Effect of Biofilm Growth on Expression of Surface Proteins of Actinomyces naeslundii Genospecies 2

Abstract

The predominant surface proteins of biofilm and planktonic Actinomyces naeslundii, a primary colonizer of the tooth surface, were examined. Seventy-nine proteins (the products of 52 genes) were identified in biofilm cells, and 30 of these, including adhesins, chaperones, and stress-response proteins, were significantly up-regulated relative to planktonic cells.

Actinomyces spp. are dominant dental plaque bacteria, and, with certain species of streptococci, they are early colonizers, attaching to the salivary pellicle coating the tooth surface and growing in biofilm (16, 18). Actinomyces spp. therefore mediate establishment of the complex plaque community, via both their interactions with host salivary glycoproteins at the tooth surface and coaggregation with later-colonizing bacteria. Bacteria growing in biofilms are phenotypically and metabolically distinct from planktonic cells, and the transition to the biofilm phenotype is mediated following sensing of key environmental parameters (5, 6). Nutrient starvation and high cell density are two characteristics that are considered, at least in part, to define the physiological characteristics of biofilm cells, and in this respect they may be considered to resemble stationary-phase planktonic cultures (2, 9). Bacteria interact with their environment via surface-associated proteins (20), and these fulfill a range of functions, including adhesion, environmental sensing, and nutrient transport. We therefore studied the effects of biofilm growth on the proteome of surface-associated proteins of Actinomyces naeslundii genospecies 2, separating proteins by two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) and achieving protein identification using liquid chromatography/tandem mass spectrometry (LC-MS/MS).

A. naeslundii genospecies 2, strain 4A7C, was isolated from the approximal plaque of a caries-free patient (21), cultured on fastidious anaerobe agar (LabM, Bury, Lancashire, United Kingdom) supplemented with 5% (vol/vol) horse blood, and incubated in an anaerobic atmosphere for 16 to 24 h. A clinical isolate was selected because of its biological relevance and because, unlike the type strain, it had not been subjected to long-term storage and repeated subculture. Saliva (unstimulated) was collected from healthy human volunteers, pooled, and sterilized as described by Söderling et al. (25). Sterile, polystyrene culture flasks with vented caps (175-cm² surface area; Sarstedt Inc., Newton, NC) were used for growth of A. naeslundii. Flasks were conditioned with saliva to coat surfaces with glycoproteins and provide attachment sites for bacteria. Excess saliva was decanted and replaced with 200 ml of brain heart infusion broth (BHI; Oxoid Ltd., Basingstoke, Hampshire, United Kingdom), which was inoculated with 5% (vol/vol) of a mid-exponential-phase culture of A. naeslundii in BHI. Flasks were secured and incubated statically under anaerobic conditions at 37°C for 7 days to model mature plaque, by which time cells had entered stationary phase, as determined by following the optical density of the planktonic culture. Culture supernatant was decanted and retained to yield planktonic cells, and residual cells that had not firmly adhered to the flask surface were removed by rinsing with water and discarded. The remaining biofilm cells were removed from the surface by physical disruption with a sterile, disposable scraper (Sarstedt). Planktonic and biofilm cells were washed and surface-associated proteins extracted by treatment with Zwittergent, essentially as previously described (29). Growth experiments and extraction of surface proteins were performed on three separate occasions, and each preparation was analyzed independently by 2D-PAGE to facilitate statistical analysis of protein expression. Proteins were suspended in isoelectric focusing (IEF) buffer (Bio-Rad Laboratories Ltd., Hemel Hempstead, Hertfordshire, United Kingdom), and 380 μg of each preparation was applied to an immobilized Pharmalyte gradient strip (17 cm, pH 4 to 7, linear; Bio-Rad). Active rehydration and focusing were performed in a Protean IEF cell (Bio-Rad) according to manufacturer's instructions. Following focusing, proteins were reduced and alkylated in-strip, as previously described (28), prior to sodium dodecyl sulfate (SDS)-PAGE in the second dimension. Gels were stained with colloidal Coomassie brilliant blue (Sigma-Aldrich), and protein expression was analyzed using 2-D Advanced software (version 5.1; Phoretix International, Newcastle upon Tyne, United Kingdom). The integrated optical density (IOD) for each spot was expressed as a percentage of the total IOD across the gel. Percent IOD values were summed for proteins that occurred as isoforms to calculate total expression. Data from triplicate cultures were analyzed, and proteins from biofilm or planktonic cells were considered to be differentially expressed if the mean percent IOD was altered by 50% or greater (i.e., >1.5-fold up- or down-regulation) and there was a statistically significant difference, as determined by Student's t test. This statistical test takes into account the means and standard deviations for individual measurements across independent replicates, and here we required the probability value (P) to be <0.05 for expression data to be considered different across the two growth conditions. Only proteins that were present on all three replicate 2D gels were analyzed by LC-MS/MS for identification. To demonstrate that surface proteins did not arise as a result of cell lysis, total protein was extracted from cells previously treated with Zwittergent and analyzed by SDS-PAGE along with preparations of surface proteins. Proteins from 2D or SDS gels were excised and digested in-gel with trypsin (28, 29). Tryptic digests were analyzed by LC-MS/MS using a ProteomeX system (Thermo Electron, Hemel Hempstead, Hertfordshire, United Kingdom), essentially as previously described (29). MS and MS/MS data were acquired, the latter in data-dependent mode with dynamic exclusion. Spectra were submitted against the prerelease translated genomic sequence database for A. naeslundii strain MG1 (The Institute for Genomic Research; http://www.tigr.org/) using Bioworks v3.1/TurboSEQUEST software (Thermo Electron). Proteins were considered to match entries in the database if (i) XCorr values for individual peptides were ≥1.5, ≥2, and ≥3 for singly, doubly, and triply charged ions, respectively; (ii) ≥3 peptides matched each protein; and (iii) there was ≥20% amino acid coverage of the protein. For clarification of putative functions, protein sequences were retrieved from the A. naeslundii database and BLAST (1) and conserved-domain (CD) (19) searches were performed against the nonredundant database at the National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov/). PSORTb, version 2.0.3 (11; http://www.psort.org/psortb/index.html), analysis gave predicted cellular locations for proteins based on sequence motifs, giving a score of 1.0 for a certain assignment and <1.0 if a protein spans different sites or the site location is unknown. All databases were last accessed August 2005.

2D-PAGE analysis of A. naeslundii surface proteins showed 106 (±13) and 92 (±12) proteins in the pH range of 4 to 7 for biofilm cells (Fig. 1) and planktonic cells (image not shown), respectively. Seventy-nine proteins that were present on all gels derived from biofilm cultures were unambiguously identified as the products of 52 genes (Tables 1 to 3). The inability to identify all proteins on gels could arise from the incomplete nature of the genomic database or the strain under study, which was not the sequenced isolate. Some proteins occurred as isoforms (i.e., migrated differently on gels but were matched to a single open reading frame [ORF]), and most were characterized by small mass changes but clearly different pI values. Proteins undergoing such posttranslational modification (PTM) include enolase, glyceraldehyde-3-phosphate dehydrogenase, and translation elongation factor Ts. The stress-induced DNA-binding protein (Fig. 1, spot 38), however, had isoforms with identical pI but substantially different masses (19 and 21 kDa), probably as a result of proteolytic processing. This is of particular note because only a single isoform was observed in extracts of planktonic cells (Fig. 2). This was the only example found where expression of a particular protein isoform was specific to bacterial growth in biofilm. In addition to proteolysis, PTM can arise due to phosphorylation/dephosphorylation, and further studies of these events in biofilm-cultured bacteria may give insights into regulatory aspects of protein function (2). The most abundant surface proteins of biofilm-grown A. naeslundii included DnaK, enolase, glyceraldehyde-3-phosphate dehydrogenase, stress-induced DNA-binding protein, HSP10, and ribosomal protein L7/L12 (Fig. 1), and these were also highly abundant surface proteins of planktonic cells (data not shown). These proteins have no signal peptide and were not considered surface localized by PSORTb. Other A. naeslundii surface proteins (e.g., the glycolytic enzyme triosephosphate isomerase and ribosome-binding factor A) also lack signal and anchor motifs. Cell lysis did not account for the presence of anchorless proteins at the surface because the abundant intracellular protein lactate dehydrogenase was found only in whole-cell extracts of biofilm-grown and planktonic cells. It is possible, however, that our method of extraction of surface proteins enriched for the anchorless class of proteins. Anchorless proteins are found at the surfaces of other gram-positive bacteria, including streptococci, and some function as adhesins (4, 22, 29). Enolase, which was up-regulated in biofilm-grown A. naeslundii, mediates binding of streptococci to plasminogen and mucin (12, 22), and glyceraldehyde-3-phosphate dehydrogenase recognizes fibronectin (22). These and other bifunctional proteins are found both at the surface and in the cytoplasm of bacterial cells. Further studies are required to determine whether the differential expression of anchorless proteins at the surface of A. naeslundii is accompanied by similar changes in expression of the cytoplasmic forms of these proteins.

FIG. 1.

Surface-associated proteins of biofilm-grown A. naeslundii genospecies 2, strain 4A7C, separated by 2D-PAGE. Proteins were separated by isoelectric focusing (pI range, 4 to 7) in the first dimension and by SDS-PAGE in the second dimension. The migration positions of molecular mass markers and pI values are shown. Spot numbers indicate proteins identified following in-gel tryptic digestion and LC-MS/MS analysis. Proteins for which multiple isoforms were observed are indicated by multiple arrows.

TABLE 1.

Surface-associated proteins significantly up-regulated in biofilm-grown A. naeslundii

Functional categorya	Putative functionb	Spot no.c	Fold up-regulationd	ORFe	Mass (kDa)f	pIg	Peptides matchedh	Sequence coveragei
Adhesins, transport	Adhesin, probable nmb0586 (putative)	13	De novoj	A00361	41	5.5	18	45
and binding	Zinc-binding protein AdcA precursor	6	3.50	A00901	43	5.3	5	20
proteins	Hypothetical ABC transporter extracellular binding protein PH1214 precursor (putative)	12	1.70	A02076	46	5.4	10	27
	Fk506-binding protein (PPIase)	48	De novo	A02772	14	4.2	4	41
	Sugar-binding periplasmic proteins/domains (putative)	9	De novo	A03920	41	4.8	10	28
Cellular process and	Chaperonin (HSP10)	49	3.50	A00150	16	9.2	4	44
stress proteins	Protein with similarity to ATP synthase B chain	35	2.70	A00280	21	4.6	5	29
	Superoxide dismutase (Fe/Mn)	34	4.40	A00939	23	5.1	3	29
	ATP-dependent Clp protease, proteolytic subunit ClpP (encoded by clpP)	43	De novo	A01935	24	5.0	3	25
	General stress protein 14 (GSP14; quinone)	37	De novo	A02237	20	4.6	4	25
	Cochaperone GrpE (encoded by grpE)	23	De novo	A03091	19	4.4	5	30
	Chaperone protein DnaK (heat shock protein 70; HSP70)	1	1.70	A03094	66	4.6	31	54
	Thioredoxin (encoded by trx)	51	2.50	A03338	12	7.8	5	51
	Peptidyl-prolyl cis-trans isomerase B (PPIase B; rotamase B; cyclophilin ScCypB)	32	De novo	A03389	18	4.5	7	67
	Stress-induced DNA-binding protein	38	1.50	A03702	17	4.4	5	44
Central and	Triosephosphate isomerase (encoded by tpiA)	21	1.60	A01121	28	4.8	14	63
intermediary	Flavodoxin (putative)	47	De novo	A01580	19	5.1	3	25
metabolism	Molybdopterin biosynthesis enzymes (PCD)	45	De novo	A02201	18	5.8	7	46
	Enolase (encoded by eno)	5	3.40	A02802	46	4.5	26	67
	Uracil phosphoribosyltransferase (encoded by upp)	28	1.50	A03612	23	4.9	9	44
	Hypoxanthine phosphoribosyltransferase (encoded by hpt)	31	De novo	A04114	19	4.6	11	71
Transcription,	Probable response regulator	52	De novo	A00940	19	5.3	7	33
translation,	Response regulator (encoded by cheY)	30	De novo	A01081	22	4.7	6	25
sensory, and	DNA-directed RNA polymerase omega chain	41	De novo	A01253	14	4.3	3	40
regulatory	Ribosome-binding factor A (encoded by rbfA)	39	De novo	A01735	18	4.6	4	38
proteins	RNA polymerase sigma-70 factor, ECF subfamily (encoded by rpoE)	20	De novo	A02173	28	4.6	8	39
	Sensory transduction protein RegX3	24	De novo	A02416	27	4.5	7	29
	Methanol dehydrogenase regulatory protein homolog (encoded by moxR)	22	De novo	A02861	34	4.9	7	29
	Transcriptional regulator, similar to Myxococcus xanthus CarD (encoded by carD)	44	De novo	A03851	18	5.1	6	35
Other	65-kDa antigen MbaA (GroEL2)	3	De novo	A02975	57	4.7	24	57

^aProteins were assigned to categories on the basis of their putative functions.

^bPutative functions were assigned on the basis of the nomenclature used by The Institute for Genomic Research (Rockville, MD) in the prerelease copy of the translated genomic sequence data for A. naeslundii MG1 (May 2004).

^cRefers to the proteins indicated in Fig. 1.

^dRatio of mean percentage integrated optical density for each protein from biofilm-grown and planktonic A. naeslundii cells (n = 3).

^eThe number assigned to each ORF in the A. naeslundii MG1 database.

^fTheoretical mass of protein as given in the A. naeslundii MG1 database.

^gTheoretical pI of protein as given in the A. naeslundii MG1 database.

^hNumber of tryptic peptides matched to protein sequence on the basis of MS/MS spectra.

ⁱThe percentage of amino acid coverage (peptides observed/theoretical number from sequence data given in A. naeslundii MG1 database).

^jDe novo indicates that a protein was detectable only in surface protein preparations of biofilm cells and was absent in planktonic cells.

TABLE 3.

Surface-associated proteins not differentially expressed in biofilm-grown A. naeslundii

Functional categorya	Putative functionb	Spot no.c	ORFd	Mass (kDa)e	pIf	Peptides matchedg	Sequence coverageh
Transport proteins	Protein-Npi-phosphohistidine-sugar phosphotransferase IIAB component (encoded by ptsL)	46	A03038	17	7.9	3	25
	ABC transporter substrate-binding protein	27	A03599	31	6.0	13	55
Cellular processes	ATP synthase F1, beta subunit (encoded by atpD)	4	A00288	52	4.7	13	28
Central and intermediary	Malate dehydrogenase	16	A02078	34	4.7	13	46
metabolism	Phosphoglycerate mutase (encoded by gpmA)	26	A03842	32	5.5	13	45
Transcription and	Translation elongation factor Tu (encoded by tuf)	7	A00029	44	5.0	10	31
translation	Ribosome recycling factor (encoded by rrf)	40	A01783	22	5.4	10	55
	Translation elongation factor Ts (encoded by tsf)	18	A01787	30	5.1	14	54
	Ribosomal protein L7/L12 (encoded by rplL)	50	A02288	13	4.5	11	91
	50S ribosomal protein (encoded by rplJ)	36	A02290	18	6.4	11	69
	Transcription elongation factor GreA (transcript cleavage factor encoded by greA)	42	A02744	17	4.7	6	43
Other	Xaa-Pro aminopeptidase I (encoded by pepP)	2	A00394	58	4.8	15	32

^aProteins were assigned to categories on the basis of their putative functions.

^bPutative functions were assigned on the basis of the nomenclature used by The Institute for Genomic Research (Rockville, MD, USA) in the prerelease copy of the translated genomic sequence data for A. naeslundii MG1 (May 2004).

^cRefers to the proteins indicated in Figure 1.

^dThe number assigned to each ORF in the A. naeslundii MG1 database.

^eTheoretical mass of protein as given in the A. naeslundii MG1 database.

^fTheoretical pI of protein as given in the A. naeslundii MG1 database.

^gNumber of tryptic peptides matched to protein sequence on the basis of MS/MS spectra.

^hThe percentage of amino acid coverage (number of peptides observed/theoretical number from sequence data given in A. naeslundii MG1 database).

FIG. 2.

Isoforms of a putative stress-induced DNA-binding protein isolated from A. naeslundii genospecies 2, strain 4A7C, resolved by 2D-PAGE. Two forms of the protein mapping to the A03702 open reading frame were observed on 2D gels of proteins extracted from cells grown in biofilms (spot 38, Fig. 1), whereas only one of these isoforms was detected in planktonic cells.

Thirty surface proteins were significantly up-regulated following growth in biofilm, 19 of these being expressed de novo (i.e., not detectable in planktonic cells; Table 1). Two adhesins (A00361 and the lipoprotein A00901) had CDs corresponding to ABC/metal-binding protein families PsaA and AdcA, respectively. These belong to the TroA superfamily, which, in addition to metal uptake and adhesion, mediates virulence and competence for genetic transformation. The substrate-binding components of two ABC-type sugar transporters (A02076 and A03920) were also up-regulated in biofilm. The former had 36% sequence identity with a glucose-binding protein from Thermus thermophilus (accession number YP_004303), but, while A03920 had domains consistent with sugar transport, bioinformatic analyses gave no indication as to the substrate. In addition to adhesion and nutrient acquisition, biofilm maturation and formation of three-dimensional architecture are important for microbial survival in the environment. Antigen MbaA, which was expressed de novo in biofilm-grown A. naeslundii, has a CD of the GroEL chaperonin family and may play a role in these processes. In Vibrio cholerae the mbaA gene regulates both synthesis of a component of the biofilm matrix (extracellular polysaccharide) and formation of a typical biofilm architecture (3), while in Clostridium difficile the chaperonin GroEL is surface associated under stress conditions and acts as an adhesin (14).

Bacteria growing as a biofilm frequently develop a general stress response, and it is suggested that cell surface stress is a signal that modulates the activity of pathways that stabilize biofilm formation (2). Bacteria sense their environment through two-component signal transduction systems which consist of a sensor (responding to a stimulus) and a regulator (controlling expression of a subset of genes), and these can be virulence determinants (10). On the surfaces of biofilm-grown A. naeslundii cells the sensory transduction protein RegX3 (A02416), a CheY-like receiver, was expressed de novo. This protein constitutes part of the senX3-regX3-encoded two-component regulatory system, which is required for virulence in Mycobacterium tuberculosis (23). The biofilm-grown cells yielded a further two response regulators (A00940 and A01081), which were also expressed de novo. The largest class of proteins up-regulated at the surface of biofilm-grown A. naeslundii were the stress response proteins/chaperonins, and these included HSP10, superoxide dismutase, GSP14, GrpE, and DnaK. PSORTb indicated that these proteins should all have a cytoplasmic location, with the exception of superoxide dismutase, which is predicted extracellular (score, 9.55) and is found at the surface of Mycobacterium avium (8). Chaperonins expressed in response to stress conditions function in the maturation of synthesized proteins and in the degradation or refolding of denatured proteins (7, 13). GSP14 and thioredoxin both play roles in protection against oxidative stress. The stress-induced DNA-binding protein (A03702) has a CD indicative of a role in DNA protection under starvation conditions, also protecting against damage by reactive oxygen species. The PSORTb prediction for this protein is cytoplasmic (score of 8.87), but homologs of this protein are highly immunogenic in bacteria including Helicobacter pylori and Treponema pallidum. Chaperones and stress proteins may also act as microbial virulence factors when expressed at the cell surface, functioning as bacterial adhesins and promoting host tissue damage (17). Clp protease (A01935) was expressed de novo in biofilm cells, and these multifunctional enzymes hydrolyze proteins in an ATP-dependent reaction, act as molecular chaperones, and play roles in bacterial stress tolerance, stationary-phase adaptive responses, and virulence (26, 27). Two peptidyl-prolyl cis-trans isomerases (A02772 and A03389) were also expressed de novo at the surfaces of biofilm-grown cells. The FK506-binding protein (A02772) is a member of the FKBP superfamily, which mediates a chaperone-type function in bacteria, facilitating protein folding and protecting against protein aggregation (15). RNA polymerase sigma-70 factor, de novo expressed in biofilm-grown cells, belongs to the extracytoplasmic function sigma factor subfamily. These proteins are activated in response to environmental stress or energy depletion and control many functions in bacteria in addition to transcription, including virulence, cell wall synthesis, and protein folding (2, 24).

Ten identified surface proteins were down-regulated in biofilm-grown A. naeslundii, including some glycolytic enzymes and a cell division protein (Table 2). In addition, seven proteins found in surface preparations of planktonic A. naeslundii were absent in biofilm-grown cells and were therefore specific to that growth state, but these proteins could not be identified (data not shown). Nutritional/energy stress is common in bacteria growing in biofilms, and we detected three components of ATPase/ATP synthase systems (A00280, A00716, and A00288), which have asserted roles in energy generation. All three proteins were expressed in both biofilm and planktonic cells, but they appeared to be differently regulated, and the archaeal-type ATPase (A00716), which functions to pump protons rather than generate ATP under physiological conditions, was down-regulated in the biofilm cultures. We also observed differential expression of the omega and alpha chains of DNA-directed RNA polymerase (A01253 [Table 1] and A00090 [Table 2], respectively). The omega polypeptide was up-regulated in biofilm cells, while the alpha chain was detected in decreased amounts. At present it is not known if these proteins are coordinately regulated, as might be expected, or if there is differential partitioning of the two proteins between the cytoplasm and the cell surface.

TABLE 2.

Surface-associated proteins significantly down-regulated in biofilm-grown A. naeslundii

Functional categorya	Putative functionb	Spot no.c	Fold down- regulationd	ORFe	Mass (kDa)f	pIg	Peptides matchedh	Sequence coveragei
Transport proteins	Glutamine ABC transporter, substrate binding protein (GlnBP)	15	3.20	A01594	33	5.7	10	47
Cellular processes	Cell division protein DivIVA-like protein (CIE)	25	1.50	A01664	22	4.7	9	52
and stress proteins	Archaeal/vacuolar type H+ ATPase subunit H (coiled-coil)	33	1.90	A00716	20	4.8	10	50
Central and	Succinate dehydrogenase subunit B	29	2.70	A00134	28	5.8	7	37
intermediary	Glyceraldehyde-3-phosphate dehydrogenase,	14	1.80	A01118	38	5.7	18	67
metabolism	type I (encoded by gap)
	Phosphoglycerate kinase (encoded by pgk)	10	1.80	A01119	42	4.8	20	57
	Ornithine carbamoyltransferase (encoded by argF)	11	2.50	A01241	37	4.9	8	30
	Glucose-1-phosphate thymidylyltransferase (encoded by rfbA)	19	2.00	A02337	32	4.9	9	30
	Fructose-bisphosphate aldolase, class II (encoded by fbaA)	17	1.50	A03939	37	5.0	13	44
Other	DNA-directed RNA polymerase, alpha subunit (encoded by rpoA)	8	1.70	A00090	36	4.6	19	65

^aProteins were assigned to categories on the basis of their putative functions.

^bPutative functions were assigned on the basis of the nomenclature used by The Institute for Genomic Research (Rockville, MD) in the prerelease copy of the translated genomic sequence data for A. naeslundii MG1 (May 2004).

^cRefers to the proteins indicated in Fig. 1.

^dRatio of mean percent integrated optical density for each protein from planktonic and biofilm-grown A. naeslundii cells (n = 3).

^eThe number assigned to each ORF in the A. naeslundii MG1 database.

^fTheoretical mass of protein as given in the A. naeslundii MG1 database.

^gTheoretical pI of protein as given in the A. naeslundii MG1 database.

^hNumber of tryptic peptides matched to protein sequence on the basis of MS/MS spectra.

ⁱPercent amino acid coverage (number of peptides observed/theoretical number from sequence data given in A. naeslundii MG1 database).

It is likely that we have described the minimal differences between the biofilm and planktonic cells because biofilm cells may have sloughed from the surface and planktonic cells may have adhered to the substratum rather than growing in biofilm. However, our study of the surface proteome of A. naeslundii has defined a subset of proteins significantly up-regulated in biofilm-grown cells, and this included adhesins, transporters, and chaperonins. Future studies will necessitate the generation of mutants of some of these key proteins to determine their roles in the different stages of biofilm formation and maturation. What is perhaps surprising is the relatively large number of anchorless proteins found at the surfaces of biofilm and planktonic A. naeslundii cells. The presence of these proteins at the cell surface can be determined only using methods such as those described in the present study and cannot be inferred by mining genomic sequence data for signature motifs. Presently, the mechanisms of export and association of these proteins with the cell surface are unknown, but it is clear that many more examples of anchorless proteins are being discovered. The functions of most of these have yet to be defined, but they clearly form an important novel class of virulence determinants, being involved in processes including adhesion (4, 22).