Identification of a sensor histidine kinase (BfcK) controlling biofilm formation

Dong Liu,Shikai Ge,Zhenyu Wang,Mengting Li,Wei Zhuang,Pengpeng Yang,Yong Chen,*,Hanjie Ying,*

1 State Key Laboratory of Materials-Oriented Chemical Engineering,College of Biotechnology and Pharmaceutical Engineering,Nanjing Tech University,Nanjing 211816,China

2 School of Chemical Engineering and Energy,Zhengzhou University,Zhengzhou 450001,China

Keywords:Clostridium acetobutylicum Biofilm Histidine kinases CA_C2730 Phosphoproteomics Flagellar motility

ABSTRACT Clostridium acetobutylicum has been extensively exploited to produce biofuels and solvents and its biofilm could dramatically improve the productivities.However,genetic control of C.acetobutylicum biofilm has not been dissected so far.Here,to identify potential genes controlling C.acetobutylicum biofilm formation,over 40 gene candidates associated with extracellular matrix,cell surface,cell signaling or gene transcription,were tried to be disrupted to examine their individual impact.A total of 25 disruptants were finally obtained over years of attempts,for which biofilm and relevant phenotypes were characterized.Most of these disruptants formed robust biofilm still,or suffered both growth and biofilm defect.Only a strain with a disrupted histidine kinase gene(CA_C2730,designated bfcK in this study)abolished biofilm formation without impairing cell growth or solvent production.Further analysis revealed that bfcK could control flagellar biogenesis and cell motility at protein levels.The bfcK also appeared to repress the phosphorylation of a serine/threonine protein kinase(encoded by CA_C0404) that might negatively regulate biofilm formation.Based on these findings,possible bfcK-mediated mechanisms for biofilm formation were proposed.This is a big step toward understanding the biofilm formation in C.acetobutylicum and will help further engineering of its biofilm-based industrial processes.

1.Introduction

C.acetobbutylicumis a well-known industrial strain capable of producing a variety of liquid fuels and important chemicals.A feature of this strain is that it forms biofilm easily on solid surfaces.Biofilm is a complex cell community wherein the cells are encased in extracellular polymeric substances (EPS) secreted by the cells themselves [1].Biofilm provides favorable environment for cell growth and could dramatically improve cell tolerance and metabolic capability [2,3].C.acetobbutylicumbiofilm has been long and extensively used to produce biobutanol[4].However,molecular basis underlying the biofilm formation is yet not clear,which hampers further development of the biofilm-based processes.Unlike model strains such asEscherichia coliandPseudomonas aeruginosawhose biofilm formation has been well-understood[5,6],C.acetobutylicumbiofilm was only depicted from morphological views.No triggering factor or genetic determinant was identified [7].This was mainly due to poor understanding ofC.acetobutylicumphysiology and its genome function.To address this issue,we previously reported for the first time a comparative transcriptomic analysis ofC.acetobutylicumbiofilm and its planktonic cells,which revealed significant modulation of gene expression in the biofilm[8].Subsequently,we analyzedC.acetobutylicumbiofilm EPS composition,which revealed extracellular heteropolysaccharides and proteins in the biofilm matrix [9].Recently,we studied the effect of the master transcription factor Spo0A on biofilm formation [10].Despite these pioneering studies,genetic mechanisms controllingC.acetobutylicumbiofilm formation remains to be elucidated.

Biofilm formation is a complex process and often associated with strain-specific characteristics [7].It is typically a result from multiple,coordinated events[11].In general,extracellular polysaccharides and proteins can serve as intercellular adhesins or biofilm structural components[12].Cell appendages like fimbriae and flagella can also contribute to biofilm formation by mediating initial cell contact with surfaces [5].Signaling processes that usually involve kinases-mediated phosphorylation of regulatory proteins,as exemplified by quorum sensing and two component systems(TCS),together with intracellular secondary messengers like cyclic dimeric guanosine 3’,5’-monophosphate (c-di-GMP) or cyclic adenosine 3’,5’-monophosphate(cAMP),are also important for biofilm formation[13].Biofilm formation could also respond to nutritional cues [12].

In this study,we attempted to find key genes underlyingC.acetobutylicumbiofilm formation by resorting to prior knowledge in other species and bioinformatics analysis.Candidates include genes associated with polysaccharide biosynthesis,extracellular protein,cell surface structure,cell signaling,transcriptional regulation,and genes selected from our transciptomic and proteomic studies of the biofilm.Among these candidate genes,CA_C2730 was identified as the most plausible and critical gene controllingC.acetobutylicumbiofilm formation.It encodes a histidine protein kinase whose role was previously not understood [14].In this study,the roles of CA_C2730 were studied and possible mechanisms of biofilm formation inC.acetobutylicumwere proposed.

2.Material and Methods

2.1.Growth,fermentation and biofilm formation

C.acetobutylicumB3 (CGMCC 5234) was used in this study.It was grown in an anaerobic chamber (Electrotek AW400SG,UK)in P2 seed medium containing 10 g∙L-1glucose.When necessary,Coomassie Brilliant Blue (CBB,2 mg∙L-1) or Congo Red (CR,2 mg∙L-1) was added to the agar to observe colony morphology.CBB can stain proteins and CR can stain amyloid fibers as well as certain polysaccharides like cellulose.Fermentation and biofilm formation were performed in P2 fermentation medium (glucose 60 g∙L-1;K2HPO40.5 g∙L-1;KH2PO40.5 g∙L-1;CH3COONH42.2 g∙L-1;MgSO4∙7H2O 0.2 g∙L-1;MnSO4∙H2O 0.01 g∙L-1;NaCl 0.01 g∙L-1;FeSO4∙7H2O 0.01 g∙L-1;p-aminobenzoic acid 1 mg∙L-1;thiamine 1 mg∙L-1;biotin 0.01 mg∙L-1) at 37 °C with 10% (vol)inoculum.To observe biofilm,a microscope slide as biofilm carrier was initially put into 50 ml of fermentation medium in a 100-ml DuranTMglass bottle and incubated statically for 48 h.For longterm development of biofilm,experiments were performed in 2-L column reactors that were packed with cotton towel and operated continuously in a repeated-batch mode,as described in our previous work [2].For quantification of biofilm biomass,cells were grown in 96-well plates with 200 μl working volume per well.Biofilm biomass was measured according to the crystal violet method described previously[10].Fermentation products and cell densities were measured as previously described [2].

2.2.Gene disruption and complementation

Disruption of genes was performed routinely with the pMTL007C-E2 plasmid using the ClosTron technology as previously described [15].Targeted sites (Table S1) and targeting DNA sequences to be inserted at the sites for each gene were all generated using the computer algorithm available at the ClosTron website (www.clostron.com).Disruptants were screened and cultured with 10 μg∙ml-1of erythromycin.For some genes,disruptants could not be obtained despite our repeated attempts.Altogether,58 targeted sites of 41 genes were tried,among which 25 genes were finally disrupted (Table S1).For some genes (CA_C2337,CA_C2318 and CA_C2341) that could not be disrupted using Clos-Tron technology,CRISPR/Cas9n tools [16] were also tried.Only the CA_C2341 gene was successfully disrupted using a 20-bp guide RNA sequence of “TATCTTTAAGGAAGAAGTTC”which resulted in deletion of the 700th -940th bp region of the gene.Gene expression or gene complementation in disruptants was performed with the pSY8 plasmid under aptbpromoter [17],with an exception ofspo0Awhich seemed to only be expressed under its own promoter.Gene expressing strains were cultured with 15 μg∙ml-1thiamphenicol.All recombinant strains were validated by colony PCR and sequencing of the PCR products.Fermentation products of the disruptants are summarized in Table S2 (Supplementary Material).Effects of all the gene disruption are generally summarized in Table 1.

2.3.Cell motility assay

Strains were grown in P2 seed medium for 12–14 h and the OD600nmwas diluted to 2.0.Then 2 μl of the culture was transferred onto a P2 semisolid plate with 0.4% agar.The agar pates were incubated at 37°C in an anaerobic chamber(Electrotek AW400SG,UK)for 12 h.

2.4.Electron microscopy

For scanning electron microscopy (SEM),biofilm attached on cotton towel or cells collected from planktonic culture were rinsed twice with PBS buffer.Samples were then fixed in 2.5% glutaraldehyde solution at 4 °C for 2 h before ethanol dehydration and lyophilization.The samples were coated with an ion-beam sputtering apparatus and observed with a JEOL JSM-6010 scanning electron microscope.For TEM observation,cells were grown on P2 agar plates for 12 h,and then they were resuspended with a small amount of water.The cells were transferred onto a copper mesh by floating the copper mesh on the resultant cell suspension for 30 s.The mesh was naturally dried at room temperature.Samples were observed with a Hitachi HT7700 transmission electron microscope(TEM) operated at 80 kV.

2.5.Proteomics and phosphoproteomics analysis

Cells grown in P2 fermentation medium were collected at 12 h(exponential growth phase)and 36 h(solventogenisis phase),then sent to Jingjie PTM BioLab (Hangzhou,China) for both proteomics and phosphoproteomics analysis according to the protocol previously reported [18].Briefly,cells were grinded by liquid nitrogen and then sonicated in lysis buffer (8 mol∙L-1urea,1% Triton-100,0.1% protease inhibitor Cocktail Set VI (Sigma),and 0.1% phosphatase inhibitor mixture (Roche,Mannheim,Germany)).Protein solution was reduced with dithiothreitol,alkylated with iodoacetamide,and then diluted.Finally,trypsin was added at a trypsin-to-protein mass ratio of 1:50 for the first overnight digestion and 1:100 for the second 4-h digestion.The resultant peptides were labeled with TMT kit (ThermoScientific) following the user’s guide.Samples were desalted,vacuum-dried,and then fractionated by high pH reverse-phase HPLC using an Agilent 300 Extend C18 column(5 μm particles,4.6 mm×250 mm).For phosphoproteomic analysis,phosphopeptide enrichment was additionally performed by incubating peptides in a 50 μl suspension of IMAC(immobilized titanium-ion affinity chromatography)microspheres(10 mg∙ml-1) in 50% acetonitrile and 6% trifluoroacetic acid with shaking for 30 min.The IMAC microspheres were collected,washed,then the enriched phosphopeptides were eluted in 100 μl of 10% (vol) NH3∙H2O.Phosphopeptides in the supernatant were collected and lyophilized for LC-MS/MS analysis.Pepetides were separated in the EASY-nLC 1000 UPLC system and analyzed by a quadrupole-orbitrap mass spectrometer (Q Exactive,Thermo Scientific).The resultant MS/MS data were processed with Maxquant(v1.5.2.8)for searching against the UniProtC.acetobutylicumproteome database containing 3847 entries.Trypsin/P was specified as a cleavage enzyme allowing for up to two missed cleavages.Mass error was set to 0.001 percent for precursor ions and 0.02 Da for fragment ions.Cysteine alkylation was set as fixed modification.Protein N-terminal acetylation and methionine oxidation were set as variable modification (for phosphopeptide search,phosphorylation on Ser,Thr and Tyr was also included).Phosphoproteomic data was normalized according to protein amounts determined by the quantitative proteomics.Proteins with a foldchange greater than 1.5 at least one time point were considered differentially regulated.The proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD021161.

3.Results

3.1.Genes/gene clusters involved in polysaccharide biosynthesis

Extracellular or surface polysaccharides commonly contribute to biofilm formation.Polysaccharide biosynthesis genes are typically located as operons or gene clusters,encoding proteins responsible for repeat unit synthesis,chain initiation and elongation,length control,transmembrane transport and process regulation [6,19].Besides the genes known for cell-wall peptidoglycan biosynthesis,genomic and transcriptomic study revealed another 4 potential polysaccharide synthesis gene clusters inC.acetobutylicum.

Gene cluster CA_C3059-CA_C3045contains 15 genes starting with CA_C3059 that encodes a sugar transferase (Fig.S1).This sugar transferase is responsible for the transfer of UDP-sugar(may be UDP-galactose) to lipid carrier (undecaprenol phosphate)to initiate polysaccharide synthesis.This cluster also contains a PHP family hydrolase(CA_C3045)and a LytR family transcriptional regulator (CA_C3046) that are typically involved in regulation of exopolysaccharide biosynthesis.However,disruption of this gene cluster at the CA_C3059 locus did not apparently affect biofilm formation or fermentation products (Table S2).

Gene cluster CA_C3073-CA_C3060is immediately adjacent and similar to the CA_C3059-CA_C3045 gene cluster (Fig.S2).The first gene CA_C3073 encodes a sugar transferase sharing 48% amino acid identity with CA_C3059.Of this gene cluster,CA_C3065 encodes a protein homologous to flippase Wzx that transports repeat units across the cytoplasmic membrane for polymerization.This cluster also encodes a Rol/Cld family membrane protein and a CPSC subfamily ATPase that possibly control polysaccharide length and polysaccharide attachment to cell wall.Here,disruption of this gene cluster at the CA_C3073 locus did not reduce the biofilm grown in 96-well plates (Fig.1),but the biofilm grown on glass slides seemed much looser (Fig.2).Meanwhile,fermentation efficiency was reduced.

Gene cluster CA_C2337-CA_C2325is characterized by several genes responsible for sugar structural transformation,and might be involved in synthesis of rhamnose-containing polysaccharides(Fig.S3).This cluster includes a phosphomannomutase gene(CA_C2337) and a UTP-glucose-1-phosphate uridylyltransferase gene (CA_C2335) sharing 49% and 61% amino acid identity withBacillus subtilis pgcAandgtaB,respectively.Mutations in either of these two genes resulted in defective biofilm formation inB.subtilis[20,21].However,despite our repeated attempts,this gene cluster could not be disrupted either at CA_C2337,CA_C2336,or CA_C2335 locus.

Gene cluster CA_C2321-2312is smaller than the other clusters but it contains necessary genes required for polysaccharide synthesis (Fig.S4).This cluster is adjacent to and may be functionally associated with the CA_C2337-CA_C2325 gene cluster,likely involved in rhamnose-containing,teichoic acid-like polysaccharide synthesis.Also,this gene cluster could not be disrupted either at CA_C2313,CA_C2317,or CA_C2318 locus.

CA_C0734encodes a glycosyltransferase.CA_C0734,CA_C0735 and CA_C0736 are homologous to three genes of thepelpolysaccharide gene cluster which is critical for biofilm formation inP.aeruginosa[6].Here,after disruption of CA_C0734 the biofilm formation on both 96-well plates and glass slides seemed not affected,but when cultured with cotton towel continuously for 7 days,the biofilm became thinner (Fig.3).Also,disruption of CA_C0734 made the cell colonies occasionally CR-stained (Fig.4).

CA_C1564is homologous to the gene CDIF630erm_02795 inC.difficile.Bioinformatics analysis predicted thatC.acetobutylicumandC.difficilehad a similar operon involved in secretion of an acetylated glucose polymer [19].Here,disruption of this operon at CA_C1564 locus had no apparent effect on biofilm.

CA_C0332encodes an extracellular beta-mannanase (ManB).Beta-mannanase was involved in biofilm dispersal in some strains.Here,disruption of this gene had no apparent effect on biofilm formation.

Fig.1.Quantification of biofilm mass formed in 96-well plates by B3 wild strain and the disruptants.Biofilm was stained with crystal violet and quantified in terms of OD570nm.

Fig.2.Biofilm formed on glass slides of selected disruptants.For each strain,there are four slides from left to right representing 12,24,48 and 72 h-old biofilm,respectively.

Fig.3.Biofilm formed on cotton towel of selected disruptants.Biofilm was cultured continuously for 7 days in circulated culture medium refreshed every 12 h.

Fig.4.Colony morphology of selected disruptants.Strains were grown on solid agar media containing 2 mg∙L-1 of Congo Red.Red color on the colonies of spo0A-disruptant,its complemented strain (Δspo0A-C) and the CA_C0734-disruptant indicates a CR-stained substance that is yet not clear.These mutants were selected because only they showed apparently different colony morphology.No other strains in this study showed similar phenomenon.

CA_C3308encodes a glycosyltransferase with a 32% amino acid identity to a glycosyltransferase inC.difficilethat was involved in flagellar glycosylation and biofilm formation[22].Here,disruption of CA_C3308 had no apparent effect on biofilm formation.

3.2.Genes involved in cell signaling and transcriptional regulation

Agr quorum sensing (QS) systemhas been demonstrated to regulate granulose formation and sporulation inC.acetobutylicum[23].QS is a vital regulatory mechanism used by many bacteria to communicate and coordinate group behavior such as biofilm formation.TheagrBgene (CA_C0078) of theagrsystem(CA_C0078-CA_C0081)inC.acetobutylicumis required for producing the AIP (autoinducter peptide) signal molecule.Here,disruption ofagrBenhanced biofilm formation (Fig.1 and Fig.3) and fermentation (Table S2),although the biofilm on glass slides seemed looser (Fig.2).Colonies of theagrB-disruptant were also much rougher and more irregular (Fig.4).

LuxS QS systemmediates signaling in both Gram-positive and Gram-negative species.This QS system involves a furanone compound as the signal molecule (AI-2) which is produced as a byproduct of the LuxS enzyme.LuxS deficiency could cause significantly impaired biofilm formation inStreptococcus mutans[24].Here,disruption ofluxSgene (CA_C2942) had no apparent effect on biofilm formation and fermentation(Fig.3 and Table S2)though it made the cell colonies much smoother (Fig.4).

spo0A(CA_C2071) encodes the master regulator of sporulation that also affects certain other physiology.InC.acetobutylicum,Spo0A serves as the response regulator of a two-component system (TCS).It consists of a CheY-like signal receiver domain and a transcription activation domain[25].Similar to CA_C0734,disruption ofspo0Aresulted in CR-stained colonies (Fig.4),although the stained substances remain to be identified.The disruptant could not form biofilm anymore (Fig.2) and its motility was markedly reduced,as characterized in our previous study[10].However,cell growth and fermentation of this disruptant were severely impaired(Table S2).So,the abolished biofilm in this disruptant could be a result of growth defect.

CA_C2730encodes a signal transduction histidine kinase whose role has not been identified.InC.acetobutylicumthere are five orphan histidine kinases and previous study showed that CA_C2730 was the only one not involved in sporulation[14].Here,disruption of CA_C2730 abolished biofilm formation (Fig.1 and Fig.2) while cell growth and fermentation were not impaired.The maximum cell density (OD600nm) during fermentation was even increased from an average of 5.0 to 7.5,and the culture broth became obviously viscous.The colonies of this disruptant were also much thicker than those of wild strains (Fig.5A and 5B).Interestingly,it was found that flagella disappeared in the disruptant(Fig.5C and 5D) and cell motility was apparently reduced.In contrast,when CA_C2730 was over-expressed,cell motility was strikingly increased (Fig.5B).When the CA_C2730-disruptant was complemented with CA_C2730 gene,the biofilm formation could be effectively restored(Fig.S5).So,the CA_C2730 gene was herein designatedbfcKand further studied in the following section.

CA_C0404encodes a serine/threonine protein kinase (STPK).It is the first gene of the CA_C0404-CA_C0410 operon encoding an ESAT-6/WXG100 secretion system (Fig.S6).Bioinformatics analysis suggested that CA_C0404 was involved in phosphorylationdependent protein–protein interactions with the ESAT-6/WXG100 secretion system [26,27].STPKs such as the PrkC inB.subtilisandB.anthraciswere critical for biofilm formation[28,29].Here,disruption of CA_C0404 slightly increased biofilm formation.However,when CA_C0404 was overexpressed,the biofilm was apparently decreased while cell growth and fermentation were not affected(Fig.S7,Table S2).This suggested that CA_C0404 might negatively regulate biofilm formation.There is another STPK(CA_C1728) inC.acetobutylicumbut it could not be disrupted in this study.

Fig.5.Changes in colony morphology,cell motility and flagella in the bfcK-disruptant.ΔbfcK,the bfcK-disruptant;B3 (psy8),B3 wild strain harboring the pSY8 plasmid as control;ΔbfcK-C,the bfcK-disruptant complemented with pSY8-bfcK plasmid;B3(psy8-bfcK),B3 wild strain overexpressing bfcK.The ΔbfcK colonies were thicker than those of B3 strain when grown on solid agar media (A).Cell motility on semi-solid agar media was increased upon expression of bfcK (B).TEM showed that B3 wild strain had flagella (C) whereas the flagella disappeared in the ΔbfcK strain (D).

CA_P0082encodes a cyclic AMP(cAMP)receptor protein(CRP).As a global transcriptional regulator,CRP regulates many biological processes including metabolism of carbon sources and biofilm formation.CRP-cAMP complex was reported to repress biofilm formation or in some cases it activated biofilm formation [12].Here,disruption ofcrpslightly reduced biofilm formation,but cell growth and fermentation were apparently impaired.

CA_C2222encodes a methylesterase CheB that can remove methyl group from methyl-accepting chemotaxis proteins (MCP).Disruption of this gene had no apparent effect on biofilm formation and fermentation.

3.3.Genes selected from transcriptomic and proteomic studies

CA_C0253encodes a nitrogenase NifH.This gene and other nitrogen fixation related genes (CA_C0253-CA_C0254) were upregulated 120-to 200-fold inC.acetobutylicumbiofilm cells,compared to those in planktonic cells at the late growth stage[8].Here,disruption of this gene only slightly reduced biofilm formation.

CA_C0377encodes an amino acid ABC transporter permease.Transcriptomic analysis showed that the CA_C0376-CA_C0378 operon genes were all up-regulated inC.acetobutylicumbiofilm,up to 30-to 34-fold [8].Proteomic analysis also showed elevated protein levels of CA_C0377 during long term biofilm development(data not published).Here,disruption of this gene did not apparently affect the biofilm formation.However,the biofilm formed on cotton towel appeared less hydrated (Fig.3).

CA_C3179encodes an oligopeptide ABC transporter component whose expression was up-regulated 5-to 15-fold in the early stages of biofilm formation[8].Many genes encoding oligopeptide ABC transporter components were up-regulated to different extents inC.acetobutylicumbiofilm cells.However,disruption of CA_C3179 did not apparently affect the biofilm formation.

CA_C3557is adjacent to CA_C3558 gene locus and both genes encode surface layer (S-layer) proteins (SLP) sharing 56% identity.Transcriptomic analysis showed that they had a similar gene expression pattern but CA_C3557 expression intensity was relatively lower than CA_C3558 [8].However,proteomic analysis showed that CA_C3557 (identical to CEA_G3563 inC.acetobutylicumEA2018) was among the most abundant extracellular proteins inC.acetobutylicumbiofilm matrix [9].InC.difficile,Slayer proteins were essential for biofilm formation perhaps because S-layer was essential for anchoring cell wall associated proteins that were required for biofilm formation [11].Here,disruption of CA_C3557 did not apparently affect biofilm formation on 96-well plates and glass slides (Fig.1 and Fig.2).However,the biofilm grown on cotton towel was much thinner and looser(Fig.3).

CA_C2517encodes an extracellular neutral metalloprotease.This gene was up-regulated inC.acetobutylicumbiofilm [8] and the protein was apparently present in the biofilm matrix[9].However,disruption of this gene had no apparent effect on biofilm formation.

Our studies also had shown that molecular chaperones and reverse rubrerythrins were very abundant inC.acetobutylicumbiofilm matrix.However,the GroEL (CA_C2703),GroES (CA_C2704),DnaK (CA_C1282),rubrerythrins or reverse rubrerythrins(CA_C3597/CA_C3598/CA_C2575) genes could not be disrupted in this study despite repeated attempts.

3.4.Other genes of functional proteins

CA_C3355encodes a type I single-modular polyketide synthases whose products were shown to act as chemical triggers of certain physiological events inC.acetobutylicum[30].Here,disruption of this gene had no apparent effect on biofilm formation,though the biofilm on glass slides seemed somewhat sporadic(Fig.2).CA_C2341encodes a collagenase.Collagen-like proteins were reported to be surface adhesins during biofilm formation in some strains[31].Although no gene has been annotated to encode collagen inC.acetobutylicum,disruption of this putative collagenase slightly enhanced biofilm formation.CA_C3085encodes a TPR repeat-containing cell adhesion protein.Disruption of this gene showed no apparent effect on biofilm formation.Biofilm formation was also characterized for strains with disruptedCA_C2229(encoding pyruvate-ferredoxin oxidoreductase PforA),CA_C2499(pyruvate-ferredoxin oxidoreductase PforB),CA_P0025(pyruvate decarboxylase),orCA_P0165(acetoacetate decarboxylase).These genes are involved in central metabolism and not supposed to play roles in biofilm formation,but their disruptants could serve as controls for assessing the impacts of above investigated genes.Results showed that although disruption of these genes decreased solvent production to different extents,they generally had no apparent effect on biofilm formation(Fig.1).All these disputants could form biofilm comparable to that of wild cells on glass slides except theadc-disruptant.Theadc-disruptant accumulated a large amount of acetic acid which dramatically impaired cell growth and fermentation(Table S2).This suggested that,in general,biofilm formation inC.acetobutylicumwas robust and not necessarily associated with solvent production.

3.5.Quantitative proteomics and phosphoproteomics study of bfcK

SincebfcKwas the most critical gene controlling biofilm formation among the gene set investigated above,it was further studied by proteomics.Results showed that whenbfcKwas disrupted,a total of 315 proteins were differentially expressed more than 1.5-fold at either acidogenic phase (12 h) or solventogenic phase(36 h) (Supplementary File 2:Table P1).Functional enrichment analysis revealed that the most notable change occurred to flagellar biogenesis and chemotaxis(Fig.6A).Proteins of these two processes were down-regulated,mostly 2-to 5-fold at both phases.This was consistent with the disappearance of flagella and impairment of motility observed in thebfcKdisruptant.Also greatly down-regulated was the S-layer protein (CA_C3558) which was down-regulated more than 10-fold at the solventogenic phase.The reverse rubrerythrin (Rbr3B),typically abundant in biofilm matrix,was down-regulated 2-fold in the disruptant.Several proteins potentially involved in polysaccharide biosynthesis(CA_C2166,CA_C2176,CA_C2174) were also down-regulated.Some uncharacterized protein (CA_C2620,CA_C2185,CA_C0052,CA_C2204,CA_C0327,CA_C2202) were down-regulated 3-to 5-fold at both phases.By contrast,proteins involved in sporulation,starch synthesis,peptidoglycan biosynthesis and cell wall macromolecule metabolism were generally up-regulated in thebfcKdisruptant(Fig.6A).Many carbohydrate metabolism proteins were differentially regulated,but few of the solventogenic genes were affected,explaining the unaffected solvent production inbfcKdisruptant.

Since BfcK is a histidine protein kinase,changes in protein phosphorylation was expected in thebfcK-disruptant.Phosphoproteomics study showed that a total of 84 sites within 64 proteins were differentially phosphorylated (fold-change greater than 1.5 at either acidogenic phase or solventogenic phase)(Supplementary Material File 2:Table P2).The top 30 most differentially phosphorylated proteins are listed in Fig.6B.Protein kinases were of particular interest because they dominate protein phosphorylation.Two histidine kinases CA_C3319 and CA_C0437 which had been shown involved in phosphorylation and de-phosphorylation of Spo0A,respectively,were down-regulated in quantitative proteomics analysis,but they were not detected as differentially phosphorylated proteins.A previously uncharacterized PrkA-like serine protein kinase (CA_C0579) had a 2-fold higher protein level but a 1.5-to 1.8-fold lower phosphorylation level at the acidogenic phase.The serine/threonine protein kinase (CA_C1728) was also down-phosphorylated at the acidogenic phase.As mentioned above,this gene could not be disrupted and its role remained to be elucidated.Among the differentially phosphorylated kinases,only CA_C0404 was apparently up-phosphorylated at both acidogenic and solventogenic phases (Supplementary Material File 2:Table P2).A protein CA_C0406 was also up-phosphorylated at solventogenic phase.CA_C0406 belongs to the CA_C0404-CA_C0410 operon and it encodes an FHA-domain that could probably bind to CA_C0404-phosphorylated phosphopeptides [26].

A phosphocarrier protein (HPr/PtsH) that is a central component of the sugar phosphotransferase system (PTS) was downphosphorylated 2-to 2.5-fold at multiple sites at the acidogenic phase.Dephosphorylated HPr usually indicate active glycolysis and could possibly favor primary metabolism [32,33].Also apparently down-phosphorylated were a DnaJ family molecular chaperone (CA_C0648),a TldD protein (CA_C1828),an uncharacterized protein (CA_C2564),an uncharacterized membrane protein(CA_C0354) and a DUF1540 domain-containing protein CA_C3540.They were down-phosphorylated predominantly at acidogenic phase.An NADP+-dependent glyceraldehyde-3-phosphate dehydrogenase (CA_C3657) showed 1.5-to 2.5-fold higher phosphorylation levels at multiple sites at the solventogenic phase.Some key proteins involved in central metabolism such as ThlA,AdhE1 and PFOR showed 1.5-to 2-fold higher phosphorylation levels at the acidogenic phase.

A large number of electron carrier and antioxidant proteins were differentially regulated in thebfcK-disruptant (Supplementary Material File 2).In particular,the bacterioferritin comigratory protein (CA_C0327),the NADH-rubredoxin oxidoreductase (NroR/CA_C2448),the Flavo-diiron protein(FprA2/CA_C2449),the desulfoferrodoxin(Dfx/CA_C2450),the thioredoxin reductase(TrxR)and a reverse rubrerythrin(Rbr3B)were all downregulated 2-to 3-fold at protein levels at both phases.These proteins are crucial for detoxification of reactive oxygen species in anaerobic bacterium[34–36].Some of these antioxidant proteins,the NroR,FprA2,Dfx and TrxR,however,showed 1.5-to 2-fold higher phosphorylation levels.

4.Discussion

In this study,41 gene candidates inC.acetobutylicumwere tried to identify their impact on biofilm formation.Some of them could not be disrupted despite of repeated attempts,probably because they are essential for cell survival or due to technical problems.Among the 25 gene disruptants,most could still form robust biofilm.Some polysaccharides biosynthesis gene disruptants(ΔCA_C3073,ΔCA_C0734) might alter biofilm composition because the biofilm showed somewhat different morphology,but they did not fundamentally affect the biofilm.Some genes that were commonly found differentially regulated inC.acetobutylicumbiofilm,such as the oligopeptide ABC transporter genes and nitrogenase genenifH,also showed little effect on biofilm formation.In general,genes involved in cell signaling and transcriptional regulation had the most notable effects on biofilm formation,suggesting that biofilm formation was a complicated process associated with global gene expression.

Overall,thebfcK,spo0Aand S-layer gene had the most apparent effect on biofilm reduction.S-layer is the outermost cell envelope dominating cell adhesion and surface properties [11].Disruption of the S-layer made the biofilm on cotton towel much thinner and looser,although the biofilm on 96-well plates and glass slides appeared less affected.S-layer proteins were also demonstrated essential for biofilm formation inC.difficile,probably because it was useful for anchoring biofilm-associated proteins [11].Disruption ofspo0Aabolished biofilm inC.acetobutylicum.Our recent study revealed that wire-like extracellular structure that crosslinked cells was eliminated uponspo0Adisruption [10].However,the problem with thespo0Adisruptant was that cell growth and solvent production were also severely impaired.So,the abolished biofilm formation uponspo0Adisruption could potentially be a result of growth defect.More elaborate experiments are therefore needed to study the role of Spo0A in biofilm formation in future.Among the genes investigated in this study,only thebfcK-disruptant abolished biofilm formation without impairing cell growth or solvent production.ThebfcKgene encodes a membrane sensor kinase which is an orphan kinase with yet unknown ligands.Homologues of BcfK exist in other Clostridia likeC.ljungdahliiandC.autoethanogenum,but have not been characterized as well.It has been reported that histidine kinases regulate biofilm formation in many bacteria.InB.subtilis,the histidine kinases KinA,KinB,KinC and KinD could phosphorylate the transcriptional regulator Spo0A.The phosphorylated Spo0A caused derepression of theepsA-epsOandyqxM-sipW-tasAgenes responsible for biofilm matrix production [37–39].Although Spo0A exists inC.acetobutylicum,similar matrix operons likeepsA-epsOandyqxM-sipWtasAcould not be identified.Furthermore,previous study showed thatbfcKwas not involved in Spo0A phosphorylation inC.acetobutylicum[14].So,thebfcKmight control biofilm formation inC.acetobutylicumin a way different from that for theB.subtiliskinases.Histidine kinases have also been reported to regulate bacterial chemotaxis and motility.InComamonas testosteroni,the histidine kinase CheA phosphorylated not only its cognate response regulator CheY (implicated in chemotaxis),but also one response regulator FlmD implicated in biofilm formation[40].InE.coli,certain histidine kinase phosphorylated ArcA.Phosphorylated ArcA then activated FliA that was necessary for the expression of class-3 flagellar genes.AnarcAmutant showed a motilitydefective phenotype [41].Similar to these studies,bfcKwas found to mediate biofilm formation through flagellar biogenesis in the present study.Disruption ofbfcKeliminated flagella and cell motility by reducing their protein levels.Flagella are critical for biofilm formation on surfaces because they could act as adhesins to mediate initial surface attachment.Flagella-dependent motility is also important for overcoming repulsive forces that might exist between the bacteria and the surfaces,and important for spread of biofilm by facilitating cell movement along surfaces [5,42].In addition,in some strains flagellar motility appeared to recruit planktonic cells into biofilm during the biofilm development[42].Disruption ofbfcKalso affected numerous proteins of electron carrier activity or antioxidant activity.In particular,the antioxidant proteins NroR,FprA2,Dfx and TrxR showed apparently reduced protein levels but elevated phosphorylation levels.In some strains,processes associated with these proteins could affect the assembly of extracellular protein appendages including flagella[43,44].Whether or not these antioxidant proteins were associated with the disappearance of flagella in thebfcK-disruptant needs further investigation.Furthermore,disruption ofbfcKreduced the Slayer protein (CA_C3558) 10 times at the solventogenic phase,which should favor the abolishment of biofilm formation.The reduction in S-layer and flagella proteins which were all extracellular inspired us to explore whether protein secretion in thebfcKdisruptant was affected.Results showed that thebfck-disruptant secreted 3-fold less protein than the control did throughout the fermentation process(27–43 mg/gDCW versus 70–135 mg/gDCW).So,Disruption ofbfcKmay generally affect protein secretion.

Fig.7.Possible mechanisms of BfcK-mediated biofilm formation in C.acetobutylicum.BfcK is a membrane sensor histidine kinase that is encoded by CA_C2730 gene.It could promote flagellar biogenesis,cell motility,S-layer formation and protein secretion,all of which facilitate biofilm formation.CA_C0404 encodes a serine/threonine protein kinase (whether or not this protein is located on the membrane is yet not clear).While the CA_C0404 protein (or possibly the phosphorylated CA_C0404 protein) seems to inhibit biofilm formation,BfcK could repress its phosphorylation.BfcK also apparently modulates the antioxidation process in C.acetobutylicum.

Disruption ofbfcKalso changed the phosphorylation levels of proteins.The most notable change occurred to the STPK CA_C0404.This STPK shares low identity with STPKs commonly studied in other species.Our previous proteomics data showed that CA_C0404 was poorly expressed in the wild strain.Disruption of CA_C0404 only slightly increased the biofilm.However,when it was overexpressed the biofilm was apparently reduced,suggesting a negative impact of CA_C0404 on the biofilm formation.WhenbfcKwas disrupted,CA_C0404 was apparently up-phosphorylated at both acidogenic and solventogenic phases.This up-regulated phosphorylation of CA_C0404 could possibly increase its activity and thus reduced biofilm formation.Furthermore,a protein encoded by CA_C0406 (located on the same operon with CA_C0404) was also up-phosphorylated at the solventogenic phase.This protein is probably phosphorylated through the action of the CA_C0404 protein.It contains an FHA domain that could bind to phosphorylated phosphopeptides.This may block its own activity and thus affect the ESAT-6/WXG100 secretion system[26] that could possibly be implemented in biofilm formation.Based on the above discussion,possible mechanisms for BfcKmediated biofilm formation inC.acetobutylicumare overviewed in Fig.7.Overall,thebfcK-disruptant abolished biofilm formation without impairing cell growth and solvent production.The abolished biofilm formation could be explained,at least in part,by impaired flagellar biosynthesis,cell motility and protein secretion in the disruptant.

C.acetobutylicumbiofilm was applied in solvent production 20 years ago[4],but molecular mechanisms underlying the biofilm formation have not been reported so far.This is the first study revealing a novel,significant genetic determinant for biofilm formation inC.acetobutylicumand establishes a model for further dissection of the biofilm formation.The physiological,metabolic,proteomic and phosphoproteomic insights intoC.acetobutylicumand its numerous mutants in this study will also help better understand and engineer this bacterium in the future.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors thank Prof.Nigel P.Minton from the University of Nottingham for kindly providing the ClosTron plasmids,and Prof.Sheng Yang from Shanghai Institutes for Biological Sciences for kindly providing pSY8 plasmid.This work was supported by the Key Program of the National Natural Science Foundation of China(Grant No.21636003),the Outstanding Youth Foundation of Jiangsu (Grant No.SBK2017010373),the National Key Research and Development Program of China (Grant No.2019YFD1101204),and the Jiangsu Synergetic Innovation Center for Advanced Bio-Manufacture.Dong Liu is supported by the Jiangsu Qinglan Talent Program.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.04.024.