Genetic Analysis of Vibrio cholerae Monolayer Formation Reveals a Key Role for ΔΨ in the Transition to Permanent Attachment

Abstract

A bacterial monolayer biofilm is a collection of cells attached to a surface but not to each other. Monolayer formation is initiated when a bacterial cell forms a transient attachment to a surface. While some transient attachments are broken, others transition into the permanent attachments that define a monolayer biofilm. In this work, we describe the results of a large-scale, microscopy-based genetic screen for Vibrio cholerae mutants that are defective in formation of a monolayer biofilm. This screen identified mutations that alter both transient and permanent attachment. Transient attachment was somewhat slower in the absence of flagellar motility. However, flagellar mutants eventually formed a robust monolayer. In contrast, in the absence of the flagellar motor, monolayer formation was severely impaired. A number of proteins that modulate the V. cholerae ion motive force were also found to affect the transition from transient to permanent attachment. Using chemicals that dissipate various components of the ion motive force, we discovered that dissipation of the membrane potential (ΔΨ) completely blocks the transition from transient to permanent attachment. We propose that as a bacterium approaches a surface, the interaction of the flagellum with the surface leads to transient hyperpolarization of the bacterial cell membrane. This, in turn, initiates the transition to permanent attachment.

Vibrio cholerae colonizes surfaces in marine environments, freshwater environments, and the mammalian small intestine (59). Surface colonization can take the form of a monolayer, which is the result of interactions between individual bacteria and the surface, or of a multilayer, which is the result of interactions between neighboring bacteria and between these bacteria and the colonized surface. Three types of V. cholerae multilayer biofilms have been defined, each of which is activated by particular environmental conditions (Fig. 1). The seawater biofilm requires only environmental Ca²⁺, which is thought to form bridges between negatively charged V. cholerae O-antigen surface polysaccharides on neighboring cells (25, 26). When expressed in vitro and in the mammalian intestine, toxin-coregulated pili on neighboring cells bundle to form the cellular aggregates that define the toxin-coregulated pilus-dependent biofilm (27, 55, 57). Finally, environments containing monosaccharides, bile, certain polyamines, or quorum-sensing autoinducers stimulate the Vibrio polysaccharide-dependent biofilm (15, 18, 24, 25, 37). There are many interesting aspects of multilayer biofilms, including regulation of matrix synthesis, matrix structure and function, differentiation of bacteria within the biofilm, and resistance of biofilm-associated bacteria to antibiotics. As a result, these three-dimensional bacterial collections have been intensely scrutinized.

FIG. 1.

Diagram of the four types of V. cholerae biofilms that have been defined. The stages of formation of the monolayer biofilm, which was the focus of this study, are shown on the right. TCP, toxin-coregulated pilus;VPS, Vibrio polysaccharide.

While the monolayer biofilm has been less well studied, it may well be the more relevant form of surface attachment in many interactions of V. cholerae and other bacteria with a host. In both commensal and pathogenic interactions with epithelial surfaces, adhesion may never progress beyond the monolayer stage because of bacterial internalization, harsh environmental conditions that preclude bacterial replication, or the absence of environmental signals that promote intercellular interactions. If a multilayer biofilm does develop, the monolayer biofilm is likely to serve as a catalytic intermediate in this process.

Visual inspection of the process of monolayer formation by a motile bacterium suggests that cells first form a loose association with the surface. This is termed transient attachment because cells frequently escape the surface to become free-swimming organisms again. Loosely attached cells, however, may transition to a more stable association with the surface known as permanent attachment. The resulting permanently attached cells comprise the monolayer biofilm (Fig. 1).

In V. cholerae, transient attachment is mediated primarily by the mannose-sensitive hemagglutinin (MSHA) type IV pilus (37). However, in chitin-rich environments, the chitin-regulated type IV pilus (ChiRP) also plays a role in transient attachment (33). Arrest of flagellar motility appears to be the defining feature of the V. cholerae transition to permanent attachment. First, flagellar filament mutants form a denser monolayer biofilm (37). Second, transcription of flagellar genes is repressed in monolayer-associated cells (37, 38). Interestingly, in Bacillus subtilis, a molecular clutch that disengages the flagellum from its rotor has recently been identified. This clutch, which is encoded by a gene in an operon containing genes required for biofilm formation, has been proposed to facilitate the transition of free-swimming bacteria to the biofilm-associated state (2). So far, no such structure has been identified in V. cholerae.

We previously reported that in a minimal medium containing only amino acids as a carbon source, wild-type V. cholerae forms a monolayer biofilm (37). Furthermore, by studying a limited number of genes, we established unique genetic requirements for monolayer biofilm formation compared with formation of the multilayer biofilm (37, 38). In the present study, we performed a large-scale, microscopy-based genetic screen to identify novel mutations that alter V. cholerae monolayer formation. This screen identified structures previously known to participate in monolayer formation, such as the MSHA pilus and the flagellum. In addition, we identified a number of proteins with diverse functions whose common theme is the ability to modulate the ion motive force of V. cholerae. To further elucidate the role of the ion motive force in monolayer formation, we utilized ionophores to demonstrate that membrane potential (ΔΨ) plays a critical role in the transition from transient to permanent attachment. Based on our results, we propose a model in which arrest of the flagellar motor during transient attachment leads to hyperpolarization of the cell membrane. This, in turn, initiates the transition to permanent attachment.

MATERIALS AND METHODS

Bacterial strains and media.

Relevant bacterial strains and plasmids used in this study and the primers used to generate them are listed in Tables 1 and 2, respectively. Strains were propagated in Luria-Bertani (LB) broth and then diluted into the previously described monolayer-specific minimal medium (MM) for quantification of monolayer formation (37). For quantification of the total attached cells, monolayers were rinsed with phosphate-buffered saline (pH 7.4) (PBS). For quantification of only permanently attached cells, monolayers were rinsed with PBS supplemented with 0.1% α-methylmannoside (PBS/AMM). Where indicated below, the medium was supplemented with streptomycin (100 μg/ml; Sigma), ampicillin (150 μg/ml; Sigma), or kanamycin (50 μg/ml; Sigma). Carbonyl cyanide m-chlorophenylhydrazone (CCCP) (Sigma) and valinomycin (Sigma) were used to test the effects of ionophores on monolayer formation. A 500 μM stock solution of CCCP was prepared using dimethyl sulfoxide. For all experiments, this solution was diluted in MM to obtain a final concentration of 0.5 μM. A 10 mM stock solution of valinomycin was prepared in ethanol and diluted to obtain the concentrations indicated below. For monolayer experiments, a stock solution of 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO) (Alexis Biochemicals), an inhibitor of sodium-pumping NADH-ubiquinone oxidoreductase (Na⁺-NQR), was prepared in ethanol and then diluted 1:1,000 in MM to obtain a final concentration of 50 μM.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Genotype	Reference or source
E. coli strains
SM10λpir	thi thr leu tonA lacY supE recA::RP4-2-Tc::MuλpirR6K Kmr	37
PW821	SM10λpir(pWM91ΔVC2704)	This study
PW819	SM10λpir(pWM91ΔVC2705)	This study
PW868	SM10λpir(pWM91ΔVCA0667)	This study
V. cholerae strains
PW249	MO10; Smr	16
PW357	MO10 lacZ::vpsLp→lacZ Smr	16
PW412	MO10 lacZ::vpsLp→lacZ ΔflaA Smr	38
PW391	MO10 lacZ::vpsLp→lacZ ΔmotX Smr	This study
PW413	MO10 lacZ::vpsLp→lacZ ΔflaA ΔmotX Smr	This study
PW477	MO10 ΔopuD Smr	22
PW478	MO10 ΔputP Smr	22
PW822	MO10 ΔsssH Smr	This study
PW820	MO10 ΔsssA Smr	This study
PW869	MO10 ΔVCA0667 Smr	This study
Plasmids
pWM91	oriR6KmobRP4 lacI pTac tnp mini-Tn10Km Kmr Apr	34
pWM91ΔflaA	pWM91 carrying a fragment of flaA harboring an internal, unmarked deletion	38
pWM91ΔmotX	pWM91 carrying a fragment of motX harboring an internal, unmarked deletion	9
pWM91ΔVC2704	pWM91 carrying a fragment of VC2704 harboring an internal, unmarked deletion	This study
pWM91ΔVC2705	pWM91 carrying a fragment of VC2705 harboring an internal, unmarked deletion	This study
pWM91ΔVCA0667	pWM91 carrying a fragment of VCA0667 harboring an internal, unmarked deletion	This study
pBAD-TOPO-motX	pBAD-TOPO carrying the gene at locus VC2601	This study
pBAD-TOPO/lacZ/V5-His	pBAD-TOPO carrying the E. coli lacZ gene with V5 and six-His tags	Invitrogen

TABLE 2.

Primers used for PCR

Primer	Description	Sequence
Construction of V. cholerae VC2705 deletion
P427	Forward primer for upstream fragment	CCTTCGCTTTAACGCTGTTT
P428	Reverse primer for upstream fragment	TTACGAGCGGCCGCAGATGCCGACCAAAATAAACG
P429	Forward primer for downstream fragment	TGCGGCCGCTCGTAAGTGCAGGATATGGTGGAAT
P430	Reverse primer for downstream fragment	CATGGATGTGCATGTGTTCA
Construction of V. cholerae VC2704 deletion
P431	Forward primer for upstream fragment	TTTCGAGGAGACAGGGCTAA
P432	Reverse primer for upstream fragment	TTACGAGCGGCCGCAGGCTTGAGCATGTTCAGATGATT
P433	Forward primer for downstream fragment	TGCGGCCGCTCGTAAGCGCTCGACAAGAAATATAACG
P434	Reverse primer for downstream fragment	TGATGTCCACTTCAAGGAAGC
Construction of V. cholerae VCA0667 deletion
P461	Forward primer for upstream fragment	GCCGTGTAATTGATGAGATAG
P462	Reverse primer for upstream fragment	TTACGAGCGGCCGCATGTATGTCGCTAAAATAGAAAC
P463	Forward primer for downstream fragment	TGCGGCCGCTCGTAAGCTCATTTGTTCTTGGTCTTT
P464	Reverse primer for downstream fragment	ATTGTCACAGCCGCGAAAACC
Construction of V. cholerae pBADmotX for rescue
pBADMotXN	N-terminal primer	GAGGAATAATAAATGAAGCTACGAACGGTAGCC
pBADMotXC	C-terminal primer	TCAATGGTGATGGTGATGATGCCAGAACGTTTCTCGGCGTTT
Arbitrary PCR: first round of PCR
Arb1	Arbitrary primer 1	GGCCACGCGTCGACTAGTACNNNNNNNNNNGATAT
Arb6	Arbitrary primer 2	GGCCACGCGTCGACTAGTACNNNNNNNNNNACGCC
P20	Tn10-specific primer	CCGCGGTGGAGCTCC
Mar5	Himar-specific primer	GGTTGGTACTATATAAAAATA
Arbitrary PCR: second round of PCR (nested primers)
Arb2	Complementary primer for Arb1 and Arb6	GGCCACGCGTCGACTAGTAC
P2	Tn10-specific primer	ATGACAAGATGTGTATCCACC
Mar3	Himar-specific primer	ATATGCATTTAATACTAGCGA

Transposon mutagenesis.

In addition to a previously generated library of mini-Tn10 transposon insertion mutants (17), a new library of Himar transposon insertion mutants (53) was constructed as follows. Wild-type V. cholerae (Sm^r) and Escherichia coli strain SM10λpir harboring the conditionally replicating plasmid pFD1 (Ap^r Km^r) were cultured on LB agar plates containing streptomycin (100 μg/ml) and LB agar plates containing ampicillin (150 μg/ml), respectively. These strains were then mated on an LB agar plate for 2 h at 37°C. V. cholerae carrying transposon insertions were isolated by growth on LB agar containing streptomycin (100 μg/ml) and kanamycin (50 μg/ml). Mutants were transferred to fresh LB agar containing streptomycin and kanamycin, grown overnight at 27°C, and used as described below to test monolayer formation.

Screen for mutants forming altered monolayers.

Transposon insertion mutants were isolated, transferred to the wells of a 96-well plate containing 100 μl of LB broth, and incubated for 6 to 8 h at 27°C. One microliter of each culture was transferred to a well of a 24-well, non-tissue-culture-treated polystyrene plate (Falcon) containing 300 μl of MM. The plates were incubated overnight (16 h) at 27°C with gentle agitation to allow monolayers to form.

At the end of the incubation, planktonic cells were removed. The remaining planktonic and transiently attached cells were separated from monolayers by vigorous agitation for 15 min in the presence of 300 μl of PBS/AMM. This procedure was repeated two times. Washed monolayers were assessed qualitatively by visualization with an Eclipse TE200-E microscope (Nikon) equipped with an IEEE1394 digital charge-coupled device camera (Hamamatsu), and mutants forming altered monolayers were stored at −80°C for further study.

Secondary screens.

Selected mutants found to be defective for monolayer formation were evaluated by using the following secondary screens. Growth curves were determined by inoculating the strains of interest into the wells of 96-well plates containing 100 μl of MM. Cultures were incubated at 27°C with gentle agitation, and the optical density at 655 nm (OD₆₅₅) was determined at various times over a 24-h period. Motility and biofilm formation in LB broth were assayed as previously described (63). Mannose-sensitive hemagglutination assays were performed as previously described (63), except that red blood cells harvested from sheep (Sigma) were used.

Arbitrary PCR.

For selected mutants demonstrating altered monolayer formation, arbitrary PCR (17, 45) was used to amplify genomic DNA neighboring the inserted transposon. In this technique, the DNA sequence surrounding the transposon insertion site is amplified by two rounds of PCR. The first round uses a primer specific to the transposon and one or more primers that are designed to anneal to arbitrary sites on the chromosomal DNA. The second round uses a nested primer unique to the transposon and a primer that is identical to the 5′ end of the arbitrary primers used in the first round. For transposon insertion mutants generated using the mini-Tn10 transposon, insertion junctions were amplified using primers P20, ARB1, and ARB6 in round 1, followed by primers P2 and ARB2 in round 2 (17). For transposon insertion mutants generated using the Himar transposon, insertion junctions were amplified using primers Mar5, ARB1, and ARB6 in round 1, followed by primers Mar3 and ARB2 in round 2. The sequences of these primers are shown in Table 2. Arbitrary PCR products were purified using a QIAquick PCR purification kit (Qiagen) and then sequenced using transposon-specific primers.

Construction of deletion mutants.

V. cholerae mutants carrying in-frame deletions in ΔflaA and ΔmotX were made as previously described using available plasmids (11, 17, 37). V. cholerae mutants carrying in-frame deletions within the genes at loci VC2704, VC2705, and VCA0667 were constructed as previously described (17). Briefly, the PCR primers listed in Table 2 were used to amplify two noncontiguous approximately 500-bp fragments including small portions of the N-terminal and C-terminal sequences of the gene of interest. The two fragments were joined using the gene-splicing-by-overlap-extension technique (17). The resulting deletion fragment was gel purified and cloned into the pCR2.1 TOPO vector (Invitrogen). Amplification of the correct fragments was confirmed by sequence analysis of the pCR2.1 insertion. The fragment was then removed from pCR2.1 by digestion with SpeI and XhoI and ligated into the suicide plasmid pWM91 to obtain the plasmids listed in Table 1. The plasmids were transformed into E. coli strain SM10λpir and transferred into V. cholerae by conjugation. V. cholerae strains harboring a deletion of the gene of interest were created by double homologous recombination as described previously (17). The VC2704, VC2705, and VCA0667 deletions removed 198, 1,599, and 1,040 nucleotides from the coding sequences of the genes, respectively.

Construction of motX rescue plasmid.

The gene at locus VC2601, which contains motX, was amplified using PCR. PCR primers were designed to remove the N-terminal leader sequence and the C-terminal V5 epitope tag and polyhistidine region flanking the pBAD vector cloning site (Invitrogen). Instead, a six-His sequence was included in the C-terminal primer. Amplified products were cloned into a pBAD-TOPO expression vector to obtain the rescue construct pBAD-TOPO-motX. The insertion was confirmed by DNA sequence analysis. In addition, the pBAD-TOPO-motX plasmid was introduced into the V. cholerae ΔmotX mutant by electroporation, and rescue of motility was confirmed by inoculation into swarm agar (data not shown).

Analysis of monolayer formation.

For formation of monolayer biofilms, the strains of interest were cultured in LB broth or MM overnight. In the morning, the cultures were diluted into 300 μl of MM in 24-well, non-tissue-culture-treated polystyrene plates (Falcon) to obtain the starting OD₆₅₅ noted below. Monolayers were formed by incubation at 27°C with gentle agitation. At the end of the incubation period, planktonic cells were removed. The remaining planktonic and transiently attached cells were separated from monolayer-associated cells by vigorous agitation for 15 min in the presence of 300 μl of PBS or PBS/AMM as noted below. This procedure was repeated two times. Washed monolayers were visualized with an Eclipse TE200-E microscope (Nikon) equipped with an IEEE1394 digital charge-coupled device camera (Hamamatsu). IPLab software (Scanalytics, Inc.) was used for image acquisition and quantification. Where noted below, images were collected at a magnification of ×400, and the total area of the surface covered by cells in each image was quantified. All measurements included at least three biological replicates and were repeated several times.

Measurement of monolayer formation and ΔΨ in the presence of chemicals.

For monolayer formation and ΔΨ measurement with chemicals, wild-type V. cholerae was cultured overnight in MM at 37°C. The cultures were diluted 1:10 into 1 ml of fresh MM prepared with the chemical indicated below. The cells were allowed to adjust to the presence of the chemical for 30 min and then transferred into six 24-well plates using 300-μl aliquots. Monolayers were allowed to form for 1 h and were then rinsed three times by vigorous agitation for 5 min in the presence of 300 μl of PBS/AMM. Surface coverage was evaluated as described above.

For ΔΨ measurement, 3,3′-diethyloxacarbocyanine iodide (Molecular Probes) was added to a final concentration of 30 μM after incubation of cells with the indicated chemical for 1 h at 27°C. The plates were incubated for an additional 30 min in the dark. The fluorescence intensity in each well was then measured using an HTS7000 spectrophotometer (Perkin Elmer) with a 485-nm excitation filter and a 535-nm emission filter.

qRT-PCR.

After growth of V. cholerae to exponential phase in MM, valinomycin was added to a final concentration of 10 μM, and the cells were incubated with gentle shaking for 90 min at 27°C. The cells were then pelleted by centrifugation, and total RNA was isolated using an RNeasy kit (Qiagen). Quantitative reverse transcription-PCR (qRT-PCR) was performed using 1 ng of total RNA in a 25-μl mixture, the relevant primer pair (Table 2), and a QuantiTect SYBR green RT-PCR kit (Qiagen). The level of the clpX (VC1921) transcript was used to normalize all qRT-PCRs. Template-free and reverse transcriptase-free reactions were included to verify that no contaminants were present. The experiments were conducted with an ABI Prism 7000 sequence detection system (Applied Biosystems) using the following steps: (i) 50°C for 30 min; (ii) 95°C for 15 min; and (iii) 40 cycles of denaturation for 15 s at 94°C, annealing for 30 s at 55°C, and extension for 30 s at 72°C. A dissociation curve was determined for each primer pair. Data were analyzed using the 7000 system software (Applied Biosystems). Measurements were performed in triplicate.

RESULTS

Genetic screen to identify monolayer mutants.

While multiple forward genetic screens designed to identify mutants forming altered biofilms have been performed for a variety of microbes, including V. cholerae (5-8, 16, 30, 43, 44, 49, 50, 52, 58, 60, 61, 63-65), this is one of the first reports of a screen for mutants forming altered monolayers. Our goal was to identify mutants that formed altered transient or permanent attachments. We previously showed that while AMM blocks reattachment of transiently attached cells, permanently attached cells are immune to the action of this compound. Therefore, in order to detect mutants that were unable to form either transient or permanent attachments, monolayers were rinsed as described above with PBS/AMM.

Mutant monolayers were examined visually by phase-contrast microscopy. A total of 5,617 transposon insertion mutants were screened. In the initial screen, we identified 327 mutants that formed monolayer biofilms that were different from the monolayer biofilms formed by wild-type V. cholerae. Monolayer formation by these mutants was retested in triplicate. Two hundred fifty-nine mutants demonstrated altered monolayer formation upon repeat testing. The types of monolayers identified and the scoring system utilized are shown in Fig. 2. In secondary screens, we tested mutants for growth in MM, for motility in swarm agar, and for hemagglutination. Mutants with similar phenotypes in the secondary screens were grouped together, and the transposon insertion sites of several mutants in each group were identified by sequence analysis of arbitrary PCR products. The transposon insertion sites of a total of 109 mutants (40%) were identified. Forty-seven transposon insertion mutants (18%) were not characterized because their behavior was indistinguishable from that of wild-type V. cholerae in secondary screens for growth, hemagglutination, and motility, and we were unable to identify the transposon insertion site by arbitrary PCR. Table 3 shows functional categories of genes whose mutation altered monolayer formation but not the growth rate. While the genes carrying the transposon insertions are listed, polar effects on neighboring genes may be responsible for the observed monolayer phenotypes of the transposon insertion mutants. For this reason, to study the monolayer phenotype in more detail, we subsequently constructed mutants having in-frame deletions in genes of interest.

FIG. 2.

Micrographs of monolayers formed by representative transposon insertion mutants belonging to each phenotypic category identified in the genetic screen. The interrupted genes and phenotypic classifications of the mutants are indicated above the images. The upward and downward arrows indicate increased and decreased monolayers, respectively, and the number of arrows indicates the severity of the defect. The aggregative phenotype is discussed in the text. Bar = 10 μm. WT, wild type.

TABLE 3.

Mutants identified in monolayer screen

Category	Gene	Description	Phenotypea
Pilus function and biogenesis	VC0402	MSHA biogenesis protein MshL	↓↓↓
	VC0403	MSHA biogenesis protein MshM	↓ (two insertions)
	VC0404	MSHA biogenesis protein MshN	↓↓↓ (three insertions)
	VC0406	MSHA biogenesis protein MshG	↓↓↓
	VC0408	MSHA pilin protein MshB	↓↓↓, ↓↓ (two insertions)
	VC0409	MSHA pilin protein MshA	↓↓↓ (two insertions)
	VC0410	MSHA pilin protein MshC	↓↓↓ (four insertions)
	VC0413	Hypothetical protein (in MSHA operon)	↓↓↓
	VC0414	Hypothetical protein (in MSHA operon)	↓↓↓ (nine insertions)
	VC0462	Twitching motility protein PilT-1	↓↓
	VC0463	Twitching motility protein PilT-2	↓↓ (four insertions)
Vibrio polysaccharide synthesis	VC0934	Polysaccharide biosynthesis glycosyltransferase	↓
Flagellar motility	Between VC2119 and VC2120	Hypothetical protein and flagellar biosynthetic protein FhlB	↑
	VC2133	Flagellar ring protein FliF	↑
	VC2135	Flagellar regulatory protein C FlrC	↑
	VC2190	Flagellar hook-associated protein FlgL	↓
	VC0892	Chemotaxis protein PomA	↓↓↓ and ↓↓
Transport	VC2705	Sodium-solute symporter, putative	↓ (two insertions)
	VC0724	Phosphate ABC transporter, permease protein, PstC-1	↓
Generation of electrochemical potential	VC2292	NADH:ubiquinone oxidoreductase, Na translocating, hydrophobic membrane protein NqrD	↓
	VC2469	l-Aspartate oxidase (nadB) (NAD biosynthesis)	↓
	VC2422	Nicotinate-nucleotide pyrophosphorylase, carboxylation (nadC) (NAD biosynthesis)	↓ (four insertions)
	VC0095	Chorismate-pyruvate lyase (ubiC) (ubiquinone biosynthesis)	↓ (three insertions)
Disulfide bond formation	VC2701	Thiol:disulfide interchange protein (DsbD)	↓↓ (four insertions)
	VC0034	Thiol:disulfide interchange protein (TcpG)	↓↓↓
Regulatory	s074	Response regulator from SXT transposonlike element	↓↓↓
	s075	Putative histidine kinase from SXT transposonlike element	↑
Amino acid biosynthesis/nitrogen metabolism	VC2438	Glutamate-ammonia ligase adenylyltransferase (glnE) (glutamine biosynthesis)	↓
Transcription/translation	VC1179	Pseudouridine synthase family 1 protein	↓
	VC0369	RplL ribosomal protein L9	↓
	VC2030	RNase E	↓
O-antigen/capsule	AAC46248	O139 lipopolysaccharide synthesis galactosyl transferase	Aggregative (three insertions)
	CAA69111	O139 lipopolysaccharide synthesis oxidoreductase	Aggregative
	BAA33602	Gene for O-antigen synthesis (wbfN)	Aggregative
Hypothetical	VC1079	Conserved hypothetical protein	↑
	VC0302	Conserved hypothetical protein	↑
	VC1637	Hypothetical protein	↓ (two insertions)
	VC2232	Hypothetical protein	↓ (two insertions)

^aSee Fig. 2.

A large number of the mutations that we identified affected synthesis of the MSHA pilus, the O-antigen, and O-antigen capsular polysaccharide. As previously shown (37), attachment was completely abolished by most mutations affecting MSHA pilus structure and function (Fig. 2). Furthermore, mutants harboring transposon insertions in genes encoding O-antigen synthesis proteins formed surface-attached bacterial aggregates that met the definition of a multilayer biofilm. In V. cholerae O139, the O-antigen is a component of both the lipopolysaccharide in the bacterial outer membrane and the capsular polysaccharide (62). We previously documented that the O-antigen and O-antigen capsule are able to mediate intercellular interactions when negative charges are neutralized by Ca²⁺ (26). The present findings suggest that mutation of the O-antigen also enables intercellular interactions. We hypothesized that in the absence of intercellular repulsion mediated by the V. cholerae O-antigen, V. cholerae forms a multilayer biofilm rather than a monolayer biofilm.

A number of mutants identified in our screen had transposon insertions in genes encoding proteins with a role in generation or utilization of the V. cholerae ion motive force (Table 3). The ion motive force maintained by V. cholerae is comprised of the sodium motive force and the proton motive force. Collectively, the ion motive force has three components, (i) ΔpH, which reflects the proton gradient maintained across the inner membrane; (ii) ΔpNa, which reflects the sodium gradient maintained across the inner membrane; and (iii) ΔΨ, which reflects the charge difference maintained across the inner membrane. The sodium and proton motive forces both contribute to ΔΨ. The proteins identified in our screen that play a role in generation of the ion motive force included subunits of Na⁺-NQR, NAD⁺ synthesis enzymes, ubiquinone synthesis enzymes, and DsbD. The proteins identified in our screen that utilize the ion motive force included subunits of the Na⁺-powered flagellar motor (4, 10, 19) and possibly the protein encoded at locus VC2705, which is annotated as a sodium-solute symporter. The experiments described below focused on the roles of these two groups of proteins in formation of the monolayer biofilm.

Evidence that Na⁺-NQR impacts monolayer formation.

Na⁺-NQR, the primary Na⁺ pump in V. cholerae, uses the energy released as a result of electron transfer from NADH to ubiquinone to power the translocation of Na⁺ across the cell membrane, generating both ΔpNa and ΔΨ. Our primary monolayer screen identified two monolayer-defective mutants harboring insertions in the nqrB and nqrD genes, which encode two of the six subunits of Na⁺-NQR. Our nqrD::TnMar mutant grew as well as wild-type V. cholerae, while the nqrB::TnMar mutant grew at a lower rate and was therefore not included in further study. The monolayer defect of the nqrD::TnMar mutant was reproducible but small (Fig. 3A).

FIG. 3.

Monolayer formation by wild-type V. cholerae and mutants harboring transposon insertions in the NADH:ubiquinone oxidoreductase gene nqrD, the NAD⁺ synthesis genes nadB and nadC, and the ubiquinone biosynthesis gene ubiC. Cultures were incubated overnight in LB broth and then used to inoculate MM at a starting OD₆₅₅ of 0.0005. (A) Quantification of monolayers formed by wild-type V. cholerae, an nqrD::TnMar mutant, a nadB mutant, and a nadC mutant in the presence or absence (−) of 1 μg/ml nicotinic acid or nicotinamide. (B) Quantification of monolayers formed by wild-type V. cholerae, an nqrD::TnMar mutant, and a ubiC mutant in the presence or absence (−) of 1 μg/ml 4-hydroxybenzoate. *, P < 0.0001 for a comparison with wild-type V. cholerae monolayers examined under similar growth conditions. WT, wild type.

NADH is an important cofactor in monolayer formation.

In our screen, we identified monolayer-deficient mutants carrying insertions in the genes encoding the V. cholerae NadB and NadC homologs. NadB (l-aspartate oxidase) and NadC (quinolinic acid phosphoribosyltransferase) catalyze the first and third steps in the pathway for de novo synthesis of NAD⁺. NADH is an essential cofactor in myriad oxidation-reduction reactions throughout the cell, including the reaction involving Na⁺-NQR. The nadB and nadC mutants displayed a monolayer defect similar to that of the nqrD mutant (Fig. 3A), suggesting that the primary role of their nadB and nadC gene products in monolayer formation is provision of a cofactor for Na⁺-NQR.

In E. coli, NAD⁺ can be synthesized de novo from l-aspartate, or it can be scavenged from environmental sources, such as nicotinamide and nicotinic acid. We reasoned that if a defect in NAD⁺ synthesis were responsible for the observed defect in monolayer formation by the nadB and nadC mutants, then the monolayer-deficient phenotype of these mutants would be rescued by addition of nicotinic acid or nicotinamide to the culture medium. Because the nqrD mutant is unable to use NADH to form an Na⁺ gradient, supplementation of the growth medium with nicotinic acid or nicotinamide should not rescue the monolayer phenotype of this mutant. Indeed, as shown in Fig. 3A, supplementation of the growth medium with nicotinic acid or nicotinamide rescued the monolayer defect of the nadB and nadC mutants but had no effect on the monolayer defect of the nqrD mutant. These findings support the hypothesis that a paucity of NADH is responsible for the defect of the nadB and nadC mutants in monolayer formation.

Role of ubiquinone in monolayer formation.

In our screen, we identified several monolayer-defective mutants carrying transposon insertions in ubiC. In E. coli, UbiC (chorismate-pyruvate lyase) catalyzes the first committed step in the ubiquinone biosynthesis pathway, which converts chorismate into pyruvate and 4-hydroxybenzoate (29). ubiC mutants had no observable growth defect in MM but formed a monolayer that was less confluent than that of wild-type V. cholerae (Fig. 3B). Quinones are electron carriers that are essential for electron transport through respiratory chains. Ubiquinone is a lipid-soluble, diffusible electron carrier that shuttles electrons and protons between sites in electron transport proteins located near the cytoplasmic and periplasmic face of the inner membrane, leading to generation of ΔpH and ΔΨ (3, 46). Ubiquinone receives electrons from a number of enzymes, including Na⁺-NQR. Because the monolayer formed by the ubiC mutant was much less dense than that formed by the nqrD mutant, we hypothesized that, in addition to Na⁺-NQR, ubiquinone may receive electrons from other enzymes that are operative in monolayer formation.

When minimal medium is supplemented with 4-hydroxybenzoate, the requirement for UbiC for the synthesis of ubiquinone is bypassed. We hypothesized that if the monolayer phenotype of the ubiC mutant were due to a defect in the ubiquinone synthesis pathway, supplementation of the growth medium with 4-hydroxybenzoate should rescue the monolayer defect of this mutant. To test this hypothesis, we compared monolayer formation by the ubiC mutant and monolayer formation by wild-type V. cholerae in the presence and absence of 4-hydroxybenzoate. As shown in Fig. 3B, in the presence of 4-hydroxybenzoate, monolayer formation by the ubiC mutant was indistinguishable from monolayer formation by wild-type V. cholerae. This result confirmed that the monolayer defect of the ubiC mutant was in fact, the result of a defect in ubiquinone synthesis.

Sodium-solute symporters in monolayer formation.

We identified two mutants in our genetic screen with transposon insertions just upstream of and directly in the gene at locus VC2705, which encodes a putative sodium-solute symporter. We designated this gene sssA (sodium-solute symporter A). SssA is homologous to the Na⁺-proline family of symporters, which includes the E. coli Na⁺-proline symporter PutP (22). Because of its similarity to sodium-proline symporters, we hypothesized that SssA dissipates both ΔpNa and ΔΨ when it is transporting its solute.

The SssA gene is in a putative operon with the gene at locus VC2704, which encodes a hypothetical 88-amino-acid protein with two predicted transmembrane helices. Homologs of this protein are found in a variety of bacterial species and are encoded by genes that are typically located adjacent to a gene encoding a sodium-solute symporter. We hypothesized that SssA might work in tandem with the VC2704-encoded protein to modulate monolayer formation and therefore designated the latter protein SssH (sodium solute symporter helper).

To study the role of SssH and SssA in monolayer formation, we first constructed strains carrying in-frame deletions in the corresponding genes and tested their abilities to form monolayers. As shown in Fig. 4A, monolayers formed by ΔsssA and ΔsssH mutants appeared to be similar to monolayers formed by wild-type V. cholerae after they were washed with PBS, but a greater proportion of the ΔsssH and ΔsssA mutant monolayer cells detached following washing in PBS/AMM. This suggests that fewer of the ΔsssH and ΔsssA mutant cells than of the wild-type V. cholerae cells had undergone the transition to permanent attachment. Thus, SssA and SssH play a role in the transition from transient to permanent attachment.

FIG. 4.

Monolayer formation by wild-type V. cholerae and sodium-solute symporter mutants. Cells cultured in LB broth overnight were used to inoculate MM at a starting OD₆₅₅ of 0.0005. (A) Quantification of monolayers formed by wild-type V. cholerae, as well as a number of sodium-solute symporter mutants. Monolayers were rinsed with either PBS or PBS/AMM. *, P < 0.0001 for a comparison with wild-type V. cholerae monolayers. (B) Quantification of monolayers formed by wild-type V. cholerae and sodium-solute symporter mutants in the presence and absence of 0.1 mM l-proline. WT, wild type.

The V. cholerae genome encodes three Na⁺-solute symporters, including SssA, the V. cholerae PutP homolog (VCA1071), and the VCA0667 protein. Our previous studies suggested that PutP and OpuD, a glycine betaine transporter encoded by VC1279, are the major transporters of proline in V. cholerae (23). Structure-function studies of E. coli PutP have demonstrated that amino acid residues D55, S340, and T341 play a role in Na⁺ binding, while S57 plays a role in proline binding (21, 48, 51). S340, T341, and S57 are present in the three V. cholerae Na⁺-solute symporters. However, D55 is displaced in SssA and is not present in the VCA0667 protein.

Based on the homology of SssA to E. coli PutP, we hypothesized that other proline transporters might be important for monolayer formation. Therefore, we constructed strains carrying in-frame deletions in the genes encoding each of these proteins, as well as OpuD, and assessed the abilities of these strains to form a monolayer (Fig. 4A). As observed for the ΔsssA mutant, the ΔputP mutant was defective for monolayer formation. However, no additive effect was observed when a strain having both putP and sssA deletions was tested, suggesting that the proteins encoded by these genes function in the same signaling pathway or attachment mechanism. The ΔopuD and ΔVCA0667 mutants had no defect in monolayer formation. We questioned whether the function of VCA0667 and OpuD in monolayer formation might be uncovered in the background of a ΔputP mutation. However, even in the background of putP and sssA deletions, VCA0667 and OpuD played no role in monolayer formation (Fig. 4A and data not shown).

Based on the homology of SssA to V. cholerae PutP and evidence that the latter protein transports proline, we considered the possibility that proline might be an environmental signal that enhances the transition from transient to permanent attachment. Because our minimal medium did not contain proline, we hypothesized that supplementation of this medium with proline might augment monolayer formation. To test this hypothesis, we compared monolayer formation by wild-type V. cholerae and Na⁺-solute symporter mutants in the presence and absence of l-proline. As shown in Fig. 4B, the presence of 0.1 mM l-proline in the medium had no effect on monolayer formation by wild-type V. cholerae or the ΔsssH, ΔsssA, ΔputP, or ΔopuD mutant. This suggested that proline transport is not the critical function of the proteins encoded by sssH, sssA, putP, and opuD during monolayer formation.

The V. cholerae flagellar motor but not flagellar motility is essential for monolayer formation.

In our screen, we identified two classes of flagellar motility mutants in which monolayer formation was altered in very different ways. While mutations in the flagellar structure generally produced a denser monolayer, mutations in the flagellar motor drastically decreased monolayer formation. This suggested that the flagellar motor might play an important role in monolayer formation.

Flagellar motility in V. cholerae is dependent on the sodium motive force, which powers the Na⁺-dependent flagellar motor responsible for turning the helical flagellar filament. Mutation of the motXY and pom AB genes, which encode components of the flagellar motor of vibrios, results in a paralyzed flagellum (9, 31, 32, 40, 41, 54, 56, 66, 67). Mutation of flaA, encoding the building blocks of the flagellar filament, results in a bacterium with no flagellar filament. To further characterize the roles of the flagellar filament and flagellar motor in monolayer formation, we constructed mutants having in-frame deletions of flaA or motX or both. We then studied the abilities of these mutants to form a monolayer. As shown in Fig. 5A, we observed that surface attachment by a ΔflaA mutant was robust, while surface attachment by the ΔmotX and ΔmotX ΔflaA mutants was severely impaired. As shown in Fig. 5B, the monolayer defect of the ΔmotX mutant was completely rescued by providing the motX gene in trans. Partial rescue of monolayer formation was also observed for the ΔflaA ΔmotX mutant. This suggests that the flagellar motor plays a critical role in monolayer formation that is distinct from the role played by the flagellar filament.

FIG. 5.

Monolayer formation by wild-type V. cholerae (MO10) and ΔflaA, ΔmotX, and ΔflaA ΔmotX mutants and rescue with the pBAD-TOPO-motX plasmid. Cells from an overnight culture in LB broth were used to inoculate MM at a starting OD₆₅₅ of 0.0005. After incubation at 27°C for the indicated times, monolayers were washed three times with PBS. (A) Quantification of monolayer formation by wild-type V. cholerae and ΔflaA, ΔmotX, and ΔflaA ΔmotX mutants. Monolayers were quantified 24 h after exposure to a surface. *, P < 0.0001 for a comparison with wild-type V. cholerae monolayers. (B) Quantification of monolayer formation by wild-type V. cholerae and ΔmotX and ΔflaA ΔmotX mutants rescued either with the control plasmid pBAD-TOPO/lacZ/V5-His (placZ) or with the rescue plasmid pBAD-TOPO-motX carrying a wild-type copy of the motX gene (pmotX). Monolayers were quantified 18 h after exposure to a surface. (C) Quantification of monolayer formation by wild-type V. cholerae and a ΔflaA mutant at various times after exposure to a surface. WT, wild type; FLA, ΔflaA mutant.

Flagellar motility is believed to enable bacteria to approach surfaces closely enough to initiate transient attachment. To evaluate the role of flagellar motility early in the course of monolayer formation, we compared total surface attachment by wild-type V. cholerae and total surface attachment by a ΔflaA mutant over time (Fig. 5C). We observed that the surface attachment by a ΔflaA mutant lagged behind that of wild-type V. cholerae at early time points. After 10 h of exposure to the surface, however, a PBS-rinsed ΔflaA mutant monolayer had a level of surface coverage that was equivalent to that of wild-type V. cholerae. This suggests that although surface attachment of ΔflaA mutant cells occurs more slowly than surface attachment of wild-type V. cholerae, over time the numbers of wild-type and mutant surface-attached cells are equal.

We concluded that flagellar motility accelerates early interactions with the surface but is not essential for transient attachment. Furthermore, we hypothesized that, independent of its role in motility, the flagellar motor plays an important role in the transition from transient to permanent attachment. We propose that this role is to modulate the ion motive force.

Chemical analysis of the role of ΔpH and ΔpNa in monolayer formation.

In the experiments described above, we obtained genetic evidence that proteins involved in generation and utilization of the V. cholerae ion motive force play a role in the transition from transient to permanent attachment. However, because so many proteins and pathways contribute to generation and utilization of the ion motive force, the defect in monolayer formation resulting from any single mutation is likely to be small. To circumvent the issue of functional redundancy, we used a chemical approach. We first examined the effects of HQNO and CCCP on monolayer formation. HQNO is a specific inhibitor of Na⁺-NQR (20, 39) and therefore blocks generation of ΔpNa and ΔΨ by Na⁺-NQR. CCCP is a weak lipophilic acid that crosses the inner membrane as CCCPH, deposits a proton in the cytoplasm, and returns to the periplasm in response to the ΔΨ, thereby completely dissipating the proton motive force, which includes ΔpH and a portion of ΔΨ.

We first confirmed that these chemicals had the expected effects on ΔΨ. ΔΨ is most conveniently measured by use of voltage-sensitive dyes. Voltage-sensitive dyes are lipophilic, cationic fluorescent dyes that are attracted to and concentrate in the negatively charged cytoplasm of polarized cells. Aggregation of such a dye in the cytoplasm alters the fluorescence of the dye. We used 3,3′-diethyloxacarbocyanine iodide, which fluoresces at 535 nm in the nonaggregated state, to compare the ΔΨ of untreated wild-type V. cholerae and V. cholerae treated with CCCP, HQNO, or both CCCP and HQNO. The resulting data were expressed as the difference between the fluorescence emitted by the dye in the presence of the chemical(s) and the fluorescence emitted by the dye in the absence of the chemical(s). Therefore, a more positive measurement indicated a greater degree of membrane depolarization. As shown in Fig. 6A, when added individually, CCCP and HQNO moderately reduced ΔΨ, while addition of both compounds had an additive effect on ΔΨ.

FIG. 6.

Effects of CCCP and HQNO on V. cholerae ΔΨ and monolayer formation. (A) Quantification of wild-type V. cholerae ΔΨ after addition of 0.5 μM CCCP, 50 μM HQNO, or both 0.5 μM CCCP and 50 μM HQNO. An increase in ΔFluorescence indicates a decrease in membrane polarization. *, P < 0.009 for a comparison with untreated monolayers. (B) Quantification of monolayers formed over 1 h by wild-type V. cholerae in the presence or absence of 0.5 μM CCCP, 50 μM HQNO, or both 0.5 μM CCCP and 50 μM HQNO. *, P < 0.0005 for a comparison with untreated wild-type V. cholerae monolayers.

We then quantified wild-type V. cholerae monolayer formation in the presence of CCCP and HQNO (Fig. 6B). In these experiments, the time of incubation with the chemicals was short to avoid reductions in viability. Incubation with CCCP alone had only a modest effect on wild-type V. cholerae monolayer formation. A slight decrease in the swimming speed was also noted. From these observations, we concluded (i) that ΔpH is not essential for monolayer formation and (ii) that V. cholerae is able to maintain ΔpNa in the absence of ΔpH. Incubation with HQNO also had a modest effect on monolayer formation and caused a slight decrease in the swimming speed. Because motility was preserved in the presence of HQNO, we concluded that HQNO does not fully dissipate ΔpNa. Two possible explanations for this are (i) that another Na⁺ transporter contributes to generation of ΔpNa in V. cholerae and (ii) that V. cholerae is able to convert a proton motive force into a sodium motive force. To distinguish between these two possibilities, we measured motility and monolayer formation in the presence of both CCCP and HQNO. We reasoned that if V. cholerae maintained ΔpNa in the presence of HQNO by converting a proton motive force into a sodium motive force, addition of CCCP would decrease flagellar motility and monolayer formation. In contrast, if another Na⁺ transporter contributed to ΔpNa in the presence of HQNO, addition of CCCP would not have this effect on flagellar motility and monolayer formation. In fact, we observed that incubation with HQNO and CCCP greatly decreased monolayer formation and abolished flagellar motility. From these observations, we concluded that V. cholerae maintains ΔpNa in the presence of HQNO by converting a proton motive force into a sodium motive force. As a result, using this approach, we were unable to determine the role of ΔpNa in monolayer formation independent of ΔpH and ΔΨ. Instead, we elected to look directly at the role of ΔΨ in monolayer formation.

ΔΨ blocks the transition from transient to permanent attachment.

To evaluate the role of ΔΨ, we examined the effect of valinomycin on monolayer formation. Valinomycin is predicted to abolish ΔΨ but not ΔpNa or ΔpH (1). We first measured the effect of valinomycin on ΔΨ as described above. We found that addition of valinomycin depolarized the membrane in a concentration-dependent manner (Fig. 7A). Furthermore, motility and transient attachment were not affected by the presence of valinomycin (data not shown). In spite of this, valinomycin blocked permanent attachment by wild-type V. cholerae in a concentration-dependent manner (Fig. 7B). We then asked whether ΔΨ might be required not only for the transition from transient to permanent attachment but also for maintenance of permanent attachment. To test this, we incubated preformed monolayers with valinomycin in PBS or PBS/AMM. As shown in Fig. 7C, addition of valinomycin to preformed monolayers did not result in dissolution of the monolayer. Therefore, we concluded that ΔΨ is required for the transition from transient to permanent attachment. Furthermore, ΔΨ is not required for maintenance of the monolayer biofilm.

FIG. 7.

Effect of valinomycin on wild-type V. cholerae ΔΨ and monolayer formation. (A) ΔΨ measurements for wild-type V. cholerae cells in the presence of various concentrations of valinomycin. An increase in ΔFluorescence reflects a decrease in ΔΨ. *, P < 0.006 for a comparison with untreated wild-type V. cholerae monolayers. (B) Quantification of wild-type V. cholerae monolayers formed over 1 h in the presence and absence of various concentrations of valinomycin. *, P < 0.0001 for a comparison with untreated wild-type V. cholerae monolayers. (C) Quantification of wild-type V. cholerae monolayers formed over 8 h and then washed with MM or MM/AMM in the presence and absence of 10 μM valinomycin.

In previous work, we showed that transcription of flagellar genes was repressed in a monolayer biofilm and transcription of sssA was activated (37, 38). To test whether modulation of ΔΨ by valinomycin might directly affect transcription of these monolayer-specific genes, we measured the transcript levels of flaA and sssA using qRT-PCR in the presence and absence of valinomycin. However, addition of valinomycin did not have a significant effect on the transcription of these two monolayer-specific genes, suggesting that the transcriptional program of monolayer-associated cells is not directly linked to modulation of ΔΨ (data not shown).

DISCUSSION

In this paper, we report the results of a screen for V. cholerae genes encoding proteins that play a role in formation of the monolayer biofilm. A large number of genes identified in this screen encode proteins that are involved in generation and dissipation of the ion motive force of V. cholerae. In particular, mutations in subunits of the flagellar motor substantially decreased monolayer formation. We used chemicals to further dissect the effect of the ion motive force on monolayer formation. These studies demonstrate that ΔΨ is required for monolayer formation but not monolayer maintenance and are consistent with a model in which modulation of ΔΨ initiates the transition to permanent attachment.

The concept that a transient increase in ΔΨ is a bacterial signal has a precedent in microbiology. Transient increases in ΔΨ following exposure to a chemoattractant, such as glucose, were documented almost 30 years ago (12-14, 35, 36). This was postulated to be part of a bacterial signal transduction mechanism that involved the chemotactic apparatus but was not essential for chemotaxis. More recently, ΔΨ has been implicated in regulation of the cidABC and lrgAB operons of Staphylococcus aureus, which control autolysis (47). Thus, we propose that control of bacterial signal transduction by ΔΨ may be a widespread and underappreciated phenomenon.

A possible model connecting the flagellar motor to modulation of ΔΨ during surface attachment is shown in Fig. 8. In planktonic and monolayer-associated cells, the flow of ions across the inner membrane is in steady state, and ΔΨ is maintained at a constant level. When V. cholerae approaches a surface, it may attach by a single tether, such as the flagellum. In this case, the bacterium spins in place. However, if it attaches by more than one tether, such as the flagellum and the cell body, the flagellar motor stops, the flow of ions through the motor ceases, and ΔΨ is transiently increased. This transient increase in ΔΨ indicates to the cell that it is on a surface and initiates the transition to permanent attachment. Our experimental results support this model since (i) cells that are defective in generation of ΔΨ, such as the cells that we isolated in our screen, also display reduced monolayer formation; (ii) cells with no flagellar motor are severely impaired in monolayer formation; and (iii) zeroing of ΔΨ by valinomycin completely blocks the transition to permanent attachment without inhibiting flagellar motility or transient attachment. In each of these cases, hyperpolarization of the cell membrane due to flagellar motor arrest should occur slowly or not at all. Recently, EpsE, a protein described as a molecular clutch, has been identified in B. subtilis. EpsE is proposed to disengage the flagellar rotor from its motor in biofilm-associated cells (2). It is interesting to speculate that engagement of a clutch might be triggered by a transient increase in ΔΨ. However, a detailed evaluation of this model requires development of reagents that allow real-time measurement of ΔΨ in single cells. Furthermore, we cannot rule out the possibility that the rotational speed of the flagellar motor regulates transcription of an as-yet-unidentified adhesin.

FIG. 8.

Model for the role of the flagellar motor and ΔΨ in the transition from transient to permanent attachment. In the planktonic state, the flagellar motor is active, and the ΔΨ is maintained at a constant level. When the cell attaches to the surface by more than one tether, such as the flagellum and the cell body, the flagellar motor stops, the flow of ions through the motor ceases, and ΔΨ is transiently increased. This initiates the transition to permanent attachment.

While similar studies of multilayer biofilm formation have identified very few proteins that are intestinal colonization factors in the infant mouse model, many of the proteins that we identified in our monolayer screen have previously been identified as colonization factors in the infant mouse intestine. Similar to the results we present here, the flagellum was not found to be necessary for colonization of the infant mouse intestine, while a flagellar motor mutant had a 10-fold decrease in colonization compared to wild-type V. cholerae (28, 64). In a separate study, genes encoding Na⁺-NQR as well as UbiC were identified in a signature-tagged mutagenesis screen targeting factors required for colonization of the infant mouse intestine (34). Lastly, a screen for genes that are preferentially expressed in the infant mouse intestine uncovered SssA (42). In competition with wild-type V. cholerae, an SssA mutant was found to have a 10-fold disadvantage in colonization. These findings suggest that the monolayer biofilm studied in this work provides a useful in vitro model for biofilm formation in the mammalian intestine.