Role of (p)ppGpp in Viability and Biofilm Formation of Actinobacillus pleuropneumoniae S8

Actinobacillus pleuropneumoniae is a Gram-negative bacterium and the cause of porcine pleuropneumonia. When the bacterium encounters nutritional starvation, the relA-dependent (p)ppGpp-mediated stringent response is activated. The modified nucleotides guanosine 5’-diphosphate 3’-diphosphate (ppGpp) and guanosine 5’-triphosphate 3’-diphosphate (pppGpp) are known to be signaling molecules in other prokaryotes. Here, to investigate the role of (p)ppGpp in A. pleuropneumoniae, we created a mutant A. pleuropneumoniae strain, S8ΔrelA, which lacks the (p)ppGpp-synthesizing enzyme RelA, and investigated its phenotype in vitro. S8ΔrelA did not survive after stationary phase (starvation condition) and grew exclusively as non-extended cells. Compared to the wild-type (WT) strain, the S8ΔrelA mutant had an increased ability to form a biofilm. Transcriptional profiles of early stationary phase cultures revealed that a total of 405 bacterial genes were differentially expressed (including 380 up-regulated and 25 down-regulated genes) in S8ΔrelA as compared with the WT strain. Most of the up-regulated genes are involved in ribosomal structure and biogenesis, amino acid transport and metabolism, translation cell wall/membrane/envelope biogenesis. The data indicate that (p)ppGpp coordinates the growth, viability, morphology, biofilm formation and metabolic ability of A. pleuropneumoniae in starvation conditions. Furthermore, S8ΔrelA could not use certain sugars nor produce urease which has been associated with the virulence of A. pleuropneumoniae, suggesting that (p)ppGpp may directly or indirectly affect the pathogenesis of A. pleuropneumoniae during the infection process. In summary, (p)ppGpp signaling represents an essential component of the regulatory network governing stress adaptation and virulence in A. pleuropneumoniae.


Introduction
Actinobacillus pleuropneumoniae is a non-motile Gram-negative bacterium causing porcine pleuropneumonia, a highly contagious respiratory disease that is transmitted through aerosols supplemented with 1 mM diaminopimelic acid (DAP) (Sigma-Aldrich, St. Louis, MO, USA) and, when required, chloramphenicol (30 μg/ml). All the bacteria were cultured at 37°C.

Construction of a (p)ppGpp synthase deletion mutant and complemented strain
The strategy used for inactivation of the relA gene in A. pleuropneumoniae was as described previously [23]. A 900-bp DNA fragment of relA (646 bp-1546 bp, encoding amino acid residues 216 to 516 of the RelA protein) was amplified from genomic DNA of A. pleuropneumoniae strain S8 with primers P1 and P2 (Table 1, Fig 1). The PCR product was cloned into the suicide plasmid pEMOC2 between sites SalI and NotI. The resulting insertional plasmid, pEMOC2-ΔrelA, was electroporated into A. pleuropneumoniae S8. Recombinants were selected on TSA plates containing chloramphenicol (5 μg/ml). The ΔrelA strain was verified to have the plasmid inserted into the relA locus by PCR using primers P3 and P4 and DNA sequencing of the resulting amplicon. To construct the complemented strain, full-length relA gene with its signal peptide sequence was amplified from S8 genomic DNA with primers P5 and P6 ( Table 1, Fig 1). The PCR product was cloned between the EcoRI and the SacI sites of the shuttle vector pLS88. The recombinant plasmid, pLSrelA, was confirmed by DNA sequencing and electroporated into the S8ΔrelA strain. Transformants were selected on TSA plates containing kanamycin (50 μg/ml) and chloramphenicol (5 μg/ml). The complemented strain was confirmed by PCR and DNA sequencing of the amplicon, and named S8HB. Table 1. Strains, plasmids and primers used in this study.

Strains
Description Reference

S-8
A. pleuropneumoniae serovar 7 clinical isolate from the lung of a diseased pig in northern China [17] S-8ΔrelA Inactivation of S8 (relA) by insertion mutant This study S-8HB The complemented strain of A. pleuropneumoniae S8ΔrelA containing the relA gene This study E. coli β2155 thrB1004 pro thi hsdS lacZ M15 (F' lacZ M15 lacI q traD36 proA + proB + )dap::erm (Erm r ) [ 19] Plasmids pEMOC2 Transconjugation vector based on pBluescript SK with mobRP4, a polycloning site, Cm r , and transcriptional fusion of the omlA promoter with the sacB gene [20] pLS88 Broad-host-range shuttle vector from Haemophilus ducreyi; Str r Sm r Km r [21] pEMOC2-ΔrelA

Detection of intracellular (p)ppGpp
The production of (p)ppGpp in response to minimal medium was assayed as previously described [7]. The strains were grown over-night in TSB to early stationary phase (12 h), diluted back to an OD 600 of 0.2 with fresh TSB, and incubated in TSB for additional 2 h, at which point all strains had reached an OD 600 of 0.3. For nutrient deprivation, 2 ml of each culture were pelleted by centrifugation at 12,000 g for 5 min, and washed once with minimal medium (50 mM MOPS (morpho-linepropanesulfonic acid) pH 7.4, 1 mM MgSO4, 0.25 mM CaCl 2 , 19 mM glutamic acid, and 0.004 mM biotin, 10 mg NAD) and resuspended in 250 μl of minimal medium, 32 P (Perkin Elmer) was added to 100 μCi/ml, and the culture labeled for 1 h at 37°C. Fifty microliters of labeled culture were added to an equal volume of 2 M formic acid and placed on ice for at least 15 min. The mixture was centrifuged for 5 min at 16,000 g, and 3 μl of the supernatant were spotted directly onto polyethyleneimine (PEI) cellulose thin-layer chromatography plates (Sigma), dried, and developed in 1.5 M KH 2 PO 4 for 2.5 h. Nucleotides were visualized by autoradiography [24].

Growth experiments
All strains (S8, S8ΔrelA and S8HB) were first grown in 2 ml of TSB for 12 h and diluted with fresh TSB to OD 600 of 0.2. The diluted cultures were incubated at 37°C. The OD 600 was determined using an Eppendorf Biophotometer (Eppendorf, Hamburg, Germany) at various time points. The effect of relA on the viability of A. pleuropneumoniae was determined by counting the CFU of A. pleuropneumoniae at the indicated time points. One hundred microliter aliquots of culture were taken and serially diluted in TSB. After plating in duplicate on TSA plates containing relevant antibiotics and incubated at 37°C for 12 h, the CFU were counted. All experiments were repeated three times.

Scanning electron microscopy
The pellets of S8, S8ΔrelA and S8HB were harvested by centrifugation at 10,000 g after growth in TSB broth for 6 h, 12 h, 24 h and 36 h at 37°C. The harvested cells were washed twice with 0.1 M PBS buffer (pH 7.2) and fixed overnight using 2.5% glutaraldehyde at 4°C. Subsequently,

Quantitative biofilm assay
The microtiter plate biofilm assay is a static assay particularly useful for examining early events in biofilm formation [25]. Overnight cultures were diluted 100× with TSB, and 200 μl of the dilution added to the wells of a sterile 96-well microtiter plate (Costar 3599, Corning, NY, USA). After incubation for 36 h at 37°C, the wells were washed with 200 μl water to remove loosely adherent cells. Excess water was removed by inverting plates several times onto new paper towels. The wells were filled with 100 μl crystal violet (0.1%) and incubated for 2 min at room temperature. After removal of the crystal violet solution, the wells were washed with 200 μl water and dried in a 37°C incubator for 30 min. Subsequently, 100 μl ethanol (70%) was added to each well. Absorbance was measured at 590 nm.

Confocal laser scanning microscopy
The biofilm assay protocol mentioned above was used except that after washing off non-adherent bacteria, instead of adding crystal violet, 200 μl diluted LIVE/DEAD @ BacLight Tm Bacterial Viability Kit solution (Molecular Probes, Eugene, Oregon, USA) were added to stain the bacterial cells. Plates were incubated for 20 min at room temperature in the dark and washed with water. The wells were examined with a confocal microscope (TCS SP5, Leica Microsystems, Hamburg, Germany). SYTO 9 nucleic acid stain was excited at 488 nm and detected using a 520 nm filter. Propidium iodide was excited at 488 nm and detected using a 572 nm filter.
Physiological and biochemical effects of (p)ppGpp All strains (S8, S8ΔrelA and S8HB) were cultured on TSA plates overnight. Subsequently, bacteria were harvested using PBS, and centrifuged at 12,000 g for 5 min. Pellets were re-suspended to an OD 600 of 0.2 in the suspension medium provided in the kit. The tests were performed using API identification systems, including API 50 CH, API 20 E, API 20 NE and API ZYM, following the manufacturer instructions (BioMerieux).

RNA isolation and qRT-PCR
Total RNA was isolated from A. pleuropneumoniae S8 and S8HB strains grown to early stationary phase (12 h) in TSB broth for analysis of relA expression. They were harvested by centrifugation at 10,000 g at 4°C and diluted back to an OD 600 of 1.0 with fresh TSB medium. The RNeasy kit (Qiagen) was used to isolate RNA. RNA concentrations were measured spectrophotometrically at 260/280 nm (IMPLEN, Germany). Complementary DNA (cDNA) was synthesized using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Japan) following the manufacturer's instructions. Real-time PCR was performed using a Stratagene3000 system (Agilent Technologies, Germany). The reaction volume was 20 μl, containing 2 μl cDNA template, 10 μl 2μSYBR Green I (TaKaRa) and 0.8 μM of forward and reverse primers. PCR reactions were set up in triplicate. For all amplifications, the cycle conditions were 95°C for 2 min, followed by 40 cycles of 95°C for 15 s and 56°C for 1 min. This experiment was done with three biological replicates, and the average values were taken as the quantitative result. The gyrA and recF gene were used as internal control. Reaction mixtures lacking RNA were used as negative controls for each set of primers. The primers for amplifying cDNAs of gyrA, recF, relA are presented in Table 1. Relative expression values were calculated as 2 −4(CT target − CT reference) , where CT is the fractional threshold cycle [22].

RNA-sequencing analysis
Cultures of S8 and S8ΔrelA were grown to early-stationary-phase (12 h) in TSB. The cells were collected at 4°C, the RNeasy kit (Qiagen) was used to isolate RNA, and the Ribo-Zero™ rRNA Removal Kit for Gram-negative bacteria (EPICENTRE Biotechnologies) used to remove rRNA. The remaining RNA was quantified and examined for protein and reagent contamination with a Nanodrop ND-1000 spectrophotometer (

Differential expression analysis
The raw sequence reads were filtered using the Illumina pipeline. All of the low-quality reads, reads with adaptor contamination, and reads with only one copy were excluded from the analysis. The clean reads remaining were mapped to the reference sequence of A. pleuropneumoniae S8 (Genbank accession No. ALYN00000000.1).
To identify the genes affected by deletion of relA, the libraries were compared. The number of reads for each coding region was determined, the number of total reads was normalized between the libraries and the ratio of S8 to S8ΔrelA reads was calculated. Differentially expressed genes were detected as previously described [26], with a false discovery rate (FDR) threshold of 0.01 [27]. Differences with FDR 0.001 and log 2 Ratio absolute value 1 were set as the threshold for significant differences in gene expression.
The Blast2GO program was used to obtain GO annotations for molecular functions, biological processes and cellular component ontologies (http://www.geneontology.org). The Kyoto Encyclopedia of Genes and Genomes pathway database (http://www.genome.jp/kegg) was used for pathway assignments. The BlastN program (http://blast.ncbi.nlm.nih.gov/) was used to compare sequences with the A. pleuropneumoniae serovar 7 strain AP76 reference sequence (Genbank accession No. CP001091.1) for annotation.

Statistical analysis
Basic statistical analyses were conducted with the SPSS software (SPSS, Inc., Chicago, IL, USA). The Student's t test was used to determine the significance of the differences in the means between multiple experimental groups. The data were expressed as the mean +/-standard deviation, and values of P<0.05 were considered to be significant.

Results
Construction of an A. pleuropneumoniae S8 ΔrelA mutant and its complemented strain The relA insertion mutant of A. pleuropneumoniae S8, constructed with plasmid pEMOC2-ΔrelA (Fig 1), was confirmed by PCR and DNA sequencing, and designated as S8ΔrelA.
Complementation of the S8ΔrelA mutant was achieved using the plasmid pLSrelA, with transformants selected on plates containing chloramphenicol and kanamycin. The complemented mutant was designated S8HB.

The ΔrelA mutant fails to produce (p)ppGpp under nutrient deprivation conditions
To determine whether the ΔrelA mutant produces (p)ppGpp, all strains were subjected to nutrient deprivation, a condition which had been shown to induce (p)ppGpp [10,[28][29][30]. S8, S8ΔrelA and S8HB were incubated in minimal medium with 32 P for 1 h. As shown in Fig 2, the S8 and S8HB strains accumulated significant amounts of (p)ppGpp upon exposure to this starvation stress. In contrast, there was almost no detectable level in the S8ΔrelA mutant extracts. Plate counts of parallel bacterial cultures after 1 h incubation in minimal medium showed that all three strains remained viable during the experiment; thus, the absence of (p)ppGpp production in the S8ΔrelA strain was not due to bacterial death. This result indicated that the relA gene product is indispensable for (p)ppGpp production in A. pleuropneumoniae S8, and disruption of the relA gene results in the absence of (p)ppGpp.

Lack of (p)ppGpp resulted in abnormal growth under nutrient limitation and decreased viability under starvation conditions
Having ascertained that deletion of relA affects the production of (p)ppGpp in A. pleuropneumoniae S8, and given the vital role of the stringent response in bacterial survival, we investigated the impact of (p)ppGpp on cell growth and viability. As shown in Fig 3, OD 600 measurements indicated that S8ΔrelA grew slower than S8 when cultured in TSB. The growth pattern of S8HB was similar to that of S8ΔrelA, indicating lack of complementation of the slow-growth phenotype by the cloned relA gene. To test whether (p)ppGpp is required for viability, we determined cell counts for S8, S8ΔrelA and S8HB cultured in TSB over 36 h. As shown in Fig 4, all strains exhibited a decline in viability from 6 h to 36 h. However, S8 exhibits a gradual decline in viable counts from 12 h to 36 h, whereas, S8ΔrelA and S8HB showed a rapid decline during late stationary phase; no viable cells were detected for either strain at 36 h. The results indicated that the relA gene expressed on plasmid pLSrelA was only able to partially restore growth of S8ΔrelA in TSB. However, it was unable to complement viability: the number of colonies decreased approximately seven orders of magnitude relative to the S8 strain. These results indicated that (p)ppGpp contributed to prolonged survival under conditions of nutrient limitation.
Cell morphology of A. pleuropneumoniae S8 is dependent on (p)ppGpp availability While examining the role of (p)ppGpp in morphology, we found that the S8ΔrelA mutant grew as a homogeneous population of non-extended (i.e. short) rods from 12 to 36 h in liquid medium (Fig 5). The non-extended phenotype is in sharp contrast to S8, which grew as a population of significantly longer rods under the same growth conditions. The extended-rod phenotype was not restored by introduction pLSrelA.
Deletion of (p)ppGpp synthases affects the biofilm formation of A. pleuropneumoniae S8 We tested whether (p)ppGpp plays a role in biofilm formation by examining the biofilm-forming ability of S8, S8ΔrelA and S8HB in polystyrene microtiter plates. As shown in Fig 6, S8 and S8ΔrelA produced minor and dense biofilms, respectively. Biofilm quantification was confirmed by confocal laser scanning microscopy (Fig 7). These results indicated that (p)ppGpp is involved in A. pleuropneumoniae biofilm production. The S8HB strain yielded less biofilm than S8ΔrelA, although it produced much more biofilm than WT S8 strain (Figs 6 and 7). This   Metabolic effects of (p)ppGpp in A. pleuropneumoniae S8 To test whether the (p)ppGpp affects the metabolic ability, S8, S8ΔrelA and S8HB were cultured in API identification systems. The S8 strain utilizes citrate and all the twelve carbon sources listed in Table 2, and it produces urease. However, S8ΔrelA neither uses those carbon sources as substrate, nor produces this enzyme. Introduction of pLSrelA into S8ΔrelA resulted in reversal to wild-type citrate utilization and production of urease, but the production of arginine dihydrolase was not restored (Table 2). No other differences in substrate utilization or enzyme production were seen between S8, S8ΔrelA and S8HB.

Validation of expression level of relA in S8 and S8HB
To compare the expression levels of relA in S8 and S8HB, qRT-PCR analysis was performed three times for RNA extracts from the three biological replicates. Two internal reference genes were used to evaluate the expression level. The results indicated that the plasmid encoded relA gene had lower expression levels in S8HB than in S8 neither the glyA nor recF gene as reference gene (Fig 8).
Global gene expression in the early stationary phase is different in A. pleuropneumoniae S8 strain compared to the relA-deficient mutant To study the effects of the effector molecule (p)ppGpp on a genome-wide scale, the transcriptomes of S8 and S8ΔrelA were compared by RNA-seq. The overwhelming majority of the differentially expressed genes in this strain were upregulated (94%). S1 Table shows the differentially expressed genes between S8 and its isogenic S8ΔrelA mutant. A total of 405 genes were found to be differentially expressed in the S8ΔrelA mutant. Among these, 380 genes were upregulated and 25 genes were downregulated compared to the WT S8. The differently expressed genes were annotated according to the COG database and are summarized in Fig 9. The most prominently upregulated transcription was in genes involved in amino acid transport and metabolism (45/405); cell wall/membrane/envelope biogenesis (45/405); translation,

Discussion
In a previous study by Lone et al. [34], genes involved in the stringent response were shown to be up-regulated in a malT mutant of A pleuropneumoniae grown in porcine bronchoalviolar (p)ppGpp Affect Actinobacillus pleuropneumoniae S8 lavage fluid and it was suggested that that this response may play a role in pathogenesis of this bacterium. However, no further investigations were made regarding the roles of individual genes. In order to investigate the role of RelA, a stringent response regulatory protein responsible for synthesis of (p)ppGpp, in A. pleuropneumoniae physiology, we constructed a relA mutant of the serovar 7 clinical isolate S8 [17]. We confirmed the mutant, S8ΔrelA, to be deficient in (p)ppGpp production compared to the WT S8 strain when incubated in minimal medium (Fig 2), and complementation of the mutant by plasmid encoded relA (pLSrelA) was able to restore WT levels of (p)ppGpp under this growth condition. The S8ΔrelA mutant had a reduced growth rate in TSB compared to S8 (Fig 3). This is consistent with previous studies describing relA-deficiency leading to growth limitation [35]. However, complementation with the relA gene (strain S8HB) only led to partial restoration of growth limitation, possibly due to differences in expression levels of relA from the plasmid. Indeed, qRT-PCR analysis of cultures grown in TSB for 12 h revealed that the expression level of relA in S8HB was less than in S8 (Fig 8), suggesting lower concentrations of (p)ppGpp in the complemented mutant under these conditions. It was also found that the viability of S8ΔrelA decreased with prolonged culture in TSB compared to S8, i.e it had a stationary phase growth defect. At the 24 h time point, the CFU of S8ΔrelA was seven orders of magnitude less compared to S8. No viable S8ΔrelA were recovered from the TSB cultures at 36 h when plated on TSA. Other studies have documented that the availability of (p)ppGpp affects bacterial survival, i.e. relA deficiency results in the death of the cells. For example, a Bordetella pertussis relA-deficient mutant also lost viability more rapidly than the wild-type [36]. In addition, (p)ppGpp is reported to be responsible for significant changes in gene expression, leading to cessation of growth and induction of specific stress responses [12,[37][38][39][40]. Our results indicated that (p)ppGpp contributes to prolonged survival of A. pleuropneumoniae S8 under conditions of nutrient limitation.
In stationary phase in rich medium (TSB), S8 exhibited a change in morphology such that it produced prolonged rods and with an increase in volume (Fig 5), as has previously been seen in other bacteria [41]. Both, S8ΔrelA and S8HB had similar morphology to that of non-stationary phase S8 cells (i.e. short rods), even at 36 h in TSB. Transcriptome analysis showed that genes involved in cell wall/membrane/envelope biogenesis were up-regulated in S8ΔrelA compared to S8, consistent with the morphological changes observed.
More recently, (p)ppGpp has been reported to be involved in a number of non-stringent processes, including virulence, and biofilm formation [7,36,42]. Bacterial biofilm formation is a complex, multifactorial process requiring genes involved in adherence, metabolism, quorum sensing, and stress responses. When we tested the ability of S8, S8ΔrelA and S8HB to form biofilms, we found increased biofilm formation by S8ΔrelA compared to S8, suggesting that (p) ppGpp negatively regulates biofilm formation of A. pleuropneumoniae S8. These results are in contrast to those found with Agrobacterium [43], Vibrio cholerae [7], E. coli [44], B. pertussis [36], where (p)ppGpp promoted biofilm formation. The reasons for this apparent discrepenancy are not clear. The transcriptome data revealed that the genes known to be important for the formation of biofilm (pgaA log 2 = 2.73, pgaB log 2 = 1.37, pgaC log 2 = 2.64) were all up-regulated in S8ΔrelA. In A. pleuropneumoniae, biofilm formation has been shown to be part of the extracytoplasmic stress response, with genes of the pga operon positively regulated by ϭ E [45]. In addition, ϭ E regulates expression of numerous genes encoding proteins involved in stress response and in reparation and maintenance of the bacterial envelope, and is itself negatively regulated by the anti-sigma factor, RseA [45][46][47]. However, the rseA gene (APP7_0419) was up-regulated (log 2 = 2.04) in S8ΔrelA, suggesting that ϭ E was not the only regulator responsible for biofilm formation in the relA mutant.
The stringent response occurs when bacteria encounter nutrient-limited environment and initiates changes in gene regulation in order to maximize the utilization of available resources [30]. In contrast to the growth and morphology results obtained in growth in TSB broth, but in agreement with the detection of WT levels of (p)ppGpp in the S8HB strain when incubated in minimal medium, the metabolic defects (including changes in carbon source utilization and all biochemical characteristics except for arginine dihydrolase; Table 2) in S8ΔrelA were complemented in S8HB as determined by API analysis, which uses a minimal medium base for inoculation of the test strips. These results suggest that expression of relA from pLSrelA may be higher in minimal medium than in late log phase in TSB.
Besides the differently expressed genes involved the metabolism, some virulence-related genes were also differently expressed (S1 Table). It has been shown that (p)ppGpp has a direct role in expression of virulence in other pathogenic bacteria [2,[31][32][33]. Some virulence factors involved in adhesion, acquisition of essential nutrients, induction of lesions, avoiding the host's defense mechanism and persistence have been up-regulated in S8ΔrelA. These results suggest that the virulence of S8ΔrelA might be higher than S8, however further experiments are required to validate this.
The gene expression profiles for S8 and S8ΔrelA were also analysed to determine if any of the differentially regulated genes could explain the various phenotypes in the relA mutant. Some significantly differentially expressed transcripts were detected. Among these, the genes involved in ribosomal structure and biogenesis, amino acid transport and metabolism, and translation cell wall/membrane/envelope biogenesis were all up-regulated in the mutant. This is consistent with a previous report that cells entering stationary phase without (p)ppGpp production have a proteomic profile similar to that during growth [48]. This suggests that the transcriptional program of S8ΔrelA is geared towards exponential growth, and may be poorly adapted to inducing systems that are required to withstand stress. Also, the (p)ppGpp deficient strain failed to suppress unnecessary gene expression under nutrient-limiting conditions, which may contribute to diminished survival. A balanced carbon flux is extremely important for viability of bacterial cells [14].
In conclusion, our results show that during stationary-phase growth in TSB, (p)ppGpp is critical for A. pleuropneumoniae S8 to adapt to the nutrient-limiting conditions. Several phenotypes were changed owing to the deficiency of (p)ppGpp. By analyzing the transcriptome of a relA mutant, we have demonstrated a set of genes that may be regulated by A. pleuropneumoniae during adaptation to changing nutrient levels. Many differently expressed genes have been identified, and further study will clarify the role of selected genes in the pathogenesis of porcine pleuropneumonia. Some of the genes identified here may serve as new targets in drug and vaccine development.
Supporting Information S1 Table. The differently expressed genes of S8ΔrelA compared to S8 by the deletion of the relA gene. All the differently expressed genes were categorized according to the COG database. (DOCX)