Microbiology Papers in Press. Published September 26, 2014 as doi:10.1099/mic.0.081992-0

The

Klebsiella

pneumoniae

yfiRNB

Operon

Affects

Biofilm

Formation,

Polysaccharide Production and Drug Susceptibility

Running title: K. pneumoniae yfiRNB genes affect biofilms and survival Contents category: Cell and Molecular Biology of Microbes Mónica G. Huertas, Lina Zárate, Iván C. Acosta, Leonardo Posada, Diana Cruz, Marcela Lozano, María M. Zambrano* Molecular Genetics, Corporación Corpogen, Carrera 5 #66A-34, Bogotá, Colombia. *corresponding author, email: [email protected]; Tel: +57-1-8050122, Fax: +57-1-348-4607

Summary word count: 215 Main text word count: 4,940 Tables: 2 Figures: 6

1

Summary Klebsiella pneumoniae is an opportunistic pathogen important in hospital-acquired infections, which are complicated by the rise of drug resistant strains and the capacity of cells to adhere to surfaces and form biofilms. In this work, we carried out an analysis of the genes in the K. pneumoniae yfiRNB operon, previously implicated in biofilm formation. The results indicate that in addition to the previously reported effect on type 3 fimbriae expression, this operon also affects biofilm formation due to changes in cellulose as part of the extracellular matrix. Deletion of yfiR resulted in enhanced biofilm formation and an altered colony phenotype indicative of cellulose overproduction when grown on solid indicator media. Extraction of polysaccharides and treatment with cellulase were consistent with the presence of cellulose in biofilms. The enhanced cellulose production did not, however, correlate with virulence as assessed using a Caenorhabditis elegans assay. In addition, cells bearing mutations in genes of the yfiRNB operon varied with respect to the wild type control in terms of susceptibility to the antibiotics amikacin, ciprofloxain, imipenem and meropenem. These results indicate that the yfiRNB operon is implicated in production of exopolyscaccharides that alter cell surface characteristics and the capacity to form biofilms, a phenotype that does not necessarily correlate with properties related with survival such as resistance to antibiotics.

Keywords: Klebsiella pneumoniae, biofilm, cellulose, c-di-GMP, yfiRNB, antibiotic resistance Introduction Healthcare-associated infections (HAI) are a major public health problem worldwide, being of particular importance the infections caused by multiresistant bacteria such as Klebsiella pneumoniae. This microorganism, which can be found in the gut and nasopharynx, is also a major opportunistic pathogen within the Enterobacteriaceae family. K. pneumoniae causes a large percentage of community-acquired diseases and HAI, like pneumonia, septicemia, urinary tract and wound infections, especially in infants, the elderly and in immunocompromised patients (Pan et al., 2013; Podschun &

2

Ullmann, 1998). An important virulence factor is its capacity to adhere to and form biofilms on different surfaces, including indwelling medical devices. It has been reported that biofilms can be involved in approximately 65% of bacterial infections, allowing cells to persist and lead to chronic infections and increased resistance to antibiotic treatment (Hoiby et al., 2011).

Bacterial biofilms are surface-associated communities characterized by the production of an extracellular matrix that contains extracellular polymeric substances (EPS) that differ in composition and physical properties, depending on the microbial species (Kostakioti et al., 2013). Biofilm formation in K. pneumoniae is affected by mutations in genes responsible for capsule biosynthesis and surface exopolysaccharides (Balestrino et al., 2008; Boddicker et al., 2006), genes for synthesis of cell surface type 3 fimbria (Di Martino et al., 2003; Langstraat et al., 2001), by mutations that affect quorum sensing (Balestrino et al., 2005) and by the intracellular levels of the second messenger bis-(3’5’)-cyclic dimeric guanosine monophosphate (c-di-GMP) (Johnson & Clegg, 2010; Johnson et al., 2011; Wilksch et al., 2011). The cellular levels of c-di-GMP are modulated by diguanylate cyclase (DGC) enzymes and phosphodiesterases (PDE), which catalyze its synthesis and degradation, respectively (Boyd & O'Toole, 2012; Romling et al., 2013). In K. pneumoniae the c-di-GMP turnover enzymes YfiN and MrkJ have been shown to modulate type 3 fimbriae expression via the transcriptional regulator MrkH, which contains a c-di-GMP-binding PilZ domain (Johnson & Clegg, 2010; Johnson et al., 2011; Wilksch et al., 2011). The gene yfiN is part of an operon together with the upstream proposed negative regulator yfiR, which codes for a hypothetical protein of 177 residues, and yfiB, an apparent outer membrane lipoprotein proposed as a sensor of the YfiBNR system in Pseudomonas aeruginosa (Malone et al., 2012). The present work analyzes in more detail genes of the yfiRNB operon in K. pneumoniae and their effect on biofilm formation, and shows that this operon also affects production of EPS cellulose, in addition to the previously identified effect on fimbria. Mutations in genes of this operon therefore result in changes in the extracellular matrix that have unexpected effects on cell properties such as resistance to antibiotics.

Methods

3

Bacterial strains, plasmids, and growth conditions. Strains and plasmids used are shown in Table 1. Cells were grown either in liquid Luria Bertani (LB), BHI (Becton Dickinson) supplemented with 2mM MOPS (Sigma-Aldrich) or Muller Hinton (MH, OxoidTM), at 37°C with agitation at 200 rpm. Solid medium was made by adding 1.5% agar (DIFCO). When needed, antibiotics were added at the following concentrations: 100µg ml-1 ampicillin (Ap), 50µg m-1 kanamycin (Km), 20 µg ml-1 tetracycline (Tc), 20µg ml-1 chloramphenicol (Cm), (Sigma-Aldrich). The concentrations of

antibiotics

imipenem

(Imp),

amikacin

(Amk),

ciprofloxacin

(Cip)

and

trimethoprim/sulfamethoxazole (TS) and meropenem (Mpm) (Sigma-Aldrich) varied depending on the assay (diluting stock concentrations), and were determined following the protocols of the Clinical and Laboratory Standards Institute (CLSI) guidelines (2013). Phenotypes on solid media were analyzed by diluting 18h overnight cultures 1/100 in fresh medium and spotting 10µL onto LB agar without salt and supplemented with 40µg ml-1 Congo red and 20µg ml-1 Coomassie blue, and on LB agar with 0.2µg ml-1 calcofluor (fluorescent brightener 28, Sigma-Aldrich) to observe cellulose fluorescence under UV light (Solano et al., 2002; Zogaj et al., 2003). Phenotypes were analyzed after growth for 18h at 37ºC and 28ºC. Biofilm formation assay K. pneumoniae mutants defective in biofilm formation were identified as described previously by screening a miniTn10Km insertion mutant library of strain CG217 (Suescun et al., 2006). All biofilm assays were done by growing cells in LB medium or BHI+MOPS for 18h at 37ºC, in 96-well, flat bottom, non-tissue culture treated polystyrene plates (MICROTESTTM 96, BD Falcon). Biofilms were stained with 0.1% crystal violet and quantified by measuring OD at 595nm, as described (O'Toole & Kolter, 1998). All assays were done using three biological replicas, and repeated independently at least once.

Mutant construction and complementation analysis Deletions of yfiR, yfiN and yfiB genes were constructed in wild type (WT) K. pneumoniae LM21gfp, which bears a egfp marker, by allelic replacement with a Km resistance gene using pKOBEG199 (Balestrino et al., 2005). Briefly, the Km resistance cassette was

4

PCR-amplified from plasmid pKD4 with 60-nt forward and reverse primers specific for each target gene that included sequences homologous to the Km resistance gene and FRT sites, as reported (Balestrino et al., 2008; Datsenko & Wanner, 2000). 100 ng of each PCR fragment, purified using the QIAquick PCR purification kit (Qiagen), were transformed into K. pneumoniae LM21gfp cells harboring the Red recombinase expression plasmid pKOBEG199 and mutants were selected at 37ºC on LB agar containing Km. In order to avoid polar effects, the Km resistance cassette was excised via the FRT sites using the FLP helper plasmid pCP20 (Cherepanov & Wackernagel, 1995). Gene deletions were verified by PCR amplification and sequence analysis. All primers used are listed in Table S1. For complementation analysis, genes were PCR-amplified from WT K. pneumoniae LM21gfp using primers bearing restriction sites (Table S1), cloned by restriction and ligation, first into pCR®2.1TOPO® vector (Invitrogen), and then released by restriction enzyme digestion and ligated to pBAD18-Cm (Guzman et al., 1995), generating pBAD18-yfiR, pBAD18-yfiN and pBAD18-yfiB (Table 1). The constructs were confirmed by PCR and sequence analysis, and then transformed into LM21gfp and each deletion mutant. All restriction enzymes were from New England Biolabs (USA). Carbohydrate measurement and cellulase treatment The polysaccharide content was measured using the anthrone – sulfuric acid method, with some modifications (Laurentin & Edwards, 2003; Raunkjær et al., 1994). Overnight cultures grown in LB broth were used to spot 20 µl colonies onto LB agar plates. After incubation at 37ºC for 48 h, cells were scraped and resuspended in 300 µl 1N NaOH, heated at 80°C for 30 min and centrifuged at 10,000 rpm, 4°C for 15 min. 40 ul of each supernatant were added to 100 ul of anthrone reagent (0.2% anthrone in 75% H2SO4), (Sigma Aldrich) placed in 96-well plates (Corning - Costar), and incubated for 5 min at 92ºC, followed by 15 min at room temperature and 15 min at 45ºC. Absorbance was measured at 595 nm. A glucose standard curve was used for quantification, providing an estimate of hexose sugar concentration in each sample. Sugar concentration was standardized based on soluble protein concentration determined using 800 uL of each cell suspension and the Bradford method (595 nm) (Barnhart et al., 2013). Assays were performed with three biological replicas and repeated independently at least once.

5

The presence of cellulose in biofilms was estimated by assessing the effect of cellulase treatment on biofilms, as described (Solano et al., 2002). Briefly, biofilms were formed by growing cells in LB for 24h in 96 well Calgary Biofilm Device (CDB) plates (NuncImmuno TSP). After incubation, the lid pegs with biofilms were removed, washed in PBS, placed in 0.05M citrate buffer, pH 4.6, with 0.1% cellulase (Sigma Aldrich), and incubated for 48h at 45ºC. After digestion, biofilms were quantified by staining with crystal violet and measuring OD at 595nm. Quantitative real time (q-RT) PCR assays Total RNA was extracted from cultures grown in LB broth for 6, 10 and 18 hours using the Direct-zol

TM

RNA MiniPrep Kit (Zymo Research), treated with RQ1 RNase-Free

DNase (Promega), quantified by spectrometry (NanoDrop 2000, Thermo Scientific) and used to synthesize cDNA (three per RNA) using SuperScript III (Invitrogen). cDNA was quantified using SYBR Green (Life Technologies), and 150 ng were used for RT PCR analysis in a LightCycler 1.5 System (Roche), by preheating at 95°C for 10 min, followed by 40 cycles of amplification and quantification: 95°C for 10s, annealing at 62°C for 10s, extension at 72°C for 60s, and quantification at 83°C for 60s. Relative expression was obtained using the 16S rRNA housekeeping gene from K. pneumoniae and the equation proposed by Pfaffl (Bustin, 2010; Pfaffl, 2001). Caenorhabditis elegans Killing Assay Assays were done as described (Fuursted et al., 2012), by spreading 50 uL drops of overnight cultures of either E. coli OP50 or K. pneumoniae LM21gfp WT, ∆yfiR, ∆yfiN and ∆yfiB mutant strains on NGM (Nematode growth media) agar plates (Stiernagle, 2006). After drying, 30 adult (stage L-4) C. elegans N2 (wild type) were added to each plate. The plates were incubated at 20ºC and scored daily for dead nematodes using a stereoscope (Motic SMZ 168 BN). Nematodes were age-synchronized with bleach, and transferred every two days to fresh plates to eliminate overcrowding by progeny until they laid no further eggs (Bialek et al., 2010; Fuursted et al., 2012; Srinivasan et al., 2012). Worm mortality was scored over time for approximately 20 days, with death defined when a worm no longer responded to touch. The experiments were performed in triplicate at least three times on separate days. Bacterial colonization of C. elegans was determined at 12 hours after feeding by picking at least 5 worms, placing them on an agar plate containing 100 µg ml-1 gentamicin to remove surface bacteria, and then 6

washing them in 5 µl of M9 buffer. Worms were placed on microscope slides in 2% agarose with sodium azide as an anesthetic, a coverslip was placed over the agarose and worms were examined by fluorescence microscopy using a Axioscop 40 microscope. Images were obtained using a 5X/1.3 objective.

MIC and MBEC assays The minimal inhibitory concentrations (MICs) were determined by microdilution in 96 well plates (Corning®Costar®), following the CLSI standards (2013). Briefly, overnight cultures were grown in MH broth at 37°C with shaking at 200 rpm, then each inoculum was adjusted to 0.5 McFarland, diluted 1/100 into fresh medium with antibiotic in each well of a 96 microtiter plate and incubated at 37ºC for 18h without agitation; all assays were performed in duplicate. The MIC was defined as the lowest concentration of antibiotic that inhibited bacterial growth.

The minimum biofilm eradication concentration (MBEC) was evaluated by adjusting overnight cultures, grown in MH broth at 37°C with shaking at 200 rpm, to 0.2 McFarland standard and inoculating 150µl of this suspension into each well of a CBD plate (Thermo NUNC). After incubation for 24h at 37˚C, the lid with pegs was removed, washed in MH broth, transferred to fresh medium with antibiotic and incubated for 8 h at 37˚C. After washing again the lid was transferred to recovery medium without antibiotic. The lowest concentration of antimicrobial agent that prevents visible growth in the recovery medium was defined as the MBEC (Harrison et al., 2010; Singla et al., 2013).

Microscopy and image analysis Biofilms were formed by placing microscope slides (Menzer Glaser, 1,0 mm thick in dimensions of 26x76 mm) in 50-ml Falcon tubes containing strains LM21 ∆yfiR, ∆yfiN, ∆yfiB and the WT LM21gfp control grown in 25ml LB broth at 37ºC for 18 hours without shaking. Slides with biofilms were analyzed by microscopy with an Olympus U-TB/90 confocal microscope equipped with an argon laser and detectors and filter sets for monitoring egfp fluorescence (excitation 488 nm, emission 517 nm). Images were obtained using a 60X/1.35 objective and 3 representative images were taken for analysis. Image visualization was obtained using the software DAIME (Digital image Analysis In Microbial Ecology) (Daims et al., 2006). Quantification of biofilm biomass

7

was performed using COMSTAT software (Heydorn et al., 2000) by taking approximately 20-50 image stacks (three image stacks from each slide), acquired randomly for each strain. Bioinformatics analyses Analysis of the original transposon insertion site, of the yfiRNB operon and of each sequenced mutant allele was done using blastn and tblastn against the NCBI database (http://www.ncbi.nlm.nih.gov/). Putative bcsA sequences were identified by first doing an in silico search for PilZ domain-containing sequences. A PilZ domain profile was generated using HMMER (http://hmmer.janelia.org) and sequences reported in the Pfam database, and this profile was then used to search for domains in four K. pneumoniae genomes: Kp342 (NC_011283), MGH 78578 (NC_009648), NTUH-K2044 (NC_012731) and HS11286 (NC_016845.1). The two identified bcsA sequences and corresponding operon genes were further analyzed by multiple alignments of these regions using the same four K. pneumoniae genomes. Blastp and TBlastn searches were carried out with all the genes from the cellulose operon in order to determine orthologous and possible paralogous genes in other K. pneumoniae strains and other bacteria belonging to the Klebsiella clade using the following genomes: E. coli. K-12 substrain MG1655 (NC_000913.2) Salmonella enterica subsp enterica serovar Enteritidis (NC_011294.1) and S. enterica subsp. enterica serovar Typhimurium str. LT2 (NC_003197.1). Homologous proteins were aligned and phylogenies were reconstructed using the Geneious software (http://www.geneious.com).

Statistical analyses All statistical analyses were performed using the SPSS statistical package (IBM). Data were analyzed by ANOVA and the Student’s t-test. A P-value < 0.05 was considered significant. Results Mutations in the yfiRNB operon affect cellulose production The yfiRNB operon (Fig. 1a) has been previously implicated in biofilm formation in K. pneumoniae due to mutations that affect synthesis of c-di-GMP, which is involved in the

8

regulation of type 3 fimbriae expression (Johnson & Clegg, 2010; Johnson et al., 2011; Wilksch et al., 2011). Consistent with this, we identified a mutant via transposon mutagenesis of K pneumoniae clinical isolate CG217 (CG217 M39) (Suescun et al., 2006) that mapped to the yfiR gene, which negatively controls the YfiN DGC protein, and thus formed a biofilm more robust than that of the parental strain (Fig. 2). This mutant strain also showed differences in cell morphology when grown on indicator media. It had a rugose and dark red colony morphology when grown on solid LB containing Coomassie blue + Congo red, and growth on solid LB supplemented with calcofluor led to fluorescent colonies distinct from the parental strain and suggestive of cellulose overproduction in the extracellular matrix (Fig. S1) (Solano et al., 2002; Zogaj et al., 2003). Deletion mutants were constructed for each of the genes of the yfiRNB operon in the K. pneumoniae wild type (WT) GFP-tagged strain, LM21gfp (Balestrino et al., 2008). The resulting mutants, LM21 ∆yfiR, LM21 ∆yfiN, and LM21 ∆yfiB, were confirmed by PCR amplification and sequence analysis of each locus. As seen in Fig. 2 strain LM21 ∆yfiR showed enhanced biofilm formation, consistent with the original CG217 M39 insertion mutant strain. In contrast, the biofilms produced by mutants LM21 ∆yfiN and LM21 ∆yfiB were similar to that of the parental WT strain. Complementation with wild type genes sequences cloned in pBAD18Cm (Guzman et al., 1995) largely restored the biofilm formation phenotype of the strain lacking yfiR and resulted in slightly enhanced biofilms when yfiN was introduced for complementation of the corresponding deletion strain, the latter probably due to an increase in gene copy number when present on pBAD18Cm (Fig. 2). The empty vector control did not have any impact on biofilm formation in the strains tested and the various mutations did not affect cell growth (data not shown). A deletion of the entire yfiRNB operon also resulted in reduced biofilm formation with respect to the parental strain, as has been previously reported (not shown) (Wilksch et al., 2011). In addition to changes in biofilm formation, strain LM21 ∆yfiR also showed colony morphologies consistent with the observations for the CG217 M39 mutant when plated on LB with either Coomassie blue + Congo red or with calcofluor, again indicative of changes in the extracellular matrix due to the presence of cellulose (Fig. S1). Confocal scanning laser microscopy (CSML) was used to analyze 18h biofilm structures. The ∆yfiR mutant produced a more robust, highly compact and dense structure when compared to the WT control cells (Fig. 3). The ∆yfiN mutant showed decreased cell

9

surface adherence and fewer attached bacteria, consistent with biofilm measurements (Fig. 2). Finally, the ∆yfiB mutant showed a uniform but thin biofilm relative to the WT strain. Images were analyzed using COMSTAT (Heydorn et al., 2000) for biomass distribution, average thickness, roughness coefficient and substratum coverage, and the data obtained support the above descriptions (Table S2).

Cellulose as part of the extracellular matrix in K. pneumoniae The observed phenotypes on solid media prompted us to examine for the presence of cellulose in the extracellular matrix. The amount of exopolysaccharide was therefore quantified using the anthrone – sulfuric acid reaction (Laurentin & Edwards, 2003; Raunkjær et al., 1994), which measures total carbohydrates by the amount of hexose released. Cells lacking the yfiR gene released more glucose than both the WT strain LM21gfp or the ∆yfiB and ∆yfiN mutant strains, indicative of more polysaccharide production. The complemented strains bearing plasmids with wild type alleles were similar to one another and even higher than the WT strain, particularly for the strain complemented with the DGC gene yfiN (Fig. 4a). The high levels observed for the yfiRcomplemented strain could be due to altered expression that leads to imbalances in control and appropriate function of the operon, as has been proposed (Malone et al., 2010). The presence of cellulose in biofilms was also assessed by treating 24h biofilms with cellulase as an indirect measurement of extracellular matrix composition (Solano et al., 2002). In all cases biofilms were affected by treatment with cellulase, as can be seen by the decrease in crystal violet staining when compared to untreated controls (Fig. 4b). The reduction in biofilm staining was slightly greater for cellulase-treated biofilms of the WT and yfiN-complemented strains, an observation that contrasts with the apparent increase of cellulose in yfiR cells. This is probably due to the enzyme’s activity being limited to the outermost surface cells and its restricted accessibility to the interior of the more highly compact multicellular structures of the yfiR mutant biofilms.

Genomic and transcriptional analysis To investigate the possible mechanism underlying the observed changes in cellulose production associated with the ∆yfiR strain, we looked at expression of the DGCencoding gene yfiN and of putative cellulose synthase genes present in the K. pneumoniae genome using q-RT PCR. Two loci containing putative cellulose synthesis genes were first identified in the sequenced K. pneumoniae genomes by looking for PilZ

10

domain-containing sequences (Fig. 1b). One of these consists of two divergently oriented operons, bcsABZC and bcsEFG, and was found to be highly conserved in four K. pneumoniae genomes (Kp342, MGH 78578, NTUH-K2044 and HS11286) and in other Enterobacteriaciae examined, E. coli str. K-12 substrain MG1655, Salmonella enterica subsp. enterica serovar Enteritidis and S. enterica subsp. enterica serovar Typhimurium str. LT2. The second locus was separated by 108 bases and consisted of an apparently duplicated copy of the bcsABZC operon, since it was found only in the four K. pneumoniae genomes and was absent from other Enterobacteriaciae analyzed. Both operons contained copies of bcsA (named bcsA1 and bcsA2), each harboring a glycosyl transferase 2 family domain, a cellulose synthase domain and a c-di-GMPbinding PilZ domain, the latter characteristic of some effector proteins from the c-di-GMP signal transduction network (Romling et al., 2013). Expression analysis of the DGC-encoding yfiN gene and both bcsA copies was carried out with RNA extracted from 6 and 10-hour WT and mutant strain cultures, when cells are likely to express genes that can affect biofilm formation. As control, we also analyzed a strain with a deletion of ycdT, another DGC-encoding gene in the K. pneumoniae genome (Cruz et al., 2012). The q-RT PCR results showed that in a strain lacking yfiR the yfiN gene was overexpressed approximately 7 and 12 times, in 6 (not shown) and 10h cultures, respectively, when compared to the WT strain (Fig. 5). Expression of yfiN was not altered significantly in either a strain bearing a deletion of ycdT or in the strain lacking yfiB. Similar results were obtained using 6h cultures (not shown). The overexpression of yfiN in the absence of yfiR indicates regulation at the transcriptional level and differs from expression of the putative cellulose synthase genes bcsA1 and bcsA2, which showed no significant difference with respect to the WT strain (Fig. 5). Together, these q-RT PCR results indicate that cells lacking the negative regulator yfiR result in overexpression of yfiN, but not of the bcsA genes putatively involved in cellulose synthesis. Cell susceptibility and survival The observed changes in cell surface characteristics raised the possibility that the yfiRNB operon could also alter the cell’s capacity to survive under adverse environmental conditions. Given that K. pneumoniae is a well-known opportunistic pathogen associated with HAI and resistance to antibiotic treatment, we wanted to see if

11

the yfiR mutant, which produces more robust biofilms in vitro and more cellulose, would also present enhanced resistance to antibiotics. We therefore assessed the various strains’ resistance to antibiotics, both during planktonic growth (MIC) and when grown as biofilms (MBEC). Antibiotics were chosen based on their different modes of action and include some that are commonly used to treat K. pneumoniae infections. Overall, resistance was greater in biofilms (MBECs) than for planktonic cells (MICs) for all antibiotics tested (Table 2). When compared against the WT strain, the yfiR mutant showed decreased resistance in biofilms for most antibiotics tested, Amk, Cip, a second generation fluoroquinolone that affects DNA gyrase, the broad spectrum β-lactam antibiotics Mpm and Imp that inhibit cell wall synthesis, and TS, which inhibits folate biosynthesis. TS did not affect either the MIC or the MBEC of any of the remaining strains when compared to WT LM21gfp (not shown). The yfiN strain was more sensitive to Cip when grown planktonically (MIC) but more resistant in biofilms. Cells bearing yifB behaved differently and were more resistant to Cip during planktonic growth yet more susceptible to both Imp and Mpm, the latter when in biofilms. Overall, susceptibility with respect to the WT strain was altered most with Cip, in both planktonic cells and biofilms. Together these results indicate that mutations in the yfiRNB operon alter susceptibility to antibiotics in a way that is not consistent with the expected effect based on biofilm formation. The yfiN mutation increased resistance in biofilms at least in one case and contrasts with the increased sensitivity of yfiR in all cases assayed, despite the fact that mutation of this gene results in more robust biofilms. The effect of the observed changes in extracellular matrix triggered by overproduction of YfiN DGC and cellulose, were next evaluated using a C. elegans infection model (Bialek et al., 2010; Fuursted et al., 2012; Srinivasan et al., 2012). Age-synchronized, adult nematodes were fed LM21gfp WT cells or isogenic mutants lacking yfiR, yfiN or yfiB and mortality was scored over time, as an indirect marker of virulence potential. All strains, including the WT strain, increased mortality (reduced survival) when compared to E. coli OP50 (p < 0.001). However host survival curves were indistinguishable from one another, and even worms fed K. pneumoniae LM21 ∆yfiR grew to maturity with no apparent adverse effects on viability compared to those fed the other LM21 strains (Fig. 6a). Likewise, no differences were observed in terms bacterial colonization (Fig. 6b). This result was contrary to our expectations and indicated that an increase in cellulose production was not involved in virulence of K. pneumoniae in this assay. This

12

observation is consistent with other reports showing that the production of cellulose is not necessarily involved in virulence, such as in S. enteritidis (Solano et al., 2002), and that virulence in many pathogens most likely requires fine regulation of c-di-GMP levels (Romling et al., 2013). Discussion Our results show that K. pneumoniae strains with mutations that affect expression of the DGC-encoding gene yfiN have altered cell wall components that affect biofilm formation and cell survival. Previous reports had shown that the K. pneumoniae yfiRNB operon is involved in biofilm formation via production of the intracellular second messenger c-diGMP and its regulation of type 3 fimbriae production (Wilksch et al., 2011). In the present study, specific deletion of the negative regulator yfiR increased expression of yfiN and resulted in altered cell matrix components. Colony phenotypes when plated on solid indicator media, as well as anthrone determination of the polysaccharides present in biofilms and the reduction in surface attached cells upon treatment with cellulase, suggested that the yfiR mutant cells contained increased cellulose. The presence of cellulose as part of the K. pneumoniae EPS contrasts with reports analyzing K. pneumoniae cell phenotypes when grown on solid indicator media (Zogaj et al., 2003). However, recent results have shown that high weight exopolysaccharides from K. pneumoniae contain various glycosyl residues and linkages, among which were both glucose and 4-linked glucopyranosyl residues (Bales et al., 2013). The increase in biofilm formation and cellulose production of the yfiR mutant strain correlates with an increased transcription of yfiN, which has been previously shown to code for a protein with proposed DGC activity (Wilksch et al., 2011). However, transcription of the two putative cellulose synthase genes identified here in the K. pneumoniae genome (bcsA1 and bcsA2) did not change in a yfiR mutant background. This result indicates that the effect of c-di-GMP on cellulose production most probably occurs via the PilZ domains in both predicted Bcs proteins, as has been shown in Gluconoacetobacter xylinus and several Enterobacteriaceae (Amikam & Galperin, 2006; Bhowmick et al., 2011; Raterman et al., 2013; Ryjenkov et al., 2006). Regulation of cellulose production by c-di-GMP has been extensively studied at the molecular level, and recent mechanistic insight has been given regarding of enzyme activity based on analysis of the crystal structure of the Rhodobacter sphaeroides c-di-GMP–activated

13

BcsA–BcsB complex (Morgan et al., 2014). The robust biofilms formed by K. pneumoniae yfiR cells are most probably due to enhanced polysaccharide production, triggered by local levels c-di-GMP that can activate BcsA cellulose synthases. This is similar to what has been reported for both E. coli and P. aeruginosa, where homologs of the yfiRNB have been shown to be involved in EPS production and biofilm formation (Malone et al., 2010; Raterman et al., 2013). Microbial cellulose is synthesized during biofilm formation in species such as S. typhimurium, S. enteritidis and E. coli and can have diverse functions. In Rhizobium and Agrobacterium species it facilitates cell adhesion and promotes interaction with plant cells, it can confer protection to organisms such as A. xylinum and Sarcina ventriculi in their natural environments, and is proposed to contribute to survival and persistence of Erwinia amylovora under stress (Ordax et al., 2010; Ross et al., 1991). Not much is known regarding cellulose production and its role in K. pneumoniae, yet the presence of the two putative synthase operons reported here for the first time indicates that this polymer could be important in certain environmental settings where its production could be triggered by niche-specific conditions not mimicked in the laboratory conditions used here. Both the extracellular matrix and biofilm formation can promote resistance to antibiotic treatment and the adverse conditions within the host and have been implicated in pathogenesis and long-term persistence in different microorganisms. Microbes defective in polysaccharide production can have increased sensitivity to antimicrobials and to the host immune system (Rendueles et al., 2013). In addition to their protective role, polysaccharides are involved in mediating cell-surface and cell-to-cell interactions and can even act as antibiofilm agents by inhibiting growth of other species (Kostakioti et al., 2013; Rendueles et al., 2013). Under our experimental conditions, enhanced K. pneumoniae biofilm formation in vitro was associated with an increase in cellulose production, but not necessarily with increased cell survival or virulence. The capacity to kill and colonize C. elegans was similar in all strains analyzed and indicated that cellulose is not involved in virulence in this model system, an observation consistent with reports for Salmonella (Solano et al., 2002), but different from the EPS-dependent resistance to ingestion observed for P. aeruginosa (Malone et al., 2010). Furthermore, robust biofilm formation and cellulose production did not necessarily lead to enhanced survival upon exposure to antibiotics. This was unexpected, as enhanced matrix

14

production has been equated with increased biofilm survival (Van Acker et al., 2014). However, it is consistent with what has been found in P. aeruginosa where increased production of Pel EPS did not increase resistance to aminoglycosides (Khan et al., 2010), while the loss of EPS did, similar to our results with the K. pneumoniae yfiN mutant biofilms, which had increased resistance to CIP. c-di-GMP is known to alter gene expression (Romling et al., 2013), and thus it is possible that the loss of the DCG YfiN could change expression of membrane proteins, such as efflux pumps or porins, that can influence drug susceptibility. It has been recently reported, for example, that a c-diGMP-bound transcriptional factor plays an important role in the synthesis of mycolic acids in Mycobacterium smegmatis, and alters permeability and drug susceptibility (Li & He, 2012).

Thus imbalances in cellular pools of c-di-GMP can affect membrane

components and transport across the membrane in our K. pneumoniae yfiRNB mutant strains. In addition, changes in expression of genes involved in protection against oxidative stress, a mechanism proposed to be important in protection against treatment with antibiotics (Van Acker et al., 2014), could also render cells more susceptible to drug treatment, particularly in biofilms. Further analysis of possible changes in membrane protein profiles, gene expression or production of reactive oxygen species could further help to understand the observed differences in these strains. In consequence, and as has been observed for other organisms (Musafer et al., 2014), in vitro analysis of K. pneumoniae biofilm formation capacity does not necessarily correlate with drug resistance, an aspect that merits more detailed studies using clinical isolates given the importance of this pathogen in HAI infections. Acknowledgements We are grateful to Dr. Damien Balestrino and Dr. Christiane Forestier at the University of Auvergne – Clermont Ferrand, France for help with mutant strain construction and for providing bacterial strain K. pneumoniae LM21gfp. We thank Dr. Federico Brown for expert assistance with the C. elegans assays, and Dr. Juan M. Anzola for help with bioinformatics analysis. This work was funded by Colciencias (Project No. 545-2011).

15

References Amikam, D. & Galperin, M. Y. (2006). PilZ domain is part of the bacterial c-di-GMP binding protein. Bioinformatics 22, 3-6. Bales, P. M., Renke, E. M., May, S. L., Shen, Y. & Nelson, D. C. (2013). Purification and Characterization of Biofilm-Associated EPS Exopolysaccharides from ESKAPE Organisms and Other Pathogens. PloS one 8, e67950. Balestrino, D., Haagensen, J. A., Rich, C. & Forestier, C. (2005). Characterization of type 2 quorum sensing in Klebsiella pneumoniae and relationship with biofilm formation. J Bacteriol 187, 2870-2880. Balestrino, D., Ghigo, J. M., Charbonnel, N., Haagensen, J. A. & Forestier, C. (2008). The characterization of functions involved in the establishment and maturation of Klebsiella pneumoniae in vitro biofilm reveals dual roles for surface exopolysaccharides. Environ Microbiol 10, 685-701. Barnhart, D. M., Su, S., Baccaro, B. E., Banta, L. M. & Farrand, S. K. (2013). CelR, an ortholog of the diguanylate cyclase PleD of Caulobacter, regulates cellulose synthesis in Agrobacterium tumefaciens. Appl Environ Microbiol 79, 7188-7202. Bhowmick, P. P., Devegowda, D., Ruwandeepika, H. A., Fuchs, T. M., Srikumar, S., Karunasagar, I. & Karunasagar, I. (2011). gcpA (stm1987) is critical for cellulose production and biofilm formation on polystyrene surface by Salmonella enterica serovar Weltevreden in both high and low nutrient medium. Microbial Pathogenesis 50, 114-122. Bialek, S., Lavigne, J. P., Chevalier, J., Marcon, E., Leflon-Guibout, V., Davin, A., Moreau, R., Pages, J. M. & Nicolas-Chanoine, M. H. (2010). Membrane efflux and influx modulate both multidrug resistance and virulence of Klebsiella pneumoniae in a Caenorhabditis elegans model. Antimic Agents Chemother 54, 4373-4378. Boddicker, J. D., Anderson, R. A., Jagnow, J. & Clegg, S. (2006). Signature-tagged mutagenesis of Klebsiella pneumoniae to identify genes that influence biofilm formation on extracellular matrix material. Infect Immun 74, 4590-4597. Boyd, C. D. & O'Toole, G. A. (2012). Second messenger regulation of biofilm formation: breakthroughs in understanding c-di-GMP effector systems. Ann Rev Cell Dev Biol 28, 439-462. Bustin, S. A. (2010). Why the need for qPCR publication guidelines?--The case for MIQE. Methods 50, 217-226. Cherepanov, P. P. & Wackernagel, W. (1995). Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibioticresistance determinant. Gene 158, 9-14. Cruz, D. P., Huertas, M. G., Lozano, M., Zarate, L. & Zambrano, M. M. (2012). Comparative analysis of diguanylate cyclase and phosphodiesterase genes in Klebsiella pneumoniae. BMC Microbiol 12, 139. Daims, H., Lucker, S. & Wagner, M. (2006). daime, a novel image analysis program for microbial ecology and biofilm research. Environmental microbiology 8, 200-213. Datsenko, K. A. & Wanner, B. L. (2000). One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 97, 66406645. Di Martino, P., Cafferini, N., Joly, B. & Darfeuille-Michaud, A. (2003). Klebsiella pneumoniae type 3 pili facilitate adherence and biofilm formation on abiotic surfaces. Res Microbiol 154, 9-16.

16

Fuursted, K., Scholer, L., Hansen, F., Dam, K., Bojer, M. S., Hammerum, A. M., Dagnaes-Hansen, F., Olsen, A., Jasemian, Y. & other authors (2012). Virulence of a Klebsiella pneumoniae strain carrying the New Delhi metallo-betalactamase-1 (NDM-1). Microbes Infect 14, 155-158. Guzman, L. M., Belin, D., Carson, M. J. & Beckwith, J. (1995). Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol 177, 4121-4130. Harrison, J. J., Stremick, C. A., Turner, R. J., Allan, N. D., Olson, M. E. & Ceri, H. (2010). Microtiter susceptibility testing of microbes growing on peg lids: a miniaturized biofilm model for high-throughput screening. Nat Protoc 5, 12361254. Heydorn, A., Ersboll, B. K., Hentzer, M., Parsek, M. R., Givskov, M. & Molin, S. (2000). Experimental reproducibility in flow-chamber biofilms. Microbiol 146 ( Pt 10), 2409-2415. Hoiby, N., Ciofu, O., Johansen, H. K., Song, Z. J., Moser, C., Jensen, P. O., Molin, S., Givskov, M., Tolker-Nielsen, T. & other authors (2011). The clinical impact of bacterial biofilms. Int J Oral Sci 3, 55-65. Institute, C. a. L. S. (2013). Performance Standards for Antimicrobial Susceptibility Testing; Twenty-Third Informational Supplement. In CLSI document M100-S23. Wayne, Pennsylvania: Clinical and Laboratory Standards Institute. Johnson, J. G. & Clegg, S. (2010). Role of MrkJ, a phosphodiesterase, in type 3 fimbrial expression and biofilm formation in Klebsiella pneumoniae. J Bacteriol 192, 3944-3950. Johnson, J. G., Murphy, C. N., Sippy, J., Johnson, T. J. & Clegg, S. (2011). Type 3 fimbriae and biofilm formation are regulated by the transcriptional regulators MrkHI in Klebsiella pneumoniae. J Bacteriol 193, 3453-3460. Khan, W., Bernier, S. P., Kuchma, S. L., Hammond, J. H., Hasan, F. & O'Toole, G. A. (2010). Aminoglycoside resistance of Pseudomonas aeruginosa biofilms modulated by extracellular polysaccharide. Int Microbiol 13, 207-212. Kostakioti, M., Hadjifrangiskou, M. & Hultgren, S. J. (2013). Bacterial biofilms: development, dispersal, and therapeutic strategies in the dawn of the postantibiotic era. Cold Spring Harb Perspect Med 3, a010306. Langstraat, J., Bohse, M. & Clegg, S. (2001). Type 3 fimbrial shaft (MrkA) of Klebsiella pneumoniae, but not the fimbrial adhesin (MrkD), facilitates biofilm formation. Infect Immun 69, 5805-5812. Laurentin, A. & Edwards, C. A. (2003). A microtiter modification of the anthronesulfuric acid colorimetric assay for glucose-based carbohydrates. Anal Biochem 315, 143-145. Li, W. & He, Z. G. (2012). LtmA, a novel cyclic di-GMP-responsive activator, broadly regulates the expression of lipid transport and metabolism genes in Mycobacterium smegmatis. Nucleic Acids Res 40, 11292-11307. Malone, J. G., Jaeger, T., Spangler, C., Ritz, D., Spang, A., Arrieumerlou, C., Kaever, V., Landmann, R. & Jenal, U. (2010). YfiBNR mediates cyclic di-GMP dependent small colony variant formation and persistence in Pseudomonas aeruginosa. PLoS pathog 6, e1000804. Malone, J. G., Jaeger, T., Manfredi, P., Dotsch, A., Blanka, A., Bos, R., Cornelis, G. R., Haussler, S. & Jenal, U. (2012). The YfiBNR signal transduction mechanism reveals novel targets for the evolution of persistent Pseudomonas aeruginosa in cystic fibrosis airways. PLoS pathog 8, e1002760. Morgan, J. L., McNamara, J. T. & Zimmer, J. (2014). Mechanism of activation of bacterial cellulose synthase by cyclic di-GMP. Nat Struct Mol Biol 21, 489-496.

17

Musafer, H. K., Kuchma, S. L., Naimie, A. A., Schwartzman, J. D., Al-Mathkhury, H. J. & O'Toole, G. A. (2014). Investigating the Link Between Imipenem Resistance and Biofilm Formation by Pseudomonas aeruginosa. Microb Ecol 68, 111-120. O'Toole, G. A. & Kolter, R. (1998). Initiation of biofilm formation in Pseudomonas fluorescens WCS365 proceeds via multiple, convergent signalling pathways: a genetic analysis. Mol Microbiol 28, 449-461. Ordax, M., Marco-Noales, E., Lopez, M. M. & Biosca, E. G. (2010). Exopolysaccharides favor the survival of Erwinia amylovora under copper stress through different strategies. Res Microbiol 161, 549-555. Pan, Y. J., Lin, T. L., Chen, Y. H., Hsu, C. R., Hsieh, P. F., Wu, M. C. & Wang, J. T. (2013). Capsular types of Klebsiella pneumoniae revisited by wzc sequencing. PloS one 8, e80670. Pfaffl, M. W. (2001). A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29, e45. Podschun, R. & Ullmann, U. (1998). Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 11, 589-603. Raterman, E. L., Shapiro, D. D., Stevens, D. J., Schwartz, K. J. & Welch, R. A. (2013). Genetic analysis of the role of yfiR in the ability of Escherichia coli CFT073 to control cellular cyclic dimeric GMP levels and to persist in the urinary tract. Infect Immun 81, 3089-3098. Raunkjær, K., Hvitved-Jacobsen, T. & Nielsen, P. H. (1994). Measurement of pools of protein, carbohydrate and lipid in domestic wastewater. Water Res 28, 251-262. Rendueles, O., Kaplan, J. B. & Ghigo, J. M. (2013). Antibiofilm polysaccharides. Environ Microbiol 15, 334-346. Romling, U., Galperin, M. Y. & Gomelsky, M. (2013). Cyclic di-GMP: the first 25 years of a universal bacterial second messenger. Microbiol Mol Biol Rev 77, 1-52. Ross, P., Mayer, R. & Benziman, M. (1991). Cellulose biosynthesis and function in bacteria. Microbiol Rev 55, 35-58. Ryjenkov, D. A., Simm, R., Romling, U. & Gomelsky, M. (2006). The PilZ domain is a receptor for the second messenger c-di-GMP: the PilZ domain protein YcgR controls motility in enterobacteria. J Biol Chem 281, 30310-30314. Singla, S., Harjai, K. & Chhibber, S. (2013). Susceptibility of different phases of biofilm of Klebsiella pneumoniae to three different antibiotics. J Antibiot 66, 61-66. Solano, C., Garcia, B., Valle, J., Berasain, C., Ghigo, J. M., Gamazo, C. & Lasa, I. (2002). Genetic analysis of Salmonella enteritidis biofilm formation: critical role of cellulose. Mol Microbiol 43, 793-808. Srinivasan, V. B., Vaidyanathan, V., Mondal, A., Venkataramaiah, M. & Rajamohan, G. (2012). Functional characterization of a novel Mn2+ dependent protein serine/threonine kinase KpnK, produced by Klebsiella pneumoniae strain MGH78578. FEBS letters 586, 3778-3786. Stiernagle, T. (2006). Maintenance of C. elegans. In The C elegans Research Community, WormBook. http://www.wormbook.org.: WormBook, ed. Suescun, A. V., Cubillos, J. R. & Zambrano, M. M. (2006). [Genes involved in fimbrial biogenesis affect biofilm formation in Klebsiella pneumoniae]. Biomedica 26, 528537. Van Acker, H., Van Dijck, P. & Coenye, T. (2014). Molecular mechanisms of antimicrobial tolerance and resistance in bacterial and fungal biofilms. Trends Microbiol 22, 326-333. Wilksch, J. J., Yang, J., Clements, A., Gabbe, J. L., Short, K. R., Cao, H., Cavaliere, R., James, C. E., Whitchurch, C. B. & other authors (2011). MrkH, a novel c18

di-GMP-dependent transcriptional activator, controls Klebsiella pneumoniae biofilm formation by regulating type 3 fimbriae expression. PLoS pathog 7, e1002204. Zogaj, X., Bokranz, W., Nimtz, M. & Romling, U. (2003). Production of cellulose and curli fimbriae by members of the family Enterobacteriaceae isolated from the human gastrointestinal tract. Infect Immun 71, 4151-4158.

19

Tables Table 1. Strains and plasmids used in this study Strains/Plasmid

Description / genotype

Reference

Strains K. pneumoniae Ap

R

CG 217 - M39

Ap

R

LM21 gfp

Clinical isolate

CG 217

Suescun et al., 2006 R

Km

Suescun et al., 2006 Balestrino et al., 2008

yfiR deletion, Sp

R

This work

LM21 △yfiN

yfiN deletion, Sp

R

This work

LM21 △yfiB

yfiB deletion, Sp

R

This work

LM21 △yfiR

LM21 △ycdT LM21/pKOBEG199 E. coli OP50

ycdT deletion, Sp Tc

R

This work

R

Balestrino et al., 2005 R

Ura-, Sm

Stiernagle, 2006

Plasmids pBAD18-cm pBAD18-yfiR pBAD18-yfiN

Arabinose inducible expression, CmR

Guzman L, et al 1995

R

This work

R

This work

R

WT yfiR gene / Cm WT yfiN gene / Cm

pBAD18-yfiB

WT yfiB gene / Cm

This work

pKOBEG119

λ Red recombinase expression vector

Datsenko and Wanner, 2000

pKD4

Template for Km cassette

Datsenko and Wanner, 2000

pCP20

FRT sites using the FLP helper plasmid Datsenko and Wanner, 2001

20

Table 2. Susceptibility to antibiotics of planktonic cultures and biofilms Strain

CIP MIC

Change

WT

0.063

∆yfiR

0.063

∆yfiN ∆yfiB

-

MBEC

AMK Change

32

0.0

MIC*

MBEC

2

256

2

64

1.0

1

0.031

0.5

64

2.0

2

0.125

2.0

2

0.1

2

IMP Change -

MIC

Change

1

-

MPM

MBEC✝

MIC*

MBEC✝

>256

2

>256

0.3

1

1.0

>256

2

128

256

1.0

1

1.0

>256

2

>256

256

1.0

0.5

0.5

>256

2

128

The MIC (minimal inhibitory concentration) and the MBEC (minimal biofilm eradication concentration) are shown for each strain incubated with ciprofloxacin (CIP), amikacin (AMK), imipenem (IMP) and meropenem (MPM). Change indicates the change with respect to the WT strain under the same treatment and is shown in bold. * indicates that there was no change in these MICs with respect to the WT; ✝ indicates that the foldchange with respect to the WT was not calculated since the precise MBEC value was not obtained.

21

Figure legends Figure 1. Schematic representation of the yfiRNB and cellulose operons The K. pneumoniae yfiRNB (a) and cellulose (b) operons are shown. Arrows indicate direction of transcription; small bent arrow indicates a putative promoter, the triangle shows the transposon insertion in mutant M39 (yfiR gene), numbers below line indicate gene coordinates; the map was obtained using SnapGene (http://www.snapgene.com/). Figure 2. Bioflm formation of K. pneumoniae strains Biofilms were quantified using crystal violet staining after 18 hours of growth of wild type (WT) strains CG217 and LM21gfp (black bars), mutant strains (grey bars) and deletion strains complemented with the indicated genes (hatched bars). Mutant strains include the transposon insertion mutant in CG217 (M39) or mutants generated by site-specific deletion in strain LM21gfp. Error bars indicate the standard deviation.

Figure 3. Biofilm structure of K. pneumoniae LM21gfp and isogenic mutants. Biofilms grown for 18h were monitored by confocal scanning laser microscopy. Images indicate a representative 3D structure of each strain (a) WT LM21gfp, (b) LM21 ∆yfiR, (c) LM21 ∆yfiN, (d) ∆yfiB. Image visualization was obtained using the software DAIME (Daims et al., 2006). The X, Y and Z axes are indicated, numbers are given in um. Figure 4. Cellulose is present in the EPS. Hexose sugars were quantified from cells grown as colony biofilms (a) and 24 hour biofilms (b) formed on lid pegs. Biofilms were treated with cellulose (hatched bars) or left untreated (black bars), and remaining attached cells were quantified using crystal violet staining. Error bars represent the standard deviation. Statistical significance was determined by ANOVA, compared to the WT strain. *P