcsB in absence of RcsA does have residual regulatory effects on biofilm formation and hms gene expression and, moreover, RcsB-dependent regulation is greatly increased with assistance of RcsA, which was consistent with previous results 19,21,22 . The above regulatory circuit leads to different expression levels of each of hmsCDE, hmsT, hmsHFRS and hmsP in the above isogenic strains and thus distinct potencies of these strains to produce c-di-GMP/biofilm (summarized in Fig. 6).
Y. pseudotuberculosis (RscB1 and RcsA1, analogous to Y. pestis strain c-rcsA in this study] has a biofilmphenotype in fleas 19,21,22 . In Y. pseudotuberculosis, biosynthesis of HmsCDE, HmsT, and HmsHFRS is tightly inhibited while HmsP is allowed to express. The pseudogenization of rcsA leads to inability of RcsAB complex in Y. pestis, which in turn alleviates RcsAB-mediated inhibition of expression of hmsCDE, hmsT, and hmsHFRS. As a prerequisite of potent Y. pestis biofilm formation, the adaptive pseudogenization of rcsA results in dramatic remodeling of hms gene expression patterns between Y. pseudotuberculosis and Y. pestis, finally enabling Y. pestis biofilm formation in fleas and thereby flea-borne transmission of this pathogen.  RcsB, RcsC, and RcsD are still functional in Y. pesits and thus, there is residual RcsB-dependent repression of biofilm formation in this bacterium 19 . Preclusion of total inactivation of Rcs phosphorelay during Y. pestis evolution might be due to the following reasons: biofilm overproduction if rcsB is inactivated would has detrimental effects on flea as vectors as well as on bacterial growth and proliferation; Rcs phosphorelay plays roles during mammalian infections 23 .
As shown previously 23 , RcsAB binds to the promoter-proximal region of hmsT to repress hmsT transcription. As disclosed in this study, RcsAB inhibits transcription of hmsCDE, hmsT, and hmsHFRS through binding to the promoter-proximal regions of all these direct RcsAB targets. RcsAB sites overlap core promoter -10 elements and transcription start sites of hmsT and hmsHFRS. Association between RcsAB and the above target promoter regions will block entry of RNA polymerase to inhibit transcription of hmsT and hmsHFRS, which has been characterized for RcsAB-mediated transcriptional repression of an array of direct target genes in other Enterobacteriaceae organisms [24][25][26] . Notably, the RcsAB site is upstream of promoter 235 element of hmsCDE and, thus, inhibitory action of RcsAB on hmsCDE transcription appears to be highly unusual, which needs to be further elucidated.

Methods
Bacterial strains. The wild-type Y. pestis Microtus strain 201 (WT) is avirulent to humans but highly virulent to mice 27 . The partial coding region of each indicated gene was replaced by a kanamycin resistance cassette by using the one-step inactivation method based on the lambda phage recombination system 28 , to generate the corresponding mutant of Y. pestis (Table 1). For in trans complementation, a PCRgenerated DNA fragment containing the coding region of each indicated gene together with its promoter-proximal region and transcriptional terminator-proximal region was cloned into the cloning vector pACYC184 (GenBank accession no. X06403), and the resulting recombinant vector was transformed into each indicated Y. pestis strain lack of the corresponding functional gene, generating the corresponding complemented mutant (Table 1). All the primers designed in this study are listed in Table S2.
Bacterial growth and RNA isolation. Overnight cell cultures in the Luria-Bertani (LB) broth with an optical density (OD 620 ) of about 1.0 were diluted 1550 into 18 ml of fresh LB broth for further cultivation at 26uC with shaking at 230 rpm to reach middle stationary phases (an OD 620 of 0.8 to 1.2), followed by cell harvest for further gene regulation or phenotypic assays. Immediately before bacterial harvest for RNA isolation, double-volume of RNAprotect reagent (Qiagen) was mixed with onevolume of cell culture, and total RNA was extracted using TRIzol Reagent (Invitrogen). RNA quality was monitored by agarose gel electrophoresis, and RNA quantity was determined by spectrophotometry.
Primer extension assay. As described in our previous studies 29, 30 , a 59-32 P-labeled oligonucleotide primer complementary to a portion of the RNA transcript of each indicated gene was employed to synthesize cDNAs from total RNA templates using Promega Primer Extension System. Sequence ladders were prepared with the same 59-32 P-labeled primers using AccuPower & Top DNA Sequencing Kit (Bioneer). Radioactive species were detected by autoradiography. If different Y. pestis strains were involved in a single experiment, equal amounts of the total RNA samples were used as the starting materials. The relative mRNA level was determined with the observed band intensity of the primer extension product of each target gene. The 59-terminus of RNA transcript (i.e., transcription start) of each target gene was mapped according to the size of primer extension product.
LacZ fusion and b-galactosidase assay. A promoter-proximal DNA region of each indicated gene was cloned into the low-copy-number transcriptional fusion vector pRW50 31 that harbors a promoterless lacZ reporter gene. Y. pestis strains transformed with the recombinant plasmid or the empty pRW50 (negative control) were grown to measure b-galactosidase activity in cellular extract using b-Galactosidase Enzyme Assay System (Promega) 29,30 .
Protein expression and purification. The entire coding region of Y. pseudotuberculosis rcsA or Y. pestis rcsB was cloned into plasmid pMAL-c4X (Invitrogen) 23 or pBADMyc-His A (New England Biolabs) 23 , respectively. The wild-type Y. pestis strain KIM61 and the rcsB null mutant of KIM61 were employed as host cells for expression of maltose-binding protein (MBP)-tagged RcsA (MBP-RcsA) and 6 3 His-tagged RcsB (His-RcsB), respectively 23 . His-RcsB and MBP-RcsA were purified under native conditions using Ni-NTA Agarose Column (Qiagen) and Amylose Agarose Column (New England Biolabs), respectively 23 . Each purified protein was dialyzed and then concentrated to a concentration of about 0.1 mg/ml in phosphate buffered saline (pH 8.0) containing 20% glycerin.
EMSA. Each indicated 59-32 P-labeled target DNA fragment was incubated with increasing amounts of purified His-RcsB, or with increasing amounts of purified His-RcsB with addition of 24 pmol of purified MBP-RcsA, for 30 min at room temperature in a binding buffer 29,30 . To achieve RcsB phosphorylation, 25 mM fresh acetyl phosphate was incubated for 30 min with His-RcsB in the binding buffer, Figure 6 | RcsAB-dependent gene expression and phenotypes. Shown were relative mRNA levels of each of hmsCDE, hmsT, hmsHFRS and hmsP in different isogenic Y. pestis strains, as well as relative potencies to produce c-di-GMP/biofilm of these strains. before labeled DNA probes were added. The resulting reactions were subjected to a native 4% (w/v) polyacrylamide gel electrophoresis. Each EMSA experiment included three controls, namely, cold probe as the specific DNA competitor (the same promoter-proximal DNA region unlabeled), negative probe as the nonspecific DNA competitor (the unlabeled coding region of the 16S rRNA gene), and nonspecific protein competitor (rabbit anti-F1-protein polyclonal antibodies) 29,30 . Detection of sequencing and radioactive species was as above.
DNase I footprinting. For DNase I footprinting 29,30 , the target DNA fragment with a single 32 P-labeled end was incubated with increasing amounts of purified His-RcsB-p with addition of 24 pmol of purified MBP-RcsA, which was followed by partial digestion of RQ1 RNase-Free DNase I (Promega). The digested DNA samples were purified and analyzed in an 8 M urea-6% polyacrylamide gel. Detection of sequencing and radioactive species was as above. Footprints were identified by comparison with sequence ladders.
Biofilm and c-di-GMP assays. As described in our previous study 32 , three different methods were used to detect Y. pestis biofilms. First, in vitro biofilm masses, attached to well walls when bacteria were grown in polystyrene microtiter plates, were stained with crystal violet. Second, percentages of fourth-stage larvae and adults (L4/adult) of C. elegans after incubation of nematode eggs on Y. pestis lawns, negatively reflecting bacterial ability to produce biofilms, were determined. Third, rugose colony morphology of bacteria grown on LB agar plates, positively reflecting bacterial ability to synthesize exopolysaccharide, was observed. In addition, intracellular c-di-GMP levels were determined by a chromatography-coupled tandem mass spectrometry (HPLC-MS/MS) method as described in our previous study 20 .
Experimental replicates and statistical methods. For LacZ fusion, crystal violet staining of biofilms, and determination of L4/adult nematodes or c-di-GMP, experiments were performed with at least three independent bacterial cultures/lawns, and values were expressed as mean 6 standard deviation. Paired Student's t-test was performed to determine statistically significant differences; P ,0.01 was considered to indicate statistical significance. For primer extension and colony morphology observation, shown were representative data from at least two independent bacterial cultures.