WO1994006917A1 - Production of desired proteins or polypeptides by culturing a transformed lactic acid bacterium - Google Patents

Abstract

The invention provides a food-grade recombinant plasmid comprising (1) a replicon and DNA sequences derived from a lactic acid bacterium (LAB), required for stable maintenance in a LAB, preferably derived from plasmid pWVO2 from L. lactis subsp. cremoris Wg2, and (2) a structural gene encoding a desired protein or polypeptide or its precursor, not being an antibiotic resistance selection marker, and optionally (3) a food-grade selection marker, whereby said plasmid has a theta-replication mechanism and is stably maintained in the transformed LAB under non-selective conditions. Also provided is a LAB transformed by said recombinant plasmid and the use of such transformed LAB in a process for producing said protein or polypeptide. Said recombinant plasmid can also be used for modifying the metabolic pathway of a LAB, e.g. for producing diacetyl or a precursor thereof by a LAB. Also claimed are food and animal feed products obtained by incorporating therein said protein or polypeptide or precursor thereof produced by said transformed LAB.

Description

Title: PRODUCTION OF DESIRED PROTEINS OR POLYPEPTIDES BY

CULTURING A TRANSFORMED LACTIC ACID BACTERIUM

Background of the invention and prior art

The invention relates to a process for producing a desired protein or polypeptide, which process comprises culturing a transformed lactic acid bacterium under conditions whereby a structural gene encoding said desired protein or polypeptide or a precursor thereof is expressed. Such a process is known from several patent publications. Several vectors suitable for transforming lactic acid bacteria to be used in such a process are described, which vectors are replicable in Escherichia coli, Bacillus subtilis and Streptococcus lactis.

For example,

in EP-A-0,157,441 (UNILEVER) a process is described in which the

transformation of lactic acid bacteria is performed by using a plasmid based on plasmid pWVO1 derived from Streptococcus cremoris Wg2, or similar plasmids from other lactic acid bacteria;

- in EP-B-0,228,726 (STICHTING NEDERLANDS INSTITUUT VOOR ZUIVEL-OΝDERZOEK, also known as ΝIZO) a similar process is described in which lactic acid bacteria are transformed with a plasmid based on the cloning vector pΝZ12, or similar vectors, containing a DNA fragment having replicon activity derived from lactic acid bacteria, e.g. from the smallest Streptococcus lactis NCDO712 plasmid pSH71 as described in Example II. Also mentioned in this publication are vectors pCKl and pCK21 carrying the replicon of the streptococcal plasmid pSH71 and that replicate in E. coli, B. subtilis and S. lactis as described in lines 18-24 on page 4;

- in EP-A-0,251,064 (VALIO MEIJERIEΝ KESKUSOSUUSLIIKE) a method for the preparation of an industrially utilizable metabolite is described, in which a microorganism is used that contains a cloning vector also derived from pSH71 and capable of transforming group Ν streptococci, E. coli, and B. subtilis. In WO 91/09131 and WO 91/09132 (both of VALIO FINNISH COOPERAΗNE DAIRIES ASSOCIATION) cloning vectors for use in lactic acid bacteria are described that do not replicate in B. subtilis or E. coli, in particular a food-grade vector indicated as pVS40. However, the experiments described in Examples 4 and 5 show that this plasmid pVS40 is not stably maintained under non-selective conditions. In fact it had to be modified by incorporation into the plasmid of an expressible lactose gene that made it possible for the resulting transformant to ferment lactose so that it could grow under the selection pressure of lactose. Further use was made of an antibiotic resistance selection marker, namely the nisin resistance gene.

Other prior art will be indicated during the discussion of the work resulting in this invention. Summary of the invention

The present invention is based on findings resulting from a research program, in which the replicating mechanism of several plasmids present in lactic acid bacteria was investigated. The results of this research program are described in Parts I-V given later in this specification.

Part I presents an investigation of the mode of replication of derivatives of both the rolling-circle-type plasmid pWVO1 from Lactococcus lactis subspecies cremoris Wg2 (formerly known as Streptococcus cremoris Wg2) and the theta-type plasmid pAMß1, a broad-host-range plasmid from Enterococcus faecalis (formerly known as Streptococcus faecalis). It turned out that the derivatives of plasmid pAMß1 were stably maintained in lactococci, whereas pWVO1 -derivatives were stably maintained only when the inserts were 8 kb or less, if the strain containing the plasmids was grown without selection pressure for several generations. The latter is a condition needed for any food-grade application or large scale fermentation. Moreover, it was found that the insert length in pWVO1-derivatives was inversely related to the copy number, resulting in a lower expression of the genes situated on the inserted fragments. This is in contrast to the pAMß1-derivatives. However plasmid pAMß1 is not a food-grade plasmid, so that it cannot be used for the construction of a food-grade expression vector. Thus a need exists for the development of food-grade vectors that might replicate via a theta-mechanism, because it may be expected that such vectors can be stably maintained in lactic acid bacteria without selection pressure. In Part II an investigation of the mode of replication in lactococci and the structure of the known plasmid pWVO2 from Lactococcus lactis subspecies cremoris Wg2 is described. The results of this investigation reveal structures that are typical for theta-replicating plasmids. Thus this specification presents the first proof of the existence of theta-replicating plasmids in lactococci. In particular, evidence is presented that the lactococcal plasmid pWVO2 replicates via a theta-mechanism. Moreover, an important DNA sequence is the 1.3 kb EcoRV fragment involved in stable

segregation of the plasmid. In Part III further investigations on the segregational stability of pWVO2 derivatives are described. The results show that the maintenance of pWVO2 and a derivative containing a 12 kb DNA insert are indeed superior to the rolling-circle-replicating (RCR) plasmid pWVO1 and similar derivatives thereof, and comparable to a pAMß1 derivative containing the same 12 kb DNA insert. Replication of all plasmids was tested without antibiotic selection for at least 120 generations. It is expected that a 100% stability of the pWVO2 derivative will be obtained by insertion of the 12 kb DNA fragment at another position in the plasmid.

In Part IV an investigation of the mode of replication in lactococci and the structure of the known plasmids pWVO4 and pWVO5 from Lactococcus lactis subspecies cremoris Wg2 is described. It appears that they also replicate according to a theta-mechanism like pWVO2 contrary to the RCR-plasmid pWVO1. Thus the majority of lactococcal plasmids belong to one family using the theta-mode of replication.

Based on these results one can subsequently convert pWVO2 into a manageable, food-grade, stable, autonomously replicating vector for homologous or heterologous gene expression, by insertion of suitable DNA elements, for example a regulon (promoter and ribosome binding site), followed by a multiple cloning site for insertion of a structural gene encoding a desired protein or polypeptide, a terminator for transcription termination and a, preferably food-grade, selection marker, e.g. the ability to grow on lactose, sucrose or raffinose as only carbon source. Such elements have been described before, e.g. in EP-B-0, 157,441 (UNILEVER), EP-B-0,228,726 (NIZO), EP-A-0,251,064 (VALIO), EP-A-0,487,159 (UNILEVER), and EP-A-0,355,036 (NIZO).

In Part V the development of a stable food-grade vector based on pWVO2 for transforming Lactococcus lactis is described. The sucrose operon from Pediococcus pentosaceus was used as a food-grade marker on the vector, allowing easy detection of the plasmid in host cells by a simple colour indicator assay. The lactococcin A operon from L. lactis was used as a model to demonstrate the successful cloning of a gene of interest. Both the cloning vector and the recombinant plasmid containing the lactococcin A operon were stably maintained when the host cells were grown in glucose-containing media, i.e. under non-selective conditions. One of the constructed vectors was developed into a food-grade expression vector for L. lactis.

Thus the invention is based on the discovery that known lactococcal plasmids like pWVO2, pWVO4 and pWVO5 replicate via a theta-mechanism. After having made this discovery a method was developed for the construction of food-grade expression vectors that can be used for the transformation of lactic acid bacteria in order to give these bacteria new or improved properties. These expression vectors thus comprise a gene encoding a desired protein or polypeptide, in addition to a replicon that works according to a theta-replication mechanism and any other DNA sequences that are responsible for stable maintenance of the plasmid vector in the lactic acid bacterium under non-selective conditions.

Although the invention is exemplified on the basis of a DNA sequence derived from plasmid pWVO2 that replicates according to a theta mechanism, the specification provides sufficient information to enable a skilled person to isolate other similar replicating sequences from Lactococcus lactis subspecies cremoris Wg2 or other lactic acid bacteria without undue burden, which replicating sequences can be used for constructing plasmids that fall within the scope of the present invention and can thus be used for transforming lactic acid bacteria, which in turn can be used in a process for producing a desired protein or polypeptide or precursor thereof according to the present invention. Consequently the present invention provides a process for producing a desired protein or polypeptide comprising culturing a transformed lactic acid bacterium under conditions whereby a structural gene encoding said desired protein or polypeptide or a precursor thereof is expressed, in which process said lactic acid bacterium is one being transformed by means of introducing a food-grade recombinant plasmid, said plasmid comprising

(1) a replicon and any DNA sequences required for stable maintenance in a lactic acid bacterium, whereby

(1.a) said plasmid has a theta-replication mechanism and

(1.b) said plasmid is stably maintained in said transformed lactic acid bacterium under non-selective conditions,

and

(2) a structural gene encoding said desired protein or polypeptide or a precursor thereof, said gene not being an antibiotic resistance selection marker,

and optionally

(3) a selection marker.

Preferably the replicon and any DNA sequences required for stable maintenance in the bacterium are derived from pWVO2 originating from Lactococcus lactis subsp. cremoris Wg2. An important DNA sequence is the 1.3 kb EcoRV fragment involved in stable segregation of the plasmid.

If the use of a selection marker is desirable, it should also be a food-grade selection marker. The use of a selection marker is advantageous in the development stage, but is not needed in the lactic acid bacterium used as production strain.

In this specification "food-grade" means that the food-grade product, e.g. a lactic acid bacterium or a plasmid or DNA sequence, should be derived from, or be identical to, a product that is acceptable in the preparation of foodstuffs, for example derived from a microorganism that is Generally Recognized As Safe (having a so-called GRAS status); or is used for many years in the preparation of foodstuffs. Preferably the food-grade product is derived from, or identical to material derived from, lactic acid bacteria.

For this embodiment the invention further provides a transformed lactic acid bacterium, suitable for use in a process for producing a desired protein or polypeptide according to the invention, said bacterium containing a food-grade recombinant plasmid comprising (1) a replicon and any DNA sequences required for stable maintenance in a lactic acid bacterium, whereby (1.a) said plasmid has a theta-replication mechanism and (1.b) said plasmid is stably maintained in said transformed lactic acid bacterium under non-selective conditions, and (2) a structural gene encoding said desired protein or polypeptide or a precursor thereof, said gene not being an antibiotic resistance selection marker, and optionally (3) a selection marker, whereby the same preferences apply as mentioned for the process described above. According to this embodiment the invention also provides a recombinant plasmid, suitable for preparing a transformed lactic acid bacterium according to the invention, comprising (1) a replicon and any DNA sequences required for stable maintenance in a lactic acid bacterium, whereby (1.a) said plasmid has a theta-replication mechanism and (1.b) said plasmid is stably maintained in said transformed lactic acid bacterium under non-selective conditions, and (2) a structural gene encoding said desired protein or polypeptide or a precursor thereof, said gene not being an antibiotic resistance selection marker, and optionally (3) a selection marker, whereby the same preferences apply as mentioned for the process described above.

Another embodiment of the invention is the use of a recombinant plasmid according to the invention for modifying the metabolic pathway of a lactic acid bacterium. For example, such use can be for increasing the production of diacetyl or a precursor thereof by a lactic acid bacterium by transforming said bacterium with a recombinant plasmid according to the invention, which plasmid is capable of expressing a gene encoding α-acetolactate synthase. The enhanced production of diacetyl or a precursor thereof by enhancing the expression of α-acetolactate synthase was described in EP-A-0,500,188 (UNILEVER). As vectors one can use the less stable, known plasmid vectors based on the food-grade lactococcal plasmid pWVO1 (EP-B-0, 157,441) and pNZ123 (EP-A-0,228,726) or the non-food-grade pAMß1 derivative pIL253 (Simon, D. and Chopin, A. (1988) Biochemie 70: 559-566). However, the present invention now provides more stable, food-grade vectors for such use. Another example of such use is one whereby the plasmid contains one or more genes encoding enzymes, that are capable of influencing the formation and structure of exopolysaccharides thereby producing functionally modified exopolysaccharides.

According to this embodiment the invention also provides a lactic acid bacterium having a modified metabolic pathway obtained as a result of such specific use according to the invention.

In this specification a "transformed lactic acid bacterium" means not only the lactic acid bacterium obtained after the actual transformation step, but also the subsequent generations still containing the additional DNA sequence providing a new or improved property to the bacterium.

Still another embodiment of the invention is a food product obtained by the use of a process for producing a desired protein or polypeptide according to the invention, or by using a lactic acid bacterium according to either embodiment of the invention. Examples of such a food product are dairy products, especially fermented dairy products like yoghurt, quark and fermented milk; ice cream and other frozen products containing milk or fermented milk; meat and meat-containing products; dressings and sandwich spreads; butter and margarine; low-fat butter- and margarine-substitutes; and mayonnaise and low-fat mayonnaise-substitutes.

In addition to food products the invention provides an animal feed product obtained by the use of a process for producing a desired protein or polypeptide according to the invention or by using a lactic acid bacterium according to either embodiment of the invention.

A further embodiment of the invention is a process for the production or

overproduction of a food-grade enzyme to be used for further enzymatic

modifications, in which a transformed lactic acid bacterium according to the invention is used under conditions whereby the enzyme is produced in a relatively high amount and can exert its action on a substrate for the enzyme. An example is the production or overproduction of a peptidase suitable for debittering of protease-treated industrial proteins, e.g. casein, whey-protein or soy-protein or gluten, for both food and non-food use, or to obtain peptide fractions with other food-functional properties.

Legends to the Figures

Figure 1. Plasmids pKS100 and pAMS100.

Only sites relevant for the construction and properties of the plasmids are shown. Open reading frames, genes and the T1T2 transcriptional terminators are indicated. Details of the construction are described in Part I of the specification.

Figure 2. Kinetics of appearance of plasmid-free cells of representative plasmids in

L. lactis.

Exponentially growing cultures of plasmid-carrying MG1363 strains in medium with

Cm were diluted into antibiotic-free medium and subcultured for approximately 140 generations. After every 10 to 20 generations, samples were plated onto non-selective agar. Colonies were subsequently assayed for resistance to Cm after transfer to selective plates.

o = pKS100 and all pAMS derivatives; ∇ = pKS108L; ▲ = pKS112L. Figure 3. Relation between plasmid size and segregational stability in L. lactis.

The percentage of Cm^R colonies obtained after 120 generations of growth in non-selective media is shown as a function of the size of the inserts. Filled symbols represent plasmids containing the original inserts; open symbols plasmids containing the 171 and 112D inserts (see Part I of the specification for explanation).

o ,• = pKS derivatives; ▲ = pAMS derivatives.

Figure 4. Analysis of plasmid DNA from L. lactis.

Total DNA was extracted, separated on agarose gels, blotted and hybridized using pKS100 or pAMS100 as probes.

Lanes 1, 2 and 3: pKS100; lanes 4, 5 and 6: pKS101C.

Lanes 1, and 4: untreated lysates;

lanes 2, and 5: total lysates digested with BglII; lanes 3, and 6: total lysates digested with an enzyme cutting the plasmids once (PstI). cc, covalently closed monomers; oc, nicked monomers;

hm, high-molecular-weight plasmid forms; lin, linearized plasmids. Figure 5. Copy numbers in circular and HMW plasmid fractions of pKS derivatives.

The copy numbers per chromosome equivalent of the circular monomeric (covalently closed plus open circular) and the total plasmid fraction (circular plus HMW) were determined. The number of plasmid copies in the HMW fraction was calculated by subtracting the copy numbers in the circular monomeric fraction from those

determined for the total plasmid DNA.

Filled bars: circular plasmid fraction; hatched bars: HMW fraction.

Figure 6. Relation between copy numbers of the circular monomers and plasmid stability.

The percentages of Cm^R cells correspond to the percentages of plasmid-containing cells after 120 generations of growth in the absence of Cm. These values were taken from the experiments described in Figures 2 and 3.

Figure 7. Relations between mass of HMW, plasmid copy numbers and plasmid stability in L. lactis.

The mass amounts of HMW, expressed per chromosome equivalent, were calculated as described in Part I of the specification. Construct names are indicated. Solid line: plasmid maintenance (percentage Cm^R colonies). Figure 8. Detection of ssDNA in total cellular DNA of strains carrying either the

RCR-plasmid pGK1 (A) or pWVO2 (B).

It is known that pGK1 has an effective minus origin. It was possible, however, to demonstrate ssDNA of this plasmid in total cellular DNA, by prolonged exposure of the film. No ssDNA of pWVO2 could be detected. The position of ssDNA is indicated. Figure 9. Physical map of pLR300.

The region representing pWVO2 is indicated with a thick line. The insert, containing the erythromycin resistance gene of pE194 was inserted into the HpaI site of pWVO2, located at position 1 of the sequence. Relevant sites are indicated. DR: 3½ 22 bp direct repeat; AT: AT-rich region. The structure of the repeated region (repeats) is given in Figure 10.

Figure 10. Organizing structure of the 1.8 kb direct repeated region, located between positions 1985 and 3786 of the sequence of pWVO2.

It consists of three different repeats of 50 bp, 100 bp or 200 bp, lying head to tail. The presence of the 5'-GATGA-3' sequence within these repeats is indicated by small dashes above the repeats. Remarkably this repeat is not found within the 200 bp repeat. Figure 11. A. Replication intermediates of pLR300 in total cellular DNA of L.

lactis subsp. lactis MG1363.

Total cellular DNA was cut with BglII and loaded for 2-D agarose gel electrophoresis and detection by ECL as described in Experimental Procedures.

B. Diagram of idealized 2-D gel patterns, generated by the four basic forms of replication intermediates.

Simple Ys (pattern b in Figure 5A) are maximally branched, and therefore maximally retarded in the second dimension, when the molecule is 50% replicated. They start from the arc of linear and they return to it as they approach 100% replication.

Bubbles (pattern a in fig 5A) are more branched, and therefore more retarded than simple Ys at all extends of replication. Double Ys (pattern c in fig 5A) start from any position on the Y arc and join a near vertical line above the 2x point at a position which depends on how far replication has proceeded. Termination structures

(presented by pattern d in Fig 5 A) arise nearly vertically from the position of 100% replication. The position depends on the location of the termination point from directly opposite of the restriction site (most retarded) to directly adjacent of the restriction site (least retarded). Figure 12. Physical maps of pLR312L and pLR100.

Relevant sites and features are indicated.

Figure 13. Kinetics of appearance of plasmid-free cells of studied plasmids in L.

lactis subsp. lactis MG1363.

Cultures of plasmid-carrying MG1363 strains were diluted into antibiotic-free medium and subcultured for a considerable number of generations. Samples were taken at regular intervals and plated onto non-selective agar. Colonies were subsequently assayed for resistance to the appropriate antibiotic after transfer to selective plates. o = pLR300 and pAMS112L; ▲ = pKS112L;

+ = pLR100/pLR200 ∇ = pLR312L;

Figure 14. Diagrams, showing the construction of plasmids pJR04 and pJR7 and the organizing structure of the plasmids they were derived from.

For details see Part IV of the specification.

Figure 15. A: Plasmid profiles of wild-type Lactococcus strains.

1. L. lactis Hpl; 2. L. lactis subsp. lactis F16;

3. L. lactis subsp. lactis H61; 4. L. lactis subsp. lactis E8S;

5. L. lactis subsp. cremoris C916; 6. L. lactis subsp. cremoris Cla2;

7. L. lactis subsp. cremoris C109; 8. L. lactis subsp. lactis L10.

B: Fluorograph obtained after hybridization of DNA in panel A with the 1 kb PCR fragment of the rep gene of pWVO2.

Figure 16. Limited restriction map of the replication regions of plasmids pWVO2, pWVO4 and pWVO5.

None of the shown restriction sites is present in all three plasmids. Some sites are found in two of three plasmids (HindIII and DdeI in pWVO2 and pWVO5; RsaI, location 479 and TaqI, location 1175 in pWVO4 and pWVO5). The restriction patterns obtained from pWVO3 resembled that of pWVO2 (results not shown). Figure 17. Restriction map of plasmid pINT123E.

The relevant genes for the sucrose metabolism are indicated: scrA, scrB, and a truncated scrR gene (scrR'), as well as the erythromycin resistance gene Em^R. Figure 18. Construction of the plasmids pLR330 en pLR331.

Only sites relevant for construction of the plasmids are shown. Open reading frames are indicated. Details of the construction are described in Part V of the specification.

Figure 19. Overview of the construction of the food-grade vector pLCN312.

Details are described in Part V of the specification.

Figure 20. Restriction maps of plasmids pORI28 (A) and pORI128 (B).

Details are described in Part V of the specification. Figure 21. Kinetics of appearance of cells free of plasmid pLCN312 in liquid SM17 media.

Inocula were taken from colonies on SM17 plates. At regular intervals, appropriate dilutions of the cultures were made in GM17 or in SM17 medium and growth was continued for about 80 generations. Samples were plated on SM17 agar with BCP. The yellow colonies were plasmid containing; white colonies had lost the plasmid.

● - - - - - - -● M17 medium plus glucose

0 - - - - - - - 0 M17 medium plus sucrose

Figure 22. Overview of the construction of the food-grade expression vector pJR201. Details are described in Part V of the specification.

Detailed description of the invention

The results of the research investigations leading to the present invention are described in Parts I-V given below, which are based on draft publications. Parts I-IV were also present in the priority document. Part I has been published during the priority year (65). Part V was not present in the priority document. PART I

THE MODE OF REPLICATION IS A MAJOR FACTOR IN SEGREGATIONAL PLASMID INSTABILITY IN Lactococcus lactis

SUMMARY

We studied the effects of the rolling-circle and theta-modes of replication on the maintenance of recombinant plasmids in Lactococcus lactis. Heterologous Escherichia coli or bacteriophage λ DNA fragments of various size were inserted into vectors based on either the rolling-circle-type plasmid pWVO1 or the theta-type plasmid pAMß1. All pAMß1 derivatives were stably maintained. pWVO1 derivatives, however, showed size-dependent segregational instability, in particular when large DNA fragments were inserted. All recombinant pWVO1 derivatives generated high-molecular-weight plasmid multimers (HMW) in amounts that were positively correlated with plasmid size and inversely correlated with the copy numbers of the monomeric plasmid forms. Formation of HMW, or reductions in copy numbers were not observed with pAMß1 derivatives. The results indicate that HMW formation and/or reduction in plasmid copy numbers are important factors in the maintenance of pWVO1 derivatives. We concluded that theta-type plasmids are superior to rolling-circle-type plasmids for cloning in lactococci.

INTRODUCTION Lactococci are important organisms in dairy and other food fermentations.

Considerable progress in studies on the molecular genetics of these bacteria has been made after the development of cloning vectors based on the strongly related lactococcal plasmids pWVO1 (21, 22) and pSH71 (12, 13). Both replicons have an extensive host range, sustaining replication in many gram-positive bacteria, like Lactococcus, Lactobacillus, Leuconostoc, Pediococcus, Propionibacterium,

Staphylococcus, Streptococcus, Clostridium, Listeria and Bacillus species, as well as in some strains of the gram-negative bacterium Escherichia coli (21, 22, 28, 33). Analyses of the nucleotide sequences and the mode of replication have revealed that pWVO1 and pSH71 belong to the class of small plasmids that generate single-stranded DNA (ssDNA) intermediates during rolling-circle-replication (16, 26). From studies with Bacillus subtilis it is known that RCR-type plasmids show several problems: they are frequently refractory to the cloning of large inserts, and recombinant plasmids often show a high degree of instability. Several studies have indicated that the formation of ssDNA intermediates is an important factor in both structural instability (3, 6, 14, 16) and segregational instability (4, 5, 6, 16). In addition to ssDNA, the formation of linear high-molecular-weight plasmid multimers (HMW) by RCR-plasmids (15, 36) has been implicated in structural instability (27) and segregational instability (16, 37).

In addition to plasmids of the RCR-type, plasmids that replicate via a theta-mechanism have been used for the construction of cloning vectors in gram-positive bacteria. Theta-type plasmids neither generate ssDNA replication intermediates, nor do they form HMW products in wild-type B. subtilis strains (15). Several cloning vectors are based on pAMß1 (19, 32), a broad-host-range Enterococcus faecalis plasmid, that replicates according to a unidirectional theta-mechanism (8).

Structurally, these plasmids are very stable in B. subtilis (19). Some variants, however, are segregationally unstable in this organism (32). The latter property was attributed to the absence of a stability determinant, assumed to specify a plasmid resolution function (34).

So far, no systematic studies have been reported on the effects of heterologous DNA inserts on plasmid maintenance in lactococci. This question is of relevance for the development of efficient and stable host/vector systems for these bacteria. The major aim of the present studies was to compare the effects of the size of DNA inserts on the maintenance of RCR- and theta-type plasmids in Lactococcus lactis. For this comparison, derivatives of the plasmids pWVO1 (RCR-type) and pAMß1 (theta-type) were chosen. Evidence is presented that the theta-type plasmid is superior to the RCR-type plasmid for cloning in L. lactis. MATERIALS AND METHODS

Bacterial strains and plasmids

Bacterial strains and plasmids used in this study are listed in Table 1.

Table 1. Bacterial strains, plasmids, and DNA inserts bacterial strain properties source or or plasmid reference or insert

Bacterial strain

E. coli JM101 supEthi Δ(lac-proAB) [F'traD36,

proAB, lacI^qlacZΔM15] (30) B. subtilis 8G5 trpC2, tyrI, met, his, nic, purA, ura, rib (4) L. lactis MG1363 ssp lactis, plasmid-free, Lac-, Prt- Lab collection * L. lactis Wg2 ssp cremoris, industrial strain Lab collection *

* Lab = Department of Genetics, Centre of Biological Sciences of Rijks- Universiteit Groningen (RUG), The Netherlands

Plasmids

pWVO1 2.2 kb; cryptic plasmid from L. lactis Wg2 (21) pIL252 Em^R, 4.7 kb; unstable pAMß1 deletion derivative (32) pMTL23 Ap^R, 2.4 kb; pUC derivative containing lacZα gene (9) pC194 Cm^R, 2.9 kb; Staphylococcus aureus plasmid (17) pE194cop6 Em^R, 3.7 kb; Staphylococcus aureus plasmid (18) pSR11 Em^R, 6.8 kb; stable pIL252 derivative This work pAMS100 Em^R, Cm^R, 8.4 kb; pAMß1/pSR11 derivative This work pAMS101 - pAMS171 pAMS100 with various heterologous

inserts in the BclI site This work pKS100 Em^R, Cm^R, 5.0 kb; pWVO1 derivative This work pKS101 - pKS171 pKS100 with various heterologous

inserts in the BclI site This work

DNA inserts

1C 1.2 kb; E. coli chromosomal DNA (4)

3C 4.2 kb; E. coli chromosomal DNA (4)

5L 5.6 kb; phage λ DNA; coordinates 22346-27972 (31)

7L 7.2 kb; phage λ DNA; coordinates 34499-41732 (31)

8L 8.4 kb; phage λ DNA; coordinates 415-8844 (31) 10L 10.2 kb; phage λ DNA; coordinates 38814-0-415 (31)

12L 12.2 kb; phage λ DNA; coordinates 41732-0-5505 (31)

16L 16.8 kb; phage λ DNA; coordinates 5505-22346 (31)

112D 8.0 kb; shortened 12L fragment This work

171 8.4 kb; 1C + 7L fragments This work

Media

L. lactis was grown at 30°C in M17 medium (35) containing 0.5 % glucose (GM17). E. coli and B. subtilis were routinely cultured in TY broth (tryptone, 10 g/l; yeast extract, 5 g/l; and NaCl, 10 g/l; pH 7.4) at 37°C. When B. subtilis was grown to competence, the media and procedures described in reference 2 were used. When required, media were supplemented with antibiotics to the following concentrations: chloramphenicol, 5 μg/ml; erythromycin, 5 μg/ml for L. lactis and B. subtilis and 100 μg /ml for E. coli; ampicillin, 50 μg/ml. For plates, media were supplemented with 1.5% agar.

Isolation of DNA

Plasmid DNA was isolated following the standard alkaline-lysis procedures (2,29). Total DNA was isolated from L. lactis as described before for B. subtilis (2). When HMW was analyzed, shearing of the DNA was avoided by using pipettes with a wide bore. Restriction enzymes, molecular cloning and gel electrophoresis

DNA- modifying enzymes were used as recommended by the suppliers (Boehringer, Mannheim, Germany; or New England Biolabs, Beverly, USA). General cloning techniques were as described (29).

Transformations

B. subtilis was grown to competence and transformed as described by Bron (2). E. coli was transformed using the CaCl₂ method (29). L. lactis was transformed by

electroporation as described by Leenhouts et al. (25), using a gene pulser apparatus (Bio-Rad Laboratories, Richmond, CA) equipped with a pulse controller unit.

Southern hybridizations

After agarose gel electrophoresis, the DNA was transferred to Gene Screen Plus filters (NEN Research Products, Dreieich, Germany) as described by Chomczymski and Qasba (10). For hybridizations, the ECL gene detection system (Amersham International, Amersham, UK) was used as recommended by the supplier.

Assay of segregational stability

Single colonies of plasmid-carrying strains, which had been purified by two successive platings, were used to inoculate GM17 medium containing chloramphenicol. At an optical density of 0.5 at 660 nm, the cultures were diluted 10⁶-fold in 100 ml of prewarmed GM17 medium without chloramphenicol. The cells were kept

exponentially growing for about 120 generations in successive batch cultures. At regular intervals appropriate dilutions of the cultures were plated on non-selective GM17 agar. At least 200 individual colonies from each sample were transferred to plates containing chloramphenicol. Resistance to the antibiotic was correlated with the presence of a plasmid of the expected size, whereas antibiotic-sensitive colonies were plasmid-free (50 colonies of each type were tested for the presence/absence of the plasmid). Determination of plasmid copy numbers

Plasmid copy numbers were estimated in two different fractions: (i), the circular monomers (covalently closed plus open circular); and (ii), all plasmid forms (circular plus linear HMW). To estimate the copy numbers per chromosome equivalent of the monomeric circular plasmid forms, plasmid-carrying L. lactis strains were grown to late exponential phase in 2-ml portions of selective GM17 medium supplemented with 10 μCi [methyl-³H]thymidine. Total DNA of these cultures was extracted, heated for 10 min at 65°C, vortexed at maximum speed for two 30-sec intervals and placed on ice for 5 min. Chromosomal and plasmid DNAs were separated in 0.5 % agarose gels, and the ethidium bromide-stained bands were excised and dissolved in a boiling water bath. Hydroluma, 15 ml, (Lumac systems Inc., Titusville, Pa.) was added and radioactivity in the samples was determined using a Mark II liquid scintillation counter (Nuclear Chicago Corp., Des Plaines, III.). From the ratios of the

radioactivities in the plasmid and chromosomal DNA fractions, the plasmid copy numbers per chromosome were estimated as described before (2). The size of the lactococcal chromosome was taken to be 2.446 mb (24). For the determination of the total plasmid copy numbers (including HMW), the densitometric procedure described by Janniere et al. (20) was used. This method involves the comparison of restriction patterns of the total DNA of cells carrying the plasmid of interest with a reference DNA mixture, consisting of a fixed amount of chromosomal DNA to which known amounts of the plasmid DNA are added.

RESULTS Construction of plasmids

For the analysis of the effects of the mode of replication on plasmid maintenance in L. lactis, pWVO1 was selected as the RCR-type model plasmid, because it is well-characterized and widely used in lactococci (21, 26). The pAMß1 derivative pIL252 (32) was chosen as the theta-type model plasmid. An undesirable property of the original pIL252 plasmid was that, due to its construction, it lacks the stability determinant orfH (32, 34). To avoid instability of the basic construct, we reinserted the 2.2-kb orfH-containing EcoRI fragment of pAMß1 into the EcoRI site of pIL252. The resulting plasmid, pSR11, carried the EcoRI fragment in the same orientation as in pAMß1 and, in contrast to pIL252, was stably maintained in L. lactis.

pSR11 and pWVO1 were subsequently provided with a set of antibiotic resistance markers. First, a cassette was constructed in the ClaI site of the multiple cloning site of the E. coli plasmid pMTL23 (9). The cassette contained the pC194-derived chloramphenicol (Cm) resistance gene (pC194 coordinates 973 to 2008 [17]) and the pE194-derived erythromycin (Em) resistance gene (pE194 coordinates 3140 to 1939 [18]). To minimize possible read-through transcription from the cassette into plasmid sequences, the T1T2 transcriptional terminators from the E. coli rrnB ribosomal RNA operon (7) were inserted on a 500-bp EcoRI fragment into the EcoRI site upstream of the Em gene. These terminators are functional in L. lactis (J. van der Vossen, personal communication). In between the two resistance genes the transcription terminator of the Em gene was present. The cassette was subsequently transferred from pMTL23 to pWVO1 and pSR11, i.e. as a BamHI-BglII fragment into the unique MboI site of pWVO1, and as a BamHI-Nrul fragment between the BamHI and SnaBI sites of pSR11. The resulting plasmids, denoted pKS100 and pAMS100, are shown in Figure 1. The copy numbers of pKS100 and pAMS100 were comparable (about 15 per chromosome equivalent; see below).

Several heterologous DNA inserts were introduced into the basic plasmids pKS100 and pAMS100. As DNA inserts, we chose the 1C and 3C E. coli BglII DNA

fragments, which were used before in plasmid stability studies in B. subtilis (4, 5), and several bacteriophage λ DNA fragments which were generated by restriction withBamHI, BglII or BclI. The various fragments were cloned into the unique BclI site in the Em^R gene of pKS100 and pAMS100, resulting in the pKS and pAMS series of plasmids (Table 1 above).

Stability of pKS100 and pAMS100 derivatives

The maintenance of the various pKS100 and pAMS100 derivatives in L. lactis was studied in the absence of selective antibiotics. The kinetics of appearance of plasmid-free cells were determined with all pKS and pAMS derivatives. Examples of representative plasmids are given in Figure 2. The results show that pKS100 was stably maintained during the entire assay period (140 generations of growth), whereas the pKS derivatives carrying λ DNA fragments of 8 or 12 kb were unstable. Similar results were obtained with several other pKS derivatives (for clarity, these were not included in Figure 2). In contrast, all pAMS derivatives were stable: no plasmid-free cells were detected within the assay period (Figure 2). In Figure 3 a summary of the results of all plasmid stability assays is shown. In this Figure the percentages of plasmid-containing cells after 120 generations of growth are given as a function of the size of the DNA inserts. The results show that most pKS100 derivatives carrying inserts smaller than about 8 kb were stably maintained. pKS103C (insert 4.2 kb) and pKS105L (insert 5 kb), however, were slightly unstable. With these plasmids about 70% (pKS103C) and 80% (pKS105L) of the cells contained the plasmid after 120 generations of growth. All derivatives with inserts exceeding about 8 kb were poorly maintained and the instability increased exponentially with the size of the DNA insert. These results suggested that insert size was an important factor in the instability of pKS derivatives. Since plasmid-free cells were not detected with any of the pAMS derivatives, these results indicated that the stability of the theta-type recombinant plasmids was superior to that of the RCR-type recombinants.

To study whether, in addition to size, specific properties of the inserts also affected the maintenance of pKS derivatives, we altered two typical inserts. The first, the 7L fragment (7.2 kb) which did not cause instability on its own (Figure 3), was increased in size to 8.4 kb by the insertion of the 1.2-kb 1C fragment. The corresponding plasmids were denoted pKS171 and pAMS171. The second fragment, 12L (12.2 kb) which caused high levels of plasmid instability in pKS12L, was shortened to 8.0 kb by deleting two BclI fragments. The resulting plasmids were denoted pKS112D and pAMS112D. The small fragments (1.6 kb and 2.6 kb) that were deleted from the 12L fragment were also cloned, separately and together, into the Bell site of pKS100, resulting in pKS102L, pKS103L, and pKS104L, respectively.

The results of stability assays (Figure 3) showed that all newly constructed pAMS derivatives remained fully stable. Plasmid pKS171, however, containing the fusion of the 1C and 7L fragments, was unstable (Figure 3). Since neither the 1C nor the 7L fragment had an effect on plasmid stability on their own, it is unlikely that a specific sequence on one of these fragments caused the instability of pKS171. Figure 3 also shows that pKS112D (8.0 kb), obtained from the unstable pKS112L (12.2 kb), was almost totally stable during the assay period. Since all small plasmids carrying fragments from pKS112L (pKS102L, -103L, and -104L) were also stable, it seems unlikely that a specific sequence on pKS112 caused its instability. These results indicate that the size of the inserts was the main factor in the instability of pKS derivatives.

Presence of HMW DNA

The cloning of heterologous DNA fragments in RCR-type plasmids often results in the formation of HMW in B. subtilis (15, 16) and E. coli (11). To study whether the plasmids used in the present studies generated HMW in L. lactis, total DNA lysates of cells carrying the various constructs were analyzed by Southern hybridization, using pKS100 or pAMS100 as probe. Hybridizing DNA migrating at the position of HMW was detected with all pKS derivatives carrying heterologous E. coli or λ DNA inserts, but not with the parental plasmid pKS100. With none of the pAMS derivatives HMW was observed. This result is illustrated for pKS100, and pKS101C in Figure 4.

Incubation of the extract containing pKS101C with a restriction enzyme for which no site was present (BglII), did not affect the migration of the DNA at the HMW position. This rules out the possibility that the hybridization signal resulted from aspecific interaction with chromosomal DNA. When the DNA was pretreated with an enzyme cutting the plasmids once, one band was detected at the position of the linearized plasmid. This indicates that the pKS-derived HMW consisted of head-to-tail plasmid multimers. We conclude that the appearance of insert-induced HMW in L. lactis depended on the mode of replication, since this DNA was only formed with RCR-type plasmids.

Relations between HMW, plasmid copy number and stability

To study possible relations between the formation of HMW, plasmid copy numbers, and stability, we determined copy numbers in exponentially growing L. lactis strains. Two assays were used; one yielding the copy numbers of all possible plasmid forms (circular plus HMW); and the other of the circular plasmid forms. In both assays the copy numbers were measured relative to chromosome equivalents. In the determination of copy numbers of the circular plasmid fraction, only monomers were considered since the fraction of dimeric and oligomeric forms was very low.

All pAMS derivatives had a copy number of 15 to 18 in both assays. This means that only circular plasmid forms were present. This result confirms the conclusion from the foregoing section that insert-induced HMW was not formed with pAMS derivatives. The results obtained with pKS derivatives are shown in Figure 5. In contrast to the copy numbers of pAMS derivatives, those of pKS derivatives were clearly affected by the presence of heterologous DNA inserts. With most pKS plasmids an inverse relation was observed between the copy numbers of the circular monomers and insert sizes. For instance, the copy number of pKS100 (no insert) was 15, whereas that of pKS116L (insert 16.8 kb) was only 2. The copy numbers of the circular plasmid forms were inversely related to the copy numbers in the HMW fraction, which increased with plasmid size. With the largest pKS derivatives, the HMW fraction amounted to about 70 plasmid copies per chromosome equivalent. The relation between HMW and insert size was not perfect, however. For instance, with the 3C insert (4.2 kb) considerably more plasmid copies were present as HMW than with the larger 5L (5.6 kb) or 7L (7.2 kb) inserts. This means that, in addition to insert size, insert-specific sequences also contributed to HMW formation.

Figure 6 shows the relation between the maintenance of pKS plasmids and the copy numbers of the circular monomers. Plasmids with copy numbers higher than about 7 were stably maintained. Plasmids with lower copy numbers were, however, unstable and the instability was inversely related to the copy numbers of the plasmid

monomers.

Since the copy numbers determined here are independent of plasmid size, they are not proportional with the mass amounts of the various plasmid forms. Since it is conceivable that the mass amounts of HMW are important for plasmid stability, we also calculated these values for the various pKS derivatives. This was done by multiplying the HMW copy numbers with plasmid size. The results (Figure 7) revealed an inverse relation between the mass amounts of HMW and the segregational stabilities of pKS derivatives. DISCUSSION

The results presented in this Part I show that in L. lactis the RCR-type plasmid pWVO1 suffers from at least one disadvantage that was previously also observed with RCR-type plasmids in B. subtilis (4, 5, 6): increased size of DNA inserts results in drastically increased segregational instability. The size-dependent instability may well explain why natural RCR-plasmids are generally relatively small (usually less than 10 kb). In contrast to pWVO1, pAMß1 derivatives with the same inserts were stably maintained, provided that they carried the stability determinant orfH (34). Since pAMß1 replicates according to a theta-mechanism (8), the difference in replication mechanism is likely to underlie the different segregational stabilities of pWVO1 and pAMß1 derivatives in L. lactis. So far, we have not studied the structural stability of recombinant pAMß1 derivatives in L. lactis in detail. We have, however, never observed deleted forms of pAMß1 derivatives upon gel electrophoresis (data not shown). This suggests that also the structural stability of these recombinant plasmids was good. High structural stability of pAMß1 derivatives was observed before in B. subtilis (19).

An additional advantage of pAMß1 is that its copy number is, apparently, not affected by DNA inserts. We speculate that also endogenous lactococcal theta-type plasmids have these advantageous properties. This would render theta-type plasmids from L. lactis attractive for cloning purposes. This is particularly valuable when selective antibiotics can not be used (for instance for food-grade purposes), or when large DNA molecules have to be cloned.

Two observations are relevant for the question why recombinant pWVO1 derivatives show size-dependent segregational instability. The first is that with these plasmids, HMW was induced in mass amounts that were related to insert size, and the second that increased amounts of HMW were paralleled by decreased copy numbers of the circular plasmid forms. Since these phenomena were not observed with pAMß1 derivatives, it is clear that insert-induced HMW formation was uniquely associated with plasmids using RCR. Similar observations have been made in other bacteria: in wild-type B. subtilis (15, 16) and E. coli (11) strains, insert-induced HMW was exclusively observed with RCR-type plasmids. Our observations indicate that HMW DNA, and/or reduced copy numbers of the circular plasmid forms, are likely to be important factors in the instability of pWVO1 derivatives. HMW formation per se was, however, not a sufficient condition for instability. This can be concluded from the observation that several derivatives produced (low) amounts of HMW, but were nevertheless stably maintained. The results rather indicated that a certain threshold level of HMW was required for the induction of instability.

At least two mechanisms can be conceived by which HMW might influence plasmid maintenance. The first is by interference with normal cell physiology. Such a mechanism was suggested before from experiments in B. subtilis (1, 27, 36). From studies in E. coli, it was concluded that under certain conditions HMW could even cause reduced cell viability (23). Since the total mass of HMW induced by the largest inserts amounted to about 40 % of the total cellular DNA, we consider reduced cellular growth rates as a realistic possibility. Growth disadvantage of HMW-containing cells would increase the rate of plasmid disappearance from the population. The second mechanism by which HMW can be conceived to reduce plasmid maintenance is through interference with the normal copy control of the circular plasmid forms. This could, for instance, be mediated through the titration of host or plasmid proteins needed for plasmid replication. Assuming that the circular pWVO1 plasmid forms are partitioned randomly to daughter cells during cell division, a reduction in plasmid copy number is expected to result in increased rates of plasmid loss. The present data do not allow us to discriminate between the two possible mechanisms. It is also possible that they both occur, and that the reduction in copy numbers of the circular plasmid forms is independent of HMW induction.

The mechanism underlying insert-induced HMW formation of pWVO1 derivatives in L. lactis is not clear. The correlation with insert size indicates that this parameter is important. Insert-induced HMW formation is generally believed to result from non-termination of leading strand displacement during RCR (11, 15, 16). Based on this idea, the effect of insert size observed in the present experiments can be explained by assuming that the probability of non-termination is directly related to insert size. In addition to size, other properties of the inserts also seemed to affect HMW formation. This was most obvious with the relatively small 3C E. coli DNA insert, which induced considerable amounts of HMW. This may, in fact, explain why pKS103C is slightly unstable. From studies conducted in E. coli, Dabert et al. (11) concluded that the nucleotide composition of the inserts affected HMW formation. Although this may also have been the case in the present studies, we have no indications for this supposition. Recent studies by Dabert et al. (P. Dabert, S.D.

Ehrlich and S. Gruss; personal communication) indicated that, in particular the presence of the chi sequence on DNA inserts stimulated HMW formation in E. coli. We consider it unlikely that similar specific nucleotide sequences were involved in the formation of the HMW described in the present Part I. This follows from the observation that insert 12L caused high levels of HMW, whereas none of its subfragments was able to do so (data not shown). Fragment 3C may form a possible exception, since considerable amounts of HMW were formed with this relatively small plasmid.

The present studies show that the theta-type plasmid pAMß1 was more stably maintained in L. lactis than the RCR-type plasmid pWVO1. We speculate that endogenous lactococcal theta-type plasmids hold good promise for the development of efficient and stable cloning vectors in lactococci.

PART II

THE LACTOCOCCAL PLASMID pWVO2 REPLICATES VIA A THETA-MECHANISM

SUMMARY

Most vectors in use today for molecular cloning in lactococci are based on rolling-circle-replicating plasmids. Plasmids replicating via this mechanism appear, however, to be unstably maintained, in particular if they carry large DNA inserts. For the development of stable vectors for cloning in lactococci we therefore searched for plas mids from these bacteria that would not replicate via the rolling-circle-mechanism. The absence of single-stranded plasmid DNA in total lysates of lactococcal cells carrying only the plasmid pWVO2 suggested that this plasmid might replicate via a theta-mechanism and be a possible candidate as cloning vector. This prompted us to analyze this plasmid further.

The entire nucleotide sequence of this 3.8 kb plasmid was determined. It revealed an ORF that showed a high degree of similarity to replication genes of several other lactococcal plasmids that have recently been described. The structure of this region of this plasmid was similar to that of the replicon of pCI305, consisting of an AT-rich region and 3½ direct repeats (22 bp each) which preceded an ORF encoding the Rep protein. The mode of replication was determined by 2-D agarose gel electrophoresis of replication intermediates. The results revealed structures that are typical for theta-replicating plasmids.

This represents the first proof of the existence of theta- replicating plasmids in lactococci.

INTRODUCTION

Most lactococcal plasmids used for cloning purposes today are based on the rolling-circle-replicating plasmids pWVO1 and pSH71 (12, 21, 26). However, it has been found that large derivatives of the theta-replicating plasmid pAMß1, which originates from Enterococcus faecalis (41), was stably maintained, whereas large derivatives of the RCR-plasmids were not (see Part I above). Conceivably, other theta-type plasmids will also be stably maintained. For the development of future lactococcal cloning vectors, also for food-grade purposes, it seems therefore desirable to use lactococcal plasmids that replicate via a theta-like mechanism.

A characteristic difference between plasmids that replicate via the RCR-mechanism and via the theta-mechanism is that the former generate single-stranded (ss) replication intermediates (16, 57, 58). Therefore, these plasmids are also referred to as ssDNA plasmids. Plasmids replicating via a theta-mechanism do not generate ssDNA intermediates. The difference in ssDNA production can be used as an indication of the plasmids mode of replication. The eventual absence of ssDNA, however, does not absolutely exclude the possibility of RC replication, since the kinetics of conversion of ss to double-stranded DNA can be very fast due to the presence of efficient initiation sites for complementary strand synthesis, so-called minus origins (6, 16). Proof for theta-replication, therefore, requires additional evidence. Two approaches can be followed for this. The first is sequencing the plasmid. All ssDNA plasmids characterized sofar appear to be highly interrelated (16). New RCR-plasmids are therefore expected to show some sequence similarity with known RCR-plasmids. The second approach is to visualize replication

intermediates, which in the case of theta-replicating plasmids show typical patterns upon 2-D agarose gel electrophoresis (38, 39). Replication of plasmid DNA by a theta-mechanism proceeds through branched intermediates. Because of their distinct topological properties, these branched forms migrate slower in agarose gels than unbranched, linear DNA molecules of equivalent mass. The resulting pattern from theta-replicating molecules differs markedly from that of RCR-molecules and can be used as a criterion to distinguish between the two possible modes of replication.

For our purpose to search for stable food-grade cloning vectors for lactic acid bacteria, we decided to characterize the plasmids present in strain L. lactis subsp. cremoris Wg2. In addition to the RCR-plasmid pWVO1, this strain contains at least four other plasmids, denoted pWVO2 through pWVO5 (52). Here we describe the characterization of the cryptic plasmid pWVO2 (3.8 kb). Other plasmids from the lactococcal strain which all appeared to be related to pWVO2, are described in Part IV below.

The results show that no homology is present between the sequence of pWVO2 and known RCR-plasmids. Extensive homology, however was found with a number of other, recently described lactococcal plasmids (44, 47, 49). The results obtained by 2-D agarose gel electrophoresis of replication intermediates give a strong indication that pWVO2 replicates via a theta-mechanism. Based on these results, we anticipated that all plasmids belonging to this family replicate in a similar manner as pWVO2. RESULTS

Absence of pWVO2 single-stranded plasmid molecules

Plasmids, that replicate via the RCR-mechanism, produce single-stranded replication intermediates (16, 57, 58), which can be visualized by Southern hybridization of total cell DNA. When run on an agarose gel under certain conditions, the ssDNA will migrate just below the covalently closed circular DNA. The presence of these ssDNA molecules gives an indication that the plasmid in question replicates via the RCR-mechanism.

Total cellular DNA of a strain carrying pWVO2 as the only plasmid was isolated and run on agarose gels. The gel was blotted onto a nitrocellulose membrane and plasmid DNA was detected with a pWVO2 derived probe. No ssDNA of pWVO2 could be detected (Figure 8). In parallel experiments with an RCR-plasmid ssDNA could be detected in spite of the presence of a minus origin. This result suggested that pWVO2 does not replicate via the RCR-mechanism.

Sequence of plasmid pWVO2

Purified pWVO2 was linearized with ClaI. This fragment was cloned into the unique ClaI site of the pBluescriptll vector pBSKII (Stratagene, La Jolla, Calif.). The resulting plasmid was denoted pJR02. A set of deletion derivatives of this plasmid was generated by using the ExoIII/Mung Bean method (see materials and methods). Deletions in one direction were created using the SacI site as the 3' protective end and the EcoRI site as the 5' starting point for Exonucleaselll. Deletions in the other direction were created with the KpnI site as the 3' protective end and the SalI site as the 5' starting point. The resulting deletion derivatives were sequenced, using the M13 universal and reverse primers (56).

The complete nucleotide sequence of plasmid pWVO2 is given below. Possible promoter and ribosome binding sites 5' upstream of the ORF, encoding the putative rep protein are indicated by dashed lines under the sequence. The deduced amino acid sequence of the rep protein is indicated by the single letter code. The direct repeats in front of the rep protein are indicated by arrows. The location of other significant regions are indicated in the specification following the sequence of pWVO2.

The obtained sequence revealed an ORF that showed extensive similarity (75% identity on the amino acid level) to the ORF encoding the replication gene of pCI305 and a number of other lactococcal plasmids (44, 47, 49). Analogous to pCI305 we denoted this ORF repB. The organization of pWVO2 was similar to that of pCI305 (44; see also Figure 9 giving the structure of pLR300, a derivative of pWVO2 containing the erythromycin resistance gene of pE194). Like in pCI305, the

replication gene is preceded by a 22 bp 3½ direct repeat (DR) upstream of which an AT-rich sequence (more than 80% AT over a length of 150 bp) is present.

Outside this region of similarity with other plasmids a 1.8 kb long stretch was found that showed three different repetitive sequences that varied in length from 50 to 200 bp. The organization of this region is shown in Figure 10. Preceding and just overlapping the first repeated sequence of 50 bp a 13 bp sequence is found that also appears in the 100 bp repeats. The overlapping 5 bp sequence, 5'-GATGA-3', is present in both the 50 bp repeat as in the 100 bp repeat, but is absent from the 200 bp repeat. No homology to any known sequence could be found for this region.

Deletion of a 1.3 kb EcoRV fragment from this region lead to a significant drop in copy number of the plasmid. Possibly as a result of this, the plasmid became segregational very unstable (see Part III below).

Analysis of replication intermediates of pWVO2 by 2-D agarose gel electrophoresis

To analyze its mode of replication we introduced into pWVO2 a selectable marker. This was done by inserting a 1.8 kb fragment containing the erythromycin resistance gene of pE194 (18) into the unique HpaI site of pWVO2 (position 1 of the sequence), located behind the 1.8 kb DR region and in front of the AT-rich sequence, resulting in plasmid pLR300 (Figure 9).

Plasmids that replicate via a theta-mechanism generate intermediates which have a typical shape, often called replication bubbles. The method of 2-D agarose gel electrophoresis separates linearized branched DNA fragments in proportion to their mass but retards their migration in the second dimension in a manner that depends on the number, length and topology of their branches. Southern hybridization subsequently permits the detection of these branched forms. To see if pWVO2 might generate such intermediates, we analyzed replicating pLR300 molecules by 2-D agarose gel electrophoresis. Upon BglII cleavage of total cellular DNA prepared from a L. lactis strain harbouring pLR300 which linearizes the plasmid, and subsequent 2-D gel electrophoresis, all possible replication intermediates, typical for theta-replication could be observed (Figure 11). Especially the presence of an arc, starting at the position of linear monomers and stretching towards the position of circular dimers, provides evidence for theta-replication. This arc represents of bubble-shaped replication intermediates which are typical for theta-replicating plasmids. 2-D agarose gel electrophoresis of total cellular DNA restricted with BamHI resulted in a similar pattern (results not shown). DISCUSSION

In this Part II we describe attempts to search for plasmids from lactococci which have the potential to be used for the development of stable food-grade, cloning vectors for lactic acid bacteria. In particular, we focused on the 3.8 kb plasmid pWVO2, which has this potential.

The complete nucleotide sequence of pWVO2 was determined. It contained an ORF that was homologous to the replication genes of a number of other recently examined lactococcal plasmids (44, 47, 49). Evidence that the described ORF in pWVO2 has similar replication functions was obtained from experiments in which attempts were made to clone pWVO2 in E. coli plasmids, like pUC19. When for this purpose restriction sites, located within this ORF were used the resulting plasmids were no longer able to support replication in L. lactis (results not shown). Like in pCI305 (44), a region with 3 ½ DRs could be identified upstream of the ORF encoding the Rep protein. Similar structures, with strongly related sequences were observed for other lactococcal plasmids (47, 49). One possible function of these repeats is that they function as the target for the Rep protein. Based on similarities with certain plasmids from Gram-negative bacteria (42, 53), one could, however, also conceive that these DRs are involved in (in)compatibility phenomena between the different plasmids of this family, different members of which are frequently present within one strain.

The results described in this Part II strongly suggest that pWVO2 replicates via a theta-mechanism. The observations supporting our conclusion are: (i) no pWVO2 ss-DNA intermediates could be detected; (ii) The absence of sequence similarities with known ssDNA (RCR) plasmids; (iii) the pattern of replication intermediates upon 2-D agarose gel electrophoresis were typical for theta-type plasmids. This is the first described proof of the existence of theta-type plasmids in lactic acid bacteria. From the patterns of intermediates observed sofar, we cannot deduce the location of the start site of replication (origin), nor the direction of replication. Both bi-directional and unidirectional replication, starting from one origin, should reveal "switchpoints" within the pattern of replication intermediates. These "switchpoints" will appear when the replication fork passes through the restriction site that was used for the analysis of linear replication intermediates. The pattern is predicted to switch from bubbles to simple Y structures in the case of unidirectional replication; and from bubbles to complex Y structures in the case of bidirectional replication. The observed patterns, however, showed no interruption for any of the structures (bubbles, simple Ys, termination structures and complex Ys). Two possible explanations can be envisaged for this observation. The first is that replication is either uni- or bidirectional, but starts at various sites on the plasmid molecule. Sofar, this type of replication has only been described for eukaryotic systems (45, 48, 50, 59). The second explanation is that replication starts from a fixed origin, but proceeds in an asymmetrical, bidirectional manner. In that case each single replicating molecule would show a switchpoint, but these would be different for each molecule. The resulting 2-D pattern of the collected replication intermediates contains all the possible molecules, giving rise to a smooth pattern. To obtain clear answers concerning this question, additional 2-D analysis in which other restriction enzymes are used for the linearization of replication intermediates will have to be performed.

Based on our results obtained sofar, we favour the following working model for the replication of pWVO2. After translation, the Rep-protein binds to the 3½ DRs. After binding to this site, the Rep protein mediates, in an unknown way, directly or indirectly, the opening of the DNA helix at several locations on the plasmid from which replication can start either bi- or unidirectionally in a theta-like manner. Once replication has started, no additional starting points are used on the same molecule. The last idea is based on the fact that structures predicted to result from multiple initiations (like bubble/Y shapes or double bubble shapes) were not observed.

In pWVO2, a 1.3 kb fragment within the repeated region appeared to be dispensable for replication. Therefore this fragment is not part of the minimal replicon.

Nevertheless this fragment is of interest,both from a fundamental and applied point of view. Our preliminary results indicated that this fragment may be involved in plasmid copy number control and plasmid maintenance. The abundance of the 5'-GATGA-3' region in the 1.8 kb direct repeated region could be significant in this respect. It is conceivable that the 5'-GATGA-3' sequence functions as a recognition site from which replication can start. We can also conceive of the possibility that this fragment plays a role in the co-existence of the various related plasmids in lactococcal strains, such as L. lactis subsp. cremoris Wg2. We will pursue the role of the 1.8 kb fragment in future studies.

Based on our observations that derivatives of the Enterococcus faecalis, theta-type plasmid pAMß1 are stably maintained in L. lactis (see Part I above), we anticipated that the lactococcal plasmid pWVO2, described here, will also be stable. As will be described in Part III below we have indeed found that the introduction of a 12 kb DNA insert in pWVO2 results in a stable replication comparable to a pAMß1 derivative having the same insertion, while insertion of this DNA fragment in a pWVO1 derivative led to a highly unstable replication. Thus, pWVO2 and related plasmids are good candidates for the construction of a set of stable, food-grade, cloning vectors for lactic acid bacteria.

EXPERIMENTAL PROCEDURES

Bacterial strains, biochemicals and growth conditions

E. coli strain JM101 (supE, thi⌂()[FtraD36, proAB, lacI^qZ⌂M15]) (60) was used for cloning and sequencing experiments. Cells were transformed as described by Mandel and Higa (51) and plated on LB agar or grown in LB medium (54). When required, the media were supplemented with 100 μg/ml erythromycin. For DNA isolations, strain L. lactis subsp. lactis MG1363 (43) was used. Cells were transformed by electroporation (46) and plated on M17 agar, containing 0.5M sucrose. Transformants were grown on M17 medium (35). When required, media were supplemented with 5 μg/ml erythromycin. Plasmid stability assays were performed as described in Part I above. Biochemicals used were obtained from Merck (Darmstadt, Ger.) or BDH (Poole, England). All enzymes were from Boehringer (Mannheim, Ger.) or Promega Biotec, Madison, USA.

DNA manipulations

Cloning procedures and the isolation of DNA were essentially as described by

Sambrook et al. (54). Double-stranded DNA sequencing was performed by the dideoxynucleotide chain termination method of Sanger et al. (55).

Isolation of pLR300 replication intermediates

Replication intermediates were isolated according to a method for the isolation of plasmid DNA from cleared lysates of B. subtilis (40) with minor modifications. We chose this method because it does not use an alkaline treatment by which only covalently closed circular DNA is recovered and not the desired replication

intermediates. The following modifications were used. Lactococcal cells were grown at 30°C, without shaking. Mutanolysin was added to the lysis solution to enhance lysis. Replication intermediates were recovered from a CsCl/ethidium bromide gradient by isolating the fraction between the band of covalently closed circular DNA and the chromosomal DNA (8). Separation and detection of replication intermediates

Total cellular DNA was cut with restriction enzymes that linearized the plasmid DNA. Two-dimensional agarose gels were run as described (39). Transfer of DNA from gels onto nitrocellulose filters was performed as described by Sambrook et al. (54). Preparation of probes, hybridization and detection of DNA were conducted by using the ECL gene detection system (Amersham, Buckinghamshire, UK) as recommended by the manufacturer. PART III SEGREGATIONAL STABILITY OF pWVO2 DERIVATIVES For the development of efficient food-grade vectors their stable maintenance is a prerequisite. Recently we have shown (see Part I above) that in L. lactis plasmids based on the theta-replicating plasmid pAMß1 (41) were stably maintained, whereas large derivatives of the RCR-plasmid pWVO1 were not. However, the fact that pAMß1 was isolated from Enterococcus faecalis, makes it unattractive for the development of food-grade vectors. For this purpose the use of endogenous lactococcal plasmids is preferred. A potential candidate in this respect would be pWVO2 since we have shown that, like pAMß1, this plasmid replicates via the theta-mode (see Part II above). To investigate whether, like pAMß1, plasmids based on pWVO2 were segregationally superior to plasmids based on the RCR-type plasmid pWVO1, the segregational stability of a number of pWVO2 derivatives was determined and compared to that of similar derivatives of pAMß1 and pWVO1.

RESULTS AND DISCUSSION The plasmids used for the stability assays are listed in Table 2.

Plasmid pLR300 was constructed as mentioned above in Part II (see Figure 9).

Plasmid pLR100 was constructed by replacing a 1.3 kb EcoRV fragment of pWVO2 (Figure 12) by a cassette, containing the erythromycin resistance gene of pE194 and the chloramphenicol resistance gene of pC194 (see Part I above). The deleted EcoRV fragment is part of the region containing large direct repeats, but does not contain elements of the minimal replicon. pLR312L (Figure 12) is based on pLR300 and contains the same 12 kb BamHI bacteriophage lambda DNA insert as pKS112L (pWVO1 derivative) and pAMS112L (pAMß1 derivative). Table 2. Plasmids used in this study plasmids properties source or

reference pWVO2 3.8 kb; cryptic plasmid from L. lactis subsp. cremoris Wg2 (52) pLR300 Em^R, 5.6 kb; pWVO2 derivative This work pLR100 Em^R, Cm^R, 4.7 kb; deletion derivative of pWVO2 This work pLR200 Em^R, Cm^R, 4.7 kb; insert in opposite

orientation from pLR100 This work pLR312L Em^R, 16 kb; pLR300 derivative This work pKS112L Em^R, Cm^R, 17.2 kb; pWVO1 derivative See Part I pAMS112L Em^R, Cm^R, 20.6 kb; pAMß1 derivative See Part I

In a culture of L. lactis strain MG1363 containing pLR300 no plasmid-free segregants could be detected after 120 generations of growth under non-selective conditions. In contrast a culture of the same L. lactis strain containing pLR100 was completely plasmid-free within 100 generations of growth (Figure 13). To study the possibility that in the latter plasmid read-through transcription from the inserted sequences into essential genes of pWVO2 interfered with stability we also measured the

segregational stability of a pWVO2 derivative in which the cassette containing the antibiotic resistance genes was reversed in comparison to pLR100 (pLR200). This plasmid was as poorly maintained as pLR100. This indicates that the removed sequence plays a role in the replication and maintenance of pWVO2. Another observation that supported this idea is that the amounts of DNA that could be extracted from cells containing these constructs was far less than those of pLR300. This suggests that the plasmid copy number was reduced in the absence of the large directly repeated region.

The role of DNA inserts on the segregational stability of pWVO2 has sofar been tested with one construct, pLR312L, containing a 12 kb BamHI fragment of bacteriophage lambda DNA. The stability of this plasmid was compared to that of pKS112L and pAMS112L. After 120 generations of growth under non-selective conditions, 37% of the cells still carried the plasmid. In a similar experiment pAMS112L was retained 100%, whereas pKS112L at this stage was entirely lost (Figure 13). These results show that the maintenance of pWVO2 is indeed superior to the RCR-plasmid pWVO1. pLR312L was only slightly less stable than the pAMß1 derivatives carrying the same 12 kb DNA insert. Thus theta-type plasmids are more stably maintained in lactococci and are, therefore, more suitable as cloning vector. Why pLR312L is slightly less stable than pLR300 or pAMS112L is not yet clear. One possible explanation is that the location of the insert interferes with replication of the plasmid. Another possibility is that read-through transcription from the 12 kb insert interferes with replication of the plasmid. Other sites further away from the replication functions of pLR300 might be preferable. These possibilities are currently under investigation. From the experiments with pLR100 and pLR200 it is clear, however, that for cloning purposes most, if not all, of the intact plasmid is required. We anticipate that with the appropriate pWVO2 constructs, this plasmid can be developed into an efficient and stable food-grade vector.

PART IV

THE MAJORITY OF LACTOCOCCAL PLASMIDS BELONG TO ONE FAMILY

SUMMARY

Most lactococcal strains carry a number of plasmids, varying in size from

approximately 2 kb to over 100 kb. Sofar, the replication mechanism of only a limited number of these plasmids has been analyzed in detail. All of these plasmids appeared to belong to the class of so-called rolling-circle-replicating plasmids which generate single-stranded replication intermediates. Proof that Lactococci may also contain plasmids which replicate via a theta-mechanism was lacking sofar. With the aim to identify such plasmids, we analyzed in the present work a number of plasmids, two of which are derived from the same strain. The replication regions of these plasmids were cloned and their nucleotide sequences determined. All appeared to contain an ORF, encoding a putative replication protein (Rep). The various Rep proteins were strongly related to each other, but differed markedly from known Rep proteins of RCR-plasmids. The family of Rep proteins described here was also strongly related to the Rep protein of a number of previously described lactococcal plasmids. Southern hybridizations were used to confirm that the majority of lactococcal plasmids belong to this class of related plasmids. One plasmid of this family is pWVO2, which was shown to replicate via a theta-mechanism (see Part II). This strongly suggests that all plasmids described here use the theta-mode of replication.

INTRODUCTION

Lactococcal strains generally carry a number of different plasmids, some of which specify traits that are of major interest for dairy industry. A disadvantage of this plasmid location is that when the plasmid is lost from the population, the desired trait is lost with it. This is a problem of both fundamental and applied interest. Plasmid loss is a frequently observed problem in Gram-positive bacteria such as Bacillus subtilis (3, 16) and Staphylococcus aureus (63). Plasmid loss has also been observed in Lactococci (52). It has been found that, in particular, large RCR-type plasmids are poorly maintained in L. lactis (see Part I above). On the contrary, the theta-type plasmid pAMß1, which originated from Enterococcus faecalis (6), appeared to be stably maintained in L. lactis.

Sofar, molecular cloning in lactococci has mainly been carried out with the well-characterized small RCR-type plasmids pWVO1 (21, 22, 26) and pSH71 (12). Although several lactococcal plasmids have recently been described that are unrelated to the known RCR lactococcal plasmids (44, 47, 49), proof that theta-type plasmids were present in lactococci has not been provided. Based on the observation that the streptococcal plasmid pAMß1 is structurally stable in B. subtilis (19) and is stably maintained in lactococci (see Part I above), we anticipated that endogenous theta-plasmids from lactococci might be extremely valuable for the development of structurally and segregationally stable (food-grade) cloning vectors in lactic acid bacteria. The aim of the present studies was to characterize and compare plasmid contents from a considerable number of lactococci. A major long-term goal of this work is to identify plasmids, presumably of the theta-replication-type, which are valuable for the development of stable, food-grade, cloning vectors. Such studies are also relevant for the understanding of mechanisms underlying plasmid instability in lactococci. One of the possible causes of plasmid loss may be found in the complex plasmid complement found in most lactococcal strains. It is conceivable that in the case of combinations of plasmids one type of plasmid out-competes another. This, for instance, would be the case if these plasmids were partially or completely incompatible. One possibility to analyze whether plasmid loss is caused by intrinsic properties of the plasmid itself or by the interaction between different plasmids found within one strain is to analyze a number of plasmids, all derived from one and the same strain. For this approach we chose Lactococcus lactis subsp. cremoris Wg2 which carries a total of five different plasmids (52). The smallest of these, pWVO1, has been studied extensively and a large set of multipurpose vectors based on it has been developed (21, 22). Of the other plasmids, denoted pWVO2 (3.8 kb), pWVO3 (7 kb), pWVO4 (19 kb) and pWVO5 (27 kb), the latter is of particular interest since it carries the gene specifying the proteinase required for casein breakdown in cheese production (52, 62). Loss of proteinase activity is frequently observed during cheese production and appears to be connected to the loss of pWVO5 (52).

In this Part IV we describe the characterization of the replication region of pWVO4 and pWVO5. In addition, we sequenced the replication region of pIL7. This plasmid is a lactococcal plasmid containing a restriction/modification system that renders carrier strains resistant to a number of bacteriophages (61). All the analyzed regions appeared to be highly related to the replication regions of pWVO2 and other lactococcal plasmids that have recently been described (44, 47, 49). Using Southern hybridization, we could demonstrate that members of this plasmid family were also present in several other lactococcal strains. We conclude that the majority of lactococcal plasmids belong to just one class. Since we have found that pWVO2 replicates according to a theta-mechanism (see Part II above) this whole family of plasmids is likely to use this replication mechanism. RESULTS

Cloning of the replication regions of plasmids pWVO4, pWVO5 and pIL7

Different approaches were used for the cloning of the replication regions of the various plasmids that were compared in this study. Plasmids pWVO4 and pWVO5 were isolated from an agarose gel on which the entire plasmid complement of L. lactis subsp. cremoris Wg2 was separated. Plasmid pIL7 was isolated from a

lactococcal strain, carrying only this plasmid. pWVO4/pJRO4

Plasmid pWVO4 was digested with BglII. This resulted in at least three fragments which were cloned into the BamHI site of pMTL23E. pMTL23E was derived from pMTL23 (9) into which the Em^R gene (1.8 kb) of pE194 (18) was cloned for selective purposes in L. lactis. Recombinants were initially selected in E. coli JM101. Only recombinant plasmids carrying a 4.5 kb pWVO4 fragment were able to support replication in L. lactis. The desired plasmid, denoted pJR04 was used for further analysis (Figure 14). pWVO5/pJRO5

The intact pWVO5 was genetically marked at its unique BglII site with the Em^R gene of pE194. The resulting plasmid was denoted pWVO500. Deletion analysis of this plasmid revealed that its replication region was contained on a 7 kb EcoRI fragment. This fragment was inserted into the EcoRI site of the pBluescriptII vector pBSKII (56), resulting in pJR05. pIL7/pJR7

A partial Sau3A digest of pIL7 was cloned into the BamHI site of pUC19E, a pUC derivative carrying the Em^R marker of pE194 in the SmaI site. The ligation mixture was used to transform E. coli and resulting transformants were pooled. Plasmid DNA isolated from the pooled transformants was used to transform L. lactis. Only plasmids carrying the replication region of pIL7 should be able to support replication in Lactococcus. Transformants were tested for plasmid contents. Plasmids with inserts of varying length could be isolated. The smallest plasmid carried a fragment of 2.4 kb. This plasmid was used for further analysis and denoted pJR7 (Figure 14).

DNA sequence of the pJR04, pJR05 and pJR7 replication regions

Based on the similarities described above between pWVO2 and several other previously described lactococcal plasmids (44, 47, 49) we reasoned that most lactococcal plasmids, other than those being of the RCR-type, might belong to the same family. To study this possibility we designed a set of ten primers (Table 3) that might enable us to sequence the replication regions of other lactococcal plasmids in both directions. The set of primers was used to sequence pJR04 (ex pWVO4), pJR05 (ex pWVO5) and pJR7 (ex pIL7). Most of the primers resulted in a clear sequence (see Table 3). The resulting nucleotide sequences of all three replication regions were similar to that of pWVO2, though not identical. Table 3. List of oligonucleotide primers

Oligonucleotides used for sequencing of replication genes from pWVO4, pWVO5 and pIL7. Nucleotides between brackets indicate different possibilities at that location.

+ or - indicates if the oligonucleotides produced a readable sequence in the dideoxy chain termination reaction.

Primer Sequence pWVO4 pIL7

pWVO5

P1 AA(TA) CAA AAG CAG GTG C - - +

P2 ATG CAA (GA)A(GA) CAA GC(CT) TTT T - + +

P3 GAA (TC)TA (TC)AA CCA ATA CG - - +

P4 (CT)AT TGT CTT TCA TAT + + +

P5 GG(GT) GTC AAA GAC CAC TTG TC + + +

P6 GAC AAG TGG TCT TTG AC (AC) CC + + +

P7 (GA)AT ATG AAA GAC AAT (GA) + + +

P8 CGT ATT GGT T(AG)T A(AG)T TC - + +

P9 AAA A(GA)G CTT G(CT)T (CT)TT GCA T + + +

P10 CAA GGT (TCA) (TA) (GA) CAC CTG CTT TT + + + In Table 4 the nucleotide sequences of the replication regions of three plasmids, all derived from the same strain are presented. These plasmids show an overall identity of 60% at the nucleotide level and of 65 % at the amino acid level. The amino acid sequence of the Rep protein that is shown was deduced from the nucleotide sequence of pWVO2. The stop codon ATG is underlined. Amino acids that are identical for all three Rep proteins are shaded. Parts of the sequence that could not be determined with the oligonucleotides, indicated in Table 3, were determined with additional oligonucleotides.

The alignment of amino acid sequences deduced from the nucleotide sequences of the replication genes of several lactococcal plasmids, all derived from different strains is given in Table 5 below. Lines and numbers above the sequences indicate the location and direction of the oligonucleotides that were designed for the sequencing of the replication regions of plasmids pIL7, pWVO4 and pWVO5. Asterisks below the sequences indicate amino acids that are identical for all plasmids. The replication region of pSK11L varies most from the other sequences.

Comparison of the deduced amino acid sequences of the replication proteins of plasmids derived from different strains revealed a high level of similarity. Especially the N-terminal and central regions show highly conserved boxes. Southern hybridizations of a number of lactococcal strains

Using the replication region of pCI305 as a hybridization probe it has already been shown that this plasmid is widely spread among lactococci (44). We have extended these studies with another eight lactococcal strains. A 1 kb PCR fragment of the replication gene of pWVO2, obtained with primers 1 and b (Table 3 above) was used as a probe. With all strains tested, at least two plasmids were found to hybridize to this fragment (Figure 15). DISCUSSION

The homology observed between the ORFs and deduced amino acid sequences of pWVO2, pWVO4, pWVO5 and pIL7 on the one hand, and those of the replication genes of pCI305 and related plasmids on the other hand strongly suggests that all these plasmids encode a strongly related family of replication proteins. These results imply that at least three out of the five different plasmids of strain L. lactis subsp. cremoris Wg2 belong to the same family. Southern hybridization analysis of eight other Lactococcus strains revealed that this conserved replication region is present in the majority of lactococcal plasmids and that therefore all of these plasmids belong to the same family. Only a limited number of plasmids did not hybridize to the pWVO2 probe. A number of small plasmids of the latter group, however, hybridized with a probe derived from pWVO1 (results not shown). This replicon which is of the RCR-type, is, apparently also widespread among lactococci. The homology of the replication region between plasmids of this latter group (12, 26, 39) appears to be even higher (over 95% identity at the DNA level) than that of the plasmids discussed in the present Part IV.

Sofar, we have not succeeded in the cloning of the replication region of pWVO3. The reason for this is, as yet, obscure. The restriction pattern of this plasmid, in contrast to that of pWVO4 and pWVO5, revealed a remarkable similarity to that of pWVO2 (Figure 16), which suggests that pWVO2 and pWVO3 are derived from each other. The observation that the replication regions of the plasmids, analyzed and described in the present studies, are all related to the replication region of pWVO2 is of particular interest. We have found that pWVO2 replicates according to the theta-mode (see Part II above). This implies that the other members of the plasmid family described here will also use the theta-mode of replication. Since the existing evidence suggests that theta-type plasmids are structurally stable in B. subtilis (19) and that at least pAMß1 is segregationally stable (see Part I above), we consider it likely that the plasmids described here can be successfully used for the construction of stable food-grade host-vector systems for lactic acid bacteria. Further work on that aspect is in progress (see Part V below).

Another question of interest which emerges from these studies is the coexistence of several closely related plasmids in one and the same strain. In such situations loss of certain plasmid types might have been anticipated. This is apparently not the case with the varied plasmid populations present in L. lactis strain cremoris Wg2 and other Lactococcus strains. EXPERIMENTAL PROCEDURES

Bacterial strains and growth conditions

E. coli strain JM101 (supE, thi⌂()[F'traD36, proAB, lacI^qZ⌂M15]) (60) was used for cloning and sequencing experiments. Cells were transformed as described by Mandel and Higa (51) and plated on LB agar or grown on LB medium (54), supplied with 100 μg/ml erythromycin or 75 μg/ml ampicillin when required. Tests for replication abilities of the constructed plasmids in lactococci were carried out in L. lactis subsp. lactis MG1363 (43). Cells were transformed by electroporation (46) and plated on M17 agar, containing 0.5M sucrose and grown on M17 medium (35), supplied with 5 μg/ml erythromycin.

DNA manipulations

Plasmids were isolated from gel using the prep-a-gene kit of Bio Rad as

recommended by the manufacturer (Bio Rad Laboratories, Richmond, Cal.). Cloning and isolation of DNA were essentially as described by Sambrook et al. (54). Double-stranded DNA sequencing was performed by the dideoxynucleotide chain termination method of Sanger et al. (55).

Southern hybridization

Southern transfer was carried out as described by Sambrook et al. (54) using Gene Screen Plus membranes as carrier (NEN-research laboratories, Boston, Ma.). DNA fragments to be used as probes were prepared by PCR using primers 1 and 10 and pJR02 as a template. Probe labelling, hybridization and detection were conducted using the ECL gene detection system (Amersham, Buckinghamshire, UK) as recommended by the manufacturer.

PART V

DEVELOPMENT OF pWVO2 AS A STABLE FOOD-GRADE VECTOR SUMMARY

A food-grade host/vector cloning system was developed for Lactococcus lactis. The vector is based on the stable theta-replicating lactococcal plasmid pWVO2. As a food-grade marker on the vector, the sucrose operon from Pediococcus pentosaceus was taken. The presence of the plasmid in host cells was detectable by a simple colour indicator assay. The lactococcin A operon from L. lactis was used as a model to demonstrate that genes of interest can be cloned using this host/vector system. The cloning vector as well as the recombinant plasmid containing the lactococcin A operon were fully stably maintained when the host cells were grown in glucose-containing media. Low levels of plasmid loss were observed in sucrose-containing liquid media. One of the constructed vectors was developed into a food-grade expression vector for L. lactis.

INTRODUCTION

In Part II of this specification we showed that the lactococcal plasmid pWVO2 replicates via a theta-mechanism. This type of replicon is widespread among lactococci and several plasmids of this family can co-exist within the same host (see 44, and Part IV of this specification). Our previous results indicated that theta-replicating plasmids are structurally more stable than rolling-circle (RC) plasmids and that their segregational stability is less susceptible to the cloning of foreign DNA (see Part I of this specification; during the priority year published: 65). We therefore anticipated that the segregationally stable plasmid pWVO2 can successfully be used for the development of stable food-grade vectors.

The aim of the present work was to exploit this possibility. Food-grade vectors were constructed which were stably maintained under certain conditions. In addition, we tested whether the constructed vectors were suitable for the cloning and expression of genes of interest. Finally, one of the plasmids was developed into an expression vector by the insertion of a DNA fragment containing suitable expression signals, a multiple cloning site and a transcriptional terminator.

As a food-grade marker we have chosen the Pediococcus pentosaceus sucrose operon containing the genes ser A (encoding enzyme II, a specific sucrose permease), scrB (encoding sucrose-6-P-hydrolase) and scrR (a regulatory gene). These genes are located on the Pediococcus pentosaceus plasmid pSRQ1 (64), specifying functions in the sucrose metabolism. For the cloning and expression of genes of interest the lactococcin A operon (71) was chosen as a model system.

RESULTS AND CONCLUSIONS

Cloning of the sucrose operon

The sucrose operon has previously been cloned from the Pediococcus pentosaceus plasmid pSRQ1 (64) as a 15 kb BamHI fragment on pSR11 (see EP-A1-0487159;

UNILEVER, published 27.05.92), and we subcloned it further on the integration vector pINT123E (Figure 17).

This plasmid was constructed as follows.

In the BamHI site of the multiple cloning site of pUK21 (75) the erythromycin resistance gene of pE194 (18) was introduced giving plasmid "A". Subsequently the

XhoI-PvuII internal fragment of the lactococcal pepXP gene (70) was ligated with the BglII- and StuI-treated plasmid "A" giving plasmid "B". From the latter the SpeI fragment was isolated, containing the Em^R gene and the pepXP internal fragment.

This fragment was provided with blunt ends with Klenow enzyme and ligated to the 0.6kb TaqI Ori⁺ fragment of pWVO1 (26), also treated with Klenow enzyme. The ligation mixture was transformed to the special Lactococcus lactis helper strain as described (67). The resulting plasmid was designated pINT23. The sucrose genes ser A, scrB and part of scrR (scrR') were isolated from the

Pediococcus pentosaceus PPE1.0 plasmid pSRQ1 (64) as a BglII-SalI fragment and blunt ends were generated using Klenow enzyme. This fragment was ligated into the ClaI site of pINT23, which was provided with blunt ends using Klenow enzyme. The resulting plasmid was designated pINT123E (Figure 17). This plasmid is able to replicate only in special helper strains, in which the Rep gene is integrated in the chromosome, while this plasmid integrates in normal Rep- L. lactis strains (67).

The cloned BglII-SalI fragment contained all the genes required for sucrose metabolism. Attempts to clone the sucrose operon in pWVO2 by the direct selection for sucrose metabolism in L. lactis were unsuccessful. Clones were obtained, however, when the sucrose genes were used together with the erythromycin resistance (Em^R) gene of plasmid pE194, which is also present on pINT123E. For this aim, pINT123E was digested with EcoRV and BglII and the fragment containing the scrA, scrB genes, a truncated scrR gene, and the Em^R gene, which was made blunt by T4 DNA polymerase treatment, was inserted into the unique HpaI site of pWVO2.

Transformants were initially selected on erythromycin-containing plates and then toothpicked onto SM17 plates containing bromocresol purple (BCP) as a colour indicator. Transformants containing the sucrose genes, which can grow on sucrose as carbon source, produce acid when they are streaked on SM17-BCP plates, and this renders colonies yellow. Transformants were analyzed for their plasmid contents. Two plasmids containing the desired DNA fragments were found (pLR330 and pLR331; Figure 18) that differed only in the orientation of the sucrose genes. These plasmids both sustained growth of L. lactis MG1363 on sucrose and carried, except for the erythromycin marker, DNA from GRAS-classified bacteria only and can therefore be marked as nearly food-grade.

Stability of pLR330 and pLR331

When pLR330- and pLR331-c-ntaining cells were grown overnight on glucose- plus erythromycin-containing M17 liquid media and then transferred to glucose-containing M17 media without erythromycin, the plasmids were stably maintained (96% of the cells were still resistant to erythromycin after about 80 generations of non-selected growth). When, however, the plasmid-containing strains were grown overnight in sucrose-containing M17 liquid media and then transferred to glucose-containing M17 medium, about half of the cells had lost the plasmid already in the starter culture. No further decrease in the percentage of plasmid-containing cells was observed after prolonged growth on glucose-containing M17 medium. In contrast to cells pre-grown in liquid sucrose-containing media, cells pre-grown in colonies on sucrose-containing M17 agar plates appeared to maintain the plasmids fully stably: all cells transferred from a colony on the plate and subsequently grown in glucose liquid M17 medium for over 80 generations were resistant to erythromycin and contained the plasmid.

A possible explanation for the low level of plasmid instability in sucrose media is increased transcription through induction of the scrA and scrB genes. Since no transcriptional terminators flank these genes in pLR330 and pLR331, it can be envisaged that transcriptional read-through into the pWVO2 replication region interferes with the replication of the plasmid. This may result in reduced copy numbers and increased rates of plasmid loss from the cells. Plasmid-free cells will remain in the population, since M17 medium containing sucrose also contains undefined additional C-sources which sustain limited growth of L. lactis. An alternative explanation is that the increased transcription of the plasmid-located sucrose genes results in altered levels of plasmid supercoiling, which may affect plasmid

maintenance. Effects of transcription on plasmid supercoiling are well-known (68, 69). When grown on glucose, the sucrose genes are likely to be repressed (due to the presence of regulatory genes on the plasmid or the chromosome of L. lactis), so that less interference with the replication of the plasmid occurs.

Construction of a food-grade vector containing the sucrose and lactococcin A operons flanked by transcriptional terminators

If the initial low levels of plasmid instability observed when the cells are grown on sucrose-M17 liquid media would result from transcriptional read-through, it might be possible to increase the stability by introducing transcription terminators at appropriate positions in the plasmid. In order to test this possibility, and at the same time to construct plasmids with which the feasibility of cloning genes of interest could be tested, the cloning strategy shown in Figure 19 was used. This strategy resulted in the vector pLCN312 in which the sucrose genes were flanked on both sides by transcriptional terminators and the lactococcin A operon was introduced as a model of genes of interest.

A. Introduction of transcriptional terminators and the lcnA operon

Using PCR techniques (see Table 6 below for primers ter2 and ter3), a DNA fragment present downstream of the replication genes of the lactococcal plasmid pWVO1 which contains a strong transcription terminator (fragment IR IV; 26) was produced which was flanked by BglII sites. This fragment was introduced into theBamHI site of the pUC-derived plasmid pMTL25. E. coli transformants were selected on X-gal plates, on which plasmids containing an insert appear as white colonies. Both orientations of the insert were obtained in different recombinant plasmids (denoted pMTL25IV in Figure 19).

In the following step a PCR fragment (see Table 6 below for primers lat2 and lat3) containing the lactococcin A (lcnA) operon flanked by BamHI sites was inserted in the BglII site of one of the pMTL25IV derivatives. In one of the derivatives

(pMTL25lcnIV) the lcnA operon is flanked on one side by its own terminator and at the other side by the IR IV terminator. Read-through transcription from sequences in between these terminators should be reduced in pMTL25lcnIV. The relevant fragment (the cassette containing the lcnA operon flanked by terminators) was exised from pMTL25lcnIV as a PstI fragment which was inserted between the EcoRI and BamHI sites of the pWVO2-derived plasmid pLR300 (these sites were made blunt by Mung Bean nuclease treatment). The ligation mixture was used to transform L. lactis IL1403. This strain carries the genes required for the export of the lactococcin (LcnA) but is sensitive to LcnA in the absence of the LciA immunity protein. Transformants were selected on erythromycin-containing plates and analyzed for their plasmid content. One of the resulting plasmids which contained the desired insert was denoted pLCN302. Table 6. List of oligo-nucleotide primers

name PCR fragment sequence ter2 IR IV GAAGATCTTGATTTATTGAGAGGAGGGATTATTG ter3 IR IV GAAGATCTGCTATTAATCGCAACATCAAACC lat2 lcnA-operon CGCGGATCCGAGTTATTAACATTTGT

lat3 lcnA-operon CGCGGATCCTACTGATTGCCTCTTCCC

pjs43 spc-gene TCCCCCGGGCATATGGATCCCCCGATTTTCGTTCG pjs44 spc-gene TGGGGTACCGCGGATCCTTTATTGTTTTCTAAAATC pjs45 regulon TCCACGCGTCCTCGGGATATGATAAG

pjs46 regulon TCCCTCGAGCTTGCGTTTGATTTTC

B. Introduction of the sucrose operon

The sucrose operon, containing scrB, scrA and the intact scrR gene, was present on plasmid pORI128 (Figure 20) which contains only sequences derived from L. lactis and P. pentosaceus.

This plasmid pORI128 was made as follows.

First pORI28 was constructed by ligating the 0.6kb TaqI Ori⁺ fragment of pWVO1 (26) to the multiple cloning site of pUK21 (75), isolated as a Spel fragment, containing in the XhoI site an 1kb fragment of plasmid pE194 (18) carrying the erythromycin resistance gene. Both fragments were made blunt end using Klenow enzyme and after ligation and transformation, plasmid pORI28 was recovered (Figure 20.A).

Secondly, the sucrose genes scrA, scrB and scrR were isolated from the Pediococcus pentosaceus PPE1.0 plasmid pSRQ1 (64) as an AatII-MluI fragment and ligated with pORI28 after treatment thereof with AatII (site at 1036) and MluI (site at 1029), resulting in plasmid pORI128 (Figure 20.B). This plasmid is able to replicate only in special helper strains, in which the Rep gene is integrated in the chromosome, while this plasmid integrates in normal Rep- L. lactis strains (67). The operon could be excised as an AatII/MluI fragment and was cloned into the AatII/MluI site of pLCN302. Transformants were initially selected on erythromycin-containing plates. To select for plasmids containing the sucrose genes, transformants were toothpicked onto SM17 plates containing the colour-indicator BCP. Yellow colonies were analyzed for their plasmid content. All selected plasmids appeared to contain a functional sucrose operon. One of the resulting plasmids (pLCN302S) was chosen for further use. To render pLCN302S food-grade, the Em^R gene was removed by digesting the plasmid with SmaI and StuI and subsequent religation, resulting in pLCN312.

Analysis of the expression of the lcnA operon

pLCN302 containing cells were streaked on GM17 plates so that about 20 to 30 colonies arose per plate after overnight incubation. Colonies were overlayed with 0.5% GM17 agar, containing L. lactis IL1403 cells (3 μl of an overnight culture in 3½ ml of 0.5% GM17 agar). When LcnA is synthesized, it will be secreted and diffuse around the producing colony. The L. lactis strain IL1403 is sensitive to LcnA in the absence of LciA and will not be able to grow around colonies which produce LcnA. In an overlay experiment this will show up as a halo around these colonies. All colonies containing the lcnA operon gave halo's, which indicates that both the lcnA and lciA genes were expressed.

Maintenance of pLCN312 in L. lactis MG1363

To study whether the expression of the sucrose genes flanked by transcriptional terminators affected the segregational stability of pLCN312, this plasmid had to be transferred into L. lactis MG1363. This assay could not be carried out with strain IL1403 since this strain is sensitive to LcnA and plasmid-free segregants would be killed after the loss of the plasmid containing the lciA gene. This, in fact, means that the pLCN312/strain IL1403 food-grade vector/host pair should be very stable, since plasmid-free cells will not survive.

MG1363 is not sensitive to LcnA, thus LcnA does not act as a selection marker in strain MG1363. For the introduction of pLCN312 into MG1363 the BCP shift from purple to yellow was used as an indicator of successful transfer. All yellow colonies turned out to carry the plasmid. The segregational stability, as determined by the yellow-white phenotype of colonies, turned out to be the same as that of pLR330 and pLR331: the plasmid was completely stable when grown on glucose, but slightly unstable when grown on sucrose. As observed with pLR330/pLR331, the maintenance of pLCN312 was much better in cells from colonies grown on sucrose-containing plates. After replating cells from resuspended colonies, all colony-forming units contained the plasmid. If, however, the cells from the resuspended colony were first grown in sucrose-containing M17 medium before being replated, the plasmid-containing cell fraction was about 45% after 25 generations. No further reduction in the plasmid-containing cell fraction was observed even after 80 generations (Figure 21).

Construction of a versatile food-grade cloning/expression vector

To develop pLCN312 into a more versatile vector, the lactococcin A operon was replaced by an expression cassette containing the P32 lactococcal promoter (74), a multiple cloning site and a transcriptional terminator. To achieve this, the following method of construction was used (Figure 22). The expression signals already present in pMG36E (72) were used. Several restriction sites present on this expression cassette would, however, not be unique in the desired final plasmid. Using the primers pjs43 and pjs44 (Table 6 above), which contained restriction sites unique for the final plasmid, the spectinomycin resistance gene from Enterococcus faecalis (66) was amplified by PCR and cloned into the SmaI/KpnI site of pMG36E, resulting in pMG36E-Spc. The region of pMG36E-Spc, containing the P32 promoter, the spectinomycin resistance gene and the transcriptional terminator of the lactococcal proteinase (prtP) gene, was subsequently amplified by PCR using primers pjs45 and pjs46 (Table 6 above) which contained sites that flanked the lcnA operon in pLCN312. The lcnA operon of pLCN312 was replaced by the modified expression cassette, resulting in plasmid pJR200. Transformants containing this plasmid were selected on GM17 plates containing 250 μg spectinomycin/ml. To render pJR200 food-grade, this plasmid was digested with BamHI to remove the spectinomycin resistance gene and religated resulting in pJR201. MATERIALS AND METHODS

Bacterial strains, growth conditions and plasmids

All strains and plasmids used in this study are listed in Table 7 below. E. coli cells were transformed as described (51) and plated on LB agar or grown in LB medium (54), supplemented with 100 μg/ml erythromycin or 75 μg/ml ampicillin when required. Stability tests of plasmids were carried out with L. lactis cells grown in M17 medium (35) containing 0.5% glucose or 0.5% sucrose. L. lactis cells were

transformed by electroporation (46) and plated on M17 agar, containing 0.5 M sucrose and appropriate antibiotics.

Table 7. Bacterial strains and plasmids bacterial strain properties source or or plasmid reference

E. coli

JM101 supE thiΔ(lac-proAB) [F'traD36,

proAB, lacI^qlacZΔM15 (Yanish-Perron et al., 1985)

L. lactis

ssp lactis MG1363 plasmid-free, insensitive to LcnA Lab collection ssp lactis IL1403 plasmid-free, sensitive to LcnA Lab collection Plasmids

pWVO2 3.8 kb; cryptic; from L. lactis Wg2 (Otto et al., 1983) pLR300 Em^R, 5.6 kb; pWVO2 derivative containing in the

HpaI site the Em^R gene from pE194 (18) (Part II) pINT123E Integration vector, containing the

sucrose operon and the Em^R gene

of pE194 This work pORI128 Special purpose vector, unable of autonomous

replication in L. lactis; carries the scrA,

scrB and scrR genes This work pLR330 pWVO2, carrying the sucrose genes

scrA, scrB and a truncated scrR This work pLR331 Like pLR331, with the sucrose

genes in the opposite orientation This work pMTL25 Ap^R, 2.4 kb; pUC derivative

containing lacZα gene (Chambers et al., 1988) pMTL25IV pMTL25, containing IR IV of pWVO1 This work pMTL25lcnIV pMTL25, carrying both IR IV and

the lcnA operon This work pLCN302 pLR300, carrying a PstI insert from

pMTL25lcnIV with both the IR IV

and the lcnA operon This work pLCN302S pLCN302, carrying a 6 kb AatII/MluI fragment of

pORI128 on which the entire sucrose operon is

present This work pLCN312 pLCN302S from which the Em^R has been removed

by SmaI/StuI digestion This work pMG36-E Em^R, 3.9 kb; expression vector (Van de Guchte et al., 1989) pJR200 pLCN312 with spc^R and without IcnA operon This work pJR201 pJR200 without spc^R but with expression cassette This work

DNA manipulations

Isolation of plasmid DNA by the alkaline lysis method, amplification of DNA by PCR and subcloning of DNA were essentially as described (54). Expression of lcnA and IciA

Expression of the lcnA genes was tested with an indicator strain that is sensitive to LcnA. For this purpose L. lactis IL1403 was chosen. L. lactis IL1403 cells, containing the lcnA operon on a pWVO2-derived plasmid, were grown overnight on plates. The colonies were overlayed with 3.5 ml of 0.5% GM17 agar, containing 3 μl of an overnight culture of plasmid-free L. lactis IL1403 cells. Expression of lcnA could be observed as halo formation around LcnA-producing cells as a consequence of growth inhibition of the indicator strain by LcnA. LcnA-producing L. lactis IL1403 cells can only grow in the presence of an active immunity protein, LciA. Therefore lciA must be expressed in LcnA-producing cells.

Assay of plasmid maintenance

Single colonies of plasmid-carrying strains, which had been purified by two successive platings on SM17-BCP agar, were used to inoculate M17 medium containing either glucose or sucrose. The cells were kept growing for about 80 generations in successive batch cultures. At regular intervals appropriate dilutions of the cultures were plated on SM17-BCP agar. The yellow colonies contained a plasmid, whereas the colourless ones did not.

NON-PATENT LITERATURE REFERENCES OF PART I

1. Alonso, J.C., J.-F. Viret, and R.H. Tailor. 1987. Plasmid maintenance in Bacillus subtilis recombination deficient mutants. Mol. Gen. Genet. 208: 349-352.

2. Bron, S. 1990. Plasmids, p. 75-175. In C.R. Harwood, and S.M. Cutting (ed.), Molecular biological methods for Bacillus. John Wiley & Sons Ltd, Chichester, UK.

3. Bron, S., S. Holsappel, G. Venema, and B.P.H. Peeters. 1991. Plasmid deletion formation between short direct repeats in Bacillus subtilis is stimulated by single- stranded rolling-circle-replication intermediates. Mol. Gen. Genet. 226: 88-96.

4. Bron, S., and E. Luxen. 1985. Segregational instability of pUB110-derived

recombinant plasmids in Bacillus subtilis. Plasmid 14: 95-244.

5. Bron, S., E. Luxen, and P. Swart. 1988. Instability of recombinant pUB110

plasmids in Bacillus subtilis: plasmid-encoded stability function and effects of

DNA inserts. Plasmid 19: 231-241. 6. Bron, S., W. Meijer, S. Holsappel, and P. Haima. 1991. Plasmid instability and molecular cloning in Bacillus subtilis. Res. Microbiol. 142: 875-883.

7. Brosius, J., A. Ullrich, M.A. Raker, A. Gray, T.J. Dull, R.R. Gutell, and H.F.

Noller. 1981. Construction and fine mapping of recombinant plasmids containing the rrnB ribosomal RNA operon of Escherichia coli. Plasmid 6: 112-118.

8. Bruand, C., S. Ehrlich, and L. Janniere. 1991. Unidirectional theta replication of the structurally stable Enterococcus faecalis plasmid pAMβ1. EMBO J. 10: 2171- 2177.

9. Chambers, S.P., S. Prior, D.B. Barstow, and N.P. Minton. 1988. The pMTL series of cloning vectors: Improved polylinker regions to facilitate the generation of sonicated DNA for nucleotide sequencing. Gene 68: 139-149.

10. Chomczynski, P., and P.K. Qasba. 1984. Alkaline transfer of DNA to plastic

membrane. Biochem. Biophys. Res. Commun. 22: 340-344.

11. Dabert, P., S.D. Ehrlich, and A. Gruss. 1991. High-molecular-weight linear

multimer formation by single-stranded DNA plasmids in Escherichia coli. J. Bacteriol. 174: 173-178.

12. de Vos, W.M. 1987. Gene cloning and expression in lactic streptococci. FEMS Microbiol. Rev. 46: 281-295.

13. de Vos, W.M. 1986. Gene cloning in lactic streptococci. Neth. Milk Dairy J. 40:

141-154.

14. Ehrlich, S.D., P. Noirot, M.A. Petit, L. Janniere, B. Michel, and H. te Riele. 1986.

Structural instability of Bacillus subtilis plasmids, p. 71-83. In J. K. Setlow and A. Hollaender (ed.), Genetic engineering, vol 8. Plenum Press, New York.

15. Gruss, A., and S. Ehrlich. 1988. Insertion of foreign DNA into plasmids from

gram-positive bacteria induces formation of high-molecular-weight plasmid multimers. J. Bacteriol. 170: 1183-1190.

16. Gruss, A., and S.D. Ehrlich. 1989. The family of highly interrelated single- stranded deoxyribonucleic acid plasmids. Microbiological Reviews 53: 231-241.

17. Horinouchi, S., and B. Weisblum. 1982. Nucleotide sequence and functional map of pC194, a plasmid that specifies inducible chloramphenicol resistance. J.

Bacteriol. 150: 815-825. 18. Horinouchi, S., and B. Weisblum. 1982. Nucleotide sequence and functional map of pE194, a plasmid that specifies inducible resistance to macrolide, lincosamide, and streptogramin type B antibiotics. J. Bacteriol. 150: 804-814.

19. Janniere, L., C. Bruand, and S.D. Ehrlich. 1990. Structurally stable Bacillus

subtilis cloning vectors. Gene 87: 53-61.

20. Janniere, L., B. Niaudet, E. Pierre, and S.D. Ehrlich. 1985. Stable gene

amplification in the chromosome of Bacillus subtilis. Gene 87: 47-55.

21. Kok, J. 1991. Special-purpose vectors for Lactococci. In G.M. Dunny, P.P. Cleary, and L.L. Mackay (ed.), Genetics and molecular biology of streptococci, lactococci, and enterococci. American Society for Microbiology, Washington, DC 20005, pp.

97-102.

22. Kok, J., J.M.B.M. van der Vossen, and G. Venema. 1984. Construction of plasmid cloning vectors for lactic streptococci which also replicate in Bacillus subtilis and Escherichia coli. Appl. Environ. Microbiol. 48: 726-731.

23. Kusano, K., K. Nakayama, and H. Nakayama. 1989. Plasmid-mediated lethality and plasmid multimer formation in an Escherichia coli recBC sbcBC mutant. J. Mol. Biol 209: 623-634.

24. le Bourgeois, P., M. Mata, and P. Ritzenthaler. 1991. Pulsed-field gel

electrophoresis as a tool for studying the phylogeny and genetic history of lactococcal strains, p. 140-145. In G.M. Dunny, P.P. Cleary, and L.L. McKay (ed.),

Genetics and molecular biology of streptococci, lactococci, and enterococci.

American Society for Microbiology, Washington, D.C.

25. Leenhouts, K.J., J. Kok, and G. Venema. 1989. Campbell-like integration of

heterologous plasmid DNA into the chromosome of Lactococcus lactis subsp. lactis. Appl. Envir. Microbiol. 2: 394-400.

26. Leenhouts, K.J., B. Tolner, S. Bron, J. Kok, G. Venema, and J.F.M.L. Seegers.

1991. Nucleotide sequence and characterisation of the broad host range lactococcal plasmid pWVO1. Plasmid 26: 55-66.

27. Leonhardt, H., and J.C. Alonso. 1991. Parameters affecting plasmid stability in Bacillus subtilis. Gene 103: 107-111. 28. Luchansky, J.B., P.M. Muriana, and T.R. Klaenhammer. 1988. Application of electroporation for transfer of plasmid DNA to Lactobacillus, Leuconostoc, Pediococcus, Bacillus, Staphylococcus, Enterococcus and Propionibacterium. Mol. Microbiol. 2: 637-646.

29. Maniatis, T., E.F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a

laboratory manual. Cold Spring Harbour Laboratory, Cold Spring Harbour, New York.

30. Messing, J. 1979. A multipurpose cloning system based on the single stranded DNA bacteriophage M13. Recombinant DNA Technical Bulletin, NIH

Publication, No. 79-99, 2: 43-48.

31. Sanger, F., A.R. Coulson, G.F. Hong, D.F. Hill, and G.B. Petersen. 1982.

Nucleotide sequence of bacteriophage lambda DNA. J. Mol. Biol. 162: 729-773.

32. Simon, D., and A. Chopin. 1988. Construction of a vector plasmid family and its use for molecular cloning in Streptococcus lactis. Biochimie 70: 559-566.

33. Suvorov, A., J. Kok, and G. Venema. 1988. Transformation of group A

streptococci by electroporation. FEMS Microbiol. Lett. 56: 95-100.

34. Swinfleld, T.-J., J. Oultram, D. Thompson, J. Brehm, and N. Minton. 1990.

Physical characterization of the replication region of the Streptococcus faecalis plasmid pAMβ1. Gene 87: 79-90.

35. Terzaghi, B.E., and W.E. Sandine. 1975. Improved medium for lactic streptococci and their bacteriophages. Appl. Microbiol. 29: 807-813.

36. Viret, J.-F., A. Bravo, and J.C. Alonso. 1991. Recombination-dependent

concatemeric plasmid replication. Microbiol. Reviews 55: 675-683.

37. Viret, J.-F. and J.C. Alonso. 1987. Generation of multigenome-length plasmid molecules in Bacillus subtilis. Nucleic Acids Res. 15: 6349-6367.

ADDITIONAL non-patent literature references of part II

38. Brewer B.J. and Fangman W.L. (1987) The localization of replication origins on ARS plasmids in S. cerevisiae. Cell 51: 463-471. 39. Brewer B.J., Sena E.P. and Fangman W.L. (1988) Analysis of replication intermediates by two-dimensional agarose gel electrophoresis. Cancer Cells 6: 229-234.

40. Bron S. (1990) Plasmids. In Molecular biological methods for Bacillus. Harwood C. R. and Cutting S. M. (eds). Chichester, England:John Wiley and Sons Ltd., pp.75-174.

41. Clewell D.B., Yagi Y., Dunny M.G. and Schultz S.K. (1974) Characterization of three plasmid deoxyribonucleic acid molecules in a strain of Streptococcus faecalis: identification of a plasmid determining erythromycin resistance. J. Bacteriol. 117: 283-289.

42. Dorrington R.A. and Rawlings D.E. (1989) Identification and sequence of the basic replication region of a broad-host-range plasmid isolated from Thiobacillus ferrooxidans. J. Bacteriol. 171: 2735-2739.

43. Gasson M.J. (1983) Plasmid complements of Streptococcus lactis NCDO712 and other lactic streptococci after protoplast-induced curing. J. Bacteriol. 154: 1-9.

44. Hayes F., Vos P., Fitzgerald G.F., De Vos W.M. and Daly C. (1991) Molecular organization of the minimal replicon of novel, narrow-host-range, lactococcal plasmid pCI305. Plasmid 25: 16-26.

45. Heck M.M.S. and Spradling A.C. (1990) Multiple replication origins are used during Drosophila chorion gene amplification. J. Cell Biol. 110: 903-914.

46. Holo H. and Nes I.F. (1989) High-frequency transformation, by electroporation, of Lactococcus lactis subsp. cremoris grown with glycine in osmotically stabilized media. Appl. Environ. Microbiol. 55: 3119-3123.

47. Horng J.S., Polzin K.M. and McKay L.L. (1991) Replication and

temperature-sensitive maintenance functions of lactose plasmid pSK11L from

Lactococcus lactis subsp. cremoris. J. Bacteriol. 173: 7573-7581.

48. Hyrien O. and Mechali M. (1992) Plasmid replication in Xenopus eggs and egg extracts: A 2D gel electrophoretic analysis. Nucleic Acids Res. 20: 1463-1469.

49. Jahns A., Schafer A., Geis A. and Teuber M. (1991) Identification, cloning and sequencing of the replication region of Lactococcus lactis ssp. lactis biovar.

diacetylactis Bu2 citrate plasmid pSL2. FEMS Microbiol. Lett. 80: 253-258. 50. Mahbubani H.M., Paull T., Elder J.K. and Blow JJ. (1992) DNA replication initiates at multiple sites on plasmid DNA in Xenopus egg extracts. Nucleic Acids Res. 20: 1457-1462.

51. Mandel M. and Higa A. (1970) Calcium-dependent bacteriophage DNA infection.

J. Mol. Biol. 53: 159-162.

52. Otto R., De Vos W.M. and Gavrieli J. (1982) The influence of pH on the

selection in lactose-limited chemostats of a mutant of Streptococcus cremoris lacking a protease plasmid. Appl. Environ. Microbiol. 43: 1272-1277.

53. Persson C. and Nordstrom K. (1986) Control of replication of the broad host range plasmid RSF1010: The incompatibility determinant consists of directly repeated sequences. Mol. Gen. Genet. 203: 189-192.

54. Sambrook J., Fritsch E.F. and Maniatis T. (1989) Molecular Cloning: A

Laboratory Manual. Cold Spring Harbor, NY.

55. Sanger F., Nicklen S. and Coulsen A.R. (1977) DNA sequencing with

chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74: 5463-5467.

56. Short J.M., Fernandez J.M., Sorge J.A. and Huge W.D. (1988) Lambda ZAP: a lambda bacteriophage expression vector with in vivo excision properties. Nucleic Acids Res. 16: 7583-7600.

57. te Riele H., Michel B. and Ehrlich S. (1986a) Are single-stranded circles intermediates in plasmid DNA replication? EMBO J. 5: 631-637.

58. te Riele H., Michel B. and Ehrlich S. (1986b) Single-stranded plasmid DNA in Bacillus subtilis and Staphylococcus aureus. Proc. Natl. Acad. Sci. USA 83:

2541-2545.

59. Vaughn J.P., Dijkwel P.A. and Hamlin J.L. (1992) Replication initiates in a broad zone in the amplified CHO dihydrofolate reductase domain. Cell 61: 1075-1087.

60. Yanish-Perron C., Vieira J. and Messing J. (1985) Improved M13 phage cloning vectors and host strains: nucleotide sequences of M13mp18 and pUC19 vectors. Gene 33: 103-119. ADDITIONAL non-patent literature references of Part IV

61. Chopin A., Chopin M-C., Moillo-Batt A. and Langella P. (1984) Two

plasmid-determined restriction and modification systems in Streptococcus lactis. Plasmid 11: 260-263.

62. Kok J., Leenhouts K., Haandrikman A.J., Ledeboer A.M. and Venema G. (1988) Nucleotide sequence of the cell wall proteinase gene of Streptococcus cremoris Wg2. Appl. Environ. Microbiol. 54: 231-238.

63. Novick R.P. (1989) Staphylococcal plasmids and their replication. Annu. Rev.

Microbiol. 43: 537-565.

ADDITIONAL non-patent literature references of Part V

64. Gonzalez, C.F. & Kunka, B.S. (1986) Evidence for plasmid linkage of raffinose utilization and associated α-galactosidase and sucrose hydrolyse activity in

Pediococcus pentosaceus. J. Bacteriol. 51: 105-109

65. Kiewiet, R., Kok, J., Seegers, J.F.M.L., Venema, G. & Bron, S. (1993) The mode of replication is a major factor in segregational plasmid instability in Lactococcus lactis. Applied and Environmental Microbiology 59: 358-364.

66. Leblanc, D.J., Lee, L.N., Inamine, J.M. (1991) Cloning and nucleotide base

sequence analysis of a spectinomycin adenyltransferase AAD(9) determinant from Enterococcus faecalis. Antimicrob. Agents Chemother. 35: 1804-1810

67. Leenhouts, K.J., Kok, J., & Venema, G. (1991) Lactococcal Plasmid pWVO1 as an Integration Vector for Lactococci. Applied and Environmental Microbiology 57: 2562-2567

68. Liu, L.F. & Wang, J.C. (1987). Supercoiling of the DNA template during

transcription. Proc. Natl. Acad. Sci. (USA) 84: 7024-7027.

69. Lynch, A.S. & Wang, J.C. (1993) Anchoring of DNA to the bacterial cytoplasmic membrane through cotranscriptional synthesis of polypeptides encoding

membrane proteins or proteins foe export: a mechanism of plasmid hypernegative supercoiling in mutants deficient in DNA topoisomerase I. J. Bacteriol. 175: 1645- 1655. 70. Mayo, B., Kok, J., Venema, K., Bockelmann, W., Teuber, M., Reinke, H., & Venema, G. (1991) Molecular Cloning and Sequence Analysis of the X-Prolyl Dipeptidyl Aminopeptidase Gene from Lactococcus lactis subsp. cremoris. Appl. Environ. Microbiol. 57: 38-44.

71. Van Belkum, M.J., Hayema, B.J., Jeeninga, R.E., Kok, J., & Venema, G. (1991) Organization and nucleotide sequences of two lactococcal bacteriocin operons. Applied and Environmental Microbiology 57: 492-498.

72. Van de Guchte, M., Van der Vossen, J.M.B.M., Kok, J. & Venema, G. (1989) construction of a lactococcal expression vector: expression of hen egg white lysozyme in Lactococcus lactis subsp. lactis. Applied and Environmental

Microbiology 55: 224-228.

73. Van der Vossen, J.M.B.M., Kodde, J., Haandrikman, A.J., Venema, G. & Kok, J.

(1992) Characterization of transcription-initiation and -termination signals of the proteinase genes of Lactococcus lactis Wg2 and enhancement of proteolysis in Lactococcus lactis. Applied and Environmental Microbiology 58(9) 3142-3149.

74. Van der Vossen, J.M.B.M., Lelie, D. & Venema, G. (1987) Isolation and

characterization of Lactococcus lactis subsp. cremoris Wg2-specific promoters. Applied and Environmental Microbiology 53: 2452-2457.

75. Vieira, J. and Messing, J. (1991) New pUC-derived cloning vectors with different selectable markers and DNA replication origins. Gene 100: 189-194.

Claims

C L A I M S

1. A process for producing a desired protein or polypeptide comprising culturing a transformed lactic acid bacterium under conditions whereby a structural gene encoding said desired protein or polypeptide or a precursor thereof is expressed, in which process said lactic acid bacterium is one being transformed by means of introducing a food-grade recombinant plasmid, said plasmid comprising

(1) a replicon and any DNA sequences required for stable maintenance in a lactic acid bacterium, whereby

(1.a) said plasmid has a theta-replication mechanism and

(1.b) said plasmid is stably maintained in said transformed lactic acid bacterium under non-selective conditions, and

(2) a structural gene encoding said desired protein or polypeptide or a precursor thereof, said gene not being an antibiotic resistance selection marker, and optionally

(3) a selection marker.

2. A process according to claim 1, in which the replicon and any DNA sequences required for stable maintenance in the bacterium are derived from pWVO2 originating from Lactococcus lactis subsp. cremoris Wg2.

3. A transformed lactic acid bacterium, suitable for use in a process as claimed in claim 1, containing a food-grade recombinant plasmid comprising

(1) a replicon and any DNA sequences required for stable maintenance in a lactic acid bacterium, whereby

(1.a) said plasmid has a theta-replication mechanism and

(1.b) said plasmid is stably maintained in said transformed lactic acid bacterium under non-selective conditions, and

(2) a structural gene encoding said desired protein or polypeptide or a precursor thereof, said gene not being an antibiotic resistance selection marker, and optionally

(3) a selection marker.

4. A bacterium according to claim 3, in which the replicon and any DNA sequences required for stable maintenance in the bacterium are derived from pWVO2 originating from Lactococcus lactis subsp. cremoris Wg2.

5. A food-grade recombinant plasmid, suitable for transforming a lactic acid bacterium, comprising

(1) a replicon and any DNA sequences required for stable maintenance in a lactic acid bacterium, whereby

(1.a) said plasmid has a theta-replication mechanism and

(1.b) said plasmid is stably maintained in said transformed lactic acid bacterium under non-selective conditions, and

(2) a structural gene encoding said desired protein or polypeptide or a precursor thereof, said gene not being an antibiotic resistance selection marker,

and optionally

(3) a selection marker.

6. A plasmid according to claim 5, in which the replicon and any DNA sequences required for stable maintenance in the bacterium are derived from pWVO2 originating from Lactococcus lactis subsp. cremoris Wg2.

7. The use of a plasmid as claimed in claim 5 for modifying the metabolic pathway of a lactic acid bacterium.

8. The use according to claim 7, in which the production of diacetyl or a precursor thereof by a lactic acid bacterium is increased by transforming said bacterium with a plasmid as claimed in claim 5, which plasmid is capable of expressing a gene encoding α-acetolactate synthase.

9. The use according to claim 7, in which the plasmid contains one or more genes encoding enzymes that are capable of influencing the formation and structure of exopolysaccharides thereby producing functionally modified exopolysaccharides.

10. A transformed lactic acid bacterium having a modified metabolic pathway obtained as a result of a use as claimed in claim 7.

11. A food product obtained by the use of a process as claimed in claim 1 or by using a transformed lactic acid bacterium as claimed in claim 3 or in claim 10.

12. A food product according to claim 11, which is selected from the group consisting of (a) dairy products, especially fermented dairy products like yoghurt, quark and fermented milk; (b) ice cream and other frozen products containing milk or fermented milk; (c) meat and meat-containing products; (d) dressings and sandwich spreads; (e) butter and margarine; (f) low-fat butter- and margarine-substitutes; and (g) mayonnaise and low-fat mayonnaise-substitutes.

13. An animal feed product obtained by the use of a process as claimed in claim 1 or by using a lactic acid bacterium as claimed in claim 3 or in claim 10.

14. A process for the production or overproduction of a food-grade enzyme to be used for further enzymatic modifications, in which a transformed lactic acid bacterium as claimed in claim 3 is used under conditions whereby the enzyme is produced in a relatively high amount and can exert its action on a substrate for the enzyme.

* * * * *