Identification of genetic determinants of hemolytic activity of Riemerella anatipestifer using random transposon mutagenesis

Riemerella anatipestifer causes epizootic infectious disease in poultry resulting in serious economic losses especially to the duck industry. In our previous study, R. anatipestifer was found to lyse duck erythrocytes in vitro. In the present study, a random Tn4351 mutagenesis library of hemolytic R. anatipestifer strain SX containing 4000 mutants was constructed to investigate the genetic basis of hemolytic activity. Thirty mutants with reduced hemolytic activity and one with increased hemolytic activity were screened and insertions in 24 genes were identified. Of these genes, four were predicted to encode outer membrane proteins, one encoded a cytoplasmic membrane protein, 11 encoded cytoplasmic proteins, and eight encoded proteins with unknown locations. Based on current annotations of the R. anatipestifer genomes, of the 24 genes, 7 (29.17%) were involved in iron utilization. The hemolytic activities of the complemented strains M2 (pRES-Riean_0790) and M18 (pRES-Riean_0653) were restored, indicating that both Riean_0653 and Riean_0790 are involved in the hemolytic activity of strain SX. However, the recombinant proteins rRiean_0317, rRiean_0790, rRiean_0653, rRiean_1027, rRiean_1143, and rRiean_1561 had no hemolytic activity, suggesting that none were hemolysins.


Introduction
Riemerella anatipestifer, the type species of the genus Riemerella in the family Flavobacteriaceae, mainly infects domestic ducks, geese, turkeys, and other birds. R. anatipestifer infection primarily causes an acute septicemic disease in younger birds and more chronic and localized lesions in older birds, resulting in significant economic losses to the duck industry worldwide [1]. However, the mechanisms underlying the pathogenesis of R. anatipestifer infection remain unclear.
Some R. anatipestifer strains caused hemolysis on 5% bovine blood agar over a 14-day examination period [2] and on Columbia agar base with 7% defibrinated sheep blood [3]. A previous study by our group showed that 29 (55.77%) of 52 R. anatipestifer strains caused hemolysis on duck blood agar and observations by phase contrast microscopy showed that all of the strains generated pores in the membranes of duck erythrocytes in suspension [4]. Although the virulence of strong hemolytic R. anatipestifer strains may not be greater than that of weak hemolytic strains [4], a mutant with enhanced hemolytic activity and increased pathogenicity to ducklings was identified (Hu et al. unpublished data), suggesting that hemolytic activity may be related to virulence in at least one R. anatipestifer strain. However, in R. anatipestifer, with the Open Access *Correspondence: huqinghai@caas.cn 1 Shanghai Veterinary Research Institute, the Chinese Academy of Agricultural Sciences, 518 Ziyue Road, Shanghai 200241, China Full list of author information is available at the end of the article exception of the cam gene, encoding cohemolysin, which confers cohemolytic activity of Staphylococcus aureus on trypticase soy agar supplemented with 5% sheep blood (Christie-Atkins-Munch-Peterson test) [5], no genes associated with hemolysis have yet been identified.
Transposon-based random mutagenesis of bacterial genomes is a powerful genetic tool for the identification of genes and regulatory elements that contribute to specific phenotypes. The most frequently used transposons in bacterial genetics are based on the Tn5, Tn10, and Himar1 mariner platforms [6]. However, transposons Tn5 and Himar1 have no or low transposition frequency in R. anatipestifer (Hu et al. unpublished data). Transposon Tn4351 has been used for transposon mutagenesis of Bacteroidetes sp. [7,8]. In our previous studies, Tn4351 was successfully used to construct random transposon mutant libraries of R. anatipestifer, which included 33 genes involved in biofilm formation in R. anatipestifer strain CH3 [9] and 28 potential virulence-associated genes in R. anatipestifer strain YZb1 [10].
In the present study, a random Tn4351 transposon mutagenesis library was successfully constructed, which included 24 genes associated with the hemolytic activity of R. anatipestifer strain SX.

Bacterial strains and culture conditions
Hemolytic R. anatipestifer strain SX was isolated from the brain of a sickened Cherry Valley duck in our laboratory in 2010 [4]. Erythromycin resistance is conferred to strain SX via the ereD gene [11]. Escherichia coli strain BW19851 carrying the plasmid pEP4351 [12] was a generous gift from Professor Mark J. McBride of the University of Wisconsin-Milwaukee (Milwaukee, WI, USA). R. anatipestifer cells were cultured at 37 °C in tryptic soybean broth (Difco Laboratories Inc., Detroit, MI, USA) or on tryptic soy agar, while E. coli cells were routinely grown in Luria broth (LB; Difco Laboratories Inc.) or on LB agar at 37 °C. For selective growth of the bacteria, the following antibiotics (concentrations) were added: ampicillin (100 μg/mL), chloramphenicol (34 μg/mL), cefoxitin (0.5 μg/mL), and kanamycin (50 μg/mL). The strains, plasmids, and primers used in this study are listed in Additional file 1.

Construction of the recombinant plasmid pEP4351-cfxA
The cfxA expression box was amplified from plasmid pCP29 by polymerase chain reaction (PCR) and ligated into the pGEM-T easy vector (Promega Corporation, Madison, WI, USA), which generated T-cfxA. Then, the cfxA gene was deleted from T-cfxA with the endonucleases PpuMI and AflII, and ligated into the plasmid pEP4351 to generate the plasmid pEP4351-cfxA, in which parts of the ermF and tetX genes in transposon Tn4351 DNA were replaced with the cfxA expression box.

Generation of a mutagenesis library of R. anatipestifer strain SX Tn4351
Tn4351 insertion mutagenesis was performed as described previously [10] with E. coli strain BW19851 (pEP4351-cfxA) as the donor and R. anatipestifer strain SX as the recipient. After conjugation, the plasmid pEP4351 from E. coli BW19851 was introduced into R. anatipestifer strain SX and then transposon Tn4351 was randomly inserted into SX genome. The transconjugants were selected on tryptic soy agar containing cefoxitin and kanamycin with the 16S rRNA + cfxA + transconjugant as the positive mutant.

Screening of mutants with decreased or increased hemolytic activity
The hemolytic activities of the wild-type (WT) strain SX and transposon insertion mutants were determined on duck blood agar plates (LB agar base supplemented with 3.5% duck blood, pH 7.5) as described previously [4]. Briefly, suspended bacterial cells were plated on blood agar plates, which were then incubated at 37 °C under an atmosphere of 5% CO 2 /95% air for 24 h, followed by incubation overnight at 4 °C. The hemolytic zones on duck blood agar generated by the mutants and WT SX strain were compared. Transposon insertion mutants with increased or reduced hemolytic activities were isolated. The hemolytic activities of the mutants were determined at least twice on duck blood agar.

Identification of transposon insertion sites
The genomic DNA of the transposon insertion mutants with decreased or increased hemolytic activity was extracted using the TIANamp Bacteria DNA Kit (Tiangen Biotech (Beijing) Co., Ltd., Beijing, China). The nucleotide sequences flanking the transposon insertion sites were determined using inverse PCR or genomic walking as described previously [9,13,14]. The sequences of the identified genes were used to search for other known homologous sequences and putative functions using the Basic Local Alignment Search Tool (BLAST) server [15] and the online PSORTb v.3.0 program [16] was used to predict the subcellular locations of the proteins.
Southern blot analysis of the Tn4351 insertions was used to identify the mutants. The genomic DNA of the tagged Tn4351 insertion mutants was digested with XbaI, separated by gel electrophoresis, and transferred to nylon membranes, essentially as described previously [9]. The transposons were identified by detection of the cfxA gene. The DIG DNA Labeling and Detection Kit (Roche Diagnostics USA, Indianapolis, IN, USA) was used to prepare the probes and perform hybridization. The number of cfxA bands on the nylon membrane represented the number of transposon insertion sites in one mutant.

Complementation of the mutant strains
To determine whether the decreased hemolytic phenotype of the mutants was due to an inactivated gene, mutants M2 and M18, in which the Riean_0790 and Riean_0653 genes were inactivated by insertion of transposon Tn4351, respectively, were used for a complementation experiment. The open reading frames of Riean_0653 and Riean_0790 were amplified from the WT SX strain by PCR and subcloned into the E. coli-R. anatipestifer shuttle plasmid pRES [17] generating pRES-Riean_0653 and pRES-Riean_0790, respectively. Then, the plasmids pRES-Riean_0790 and pRES-Riean_0653 were introduced into the mutants M2 and M18 by conjugation generating the complemented strains M2 (pRES-Riean_0790) and M18 (pRES-Riean_0653), respectively.

Animal experiments
One-day-old Cherry Valley ducklings were obtained from the Lijia Duck Farm (Wuxi, Jiangsu province, China). Serum samples of the ducklings were free of antibodies against R. anatipestifer as detected with an indirect enzyme-linked immunosorbent assay using whole cells of strain SX as the coating antigen. All animal experiments were conducted in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of Shanghai Veterinary Research Institute, the Chinese Academy of Agricultural Sciences (CAAS), Shanghai, China. The study protocol was approved by the Committee on the Ethics of Animal Experiments of the Shanghai Veterinary Research Institute, CAAS (permit no. SHVRI-SV-20201119-02).
To determine whether the genes inactivated in the mutants that caused defects in hemolytic activity had an influence on virulence, the median lethal dose (LD 50 ) values of the WT SX strain, the mutant strains M2 and M18, and the complemented strains M2 (pRES-Riean_0790) and M18 (pRES-Riean_0653) were measured using 8-day-old Cherry Valley ducklings as described previously [18].

Expression of rRiean_1561, rRiean_1143, rRiean_0790, rRiean_0653 and rRiean_0317 in E. coli cells
The open reading frames of Riean_1561, Riean_1143, Riean_0790, Riean_0653, and Riean_0317 from the WT SX strain were amplified, cloned into the expression vector pET30a ( +) (Novagen, Inc., Madison, WI, USA), and then expressed in E. coli BL21 (DE3) cells (Novagen, Inc.). The bacteria were collected by centrifugation and lysed by sonication. The recombinant proteins rRiean_0317, rRiean_0653, and rRiean_1143 were expressed in E. coli cells mainly in the soluble form, while rRiean_1561 and rRiean_0790 were expressed mainly in inclusion bodies. The soluble recombinant proteins rRiean_0317, rRiean_0653, and rRiean_1143 were purified by Ni-iminodiacetic acid affinity chromatography (Detai Bio-Tech (Nanjing) Co., Ltd., Nanjing, China), and the recombinant fusion proteins rRiean_1561 and rRiean_0790 were obtained after washing and dissolution of the inclusion bodies, purification by affinity chromatography, and refolding as previously described [19]. All purified recombinant proteins were soluble in solution buffer (50 mM Tris [pH 8.0] and 150 mM NaCl) and the concentration of each recombinant protein solution was measured using the BCA Protein Assay kit (Pierce Biotechnology, Waltham, MA, USA). The proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and identified by Western blot analysis using anti-His tag antibodies (Beyotime Institute of Biotechnology, Haimen, China).

Hemolytic activities of the complemented strains and recombinant proteins
The hemolytic activities of the WT SX strain, mutant strains M2 and M18, and complemented strains were measured on duck blood agar as described previously [4], and the hemolytic zones generated by these strains were observed. The hemolytic activities of the recombinant proteins rRiean_0317, rRiean_0790, rRiean_1027 (rOmpA1467) [19], rRiean_0653, rRiean_1143, and rRiean_1561 were also monitored on duck blood agar as described previously [4]. The solution buffer was used as a negative control. In addition, to further determine whether these recombinant proteins could perforate the duck erythrocyte membrane, 2% (v/v, in Williams' Medium E; Sigma-Aldrich Corporation, St. Louis, MO, USA) duck red blood cells were mixed with each recombinant protein in the wells of a 96-well round-bottomed cell culture plate (Corning Incorporated, Corning, NY, USA) and incubated for 12 h at 37 °C. Duck blood cells alone in Williams' Medium E were used as a negative control. The pelleted non-lysed duck blood cells were washed twice with Williams' Medium E and the morphologies of the cells were observed under a phase contrast microscope (Eclipse Ci; Nikon Corporation, Tokyo, Japan).

Screening of the hemolytic activities of the R. anatipestifer SX mutant strains
A random Tn4351 transposon mutagenesis library of R. anatipestifer strain SX containing 4000 mutants was constructed with biparental mating of E. coli strain BW19851 (pEP4351-cfxA) with R. anatipestifer strain SX. The hemolytic activities of all mutants and the WT SX strain were measured on duck blood agar. The mutants that exhibited no or very weak hemolytic zones on duck blood agar were characterized as hemolytic-defective mutants, while those with obvious increased hemolytic zones were characterized as hemolytic-increase mutants (Fig. 1). On the basis of these criteria, 30 of the 4000 transposon mutants with no or very weak hemolytic zones and one mutant with an increased hemolytic zone on duck blood agar were identified. These mutants displayed reproducible hemolytic activity on duck blood agar. Southern blot analysis confirmed that each of the 31 mutants had a single Tn4351 insertion.

Identification of disrupted genes
The GenBank database was searched for genes homologous to the DNA sequences flanking the transposon Tn4351 insertion sites of the 31 mutants. The results of this analysis are shown in Table 1, along with the details of the identified 24 mutated genes. Of these, five genes were found to be inserted by the transposon Tn4351 in different mutants. In mutants M18, M27, M29, and M31, Riean_0653 was inserted by Tn4351 at three sites; in mutants M2 and M25, Riean_0790 was inserted at different sites; in mutants M3 and M26, Riean_1027 was inserted at the same site; in mutants M7 and M28, Riean_0317 was inserted at different sites, and in mutants M21 and M30, Riean_1143 was also inserted at the same site (Additional file 3). The two mutants M3 and M26, in which transposon Tn4351 was inserted at the same site on the genome, were isolated from different experiments, as with M27 and M31, M21 and M30. Therefore, none of the mutants shown in Fig. 1 appear to be siblings. In addition, in six mutants (M5, M11, M14, M16, M24 and M28), the transposon Tn4351 was inserted into the putative promoter of the affected gene (Additional file 3). Surprisingly, seven of the 24 genes identified in this study were predicted to be involved in iron utilization based on the current annotations of R. anatipestifer genomes.
The subcellular locations of the 24 protein-encoding genes were predicted using PSORTb software [16]. Of these 24 genes, four were predicted to encode outer membrane proteins (OMPs), one encoded a cytoplasmic membrane protein, 11 encoded cytoplasmic proteins, and eight encoded proteins with unknown locations (Table 1). These proteins were further categorized based on putative functions. Of the 24 proteins, nine were classified into the cellular processes and signaling category (M, T, O, U, V. and P), six into the metabolism category (C, F, H, I, and Q), one into the information storage and processing category (J), two as poorly characterized  clusters of orthologous groups (R), and seven could not be categorized.

Hemolytic activities of the mutants M2 and M18 were restored by plasmid complementation
As shown in Fig. 2, on duck blood agar, there was no or very weak hemolytic activities of the mutants M2 and M18, while the hemolytic activities of the complemented strains M2 (pRES-Riean_0790) and M18 (pRES-Riean_0653) were almost completely restored to that of the WT SX strain, suggesting that both Riean_0653 and Riean_0790 are involved in the hemolytic activities of R. anatipestifer strain SX.

Determination of the virulence of mutants M2 and M18
Eight-day old Cherry Valley ducklings were infected with the WT SX strain, mutant strains M2 and M18, and the complemented strains to determine the effect of the mutations on the virulence of the bacterium. The results showed that the LD 50 values of the WT SX, M2, M18, M2 (pRES-Riean_0790), and M18 (pRES-Riean_0653) were 4.48 × 10 8 , 5.46 × 10 9 , 5.97 × 10 9 , 3.99 × 10 9 , and 4.64 × 10 9 CFU, respectively. The LD 50 values of the mutants M2 and M18 were about 12 to 13 fold greater than that of the WT SX strain, suggesting that disruption of Riean_0790 or Riean_0653 may result in attenuation of the virulence of R. anatipestifer. The LD50 of the mutants M2 and M18 complemented with the shuttle expression plasmid pRES-Riean_0790 and pRES-Riean_0653 respectively, was not clearly different from that of the non-complemented mutant strains. In addition, as shown in Fig. 3, the bacterial loads in the blood of ducklings infected with the mutant strain M2 or M18 were decreased significantly as compared to those infected with the WT SX strain at 24 h post-inoculation (p < 0.0001), and the bacterial loads in the blood of ducklings infected with the complemented strain M2 (pRES-Riean_0790) or M18 (pRES-Riean_0653) were significantly restored (p < 0.0001), consistent with the restored hemolytic activities of the two complemented   Fig. 2 Hemolytic activity of the mutants M2 and M18 were restored by plasmid complementation. The WT SX strain, the mutant strains M2 and M18, and the complemented strains M2 (pRES-Riean_0790) and M18 (pRES-Riean_0653) were grown in TSB to the mid-logarithmic phase. The bacterial cell pellets were resuspended in phosphate-buffered saline and 10 μL of the cell suspensions were plated on blood agar. The hemolytic zones generated by these strains were observed. Yb2 (dba − strain) was used as a negative control showed that the bacterial loads of blood of ducklings infected with the complemented strain M2 (pRES-Riean_0790) or M18 (pRES-Riean_0653) were significantly restored as compared to those infected with M2 or M18. The error bars represent means ± standard deviations of eight ducklings. Asterisks indicate statistically significant differences between two groups (***p < 0.0001) strains. These results indicate that the hemolytic activities of these strains might not be directly related to virulence.

None of the recombinant proteins exhibited hemolytic activity
The results of our previous study showed that both the culture supernatants and OMPs of R. anatipestifer had hemolytic activities, indicating that the culture supernatants and OMPs may contain hemolysins [4]. To determine whether the identified genes encoded hemolysins, six genes predicted to encode OMPs or cytoplasmic proteins, especially those inserted in different mutants by transposon Tn4351, were expressed in E. coli cells, purified (Additional file 2), and the hemolytic activities were measured. The results showed that all of the tested recombinant proteins (rOmpA1467 (rRiean_1027), rRiean_1561, rRiean_1143, rRiean_0790, rRiean_0653, and rRiean_0317) had no hemolytic activity on duck blood agar. In addition, these six recombinant proteins did not generate pores in the duck erythrocyte membrane (Additional file 4), suggesting that none were hemolysins.

Discussion
In the present study, a total of 24 genes were identified and 31 mutants with reduced and one with increased hemolytic activity were screened by random Tn4351 transposon mutagenesis using R. anatipestifer strain SX as the parent strain. These genes or the encoded proteins may be directly or indirectly involved in the hemolytic activity of R. anatipestifer. The encoded proteins may act as hemolysin/cytotoxin molecules to directly lyse duck red blood cells. More likely, these genes or the encoded proteins may indirectly lyse the membranes of duck erythrocytes by (1) the production of hemolysin [20], (2) the export or secretion of hemolysin, (3) acting as a regulator of hemolytic activity; or (4) are other hemolysis-related proteins [21]. However, the exact roles of the genes identified in the present study or the encoded proteins conferring hemolytic activity to R. anatipestifer remain to be further confirmed. The ability of pathogenic bacteria to acquire iron in an animal host cell is a critical determinant for the establishment of infection. Increased levels of hemolytic activity and hemolysins were detected in the supernatants of iron-limited Vibrio cholerae cells [22]. R. anatipestifer has been reported to acquire iron and/or other nutrients from lysed erythrocytes to promote growth and proliferation both in vitro and in vivo [4]. Seven of the 24 genes identified in this study were predicted to be involved in iron utilization, suggesting that the ability of R. anatipestifer to acquire iron is related to its hemolytic activity.
In the mutant M4, Tn4351 was inserted into Riean_1561, which encodes a TonB-dependent receptor plug. TonB-dependent receptors of some bacteria are involved in iron acquisition and host infection. The TonB-dependent receptor TbdR1 of R. anatipestifer strain CH3 is involved in hemin iron acquisition and necessary for optimal bacterial virulence [23]. According to the Conserved Domain Database [24], the conserved domain Cir (COG1629) of Riean_1561 is predicted to be mostly involved in iron transport, while OMP_RagA_ SusC may be involved in iron utilization of R. anatipestifer strain SX.
In the mutant M10, Riean_1262 (yifB gene), which encodes a Mg chelatase, was inserted with the transposon Tn4351, which resulted in significantly decreased hemolytic activity as compared to that of the WT SX strain. However, the exact underlying mechanism involved in this process remains unclear. In Bacteroides fragilis lacking the yifB gene, ferrochelatase activity was significantly increased by approximately 2.2-fold as compared with the parent strain. In addition, the ability of B. fragilis to utilize heme or protoporphyrin IX for growth was greatly reduced in a ΔuroS mutant and this ΔuroS suppressive effect was enhanced by deletion of the yifB gene [25].
In the mutant M13, Tn4351 insertion in Riean_1661 resulted in significantly decreased hemolytic activity. Riean_1661 encodes the riboflavin biosynthesis protein RibD. In bacteria, the expression of genes involved in riboflavin biosynthesis is influenced by iron availability. Riboflavin is involved in overcoming iron-restrictive conditions in some species, which is probably achieved through increased iron bioavailability by reduction of extracellular iron, activities of iron uptake pathways, and hemolytic activity [26].
In the mutant M15, Tn4351 was inserted in Riean_1808, which encoded the RelA/ SpoT domain protein. In E. coli, RelA produces pppGpp (or ppGpp) from adenosine triphosphate and guanosine triphosphate, or guanosine diphosphate, and SpoT degrades ppGpp, but may also act as a secondary ppGpp synthetase. Iron was found to be involved in the regulation of the two activities of SpoT: synthesis and hydrolysis of (p)ppGpp. Iron limitation causes SpoT-dependent accumulation of ppGpp, which in turn stimulates the iron uptake capacity of the cell [27].
Tn4351 insertion in Riean_0932 in the mutant M19 led to defective hemolytic activity. ExbD1 encoded by Riean_0932, which is part of the TonB/ExbB/ExbD complex, is required for high-affinity iron transport in Gram-negative bacteria. The results of our previous study showed that two TonB/ExbB/ExbD systems of R. anatipestifer are involved in iron acquisition by hemin, hemoglobin, and holo-transferrin, and are necessary for optimal bacterial virulence [22].
In the mutant M23, Tn4351 was inserted in Riean_1386, which encodes a ferritin Dps family protein. Ferritin is the primary iron storage protein and is critical for iron homeostasis. Ferritin makes iron available for critical cellular processes, while protecting lipids, DNA, and proteins from the potentially toxic effects of iron [28]. Tn4351 insertion in Riean_1386 in the mutant M23 may lead to disordered iron storage and homeostasis.
In the mutant M24, transposon Tn4351 insertion in Riean_1546 led to an increase in hemolytic activity. Riean_1546 (hmgA gene) encodes homogentisate 1,2-dioxygenase, which can degrade homogentisic acid (HGA). In the absence of HmgA (hmgA mutant), HGA accumulates and oxidizes to form of HGA-melanin, and the HGA-melanin pigment secreted by Legionella pneumophila was found to confer ferric reductase activity, resulting in excessive reduction of iron [29,30]. In the present study, Tn4351 insertion in R. anatipestifer Riean_1546 may also lead to excessive reduction of iron, and the resulting iron deficiency may upregulate the expression of iron uptake-related genes, including hemolysin, and hemolytic activity, etc. The exact mechanism of negative regulation of Riean_1546 deletion on increased hemolytic activity of R. anatipestifer SX will be investigated in a future study.
In addition, in the present study, six recombinant proteins, including rRiean_1027, rRiean_1561, rRiean_1143, rRiean_0790, rRiean_0653, and rRiean_0317, showed no hemolytic activity, indicating that these proteins were not hemolysins. Whether one or more other proteins identified in this study were hemolysins remains unclear, but two predicted hemolysins (Riean_0620 and Riean_0415), according to the genomic annotations of the type strain ATCC11845 [31], were not identified, which may due to the insufficiency of the random transposon mutagenesis library used in this study. Moreover, it seems that transposon Tn4351 preferentially inserts into "TA" nucleotide sequences within the genome of R. anatipestifer, which led to the same Tn4351 insertion sites of the genomes of different mutants, as shown in this study. In addition, insertion of a transposon within an operon could result in transcription polarity of the downstream genes. Therefore, although random transposon mutagenesis is a useful tool to identify new functional genes in bacteria, this technology has its limitations.
In the present study, 24 genes were involved in the hemolytic activity of R. anatipestifer strain SX and seven were predicted to be related to iron utilization. These results will be helpful to clarify the molecular mechanism underlying the hemolytic activity of R. anatipestifer.