Genetics of lactobacilli: plasmids and gene expression - Plasmid vectors(진행중)

Plasmid vectors

General properties

Broad-host-range vectors like the lactococcal plasmid pGK12, which has been shown to replicate in a variety of bacterial strains (Kok et al. 1984), can also replicate in different Lactobacillus species (Bringel et al. 1989; Posno et al. 1991a). These vectors have been instrumental in the development of genetransfer systems for Lactobacillus, at the time that vectors based on Lactobacillus replicons were not yet available. To date a spectrum of plasmid vectors with Lactobacillus replicons has been constructed, allowing genetic manipulation of a wide variety of Lactobacillus species. Most plasmid vectors are derived from small cryptic plasmids of different Lactobacillus species, which replicate through a mechanism of RCR. This may affect the stability of recombinant plasmids, as will be shown in a subsequent section.

 

 

 

Table 2 presents a list of plasmid vectors currently in use in various laboratories. Vectors contain either the erythromycin-resistance (ery) gene from pE194 or pAM~I, or the chloramphen- |col-resistance (cml) gene from pC194 or pBR328 as selection marker. Since lactobacilli are intrinsically resistant to relatively high concentrations of kanamycin/neomycin, these markers cannot be used for vector construction. Also ampicillin resistance cannot be used as selection marker in Lactobacillus (Shimizu-Kadota et al. 1991). Most vectors, as for example pLP825 and pLPE323, display a wide-host-range phenotype, since they can be propagated in a wide variety of Lactobacillus species (Posno et al. 1991a). Moreover, their copy number in different Lactobacillus strains does not significantly differ, indicating that also control mechanisms for DNA replication in different host bacteria operate in a similar way. The average copy number of these plasmid vectors is estimated at 30-50 copies per cell. Copy-number mutants, showing a 3-5 fold elevated copy number, can arise spontaneously during vector construction, probably as a result of selection pressure (Fig. 2).

Fig. 2. Effect of plasmid copy number on the activity of the bacteriocin acidocin B. The acidocin B activity is visualized using Clostridium sporogenes as indicator bacteria, immobilized in agar. To the wells were applied neutralized supernatants of (A) parent strain L. acidophilus M46 with acidocin B gene on a low-copy number plasmid, (B) L. plantarum with low-copy vector (pGKV21) containing acidocin B gene, (C) L. plantarum with high-copy mutant (pLPE24M) containing acidocin B gene. In L. acidophilus M46 and L. plantarum/pGKV12-acidocin B, the copy number is 5-15. The copy number in L. plantarum/ pLPE24M-acidocin B is 50-100.

 

 

 

The mutations, which give rise to the cop phenotype, have not been mapped. A cop mutant of pLPE323, which was purposely made by deletion of the replication repressor gene, shows a 5-10-fold increase of the copy number (Pouwels et al. 1993). Some of the Lactobacillus vectors lacking E. coli sequences can nevertheless replicate in this host organism. For example, plasmids of the pPSC series and pAl-derived vectors are promiscuous plasmids that can be propagated in different Lactobacillus species, in Bacillus and in E. coli (Cocconcelli et al. 1991; Vujcic & Topisirovic 1993). Similar findings have been reported for lactococcal plasmids like pSH71 and pWV01 (Kok et al. 1984; Gasson & Anderson 1985; de Vos 1987). Replication proteins of RCR plasmids generally are expressed in E. coli, suggesting that replication of these plasmids in E. coli depends on the presence of host factors that stabilize ss-DNA intermediates and/or initiate replication at the minus origin. Recently, a series of broadhost-range vectors was constructed for Lactococcus based on the replication functions of the theta-type plasmid pAM[31 (Simon & Chopin 1988; O'Sullivan & Klaenhammer 1993). They might be useful for cloning in Lactobacillus as plasmids replicating by a theta-type mechanism show structural and segregational stability (Swinfield et al. 1991; Br0ckner 1992). Recently, an improved version of vector pLPE323, named pLPE23M (Fig. 3) was obtained by introduction of a multi-linker region with 19 unique restriction enzyme sites. The usefulness of the vector has been demonstrated by cloning and overexpression of several proteins in Lactobacillus. Also a vector with narrow host-range has been described. Plasmid pLUL631 from L. reuteri carrying an erythromycin-resistance gene was found to replicate in L. reuteri and in a strain of L. fermentum among several lactobacilli and other Gram-positive bacteria tested (Ahrn6 et al. 1992). Similarly, a 3.6 kb plasmid replicon from L. crispatus was found to replicate only in the host strain from which it was derived (Posno pers. comm.). The latter type of vectors offers attractive properties with regard to safety aspects associated with the use of live recombinant DNA organisms in e.g. food products. Vectors with a narrow host range are less likely to be horizontally transferred to other bacterial species than vectors based on broad-host-range replicons, and are, consequently, intrinsically more safe. Vectors have also been described which, potentially, are useful for the development of food-grade vectors. Plasmid pBG10, which carries the L. bulgaricus gene encoding J3-galactosidase under control of the promoter of the ery gene from pAM[31, might be useful for the selection of transformants in milk where lactose is the sole energy source (Hashiba et al. 1992). Plasmid pLP3537xyl, which contains genes from L. pentosus involved in the catabolism of xylose, is capable of conferring to lactobacilli the capacity to utilize xylose as sole energy source, a trait which is infrequently found in lactobacilli (Posno et al. 1991b).

 

Incompatibility

Lactobacillus strains are cured from the endogenous plasmid when transformed by a vector with a replicon from that plasmid (Bringel et al. 1989; Posno et al. 1991a; Leer et al. 1992). This finding, which can be easily explained by selective advantage of vector DNA over the resident plasmid, can be exploited to cure strains from plasmids that are otherwise difficult to eliminate. Also functional relationships between replication functions in different plasmids can be assessed in this way. For example, the lactococcal plasmid pGK12 was found to cure L. pentosus MD353 from an 1.7 kb plasmid, indicating that the plasmids share replication functions, although they show little DNA homology (Posno et al. 1991a). L. plantarum ATCC8014 and L. pentosus MD353 were reciprocally cured of the endogenous plasmids p8014-2 and p353-2, when they were transformed by vectors based on these plasmids. The assumption that the two plasmids have similar replication functions was verified by sequence analysis, showing that the two plasmids encode replication proteins displaying 94 % similarity and have identical target sites for these proteins (Leer et al. 1992).

 

Fig. 3. Structure of plasmid pLPE23M. Plasmid pLPE23M is constructed by insertion of an erythromycin-resistance gene from pE194 and a multi-cloning oligonucleotide into the XbaI site of plasmid p353-2 from L. pentosus. In plasmid pLPE24M (see Fig. 2) the erythromycin-resistance gene was cloned in the opposite orientation.

 

Structural stability

For application in industrial fermentation processes, it is essential that the vector remains structurally intact (structural stability) and can be maintained in the host cell in the absence of selective pressure during the fermentation process (segregational stability). Most plasmid vectors remain structurally intact when the bacteria are cultivated in the presence or absence of the selective agent. No structural instability in L. casei or L. pentosus was observed after insertion into a Lactobacillus vector of DNA fragments derived from bacteriophage )~ varying in size from 2 to 9 kb, irrespective of whether the bacteria were cultivated in the presence or absence of the selective agent (Leer et al. 1992). Complete structural stability was also observed when a 4 kb DNA fragment coding for a hybrid protein consisting of a foot-and-mouth disease virus epitope and E. coli [3-galactosidase or a 7.2 kb DNA fragment from L. pentosus comprising genes involved in xylose catabolism were cloned in a Lactobacillus vector (Posno et al. 1991b). However, attempts to clone the Hepatitis delta surface protein under control of a Lactobacillus promoter in L. casei did result in transformants with plasmids of the expected size only, when the cloned gene lacked proper translation-initiation signals (Jore, pers. comm.). Apparently, the expression of a protein which is harmful to lactobacilli can be obviated by the occurrence of deletions. Recombinant DNA could also be stably maintained without structural changes when DNA was inserted into the chromosome (Scheirlinck et al. 1989; Bhowmik & Steel 1993; Leer et al. 1993). It appears that except for cases where expression of the cloned gene results in a deleterious protein, cloning of homologous and heterologous DNA into Lactobacillus offers no serious problems. Even relatively large fragments can be stably maintained without the occurrence of detectable deletions or rearrangements.

 

Segregational stability

Accurate copy number control is required for stable plasmid maintenance. Since RCR plasmids appear to lack a partitioning function, plasmids most probably are randomly distributed over daughter cells (Novick 1987). Large fluctuations in copy number as a result of a disturbance of copy number control systems may lead to daughter cells receiving no plasmid molecules. For the development of vectors that can be stably maintained, knowledge about plasmid or host factors that contribute to copy number control thus is of paramount importance. Most vectors are rapidly lost (50-> 95 % loss after 100 generations) when lactobacilli are cultivated in the absence of the selective agent (Bringel et al. 1989; Posno et al. 1991a; Shimizu-Kadota et al. 1991). Vectors with a replicon from Lactococcus or Staphylococcus are even less stable in Lactobacillus, showing segregation rates of several percent per generation (Posno et al. 1991a; Shimizu-Kadota et al. 1991). Some vectors with Lactobacillus replicons can, however, be stably maintained for more than 100 generations in Lactobacillus in the absence of selective pressure (Posno et al. 1991a; Cocconcelli et al. 1991; Leer et al. 1992). Of particular interest is plasmid pLPE323 which was found to be segregationally fully stable in all except one Lactobacillus strain (Leer et al. 1992). Plasmid pLP3537, which is rapidly lost under non-selective conditions, harbours the same replicon as pLPE323 but differs from it by the presence of E. coli vector sequences. The difference in stability between the two plasmids can be fully accounted for by E. coli sequences. After removal of ~ 95% of the E. coli sequences from pLP3537, the resulting vector had become segregationally completely stable. A similar result has been obtained for pLP825 which comprises pBR322 sequences and a replicon from L. plantarum. Segregational instability increases as the size of the inserted DNA fragment increases (Leer et al. 1992), like a phenomenon also observed in B. subtilis for plasmids derived from pUB110 (Bron & Luxen 1985; Bron et al. 1988a, b). In studies with RCR plasmids from S. aureus (Gruss et al. 1987), Streptococcus pneumoniae (Del Solar et al. 1987), Streptomyces lividans (Deng et al. 1988) and B. subtilis (Bron 1990), a functional minus origin has been implicated in plasmid segregational stability, probably because of inefficient synthesis of the minus strand of the plasmid. A plasmid with a deletion removing one-half of the stem-loop, which is involved in the conversion of ss-DNA to ds-DNA, is considerably less stable than the parent plasmid in two Lactobacillus species, indicating that also in Lactobacillus a functional minus-origin of replication is important for segregational stability (Leer et al. 1992). Likewise, Shimizu-Kadota and coworkers have observed that introduction of a palA-type or M13 minus origin into a pUB110-derived vector increases the segregational stability of the vector in L. casei (Shimizu-Kadota et al. 1991). Increased segregational stability was positively correlated with a decrease of the amount of ss-DNA intermediates (Shimizu-Kadota et al. 1991; Leer et al. 1992), indicating that inefficient conversion of ss-DNA into ds-DNA in Lactobacillus disturbs copy number control and, consequently, the segregational stability.