A BolA-like morphogene from the alga Chlamydomonas reinhardtii changes morphology and induces biofilm formation in Escherichia coli


Correspondence: Jacinta S. D'Souza, UM-DAE Centre for Excellence in Basic Sciences, University of Mumbai, Kalina Campus, Santacruz (E), Mumbai 400 098, India. Tel.: +91 22 26524978; fax: +91 22 26524982; e-mail: jacinta@cbs.ac.in


Escherichia coli BolA protein is a stress-inducible morphogene, regulates transcription, forms biofilms and interacts with monothiol glutaredoxins. Its presence has been documented in plants but its role remains enigmatic. This study attempts to functionally dissect the role of a BolA-domain-containing protein in the alga Chlamydomonas reinhardtii. Of the five C. reinhardtii bolA-like genes annotated for the presence of BolA-domain, the open reading frame with the highest similarity to algal systems was cloned and the protein over-expressed in E. coli. This over-expression did not affect E. coli growth but induced biofilm formation and changed its morphology, indicating functional conservancy. This is the first compelling evidence depicting the role of a plant BolA-like protein in morphogenetic pathway and biofilm formation. The implications of the phenotypic consequences of this heterologous expression are discussed.


Cellular stress can induce substantial physiological and molecular adaptations to ensure survival. Depending on the stress dose and length of exposure, the flagella of the unicellular chlorophyte Chlamydomonas reinhardtii are paralyzed or lost (Hessen et al., 1995; Dharmadhikari et al., 2006), and later cells manifest ‘palmelloids’ (Jamers & De Coen, 2010 and references therein) or apoptose (Moharikar et al., 2006, 2007; Yordanova et al., 2010). Palmelloids are a stress-responsive temporary, dormant ‘colonial’ stage with characteristic physiological changes such as exopolysaccharide secretion, clustering of cells embedded in an exopolysaccharide mesh and a common membrane, individual cell wall thickening, abnormal cell division, change in cell morphology and decrease in cell viability (Nakamura et al., 1976; Lewin, 1984; Visviki & Santikul, 2000; Jamers & De Coen, 2010). Microarray analysis of paraquat-treated Creinhardtii palmelloids showed differential regulation of genes (Jamers & De Coen, 2010) indicating alterations in both cell phenotype and transcriptome, facilitating cells to devise strategies for survival or death. Although the molecular mechanisms that regulate palmelloid formation in Chlamydomonas species remain elusive, some of the physiological hallmarks are conserved. Like most other eukaryotes, C. reinhardtii harbors conserved stress-responsive gene families, making it a suitable system to study the molecular mechanisms that underlie palmelloid formation. Akin to palmelloidy, biofilms are defined as a structured community of bacterial cells enclosed in a self-produced polymeric matrix that adhere to a living or inert surface (Watnick & Kolter, 2000) and exhibit altered phenotype, changing growth rate and gene transcription. Parallels between biofilms and palmelloids are: (1) secretion of high levels of exopolysaccharides (Kobayashi, 2007; Shemesh et al., 2010; Yoon et al., 2011); (2) alterations in cell morphology (Kobayashi, 2007; Yoon et al., 2011); (3) effects on cell wall biogenesis (Shemesh et al., 2010; Daher et al., 2011); (4) relations with flagella (Prigent-Combaret et al., 2000; Klausen et al., 2003; Lemon et al., 2007); and (5) stress induction (Rode et al., 2007; Kaplan, 2011). Microarray analysis to study biofilm formation in Escherichia coli has reported ∼ 150 differentially expressed genes (Niba et al., 2007). Furthermore, biofilm formation in E. coli at 23 °C was regulated by adrA, nhaR, mlrA, csgA and bolA genes (White-Ziegler et al., 2008). A search for homologs in Creinhardtii resulted in five putative bolA-like genes. In E. coli, the BolA protein was first studied as a morphogene that could promote round morphology in cells when over-expressed (Aldea et al., 1988). The gene was found to be regulated by growth phase-dependent promoter P1, induced during the transition to stationary phase of growth. It was found to be under the control of RpoS (Hengge-Aronis et al., 1993). Exposure to several stresses up-regulated bolA gene even in the early logarithmic phase (Santos et al., 1999), suggesting an important role in stress response. As a transcriptional regulator, it was found to independently regulate the transcription of dacA and dacC carboxypeptidases, a β-lactamase ampC (Santos et al., 2002; Guinote et al., 2011) and the actin-like gene mreB (Freire et al., 2009); genes involved in different activities of cell morphology and division. The 3-D structures of BolA protein from Mus musculus and Xanthomonas campestris pv. campestris confirmed the presence of the DNA-binding Class II KH fold, supporting a regulatory role (Kasai et al., 2004; Chin et al., 2005). BolA protein over-expression in E. coli facilitated biofilm formation in a stress-responsive manner (Vieira et al., 2004). In Shizosaccharomyces pombe, uvi31+, a bolA homolog was up-regulated under UV light and an involvement in the regulation of cell septation and cytokinesis was suggested (Kim et al., 1997, 2002). Although in silico searches of three human homologs (HsBolA) showed no signal peptide for secretion, experimental analysis showed secretion of the proteins from Cos-7 cells (Zhou et al., 2008). The widespread occurrence and conservation of the bolA gene family across different genera prompted us to explore the function of an as yet elusive algal bolA-like gene. The current study attempts to functionally dissect the role of a C. reinhardtii BolA-like protein by cloning the gene (CrbolA), over-express the protein (CrBolA) in E. coli, and study the possible effect on morphology and biofilm formation.

Materials and methods

Reagents and media components were obtained from Amresco, SRL (India) and Merck (India). Bacterial strains, plasmids, primers details and websites for in silico work are outlined in Table 1.

Table 1. Plasmids, strains and primers used in this study
Plasmids, strains, primers, websitesCharacteristicsSource/reference
pQE30UA3.5 kb cloning and expression vector. AmprQiagen
pJExpress2042.8 kb cloning and expression vector. AmprDNA 2.0
pCBS-3pJExpress204 harboring 282 bp fap174This study
pCBS-4apQE30UA harboring 489 bp CrbolAThis study
pCBS-4bpJExpress204 harboring 489 bp CrbolAThis study
XL1bluerecA1endA1gyrA96thi1hsdR17supE44relA1lac [F′proAB,lacIqZDM15,Tn10(tetR)]Stratagene
MG1693 (wild-type) thyA715 Santos et al. (1999)
bolA mutantMG1693ΔbolA2::KanrFreire et al. (2009)
XL1bluepQENTXL1blue + pQE30UAThis study
CBS-3XL1blue + pCBS-3, over-expressing (His)6CrFAP174This study
CBS-4aXL1blue + pCBS-4a, over-expressing (His)6CrBolAThis study
CBS-4bXL1blue + pCBS-4b, over-expressing (His)6CrBolAThis study
CBS-4cMG1693(WT) + pQE30UAThis study
CBS-4dMG1693(WT) + pCBS-4aThis study
CBS-4eMG1693ΔbolA2::Kanr + pQE30UAThis study
CBS-4fMG1693ΔbolA2::Kanr + pCBS-4aThis study
CMA50BL21(DE3) over-expressing (His)6EcBolAFreire et al. (2009)
Primers (for cloning)
Websites for in silico work
Expasy Protparam http://www.expasy.ch/tools/protparam.html Gasteiger et al. (2005)
oligocalc software http://www.basic.northwestern.edu/biotools/oligocalc.html Kibbe (2007)
NCBI-Special BLAST-Conserved Domains http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi Marchler-Bauer et al. (2011)
ncbi-blastp http://blast.ncbi.nlm.nih.gov/Blast.cgi Altschul et al. (1990)
clustalw2 http://www.ebi.ac.uk/Tools/msa/clustalw2 Larkin et al. (2007)
TargetP http://www.cbs.dtu.dk/services/TargetP/ Emanuelsson et al. (2000)

Media and growth conditions

Compositions of Luria–Bertani (LB) and M9 media followed standard protocols. Whenever necessary, these media were supplemented with thymine (50 μg mL−1), ampicillin (100 μg mL−1) and kanamycin (50 μg mL−1). Cells were incubated at 37 °C for all experiments.

In silico analyses

Gene information and protein parameters were generated from the raw nucleotide and amino acid sequences using expasy protparam and oligocalc software. The BolA domain on the protein sequence was defined using NCBI-Special BLAST-Conserved Domains. The protein sequence was subjected to NCBI-BLASTP and homologs from other organisms selected using default settings (see Supporting Information, Table S2). These were subjected to clustalw2 for multiple alignments and phylogenetic analysis using the maximum likelihood algorithm. The in silico sub-cellular localization was predicted using targetp.

Cloning and over-expression of CrBolA in E. coli

In silico similarity analyses of Ecoli bolA gene with C. reinhardtii genome (www.chlamy.org) revealed five putative bolA-like genes annotated as BolA-like proteins since they harbor a BolA-domain in their primary sequence, and they have been referred to in this study as CrBolA-1 of 108 aa C_1730026, CrBolA-2 of 99 aa C_1350011, CrBolA-3 of 100 aa C_2020005, CrBolA-4 of 162 aa C_330042 and CrBolA-5 of 345 aa C_90207. The 489-bp C. reinhardtii bolA-like ORF was cloned in pJExpress204 (Dongre et al., 2011), sub-cloned into resultant plasmids (pCBS-4a and pCBS-4b), and positive clones as confirmed by PCR and nucleotide sequencing were grown to an optical density (OD)600 nm of 0.6, they were induced with 0.5 mM IPTG for 3 h and over-expression was confirmed by SDS-PAGE.

Morphology assay

Escherichia coli was grown in 5 mL of LB at 37 °C from overnight seed cultures. Induced (0.5 mM IPTG) and uninduced cultures were harvested at 1, 2 and 18 h post-induction, cell pellets washed in 1× phosphate-buffered saline (PBS, pH 7.0), fixed for 1 h at 4 °C in 1× PBS containing 0.75% formaldehyde, washed, re-suspended in 1× PBS and analyzed on a Nikon 90i microscope using phase contrast objective at 100× magnification.

Microtiter plate biofilm assay

Escherichia coli master cultures were grown until the logarithmic phase when OD600 reached 0.6, after which cells were harvested, and pellets washed twice and re-suspended in M9 medium. IPTG at a concentration of 0.5 mM was added to the cultures harboring plasmids pCBS-4a and pCBS-4b. Cell suspensions of 200 μL were transferred to polystyrene microtiter plate wells in triplicates (Laxbro Bio-Medical Aids Pvt. Ltd, India). Plates were incubated at 37 °C for 24 h with gentle shaking; three wells of medium without culture served as controls. Post-incubation, the medium was decanted, wells washed three times with double distilled water (DDW, to remove any unattached bacteria) and air-dried for 30 min. Biofilms were stained for 30 min with 1% crystal violet made in 10% ethanol, excess stain discarded, washed three times with DDW and air-dried. 200 μL ethanol was added to each well and the color estimated by measuring the OD595 nm.

Environmental Scanning Electron Microscopy (ESEM) study for biofilm formation

Cells were cultured on grease-free, sterile glass cover-slips (triplicates) placed in sterile petri-plates with M9 medium and incubated at 37 °C for 24 h with gentle shaking. The medium was discarded, unattached bacteria removed by washing with DDW, fixed for 4 h at 4 °C with 2.5% glutaraldehyde in 0.1 M HEPES buffer (pH 6.0), rinsed with HEPES buffer and air-dried. ESEM was performed by mounting the cover-slips on aluminum stubs and analyzing on FEI Quanta 200 at 20 kV, under 1.5 Torr using a gaseous state electron detector.

Results and discussion

Serine-rich, evolutionarily divergent CrBolA protein – in silico analyses

The BolA family of proteins is widespread and harbors a conserved BolA domain with a number of genes varying from one to several per genome. As mentioned in 'Materials and methods', in silico similarity analyses of Ecoli bolA gene with C. reinhardtii genome (www.chlamy.org) revealed five putative bolA-like genes (CrBolA-1, CrBolA-2, CrBolA-3, CrBolA-4 and CrBolA-5); their role in stress response and tolerance remains elusive. All have a C-terminal BolA domain but CrBolA-5 additionally harbors a SufE domain at its N-terminus (59th–181st residues). The identity of these protein sequences individually with the EcBolA protein ranges from 24% to 41% (for more details, see Table S1). Since bolA genes in E. coli are associated with stress, it is quite possible that these genes play a similar role in plants; the response being via alteration of morphology, transcriptional regulation of other genes, signal transduction by the BolA-like protein, biofilm formation due to accumulation of polyphosphates or a yet elusive pathway. The present study aims to determine the possible function of bolA-like genes in C. reinhardtii.

The Arabidopsis thaliana genome is thus far the best annotated plant genome. Therefore, as a starting point, we have compared the five C. reinhardtii BolA-like sequences with the three annotated A. thaliana BolA proteins (At1G55805, At5G09830, and At5G17560). A phylogenetic analysis reveals that CrBolA-1, -3 and -5 are orthologous to AtG55805 with an identity ranging from 42% to 53% (Fig. 1a). CrBolA-2, on the other hand, seems to have evolved from a common ancestor in parallel with At5G09830 and they may be considered orthologous to each other with an identity of 73%. CrBolA-4, the protein under current investigation, is orthologous to At G17560 with an identity of 49%. Of all these five CrbolA-like genes, we selected to work with CrbolA-4 since its ORF shared the highest similarity to algal systems (Fig. 1b). In silico analyses of its protein sequence exhibited a theoretical Mr and pI of 17.34 kDa and 8.41, respectively, the protein being serine-rich (∼ 12%). Known BolA proteins have an average Mr of ∼ 14 kDa, although plant BolA-like sequences are ∼ 20 kDa with pIs between 5 and 10. Several BolA protein sequences have secondary structures constituting α helices and β sheets. CrBolA has the propensity of such structures mainly in its BolA domain, which spans the C-terminus from 91 to 161st aa residues (Fig. 1c). Multiple alignments using similar sequences from other organisms (obtained from NCBI-BLASTP searches) revealed the presence of short stretches of highly conserved amino acids in the C-terminus BolA domain of the CrBolA protein (Larkin et al., 2007; Fig. 1c). The protein is more identical to Viridiplantae BolA-like sequences with amino acid level identities between 50% and 65%; and identities with E. coli and human sequences at 38% and 32%, respectively (see Table S2). A phylogenetic tree generated from these sequences showed evolutionary divergence similar to the other plant members of the group (Fig. 1b), suggesting a new diverged class of proteins. We have carried out an in silico study using targetp software (Emanuelsson et al., 2000) for predicting the sub-cellular localization of the five CrBolA homologs. While CrBolA-1 has a signal peptide for the secretory pathway, CrBolA-2 and -3 have been predicted to be cytosolic. CrBolA-4, which is at present under investigation, and the CrSufE BolA-domain containing protein, CrBolA-5, are both predicted to be targeted to the chloroplast; all these predictions have confidence levels of 0.405–0.754. Whereas BolA-like protein from A. thaliana has also been found to be chloroplast localized (Ye et al., 2006), Fra2 from Saccharomyces cerevisiae has been shown to be cytosolic (Kumanovics et al., 2008). The assortment in the sub-cellular localization indicates the multiplicity in function of the CrBolA proteins in C. reinhardtii.

Figure 1.

In silico analyses of Chlamydomonas reinhardtii bolA gene and protein. (a) Phylogenetic analysis showing clustering of five CrBolAs with the respective Arabidopsis thaliana BolAs. (b) Phylogenetic analysis of CrBolA with other algal and plant sequences showing shared lineage. (c) Multiple alignment of the CrBolA protein with 16 BolA and/or BolA-like protein sequences. (*)Positions which have a single, fully conserved residue. Colon (:) indicates conservation between groups of strongly similar properties; period (.) indicates conservation between groups of weakly similar properties. α (2, 3 and 4) and β (2) are the predicted helices and sheets in the protein sequence, respectively.

CrBolA over-expression does not affect E. coli growth

From among five bolA-like genes in C. reinhardtii, the ORF (489 bp) of the one with the highest similarity to algal systems was cloned, transformed into E. coli XL1blue cells, positive clones confirmed by nucleotide sequencing, and the protein over-expressed using IPTG (see Fig. S1). CrBolA over-expressed in E. coli was insoluble; the gene was subsequently sub-cloned in several compatible over-expression vectors and its effect on E. coli growth examined. The effect of EcBolA over-expression on E. coli growth rate has been observed earlier (Guinote et al., 2011), with the demonstration of differential impairment in the growth depending on the origin of the starting batch culture (plate or liquid). The cultures initiated from liquid medium showed more growth rate impairment as compared with those grown from plates. Taking into account these results, we used a starting liquid batch culture of E. coli (∼ 16 h) that reached stationary phase and over-expressed EcBolA (Guinote et al., 2011). In such an inoculum, when used for studying the effect of CrBolA over-expression with IPTG as an inducer, the growth was not affected by two wild-type E. coli strains (XL1blue and MG1693; Fig. 2). It thus appears that whereas over-expression of EcBolA impairs growth of E. coli cells, the same effect is not seen when CrBolA is over-expressed in E. coli. When present in very high quantities, EcBolA interferes with the cell cycle progression; however, CrBolA may not necessarily follow the same pathway or the amount of active protein available for this function may be low. The reduced growth rate of E. coli bolA deletion mutant with the empty plasmid remains to be determined.

Figure 2.

Growth curve analysis of differing Escherichia coli strains with or without the over-expression of CrbolA. (a) CrbolA gene was inserted into two backbone vectors, viz. pQE30UA and pJExpress. These plasmids were individually transformed into E. coli XL1 blue strain and monitored for cell growth. Appropriate controls such as pQE30UA without the CrbolA insert or an irrelevant Chlamydomonas reinhardtii gene, viz. fap174, were used. On similar lines, the (b) wild-type E. coli strain MG1693 and (c) bolA mutant was transformed with CrbolA. No significant differences in the growth rates of test and control samples were observed.

CrBolA protein induces changes in the morphology of E. coli cells

RpoS sigma factor is a regulator of stress-related genes such as bolA (Aldea et al., 1988; Hengge-Aronis et al., 1993). EcBolA independently regulates dacA, dacC carboxypeptidases and mreB in E. coli (Santos et al., 2002; Freire et al., 2009), which in turn are involved in cell wall biosynthesis and maintenance, thus contributing to generate the EcBolA-induced round morphology.

To evaluate the effect of CrBolA over-expression on E. coli cell morphology, the protein was over-expressed in various E. coli strains and the morphology scored. A definitive change in morphology was observed; cells became shorter, wider, and tended to become round. When ∼ 300 cells were scored for their lengths, an average size of 1.4 μm in the CrBolA over-expressed cells vs. 2.4 μm in control cells was observed and representative images are depicted in Fig. 3. When compared with relevant controls (WT, XL1blue, pQE30UA transformed strain, pJExpress with fap174 insert where FAP174 is a flagellar associated protein 174 with no relevance to round morphology and/or biofilms) at 1 h post-induction, the number of cells with an average length of 1.4 μm increased three- to sixfold. It is noteworthy that at 18 h post-induction, CrBolA protein partitions into an insoluble fraction (inclusion bodies); the number of cells with reduced length concomitantly decrease to ∼ nil (Fig. 3a). A similar change in the morphology of cells was observed when CrBolA was over-expressed in WT E. coli MG1693 cells but not in the bolA mutant (Fig. 3b and c). As a positive control, when EcBolA was over-expressed in E. coli BL21(DE3) cells, almost all cells showed a round morphology within an hour (Fig. 3d). However, at 18 h when cells are far beyond the stationary phase, uninduced E. coli cells show round morphology; however, with induction the EcBolA seemed to partition into the insoluble fraction (induced panel of Fig. 3d); such cells do not exhibit a round morphology. As far as imparting a round morphology to Ecoli cells is concerned, it appears that CrBolA does not serve as a complete replacement for this function of EcBolA. However, the CrBolA over-expressed cells do reduce in size and appear shorter than the controls (Fig. 3a and b); implying some morphogenetic role.

Figure 3.

Effect of over-expression of CrBolA on Escherichia coli cell morphology. Escherichia coli cells of both wild-type and bolA mutant were individually transformed with plasmids harboring the CrbolA gene. Induced and uninduced cells were observed at 0, 1 and 18 h. Note the change in morphology at 1 h in CBS-4a (a), -4b (a), -4c (b) and EcBolA (induced, d).

CrBolA over-expression in E. coli induces biofilm formation

In a hostile environment or during erratic cell division, algae form biofilms in the form of palmelloids, independently or with other microorganisms (Flora et al., 1995; Suseela & Toppo, 2007). Palmelloids are a group of tightly bound cells symmetrically arranged and embedded in a jelly-like substance (Visviki & Santikul, 2000 and references therein). Since EcBolA has been implicated in biofilm formation (Vieira et al., 2004), it is highly possible that bolA-like genes may show a similar effect on the formation of palmelloids in Creinhardtii. Therefore, the effect of over-expressing CrBolA into related E. coli strains for biofilm formation was assayed by the crystal violet staining method (Fig. 4a); an approximately two- to threefold increase in biofilm formation over controls was observed. However, over-expression of CrbolA in the E. coli bolA mutant did not form biofilms, showing that CrBolA cannot completely replace E. coli BolA functionally. An unrelated C. reinhardtii flagellar gene (Crfap174) when over-expressed in E. coli XL1blue strain does not form biofilms (Fig. 4a). It appears that a small pool of E. coli BolA protein is required to initiate biofilm formation and an over-expressed CrBolA enhances it further. Controls (cells without plasmid and with plasmid not harboring CrbolA gene) did not show biofilms within 24 h, in either the quantitative or the qualitative assays (Fig. 4).

Figure 4.

Chlamydomonas reinhardtii BolA forms biofilms in Escherichia coli cells. (a) Quantitative microtiter plate biofilm-formation assay at 24 h in different E. coli strains. (b) Qualitative ESEM analyses at 24 h.


EcBolA protein leads to changes in cell morphology, promotes biofilm formation and induces modifications that allow the cell to cope with stress. One of the five C. reinhardtii CrBolA proteins may not completely replace EcBolA in function but upon over-expression is sufficient to trigger biofilm formation and to induce rounder cell morphology. In conclusion, CrBolA protein seems to play a role in morphogenesis and biofilm formation; depicting functional conservation. This is the first compelling evidence depicting the role of a plant BolA-like protein in the morphogenetic pathway and biofilm formation.


J.S.D. thanks Dr. Erik Hom, MCB, Harvard University, USA, for the CrBolA annotation information. Work in the J.S.D. laboratory and the D.K.K. scholarship was funded by the Department of Atomic Energy, India. Work in the C.M.A. laboratory was funded by grants from Fundação para a Ciência e Tecnologia (FCT), Portugal, including grant PEst-OE/EQB/LA0004/2011.