T. forsythia was first isolated in the mid-1970s from subjects with progressing advanced periodontitis and described as ‘fusiform Bacteroides’ by Tanner et al. [8]. Initially, its taxonomic affiliation was unclear because it did not resemble described species of oral or enteric Gram-negative anaerobic rods [9]. The phylogeny of oral Bacteroides species in the Cytophaga–Flavobacterium–Bacteroides family was reorganized after Bacteroides forsythus had been described and eventually clarified in the phylogenetic studies comparing 16S rRNA sequence data [10]. Subsequently, B. forsythus was affiliated to the genus Tannerella [11]. Here it was first formally classified to Tannerella forsythensis but then reclassified to T. forsythia [12].

Recent taxonomic analyses have revealed that Bacteroides, Porphyromonas and Tannerella are all contained in the order Bacteroidales and that Porphyromonas and Tannerella are phylogenetically even more closely related, as both are affiliated to the Porphyromonadaceae family [13].

A preliminary rapid identification of human-derived T. forsythia strains can be based on the following eight criteria [14]: positive activity for (i) a‑glucosidase; (ii) b-glucosidase; (iii) sialidase; (iv) trypsin-like enzyme; (v) negative indole production; (vi) requirement for N‑acetylmuramic acid; (vii) colonial morphology; and (viii) Gram-stain morphology from blood agar medium deficient in N-acetylmuramic acid.

The full genome sequence of T. forsythia ATCC 43037 is available through the Oral Pathogen Sequence Databases at Los Alamos National Laboratory Bioscience Division [15]. The genome consists of 3,405,543 base pairs with 3,034 predicted open reading frames.

T. forsythia meets the criteria for periodontal pathogens postulated by Socransky [16] and Socransky et al. [17], (i) because this bacterium is present in increased levels in periodontitis [17]; (ii) there is evidence for host response to its antigens [18,19,20]; (iii) it is able to cause disease in animal models [21,22,23]; and (iv) it expresses virulence factors that can potentially contribute to the disease process [24]. Among the characterized T. forsythia virulence factors related to the field of glycobiology are a sialidase [25], an a-D-glucosidase and an N-acetyl-b-glucosaminidase [26], as well as the glycosylated surface (S-) layer [27,28,29].

Despite the growing evidence implicating this bacterium in the pathogenesis of periodontitis, it is up to now a rather poorly studied organism, a fact that can be attributed to its fastidious growth requirements as well as to the lack of molecular tools for genetic manipulation [24].

Recent evidence suggests that glycobiology is important for defining the life of the periodontal pathogen T. forsythia. This review compiles the current state of knowledge about the glycobiology aspects of T. forsythia, with relevance for the bacterium’s cellular integrity, its life-style and metabolism as well as its virulence potential.

Many bacteria from all phylogenetic lineages are covered by regularly arrayed superficial layers, termed S-layers. The protomeric units of S-layers usually consist of high-molecular-mass proteins or glycoproteins [30,31]. Particularly in Gram-negative bacteria it has been difficult to isolate these structures in a form which truly represents what might occur in the living organism, although some attempts have been made [32,33,34]. Different functions, such as protective covering, molecular sieve and ion trap, phage receptor, providing an adhesion and surface recognition mechanism, as well as involvement in mediation of virulence, have been attributed to S-layers [30,31].

About 25 years ago, Kerosuo [35] reported for the first time on the ultrastructure of the T. forsythia ATCC 43037^T S-layer. Later, Sabet et al. [27] published their findings regarding the isolation, purification, and initial studies on the virulence potential of the S-layer from T. forsythia strains. SDS-PAGE analysis revealed the presence of two high molecular-mass, glyco-positive protein bands [36,37,38], which were later confirmed by our research group to be the components of the S-layer [39]. The ~135-kDa Tannerella S-layer protein TfsA and the ~152-kDa S-layer protein TfsB are encoded by the genes tfsA (TF2661-2662) and tfsB (TF2663), respectively, which are co-transcribed from a single promoter [38]. The two S-layer proteins share 24% amino acid similarity. They do not show overall homology to any other S-layer protein sequence deposited in databases, except for their C‑terminal regions, which have profound similarity to putative S‑layer glycoproteins of the phylogenetically closely related bacterium Bacteroides distasonis [40]. The Tannerella S‑layer proteins exhibit C‑terminal sequence similarity to the CTD (C-Terminal Domain) family proteins of Porphyromonas gingivalis which supports the assumption of a novel CTD Bacteroidales secretion pathway [41,42].

Our interest in the S-layer of T. forsythia was aroused by the facts that (i) it represents the first glycosylated S-layer of a Gram-negative organism and (ii) it is structurally unique due to the simultaneous presence of two S-layer proteins [38]. Interestingly, in periodontal lesions, another S‑layer-carrying bacterium, namely Campylobacter rectus is found. This organism is thought to be capable of inducing pro-inflammatory cytokines and its S-layer may temper this response to facilitate the survival of C. rectus at the site of infection [43].

Recently, Sekot et al. [39] extended the previous structural characterization of the T. forsythia S‑layer [35] by a combined ultrastructural/immunological approach, showing that the two S-layer glycoproteins TfsA-GP and TfsB-GP are intercalated to form a monolayer on the cell surface of T. forsythia. The lattice spacing for the square S-layer lattice was determined to be 10.1 ± 0.7 nm (Figure 1) in freeze-etching, freeze-drying, and negative staining experiments, and a thickness of the layer of approximately 22 nm was found in ultrathin sections.

In that study, sheared flagella with intact hook regions have been identified by freeze-etching [39], which somehow challenges the description of T. forsythia as a non-motile species [44]. However, their role in bacterial motility and a possible impact on co-aggregation of T. forsythia with other species in the biofilm is still unclear. In this context it should also be mentioned that according to a recent partial annotation of the open reading frames of the T. forsythia genome, homologous genes for hook and other pilus/flagella forming units have not been identified [45].

As mentioned above, Lee et al. [38] inferred from SDS-PAGE that the two T. forsythia S-layer proteins are glycosylated. However, no further compositional or structural details were provided. In our laboratory we focused the efforts on the characterization of the glycoproteome of T. forsythia.

By SDS-PAGE, glycosylation of the two S-layer proteins of T. forsythia, the 153-kDa TfsA-GP and the 180-kDa TfsB-GP (calculated molecular masses), could be confirmed. In-gel reductive b-elimination of purified S-layer glycoproteins was employed to release the glycans from the protein backbone [29]. Mass spectrometric analysis revealed a dominant glycan structure of 1,621 Da. Upon CID fragmentation analysis a hetero-oligomer consisting of eight different sugar residues was observed. Mass increments for one pentose, one deoxyhexose, three uronic acids (modified or free), one methylhexose, and one reduced hexose in addition to one, yet non-described, 361-Da sugar residue were identified. Further, some of the oligosaccharides were substituted by one dideoxyhexose and one additional deoxyhexose. Interestingly, the glycans released from either S-layer glycoprotein showed identical glycan mass profiles, indicating the presence of a uniform glycosylation pattern.

NMR spectroscopy of the glycan samples was used to identify the methyl-hexose as methyl-galactose and the two N-acteylhexosaminuronic acid residues as N-acetylmannosaminuronic acid and as O-methyl-N-acetylmannosaminuronic acid, respectively. The dideoxyhexose that was already observed in the MS/MS fragmentation profile was identified as a digitoxose residue. The identity of the 361-Da sugar residue, however, remained unclear even after high resolution MS/MS and NMR analysis of the intact O-glycan. In a screen of nucleotide-activated sugars from the bacterial cytoplasm large amounts of CMP-activated substance matching the 361-Da unit were observed. This was indicative for the presence of an a-keto sugar. Combined NMR analyses of the purified DMB (1,2-diamino-4,5-methylene dioxybenzene dihydrochloride)-labelled 361-Da sugar residue and of the complete oligosaccharide revealed the presence of a non‑2-ulosonic acid carrying two substituents on the amino functions at carbons 5 and 7. Comparison of all recorded NMR spectroscopic data of the O-glycan with those of the isolated C₁₄H₂₅O₉N₃ compound showed good concordance. The configuration of the non-2-ulosonic acid has been proven by the few detectable coupling constants and NOEs and is in better accordance with those of a pseudaminic acid residue than with those of legionaminic acid. The whole C₁₄H₂₅O₉N₃ compound can hence be considered as Pse5Am7Gc, where Am is an amidinyl and Gc a glycolyl group [29]. The structure of the highly complex decaglycan of T. forsythia including the so far determined glycosidic linkages is given in Figure 2. The heterosaccharide is O‑glycosidically linked via the reducing-end a‑galactose to serine and threonine residues of both the TfsA and TfsB S‑layer proteins. Interestingly, all identified glycosylation sites match the D(S/T)(A/I/L/M/T/V) three-amino acid motif that was recently described for the general protein O-glycosylation system in B. fragilis [46].

Carbohydrate-stained SDS-PAGE gels of T. forsythia whole cell lysates indicated, besides the S-layer glycoproteins, the presence of several other glycosylated proteins in the molecular mass range between 60 and 250 kDa [29]. Some of these bands were also reactive with the fucose-specific Aleuria aurantia lectin, which readily detects the S-layer glycan. Upon closer examination, all bands were shown to carry the same glycosylation profile as the S-layer glycoprotein bands, with slight variations detected in their relative ratios and in O-methylation of the mannosaminuronamide residue. This demonstrates that the S-layer O-glycosylation system is also involved in the glycosylation of other T. forsythia proteins [29].

Among these proteins are the predicted outer-membrane proteins TF2339 and its paralog TF1259, as well as the predicted lipoproteins TF1056 and TF0091. The former two proteins exhibit similarity to the CTD family of P. gingivalis [41,42]; the latter show similarity to TonB-dependent receptor associated proteins [47]. Interestingly, all of these glycoproteins are antigenic upon probing with an antiserum raised against a T. forsythia outer membrane preparation [45]. The finding that several abundant proteins in T. forsythia are modified with the S‑layer glycan is fueled by the recent identification of a rich outer membrane glycoproteome in T. forsythia [45].

Initial information about glycosylation in T. forsythia was provided by the description of a so-called exopolysaccharide operon [48], of which WecC (TF2055) coding for a predicted UDP-N-acetylmannosaminuronic acid dehydrogenase is part of. Closer inspection of that genomic region revealed the presence of a 6.8‑kb gene locus of T. forsythia spanning TF2055-TF2049, encoding in addition to WecC (TF2055), a predicted UDP-N-acetylglucosamine 2-epimerase (NeuC, TF2054), three predicted glycosyltransferases (TF2053; TF2050; TF2049), a predicted acetyltransferase (TF2052) and one ORF with yet unassigned function (TF2051) [29].

Since deletion of TF2055 caused truncation of S-layer protein glycans [48] by lacking the 809-Da Pse-containing trisaccharide side branch [29], it is evident that this genomic region carries crucial information for proper O‑glycan assembly (see Section 3.2. and Section 6.2.). Interestingly, similar glycosylation loci are also present in other phylogenetically related species, including, for instance, Bacteroides fragilis NCTC 9343, Bacteroides thetaiotaomicron VPI 5492, Bacteroides uniformis ATCC 8492, Porphyromonas gingivalis ATCC 33277, and Parabacteroides distasonis ATCC 8503.