ro ui ui

Several other IS elements have since been identified in many dairy LAB, including other species of Lactobacillus, Lc. lactis, S. thermophilus, and leuco-nostocs (see Table 2). Nucleotide sequence analysis and DNA-DNA hybridizations have established that several of these elements are present in multiple copies throughout the LAB genome (plasmids and chromosome), and that related sequences are present in most (and probably all) industrially important LAB. Existence of iso-IS elements (e.g., ISSi) in many different LAB species, and proximity of these elements to plasmid-borne genes encoding important milk fermentation properties (e.g., lactose and citrate utilization, proteinase production, and phage resistance) suggests that IS were probably important in the evolutionary adaptation of LAB to a milk environment (Bourgoin et al., 1999; Davidson et al., 1996; Magni et al., 1996).

2. Transposons

Two types of transposons can be distinguished in dairy LAB: composite transposons and conjugative transposons. Composite transposons typically consist of a nonmobile central region that is flanked on each side by complete IS elements that provide the transposition factors. Given the frequency at which some IS occur in the chromosome and plasmid DNAs of Lc. lactis and other dairy LAB (see Table 2), elements that satisfy the structural definition of a composite transposon may be quite common in genomes of these bacteria. A few putative elements have been identified in Lc. lactis and S. thermophilus, but conclusive proof for intracellular transposition by any naturally occurring composite transpo-son in food-grade LAB is still lacking (Bourgoin et al., 1999; Duan et al., 1996; Huang et al., 1993; Romero and Klaenhammer, 1991; Teuber et al., 1999). Nonetheless, transposition of an artificial composite transposon that was assembled with IS946 elements has been demonstrated in Lc. lactis (Romero and Klaenhammer, 1991), and functional mobility for some native elements is evidenced by the fact that they contain IS elements (and intervening DNA regions) that have clearly been acquired through horizontal gene transfer (Bourgoin et al., 1999; Teuber et al., 1999). Readers should also recognize that several functionally active composite transposons have been identified in enterococci and streptococci, where these elements contribute to the problematic spread of antibiotic resistance genes (Horaud et al., 1996; Teuber et al., 1999; Woodford 1998).

a. Conjugative Transposons With a size range of 18-70 kb, the conjugative transposons of gram+ bacteria are generally larger and more complex mobile elements than composite transposons. Conjugative transposons were originally discovered in the late 1970s in pathogenic LAB, and current models for their transposition are derived largely from studies of the enterococcal transposon Tn9i6 and other Tn9i6-like elements (Salyers and Shoemaker, 1997). Members of the Tn916 family of transposons have a very broad host range that extends to more than 50 species in 24 bacterial genera (Jaworski and Clewell, 1995).

Like IS and composite transposons, conjugative transposons are able to excise from and insert into chromosomal or plasmid DNA, but some aspects of their transposition are more akin to plasmids and temperate bacteriophages than to other transposable elements. Excision of a conjugative transposon, for example, is followed by its conversion into a plasmid-like covalently closed circular DNA molecule (which is, however, incapable of autonomous replication) that can be transferred by conjugation in single-stranded form into another (recipient) cell. Moreover, mechanisms for integration and excision of the circular DNA intermediate are phage-like in that they require a transposon-encoded integrase, and excision is stimulated by the transposon's xis gene product (Salyers et al., 1995).

As was hinted above, several conjugative transposons have been identified in enterococcal and streptococcal clinical isolates, and these elements are now recognized for their integral role in dissemination of antibiotic resistance genes to many species of bacteria (Teuber et al., 1999). In contrast, the only conjugative transposons to be conclusively identified thus far in food-grade LAB are the very large (approximately 70 kb) and genetically related nisin-sucrose transposons of Lc. lactis. These elements do not appear to encode antibiotic resistance genes and seem to be far less promiscuous than their enterococcal and streptococcal counterparts. Evidence for the latter assertion comes from the observation that although intraspecific conjugation of these transposons has been demonstrated by several groups, genetic proof for intergeneric transfer has only been documented once (Broadbent and Kondo, 1991; Broadbent et al., 1995).

Interest in lactococcal nisin-sucrose transposons stems from the finding that they encode genes for nisin biosynthesis and immunity. Nisin is a broad-spectrum lantibiotic that is widely used as a preservative to combat gram+ spoilage and pathogenic bacteria in food (Horn et al., 1991). Structural characterization of several nisin-sucrose transposons has revealed that conjugative elements can be separated into two classes, designated I and II, whose structures are represented by Tn5276- and Tn527S-like transposons, respectively (Rauch et al., 1994). A third group of nisin-sucrose ''transposons,'' class III elements, appear to be derived from class II transposons, but the former elements cannot be transferred by conjugation and probably lack transpositional mobility.

As shown in Fig. 2, group I nisin-sucrose transposons have an IS904 element near their left junction, just upstream of genes for biosynthesis of nisin A (one of two natural nisin variants). Another IS, IS981, lies downstream of the nisin gene cluster and adjacent to genes for sucrose metabolism via a sucrose-specific phosphoenolpyruvate-dependent phosphotransferase system. Like Tn916, genes involved in Tn5276 excision and integration are located near the nisABTCIPRKFEG sacAR sacBK xis int nisABTCIPRKFEG sacAR sacBK xis int

IS904 IS981

Figure 2 Genetic organization of the lactococcal group I nisin-sucrose transposon Tn5276. The transcriptional orientation of genes for nisin biosynthesis (nisA-G), sucrose utilization (sac), excision (xis), and integration (int) are illustrated by black arrows above the element, and orientations of putative transposase genes in IS904 and IS981 are indicated by the thick gray arrows. The hatched area represents the region of the transposon for which DNA sequence information has not yet been reported. Map is not to scale.

IS904 IS981

Figure 2 Genetic organization of the lactococcal group I nisin-sucrose transposon Tn5276. The transcriptional orientation of genes for nisin biosynthesis (nisA-G), sucrose utilization (sac), excision (xis), and integration (int) are illustrated by black arrows above the element, and orientations of putative transposase genes in IS904 and IS981 are indicated by the thick gray arrows. The hatched area represents the region of the transposon for which DNA sequence information has not yet been reported. Map is not to scale.

right end of the transposon (De Vos et al., 1995). Much of the region between xis and sacBK (see Fig. 2) has not been described, but phenotypic characterization of transconjugants suggests it may include genes for conjugative self-transfer, resistance to certain bacteriophages, and synthesis of N5-(carboxyethyl)ornithine (Gonzales and Kunka, 1985; Thompson et al., 1991). The structure of group II nisin-sucrose transposons has not been as extensively characterized, but they are known to lack the left-end copy of IS904 and to encode genes for biosynthesis of nisin Z instead of nisin A (Rauch et al., 1994).

Very recently, Burrus and coworkers (2000) described an element in S. thermophilus that appears to represent a new species of conjugative transposon in LAB. This element, termed ICEStl, is 35.5 kb in length and encodes, near its right terminus, genes whose products show extensive homology to proteins involved in conjugation, excision, and integration of other conjugative transposons, including Tn916 and Tn5276. Although conjugal transfer of ICEStl has not yet been confirmed, strong evidence for in vivo and int-dependent excision of the element into a circular intermediate form has been presented. Interestingly, ICESt1 also contains a truncated copy of the lactococcal element IS981, which suggests that conjugal transposition of ICESt1 (or a larger transposon from which it was derived) may have facilitated horizontal gene transfer between Lc. lactis and S. thermophilus (Burrus et al., 2000).

3. Group I and Group II Introns

Group I and group II introns are ribozymes that catalyze a self-splicing reaction from mRNA species that contain the intron, and many of these sequences also function as mobile genetic elements. The most common type of transposition event noted for group I and group II introns is termed ''homing,'' wherein the intron will insert itself into an allele that lacks the cognate element. However, group II (and perhaps group I) introns can also effect transposition to other locations in the genome (Lambowitz and Belfort, 1993). Although they were once thought to be confined exclusively to eukaryotic cells, introns are now known to occur in a wide range of prokaryotes (Belfort et al., 1995). Self-splicing representatives of both groups have now been identified in LAB, where their discovery and characterization have shed new light on intron evolution and biology (Foley et al., 2000; Mikkonen and Alatossava, 1995; Mills et al., 1996).

a. Group I Introns In addition to self-splicing activity, many group I in-trons encode a site-specific endonuclease that confers homing mobility on the intron. The mechanism for homing in group I introns is reasonably well understood, and is thought to occur through a recombination repair process that resembles gene conversion. Homing is initiated by the endonuclease, which creates a double-strand break at a specific target in the intron-free allele, then cleaved DNA strands of the recipient are partially degraded by exonucleases. The gap created in recipient DNA is filled in using the donor strand as the template, which results in coconversion of any exon sequences that were lost to nucleolytic degradation. It is important to note that mobile, endonuclease-encoding group I introns appear to be confined to multicopy genomes such as mitochondria, chloroplasts, and bacteriophages, and this observation has led to suggestions that inefficient doublestrand break repair may limit viability in hosts with a single-copy genome (Lam-bowitz and Belfort, 1993). Given this background, it is not unexpected to learn that all mobile group I introns identified to date in LAB reside in bacteriophage genomes (Foley et al., 2000; Mikkonen and Alatossava, 1995; Van Sinderen et al., 1996).

The first group I intron identified in a LAB was located in a gene encoding the large terminase subunit of the Lb. delbrueckii subsp. lactis virulent bacteriophage LL-H (Mikkonen and Alatossava, 1995). The LL-H intron is 837 bp in length and encodes a 168-amino acid (aa) protein that has good homology to intron-encoded DNA endonucleases found in B. subtilis phages. Although the extreme 3' nucleotide of the intron was reported to contain an A instead of the G found in all other group I introns, in vivo autocatalytic activity was confirmed by polymerase chain reaction (PCR) analysis of terminase gene cDNA (Mikko-nen and Alatossava, 1995).

A second putative group I intron has since been located in the genome of the Lc. lactis temperate bacteriophage r1t (Van Sinderen et al., 1996), and Foley and coworkers (2000) recently showed that many genetically and ecologically unrelated S. thermophilus phages contain a functional group I intron in their lysin gene. The latter work showed that although location of the intron was conserved among different phages, nucleotide sequence analysis revealed the existence of several variant introns. Two of these elements, represented by the 1013-bp introns in phages S3b and ST3, differ by a single nucleotide substitution and contain an open reading frame (ORF) encoding a 253-aa protein with good homology to other intron-encoded endonucleases. Three other variant introns, typified by the elements in phages Sfi6A, S92, and ST64, had deletions in the intron-encoded

ORF that yielded elements of 519,443, and 316 bp, respectively. Another variant intron in phage DT1 differed from the S92 element by one nucleotide substitution (Foley et al., 2000). Since the intron-encoded endonuclease is required for mobility (Lambowitz and Belfort, 1993), it seems unlikely that any of the four latter variants would display homing activity.

b. Group II Introns Like their group I counterparts, many group II introns also contain ORFs. In contrast to the former class of introns, however, ORFs encoded by group II introns produce multidomain proteins with maturase and reverse transcriptase activities that are involved in self-splicing and mobility reactions, respectively (Dunny and McKay, 1999; Lambowitz and Belfort, 1993). Only one functional group II intron has been identified to date in LAB, but putative elements have also been identified in En. faecalis and S. pneumoniae (Dunny and McKay, 1999). The group II intron whose function has been studied is designated Ll.ltrB, and it was independently discovered in a gene (ltrB) encoding con-jugative relaxase by researchers studying the conjugative sex factor of Lc. lactis strains ML3 (pRS01) and 712 (Mills et al., 1996; Shearman et al., 1996). Both groups showed Ll.ltrB had in vivo self splicing activity, and Mills et al. (1997) also demonstrated homing of the intron into an intron-free ltrB allele in Lc. lactis.

The latter observations are particularly significant, because Ll.ltrB was the first functional group II intron to be identified in any bacterium, and its discovery has significantly advanced current understanding of group II intron biology. Analysis of the Ll.ltrB homing pathway in Es. coli and Lc. lactis, for example, has provided new insight into the mechanism for group II intron mobility in bacteria (Dunny and McKay, 1999). The model that emerged from those studies suggests that homing occurs through a novel pathway that is initiated by staggered, doublestrand DNA cleavage at the target site. Two endonuclease activities are required in this reaction: cleavage of the antisense strand is effected by the Ll.ltrB intron-encoded protein and the sense strand is cut at the intron insertion locus by reverse splicing of the intron RNA. Both activities are found in ribonucleoprotein particles formed by the intron-encoded protein and intron RNA. After reverse splicing into the cut site, transposition is completed by cDNA synthesis from the intron template. Unlike group I intron homing, these reactions result in precise integration of the group II intron without coconversion of flanking 5' exon sequences. Group II intron mobility also differs in that it does not require RecA protein and has very relaxed requirements for flanking exon homology (Cousineau et al., 1998; Dunny and McKay, 1999).

Finally, these studies have also shown that domain IV of Ll.ltrB is not essential for self-splicing and can accommodate foreign DNA inserts greater than 1 kb in length. This feature, coupled with the intron's relatively relaxed target specificity and absence of exon coconversion during transposition, indicate that Ll.ltrB may be a useful tool for genetic engineering in bacteria and higher cells (Dunny and McKay, 1999).

C. Bacterial Chromosome

Genes encoding all of the essential housekeeping, catabolic, and biosynthetic activities of the cell are housed in the chromosome. As such, knowledge of chromosomal structure and organization in dairy LAB has great fundamental and applied value to the dairy industry. Recent advances in chromosomal mapping technologies and in nucleotide sequencing resources has sparked an intense interest in bacterial genome analysis, and chromosomes of LAB are certainly no exception.

Efforts to characterize chromosomes of LAB were begun in the early 1970s and 1980s by researchers who used DNA-DNA renaturation kinetics to estimate the genome size (in daltons) of En. faecalis, Lc. lactis, and pathogenic streptococci (Bak et al., 1970; Jarvis and Jarvis, 1981). Classic methods for gene exchange such as transduction and conjugation (see Sec. III.A and III.C) are not well suited to chromosomal mapping in LAB, so more detailed genome studies were not feasible until the advent of pulsed-electric field gel electrophoresis (PFGE) technology in the early 1980s (Le Bourgeois et al., 1993). This methodology allows one to purify relatively intact bacterial chromosomes, digest them with rare-cutting restriction endonucleases, then resolve the large molecular weight restriction products by electrophoresis in an alternating electric field. If appropriate size standards are included in the gel, summation of individual restriction fragments after PFGE provides a rapid and relatively accurate means to estimate genome size. By this approach, genome size estimates have now been collected for strains representing more than 15 species of LAB. These data show that LAB, like other nutritionally fastidious eubacteria, have a relatively small (approximately 1.8 to 3.4 megabase pairs) chromosome (Davidson et al., 1996). One of the practical observations to emerge from this work was that restriction fragment polymorphisms are common in the PFGE profiles from different strains of the same LAB species. This finding has led industry and academia to employ PFGE as a DNA fingerprinting tool for strain identification and for evaluation of strain lineage (Le Bourgois et al., 1993).

Another important outcome of PFGE technology has been its use, in combination with other procedures such as Southern hybridization with specific gene probes, to assemble modest physical and genetic maps of LAB chromosomes. This strategy has been used to procure maps for chromosomes of several industrially important LAB, including Lc. lactis (Davidson et al., 1995; Le Bourgeois et al., 1992a; Tulloch et al., 1991), O. oeni (Ze-Ze et al., 1998), and S. thermophilus (Roussel et al., 1994), and for many of the pathogenic streptococci (Dmitriev et al., 1998; Gasc et al., 1991; Hantman et al., 1993; Suvorov and Ferretti, 1996). These maps have confirmed that individual species and even strains may differ in genomic size and organization, and show that all LAB characterized to date possess a single and circular chromosome.

Finally, PFGE has also facilitated the study of chromosomal geometry and intraspecific polymorphisms in Lc. lactis and S. thermophilus. Those investigations identified intraspecific genomic polymorphisms that have arisen by DNA inversions, insertions, deletions, and translocations, and they provided evidence that IS elements were involved in many of these events (Davidson et al., 1996; Leblond and Decaris, 1998; Roussel et al., 1997). As was noted in Sec. II.B, subsequent work has confirmed that a large genomic inversion in the chromosome of Lc. lactis ML3 was in fact produced by homologous recombination between IS905 elements.

As outlined in the preceding paragraphs, development and commercialization of PFGE technology gave rise to a new microbiological discipline whose subject involves structural, functional, and comparative analyses of bacterial genomes. Although PFGE analysis is still an important component of genome research, the most exciting and innovative work in this rapidly growing field is now being fueled by nucleotide sequence analysis of complete genomes.

1. Comparative Genomics

Compilation and annotation of entire genome sequences has revolutionized bacteriology and microbial genetics, and has created almost unimaginable opportunities to study bacterial evolution, genetics, physiology, and metabolism. Entire nucleotide sequences for more than 30 different microbial genomes have been published since 1995, and sequencing projects for over 100 other species are underway (see The dramatic growth in genome sequence research is largely the result of technical improvements in automated DNA sequencers, molecular biology tools, personal computers, and computer software, which now allow even small laboratories to engage in a bacterial genome project (Frangeul et al., 1999).

For obvious reasons, most of the microbial genome sequencing projects have focused on species with human clinical significance. It is therefore no surprise that sequencing projects among the LAB have targeted several important pathogens in this group (e.g., En. faecalis, S. mutans, S. pyogenes, and S. pneu-moniae). Nonetheless, low-redundancy sequencing of the entire Lc. lactis genome was recently reported by Bolotin et al. (1999), and genome sequencing projects are underway for other important dairy LAB, including Lb. acidophilus and Lb. helveticus.

The value of genome sequence information from both food-grade and pathogenic LAB species to LAB research cannot be overstated. Such comprehensive knowledge will endow industry and academia with unprecedented power to determine the means by which LAB have evolved in, interact with, and respond to milk and cheese environments. It is important to note that sequence acquisition and annotation are only the first steps in functional genomics research. The physi ological role and regulation of most of the deduced ORFs must still be confirmed or identified, and this task could span several decades. Moreover, the speed at which LAB genomics research can progress will also hinge upon the time and degree to which genome sequences are made available to the general scientific community. Nonetheless, the fundamental and applied payoffs of genomic research to the dairy industry are too numerous to list, and many probably cannot yet be envisioned. A few examples of research outcomes that should be possible through this exciting work include:

  1. Knowledge of global gene regulation and integrative metabolism in LAB would help answer long-standing questions regarding mechanisms for the health-promoting benefits of certain LAB; identify means by which some species grow in harsh environments; highlight the most rational strategies for metabolic and genetic improvements to industrial strains; and improve molecular biology resources for genetic manipulation of many dairy LAB species.
  2. Comparative genomics will build a fundamental understanding of LAB evolution and taxonomy that will facilitate safety assurance evaluations of food-grade, genetically modified LAB; provide novel methods for isolation of new starter and adjunct LAB from different environments; and yield new strategies to combat the spread of virulence factors by pathogenic enterococci and streptococci.
  3. Bacteriophages

Bacteriophages, or phages for short, are viruses that attack and destroy bacterial cells. The inhibitory effect of these obligate parasites on dairy starter bacteria has been recognized for more than 60 years, and their destructive impact on the cheese and yogurt industries has focused worldwide attention on molecular genetics and evolution of LAB phages. Because industrial fermentations with Lc. lactis and S. thermophilus starters suffer greatest economic losses, current understanding of LAB phage biology stems largely from phages infecting these two species (Bruissow et al., 1998; Garvey et al., 1995). However, several groups have described bacteriophages infecting other industrially important LAB species, including many dairy lactobacilli, and some of these phages have even been characterized at the genome sequence level (Altermann et al., 1999; Kodaira et al., 1997; Mikkonen et al., 1996). Taxonomically, a few phages with contractile tails (family Myoviridae) or very short tails (family Podoviridae) have been isolated from LAB, but most bacteriophages infecting these species belong to the Sipho-viridae family (phages with long noncontractile tails) of the order Caudovirales (Brussow et al., 1998; Caldwell et al., 1999; Davis et al., 1985; Diaz et al., 1992; Garcia et al., 1997; Jarvis et al., 1991, 1993; Manchester, 1997; Park et al, 1998;

Sechaud et al., 1988; Trevors et al., 1983). A detailed description of LAB phage morphology, infectious cycles, and host range properties are provided in Chapter 6 of this volume and will not be addressed any further here. Instead, this section will highlight some of the exciting outcomes from molecular genetic research of LAB bacteriophages.

Unlike the LAB chromosome, where the promise of genomics research remains largely untapped, the structural, organizational, and evolutionary study of LAB bacteriophage genomes has progressed rapidly in recent years. The obvious reason for this difference is that phage genomes are much smaller (sizes range from 18 to 134 kb) (Prevots et al., 1990) than a bacterial chromosome, and can therefore be sequenced far more rapidly (and inexpensively). Two of the most significant outcomes of phage genetics and genomics studies include (1) a more comprehensive view of bacteriophage diversity and evolution in LAB and (2) application of phage-derived elements to enhance bacteriophage resistance in dairy starter bacteria and for genetic manipulation of these species.

1. On the Origin of Phages

The design of effective phage-control strategies for the dairy fermentation industry depends, to a large degree, on sound knowledge of bacteriophage diversity and evolution. The origin of phages in dairy plants has therefore been the subject of considerable research and debate, and one of the focal points of this discussion has been the role of lysogeny in evolution of virulent phages. As was outlined earlier (see Sec. II.B.1), Shimizu-Kadota and coworkers (1985) showed that a virulent Lb. casei phage clearly was derived from a prophage in the host starter bacterium by insertional transposition of ISLi. Discovery that lysogeny is quite common in dairy LAB, and especially in Lc. lactis, led to speculation that pro-phages may be an important reservoir of lytic bacteriophages in the dairy industry (Davidson et al., 1990). We now know that although virulent Lc. lactis phages can evolve from temperate phages (Davidson et al., 1990), most of the lytic and temperate phages that infect this species share very little DNA homology and therefore are not closely related (Garvey et al., 1995). An important exception involves lytic phages from the P335 species, which do exhibit DNA homology with temperate bacteriophages and whose frequency in cheese plants is increasing (Dumaz and Klaenhammer, 2000; Moineau et al., 1994; Walker et al., 1998). More significantly, new P335 lytic phages evolve by acquisition of host chromosomal DNA, and nucleotide sequence analysis of one of these fragments has confirmed it was derived from prophage components (Dumaz and Klaenhammer, 2000; Moineau et al., 1994).

In contrast to the situation in Lc. lactis, all lytic and temperate S. thermophilus bacteriophages characterized to date belong to a single DNA homology group (Briissow et al., 1998), and comparative genomics has revealed that deletions in the lysogenic module of temperate phages probably plays a key role in evolution of lytic phages (Lucchini et al., 1999a; Tremblay and Moineau, 1999). Fortunately, lysogeny appears to be quite rare in this species (Le Marrec et al., 1997). Lysogens are more common in dairy lactobacilli (Davidson et al., 1990), however, and a genetic relationship between lytic and temperate phages from some of these species has also been established (Auad et al., 1999; Lahbib-Mansais et al., 1988; Mikkonen et al., 1996; Shimizu-Kadota et al., 1985). As a whole, these data clearly show that lysogeny has an important (but not exclusive) role in evolution of new lytic phages in the dairy fermentations industry, and they argue for development of prophage-cured starter LAB (Shimizu-Kadota and Sakurai, 1982).

From a more fundamental perspective, comparative genomics studies of LAB Siphoviridae have also yielded rewarding insight into bacteriophage evolution and taxonomy. As is typical of tailed phages, all LAB phage genomes characterized thus far comprise a linear, double-stranded DNA molecule whose G + C content is parallel to that of the host (Ackermann, 1999). Depending upon the mechanism by which it is packaged into the capsid (which may differ even between very closely related bacteriophages), genomes from LAB Siphoviridae possess cohesive ends or circular permutation with terminal redundancy. Most phage ORFs appear to be transcribed from a common strand, except in temperate phages, where a cluster of genes associated with lysogeny is transcribed divergently from those that encode the lytic cycle (Altermann et al., 1999; Garvey et al., 1995; Klaenhammer and Fitzgerald, 1994; Kodaira et al., 1997; Le Marrec et al., 1997; Lucchini et al., 1999c; McShan and Ferretti, 1997; Mikkonen et al., 1996; Venema et al., 1999).

Efforts to further elucidate structure-function properties of LAB bacterio-phage genomes have been hindered by the experience that protein homology searches rarely yield useful matches for more than a fourth of the phage-encoded ORF products (Desiere et al., 1999). Nonetheless, the structural organization of genes whose function is known or to which a putative role can be assigned has revealed that functionally related genes are distributed into clusters or modules whose order is highly conserved among very different phages (Altermann et al., 1999; Auad et al., 1999; Kodaira et al., 1997; Luccini et al., 1999b, 1999c; McShan and Ferretti, 1997; Mikkonen et al., 1996; Venema et al., 1999). In this regard, LAB bacteriophage genomic structure is quite consistent with the prevailing theory on phage evolution. This theory, termed the modular theory for phage evolution, was formulated to address the highly recombinogenic nature of bacteriophages which, of course, makes evolution by linear descent implausible (Botstein, 1980). By the modular theory, the product of evolution is not a particular virus but instead a family of interchangeable genetic modules which individually perform a specific biological function. Thus, individual viruses represent a combination of modules that have been selected for their singular and coordinated ability to fill a particular niche. Exchange of one module for another with similar function occurs by recombination between bacteriophages that exist within a common, interbreeding population (and these viruses can differ widely in any characteristic except modular construction). Experience now suggests that single modules may be as small as one gene or even a gene fragment encoding the single domain of a protein (Luccini et al., 1999c).

A modular mechanism for LAB phage evolution is clearly evidenced by the recent work of Lucchini et al. (1999b), who showed that genomes of temperate Siphoviridae from all gram+ bacteria with a low G + C content display the following organization in their morphogenesis and lysogeny modules: DNA packaging-head morphogenesis-tail morphogenesis-tail fiber morphogenesis-lysis-lysogeny-DNA replication-followed by a module whose function has not been identified. The workers also noted that these phages may comprise a unique genus within the Siphoviridae family because even though their morphogenesis module is evolutionarily closest to the lambda-like Siphoviridae, their lysogeny module is actually more closely related to that of the P2-like Myoviridae.

Finally, it is important to recognize that since module structure is more highly conserved than nucleotide or amino acid sequences, similarities that exist between morphogenesis modules of lambdoid and LAB Siphoviridae can be exploited to assign putative functions to many LAB phage genes (Chandry et al., 1997; Desiere et al., 1999). Recent validation of this strategy by Desiere et al. (1999) should encourage structure-function research in LAB phage genomes that will eventually provide exciting new insight into the biology of LAB bacterio-phages and phage-host interactions.

2. New Tools for Biotechnology of Lactic Acid Bacteria

Bacteriophage genomics research has also produced several novel phage defense mechanisms for dairy starter cultures. The first system to be described involved insertion of a bacteriophage origin of replication into a streptococcal shuttle vector (Hill et al., 1990). Lactococcal host cells that carry the recombinant plasmid display an abortive phage resistance phenotype called Per (for phage encoded resistance) that is proposed to act by titration of phage replication proteins away from true phage ori sequences during the early stages of infection (Hill et al., 1990; McGrath et al., 1999). Although the efficacy of Per-mediated phage resistance was originally established in Lc. lactis, recent work by Foley and coworkers (1998) suggests Per systems may actually have greater value in S. thermophilus. The reasons for this are twofold: first, very few natural phage defense systems are available for this species; and second, Per-type systems appear to confer relatively broad resistance against S. thermophilus phages (Foley et al., 1998). Other examples of phage defense systems that have been derived from bacteriophage genetics include (1) application of antisense mRNA against highly conserved Lc. lactis phage sequences (Kim and Batt, 1991; Walker and Klaenhammer, 2000); (2) a system for Lc. lactis that places a suicide gene under control of a strictly phage-inducible promoter to trigger death of host cells upon infection (Djordjevic et al., 1997); and (3) a mechanism that imparts immunity to temperate phage superinfection in Lb. casei by constitutive host expression of the phage's gene for repressor protein (Alvarez et al., 1999).

Functional genomic analysis of LAB phages has also yielded a variety of useful tools for molecular genetic manipulation of dairy starter bacteria (Venema et al., 1999). For example, the integrase gene (int) and attachment sequence (attP) that mediate site-specific integration of temperate phages into the host chromosome have been utilized to develop integration vectors that insert foreign DNA into a specific locus (attB) on the bacterial chromosome. The int-attP integration systems offer several important advantages over counterparts that rely upon hostmediated homologous recombination. These include (1) integration occurs at attB, the locus normally used for prophage insertion, and is thus less likely to disrupt cellular functions or viability; (2) integrant stability is usually high under nonselective conditions; and (3) conservation of the attB sequence in different bacteria, or flexibility in its recognition by the int-attP cassette, permits use of these systems in a wide range of bacterial species (Alvarez et al., 1998; Auvray et al., 1997; Van de Guchte et al., 1994; Venema et al., 1999).

Bacteriophage regulatory sequences and lysin genes can also be useful elements for biotechnology. As an example, a rapidly inducible and efficient heterol-ogous gene expression system for Lc. lactis has been developed by incorporation of a phage origin and middle promoter into a low copy number expression vector (O'Sullivan et al., 1996). The system is triggered by deliberate infection with an appropriate bacteriophage, which results in explosive vector replication (i.e., target gene amplification) coupled with phage-induced transcription of the target DNA. Other workers have isolated a phage repressor-operator region that encodes a mutant, temperature-sensitive repressor protein and demonstrated its use for temperature-inducible gene expression in Lc. lactis (Nauta et al., 1997). Finally, model studies indicate that phage lysin genes may have application in tightly regulated suicide cassettes designed to induce starter lysis for accelerated cheese maturation (De Ruyter et al., 1997).

Was this article helpful?

0 0

Post a comment