Bacteria are often found in multicellular communities known as biofilms, which constitute a resistance form against environmental stresses. Extracellular adhesion and cell aggregation factors, responsible for bacterial biofilm formation and maintenance, are tightly regulated in response to physiological and environmental cues. We show that, in Escherichia coli, inactivation of genes belonging to the de novo uridine monophosphate (UMP) biosynthetic pathway impairs production of curli fibers and cellulose, important components of the bacterial biofilm matrix, by inhibiting transcription of the csgDEFG operon, thus preventing production of the biofilm master regulator CsgD protein. Supplementing growth media with exogenous uracil, which can be converted to UMP through the pyrimidine nucleotide salvage pathway, restores csgDEFG transcription and curli production. In addition, however, exogenous uracil triggers cellulose production, particularly in strains defective in either carB or pyrB genes, which encode enzymes catalyzing the first steps of de novo UMP biosynthesis. Our results indicate the existence of tight and complex links between pyrimidine metabolism and curli/cellulose production: transcription of the csgDEFG operon responds to pyrimidine nucleotide availability, while cellulose production is triggered by exogenous uracil in the absence of active de novo UMP biosynthesis. We speculate that perturbations in the UMP biosynthetic pathways allow the bacterial cell to sense signals such as starvation, nucleic acids degradation, and availability of exogenous pyrimidines, and to adapt the production of the extracellular matrix to the changing environmental conditions.
Citation: Garavaglia M, Rossi E, Landini P (2012) The Pyrimidine Nucleotide Biosynthetic Pathway Modulates Production of Biofilm Determinants in Escherichia coli. PLoS ONE 7(2): e31252. doi:10.1371/journal.pone.0031252
Editor: Martin G. Marinus, University of Massachusetts Medical School, United States of America
Received: October 8, 2011; Accepted: January 5, 2012; Published: February 16, 2012
Copyright: © 2012 Garavaglia et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by the CHEM-PROFARMA-NET (Project RBPR05NWWC_004) and PRIN (Project 2008K37RHP) Research Programs of the Italian Ministry for University and Research. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Bacteria are able to switch between two different “lifestyles”: single planktonic cells and sessile microbial communities, or biofilms. Biofilm cells are characterized by production of adhesion factors and extracellular polysaccharides (EPS) constituting the so-called “biofilm matrix” that, in addition to promoting cell-cell aggregation and cell-surface adhesion, can confer bacterial cell resistance to various environmental stresses –. Transition from planktonic cells to biofilm, as well as biofilm maturation and dispersal, responds to environmental and physiological cues, usually relayed to the bacterial cell by signal molecules. Accumulation of signal molecules triggers biofilm formation and maintenance by stimulating the production of adhesion factors, either by activating transcription of corresponding genes or by increasing activity of EPS biosynthetic enzymes. In Gram negative bacteria, the modified nucleotide cyclic-di-GMP (c-di-GMP) plays a pivotal role in biofilm formation and maintenance by stimulating production of EPS and adhesion factors –, while negatively affecting cell motility , . Another class of signal molecules, homoserine lactones, can promote biofilm formation in the opportunistic pathogen Pseudomonas aeruginosa by promoting production of biosurfactants , , and by stimulating production of extracellular DNA  and of lectins, proteins able to promote cell adhesion to sugar moieties . In addition to dedicated signal molecules, intermediates and products of different metabolic pathways can also affect biofilm formation: for instance, indole, a product of tryptophan degradation, stimulates EPS production in Vibrio cholerae . Likewise, glucose and glycolysis intermediates can greatly impact adhesion factors' production through different regulatory mechanisms (reviewed in ).
In Escherichia coli and other enterobacteria, curli amyloid fibers greatly enhance cell aggregation and adhesion to surfaces. Genes involved in curli biosynthesis are clustered in the csgBAC operon, encoding curli structural components, and the csgDEFG operon, encoding the CsgD transcription regulator and proteins involved in curli assembly and transport , . The CsgD protein activates transcription of the csgBAC operon and of several genes involved in production of cell surface-associated structures and in cell adaptation to the biofilm lifestyle –, including the adrA gene, encoding a diguanylate cyclase able to trigger cellulose production via c-di-GMP synthesis , . Thus, curli, cellulose and other cell surface-associated structures are co-produced in a CsgD-dependent fashion to constitute the biofilm extracellular matrix. Expression of the csg operons takes place in response to a combination of environmental conditions: low growth temperature (<32°C), low osmolarity, and slow growth , and it is strongly dependent on the signal molecule c-di-GMP , . A number of regulators, including OmpR, IHF, H-NS, CpxR, Crl, and the RpoS protein, play a role in curli gene expression , –. However, several aspects of curli regulation are still unclear: for instance, the molecular mechanisms of temperature dependence have not yet been fully elucidated, and no c-di-GMP sensor element involved in csg activation has been identified so far.
In this work, we show that curli and cellulose production are tightly linked to nucleotide biosynthetic pathways. In particular, transcription of the curli operons is strongly affected by pyrimidine nucleotide availability, while cellulose production is activated in the presence of exogenous uracil. Our observations suggest that production of cellulose and curli, usually co-regulated, can be unbalanced depending on the activity of different UMP biosynthetic pathway. Coupling of curli and cellulose production to UMP biosynthesis modulates formation of extracellular structures in response to physiological and environmental cues, such as starvation, nucleic acid turnover, and availability of exogenous pyrimidines.
Mutations in the carB gene affect curli production
Amyloid fibers such as curli bind to the dye Congo red very efficiently ; thus, phenotype on Congo red-supplemented agar medium (CR medium, see Materials and Methods) provides a convenient method for curli detection and an easy way to screen mutants affected in curli production (Figure 1). To identify novel genes involved in curli regulation, we carried out transposon mutagenesis in the E. coli strain MG1655; mutants were screened for their phenotype on CR medium both at 30°C and 37°C, i.e., at permissive and non-permissive temperature for curli production. Several mutants were isolated that showed altered phenotype on CR medium (data not shown): one mutant displaying a dark red phenotype at 30°C and a weak red coloration at 37°C, suggesting increased curli production (Figure 1), was further characterized. Mapping of the Tn5<R6Kγori/KAN-2> transposon indicated that the insertion site lay in the carB gene, encoding a subunit of carbamoyl phosphate synthetase, which catalyzes the first step in the de novo pyrimidine nucleotide biosynthetic pathway (Figure 2). To verify that changes in phenotype in the carB::Tn5kan mutant of MG1655 were indeed due to altered curli production, we transduced the mutation in a strain unable to produce curli: the MG1655carB::Tn5kan ΔcsgA::cat double mutant displayed a white phenotype on CR medium both at 30°C and at 37°C (Figure 1), thus indicating that the dark red phenotype of the MG1655carB::Tn5kan mutant is totally dependent on curli fibers.
Phenotypes on CR medium of MG1655 (wild type strain), AM70 (csgA deletion mutant, unable to produce curli), MG1655carB::Tn5kan, MG1655carB::Tn5kan ΔcsgA::cat and MG1655ΔcarB::cat. Strains were grown either at 30°C (for 24 hours) or at 37°C (for 18 hours). Plates were incubated for 48 hours at 4°C to enhance Congo red binding.
Adapted from Ecocyc (http://ecocyc.org/).
Several pieces of evidence indicated that the carB::Tn5kan mutation does not result in the inactivation of carbamoyl phosphate synthase activity: the MG1655carB::Tn5kan mutant was not auxotrophic for pyrimidines, nor did it show any defect in growth rate on minimal medium. Finally, its phenotype on CR medium was not reversed by complementation with the wild type carB allele on a multicopy plasmid (data not shown). The insertion site for the Tn5kan transposon occurs at nucleotide 2720 of the carB gene, corresponding to the 907th codon, likely resulting in the production of a truncated form of the CarB protein lacking its regulatory domain involved in allosteric inhibition of protein activity by UMP . Loss of the regulatory domain suggests that CarB protein activity might be increased in the MG1655carB::Tn5kan mutant strain. To verify this hypothesis, we constructed a carB mutant in which the portion of the gene encoding the catalytic domain of the CarB protein had been deleted (MG1655ΔcarB::cat). As expected, this mutant was auxotrophic for pyrimidines, and showed reduced growth rate in LB1/4 medium (data not shown). Addition of uracil at 0.25 mM to LB1/4 medium (LB1/4(ura)) fully overcame MG1655ΔcarB::cat partial growth defect (data not shown). The MG1655ΔcarB::cat mutant displayed a white phenotype on CR medium, suggesting inability to produce curli (Figure 1), thus confirming the hypothesis that the carB::Tn5kan mutation does indeed result in enhanced carbamoyl phosphate synthetase activity.
Inactivation of UMP biosynthetic genes inhibits curli production at gene transcription level
To investigate whether the effects of carB inactivation could also be observed for other genes belonging to the de novo UMP biosynthetic pathway, we constructed knock out mutants in the pyrB, pyrC and pyrE genes, and tested them for their phenotypes on CR medium. As shown in Figure 3, inactivation of any UMP biosynthetic gene resulted in white phenotype on CR medium, indicating that curli production is inhibited by pyrimidine nucleotide starvation, rather than by lack (or accumulation) of any specific intermediate in the UMP biosynthetic pathway. Consistent with this result, strains impaired in de novo UMP biosynthesis were deficient in surface attachment experiments (Figure S1). To elucidate the mechanism of curli inhibition by perturbation of UMP biosynthesis, we measured transcript levels of the csgD and csgB genes, representatives of the two curli biosynthetic operons, using quantitative Real Time PCR (Table 1). Transcript levels of both csgD and csgB genes were dramatically decreased in every mutant deficient in UMP biosynthesis; in contrast, they were increased by approximately 3.5-fold in the MG1655carB::Tn5kan, in agreement with the dark red phenotype observed in this mutant (Figure 1). Consistent with inhibition of csgDEFG transcription, transcript levels of the CsgD-dependent adrA gene were also reduced by roughly 10-fold by mutations negatively affecting de novo UMP biosynthesis (Table 1).
The MG1655 strain and isogenic mutants deficient in UMP biosynthetic genes were spotted on either CR medium or CR(ura) medium (CR medium supplemented with 0.25 mM uracil) and grown for 24 hours at 30°C. Plates were incubated for 48 hours at 4°C to enhance Congo red binding.
In contrast to the genes belonging to the CsgD regulon, relative amounts of 16S rRNA, used as reference gene in Real Time PCR experiments, were similar in MG1655 and in the strains carrying non-functional alleles of UMP biosynthetic genes (data not shown), as were transcript levels of the cellulose biosynthetic bcsA gene, which is not regulated by the CsgD protein ,  (Table 1). These results strongly suggest that pyrimidine starvation leads to a reduction in csgD and csgB transcript levels via a specific mechanism rather than through a general inhibition of transcription. To determine whether reduction in csgD transcript levels could depend on decreased mRNA stability, we performed mRNA decay experiments, which did not show any significant difference in csgD mRNA half-lives in MG1655ΔcarB::cat in comparison to MG1655 (data not shown), suggesting that knock out mutations in the de novo UMP biosynthetic pathway affects csgD expression at the transcription initiation step.
Upon addition of 0.25 mM uracil to LB1/4 medium (LB1/4(ura) medium) transcription of both csgD and csgB was re-established in mutant strains affected in de novo UMP biosynthesis (Table 1), thus confirming that csgDEFG expression is repressed by pyrimidine starvation. However, surprisingly, addition of uracil to CR medium (CR(ura) medium) failed to restore the curli-dependent red phenotype in the carB and pyrB strains (Figure 3), in apparent contradiction with the results of the gene expression experiments. In contrast, the MG1655 strain, as well as the MG1655ΔpyrC and MG1655ΔpyrE mutants, displayed a red phenotype on CR(ura) medium, which was not affected by supplementing uracil up to a final concentration of 1 mM (data not shown). Surface adhesion experiments showed that growth in LB1/4(ura) only partially restored ability to form biofilm in the MG1655ΔcarB::cat and MG1655ΔpyrB::cat strains (Figure S1). These results could suggest that, although curli operon transcription was fully resumed in the presence of additional uracil, curli subunit production might still be impaired in the MG1655ΔcarB::cat and MG1655ΔpyrB::cat strains. However, determination of curli fibers' production using the SDS-agarose electrophoresis method  performed on MG1655ΔcarB::cat showed that was fully competent for curli production when grown in LB1/4(ura) solid medium (Figure S2), in agreement with gene expression experiments (Table 1).
Effects of regulatory proteins affecting pyrimidine metabolism and of inhibition of purine biosynthesis on curli production
We investigated whether pyrimidine starvation might affect curli production and csg gene expression via known pyrimidine-sensing regulatory proteins. To this aim, we constructed isogenic mutants of MG1655 in which either the cytR or the rutR gene were inactivated. The CytR protein is a repressor of genes involved in pyrimidine uptake and degradation; negative regulation by CytR is relieved by high intracellular concentrations of cytidine . Interestingly, in Vibrio cholerae, a CytR-like protein negatively controls biofilm formation by repressing EPS production . DNA binding by RutR, a regulator of genes involved both in pyrimidine biosynthesis and degradation, is inhibited by uracil . Thus, both CytR and RutR proteins regulate gene expression in response to intracellular pyrimidine concentrations. We tested the effects of the cytR and of the rutR mutations on CR phenotype, either in the presence or in the absence of exogenous uracil (Figure 4A): inactivation of the rutR gene did not affect CR phenotype, while, in contrast, the MG1655ΔcytR mutant strain displayed a white phenotype both on CR and on CR(ura) medium, indicative of reduced curli production. csgD transcript levels are reduced by roughly 5-fold in the ΔcytR mutant strain grown in LB1/4 medium, but they are restored to wild type levels by addition of 0.25 mM uracil (Figure 4B). In contrast, expression of the CytR-dependent udp gene, used as a control in gene expression experiments, are increased in the cytR mutant regardless of the presence of exogenous uracil, as expected (Figure 4B). Thus, the behavior of the cytR mutant with respect to curli production and csgD gene expression strongly resembles the MG1655ΔcarB::cat and MG1655ΔpyrB::cat strains (see Figure 3 and Table 1). These observations suggest that the CytR protein does not mediate pyrimidine-dependent regulation of the csg operons directly; however, lack of a functional cytR gene likely results in altered intracellular pyrimidine concentrations, which would in turn affect csgDEFG expression and curli production.
4A. The MG1655 strain and its isogenic mutants in the purH, cytR and rutR genes were spotted either on CR medium (left panel) or on CR(ura) medium (right panel) and grown for 24 hours at 30°C. Plates were incubated for 48 hours at 4°C to enhance Congo red binding. Determination of transcript levels. 4B. Relative expression of either the csgD gene (left panel) or the udp gene (right panel) was determined by Real-Time PCR on RNA extracted from overnight cultures of MG1655 and of its isogenic purH and cytR mutants. 16S RNA transcript was used as reference gene. ΔCt values between the genes of interest and 16S RNA were set at 1 for MG1655 in LB1/4 medium, and transcript levels in other strains and/or growth conditions are expressed as relative values. Experiments were repeated at least three times, each time in duplicate; standard deviations were always lower than 5%.
Our results show that pyrimidine starvation-dependent downregulation of csgDEFG expression and of curli production is not mediated by regulatory proteins directly involved in sensing intracellular pyrimidine concentrations. Thus, we hypothesized that pyrimidine starvation might downregulate csgDEFG expression through a general effect on intracellular nucleotide pools. As an initial verification of this hypothesis, we tested the effects of purine starvation on curli production and csgDEFG expression. Inactivation of the purine biosynthetic gene purH resulted in white phenotype on CR medium (Figure 4A) and in a 7-fold reduction of csgDEFG transcript levels (Figure 4B). Similar to what observed for mutations in de novo pyrimidine biosynthesis, purH inactivation does not result in a non-specific downregulation of transcription, since transcript levels of the CsgD-independent udp gene were unaffected in the purH mutant strain (Figure 4B). As expected, addition of 0.25 mM uracil did not revert the effects of the purH mutation (Figure 4), indicating that uracil can only counteract the effects of mutations specifically affecting UMP concentrations.
Since curli production and csgDEFG expression are strongly dependent upon the signal molecule c-di-GMP , , it is conceivable that changes in the nucleotide pools due to mutation in nucleotide biosynthetic genes could affect c-di-GMP production. This would be in agreement with our previous observations that sulfathiazole, a sulfonamide drug interfering with nucleotide biosynthesis, can inhibit c-di-GMP biosynthesis . Determination of intracellular c-di-GMP concentrations did not show significant differences in MG1655ΔcarB::cat and the MG1655carB::Tn5kan strains in comparison to MG1655 (data not shown); however, c-di-GMP concentrations in MG1655 cells are in the nanomolar range , making a precise determination of c-di-GMP in cell extracts rather difficult. In addition, it must be pointed out that induction of one specific diguanylate cyclase, sufficient for activation of its corresponding target, might not result in any significant increase in the overall concentration of intracellular c-di-GMP.
Uracil triggers cellulose production
Results presented in this work (Figure 3, Table 1, Figure S1) suggest that, in the MG1655ΔcarB::cat and MG1655ΔpyrB::cat strains grown in LB1/4(ura) medium, exposure of curli fibers on the cell surface might be hindered by production of additional extracellular structures. Indeed, it is known that overproduction of cellulose and other EPS can prevent curli-mediated Congo red binding and cell adhesion –. To test the possibility that exogenous uracil might affect phenotypes on CR medium in the MG1655ΔcarB::cat and the MG1655ΔpyrB::cat strains via cellulose overproduction, we inactivated bcsA, the first gene of the cellulose biosynthetic operon, in these genetic backgrounds. Deletion of the bcsA gene restored, albeit partially, the red phenotype on CR(ura) medium (Figure 5, data not shown), suggesting that the white phenotype on CR(ura) medium might indeed depend on EPS overproduction. Likewise, it resulted in efficient surface attachment by the MG1655ΔcarB::cat ΔbcsA::kan double mutant (Figure S1). In contrast, deletion of the bcsA gene did not affect either CR phenotype or surface attachment in the MG1655 strain (data not shown), in agreement with previous observations . To confirm our hypothesis further, we determined cellulose amounts in the MG1655, MG1655ΔcarB::cat and MG1655ΔpyrC::tet strains grown either in LB1/4 or in LB1/4(ura). Although growth in LB1/4(ura) enhanced cellulose production in all strains tested, this effect was much stronger in MG1655ΔcarB::cat, leading to production of a 3.5-fold higher amount of cellulose in comparison to MG1655 grown in the same conditions (Figure 6).
Phenotypes on CR medium of MG1655, MG1655ΔcarB::cat, MG1655ΔcarB::cat ΔbcsA::kan, MG1655ΔcarB::cat ΔyedQ::kan, MG1655ΔcarB::cat ΔadrA::kan. Strains were spotted on either CR medium or CR(ura) medium (CR medium supplemented with 0.25 mM uracil) and grown for 24 hours at 30°C. Plates were incubated for 48 hours at 4°C to enhance Congo red binding.
Strains MG1655, MG1655ΔcarB::cat, and MG1655ΔpyrC::tet were grown 48 hours at 30°C on either LB1/4 agar (no added uracil, Cellulose extraction and determination was performed as described . Data shown are the average of two independent experiments giving very similar results. For strains MG1655ΔcarB::cat and MG1655ΔpyrC::tet grown on LB1/4 agar no glucose was detectable in the assays. A value of 0.5 nmol glucose, corresponding to the lowest detectable concentration in the assay, as determined by a glucose standard curve, was thus arbitrarily assigned to these strains.
Enzymatic activity of the cellulose biosynthetic machinery is subject to regulation by the signal molecule c-di-GMP. Two distinct c-di-GMP synthetases, the AdrA and YedQ proteins, can activate cellulose production; although AdrA overexpression has been reported to affect curli production , , in physiological conditions AdrA sole function is to activate cellulose production , . AdrA and YedQ act independently and belong to two distinct regulatory circuits , , : indeed, while AdrA is encoded by a CsgD-dependent gene , thus presiding to co-ordinated production of curli and cellulose, YedQ expression and activity are independent of CsgD , . We tested the hypothesis that exogenous uracil might affect cellulose production via c-di-GMP synthesis by either AdrA or YedQ. Interestingly, inactivation of the yedQ gene, but not of adrA, in either the MG1655ΔcarB::cat (Figure 5) or the MG1655ΔpyrB::cat (data not shown) genetic backgrounds partially restored red phenotypes on CR medium, similar to the MG1655ΔcarB::cat ΔbcsA double mutant, thus suggesting that cellulose overproduction in the presence of exogenous uracil is mediated by the YedQ protein.
In this work, we have shown that mutations in genes belonging to de novo nucleotide biosynthetic pathways strongly affect csgDEFG expression and curli production in E. coli (Figure 1, Figures 3–4, Table 1, Figures S1, S2). Interplay between nucleotide metabolism and biofilm appears to be conserved in different bacteria; however, specific effects and mechanism may vary substantially. Indeed, although our results are consistent with previous findings showing that active de novo UMP biosynthesis is necessary for biofilm formation in P. aeruginosa , , in this bacterium inhibition of purine biosynthesis through inactivation of the purH gene does not affect adhesion factors' production , in contrast to what observed in E. coli (Figure 4). Likewise, pyrimidines appear to control EPS production and biofilm formation in V. cholerae through the dedicated regulator CytR , which does not appear to play a direct role in curli regulation in E. coli (Figure 4). Despite these differences, it seems that absence of de novo pyrimidine biosynthesis can act as a signal for severe nutrient starvation, which can in turn prevent biofilm formation and promote biofilm dispersal .
In E. coli, the effects of mutations in the de novo UMP biosynthesis on curli production can be complemented by supplementing growth medium with uracil, thus suggesting that pyrimidine nucleotide availability, regardless whether it is achieved via de novo UMP biosynthesis or the pyrimidine salvage pathway, allows efficient csgDEFG transcription and expression of the CsgD regulon (Table 1, Figure S2). Regulation of csgDEFG expression by intracellular nucleotide concentrations might take place by direct modulation of transcription initiation by RNA polymerase, similar to transcription control by GTP availability described for ribosomal promoters , or through not yet identified nucleotide-sensing regulatory proteins. Alternatively, perturbations in nucleotide pools might affect accumulation of c-di-GMP, a signal molecule necessary for csgDEFG expression , , possibly by impairing diguanylate cyclases' enzymatic activity.
Diguanylate cyclases play a role in pyrimidine-dependent regulation of cellulose production. Cellulose production is regulated by a more complex mechanism since, in addition to pyrimidine availability, it seems to respond to the relative activity of the two UMP biosynthetic pathways. Indeed, MG1655 produces twice as much cellulose when grown in the presence of exogenous uracil (Figure 6), i.e., in conditions in which UMP biosynthesis is mostly carried out via the pyrimidine salvage pathway and de novo UMP biosynthesis is inhibited , . Induction of cellulose production by exogenous uracil is further enhanced in mutants carrying non-functional carB or pyrB alleles (Figures 3, 5–6 and data not shown): in contrast, strains carrying mutations in later steps of the de novo UMP biosynthetic pathway, such as MG1655ΔpyrC::tet, do not overproduce cellulose in response to uracil (Figure 3, Figure 6). These observations suggest that bacterial cells might sense the molecular ratio between UMP and intermediates in the de novo UMP biosynthesis such as carbamoyl-L-aspartate, which accumulates in the pyrC mutant strain, as a signal of the relative balance between the two UMP biosynthetic pathways. An unbalance towards UMP biosynthesis via the pyrimidine salvage pathway triggers cellulose production, and this effect relies on the activity of the diguanylate cyclase YedQ (Figure 5).
The interplay between nucleotide salvage pathway and cellulose production might be connected to the role of cellulose and other EPS in the response to environmental stresses such as desiccation and resistance to bacteriophages , , . In bacterial biofilms, events leading to extensive cell lysis, such as exposure to antibiotics or attack by bacteriophages, would release cell components into the local environment: thus, a sudden increase in concentrations of exogenous nucleotides due to bacterial lysis might function as an “alarm signal” to neighboring cells, which would react by producing EPS as a defense mechanism against environmental stresses. For intracellular pathogenic enterobacteria, sensing an increase of exogenous nucleotide concentration might instead signal stress events in the host cell, such as leakage of nucleotides from the nuclear compartment. Consistent with our observations, it has been reported that allosteric inhibition of the CarB protein by exogenous uracil strongly influences production of extracellular structures and negatively affects expression of type III secretion systems in the intracellular pathogen Shigella flexneri . In Pseudomonas fluorescens, a spontaneous mutation in the carB gene affects the proportion of capsulated and non-capsulated subpopulations via yet unknown molecular mechanisms . Our results complement and expand these observations, and underline the importance of the interplay linking biofilm formation, bacterial virulence, production of extracellular structures, and nucleotide biosynthetic pathways: better understanding of these connections at the molecular level will allow us to improve our strategies in preventing (or promoting) bacterial biofilms. In this perspective, our results provide strong evidence to confirm previous findings suggesting that drugs targeting nucleotide biosynthetic pathways have a strong potential as antibiofilm agents , .
Materials and Methods
Bacterial strains and growth conditions
Bacterial strains used in this work are listed in Table 2. For strain construction and manipulation, bacteria were grown in LB medium (10 g/L Tryptone, 5 g/L Yeast Extract, 5 g/L NaCl). For adhesion assays and gene expression regulation studies, bacteria were grown in LB medium diluted 1∶4 in H2O (LB1/4). The LB1/4 medium was used since it allows efficient induction of the CsgD regulon  and provides sufficient pyrimidines to partially overcome the growth defect of strains mutated in de novo UMP biosynthetic pathway. When required, LB1/4 medium was supplemented with 0.25 mM uracil (LB1/4(ura) medium); uracil was dissolved in 50% dimethyl sulfoxide (DMSO) in water at a concentration of 10 mM. DMSO to a 1.25% final concentration was always added to control cultures.
For Congo red (CR) binding assays, overnight cultures were spotted, using a replicator, on LB1/4 agar medium to which 0.004% Congo red and 0.002% Coomassie blue were added after autoclaving (CR medium). Bacteria were grown for 20 h at 30°C; phenotypes were better detectable after 24–48 h incubation at 4°C. When needed, antibiotics were used at the following concentrations: ampicillin, 100 µg/ml; chloramphenicol, 50 µg/ml; kanamycin, 50 µg/ml; tetracycline, 25 µg/ml; rifampicin, 100 µg/ml.
Transposon insertion mutagenesis was carried out using the EZ-Tn5<R6Kγori/KAN-2> transposome (Epicentre). Transposon mutagenesis and determination of transposon insertion site by rescue cloning were carried out according to the manufacturer's instructions. E. coli MG1655 mutant derivatives were constructed either using the λ Red technique  or by bacteriophage P1 transduction . The list of primers used for gene inactivation and for confirmation of target gene disruption by PCR is presented in Table S1. Construction of the pCR2.1-carB plasmid was carried out by PCR amplification of the carB gene from the MG1655 genome followed by direct cloning of the PCR product into the pCR2.1vector (Invitrogen).
Gene expression studies
Determination of relative gene expression levels was performed by quantitative Real Time PCR, using bacterial cultures grown either in LB1/4 or in LB1/4(ura) at 30°C, and harvested either from overnight cultures or from exponential phase (OD600 nm = 0.6 for MG1655 and MG1655carB::Tn5kan, OD600 nm = 0.2 for strains carrying null mutations in UMP biosynthetic genes). Primers for Real-Time PCR are listed in Table S1. mRNA stability was measured by Real-Time PCR experiments in the presence of rifampicin as described . 16S RNA was always used as reference gene.
Detection of curli amyloid fibers was performed using the SDS-agarose electrophoresis method as described . Cellulose amount was estimated on bacterial cultures grown on solid medium for 48 hours; cells were collected, resuspended in H2O and centrifuged at 12.000×g for 10 minutes; cellulose was determined as glucose released from cellulase treatment on culture supernatants as previously described . Biofilm formation was determined with the surface attachment assay in microtiter plates  performed on bacterial cultures grown overnight in LB1/4 at 30°C.
Surface adhesion on polystyrene microtiter plates. Surface adhesion experiments were performed as previously described . White bars: overnight cultures grown in LB1/4 medium; grey bars: overnight cultures grown in LB1/4(ura) medium. Three independent experiments were performed and standard deviations are shown.
SDS-agarose gel. Curli production was detected using the SDS-agarose gel method . The same amount of total protein was loaded in each sample. Insoluble material, mostly constituted by curli amyloids, cannot migrate into the agarose gel and is stained by Coomassie blue. Cultures were grown on solid medium (LB1/4 agar or LB1/4(ura) agar) for 24 hours at 30°C.
Primers used in this work.
We would like to thank Louise Gourlay for correcting the manuscript, Letizia Tagliabue for performing the mRNA decay experiments, and Nadia Raffaelli and Paola Bocci (Università Politecnica delle Marche, Ancona, Italy) for performing c-di-GMP determination experiments.
Conceived and designed the experiments: MG ER PL. Performed the experiments: MG ER PL. Analyzed the data: MG ER PL. Contributed reagents/materials/analysis tools: PL. Wrote the paper: PL.
- 1. Costerton JW, Lewandowski Z, Caldwell DE, Korber DR, Lappin-Scott HM (1995) Microbial biofilms. Annu Rev Microbiol 49: 711–745.
- 2. Anderl JN, Franklin MJ, Stewart PS (2000) Role of antibiotic penetration limitation in Klebsiella pneumoniae biofilm resistance to ampicillin and ciprofloxacin. Antimicrob Agents Chemother 44: 1818–1824.
- 3. White AP, Gibson DL, Kim W, Kay WW, Surette MG (2006) Thin aggregative fimbriae and cellulose enhance long-term survival and persistence of Salmonella. J Bacteriol 188: 3219–3227.
- 4. Harrison JJ, Ceri H, Turner RJ (2007) Multimetal resistance and tolerance in microbial biofilms. Nat Rev Microbiol 5: 928–938.
- 5. Ross P, Mayer R, Benziman M (1991) Cellulose biosynthesis and function in bacteria. Microbiol Rev 55: 35–58.
- 6. Simm R, Morr M, Kader A, Nimtiz M, Römling U (2004) GGDEF and EAL domains inversely regulate cyclic di-GMP levels and transition from sessility to motility. Mol Microbiol 53: 1123–1134.
- 7. Kader A, Simm R, Gerstel U, Morr M, Römling U (2006) Hierarchical involvement of various GGDEF domain proteins in rdar morphotype development of Salmonella enterica serovar Typhimurium. Mol Microbiol 60: 602–616.
- 8. Weber H, Pesavento C, Possling A, Tischendorf G, Hengge R (2006) Cyclic-di-GMP-mediated signalling within the sigma network of Escherichia coli. Mol Microbiol 62: 1014–1034.
- 9. Méndez-Ortiz MM, Hyodo M, Hayakawa Y, Membrillo-Hernández J (2006) Genome-wide transcriptional profile of Escherichia coli in response to high levels of the second messenger 3′,5′-cyclic diguanylic acid. J Biol Chem 281: 8090–8099.
- 10. Boehm A, Kaiser M, Li H, Spangler C, Kasper CA, et al. (2010) Second messenger-mediated adjustment of bacterial swimming velocity. Cell 141: 107–116.
- 11. Davey ME, Caiazza NC, O' Toole GA (2003) Rhamnolipid surfactant production affects biofilm architecture in Pseudomonas aeruginosa PAO1. J Bacteriol 185: 1027–1036.
- 12. Pamp SJ, Tolker-Nielsen T (2007) Multiple roles of biosurfactants in biofilm development by Pseudomonas aeruginosa. J Bacteriol 189: 2531–2539.
- 13. Allesen-Holm M, Barken KB, Yang L, Klausen M, Webb JS, et al. (2006) A characterization of DNA release in Pseudomonas aeruginosa cultures and biofilms. Mol Microbiol 59: 1114–1128.
- 14. Diggle SP, Winzer K, Chhabra SR, Worrall KE, Cámara M, et al. (2003) The Pseudomonas aeruginosa quinolone signal molecule overcomes the cell density-dependency of the quorum sensing hierarchy, regulates rhl-dependent genes at the onset of stationary phase and can be produced in the absence of LasR. Mol Microbiol 50: 29–43.
- 15. Mueller RS, Beyhan S, Saini SG, Yildiz FH, Bartlett DH (2009) Indole acts as an extracellular cue regulating gene expression in Vibrio cholerae. J Bacteriol 191: 3504–3516.
- 16. Karatan E, Watnick PI (2009) Signals, regulatory networks, and materials that build and break bacterial biofilms. Microbiol Mol Biol Rev 73: 310–347.
- 17. Hammar M, Arnqvist A, Bian Z, Olsen A, Normark S (1995) Expression of two csg operons is required for production of fibronectin- and congo red-binding curli polymers in Escherichia coli K-12. Mol Microbiol 18: 661–670.
- 18. Romling U, Sierralta WD, Eriksson K, Normark S (1998) Multicellular and aggregative behaviour of Salmonella typhimurium strains is controlled by mutations in the agfD promoter. Mol Microbiol 28: 249–264.
- 19. Latasa C, Roux A, Toledo-Arana A, Ghigo JM, Gamazo C, et al. (2005) BapA, a large secreted protein required for biofilm formation and host colonization of Salmonella enterica serovar Enteritidis. Mol Microbiol 58: 1322–1339.
- 20. Gibson DL, White AP, Snyder SD, Martin S, Heiss C, et al. (2006) Salmonella produces an O-antigen capsule regulated by AgfD and important for environmental persistence. J Bacteriol 188: 7722–7730.
- 21. Gualdi L, Tagliabue L, Landini P (2007) A biofilm formation-gene expression relay system in Escherichia coli: modulation of σS-dependent gene expression by the CsgD regulatory protein via σS protein stabilization. J Bacteriol 189: 8034–8043.
- 22. Romling U, Rohde M, Olsen A, Normark S, Reinkoster J (2000) AgfD, the checkpoint of multicellular and aggregative behaviour in Salmonella typhimurium regulates at least two independent pathways. Mol Microbiol 36: 10–23.
- 23. Pratt LA, Silhavy TJ (1998) Crl stimulates RpoS activity during stationary phase. Mol Microbiol 29: 1225–1236.
- 24. Prigent-Combaret C, Brombacher E, Vidal O, Ambert A, Lejeune P, et al. (2001) Complex regulatory network controls initial adhesion and biofilm formation in Escherichia coli via regulation of the csgD gene. J Bacteriol 183: 7213–7223.
- 25. Gerstel U, Park C, Romling U (2003) Complex regulation of csgD promoter activity by global regulatory proteins. Mol Microbiol 49: 639–654.
- 26. Delannay S, Charlier D, Tricot C, Villeret V, Piérard A, et al. (1999) Serine 948 and threonine 1042 are crucial residues for allosteric regulation of Escherichia coli carbamoylphosphate synthetase and illustrate coupling effects of activation and inhibition pathways. J Mol Biol 286: 1217–1228.
- 27. Zogaj X, Nimtz M, Rohde M, Bokranz W, Römling U (2001) The multicellular morphotypes of Salmonella typhimurium and Escherichia coli produce cellulose as the second component of the extracellular matrix. Mol Microbiol 39: 1452–1463.
- 28. Robbe-Saule V, Jaumouille V, Prevost MC, Guadagnini S, Talhouarne C, et al. (2006) Crl activates transcription initiation of RpoS-regulated genes involved in the multicellular behavior of Salmonella enterica serovar Typhimurium. J Bacteriol 188: 3983–3994.
- 29. Sitaras C, Naghavi M, Herrington MB (2011) Sodium dodecyl sulfate-agarose gel electrophoresis for the detection and isolation of amyloid curli fibers. Anal Biochem 408: 328–331.
- 30. Valentin-Hansen P, Søgaard-Andersen L, Pedersen H (1996) A flexible partnership: the CytR anti-activator and the cAMP-CRP activator protein, comrades in transcription control. Mol Microbiol 20: 461–466.
- 31. Haugo AJ, Watnick PI (2002) Vibrio cholerae CytR is a repressor of biofilm development. Mol Microbiol 45: 471–483.
- 32. Nguyen Ple M, Bervoets I, Maes D, Charlier D (2010) The protein-DNA contacts in RutR•carAB operator complexes. Nucleic Acids Res 38: 6286–6300.
- 33. Antoniani D, Bocci P, Maciag A, Raffaelli N, Landini P (2010) Monitoring of di-guanylate cyclase activity and of cyclic-di-GMP biosynthesis by whole-cell assays suitable for high-throughput screening of biofilm inhibitors. Appl Microbiol Biotechnol 85: 1095–1104.
- 34. Wang X, Rochon M, Lamprokostopoulou A, Lunsdorf H, Nimitz M, et al. (2006) Impact of biofilm matrix components on interaction of commensal Escherichia coli with the gastrointestinal cell line HT-29. Cell Mol Life Sci 63: 2352–2363.
- 35. Gualdi L, Tagliabue L, Bertagnoli S, Ieranò T, De Castro C, et al. (2008) Cellulose modulates biofilm formation by counteracting curli-mediated colonization of solid surfaces in Escherichia coli. Microbiology 154: 2017–2024.
- 36. Ma Q, Wood TK (2009) OmpA influences Escherichia coli biofilm formation by repressing cellulose production through the CpxRA two-component system. Environ Microbiol 11: 2735–2746.
- 37. Tagliabue L, Antoniani D, Maciag A, Bocci P, Raffaelli N, et al. (2010) The diguanylate cyclase YddV controls production of the exopolysaccharide poly-N-acetylglucosamine (PNAG) through regulation of the PNAG biosynthetic pgaABCD operon. Microbiology 156: 2901–2911.
- 38. Da Re S, Ghigo JM (2006) A CsgD-independent pathway for cellulose production and biofilm formation in Escherichia coli. J Bacteriol 188: 3073–3087.
- 39. Romling U (2005) Characterization of the rdar morphotype, a multicellular behaviour in Enterobacteriaceae. Cell Mol Life Sci 62: 1234–1246.
- 40. Attila C, Ueda A, Wood TK (2009) 5-Fluorouracil reduces biofilm formation in Escherichia coli K-12 through global regulator AriR as an antivirulence compound. Appl Microbiol Biotecnol 82: 525–533.
- 41. Ueda A, Attila C, Whiteley M, Wood TK (2009) Uracil influences quorum sensing and biofilm formation in Pseudomonas aeruginosa and fluorouracil is an antagonist. Microbial Biotech 2: 62–74.
- 42. Gjermansen M, Ragas P, Sternberg C, Molin S, Tolker-Nielsen T (2005) Characterization of starvation-induced dispersion in Pseudomonas putida biofilms. Environ Microbiol 7: 894–906.
- 43. Gaal T, Bartlett MS, Ross W, Turnbough CL Jr, Gourse RL (1997) Transcription regulation by initiating NTP concentrations: rRNA synthesis in bacteria. Science 278: 2092–2097.
- 44. Turnbough CL Jr, Switzer RL (2008) Regulation of pyrimidine biosynthetic gene expression in bacteria: repression without repressors. Microbiol Mol Biol Rev 72: 266–300.
- 45. Hughes KA, Sutherland IW, Jones MV (1998) Biofilm susceptibility to bacteriophage attack: the role of phage-borne polysaccharide depolymerase. Microbiology 144: 3039–3047.
- 46. Durand JM, Bjork GR (2009) Metabolic control through ornithine and uracil of epithelial cell invasion by Shigella flexneri. Microbiology 155: 2498–2508.
- 47. Beaumont HJ, Gallie J, Kost C, Ferguson GC, Rainey PB (2009) Experimental evolution of bet hedging. Nature 462: 90–93.
- 48. Perrin C, Briandet R, Jubelin G, Lejeune P, Mandrand Berthelot MA, et al. (2009) Nickel promotes biofilm formation by Escherichia coli K-12 strains that produce curli. Appl Environ Microbiol 75: 1723–1733.
- 49. Datsenko KA, Wanner BL (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 97: 6640–6645.
- 50. Miller JH, editor. (1972) Experiments in molecular genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
- 51. Wang X, Dubey AK, Suzuki K, Baker CS, Babitzke P, et al. (2005) CsrA post-transcriptionally represses pgaABCD, responsible for synthesis of a biofilm polysaccharide adhesion of Escherichia coli. Mol Microbiol 56: 1648–1663.
- 52. Blattner FR, Plumkett G III, Bloch CA, Perna NT, Burland V, et al. (1997) The complete genome sequence of Escherichia coli K-12. Science 277: 1453–1474.