Modern genetics flows from the ability to manipulate living cells in ways that heritably alter their physiological properties. This achievement has become possible through discovery and refinement of gene transfer mechanisms in bacteria and higher cells. In this section, we will examine four types of gene transfer processes that have been established in dairy LAB: transduction, protoplast fusion, conjugation, and transformation. Although each has played some role in the genetic analysis of dairy LAB, transformation and, to a lesser extent, conjuga tion have clearly emerged as the most useful methods for genetic manipulation in these species.
Transduction is a form of gene transfer which can result from inadvertent packaging of host DNA within a bacteriophage virion during phage replication. Genetic exchange is effected when the phage particle injects this DNA into another bacterium. Phage-mediated gene exchange in LAB was first described by Sandine et al. (1962), who noted transduction of tryptophan biosynthesis and streptomycin resistance markers by a virulent Lc. lactis bacteriophage. This work was significant in that it not only provided the first report of transduction in any LAB, it also represented the first gene transfer system to be identified in a species that was important to the fermented foods industry.
As a mechanism for gene transfer, transduction has been very useful for genetic studies in many bacteria, and it supported some of the first genetic experiments in the industrially important LAB. Researchers at the University of Minnesota, for example, used transducing temperate phages to establish that two industrially critical traits, lactose-fermenting ability (Lac+) and proteinase activity (Prt+), were encoded by plasmid DNA in Lc. lactis (McKay and Baldwin, 1974; McKay et al., 1976). This observation was important because (1) it provided a biological explanation for industry problems with stability of the acid-producing phenotype (which requires Lac+ and Prt+) in many dairy starter cultures (Sandine et al., 1962); and (2) it presented a simple genetic strategy to alleviate the problem. The latter point is illustrated by the follow-up work of McKay and Baldwin (1978), who isolated Lc. lactis transductants in which the lactose and proteinase genes had integrated into the chromosome, and demonstrated that integration dramatically enhanced the stability of these traits.
Plasmid transduction by virulent or temperate phages has also been demonstrated in S. thermophilus, Lb. salivarius, and Lb. gasseri (Mercenier et al., 1988; Raya et al., 1989; Toyama et al., 1971), but even though this form of gene transfer helped to establish important genetics principles in Lc. lactis, it has not found similar applications in other food-grade LAB. Much of the current disinterest in transduction as a tool for genetic studies or improvements in LAB stems from the relatively narrow host range of transducing phages and, more importantly, development of more effective gene transfer systems such as conjugation and transformation.
The protoplast fusion method of gene transfer is founded upon three key observations: (1) microbial or plant cell walls can be enzymatically removed without deleteriously affecting viability; (2) intercellular membrane fusion can be effected in the presence of polyethylene glycol; and (3) fusants can regenerate a new wall on an appropriate medium. Appropriate selection after cell wall regeneration yields hybrid cells with phenotypic attributes from both parental cell types (Al-foldi, 1982). Gene transfer by protoplast fusion was first demonstrated in plants by Kao and Michayluk (1974), but the technology was soon extended to bacteria (Fodor and Alfoldi, 1976; Schaeffer et al., 1976).
The first protoplast fusion studies in LAB demonstrated exchange of both plasmid-encoded and chromosomally encoded traits between strains of Lc. lactis (Gasson, 1980; Okamoto et al., 1983). Reports of interspecific and even interge-neric gene exchange among LAB followed (Cocconcelli et al., 1986; Iwata et al., 1986; Kanatani et al., 1990; Smith, 1985), and the method has even seen limited application for strain improvement (Stoianova et al., 1988). Overall, however, interest in protoplast fusion technology has never been high because of the need to establish stringent protoplast formation and regeneration conditions for individual strains (Alfoldi, 1982). Nonetheless, protoplast fusion may still be a useful method to combine desirable traits (e.g., production of inhibitors or phage defense systems) from distinct strains, species, or even genera into a single novel bacterium.
Conjugation is a natural form of gene transfer in bacteria that requires physical contact between viable donor and recipient cells. Because it facilitates horizontal gene exchange among populations of both related and unrelated microorganisms, conjugation has weighty implications on bacterial evolution and adaptation (Arber, 2000; Firth et al., 1996). Genes required for conjugative transfer are typically located on self-transmissible plasmids and conjugative transposons, but transfer of nonconjugative plasmids can also be effected via processes termed donation and conduction (Steele and McKay, 1989). The former process applies to nonconjugative plasmids that possess a specific sequence, called the origin of transfer (oriT), that is required for DNA mobilization. Transfer of these plasmids relies only upon trans-acting gene products from a conjugative element and not on cointegrate formation between the nonconjugative and conjugative elements. In contrast, plasmid transfer by conduction does require cointegration, because the nonconjugative molecule lacks a functional oriT. Conclusive evidence for plasmid mobilization by conduction is generally based on presence of cointegrate plasmids in recipient cells (Steele and McKay, 1989).
As a genetics tool for dairy LAB, conjugation has proved especially useful to study plasmid biology in Lc. lactis (Kondo and McKay, 1985; Steele and McKay, 1989). An important outcome of this work has been the finding that many industrially important traits, including lactose and casein utilization, bacte-
riophage resistance, and bacteriocin production, can be transferred by conjugation (Gasson and Fitzgerald, 1994). This situation is of great practical value to the dairy industry, because dairy LAB that are genetically improved by a ''natural'' process like conjugation are not subject to the regulatory and social constraints that shackle the application of recombinant DNA. As a result, several groups have used conjugation to genetically enhance bacteriophage resistance in commercial Lc. lactis starter cultures (Klaenhammer and Fitzgerald, 1994) (see Sec. IV.A for additional details).
Conjugation of native plasmids and chromosomal genes has not been documented as frequently among other dairy LAB, but the ability of many species to participate in conjugation has been established through interspecific and interge-neric transfers of broad host range plasmids such as pAMpi. These observations imply that conjugation may help to support genetics research in the many strains of dairy LAB that still cannot be efficiently or reproducibly transformed (Gasson and Fitzgerald, 1994; Thompson et al., 1999). In addition, conjugation appears to be less sensitive than transformation to the size of the DNA to be transferred, so mobilizable cloning vectors for LAB should also facilitate experiments with relatively large DNA molecules. Systems for delivery of gene-cloning vectors by conjugation have been developed, but efforts fully to exploit the versatility of conjugation as a tool for dairy strain improvement would clearly benefit from a more holistic understanding of conjugal mechanisms in dairy LAB (Romero et al., 1987; Thompson et al., 1999).
At present, the most complete models for conjugation have emerged from studies of the fertility (F) plasmids in gram" bacteria (Firth et al., 1996). From those and other models, we can divide conjugal gene transfer into four basic stages: (1) stable mating pair formation; (2) DNA mobilization; (3) DNA transfer; and (4) mating pair resolution. In gram" cells, formation of stable cell-cell contact requires sex pili which are produced by the donor cell. Gram+ bacteria do not produce pili, however, so stable mating pair formation between LAB must be achieved through other mechanisms. In contrast, homologies between conjugation gene products and noncoding sequences required for DNA transfer suggest that DNA processing and transfer events which follow stable cell-cell contact may occur by similar mechanisms in gram" and gram+ bacteria. This hypothesis is further supported by the fact that conjugation between gram" and gram+ bacteria can occur bidirectionally (Trieu-Cuot et al., 1987, 1988).
In gram" hosts, establishment of a stable mating pair is believed to produce an intracellular signal that initiates DNA mobilization. One strand of the conjugative DNA is cleaved by a conjugative relaxase at a specific locus (nic) within oriT, and a DNA helicase unwinds the nicked strand in the 5'-3' direction. The displaced strand is then transported into the recipient cell in single-stranded form, 5'-3', through a mating bridge that spans both cell membranes. Complementary strand synthesis in the donor and recipient relies on host enzymes and is thought to occur as DNA transfer proceeds. Once DNA transfer is complete, the mating pair actively dissociates and the recipient assumes the conjugative phenotype of the donor cell (for a detailed discussion of conjugal mechanisms in gram" bacteria, see Firth et al., 1996).
The biochemistry of DNA processing and transfer is not nearly as well understood in gram+ bacteria, and much of the information that is available is built from assumptions based on protein and nucleic acid sequence homologies. Two important exceptions to this theme involve mechanisms for efficient mating pair formation and DNA mobilization. In mating pair formation, very good models have emerged from studies of pheromone-inducible plasmid transfer in En. faecalis and, to a lesser extent, from the Lc. lactis sex factor (Dunny and Leonard, 1997; Gasson et al., 1995; Mills et al., 1998). Sound models for DNA mobilization in LAB have also come forward through studies of the streptococcal plasmids pIP501 and pMV158 (Grohmann et al., 1999; Wang and Macrina, 1995).
1. Mating Pair Formation in Lactic Acid Bacteria: Pheromones and Sex Factors
Unlike Lc. lactis and other dairy LAB, En. faecalis is a significant cause of human morbidity and mortality, and conjugation in this species is intimately associated with dissemination of antibiotic resistance genes and virulence factors (Dunny and Leonard, 1997). For these reasons, conjugation in En. faecalis has been studied intensively for more than two decades, and the pheromone-induced plasmid transfer system in this species is now one of the most thoroughly understood mechanisms for efficient mating pair formation in gram+ bacteria. Although En. faecalis is not and should not be used as a dairy starter bacterium, this mechanism has similarity to that used in lactose plasmid conjugation by Lc. lactis and therefore warrants some discussion here.
Several plasmid families and their distinct pheromones have been identified in En. faecalis, but the most thoroughly characterized plasmids are pAD1 and pCF10, which encode production of hemolysin and tetracycline resistance, respectively. Stable mating pair formation in En. faecalis cells containing one of these or another pheromone-induced conjugative plasmid is achieved by a protein-protein interaction that involves aggregation substance (AS) on donor cells and enterococcal binding substance on the recipients. The genetic determinant for AS production is located on pAD1, pCF10, and other pheromone-inducible plasmids, and its expression is induced (along with genes for other conjugative functions) by the presence of recipient-produced pheromone in the growth medium (Dunny and Leonard, 1997).
The En. faecalis sex pheromones are small (seven to eight amino acids in length), hydrophobic, and chromosomally encoded peptides. Most strains produce a number of distinct pheromones that individually can only act on cells that contain a particular plasmid family member. Induction of pAD1 or pCF10 transfer is initiated by internalization of its cognate pheromone (cAD1 or cCF10) into donor cells via pAD1- or pCF10-encoded oligopeptide-binding proteins TraC or PrgZ, respectively, and the chromosomally encoded oligopeptide transport system (Opp). Once inside, the pheromone binds to an intracellular regulatory molecule, which then directs expression of pAD1- or pCF10-encoded conjugation genes. Interestingly, even though regulatory genes on pAD1 and pCF10 have a similar organization and even some DNA sequence homology, induction of plas-mid-coded conjugation genes apparently occurs through very distinct routes (Dunny and Leonard, 1997). Nonetheless, induction results in AS production by donor cells, which leads to rapid cell aggregation and mating pair formation. After a recipient has successfully acquired any member of a particular plasmid family, production of the cognate pheromone for that family is essentially blocked, and the recipient assumes a conjugative phenotype identical to that of the original donor (Dunny and Leonard, 1997).
a. Lactococcal Sex Factor Sex pheromone production has not been detected in Lc. lactis or other dairy LAB, but efficient mating pair formation in the former bacterium is effected by a 135-kD cell surface protein, CluA, that has significant homology to the En. faecalis AS protein (Godon et al., 1994). The gene encoding this protein, cluA, is located on the conjugative plasmid pRS01 in strain ML3 and on a homologous but chromosomally integrated sex factor in the closely related strain 712. This element also encodes a conjugative relaxase whose gene (ltrB, which contains the group II intron described in Sec. II.B.3) lies just upstream of the pRS01 origin or transfer, as well as an enzyme (TraD) that has homology to an Es. coli F plasmid product believed to facilitate transportation of ssDNA into recipient cells (Firth et al., 1996; Gasson et al., 1995; Mills et al., 1998). The lactococcal sex factor shows great promise as a genetics tool for LAB, because it can consummate intergeneric conjugation between Lc. lactis and lactobacilli, leuconostocs, pediococci, O. oeni, and S. thermophilus (D.A. Mills, personal communication). Furthermore, as an integrated element in the host chromosome, the sex factor can reportedly mobilize chromosomal gene transfer in a counterclockwise direction (Gasson et al., 1995).
Discovery and characterization of the sex factor evolved from detailed studies of lactose plasmid conjugation in Lc. lactis ML3 and 712 (Dunny and McKay, 1999; Gasson et al., 1995). Lactose-fermenting ability (Lac+) in these two strains (and several others) is encoded by a nonconjugative 55-kb plasmid, but Lac+ can be transferred by conjugation to other lactococci at low frequency. Some Lac+ transconjugants from these donors form very tight cell aggregates (Clu+) and are able to transfer Lac+ in secondary matings at frequencies 102- to 105-fold higher than those obtained with the parental strains. Genetic analysis revealed that all Clu+ and some Clu~ transconjugants contained a novel 104-kb plasmid formed by ISSI -mediated cointegration between the lactose plasmid (which carries two copies of the IS) and the sex factor. Further study showed lactose plasmid cointegration with the sex factor could occur in more than one orientation, and it was this feature that appeared to determine whether or not a transconjugant was Clu+ (Dunny and McKay, 1999; Gasson et al., 1995). The mechanism(s) by which cointegrate formation induces cluA expression is not yet clear, but the absence of a consensus lactococcal promoter sequence immediately upstream of the cluA gene has led to speculation that it may involve a promoter in ISSI (Gasson and Fitzgerald, 1994; Godon et al., 1994). Other factors must also affect cluA expression, however, because high-frequency transfer of the sex factor itself has also been documented (Gasson, 1995). Nonetheless, the role of CluA in cell aggregation, and the influence of aggregation on conjugation efficiency, are well established (Anderson and McKay, 1984; Godon et al., 1994; Wang et al., 1994).
Additional evidence for a functional analogy between En. faecalis and Lc. lactis mechanisms for efficient mating pair formation was provided by Van der Lelie et al. (1991), who showed the Clu+ phenotype in Lc. lactis may actually involve an interaction between CluA and another lactococcal cell surface component called aggregation substance (Agg). The genetic determinant(s) for Agg has not yet been identified, but the substance appears to be synthesized constitutively by many, although not all, lactococci. Thus, self aggregation only occurs when both cell surface components are expressed by the same bacterium, but efficient mating pair formation can occur between CluA~Agg+ recipients and donor cells that are either CluA+Agg+ (phenotypically Clu+) or CluA+Agg~ (phenotypically Clu"). Taken together, these and other reports of efficient conjugation systems in gram+ bacteria (Jensen et al., 1996; Reniero et al., 1992) indicate that proteinmediated donor and recipient aggregation may be an important mechanism for efficient mating pair formation in bacteria that do not produce pili.
In contrast to mechanisms for mating pair formation, DNA mobilization in LAB and other gram+ bacteria appears to occur through a process very similar to that used by gram" cells (Climo et al., 1996; Grohmann et al., 1999; Guzman and Espinosa, 1997; Wang and Macrina, 1995). Mobilization begins with binding of a conjugative relaxase (frequently called a mobilization or Mob protein) at oriT to form a nucleoprotein complex called a relaxosome, which may or may not include additional proteins. All self-transmissible elements possess an oriT, and as was noted earlier, this cis-acting locus is also found on nonconjugative mobiliza-ble plasmids (which also usually encode a trans-acting relaxase) that can be transferred by donation. The relaxosome initiates DNA transfer by cleaving one strand of the DNA at the nic locus, and then the relaxase remains bound to the 5' end of the oriT locus as DNA transfer proceeds. Biochemically, reactions surrounding nucleophilic attack by the relaxase on a specific phosphodiester bond in nic bear a strong resemblance to those performed by Rep protein during initiation of rolling-circle plasmid replication (see Sec. II.A.1) (Guzman and Espinosa, 1997).
Genes encoding conjugative relaxases and oriT regions (which typically are very close to one another) have been identified on self-transmissible and mo-bilizable elements in several LAB species (An and Clewell, 1997; Dougherty et al., 1998; Guzman and Espinosa, 1997; Jaworski and Clewell, 1995; Mills et al., 1998; Van Kranenburg and De Vos, 1998; Wang and Macrina, 1995). Like oriT regions from other bacteria, most LAB oriT sequences contain a short conserved sequence that can be used to classify these elements into one of three homology groups represented by the nic regions from gram-F-like, IncP, and IncQ plasmids. Exceptions to this observation include the streptococcal plasmid pMV158 and a few other RCR plasmids in LAB, whose oriT regions encompass a homologous sequence named RSA that is also involved in RCR plasmid cointegration (Guzman and Espinosa, 1997). Nonetheless, all of the oriT sequences that have been characterized in LAB (including members of the pMV158 family) contain a noncon-served inverted repeat immediately upstream of the conserved nic region (Table 3). A similar structural arrangement exists in the oriT regions of gram" plasmids, where the inverted repeat is thought to be involved in termination of DNA transfer (Lanka and Wilkins, 1995).
Mobilization of nonconjugative DNA in LAB can also occur by conduction. The most extensively characterized event of this type in dairy LAB is lactose plasmid conduction by the Lc. lactis sex factor (see Sec. III.C.1), where plasmid cointegration is mediated by either of two ISSi elements on the lactose plasmid. Natural conduction of other plasmids following IS-mediated cointegration has also been reported in this species (Romero and Klaenhammer, 1990).
In addition, plasmid cointegrates can be produced by homologous recombination between conjugative and nonconjugative elements, and systems based on this type of plasmid mobilization have been used to transfer gene cloning vectors to various LAB that resist transformation (Romero et al., 1987; Smith and Clewell, 1984; Thompson et al., 1999). Very efficient plasmid conduction can also be induced through cointegration of the conjugative streptococcal plasmid pIP501 with nonconjugative plasmids that are provided with a short, palindromic, recom-binational ''hot spot'' from pIP501 (Langela et al., 1993).
In summary, conjugation is an important instrument for biotechnology in dairy LAB because it provides researchers with a food-grade mechanism for genetic strain improvements, and because it can facilitate genetics research in strains that are difficult to transform. As was noted at the beginning of this section, however, efforts to exploit the versatility of conjugation for these purposes would be served from a more complete understanding of conjugation in LAB. Though much can be inferred from protein and nucleic acid sequence homologies that exist between conjugation systems of gram+ and gram" bacteria, it is important
Table 3 Representative Structures for the Origin of Conjugative Transfer (oriT) in Lactic Acid Bacteriaa
Host genus and element
Nucleotide sequence (5'-3')
oriT Familyc (reference)
Was this article helpful?