Inducible Promoter-Repressor System from the Lactobacillus casei Phage {phi}FSW

With the aim to extend the presently available inducible geneexpression systems for lactobacilli, we have isolated a thermoinduciblepromoter-repressor cassette from the temperate Lactobacilluscasei phage ${phi}$ FSW-TI in Escherichia coli. The ${phi}$ FSW-TI promoterfragment was abutted to the plasmid-borne promoterless ß-glucuronidase(gusA) reporter gene and shown to direct its transcription inL. casei. In addition, the functionality of the promoter-repressorsystem was verified in the L. casei ${phi}$ FSW-TI lysogen by showingthat the gusA reporter gene, controlled by the isolated ${phi}$ FSW-TIpromoter, was repressed at 28°C and expressed at 42°C.Moreover, a homology search revealed that the C terminus ofthe isolated ${phi}$ FSW repressor shows a high similarity to the smallmutS-related domain of the MutS2 protein family that is unprecedentedfor phage-encoded repressor proteins.

INTRODUCTION

Top

Introduction

Given the economic importance of lactic acid bacteria, and morerecently their prospective use as safe vaccines, a number ofphysiological traits as well as food-grade expression toolsof lactobacilli and lactococci have been studied (9). Variousinducible gene expression systems have been developed for Lactococcuslactis (reviewed in reference 3). In general, these transcriptionalinduction systems require the exogenous supply of inducers (18,19, 21). In addition, a heat-inducible homologous expressionsystem has been developed for L. lactis based on a geneticallyengineered thermolabile repressor of the lactococcal phage r1t(17). In lactobacilli, studies on the regulation of gene expressionhave focused mainly on carbon catabolism pathways, and an inducible-repressibleexpression system derived from the lactose operon of Lactobacilluscasei has been described (5 and references therein).

The temperate phage ${phi}$ FSW-TI was isolated from L. casei strainShirota (S1) after chemical mutagenesis of a lysogen followedby heat induction (22, 24). Since the mutation responsible forthe thermoinducibility was shown to reside in the phage DNA,it seemed likely that it could affect the stability of the phagerepressor at the elevated temperature. In this study, we havecharacterized a promoter and a cognate repressor of ${phi}$ FSW-TI withthe potential to be exploited in a thermoinducible food-gradeexpression system for L. casei.

Isolation of a promoter and a cognate heat-labile repressor of ${phi}$ FSW-TI in Escherichia coli.

Top

Isolation of a promoter...

The following strategy was used to isolate the thermoinduciblepromoter-repressor cassette of phage ${phi}$ FSW-TI. The ${phi}$ FSW-TI DNAwas first shotgun cloned into the E. coli promoter test vectorpKK232-8 (Amersham Pharmacia) carrying a promoterless cat gene.Chloramphenicol (Cm)-resistant clones were cotransformed witha ${phi}$ FSW-TI library generated in the compatible plasmid pK194 (7).The DNA sequence encoding the protein with apparent repressoractivity and its cognate promoter sequence were then obtainedfrom clones that were sensitive and resistant to the antibioticat 28 and 42°C, respectively.

Phage ${phi}$ FSW-TI DNA was prepared as described earlier (23). TheDNA was cleaved with either HaeIII, RsaI, HincII, or EcoRV,and the corresponding fragments were cloned into the SmaI siteof the promoter test vector pKK232-8. The E. coli strain MC4100F'(25) was transformed with the ligation mix, and 51 candidateclones were selected on Luria-Bertani agar plates (14) containing50 µg of Cm/ml.

To isolate the cognate repressor gene, ${phi}$ FSW-TI DNA was partiallydigested with the same enzymes used for the procedure describedabove. The restriction fragments were cloned into the SmaI siteof plasmid pK194, and the ligation mix was transformed intoE. coli strain MC4100F'. From the pool of the transformantsobtained on Luria-Bertani agar supplemented with 50 µgof kanamycin (Km)/ml, a batch plasmid isolation was performed.Then each of the 51 candidate promoter clones was transformedwith the batch plasmid preparation. Cotransformants were selectedon Cm/Km plates at 42°C, a temperature at which the putativerepressor was predicted to be inactive, and thus, transcriptionof the cat gene in plasmid pKK232-8 was not supposed to be repressed.One hundred and fifty cotransformants were replica plated onCm/Km at 28°C. The recombinant plasmids of five cotransformantsunable to grow at 28°C were isolated from the correspondingcolonies on the master plate and retransformed into MC4100F'in order to obtain colonies containing the individual plasmids.The inserts of two corresponding plasmid pairs were sequencedusing an Applied Biosystems DNA sequencer. Both pairs containedthe same core sequences. The plasmids with the shortest insertswere designated pKK232-8P (524-bp insert containing the putativepromoter) and pK194R-1 (1,557-bp insert containing the putativerepressor gene) and were used in further experiments.

Characterization of the promoter region.

Top

Characterization of the promoter...

The DNA sequence of the 524-bp insert (National Center for BiotechnologyInformation [NCBI] accession no. AF458097) present in plasmidpKK232-8P was analyzed for putative promoter sequences usingthe promoter search program available on the website http://www.dot.imgen.bcm.tmc.edu:9331/cgi-b.The highest score was obtained for a putative promoter spanningbp 275 (-35 hexamer) to bp 309 (predicted transcriptional start).To determine the transcriptional start site, primer extensionwas performed as described earlier (6) with total RNA preparedfrom E. coli MC4100F' harboring plasmid pKK232-8P containingthe sequenced 524-bp insert with the putative promoter. Theoligonucleotide V10, 5'-GACATGGGAGGCACCAGC-3', which binds 58bp downstream of the putative -10 region, was used as a primerfor the reverse transcriptase reaction. As shown in Fig. 1A,the primer extension analysis revealed that the transcript startswith the G located 6 bp downstream of the -10 hexamer. The alignmentwith the previously defined Lactobacillus consensus sequence(20) showed five matches for both the -35 and -10 hexamers,which are separated by 18 bp (Fig. 1B). The distance betweenthe -10 hexamer and the transcriptional start site is 5 bp (Fig.1B), which is within the range found for other Lactobacilluspromoters (13).

View larger version (29K):
[in this window]
[in a new window]
FIG. 1. (A) Determination of the transcriptional start site of the isolated ${phi}$ FSW promoter ( ${phi}$ FSW-P). Lane 1, the oligonucleotide V10 complementary to the sequence downstream of the predicted start point of ${phi}$ FSW-P was used to prime cDNA synthesis of RNA isolated from strain MC4100F'(pKK232-8P). The 3'-terminal nucleotide of the cDNA complementary to the 5' end of the transcript assigned to a G residue is marked by an arrow. The -10 hexamer is shaded. Reference sequencing reactions were performed with the same primer on double-stranded plasmid pKK232-8P as template. (B) Comparison of ${phi}$ FSW-P with the promoter consensus sequence for lactobacilli. Boldface type indicates identity with consensus (20). The transcriptional start site (TS) is underlined. Dots are used for the alignment of the -35 and -10 hexamer sequences.

The functionality of the isolated ${phi}$ FSW promoter was verifiedin L. casei S1 (23) by transforming this strain with a plasmidin which the promoter (including upstream and downstream sequences:bp -99 to +35 with regard to the transcriptional start [+1])was abutted to the gusA reporter gene. The selected ${phi}$ FSW promoterwas cloned 5' of the gus gene in plasmid pNZ8008 (18) as follows:the promoter fragment was obtained by PCR amplification usingplasmid pKK232-8P and the forward primer 5'-TATCTGCAGTG-AATCAAAGCCGTGGCGTTA-3'together with the reverse primer 5'-CCGAATTCGCGGCCAGCGCAATGAGCTT-3'containing a PstI site and an EcoRI site, respectively. Thepurified PCR fragment was cleaved with PstI and EcoRI and clonedinto the corresponding sites of plasmid pNZ8008, resulting inplasmid pNZFSWP. The plasmid was transformed in L. casei S1as described earlier (2). Colonies of L. casei S1 transformedwith pNZFSWP showed ß-glucuronidase activity, whereascolonies of L. casei S1 transformed with the parental vectorpNZ8008 did not, demonstrating that the ${phi}$ FSW promoter was directingtranscription of the gus reporter gene in the L. casei strain(not shown).

Characterization of the cognate repressor.

Top

Characterization of the cognate...

As mentioned above, plasmid pK194R-1 contained a 1,557-bp insertof ${phi}$ FSW-TI DNA comprising five open reading frames (ORFs) (Fig.2). To identify the repressor gene, four different restrictionfragments (Fig. 2) were isolated from plasmid pK194R-1 and reclonedinto plasmid pK194. Plasmid pK194R-2 contained the 3' end ofthe 1,557-bp fragment covering ORF4 and ORF5, whereas plasmidpK194R-3 contained ORF3 in addition to ORF4 and ORF5. PlasmidpK194R-4 harbored the 5'-proximal reading frames ORF1 to ORF3,while plasmid pK194R-5 contained only ORF1. These four plasmidswere individually transformed into E. coli strain MC4100F'(pKK232-8P).The presence of either plasmid pK194-R-4 or pK194R-5 in strainMC4100F'(pKK232-8P) conferred the same levels of Cm resistanceat 42°C (expression of the cat gene due to thermal inactivationof the repressor) and resulted in Cm sensitivity at 28°C(repression of the cat gene due to the presence of active repressorprotein). The repressor gene was therefore assigned to ORF1.

View larger version (11K):
[in this window]
[in a new window]
FIG. 2. Identification of the ORF encoding the repressor by deletion analysis. As described in the text, plasmid pK194R-1 bears a 1,557-bp DNA insert of ${phi}$ FSW containing five putative ORFs. The restriction sites used for isolation of the different subfragments and for reinsertion into plasmid pK194 are indicated on top. The restriction sites are indicated by capital letters: H, HindIII; B, BamHI; X, XmaI; E, EcoRI; and S, SmaI. The lengths of the DNA fragments present in plasmids pK194R-2 to -5 are schematically shown. + and - signs signify repressor activity (i.e., Cm sensitivity at 28°C) in strain MC4100F'(pKK232-8P) cotransformed with either of the depicted plasmids. ORF1 (indicated by a black bar) encodes the Smr-homologous ${phi}$ FSW repressor.

To test qualitatively whether the identified ${phi}$ FSW-TI promoteris controllable as well by the ${phi}$ FSW-TI repressor in L. casei,the L. casei ${phi}$ FSW-TI lysogen was transformed with plasmid pNZFSWPbearing the gusA reporter gene under transcriptional controlof the isolated ${phi}$ FSW-TI promoter. Since no ß-glucuronidaseactivity was detected at 28°C, whereas thermal inductionat 42^o resulted in expression of the reporter gene (not shown),we concluded that the thermoinducible promoter-repressor systemis functional in L. casei.

As revealed by a PSI-Blast search available on the website http://www3.ncbi.nlm.nih.gov/cgi-bin/BLAST/nph-psi_blast,the isolated 113-amino-acid (aa) ${phi}$ FSW-TI repressor protein (NCBIaccession no. AF406970) shows a high similarity (E = 2 x 10^-18)to the small mutS-related (Smr) consensus motif (15). The Smrprotein in turn has significant homology to the C terminus ofMutS2 proteins (Fig. 3), which are presumably involved in DNAmismatch repair (4, 15). Smr homologs have been found in a numberof proteobacteria. The apparent absence of Smr proteins in allspecies that have mutS2 genes and their presence in speciesthat have mutS1-homologous genes have led to the speculationthat the Smr protein may play a role in mismatch repair throughan interaction with MutS1 (15). Malik and Henikoff (12) havesuggested that the Smr domain in MutS2 proteins and Smr proteinlinked via MutL to MutS might act as a nicking endonuclease,thus replacing the E. coli MutH nuclease.

View larger version (60K):
[in this window]
[in a new window]
FIG. 3. Primary sequence of the Smr-homologous repressor of phage ${phi}$ FSW ( ${phi}$ FSWR) and multiple alignment of the sequence shared by the repressor and the carboxy-terminal domain (Smr domain) of the MutS-like or MutS2 proteins from Streptococcus pyogenes (AE006610; 776 aa), Streptococcus pneumoniae (AE007352; 778 aa), L. lactis (AE006395; 776 aa), Bacillus halodurans (AP001517; 785 aa), Staphylococcus aureus (AP003132; 782 aa), and the Smr-like protein of Listeria monocytogenes (AJ133006; 131 aa). Given in parentheses are the NCBI accession number and the length of the corresponding proteins in number of amino acids. The Smr consensus sequence is shown at the bottom. The homologous regions of the protein domains were identified by BlastP 2.2.1 (1). Multiple alignments were done with CLUSTALW (26) and created using BOXSHADE. Numbers at the left correspond to distances (in number of amino acids) from the N terminus of the respective proteins. Identical amino acids are highlighted in black, whereas amino acids in grey symbolize conservative exchanges.

Several E. coli phage ${lambda}$ -homologous repressors have been identifiedfrom temperate phage of lactobacilli (8, 10) and L. lactis (16).In general, these repressor proteins undergo self-cleavage uponDNA damage via a RecA-mediated pathway, which results in prophageinduction (11). In contrast, the ${phi}$ FSW prophage was not inducibleby either UV irradiation or mitomycin treatment (22), underliningthat the prophage status of ${phi}$ FSW-TI is not controlled by a canonicalrepressor. To our knowledge this is the first time that a repressoractivity has been attributed to an Smr-homologous protein. Sincethe Smr protein and/or domain has apparent DNA binding activity,it is conceivable that the phage has acquired from its hostthe smr gene, which may then have gained repressor function.It remains to be seen whether the isolated ${phi}$ FSW-TI repressorrepresents the founding member of a new family of repressorscontaining a region of homology to Smr proteins.

ACKNOWLEDGMENTS

We are grateful to M. Shimizu-Kadota for providing phage ${phi}$ FSW-TIand L. casei strain S1, to O. P. Kuipers for providing plasmidpNZ8008, and to A. Witte for computer help.

This work was supported by grant 10006 from the Austrian ScienceFund (FWF) and grant 6170 from the Austrian National Bank toU.B.

FOOTNOTES

* Corresponding author. Mailing address: Institute of Microbiology and Genetics, Vienna Biocenter, Dr. Bohrgasse 9, 1030 Vienna, Austria. Phone: 43-1-4277-54609. Fax: 43-1-4277-9546. E-mail: Udo.Blaesi{at}univie.ac.at.

REFERENCES

Top