| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Western Dairy Center,
Department of Nutrition and Food Sciences and Department of
Biology Utah State University, Logan 84322-8700
Department of Microbiology, Weber State University, Ogden UT 84408-2506
|| Département de Biochimie et de Microbiologie Université Laval, Québec, Canada, G1K 7P4
Corresponding author:
J. R. Broadbent; Tel: (435) 797-2113, Fax: (435) 797-2379, e-mail:
broadbnt{at}cc.usu.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: Streptococcus thermophilus • exopolysaccharide • lactic acid bacteria
Abbreviation key: CPS = capsular exopolysaccharide, CPS+/-= ability (+) or inability (-) to produce capsular exopolysaccharide, EPS = exopolysaccharide, EPS+/-= ability (+) or inability (-) to produce capsular or secreted exopolysaccharide, LAB = lactic acid bacteria
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Production of EPS has not been shown to confer any obvious advantage to growth or survival of S. thermophilus in milk. However, in situ EPS production by this species or other LAB starters typically imparts a desirable "ropy" or viscous texture to fermented milk products (Hassan et al., 1996a, 1996b; Laws and Marshall, 2001). Because EPS+ strains of S. thermophilus and Lactobacillus delbrueckii subsp. bulgaricus can reduce syneresis and enhance product texture and viscosity, these types of cultures are commonly used as a substitute for commercial stabilizers in yogurt manufacture (Cerning, 1995). Although use of EPS+ LAB starters is far less common in cheese manufacture, recent work has shown that EPS+ S. thermophilus can improve the functional properties of lowfat or part-skim Mozzarella cheese (Perry et al., 1997; Low et al., 1998; Petersen et al., 2000). Biotechnological advances in molecular and genetic analyses of bacterial EPS have sparked renewed interest in EPS production by S. thermophilus and other LAB. As understanding of the genetics, physiology, and functionality of LAB EPS continues to improve, novel applications for polysaccharides and EPS+ cultures are likely to emerge inside and outside the dairy industry. In this communication, we examine current knowledge related to EPS production by S. thermophilus.
Heteropolysaccharide Structure and Biosynthesis
Compositional and structural analyses have revealed that S. thermophilus EPS are primarily composed of galactose, glucose, and rhamnose, but N-acetyl-galactosamine, fucose, and acetylated galactose have also been identified in these molecules (Figure 1
). As shown in Figure 1, the basic repeating structure of S. thermophilus EPS may vary considerably, and variability in the molecular mass of structurally identical polymers has also been reported (Faber et al., 1998).
|
Considerable effort has also been directed toward understanding cellular mechanisms for EPS assembly in pathogenic streptococci such as Streptococcus pneumoniae, Streptococcus pyogenes, and Streptococcus agalactiae (Garcia et al., 1997; Garcia and Lopez, 1997; Kolkman et al., 1997; Yother, 1999; Chaffin et al., 2000; Bender and Yother, 2001). As a whole, those studies have demonstrated that streptococci possess at least two distinct mechanisms for EPS production. Synthesis of simple EPS, like S. pyogenes hyaluronic acid or S. pneumoniae type 3 capsule (which have linear, disaccharide basic repeating units) does not include formation of lipid-linked intermediates, whereas production of more complex EPS does (Yother, 1999). Functional genetic studies and homology comparisons of EPS-related gene products from S. thermophilus and L. lactis strongly suggest that cellular mechanisms for EPS synthesis in these species are similar to streptococcal and gram-negative systems that require formation of lipid-linked intermediates (Stingele et al., 1996, 1999; van Kranenburg et al., 1997, 1999a, 1999b; Low et al., 1998; Bourgoin et al., 1999; Almiron-Roig et al., 2000; Broadbent et al., 2001; Germond et al., 2001). Work by Stingele et al. (1999) has shown that EPS biosynthesis in S. thermophilus Sfi6 involved a lipid carrier, upon which the sugar monomers that comprise the basic repeating unit are assembled. The first step is catalyzed by EpsE, a galactosyl-1-phosphate transferase that attaches galactosyl-phosphate to a lipid-phosphate carrier. This reaction is followed by the action of three other glycosyltransferases, termed EpsG, EpsI, and EpsF, which sequentially attach N-acetylgalactosamine, glucose, and a branching galactose moiety to the basic repeating unit (Figure 2). Unfortunately, very little is known about EPS polymerization, translocation, or capsule formation in S. thermophilus or any other gram-positive bacteria (Roberts, 1996; Garcia and López, 1997). Amino acid homologies between enzymes involved in these processes in gram-negative and gram-positive bacteria suggest that they occur via similar mechanisms, but this hypothesis awaits experimental confirmation.
|
As is the case with most LAB, the ability to produce EPS in S. thermophilus is both strain dependent and significantly affected by media and growth conditions (e.g., temperature, carbon:nitrogen ratio, pH, etc.) In milk, for example, the amount of EPS produced by S. thermophilus may range from 50 mg to 1.5 g/L (Degeest et al., 2001a). While these yields are comparable to those obtained for EPS from other LAB, they are quite low when compared to production of dextran or xanthan gums (Degeest et al., 2001a). For this reason, there is considerable interest in the metabolic engineering of S. thermophilus and other "food-grade" LAB to enhance EPS production. Potential targets for this work include housekeeping enzymes whose product availability may be limiting to EPS.
As noted previously, the lipid carrier for EPS assembly in gram-negative bacteria is undecaprenyl phosphate, and participation of this molecule in EPS synthesis by LAB may be significant because it is also required for assembly of peptidoglycan, teichoic acids, and lipoteichoic acids (Cerning, 1990; Delcour et al., 1999). Specifically, if the cellular concentration of undecaprenyl phosphate is limited, then competition for lipid carrier may be a key limiting factor in EPS biosynthesis. This hypothesis is supported by the frequent observation that incubation under conditions that stimulate bacterial growth and division reduce EPS production, while growth conditions which reduce the demand for new cell wall synthesis (such as incubation at a suboptimal temperature) tend to enhance EPS production (Sutherland, 1972; Cerning, 1995). Unfortunately, the identity of the lipid carrier for EPS assembly in Gram-positive bacteria has been elusive. It is interesting to note that EPS production in mesophilic LAB such as L. lactis typically follows the pattern described above, but production in S. thermophilus and other thermophilic LAB is often directly coupled to growth rate (i.e., maximal yield at maximal growth rate; De Vuyst et al., 2001). The latter observation and the finding that bacitracin (a compound that blocks undecaprenyl phosphate recycling) does not inhibit EPS production in S. thermophilus (Stingele et al., 1999), suggests that this species either utilizes an alternative carrier molecule or that undecaprenyl phosphate availability is not limiting.
Competition for sugar nucleotide precursors is another potential limitation to high EPS yields from S. thermophilus and other EPS-producing LAB (Boels et al., 2001). Many of the sugar nucleotide precursors used for EPS occur as intermediates in sugar catabolic pathways or, like undecaprenyl phosphate, are needed for the assembly of peptidoglycan and other glycan-containing cellular polymers (Poolman, 1993; Delcour et al., 1999). For example, UDP-glucose 4-epimerase (GalE) catalyzes the interconversion of UDP-galactose and UDP-glucose residues, each of which may be needed for assembly of S. thermophilus EPS and other cell polymers (Delcour et al., 1999; Stingele et al., 1999). This enzyme is also a component of the Leloir pathway for galactose catabolism (Whitfield and Keenleyside, 1995), but most strains of S. thermophilus are phenotypically galactose-negative. Galactose-fermenting mutants can be isolated after incubation in media with galactose as the sole carbohydrate source, however, and genes and low-level enzyme activities for GalE and all other Leloir pathway enzymes have been detected in this species (Thomas and Crow, 1984; Hutkins et al., 1985; Vaughan et al., 2001; Vaillancourt et al., 2002). These findings suggested that in S. thermophilus, the primary role for Leloir pathway enzymes is the production of sugar nucleotides for assembly of EPS and other polymers (Poolman, 1993). This hypothesis has since been confirmed for lactose-grown cells (Levander and Rådström, 2001).
Because low levels of sugar precursors may be a limiting factor on EPS yields, greater understanding of S. thermophilus metabolic flux in sugar metabolism may provide new strategies to enhance EPS production. Several researchers have demonstrated a correlation between EPS production and intracellular activities of enzymes from the Leloir and glycolytic pathways, but strain-to-strain variations are also apparent (Petit et al., 1991; Escalante et al., 1998; Degeest and De Vuyst, 2000; Levander and Rådström, 2001). Although the basis for these differences is not yet clear, Levander et al. (2002) recently showed such knowledge can be used to enhance EPS production in S. thermophilus by metabolic engineering of central carbon metabolism. In considering strategies for genetically enhancing EPS yields in S. thermophilus, it is also important to recognize that EPS production is an energy-intensive process, and that S. thermophilus have a limited number of catabolic pathways from which it can derive energy. Because EPS yields in most S. thermophilus are growth associated, efforts to dramatically increase production levels are likely to require novel strategies to enhance biomass production (Degeest et al., 2001a).
Genetics of EPS Production in S. thermophilus
Four distinct EPS gene clusters have been characterized to date in S. thermophilus (Stingele et al., 1996; Bourgoin et al., 1999; Almiron-Roig et al., 2000; Broadbent et al., 2001; Germond et al., 2001) (Figure 3), and recent work suggests a dozen or more unique eps clusters may occur in this species (Rallu et al., 2002). Thus far, the distribution of regulatory and structural genes within each of these clusters shows a modular organization that is conserved in other Streptococcus spp. (Yother, 1999; Chaffin et al., 2000). Although the function of most EPS-related genes (currently designated eps or cps) and gene products are only inferred from sequence or structural homologies, the 5' region of each cluster appears to encode proteins involved in regulation of EPS synthesis, chain length determination, and membrane translocation. These open reading frames are followed by genes encoding the glycosyl-1-phosphate transferase and glycosyltransferases required for assembly of the basic repeating unit, and enzymes involved in repeat unit polymerization. Finally, the 3' end of these clusters typically contain genes for additional proteins involved in membrane translocation of the polymer subunits, and enzymes needed for the production of sugar nucleotide precursors (e.g., N-acetyl-D-galactosamine; Stingele et al., 1996; Degeest et al., 2001b) that are unique to the EPS (i.e., not found in other cell polymers) (Chaffin et al., 2000).
|
). In S. thermophilus, the 5' end of the cluster is preceded by deoD, which encodes a homolog to purine nucleoside phosphorylase presumably involved in the biosynthesis and catabolism of nucleotides (Figure 3). In strain Sfi39, however, the intergenic region between deoD and epsA has been interrupted by transposition of a single copy of IS905 (Germond et al., 2001). Several authors have suggested that the 3' end of the cluster is defined by orf14.9 (Stingele et al., 1996; Almiron-Roig et al., 2000; Germond et al., 2001), and this hypothesis was clearly supported for strain Sfi39 by heterologous expression of a structurally identical EPS in L. lactis (Germond et al., 2001). As shown in Figure 3, however, two of the four clusters characterized to date have genes encoding putative glycosyltransferases and other EPS-related enzymes downstream of orf14.9 (and in opposite orientation of other eps genes). Moreover, one of the downstream open reading frames in MR-1C (cpsU) is a homolog to the Sfi39 epsI gene located upstream of orf14.9 in the latter strain (Figure 3). These observations strongly suggest that some S. thermophilus eps clusters extend beyond orf14.9. Sequence extension beyond orf14.9 by our group has identified another gene, cpsX, whose presence downstream of the eps cluster is conserved in all strains examined to date (Broadbent et al., 2001; Welker and Broadbent, 2002a, 2002b; Welker et al., 2002b). The role, if any, of CpsX in polysaccharide synthesis is unknown, but this protein does show 50% similarity to a S. pneumoniae membrane protein that is genetically linked to a low molecular weight penicillin-binding protein in that species (Krauss and Hakenbeck, 1997). The region downstream of cpsX encodes a truncated ß-galactosidase and a putative protein related to an ABC transport system (Welker and Broadbent, 2002a, 2002b, 2002c; Welker et al., 2002b). The first four genes in the 5' region of S. thermophilus eps clusters, epsA-D, are highly conserved among this and other EPS+ Streptococcus spp. (Cieslewicz et al., 2001) and appear to contribute regulation (epsA and epsB), polymerization (epsC), and membrane translocation (epsD) functions to EPS synthesis. Interestingly, these four genes are also found in many S. thermophilus strains that do not produce EPS (Stingele et al., 1996; Bourgoin et al., 1999; Pluvinet et al., 2000). Functional studies on epsA-D genes of S. thermophilus have not been reported, but the epsA gene product of this species and other streptococci show significant homology to Bacillus subtilis LytR, a small, basic protein that negatively regulates transcription of the lytABC operon encoding autolysin (Lazarevic et al., 1992). Inactivation of the epsA homolog in S. agalactiae, cpsIaA, confirmed that this gene is also involved in regulation of eps gene expression, but instead of increasing EPS production, the disruption mutant produced 24% less CPS and 2 to 3 fold less cps transcript than wild-type cells (Cieslewicz et al., 2001). Introduction of an inframe deletion mutation in cpsA also decreased CPS production in S. pneumoniae (Morona et al., 2000). These data suggest that CpsA acts as a transcriptional activator instead of an attenuator in streptococci. As noted by Cieslewicz et al. (2001), streptococcal EpsA homologs differ from LytR in several features that may contribute to the distinct modes of action seen with each protein. For example, the N-terminal region of streptococcal EpsA homologs contain predicted membrane-spanning domains that are absent in LytR, and streptococcal EpsA proteins share conserved hydrophobic regions that have no counterpart in LytR. In addition, lytR and lytABC are transcribed from divergent promoters, whereas all streptococcal epsA genes appear to be transcribed as part of the eps operon (Lazarevic et al., 1992; Cieslewicz et al., 2001).
Like EpsA, streptococcal EpsB protein is also involved in regulation of EPS synthesis, but via an interaction with EpsC and EpsD (Bender and Yother, 2001; Morona et al., 2002). Inactivation of CpsC or CpsD in S. agalactiae reduced EPS chain length (i.e., the average molecular mass), which is consistent with the proposed role of these proteins in EPS polymerization and export (Cieslewicz et al., 2001). Streptococcal EpsC and EpsD proteins display sequence and structural homology to amino- and carboxy-terminal regions, respectively, of the E. coli autophosphorylating protein-tyrosine kinase Wzc, which also is involved in polymerization and export of capsular K antigen and colanic acid EPS (Figure 4; Vincent et al., 1999; Morona et al., 2000; Joly and Stingele, 2001; Wugeditsch et al., 2001). The E. coli Wzc and related proteins possess distinct transmembrane and nucleotide-binding domains. In Wzc, these domains are present in the same protein, but they may also be divided between two proteins such as streptococcal CpsC and CpsD homologs. Autophosphorylation of Wzc is required for CPS assembly, and this activity is likely regulated at the posttranslational level by a phosphotyrosine-protein phosphatase, Wzb (Vincent et al., 1999, 2000; Wugeditsch et al., 2001). Interestingly, the mechanism by which regulation occurs may differ even within a species; for example, phosphorylation of Wzc in E. coli enhances synthesis of capsular K antigen but inhibits production of colanic acid (Vincent et al., 2000; Wugeditsch et al., 2001). Like their functional homologs in streptococci, genes encoding Wzb and Wzc are found in the eps operon for the polymer whose synthesis is regulated by their activity (Rahn et al., 1999; Wugeditsch et al., 2001).
|
P and colanic acid synthesis in E. coli, tyrosine phosphorylation of CpsD appears to negatively regulate CPS production in S. pneumoniae (Morona et al., 2000). In streptococci, dephosphorylation of CpsD is performed by CpsB, a novel manganese-dependent phosphotyrosine phosphatase (Bender and Yother, 2001; Morona et al., 2000, 2002). More specifically, CPS production in S. pneumoniae involves a complex that includes, at minimum, CpsB, CpsC, CpsD, and ATP (Morona et al., 2000; Bender and Yother, 2001). Maximal levels of CPS biosynthesis/polymerization occur when ATP is bound to CpsD, and ATP binding requires CpsD and CpsC interaction. However, autophosphorylation of CpsD tyrosine residues by bound ATP results in (or may occur as a consequence of) dissociation of CpsD from the complex and reduced CPS production. Finally, dephosphorylation of CpsD by CpsB regenerates active CpsD for complex formation and allows for increased CPS biosynthesis (Morona et al., 2000; Bender and Yother, 2001). Interestingly, CpsB also inhibits autophosphorylation of CpsD, and the dual functions of this protein are suggestive of a central role in regulation of CPS biosynthesis (Bender and Yother, 2001). Although functional studies of S. thermophilus EpsB, EpsC, and EpsD proteins have not been performed, the existence of strains that produce comparable amounts of EPS with identical repeating units but different average molecular mass (Faber et al., 1998) strongly suggests that this species uses one or more mechanisms to regulate EPS polymerization and chain length. If streptococcal EpsB, EpsC, and EpsD perform similar functions in each species, then it may be possible to enhance the production level or molecular mass of S. thermophilus EPS by mutagenesis of the CpsD carboxy-terminal (YGX)n motif required for phosphorylation (Morona et al., 2000). However, the subtleties noted above for WzcP regulation of CPS production in E. coli indicate that the role (if any) of EpsD phosphorylation in S. thermophilus must first be determined experimentally.
The fifth gene in S. thermophilus eps clusters, epsE, encodes a glycosyl-1-phosphate transferase that catalyzes the first step in assembly of the EPS basic repeating unit: addition of hexose-1-phosphate to the lipid-phosphate carrier (Stingele et al., 1999; Almiron-Roig et al., 2000; Figure 2). Functional analyses have demonstrated that at least two enzymes of this type occur in S. thermophilus: a galactosyl-1-phosphate transferase (Mr = 26,000) that attaches galactose-phosphate to phosphorylated lipid carrier and is represented by the EpsE protein of strain Sfi6 (Stingele et al., 1999), and a much larger (Mr = 52,000) glucosyl-1-phosphate transferase that catalyzes glucose-phosphate addition to the lipid carrier and is represented by the NCFB 2393 CpsE protein (Almiron-Roig et al., 2000). As is shown in Figure 3, DNA and predicted protein sequence alignments indicate the EpsE protein of strains MR-1C and Sfi39, respectively, are homologs (>90% amino acid identity) to the Sfi6 galactosyl-1-phosphate transferase and NCFB 2393 glucosyl-1-phosphate transferase. As might be expected from their mass discrepancies, the catalytic activity of S. thermophilus glycosyl-1-phosphate transferases may differ in more ways than simple sugar-nucleotide specificity. For example, the WbaP protein of enteric bacteria is a bifunctional hexose-1-phosphate transferase that participates in the glycosyltransferase activity and in the release of the basic repeating unit from the lipid carrier (Wang et al., 1996). The domain associated with lipid carrier release is located in the amino-terminal region of WbaP, whereas glycosyl-1-phosphate transferase activity is provided by the carboxy-terminal half of the protein. Almiron-Roig et al. (2000) showed that CpsE from NCFB 2393 and WbaP share topological similarity over their entire lengths and suggested that the NCFB 2393 glucosyl-1-phosphate transferase may also be a bifunctional protein. In contrast, the much smaller galactosyl-1-phosphate transferase found in Sfi6 and MR-1C only shares homology to the carboxy-terminal region of WbaP, and is, therefore, unlikely to participate in lipid carrier release. The latter observation suggests that in S. thermophilus strains encoding galactosyl-1-phosphate transferase, lipid carrier release may involve a second, as yet unidentified protein.
Genes downstream of epsE appear to encode glycosyltransferases, export/polymerization functions, sugar biosynthesis, and a few enzymes whose function is unknown. Genes encoding a variety of 
- and ß-glycosyltransferases have been identified in S. thermophilus and other LAB, and while clear functional assignments for LAB glycosyltransferases remain limited, the sugar specificity of many enzymes can be postulated from protein sequence comparisons (see Jolly and Stingele, 2001). Nonetheless, more work is needed to confirm these assignments and to establish the function of many other genes located in the 3' end of S. thermophilus eps clusters.
As is evident from the preceding discussion, many S. thermophilus eps genes are organizationally, structurally, and (presumably) functionally conserved in eps gene clusters from different S. thermophilus strains (Figure 3) and even among other streptococci and Gram-positive bacteria (Jiang et al., 2001; Jolly and Stingele, 2001). Sequence comparisons and functional studies of these genes and gene clusters are thus expected to provide great insight on the evolution, physiological role, and biology of EPS production in S. thermophilus and other LAB. In the dairy industry, such knowledge may ultimately allow for genetic engineering of novel EPS with predictable and unique value-added food properties. The functional and comparative genetic studies needed to attain these advances would be facilitated by implementation of a unified system for eps gene nomenclature in LAB. One avenue toward this goal would be to adopt the bacterial polysaccharide gene nomenclature system, wherein homologous genes are assigned the same symbol without regard to species or strain differences (Reeves et al., 1996; Jiang et al., 2001) (for current details, see http://www.microbio.usyd.edu.au/BPDG/default.htm). By this approach, all of the S. thermophilus epsA-D (or cpsA-D) genes would be designated wzg, wzh, wzd, and wze, respectively; and homologs to the NCFB 2393 cpsE and cpsF genes would be termed wchA and wchF, respectively (Reeves, 2002). In addition, protein sequence alignments and hydrophobicity profiles indicate Sfi6 epsM and MR-1C cpsU (designated as epsJ in strain Sfi39) genes encode homologs to the conserved "flippase" Wzx (Figure 5), while Sfi6 epsJ, MR-1C cpsS, and Sfi39 epsF* genes encode homologs to the conserved EPS polymerase Wzy (Figure 6). It is interesting to note that the absence of a wzy homolog in the NCFB 2393 cps cluster suggests DNA sequence information for that strain’s cluster may not be complete. Additional information related to the application of the bacterial polysaccharide gene nomenclature system in S. thermophilus is available from Welker and Broadbent (2002a and 2002c) and Welker et al. (2002b).
|
|
), and comparative genetic analysis of eps clusters from CNRZ368 (Bourgoin et al., 1999) and MR-1C (Broadbent et al., 2001) showed that general genomic instability was likely responsible for loss of the EPS+ phenotype in CNRZ368 (Welker et al., 2002a).
Physiological Role of EPS
In nature, bacterial EPS are thought to fulfill diverse functions that may include cell protection against toxic or limiting environments and other antagonisms, sequestering of essential cations, colonization, and in cellular recognition (Roberts, 1996; De Vuyst et al., 2001; Krinos et al., 2001; Looijesteijn et al., 2001). The excellent water-binding properties of EPS, for example, may protect bacteria in low-moisture environments, and capsular EPS can inhibit amoebic attack or phagocytosis (Sutherland, 1972; Kang and Cottrell, 1979; Roberts, 1996). The adhesive properties of EPS are of considerable interest because of the key role EPS plays in biofilm formation and dental caries pathogenesis (Cerning, 1990; Whitfield and Keenleyside, 1995). Since complex carbohydrates are implicated in a multitude of biological recognition reactions, EPS may also be involved with modulation of the immune system and/or host-pathogen interactions (Jolly and Stingele, 2001). Recently, Krinos et al. (2001) showed that the gram-negative human colonic microorganism, Bacteroides fragilis, is able to modulate its surface antigenicity by producing at least eight distinct capsular polysaccharides, and that the synthesis of these polymers is regulated in an on-off manner by the reversible inversion of DNA segments containing the gene promoters for CPS expression. This mechanism for generating surface diversity allows the organism to exhibit a wide array of distinct surface polysaccharide combinations, and may have implications on how bacteria are maintained in a specific ecological niche such as the intestinal tract (Krinos et al., 2001). One function that EPS production apparently does not provide is that of an energy reserve, since most EPS+ LAB are unable to catabolize the polymers they produce (Cerning, 1990).
The protective effect of EPS against low-moisture environments led Petersen (2001) to investigate the effect of a CPS+ phenotype in resistance of S. thermophilus to freezing and lyophilization. Cell counts collected before and after each challenge showed that viability of S. thermophilus MR-1C (CPS+) was numerically higher than that of its CPS- isogen, DM10, but the differences were not significant (P > 0.05). Interestingly, microscopic examination of rehydrated cells revealed that the mean chain length was reduced by lyophilization, and that the chain length of MR-1C did not decrease as much after freeze drying as did that of DM10. Even when chain length was taken into account, however, there was no significant difference (P > 0.05) in survivability between the two strains after lyophilization. Similar experiments with ropy EPS+ and EPS- strains led the author to conclude that strain differences were more important than the presence or absence of EPS on resistance of S. thermophilus to freezing or lyophilization (Petersen, 2001).
Bacteriophage Resistance
Virulent bacteriophages, or phages for short, are the most significant cause of fermentation failures in the dairy industry worldwide (Moineau, 1999). As indicated previously, EPS are thought to protect cells against detrimental environmental conditions, and thick-layered CPS have been shown to protect bacteria against phage infection (Kang and Cottrell, 1979). However, the involvement of EPS in the phage infection process of LAB is still unclear.
For some gram-negative bacteria, the presence of EPS is essential to phage infection. In Vibrio cholera, for example, phages that specifically lyse CPS+ strains have been isolated and used to identify variants that lacked CPS (Albert et al., 1996). Specific phages infecting E. coli strains that possess CPS have also been isolated (Dietz et al., 1986; Clarke et al., 2000; Scholl et al., 2001). For other gram-negative bacteria such as Rhizobium meliloti, EPS have been demonstrated to inhibit phage adsorption by coating the cells (Defives et al., 1996). A number of studies have suggested that EPS may also protect LAB cells against phage infection (Vedamuthu and Neville, 1986; Cerning, 1990; Moineau et al., 1996; De Vuyst and Degeest, 1999). Looijesteijn et al. (2001) showed that cell-associated EPS provided Lactococcus strains with a weak protection against phages, and Forde and Fitzgerald (1999a) showed that loosely bound EPS produced by some lactococcal strains slightly inhibited phage adsorption. The latter effect was presumably caused by obstruction of the interaction between the phage and the cell surface phage receptors (Forde and Fitzgerald, 1999a).
There are very few studies on the isolation and characterization of phages infecting EPS+ LAB strains. Furthermore, to our knowledge, all such reports are related to the characterization of phages infecting EPS+ Lactococcus strains. The first study was published more than 20 yr ago and described the isolation of two lactococcal phages (Sozzi et al., 1978). However, these phages, as well as the interactions with their EPS+ hosts, have not been characterized. The second study described the isolation of the lactococcal phage KSY1, which was isolated from the Finnish fermented ropy milk viili (Saxelin et al., 1979; Figure 7). Phage KSY1 is the reference phage for the lactococcal phage species that bears its name (Jarvis et al., 1991). This morphotype is very rare, as only two KSY1-like phages have been reported in lactococci (Saxelin et al., 1979, 1986). Saxelin et al. (1986) also reported the isolation and characterization of 13 morphological types of phages from viili. However, it is not known whether these phages infect EPS+ strains.
|
We have also tested the phage sensitivity of the CPS+ S. thermophilus strain MR-1C as well as a CPS- derivative of MR-1C, strain DM10 (Moineau et al., 2002). Both strains were sensitive to three phages, Q1 and Q3 (Moineau et al., 1995) as well as phage Sfi11 for which the complete genomic sequence is now available (Lucchini et al., 1999). These three phages are classified within the pac-type group of S. thermophilus phages (Siphoviridae family, LeMarrec et al., 1997). These results indicate that the MR-1C CPS is not sufficient to protect those cells against these three phages and that it was not required for phage infection.
It remains to be seen whether the structure of some EPS produced by LAB could be involved in phage sensitivity or insensitivity. For example, it is known that the lactococcal phage infection process starts by the adsorption of the phage tail to a carbohydrate component such as galactose and rhamnose on the cell wall (for a review, see Forde and Fitzgerald, 1999b). EPS produced by S. thermophilus and other LAB are known to contain these sugars (Figure 1). It is, therefore, tempting to speculate that, in some cases, EPS could serve as a "decoy" for phage adsorption. Moreover, EPS production could increase the viscosity of the medium, and this molecular crowding could slightly inhibit the spread of the phage infection. It is also possible that the phage insensitivity reported in the literature for some EPS+ LAB strains (Moineau et al., 1996) may simply be related to the absence of specific phage receptors on the cell surface of these strains. Nonetheless, it is now clear that processors cannot rely solely on EPS production to protect starter cultures against phage attack.
Applications in Yogurt and Fermented Milks
Incorporation of EPS or EPS+ cultures in dairy foods can provide viscosity, stability, and water-binding functions (De Vuyst and Degeest, 1999) that may contribute positively to the mouth-feel, texture, and taste perception of fermented dairy products (Duboc and Mollet, 2001). For many years, European processors have exploited the differences among ropy EPS+ LAB to produce a variety of fermented milks with unique properties (Cerning, 1995). Dairy processors have also utilized EPS+ starter cultures to control syneresis in yogurt, and this practice is particularly widespread in countries where the addition of plant- or animal-derived stabilizers is prohibited (Cerning, 1995). The latter application is also important in light of modern consumer preferences for products with lower sugar and fat levels, and with fewer food additives.
Yogurt manufacture remains the most important commercial application for EPS+ LAB in dairy foods, and several investigators have studied the impact of EPS+ starters on the physical properties of yogurt. Hassan et al. (1996a) found that both ropy- and nonropy-encapsulated (CPS+) strains of S. thermophilus increased yogurt viscosity more than nonropy-nonencapsulated (i.e., EPS-) strains. Additionally, ropy (slime-producing) strains produced a more stretchable yogurt than nonropy encapsulated strains. When "yield stress" was measured (the force necessary to initiate flow), yogurt made with a nonropy CPS+ strain had lower values than yogurt made using an EPS- strain. Rawson and Marshall (1997) also found that ropy starter bacteria increased yogurt viscosity, but firmness or elasticity was not always increased. The same study also found that yogurt made with a nonropy S. thermophilus and a ropy L. debrueckii ssp. bulgaricus recovered its viscosity more rapidly after destructive testing than yogurt made from two ropy cultures. The authors (Rawson and Marshall, 1997) suggested that this effect was due to EPS interference with protein bonding, which led to the formation of larger voids within the protein matrix.
Other studies indicated nonropy CPS+ strains of S. thermophilus bound more water in yogurt than EPS- strains, but ropy strains bound even more water than CPS+ strains (Hassan et al., 1996b). As expected, increased levels of water binding made the yogurt less susceptible to syneresis. Both ropy and nonropy CPS+ strains bound sufficient additional water to affect yogurt texture. Stirred yogurt products possess a smooth, creamy texture, and are made by mild homogenization of yogurt coagulum after fermentation. The presence of EPS in stirred-type yogurt makes the product less susceptible to mechanical damage from pumping, blending, and filling machines (Robinson, 1981; De Vuyst and Degeest, 1999). Mechanical processing steps also increase syneresis of the final product, but use of EPS+ cultures can help to control this defect (Duboc and Mollet, 2001).
Besides yogurt, other dairy products in which S. thermophilus EPS has been shown to affect product quality include sour cream and traditional fermented milks in Nordic countries. Adapa and Schmidt (1998) showed that less syneresis occurred in sour cream made with S. thermophilus strains that produced larger versus smaller CPS. Strains producing the larger CPS also imparted a higher apparent viscosity as well as greater adhesiveness and gumminess to sour cream, which increased the "stringy" nature of the product.
Ropy mesophilic LAB cocci are the dominant flora of Nordic ropy milk, but EPS+ strains of S. thermophilus are used in some types to provide proper texture. Strains of EPS+ S. thermophilus have also been isolated from kefir grains, where they are suggested to affect the texture of kefir (Duboc and Mollet, 2001). EPS-producing strains of S. thermophilus have even been used in an attempt to reduce freezing mortality when viable LAB are incorporated into frozen dairy desserts as a source of ß-galactosidase (Hong and Marshall, 2001). These researchers found that encapsulated strains survived better than nonencapsulated strains after freezing and frozen storage at -29°C for 16 d.
Research shows that ropy EPS+ and nonropy CPS+ strains of S. thermophilus can influence the rheological properties of soft set milk products, but each type may have different effects. Although the amount of polymer produced certainly affects milk viscosity, Tunier (1999) showed that the functional impact of EPS on milk products was largely determined by the polysaccharide’s molar mass, monosaccharide composition, and linkage type. In addition, EPS interaction with milk proteins also helps determine the viscosity of a fermented product. The practical significance of these observations may be illustrated by the work of Faber et al. (1998) with S. thermophilus Sts and Rs. Strains Sts and Rs are ropy and non-ropy cultures, respectively, that produce polysaccharide with an identical heptasaccharide basic repeating unit (see Figure 1) but different average molecular mass. The two cultures also have the same effect on the milk protein-polysaccharide network, but stirred milk cultures of Rs appeared "thin and lumpy," whereas those of Sts looked "very thick" (Faber et al., 1998). Further characterization of EPS functionality, structure, and genetics will ultimately allow dairy processors to select or perhaps even construct EPS+ cultures that impart unique and predictable properties to a particular food product.
Application in Mozzarella Cheese
While the application of EPS+ starter lactic acid bacteria has been largely confined to the production of yogurt and fermented milks, some companies have used ropy strains to improve the texture and functionality of cheese (Pruss and Bahrs, 1981; Nauth and Hayashi, 1995). Interest in EPS+ starters for cheesemaking has been restricted, however, by the fact that EPS often accumulates in the cheese whey, thereby increasing its viscosity. This outcome is undesirable because it retards the efficiency of membrane processing and slows whey protein concentration and drying processes (Petersen et al., 2000).
Consumer demand for reduced-fat foods has created interest in the development and manufacture of reduced-fat, low-fat, and nonfat Mozzarella cheese, but fat removal has several undesirable effects on cheese physical properties (McMahon et al., 1993). Specifically, Mozzarella cheese becomes tough and rubbery, more heat is required for melting, and the cheese loses pliability rapidly during cooling (Mistry and Anderson, 1993). Research has revealed that these properties are heavily influenced by cheese moisture level (McMahon and Oberg, 1998). Examination of cheese microstructure shows that in full fat or part-skim Mozzarella cheese, much of the water is contained in channels formed in the protein matrix by entrapped fat globules (McMahon et al., 1993, Oberg et al., 1993). Because there are very few fat globules to break up the protein matrix in low-fat Mozzarella, these channels become much narrower with less space available for water retention. This results in cheese with a lower moisture level (Oberg et al., 1993) and, as a consequence, a tough, rubbery texture and poor melt and stretch properties (McMahon and Oberg, 1998).
Because EPS have excellent water-binding properties, and moisture retention is vital to functionality in low fat cheese (McMahon and Oberg, 1998), Perry et al. (1997) investigated the influence of an EPS-producing starter pair on the moisture and melt properties of low fat (6%) Mozzarella cheese. Each of the bacteria used in that study, S. thermophilus MR-1C and L. bulgaricus MR-1R, produced EPS. As predicted, low-fat Mozzarella cheese manufactured with MR-1C and MR-1R contained significantly (P < 0.05) more moisture and had better melting properties than cheese made with an EPS- commercial starter pair (S. thermophilus TAO61 and Lactobacillus helveticus LH100). Subsequent work by Low et al. (1998) showed this effect was caused exclusively by S. thermophilus MR-1C, and that the water-binding properties of this bacterium were imparted by its large CPS.
While the ability to enhance cheese yield (through increased moisture retention) and functionality are strong incentives to use EPS+ cultures in Mozzarella cheese, concerns over EPS accumulation in cheese whey continue to limit commercial interest in EPS+ cheese starters. However, it is important to recognize that while EPS accumulation in cheese whey may be expected from ropy starters, CPS+ bacteria are less likely to affect whey viscosity because the polymer is covalently linked to the cell wall (Deng et al., 2000). Petersen et al. (2000) investigated the effect of capsular and ropy EPS+ S. thermophilus starters on part-skim Mozzarella cheese and whey. They showed that cheese moisture levels were significantly higher in Mozzarella made with EPS+ versus EPS- streptococci, but whey from cheese made with the ropy starter was significantly more viscous (P < 0.01) than whey from cheeses made with the CPS+ or EPS- strains. More importantly, no significant difference was detected in the viscosity of 5-fold concentrated whey from cheese made with CPS+ or a commercial EPS- starter (Petersen et al., 2000). These findings indicate that CPS+, but not ropy, S. thermophilus starter bacteria can be used to enhance Mozzarella cheese moisture level, yield, and melt properties without deleteriously affecting whey viscosity.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Received for publication May 10, 2002. Accepted for publication June 25, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
-phosphoglucomutase, UDP-galactose 4-epimerase, and UDP-glucose pyrophosphorylase with exopolysaccharide biosynthesis by Streptococcus thermophilus LY03. Appl. Environ. Microbiol. 66:3519–3527.
