Structural diversity and biological significance of lipoteichoic acid in Gram-positive bacteria: focusing on beneficial probiotic lactic acid bacteria.

Bacterial cell surface molecules are at the forefront of host-bacterium interactions. Teichoic acids are observed only in Gram-positive bacteria, and they are one of the main cell surface components. Teichoic acids play important physiological roles and contribute to the bacterial interaction with their host. In particular, lipoteichoic acid (LTA) anchored to the cell membrane has attracted attention as a host immunomodulator. Chemical and biological characteristics of LTA from various bacteria have been described. However, most of the information concerns pathogenic bacteria, and information on beneficial bacteria, including probiotic lactic acid bacteria, is insufficient. LTA is structurally diverse. Strain-level structural diversity of LTA is suggested to underpin its immunomodulatory activities. Thus, the structural information on LTA in probiotics, in particular strain-associated diversity, is important for understanding its beneficial roles associated with the modulation of immune response. Continued accumulation of structural information is necessary to elucidate the detailed physiological roles and significance of LTA. In this review article, we summarize the current state of knowledge on LTA structure, in particular the structure of LTA from lactic acid bacteria. We also describe the significance of structural diversity and biological roles of LTA.

OVERVIEW OF TEICHOIC ACIDS

The cell surface of bacteria comprises the cell membrane and cell wall peptidoglycan as its main components. The cell membrane and cell wall play numerous physiologically relevant roles, such as separation of the intra- and extracellular microenvironments, maintenance of homeostasis, and protection against many environmental stresses. Teichoic acids (TAs) are specific polymers on Gram-positive bacterial cell surfaces and are not found in Gram-negative bacterial cells. The word “teichoic” originates from the Greek word teîkhos (τεῖχος), meaning “wall.” TAs comprise up to 50% of the cell wall dry weight [1, 2]. Thus, they are believed to play important physiological roles.

Two distinct types of TAs, a wall-teichoic acid (WTA) attached to the cell wall and lipoteichoic acid (LTA) anchored to the cell membrane, have been identified (Fig. 1). WTAs were initially discovered by Armstrong et al. in 1958 in cell wall fractions of Lactobacillus plantarum (formerly Lactobacillus arabinosus), Bacillus subtilis, and Staphylococcus aureus [3, 4]. LTAs were identified by Kelemen et al. in 1961 as structurally similar molecules to WTAs in cell membrane fractions [5]. WTA and LTA backbones are generally anionic polymers consisting of repeating polyol phosphate units that, in rare cases, also contain sugar phosphate. In most LTAs, the backbone is comprised of poly-glycerol phosphate (poly-GroP). By contrast, the WTA backbone varies between bacterial species and strains. Typically, it is comprised of poly-GroP or poly-ribitol phosphate (poly-RboP). In WTA, the backbone consisting of repeating units is covalently linked to C-6 of the cell wall N-acetylmuramic acid residue via disaccharide phosphate residues (N-acetylmannosaminyl-N-acetylglucosamine phosphate, N-acetylmannosaminyl-glucosamine phosphate, or glucosyl-N-acetylglucosamine phosphate) as linkage units (Fig. 2) [6]. In the case of LTA, poly-GroP is covalently linked to hexose residues of cell membrane glycolipids (Fig. 2) [1]. The glycolipids connected to the poly-GroP backbone are called glycolipid anchors. A typical glycolipid anchor is a dihexosyldiacylglycerol (Hex2DAG) comprised of two hexose residues [typically glucose (Glc), and/or galactose (Gal)] and diacylglycerol.

Fig. 1.

Overview of the Gram-positive cell surface architecture.

Lipoteichoic acids are anchored to the cell membrane via glycolipids, while the wall-teichoic acids are covalently bound to cell wall peptidoglycan.

Fig. 2.

Typical structures of wall-teichoic and lipoteichoic acids.

In most cases, R is H, d-alanine, and/or hexose. Typical hexose residues include glucose, galactose, and N-acetylglucosamine. GlcNAc, N-acetylglucosamine; MurNAc, N-acetylmuramic acid.

Gram-positive bacteria generally possess both WTAs and LTAs. However, Lactobacillus casei ATCC 334 has only LTAs [2]. LTAs (and related amphipathic macromolecules) have not been found in Bacillus circulans AHU 1363, AHU 1365, and AHU 1646, and Paenibacillus polymyxa (formerly Bacillus polymyxa) AHU 1231 and AHU 1385 [7]. Genetically engineered mutants lacking LTA have been constructed in S. aureus [8] and Lactobacillus acidophilus [9]. These results suggested a dispensable nature of WTA or LTA in Gram-positive bacteria. Alternatively, other macromolecules in the cell surface may fill in for TA functions.

Structural information on WTAs and LTAs from a single strain is available in a few cases, and structural variation of WTAs is greater than that of LTAs [10]. It appears that the physiological roles of LTAs are different from those of WTAs. Polarity constitutes the major difference between these polymers. In contrast to the hydrophilic nature of WTAs, LTAs are amphipathic molecules comprising hydrophilic poly-GroP and a hydrophobic glycolipid anchor. In general, the free hydroxyl groups in GroP and RboP repeating units are often substituted by d-alanine (d-Ala), Glc, Gal, and/or N-acetylglucosamine (GlcNAc) (Fig. 2). d-Ala is more frequently found as a substituent than other compounds. The structural diversity of TAs is mainly associated with the types of substituents and substitution ratios of the repeating units [7, 11, 12].

Phosphate residues of the repeating units, in addition to those of cell membrane phospholipids, impart a negative charge to the cell surface. On the other hand, d-Ala, which partly substitutes hydroxyl groups in the repeating units, imparts a positive charge. TAs are therefore zwitterionic polymers. Partial and heterogeneous d-Ala substitutions affect the distribution of charge, hydrophobicity, and TA stereostructure. In particular, positive charges of d-Ala residues are involved in the reduction of negative charge of the cell surface. TAs play very important roles in bacterial physiology, e.g., preservation of divalent cations, including Mg2+, for growth [13], maintenance of proton gradient across the cell membrane for energy metabolism [14], and protection against cationic antimicrobial peptides via three-component peptide-sensing systems [15]. WTA is involved in protecting peptidoglycan from bacteriolysis by lysozyme [16] and in the control of lytic enzyme localization during cell division [17]. LTA is also involved in the progression of normal cell division [18]. Cell wall glycopolymers and also LTA, including those derived from lactic acid bacteria, are receptors for bacteriophages [19].

STRUCTURAL DIVERSITY IN LTAS

Structural information is available for LTAs from many bacteria. However, most of the available information is limited to specific bacterial genera/species, including non-opportunistic or opportunistic pathogens: Bacillus spp., Clostridium spp., Enterococcus spp., Listeria spp., Staphylococcus spp., and Streptococcus spp. (Table 1). Little is known about the LTA structure in probiotic and related bacteria (Fig. 3). LTA is a candidate immunomodulatory molecule not only in pathogenic bacteria but also in probiotic and commensal bacteria. Thus, structural information concerning LTA derived from probiotics is important. To the best of our knowledge, structures of both the repeating unit and glycolipid anchor of LTA have been identified in 91 strains from 11 genera/53 species (Tables 1 and 2, Fig. 3). In addition, structural information concerning unspecified strains and/or partial structures of either the repeating unit or the glycolipid anchor have also been reported. Typical LTA structures in most of these bacteria comprise GroP-repeating units as the backbone, with d-Ala, hexose, and/or hexosamine residues as substituents, and a glycolipid Hex2DAG anchor unit (Fig. 2). In the following sections, the diversity of LTA structures in bacteria other than Lactobacillus spp. (Table 1) and Lactobacillus strains (Table 2) will be presented.

Table 1.

Structures of lipoteichoic acids from Gram-positive bacteria

Fig. 3.

Bacterial species whose complete lipoteichoic acid structures are known.

Details of lipoteichoic acid of such beneficial lactic acid bacteria as Lactobacillus spp., Lactococcus spp., and Leuconostoc spp. are relatively sparse compared with non-opportunistic and opportunistic pathogens.

Table 2.

Structures of lipoteichoic acids from Lactobacillus spp.

Bacterial species (former name)	Strain name	Glycolipid anchor structure	Repeating unit structure (number of units)	Substituent (substitution ratio)	Extraction method	Reference
Lactobacillus brevis	ATCC 8287	Unknown	Gro-P	Ala,	BuOH	[47]
				Glc,
				AlaGlc
Lactobacillus casei	BL23	Unknown	Gro-P (42)	Ala (64%)	BuOH	[56]
Lactobacillus delbrueckii subsp. lactis	ATCC 15808	Glc-Glc-Glc-DAG,	Gro-P (29–37)	Ala (21–27%),	Phe/H2O	[90]
		Glc-Glc-AcylGlc-DAG		Glc (3%)
	Ads-5	Glc-Glc-Glc-DAG,	Gro-P (31–36)	Ala (42–49%)	Phe/H2O	[90]
		Glc-Glc-AcylGlc-DAG
	LL78	Glc-Glc-Glc-DAG,	Gro-P (30–34)	Ala (25–28%),	Phe/H2O	[90]
		Glc-Glc-AcylGlc-DAG		Glc (26–27%)
Lactobacillus gasseri	JCM 1131T	(Gal-Gal-Gal-Glc)*1-DAG,	Gro-P (20–30)	Ala (31%)	BuOH	[49]
		(Gal-Gal-Gal-Glc-Acyl)*1,2-DAG
Lactobacillus helveticus	DSM 20075T	Glc-Gal-Glc-DAG,	Gro-P (24)	Ala (57–64%)	Phe/H2O	[43]
		Glc-Gal-AcylGlc-DAG
Lactobacillus pentosus (Lactobacillus plantarum)	DSM 20314T	Glc-Gal-Glc-DAG,	Gro-P (22)	None	Phe/H2O	[32, 42]
		Glc-Gal-AcylGlc-DAG
Lactobacillus plantarum	JCM 1149T	Hex-Hex-Hex-DAG,	Gro-P (110)	Ala (42%),	BuOH	[91]
		(Hex-Hex-Hex-Acyl)*2-DAG		Glc (10%)
	L-137	Hex-Hex-Hex-DAG,	Gro-P (96)	Ala (50%),	BuOH	[91]
		(Hex-Hex-Hex-Acyl)*2-DAG		Glc (2%)
	KCTC 10887BP (K8)	Glc-Gal-Glc-DAG,	Gro-P	Ala,	BuOH	[46, 48]
		Glc-Gal-AcylGlc-DAG		Glc,
				Gal
	NCIMB 8826	Unknown	Gro-P	Ala (42%)	BuOH	[66]
Lactobacillus reuteri	100-23	Unknown	Gro-P (20)	Ala (74–79%),	BuOH	[92]
				Glc (6%)
Lactobacillus rhamnosus (Lactobacillus casei)	DSM 20021T	Glc-Gal-Glc-DAG,	Gro-P (40)	None	Phe/H2O	[32, 42, 93]
		Glc-Gal-AcylGlc-DAG
	GG (ATCC 53103)	Glc-Gal-Glc-DAG	Gro-P (30–78)	Ala (72–74%)	BuOH	[50, 51]
Lactobacillus sakei	KCCM 11175P (K101)	Hex-Hex-DAG,	Unknown	Unknown	BuOH	[48]
		(Hex-Hex-Acyl)*2-DAG

Ala: alanine; AlaGlc: alanyl-glucose; DAG: diacylglycerol; Gal: galactose; Glc: glucose; Gro: glycerol; Hex: hexose; P: phosphate; BuOH: butanol; Phe: phenol.

*1The order of Gal and Glc is unknown.

*2The linkage position of the hexose-bound acyl group is unknown.

i) Gram-positive bacteria other than Lactobacillus spp.: d-Ala, Glc, Gal, and GlcNAc residues are generally present as C-2 hydroxyl group substituents of the GroP-repeating unit (Fig. 2), but other sugars can also be present, albeit less frequently. Glc oligosaccharides, including di-, tri-, and tetrasaccharides, are found in some Enterococcus spp. (formerly Streptococcus spp.) [20,21,22,23,24,25,26,27,28,29] and Streptococcus sanguinis DSM 20567T [21, 29, 30] and DSM 20068 [30] (Table 1). On the other hand, unique repeating units other than GroP have also been reported. A Gal-Gal-GroP-repeating unit was detected in Lactococcus garvieae (formerly Streptococcus lactis) [22, 27, 29, 31,32,33,34]. Several rare repeating units have been identified in Clostridium difficile [35], Clostridium innocuum ATCC 14501T [34], Streptococcus oralis Uo5 [36], and Streptococcus pneumoniae R6 (ATCC BAA-225) [34, 37,38,39,40] and R36A (ATCC 12214) [37] (see Table 1 for detailed structures).

Dihexosyl glycerol is a typical saccharide moiety of the LTA glycolipid anchor (Fig. 2). However, some Streptococcus spp. use monohexosyl glycerol [36, 37, 41], and trihexosyl glycerol has been detected in C. difficile [35] (Table 1). Glc and Gal are the most commonly found glycolipid anchor residues in LTA in Gram-positive bacteria. In Bacillus coagulans AHU 1366 and AHU 1634 [11] and Bacillus megaterium ATCC 14581T [29, 42], AHU 1373, and AHU 1375 [11], GroP polymer directly binds to mono- or diacylglycerol; that is, no hexose residues intervene between the repeating units and the lipid anchor (Table 1). Generally, the glycolipid anchor has two acyl groups. However, dihexosylmonoacylglycerol has been reported in some Bacillus spp. and Paenibacillus thiaminolyticus (formerly B. subtilis) [7]. Other bacteria, such as Lactococcus spp. [22, 27, 29, 31,32,33,34, 42, 43] and Leuconostoc mesenteroides [24, 42], have a third acyl group attached to a hexose residue of the glycolipid anchor (Table 1). This type of glycolipid anchor is termed acyldihexosyldiacylglycerol (AcylHex2DAG). A glycolipid anchor that contains phosphatidic acid with two acyl groups attached to the hexose residue (i.e., with four acyl groups per LTA molecule) has been reported in Enterococcus faecalis DSM 20371 (Kiel 27738) (formerly Streptococcus faecalis) [22, 24], Enterococcus hirae ATCC 9790T (NCIB 8191T) (formerly E. faecalis, S. faecalis, or Streptococcus faecium) [22,23,24,25, 27], and many Listeria monocytogenes strains [44, 45] (Table 1).

ii) Lactobacillus strains: Lactobacillus spp. typically comprise beneficial probiotic bacteria. To the best of our knowledge, LTA structures have been reported for 16 strains from 10 species. Information regarding both repeating unit and glycolipid anchor structures is available in 11 out of 16 strains (Table 2).

The Lactobacillus spp. LTA repeating unit consists of a typical poly-GroP backbone with d-Ala as a common substituent. In most cases, Glc is also found as another substituent together with d-Ala. In the case of L. plantarum KCTC 10887BP (K8), Gal has been detected in addition to d-Ala and Glc [46]. One unusual exception is d-Ala at C-6 of Glc in Lactobacillus brevis ATCC 8287 [47] (Table 2). Aminosugars, such as GlcNAc, have not been identified in a substituent of the GroP-repeating unit. It is interesting that all Lactobacillus spp., except for Lactobacillus sakei KCCM 11175P (K101) [48], do not possess a typical dihexosyl glycerol but, instead, have trihexosyl glycerol as the glycolipid anchor saccharide moiety (Tables 2 and 3). Moreover, we recently found that Lactobacillus gasseri JCM 1131T, an intestinal lactic acid bacterium, has tetrahexosyl glycerol comprising Gal and Glc at a molar ratio of 3:1 [49] (Table 2). This is the first demonstration of a tetrasaccharide in the LTA glycolipid anchor. As L. gasseri is often commercially employed as a probiotic, the prevalence of this novel anchor structure in L. gasseri and its related species need to be investigated.

Table 3.

General architecture of lipoteichoic acids from beneficial probiotic lactic acid bacteria

All the reported Lactobacillus spp. LTA glycolipid anchor structures, except for Lactobacillus rhamnosus GG (ATCC 53103) [50, 51], contain not only the typical HexxDAG but also AcylHexxDAG, with a hexose-attached third fatty acid residue (Table 2). Interestingly, AcylHex2DAG is also found in Lactococcus spp. and L. mesenteroides, as described above (Table 1). Therefore, a glycolipid structure containing three acyl groups may be characteristic for beneficial probiotic lactic acid bacteria, and it might be defined as the lactic acid bacteria-specific LTA (Table 3). However, a glycolipid anchor with three acyl groups has not been found in L. rhamnosus GG LTA [50, 51]. A more detailed structural analysis might be required before we can conclude that AcylHexxDAG is absent from strain GG. It should also be noted that Lactobacillus spp. are commonly equipped with glycolipid anchors comprising trisaccharides or tetrasaccharides (for the moment, tetrasaccharides are found only in L. gasseri JCM 1131T) (Table 2). Structural characteristics of Lactobacillus spp. glycolipids (Table 3) may potentially affect physiological properties of the cell surface.

The significance of the structural diversity of LTA in Gram-positive bacteria remains unclear. Current knowledge about the structural characteristics of LTA raises interesting questions regarding the relationship between the LTA structural diversity in pathogenic and beneficial bacteria and microbial virulence, pathogenicity, and probiotic functions. However, more data on the structural and biological characteristics of LTA from a broader range of species and strains are required to answer these questions.

FACTORS INFLUENCING LTA STRUCTURE

i) Environmental factors: Environmental stresses imposed by different growth conditions may affect the structural diversity of LTA. For example, a microarray analysis revealed a significant activation of the Staphylococcus epidermidis 1457 dlt operon, encoding genes for d-Ala substitution of TA, after exposure to human cationic antimicrobial peptide β-defensin 3 [15]. A Lactococcus lactis mutant defective in d-Ala substitution (ΔdltD mutant) displayed increased susceptibility to a cationic antimicrobial peptide nisin, and to lysozyme in comparison with the parental strain, L. lactis MG1363, while a dltD-overexpressing mutant showed increased resistance to nisin [52]. Thus, these reports indicate that d-Ala substitution in the repeating unit plays a role in stress resistance. Furthermore, it is thought that the ratio of free hydroxyl groups to d-Ala residues varies not only at species and strain levels but also in response to growth conditions, including pH [53], NaCl concentration [54], and temperature [55]. The number of repeating units and the d-Ala substitution ratio decreased under NaCl-exerted osmotic stress in L. casei BL23 LTA [56]. On the other hand, such functions of hexose substituents and the effect of growth conditions on the degree of hexose substitution have not been reported.

ii) Extraction procedures: Hot phenol/water extraction has been used for many years as a typical LTA extraction procedure [57, 58]. A similar hot phenol/water extraction procedure constitutes a conventional extraction procedure for lipopolysaccharides, amphipathic glycolipids of the cell surface of Gram-negative bacteria. The extraction yields an LTA- and nucleic acid-containing water phase, while the phenol phase contains denatured proteins and residual insoluble material. LTAs are further purified by anion exchange chromatography and/or hydrophobic interaction chromatography to eliminate contaminants, such as nucleic acids. Butanol extraction with 1-butanol was proposed in 2001 [59], and this method recently became a common method of LTA extraction. Butanol extraction relies on the polarity of LTA molecules, similar to hot phenol/water extraction, but the principles of butanol extraction are not understood in detail. Whereas careful attention is necessary for handling of phenol because of its toxicity, butanol has the advantage of easy handling. Furthermore, hot phenol may cause partial LTA destruction, such as degradation of the GroP polymer and elimination of the substituents. Morath et al. compared chemical structures of S. aureus DSM 20233 LTA prepared by phenol or butanol extraction methods [59]. Compared with LTA extracted with butanol, LTA polymer extracted with phenol was shorter, with less than ten GroP-repeating units and a reduced number of d-Ala and GlcNAc substitutions. In particular, the degree of d-Ala substitutions in phenol-extracted LTA was less than half that in butanol-extracted LTA. Thus, butanol extraction can yield a less damaged LTA than the hot phenol/water extraction procedure. Lines of evidence indicate that the extraction procedures clearly affect the results of analysis of the LTA chemical structure. Numerous determinations of LTA chemical structures have been performed with LTA preparations obtained by the conventional hot phenol/water method (Tables 1 and 2). Importantly, commercial LTA preparations have been shown to be inhomogeneous and decomposed. Furthermore, significant amounts of contaminants having immunostimulatory activity are present in the preparations [60]. Thus, the immunomodulation of commercial LTA preparations is quite different compared with that of butanol-extracted LTA. Therefore, comparison of LTA structure between species or strains requires careful consideration, taking into account structural alteration and degradation associated with the different extraction protocols.

LTA-MEDIATED BACTERIA-HOST INTERACTIONS

i) LTA-mediated host adhesion: The interaction between LTA and the host is important, and LTA can act as an adhesion molecule. LTA is involved in the adhesion of Lactobacillus johnsonii La1 to human intestinal epithelium Caco-2 cells [61]. Cell-free LTA of Streptococcus pyogenes can modulate the attachment of bacterial cells to the host cell surface through cross-linking between a bacterial M protein and host fibronectin [62]. Thus, fibronectin is a candidate host LTA receptor. Recently, Baur et al. reported that a nasal epithelial cell scavenger receptor expressed by endothelial cell-I (SREC-I) was a host receptor for S. aureus, binding the WTA GroP polymer [63]. They also showed that colonization of the rat nasal cavity by S. aureus was inhibited by an anti-SREC-I antibody. Therefore, the GroP polymer, found in both WTA and LTA, might play an important role in bacterial colonization of the host.

ii) Host receptors and LTA immunomodulation: The structural heterogeneity of LTA was suggested to impact host immune response. Host cells recognize LTA via Toll-like receptor 2 (TLR2), a pattern recognition receptor for pathogen-associated molecular patterns that transduces cellular signals to induce proinflammatory cytokines [64,65,66,67,68]. However, the reports concerning immunomodulating activities of LTA via TLR2 have been contradictory. It has been reported that LTA does not induce TLR2-mediated cytokine production [69], and conflicting reports on LTA cytokine-inducing activity have been published. Suda et al. reported that, in conventional E. hirae ATCC 9790T LTA preparations, the cytokine-inducing activity fraction can be separated from LTA by a combination of hydrophobic interaction and anion-exchange chromatography [70]. Hashimoto et al. verified that purified E. hirae ATCC 9790T LTA has no cytokine-inducing activity [20], and the authors concluded that the activity may have been due to the contaminating lipoproteins in such conventional LTA preparations [71, 72]. They confirmed that LTA obtained from a lipoprotein-deficient mutant (Δlgt mutant) had no detectable TLR2 cytokine-inducing activity [69]. The controversy surrounding the cytokine-inducing activity of LTA is still debated [73,74,75]. Most likely, the interaction between LTA and host cells is very complicated. To clarify it, it is necessary to unify the experimental materials and conditions employed, such as the LTA preparations (LTA from pathogenic or beneficial bacterial or chemically-synthesized LTA), contaminating molecules of the LTA preparations, host immune cells (whole blood, peripheral blood mononuclear cells, dendritic cells, or cell lines originating from macrophages, monocytes, and intestinal epithelial cells), and target molecules for measurement (NF-κB activation and production of IL-1β, IL-8, TNF-α, IL-10, IL-12, and IFN-γ). Several candidates for LTA receptors other than TLR2 have been reported: a lipopolysaccharide-binding protein and CD14, both of which are involved in lipopolysaccharide recognition [21]; a mannose-binding protein [76]; l-ficolin [77]; a type I macrophage scavenger receptor, which is expressed by phagocytes [78]; and paired immunoglobulin-like receptor-B, which is expressed by many immune cells [79]. All these reports suggest that LTA significantly contributes to host immune modulation during bacteria-host interactions. In the future, unambiguous details of the interaction of LTA with human immune response should be understood for application of bacteria as synbiotics.

iii) Structural requirements of LTA for modulation of the host immune response: The relationship between LTA structure and the host immune response has been investigated. The importance of d-Ala substitution of GroP-repeating units for cytokine induction in vitro has been reported [51, 80, 81]. Deininger et al. evaluated the minimum structural requirements of LTA for cytokine-inducing activity using chemically-synthesized LTA. More than three GroP-repeating units with d-Ala substitutions were required for the induction of proinflammatory cytokines [81]. Different inflammatory responses including proinflammatory cytokine production were induced in vivo by d-Ala substitution-deficient mutants (dlt operon mutants) as compared with those induced by the parental strain L. plantarum NCIMB8826 [66]. Smelt et al. constructed an L. plantarum WCFS1 ΔdltX-D mutant defective in d-Ala substitutions of the GroP-repeating units. Mutant immunomodulatory activities, especially TLR2-dependent NF-κB activation in vitro and differentiation of dendritic cells and T-cell populations in vivo, were different from those of the parental strain [82]. Acyl groups of the glycolipid anchor are also considered to be important for the immunomodulatory activity of LTA. It was reported that L. plantarum KCTC 10887BP LTA, but neither heat-inactivated cells nor peptidoglycan, inhibited Pam2CSK4-induced IL-8 expression and that d-Ala substitutions and lipid moieties of the LTA are required for the agonistic activity [83]. Cytokine-inducing activity was altered by elimination of acyl groups from LTA extracted from L. rhamnosus GG [51] and S. aureus DSM 20233 [84]. Fatty acid residues (i.e., molecular species, residue number) vary with each LTA molecule. Lines of evidence indicate that LTA is a cytokine-inducing factor of intestinal Gram-positive bacteria, and heterogeneous LTA structures are potentially a key factor in host immunomodulation. On the other hand, it was also reported that d-Ala substitutions of LTA GroP-repeating units [50] and the glycolipid anchor [81] are not important for the induction of cytokines. The cytokine-inducing activity of defined structural elements of LTA has to be clarified. Details of the LTA recognition mechanism by the host will reveal the significance of LTA structural diversity in bacterial-host interactions.

Recently, an LTA-deficient L. acidophilus mutant lacking a phosphoglycerol transferase gene (LBA0447) was constructed by a double-crossover gene replacement strategy [9]. The parental L. acidophilus NCFM strain and LTA-deficient mutant were examined in a mouse model of dextran sulfate sodium (DSS)-induced colitis [9]. When a viable LTA-deficient L. acidophilus mutant was administered orally before the administration of DSS, DSS-induced colitis was significantly suppressed compared with the effect of parental strain administration. Administration of LTA-deficient mutant cells also facilitated the resolution of inflammation of the DSS-induced colitis. Reduced production of proinflammatory cytokines IL-12 and TNF-α and enhanced production of the anti-inflammatory cytokine IL-10 were observed in dendritic cells derived from mice inoculated with the LTA-deficient mutant. It is suggested that the suppression of inflammation in mice inoculated with LTA-deficient L. acidophilus was caused by an altered induction of cytokines. In addition, regulation of the LTA-deficient mutant-induced IL-10 production was suggested to be mediated by the Erk1/2 mitogen-activated protein kinase signaling pathway [85], and LTA-deficient mutant administration resulted in increased numbers of regulatory dendritic cells and activated regulatory T-cells (FoxP3+ Tregs) [86]. Findings from experiments with the LTA-deficient mutants strongly suggest that LTA affects host immune response. However, the detailed mechanism of host-LTA interaction remains to be elucidated. Noh et al. showed that L. plantarum KCTC 10887BP LTA inhibited Pam2CSK4-induced IL-8 expression more potently than LTAs from S. aureus, E. faecalis, and Streptococcus mutans [83]. The difference in immunomodulatory effects between Lactobacillus spp. LTA and other pathogenic bacterial LTAs is interesting when we consider the structural characteristics of Lactobacillus spp. LTA (Tables 2 and 3). Thus, information on the LTA structure might provide a solution to the problem; for example, large numbers of hexoses and acyl groups in the glycolipid anchor and no aminosugar substitution in GroP-repeating units.

CONCLUSION

LTA is regarded as an important cell surface molecule of Gram-positive bacteria, with roles in bacterial physiology and bacterial interaction with the host. Data on the LTA chemical structure, extraction procedures, and LTA immunomodulatory activities are accumulating, and detailed physiological and biological roles of LTA are increasingly understood. On the other hand, numerous questions have been raised. For example, questions about how and why the LTA structural diversity is generated and about the significance of LTA structural diversity for bacterial physiology and host interactions. Full knowledge of LTA chemical structures and biological activities has to be obtained before these questions can be answered.