Genome sequence of Streptococcus mutans bacteriophage M102

Abstract

Bacteriophage M102 is a lytic phage specific for serotype c strains of Streptococcus mutans, a causative agent of dental caries. In this study, the complete genome sequence of M102 was determined. The genome is 31 147 bp in size and contains 41 ORFs. Most of the ORFs encoding putative phage structural proteins show similarity to those from bacteriophages from Streptococcus thermophilus. Bioinformatic analysis indicated that the M102 genome contains an unusual lysis cassette, which encodes a holin and two lytic enzymes.

caries, Streptococcus mutans, lysin, bacteriophage, holin

Introduction

Dental caries is a frequent disease caused by microorganisms present on the tooth surface that convert carbohydrates present in the diet to lactic acid. The acid production results in demineralization of the tooth enamel and tooth dentin. Streptococcus mutans, a gram-positive facultative anaerobic bacterium, has a major role in this demineralization process (Loesche, 1986).

Prevention of dental caries by reduction of S. mutans from dental plaque involves nonspecific approaches such as tooth brushing and the use of antimicrobial agents like chlorhexidine. Specific approaches include active or passive immunization and replacement of S. mutans by noncariogenic competitor organisms (Samaranayake, 2002), although none of these concepts has been realized thus far. In theory, bacteriophages or, preferably, lytic enzymes from bacteriophages might be used to specifically remove S. mutans from dental plaque as well. The application of bacteriophage and bacteriophage-encoded lytic enzymes as antibacterial agents has recently regained interest (Fischetti, 2001, 2005; Loeffler et al., 2001; Loessner, 2005).

Relatively limited information is available with regard to the ecological role of bacteriophages in the oral cavity. The isolation of bacteriophages from human saliva or dental plaque has had variable success (Tylenda et al., 1985; Armau et al., 1988; Bachrach et al., 2003; Hitch et al., 2004). Lysis of and virus release by several strains of S. mutans after mitomycin treatment or UV exposure has been observed, indicating that these strains contained prophages (Greer et al., 1971; Klein & Frank, 1973; Higuchi et al., 1982). However, none of these phages have been purified and further characterized. By screening more than 1000 plaque samples for lytic activity against strains of S. mutans and Streptococcus sobrinus, Armau (1988) isolated 16 lytic bacteriophages. Three of these, phages M102, e1 and f2 were found to be specific for S. mutans of serotype c, e and f, respectively (Delisle & Rostkowski, 1993). In the present report, the sequencing and analysis of the genome of phage M102 is described. In addition, the sequence of the lysis genes of phage M101, which is similar to phage M102, was determined.

Materials and methods

Bacterial strains, bacteriophages and growth conditions

Bacteriophages M101 and M102 and their host strain S. mutans OMZ381 were acquired from G. Tiraby (Université Paul Sabatier, Toulouse, France). For phage propagation, S. mutans OMZ381 was grown in M1D medium, which consisted of 10 g Bacto tryptone, 5 g Bactopeptone, 5 g yeast extract, 5 g NaCl, 2.5 g MOPS (4-morpholinepropanesulfonic acid) and 2 g glucose L⁻¹. The pH of M1D was adjusted to pH of 7.4 with NaOH. Solid media contained 1.5% agar (for plates) or 0.7% agar (for soft agar). Phage lysates of M101 and M102 were obtained by infection of exponentially growing cultures of S. mutans OMZ381 and subsequent incubation for 16 h at 37°C. To remove debris, lysates were centrifuged for 10 min at 7000 g and passed through a 0.4 µm filter.

Electron microscopy

A drop of phage solution (about 10⁹ PFU mL⁻¹) was applied for 3 min to 400 mesh copper grids coated with Formvar (Electron Microscopy Sciences) and stabilized with carbon. The grids were air-dried and subsequently a drop of 1% phosphotungstic acid (PTA) pH 4.4 was applied for 45 s. The PTA solution was removed with filter paper, the grids air dried, and examined with a Philips EM 400T TEM at 80 kV.

Isolation of phage DNA, DNA manipulations and sequencing

Phage DNA was isolated using the Lambdaprep kit (Promega, Wallisellen, CH). A shotgun library of phage M102 genomic DNA was prepared by GATC Biotech (Konstanz, Germany) as follows. Genomic DNA from phage M102 was sheared by nebulization, blunted with T4 DNA polymerase and Klenow polymerase, and then cloned into pCR4Blunt-Topo (Invitrogen). Plasmids from resulting clones were isolated and sequenced with the T7 and T3 primer. Sequencing was performed using dye terminator technology on a model 3100 sequencer (Applied Biosystems). Sequences were assembled with the program Seqman of the Lasergene package (GATC Biotech). Gaps were closed using PCR and by direct sequencing of the resulting products. A total of 320 sequencing reactions were performed to obtain the complete sequence of M102 with an average coverage of 8.15. Digestion of phage DNA with several different restriction enzymes (PstI, SalI, XhoI, HindIII, EcoRV) yielded fragments whose sizes were in accordance with the genome sequence.

Partial sequences of phage M101 were obtained by direct sequencing of PCR products, obtained with primers derived from the M102 sequence.

Sequence analysis

The assembled sequence of M102 was analyzed for the presence of ORFs using the program genemark.hmm (Lukashin & Borodovsky, 1998), which uses a cutoff of 42 nucleotides for the minimal coding region. Further sequence analysis used the programs from the GCG package (Accelrys, Cambridge, UK). Blast sequence similarity searches were carried out at http://nbc3.biologie.uni-kl.de/. For multiple sequence alignments, clustalw was used (http://www.ebi.ac.uk/clustalw/).

Phylogenetic distances between phage proteomes were essentially calculated as described (Rohwer & Edwards, 2002). In brief, ORFs from M102 were compared pariwise with all the ORFs of 17 selected Streptococcal or Lactococcal phages by blast analysis (cutoff 0.1). Similar proteins were then aligned using clustalw (gap opening penalty of 10.00 and gap extension penalty of 0.2). The output (in phylip format) was used to determine the phylogenetic distance between each ORF from phage M102 and the corresponding ORF from the other phages with the program protdist (http://artedi.ebc.uu.se/programs/protdist.html). In case of no matching ORF, a penalty of 10 was used. For each phage compared with M102, the sum of the protdist values divided by the number of ORFs used in the comparison was calculated. Calculations were carried out independently for the complete M102 proteome, for Orf1 to Orf18 (structural module) and for Orf19 to Orf41 (lysis and replication module).

Results and discussion

Morphology of bacteriophage M102

Electron microscopic analysis (Fig. 1) showed that M102 had a tail with a length of 269±33 nm (n=69) and a width of 9.5±1.2 nm (n=32). The uniformity of tail lengths indicates that it is noncontractile. The icosahedral phage head had a diameter of 63±3 nm (n=50). These values differ somewhat from those reported previously (Delisle & Rostkowski, 1993). The tail was segmented and consisted of segments of 4.0±0.2 nm (n=56). These results indicate that phage M102 belongs to the family of Siphoviridae with morphotype B1.

1

Electron microscopy images of phage M102.

Genome sequence of M102

The genome of bacteriophage M102 was 31 147 bp in size, which is slightly smaller than the size that was determined previously by restriction enzyme analysis (Delisle & Rostkowski, 1993). The GC content was 39.21%, close to the previously reported value (Delisle & Rostkowski, 1993), but somewhat higher than the value of 36.82% for the genome of S. mutans UA159 (Ajdic et al., 2002). Analysis of the genome sequence revealed the presence of 41 ORFs, all transcribed in the same direction (Fig. 2). Most of the ORFs had an ATG startcodon, but there were three ORFs with a GTG start codon and three with a TTG start codon (Table 1). The sequences of the ORFs were compared with sequences from protein databases using protein–protein blast. Based on these comparisons, most of the ORFs could be assigned to the different functional groups (Fig. 2). The ORFs from the same functional groups clustered together on the genome.

2

Organization of Streptococcus mutans phage M102 genome and comparison with the genome from Streptococcus thermophilus phage DT1 (Accession number NC_002072). For clarity, Orf45 from DT1 is shown at the left site. Putative promoters are indicated with horizontal arrows, cos sites by vertical arrows. Putative terminators are depicted by ΩOrfs with similarity are connected by shading and the similarity (in % amino acid identity) is indicated.

1

Characteristics of Orfs encoded by bacteriophage M102

ORF	Start	Start codon	Stop	Size (aa)	pI	% identity (over aa)	e-value	With [organism] (Genbank accession number)	Putative function (conserved domain)
1	181	ATG	549	122	9.61	36 (95)	3e-12	Restriction endonuclease [Pediococcus pentosaceus ATCC 25745] (ABJ68063)	Endonuclease (HNHc)
2	571	ATG	909	112	5.29	27 (96)	0.046	Putative small subunit terminase [S. thermophilus bacteriophage Sfi19] (AAD44055)	Small subunit terminase
3	896	ATG	2770	624	5.17	40 (614)	1e-133	ORF22 [S. thermophilus bacteriophage 7201] (AAF43515)	Large subunit terminase (terminase_1)
4	2840	ATG	3784	314	5.44	37 (295)	1e-48	Putative portal protein [S. thermophilus bacteriophage DT1] (AAD21882)	PORTAL protein (Phage_portal)
5	3781	TTG	4677	298	4.09	31 (195)	6e-19	Predicted Clp-protease [S. thermophilus bacteriophage Sfi21] (AAD41032)	CLP protease (CLP_protease)
6	4697	ATG	5833	378	4.82	24 (368)	2e-15	Major structural protein [Lactococcus phage BK5-T] (AAK56807)	Phage capsid (phage capsid)
7	5876	GTG	6187	103	4.15	33 (93)	5e-04	DNA packaging, phage associated [Bacteriophage Sal2] (YP_535183)	DNA packaging protein
8	6184	ATG	6528	114	9.35	27 (62)	0.06	ORF28 [S. thermophilus bacteriophage 7201] (AAF43521)	Unknown
9	6521	ATG	6913	130	6.10	32 (115)	2e-09	ORF29 [S. thermophilus bacteriophage 7201] (AAF43522)	Unknown
10	6897	ATG	7253	118	4.30	30 (92)	3e-05	Probable tail component protein 123 [S. thermophilus phage Sfi19] (T09268)	Probable tail component
11	7271	ATG	7954	227	5.52	38 (194)	3e-34	MPS-7201 [S. thermophilus bacteriophage 7201] (AAB71820)	Probable major tail protein (phage tail)
12	8127	ATG	13 328	1733	10.13	43 (1754)	0	Putative tail component protein [S. thermophilus bacteriophage DT1] (AAD21891)	Tail protein (CHAP; pfam05257) (LT-GEWL; pfam1464) (Phage-related tail protein)
13	13 328	ATG	14 854	508	5.55	28 (512)	4e-59	Putative tail component protein [S. thermophilus bacteriophage Sfi21] (AAC39282)	Probable tail component
14	14 851	ATG	17 007	718	4.81	29 (570) 39 (215)	4e-58 1e-34	Host specificity protein [S. thermophilus bacteriophage DT1] (AAD21894)	Host specificity protein
15	17 008	ATG	18 765	585	4.94	27 (577)	8e-50	Structural protein [Streptococcus phage 2972] (AAW27943)	Minor structural protein
16	18 786	ATG	19 289	167	7.48	–	–	–	Unknown
17	19 306	TTG	19 578	90	9.43	27 (83)	0.001	Hypothetical protein EJ-1p68 [Bacteriophage EJ-1] (NP_945307)	Unknown
18	19 575	ATG	20 042	155	7.60	30 (107)	0.003	Phage holin, LL-H [S. suis 89/1591] (EAP40817)	Holin
19	20 058	ATG	20 879	273	4.65	59 (220)	8e-70	Peptidoglycan endolysin [S. agalactiae bacteriophage B30] (AAN28166)	Cell wall hydrolase (Glyco_25)
20	20 879	ATG	21 394	171	5.90	45 (140)	2e-20	Phage-associated cell wall hydrolase [S. pyogenes MGAS10394] (AAT87679)	Cell wall hydrolase (CHAP)
21	21 720	ATG	22 538	272	9.60	54 (266)	8e-70	Hypothetical protein PEPE_1019 [Pediococcus pentosaceus ATCC 25745] (YP_804517)	Unknown
22	22 571	ATG	22 753	60	9.42	–	–	–	Unknown
23	22 973	ATG	23 731	252	9.17	49 (259)	2e-63	ORF5 [S. thermophilus bacteriophage 7201] (AAF26604)	DnaC homolog (DnaC; COG1484.1)
24	23 731	ATG	23 934	67	10.43	55 (49)	1e-07	Hypothetical protein lmo2272 [Listeria monocytogenes EGD] (CAD00350)	Unknown
25	23 931	ATG	24 101	56	9.81	–	–	–	Unknown
26	24 094	ATG	24 342	82	3.96	32 (83)	5e-04	Hypothetical protein [Temperate phage PhiNIH1.1] (NP_438123)	Unknown
27	24 352	ATG	25 098	248	7.29	35 (265)	7e-35	Conserved hypothetical protein – phage associated [S. pyogenes M1 GAS] (NP_269147)	(RecT; pfam03837)
28	25 113	ATG	26 042	309	5.71	45 (340)	4e-79	Phage protein [Bacteriophage 9429.2] (ABF32000)	(PolC; COG2176.1)
29	26 043	ATG	26 339	98	10.32	39 (92)	3e-08	Hypothetical protein, phage-plasmid associated [S. thermophilus CNRZ1066] (AAV62371)	Unknown
30	26 477	ATG	26 821	114	9.88	29 (120)	4e-05	gp30 [Bacteriophage A118] (NP_463491)	Unknown
31	26 912	ATG	27 193	93	10.38	–	–	–	Unknown
32	27 308	ATG	27 535	75	10.25	–	–	–	Unknown
33	27 541	ATG	27 936	131	10.39	–	–	–	Unknown
34	27 940	GTG	28 419	159	4.74	–	–	–	Unknown
35	28 416	ATG	28 811	131	4.84	87 (110)	1e-48	Putative single-stranded DNA-binding protein [S. mutans] (AAN59480)	Single-stranded DNA binding protein (SSB; pfam00436)
36	29 074	ATG	29 511	145	8.56	40 (156)	1e-24	ORF10 [S. thermophilus bacteriophage 7201] (NP_038311)	Unknown
37	29 504	ATG	29 908	134	7.26	38 (75)	5e-06	gp178 [S. thermophilus bacteriophage Sfi11] (AAF63066)	Unknown
38	29 908	ATG	30 009	33	10.54	–	–	–	Unknown
39	30 367	ATG	30 711	114	4.97	–	–	–	Unknown
40	30 720	TTG	30 911	63	3.92	–	–	–	Unknown
41	31 027	GTG	31 101	24	10.32	–	–	–	Unknown

Structural module of M102

The first ORF located downstream of the M102 cos site, a probable HNH endonuclease, is similar to Orf45 from Streptococcus thermophilus bacteriophage DT1 (Tremblay & Moineau, 1999) (Accession number NC_002072). But in DT1, and also in other Streptococcal and Lactococcal phages, e.g. phages Sfi21 (NC_000872), SM1 (NC_004996) and BK5-T (NC_002796), this ORF is located upstream of the cos site. Orf2 through Orf15 constitute the DNA packaging and morphogenesis module. In general, the ORFs from this module showed high similarity to those of phages from related organisms, e.g. Streptococci, Lactococci and Staphylococci. Most of the structural proteins showed similarity to structural proteins from S. thermophilus phage DT1 (Fig. 2)

Lysis module

The packaging and morphogenesis module is followed by the lysis module. Orf18 showed weak similarity to a putative holin from Streptococcus suis (Table 1). Orf19 contained two glycohydrolase domains, which indicates that this protein could act as endolysin and cleave the glycosidic N-acetylmuramoyl-(β1,4)-N-acetylglucosamine bond of the sugar backbone of peptidoglycan. This is supported by the similarity over the first c. 200 amino acids to muramidases encoded by bacteriophages from low GC gram-positive organisms (Fig. 3), including the lytic enzyme from Streptococcus pneumoniae bacteriophage Cp-1, whose structure has been solved (Hermoso et al., 2003). The acidic residues of Cpl-1 thought to function in catalysis were conserved in Orf19. In general, the C-terminal domains of endolysins specify binding to the cell wall (Loessner, 2005). The C-terminal part of Orf19 showed no similarity to other proteins, except for the C-terminal part of a putative endolysin from Streptococcus pyogenes MGAS10394 (accession number YP_060444).

3

clustalw sequence alignment of Orf19 with bacteriophage lytic enzymes. The amino acid sequence of Orf19 from Streptococcus mutans phage M102 was compared with: Lys44, endolysin from Oenococcus oeni bacteriophage fOg44 (AAD10705.2) (Sao-Jose et al., 2000); Cpl-1, endolysin from Streptococcus pneumoniae phage Cp-1 (NP_044837.1) (Hermoso et al., 2003); LysB30, endolysin from Streptococcus agalactiae bacteriophage B30 (AAN28166.2) (Baker et al., 2006); Lysmv1, endolysin from Lactobacillus bulgaricus bacteriophage mv1 (P33486) (Boizet et al., 1990); Lysadh, lysin from Lactobacillus acidophilus bacteriophage ADH (NP_050170.1)(Henrich et al., 1995). Identical residues in all sequences are in white on a black background. The acidic residues of Cpl-1 proposed to function in catalysis (Hermoso et al., 2003) are indicated. Note that only partial sequences are shown.

The stop codon of Orf19 overlapped with the startcodon of Orf20, which was similar to putative endolysins that appear to be encoded predominantly in bacteriophages from pyogenes Streptococci and Staphylococci (Fig. 4). Orf20 contains a CHAP (cysteine, histidine-dependent amidohydrolases/peptidases) domain, which is present in a variety of peptidoglycan cleaving enzymes with l-muramoyl-l-alanine amidases or d-alanyl-glycyl endopeptidase activity (Bateman & Rawlings, 2003).

4

clustalw sequence alignment of Orf20 with bacteriophage lytic enzymes. Orf20 from Streptococcus mutans phage M102 was compared with: Skl, N-acetylmuramoyl-l-alanine amidase from Streptococcus mitis phage SK137 (CAJ13672) (Llull et al., 2006); PlyTW, N-acetylmuramoyl-l-alanine amidase from Staphylococcus aureus phage Twort (CAA69021) (Loessner et al., 1998); LysWMY, lysin from Staphylococcus warneri M phage φWMY (BAD83402) (Yokoi et al., 2005); Ply187, cell wall hydrolase Ply187 from S. aureus bacteriophage 187 CAA69022 (Loessner et al., 1999). Identical residues in all sequences are in white on a black background, conserved cysteine and histidine residues are indicated by an asterisk. The putative signal sequence in Orf20 is underlined. The complete sequence of Orf20 from M102 is shown, but partial sequences are shown from the other proteins.

Phage M102 therefore might encode two endolysins of different substrate specificity, a muramidase and an amidase or endopeptidase. The presence of two separate endolysins is rather unusual, although in many phages different specificities are combined in one polypeptide. A further peculiarity is the presence of a possible N-terminal signal sequence in Orf20, as predicted by the program signalp (http://www.cbs.dtu.dk/services/SignalP/) (Fig. 4). The genes encoding Orf19 and Orf20 could be expressed in Esherichia coli, but both proteins accumulated in the insoluble fraction, which precluded confirmation of their role in host cell lysis (results not shown).

Replication module

The region downstream from the lysis module most probably constitutes the replication module. Five putative promoters with high similarity to the −35 and −10 regions of the E. coliσ⁷⁰ consensus promoter were identified (Fig. 2). The putative promoters were located in the intergenic regions of the module that encodes proteins required for DNA replication. The sequences of four of these promoters were highly conserved among each other. The DNA replication module also contained three putative Rho-independent termination signals. They were located downstream of ORFs 20, 29 and 38 (Fig. 2). About half of the ORFs from the replication module had no homologs in the database. The ORFs showed no sequence similarity to ORFs from DT1, but some were similar to other Streptococcal bacteriophage ORFs that are implicated in replication (Table 1).

Proteome comparison

The deduced amino-acid sequences of the ORFs encoded by M102 were compared with deduced amino-acid sequences from a collection of Streptococcal and Lactococcal phages (Table 2). Over the complete proteome, bacteriophage M102 was most closely related to bacteriophages 7201, Sfi21, Sfi19 and DT1 from S. thermophilus, which indicates that M102 belongs to the Sfi21-like siphophage group (Rohwer & Edwards, 2002). However, the similarity was largely confined to the similarity between the structural proteins (ORFs 1–18). For the remaining ORFs, phage M102 was more similar to S. pneumoniae phages MM1 and Ej-1 and to S. pyogenes phage PhiNIH1.1.

2

Proteomic comparison of M102 with Streptococcal and Lactococcal phages

			Mean of protdist values
Phage	Host	Accession number	1–41	1–18	19–41
7201	S. thermophilus	NC_002185	4.92	3.14	7.37
Sfi21	S. thermophilus	NC_000872	6.02	3.54	9.45
Sif19	S. thermophilus	NC_000871	6.05	4.05	8.82
DT1	S. thermophilus	NC_002072	6.24	3.52	10.00
BK5-T	L. lactis cremoris	NC_002796	6.36	5.39	7.70
MM1	S. pneumoniae	NC_003050	6.90	7.25	6.41
Lc-Nu	L. lactis cremoris	NC_007501	7.67	6.92	8.69
EJ-1	S. pneumoniae	NC_005294	8.41	9.17	7.37
315.1	S. pyogenes	NC_004584	8.77	8.80	8.73
O1205	S. thermophilus	NC_004303	8.85	8.87	8.83
PhiNIH1.1	S. pyogenes	NC_003157	8.85	10.00	7.26
Sfi11	S. thermophilus	NC_002214	8.99	8.72	9.37
2972	S. thermophilus	NC_007019	9.02	8.31	10.00
Tuc2009	L. lactis cremoris	NC_002703	9.15	10.00	7.97
SM1	S. mitis	NC_004996	9.45	9.55	9.32
bIL170	L. lactis lactis	NC_001909	9.47	9.09	10.00
Cp-1	S. pneumoniae	NC_001825	9.71	10.00	9.30

The sum of all Protdist values was divided by the number of proteins used for the analysis. 1–41: all Orfs; 1–18: Orf1–Orf18; 19–41: Orf19–Orf41.

Determination of the cos site

The cos site of phage M102 was estimated by comparison of the restriction enzyme pattern of heat-treated and nonheat-treated phage DNA. Digestion of nonheat-treated phage DNA with PstI gave one fragment of 6.6 kb, which resolved in two fragments of 1.4 and 5.2 kb upon heating. Digestion with XhoI gave a fragment of 4.0 kb, which became 0.7 kb smaller upon heating. Using a computer-generated restriction map of the sequence, the cos site could be mapped within a region of about 1 kb in size. For a more precise determination, heat-treated phage DNA and ligated phage DNA were directly sequenced using primers that were expected to hybridize closely to and in the direction of the expected cos site. Whereas the ligated phage DNA showed a contiguous sequence, the sequences of heat-treated phage DNA terminated, leaving a gap of 11 nucleotides (Fig. 5). The cos site of M102 thus has a 3′ overhang of 11 nucleotides (5′-ccgcgtgaata-3′). The stretch of the nine first nucleotides of this sequence is 89% identical to the last nine nucleotides from the cos site of Streptococcus mitis prophage SM1 (5′-gtgacggcgtgaa-3′) (Siboo et al., 2003).

5

Determination of the cos site of phage M102. The last base at each termination point of the sequence (T in the upper sequence and A in the lower sequence) is not present in the ligated DNA and caused by the DNA polymerase used for sequencing, which adds an A residue at the 3′ end.

Comparison of bacteriophages M102 and M101

Bacteriophage M101 has the same host strain as M102, but the restriction pattern of M101 genomic DNA is different from that of M102 (results not shown). The lysis cassette and the cos site of M101 were amplified by PCR using primers derived from the M102 sequence sequenced and compared with M102. The lysA nucleotide sequence was 85.2% identical, whereas that of lysB was 82.6% identical. The LysA protein was 91.9% identical whereas the LysB protein was 91.8% identical. The cos sequence of M101 was identical to that of M102 (results not shown).

Conclusions

In conclusion, the genome sequence of M102, which is the first from a bacteriophage that has S. mutans as host, shows high similarity to bacteriophages from S. thermophilus in the morphogenesis module, but less so or not in the lysis and replication modules. The lysis cassette of M102 is unusual in that it contains two lytic enzymes, one of which has probably an N-terminal signal sequence.

Nucleotide accession number

The nucleotide sequence reported here is deposited in the embl database under Accession number AM749121.

Acknowledgements

G. Tiraby is thanked for phages and host strain. The author appreciates the excellent technical assistance of Verena Osterwalder and Steve Reese. Bernard Guggenheim and Rudolf Gmür are acknowledged for support and critical reading of the manuscript, respectively. This work was supported by grant 215 from the Swiss Dental Association SSO.

Author notes

Editor: Oscar Kuipers