OUP user menu

Effect of growth conditions on poly-N-acetylglucosamine expression and biofilm formation in Escherichia coli

Nuno Cerca, Kimberly K. Jefferson
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01142.x 36-41 First published online: 1 June 2008


Escherichia coli contains a four-gene operon, pgaABCD, which encodes the proteins necessary for the synthesis of polymeric N-acetylglucosamine, or PGA. Poly-N-acetyl-glucosamine was first described in Staphylococcus aureus and Staphylococcus epidermidis and was found to have important roles in biofilm formation and immune evasion. PGA also plays a role in biofilm formation in E. coli, but its role in immune evasion has not been thoroughly studied. We previously reported that E. coli PGA cross-reacts with an opsonic-antibody raised against S. aureus PNAG and this is the basis for an ongoing investigation regarding the development of a vaccine against both pathogens. In this paper we investigated pga expression in wild type and csrA or nhaR deletion mutant strains during different growth phases and temperatures, and in response to chemical stimuli using a pga promoter-reporter fusion construct, real-time reverse transcriptase-PCR, immunoblotting, and biofilm assays. Expression of pga and polysaccharide synthesis were induced by glucose, NaCl, and ethanol, but only glucose augmented biofilm formation. The regulatory factor NhaR was required for NaCl-induced pga expression, whereas the effects of glucose and ethanol were independent of CsrA and NhaR.

  • biofilm
  • poly-N-acetylglucosamine
  • intercellular adhesion


Biofilm formation is recognized as an important virulence factor in many bacterial species. The ubiquitous exo-polysaccharide, β-1,6-poly-N-acetylglucosamine (referred to as PNAG in Staphylococcus aureus and PGA in Escherichia coli) appears to play an important role in biofilm formation, immune evasion, and pathogenesis in a variety of bacterial species including: S. aureus (Cramton et al., 1999; Kropec et al., 2005; Cerca et al., 2007a), Staphylococcus epidermidis (Mack et al., 2000; Vuong et al., 2004; Cerca et al., 2006), Actinobacillus species (Kaplan et al., 2004; Izano et al., 2007) and E. coli (Wang et al., 2004; Agladze et al., 2005). PNAG was first described in S. epidermidis in which it is encoded by the intercellular adhesin (ica) locus (Heilmann et al., 1996). In E. coli, PGA is synthesized by the four proteins encoded within a homologous locus, pgaABCD (Wang et al., 2004). We recently have shown that E. coli PGA is immunologically similar to S. aureus PNAG, and this is the basis for an ongoing investigation regarding the development of a vaccine against both pathogens (Cerca et al., 2007b). It has been shown that posttranscriptional control of pgaABCD expression by the RNA-binding protein carbon storage regulator A (CsrA) regulation leads to the inhibition of biofilm formation (Wang et al., 2005). Recently it was also shown that high salt concentrations and alkaline pH increase biofilm formation and this effect was dependent on the cation-responsive regulatory protein NhaR (Goller et al., 2006). However, our understanding of the regulation of this important virulence factor remains somewhat limited and we sought to further characterize this process. In this study we investigated the expression of PGA at different stages of growth, and in response to a variety of chemical stimuli at the levels of transcription, translation, polysaccharide production, and biofilm formation.

Materials and methods

Strains and culture conditions

The strains were grown in tryptic soy broth (TSB) (17 g casein peptone L−1, 3 g soymeal peptone L−1, 5 g sodium chloride L−1, 2.5 g dipotassium hydrogen phosphate L−1, 2.5 g glucose L−1) or in modified TSB lacking glucose (TSB−), at 37 °C with shaking at 200 r.p.m., overnight, except where otherwise noted. Escherichia coli CFT073 is a clinical strain originally isolated from a case of pyelonephritis (Mobley et al., 1990) TOP10 is a laboratory strain used for cloning (Invitrogen, Carlsbad, CA), and E. coli UTI-J and E. coli UTI-U are two uropathogenic strains isolated in Boston, MA (Cerca et al., 2007b). UTI-JΔpga is an isogenic mutant in which the pga locus was replaced with a chloramphenicol acetyltransferase cassette (Cerca et al., 2007b).

β-Gal reporter assays

The entire noncoding region between the pga locus and the ycdT gene was amplified from total DNA isolated from E. coli CFT073 by PCR using primers pgaproFWD (TTTTGCTGCAAGGCAGGCTTG) and pgaproREV (TGAATTCCCTGTATTACTCCATGTATTGCC) and cloned into the pBLUE-TOPO vector (Invitrogen) to yield the construct pPGA. The pBLUE vector was used as a negative control. Both constructs were sequenced at the Microbiology Core Facility (Harvard Medical School, Boston, MA) to confirm the absence of mutations. Escherichia coli TOP10 cells served as the host strain. The strains were grown in TSB−. A growth curve was performed at 37 °C. We verified that strains containing both plasmids had similar growth kinetics (data not shown). pga promoter activity was analyzed at 21 and 37 °C, at different growth stages (early, mid and late exponential and stationary phase) and in the presence of 1% glucose, 1% ethanol, 1% NaCl, 1 mM MnCl2 and 5% fetal bovine serum (FBS). Approximately 1 × 109 bacteria were collected by centrifugation, resuspended in lysis buffer (β-gal kit, Invitrogen), lysed by bead-beating, and the supernatant was cleared by centrifugation. Total protein concentrations in the supernatants were analyzed by the Bradford assay, and β-galactosidase was quantified by standard o-nitrophenyl-β-d-galactopyranoside (ONPG) assay, following the manufacturer instructions (Invitrogen). To calculate relative β-gal activity under different conditions, we subtracted the basal activity per mg of extracted protein of strain pBLUE from the activity per mg protein of strain pPGA. For growth stage and temperature assays maximum activity was defined as 100% activity. For assays in which the effect of compounds was tested, activity in TSB alone was defined as 100%. For each condition studied, three separate experiments with duplicates samples were performed.

Isolation of total RNA

Total cellular RNA was prepared using the FastRNA Pro Blue Kit (MP Biomedicals, Solon, OH) accordingly to manufacturer instructions. Contaminating DNA was removed by two sequential treatments with Turbo DNAse (Ambion, Austin, TX) of 20 min at 37 °C. The enzyme was removed using DNAse inactivation reagent (Ambion). RNA was quantified by A260 nm and A280 nm, and stored at −80 °C.

Construction of E. coli nhaR and csrA-deleted strains

The nhaR and csrA genes were replaced with a chloramphenicol acetyltransferase (CAT) cassette by the PCR-mediated one-step method of gene inactivation as described by Datsenko & Wanner (2000). To accomplish this, the clinical isolate UTI-J was first transformed with the Red recombinase expression plasmid pKD46. The recombinase gene was induced with 10 mM l-arabinose and the bacteria were electroporated with a PCR product containing CAT cassette flanked by 50-nt of homology to the csrA or nhaR genes. For PCR the template pKD3 was amplified with the primer pair CsrAdelFWD (CACCGATAAAGATGAGACGCGGAAAGATTAGTAACTGGACTGCTGGGATTGTGTAGGCTGGAGCTGCTTC) and CsrAdelREV (CAGAGAGACCCGACTCTTTTAATCTTTCAAGGAGCAAAGAATGCTGATTCACATATGAATATCCTCCTTAG) or NhaRdelFWD (CCTTTTCATTGTTATCAGGGAGAGAAATGAGCATGTCTCATATCAATTACGTGTAGGCTGGAGCTGCTTC) and NhaRdelREV (CAAATGTTTATTTTGAAGCTGGAGTAAACAGCGCAGAATAGTCTGTATTGCACATATGAATATCCTCCTTAG). Transformants were selected on Luria–Bertani (LB) agar containing 30 μg chloramphenicol mL−1 and analyzed by PCR to confirm the mutation.

Quantitative real-time reverse transcriptase (RT)-PCR analysis of pgaA mRNA

The primers used to amplify pgaA and 16SrRNA were: pgaAFW (AGGGACTGCGCATTGATTAC), pgaAREV (GTTCAGGTTCGACAACATCG), 16SFW (GATAACTACTGGAAACGGTAG) and 16SREV (ACCTACTAGCTAATCCCATCTG) respectively. RNA samples were reverse transcribed in the presence of pgaAREV, 16SREV and Superscript II Reverse Transcriptase (Invitrogen). Control reactions lacked reverse transcriptase enzyme. For amplifying pgaA, 1 : 50 dilutions of cDNA and no-RT controls were used and for 16S rRNA gene 1 : 5000 dilutions were used. Real-time RT-PCR reactions contained 1 μL diluted cDNA or no-RT control, 10 pmol of each primer, 9.5 μL nuclease free deionized H2O, and 12.5 μL Sensimix Plus SYBR Green mix (Bioline, Randolph, MA). Real-time RT-PCR was performed under the following conditions: 95 °C for 3 min, 40 cycles of 95 °C for 10 s, 55 °C for 15 s, 70 °C for 15 s. To monitor the specificity, final RT-PCR products were analyzed by melting curves and electrophoresis. The amount of pgaA transcript was expressed as the n-fold difference relative to the control gene (2ΔCT, where ΔCT represents the difference in threshold cycle between the target and control gene). For each condition studied, three separate experiments with duplicates were performed.

Detection of PGA by immunoblotting

Bacteria were grown overnight in 10 mL of TSB with or without 1% glucose, 1% ethanol, 1% NaCl, 1 mM MnCl2 or 5% FBS. The cultures were diluted in TSB to produce an OD600 nm=1.5. Bacteria were collected from 1 mL of each suspension by centrifugation, resuspended in 300 μL of 0.5 M EDTA (pH 8.0), and incubated for 5 min at 100 °C. Cells were harvested by centrifugation at 10 500 g, 6 min and 100 μL of the supernatant was incubated with 10 μL of proteinase K (20 mg mL−1; Qiagen) for 60 min at 60 °C. Proteinase K was heat inactivated for 30 min at 80 °C. This solution was then diluted threefold in of Tris-buffered saline [20 mM Tris-HCl, 150 mM NaCl (pH 7.4)] and 200 μL of each dilution were immobilized on a nitrocellulose filter that was then blocked with 1% bovine serum albumin, and incubated for 2 h with a affinity purified rabbit antibody raised to S. aureus PNAG (Maira-Litran et al., 2005) previously shown to cross-react with E. coli PGA (Cerca et al., 2007b). Horseradish peroxidase-conjugated secondary antibody against rabbit IgG (SouthernBiotech) was diluted 1 : 6000 and PGA was detected with the ECL Plus (enhanced chemiluminescence) Western blotting system (GE Healthcare, Piscataway, NJ).

Biofilm assay

Biofilm formation was assayed semi-quantitatively as described previously (Cerca et al., 2006) with some modifications. Briefly, bacteria grown in 96-well polystyrene plates for 24 H at 37 °C (Sarsted, Germany) were washed twice with 0.9% NaCl solution, dried in an inverted position, and stained with 0.4% safranin for 10 min. The plates were washed with distilled water, and dried overnight. For each condition studied, three separate experiments with duplicates were performed.

Results and discussion

Analysis of pga promoter activity using a β-galactosidase reporter

To determine pga promoter activity during different growth phases, and in response to chemical stimuli, we fused the promoter of the pgaABCD locus to a β-gal gene. β-Gal expression, as measured by ONPG hydrolysis per mg total protein was significantly higher at 37 °C than at 21 °C (P<0.05, paired samples t-test) (Fig. 1a) suggesting that in the human host the pga locus could be induced. Expression of the pga promoter was also significantly greater (P<0.05, paired samples t-test) during stationary phase (Fig. 1b). We subsequently grew all cultures to stationary phase at 37 °C.

Figure 1

Expression of the pga promoter in modified TSB medium without glucose; (a) growth at 21 or 37°C, at the beginning of the stationary phase; (b) growth at 37°C, at different points of the growth phase: early log phase (4 h), mid log phase (6 h), late log phase (8 h) and late stationary phase (14 h). Maximum activity was defined as 100% activity. Error bars represent the SD. *Represents a statistical difference (P<0.05, paired samples t-test).

High osmolarity induces a stress response in bacteria, and in S. aureus PNAG is synthesized as part of the response (Lim et al., 2004). Similarly, we found, in agreement with a recent study demonstrating a role in pga transcription for the sodium stress-response protein, NhaR (Goller et al., 2006) that addition of 1% NaCl to the growth medium resulted in a significant increase (P<0.05, paired samples t-test) in β-gal activity. Addition of another stress-inducing factor, 1% ethanol, also induced pga promoter driven expression of β-gal (P<0.05, paired samples t-test). In S. epidermidis, NaCl and ethanol were found to stimulate biofilm formation by two independent pathways (Knobloch et al., 2001) so it is likely that ethanol-induced pga activity is independent of NhaA. Divalent cations appear to play a role in surface association of exopolysaccharides but there are no reports of their influence on pga promoter activity. We did not see a significant increase in β-gal activity in the presence of 1 mM MnCl2. We hypothesized that because S. aureus expresses PNAG in vivo, that serum would induce PGA expression in E. coli. However, we found that 5% FBS did not exert a significant effect (P>0.05, paired samples t-test) (Fig. 2). We were interested in the effect of glucose on pga expression but glucose has a strong inhibitory effect on β-gal expression and we have previously observed glucose-mediated inhibition of other β-gal reporters (unpublished observation) so the effect of glucose was not determined using our reporter assay.

Figure 2

Expression of the pga promoter at 37°C and stationary phase, due to the effect of 1% glucose, 1% ethanol, 1% NaCl 1 mM MnCl2 and 5% FBS. The activity obtained with the modified TSB medium (TSB−) without added factors was defined as 100% activity. Error bar represent the SD. *Represents a statistical difference to the control (P<0.05, paired samples t-test).

Analysis of pga expression in clinical isolates by real-time RT-PCR

To verify the effects of different compounds on pga promoter expression in clinical strains, and to assess the relationship with the known pga repressor CsrA and inducer NhaR we analyzed UTI-J, and its isogenic mutants UTI-JΔpga (as a negative control), UTI-JΔcsrA, UTI-JΔnhaR (Cerca et al., 2007b). We analyzed pgaA mRNA levels in bacteria grown in the presence of 1% glucose, 1% ethanol, 1% NaCl, 1 mM MnCl2 or 5% FBS. pgaA transcript levels in response to the various stimuli corresponded with promoter activity as determined with the reporter system: addition of ethanol and NaCl to the growth medium resulted in a significant increase (P<0.05, paired samples t-test) in mRNA levels, whereas 1 mM MnCl2 had a less pronounced effect and 5% FBS (P<0.1, paired samples t-test) did not have a significant effect (Table 1). Goller et al. (2006) found in a previous study, that sucrose inhibits pga expression. However, we found that 1% glucose also induced pgaA transcript levels (P<0.05, paired samples t-test). The csrA deletion mutant exhibited an impressive 3000-fold increase in pgaA mRNA expression with respect to wild type. Strong repression of pga by CsrA has been reported previously and clearly, in our clinical isolate, CsrA had a strong inhibitory effect (Wang et al., 2004). Interestingly, glucose, NaCl and ethanol further augmented expression of pgaA in JΔcsrA (Table 1) suggesting that the effect of these factors is CsrA-independent. Deletion of nhaR did not affect pgaA expression in TSB- or in the presence of glucose of ethanol but ablated NaCl-induced pgaA expression indicating that induction of pgaA expression by NaCl is dependent upon NhaR. We investigated the possible role of other regulators in pga expression, and analysis of the pga promoter region using the BPROM online promoter analysis tool, revealed possible binding sites for the transcription factors Lrp, Crp, and ArgR, which are global regulators in E. coli (Paul et al., 2007). The role of these factors in controlling pga expression will need be analyzed in future studies.

View this table:
Table 1

pgaA mRNA expression by quantitative real-time RT-PCR of strains J, JΔcsrA and JΔnhaR in TSB− supplemented with 1% glucose, 1% ethanol, 1% NaCl, 1 mM MnCl2 and 5% FBS

StrainTSB−+ Glucose+ Ethanol+MnCl2+NaCl+FBS
J1.2 (± 0.2)3.2 (± .0.3)3.0 (± 0.5)1.6 (± 0.1)3.2 (± 0.2)1.4 (± 0.3)
JΔcsrA2.7 (± 0.3)E314.1 (± 0.4)E34.4 (± 0.4)E3N/D6.4 (± 1.3)E3N/D
JΔnhaR0.7 (± 0.1)4.6 (± 0.4)2.1 (± 0.2)N/D0.9 (± 0.1)N/D
  • Values of a representative experiment. SD in parentheses.

  • * Represents a statistical difference (P<0.05, paired samples t-test). N/D, not done.

Analysis of PGA production by immunoblot and biofilm formation

In support of our finding that pgaA expression was higher in the presence of glucose, immunoblots revealed that PGA production was elevated in the presence of glucose (Fig. 3), as was biofilm formation (Fig. 4). Ethanol and NaCl induced a small but reproducible increase in PGA production, as detectable by the immunoblot (Fig. 3). Interestingly, of all the factors we used, glucose was the only that significantly enhanced biofilm formation (Fig. 4). We repeated the PNAG blot and biofilm formation assay with a second clinical strain, UTI-U, and found similar results (data not shown).

Figure 3

Immunoblot analysis of PGA production in the presence of 1% glucose, 1% ethanol, 1% NaCl, 1 mM MnCl2 and 5% FBS in TSB.

Figure 4

Biofilm formation in microtiter plates, due to effect of 1% glucose, 1% ethanol, 1% NaCl 1 mM MnCl2 and 5% FBS, in regular TSB, determined by safranin staining. The values shown here are from a representative experiment. *Represents a statistical difference (P<0.01 Student's t-test and P<0.05, paired samples t-test).

While it seems that PGA plays an important role in E. coli biofilms (Itoh et al., 2005), it is clearly not the only factor involved in biofilm formation in E. coli, because factors that induced PGA synthesis did not always increase biofilm thickness. Biofilm formation by many E. coli strains is characteristically poor in vitro (Wang et al., 2004). There are a number of genes that affect E. coli biofilm formation, some of which are independent of the pgaABCD operon (Corona-Izquierdo & Membrillo-Hernandez, 2002). It was found that curli biosynthesis is also essential for initial adhesion and biofilm formation in E. coli strains (Prigent-Combaret et al., 2000). Furthermore, there are other proteins that are involved in intercellular adhesion but do not appear to play relevant roles on biofilm formation, such as the Hek outer membrane protein (Fagan & Smith, 2007). A similar nonfundamental but still enhancing effect was found with the production of colanic acid (Prigent-Combaret et al., 2000).

While other factors may be required under certain conditions, PGA clearly plays an important role in E. coli biofilm formation, and it can act as a target for killing by opsonophagocytosis (Cerca et al., 2007b). Furthermore, the finding that both pga promoter expression and PGA production where induced by ethanol and high osmolarity is consistent with the idea that biofilm formation occurs under stressful environmental conditions (Costerton et al., 1995).


This work was supported by NIH grant R21AI61590. We thank Gerald B. Pier and Tomás Maira-Litran for providing the antiserum used to detect PNAG/PGA.


  • Editor: Robert Gunsalus


View Abstract