Transcriptional Analysis and Regulation of Expression of the ScrFI Restriction-Modification System of Lactococcus lactis subsp. cremoris UC503

Abstract

ScrFI is a type II restriction-modification system from Lactococcus lactis which recognizes the nucleotide sequence 5′-CC↓ NGG-3′, cleaving at the point indicated by the arrow, and it comprises an endonuclease gene that is flanked on either side by genes encoding two 5-methylcytosine methylases. An open reading frame (orfX) of unknown function is located immediately upstream of these genes. In this study Northern analysis was performed, and it revealed that orfX, scrFIBM, and scrFIR are cotranscribed as a single polygenic mRNA molecule, while scrFIAM is transcribed independently. 5′ extension analysis indicated that the start site for the scrFIAM promoter was a thymine located 4 bp downstream of the −10 motif. The transcriptional start site for the orfX promoter was also found to be a thymine which is more atypically located 24 bp downstream of the −10 motif proximal to the start codon. A helix-turn-helix motif was identified at the N-terminal end of one of the methylases (M.ScrFIA). In order to determine if this motif played a role in regulation of the ScrFI locus, M.ScrFIA was purified. It was then employed in gel retardation assays using fragments containing the two promoters found on the ScrFI operon, one located upstream of orfX and the other located just upstream of scrFIAM. M.ScrFIA was found to bind to the promoter region upstream of the gene encoding it, indicating that it may have a regulatory role. In further studies the two putative promoters were introduced into a vector (pAK80) upstream of a promoterless lacZ gene, and cloned fragments of the ScrFI locus were introduced in trans with each of these promoter constructs to investigate the effect on promoter activity. These results implicated M.ScrFIA in regulation of both promoters on the ScrFI locus.

The potentially lethal endonuclease activity characteristic of all restriction-modification (R/M) systems necessitates maintenance of the inherent coordination between the restriction and methylation activities. Consequently, the genetic elements of R/M systems are usually tightly regulated, and the regulatory components are generally found close to the functional genes. This organization facilitates transfer of these elements as a single genetic unit. Two modes of control are commonly found; in some cases expression is directed by the product of a small open reading frame (ORF) which is separate from the functional genes of the operon (10, 16, 22). In other instances, the systems are controlled by a helix-turn-helix (H-T-H) motif situated at the N-terminal end of a methylase which regulates the system by interacting with the promoter region (5). To date, 11 R/M systems have been characterized to the DNA sequence level in Lactococcus, and the recognition sites of two more systems have been established, although the LlaI system is the only one for which a regulatory mechanism has been deduced (15, 16). Expression of this system is modulated by the product of a 254-bp ORF positioned at the start of the operon.

It has previously been demonstrated that the ScrFI system of Lactococcus lactis subsp. cremoris UC503 (Table 1) is encoded by an operon consisting of four ORFs (Fig. 1) (4, 23, 24). The entire operon has been sequenced and has previously been published (accession no. LI2227) (24). An endonuclease gene capable of encoding a 34-kDa protein is flanked by two 5-methylcytosine methylase genes, designated scrFIAM and scrFIBM, encoding 44.5- and 41.8-kDa proteins, respectively. Although the methylases recognize the same target sequence and exhibit significant sequence similarity, there is no definitive evidence to suggest that the two genes are the result of a recent duplication event. The difference between the methylases is illustrated by the fact that while M.ScrFIB exhibits significant sequence identity with M.ScrFIA (29.7%), it actually exhibits greater sequence identity with M.SssI (39.5%) (23). The orfX gene (size of putative product, 18.5 kDa; accession no. LI2227) upstream of these genes does not show significant homology to any other ORF in the gene database, and its function is not known yet (24).

TABLE 1.

Bacterial strains and cultures

Bacterial strain or plasmid	Relevant genotype or phenotype	Reference(s) or source
E. coli strains
M15(pREP)	Nals Strs Rifs Lac− Ara− Gal−mtl F−recA+ Uvr+	26
DB001	M15 derivative containing pDB001	This study
L. lactis subsp. cremoris strains
MG1363	Plasmid-free derivative of L. lactis subsp. cremoris 712	7
UC503	Isolate from Irish cheddar cheese mixed-strain starter encoding ScrFI R/M activity	4
DB002	MG1363 derivative containing pCI921	This study
DB003	MG1363 derivative containing pCI922	This study
Plasmids
pCI372	5.7-kb E. coli-L. lactis suttle vector, Cmr	8
pCI931m	pBR322-based chimeric construct containing scrFIAM modification gene from L. lactis subsp. cremoris UC503, Apr	2
pCI932m	pBR322-based chimeric construct containing scrFIBM modification gene from L. lactis subsp. cremoris UC503, Apr	2, 23
pREP4	Kanr, constitutively expresses the lac repressor protein	Qiagen
pQE60	Ampr, IPTG-inducible expression vectora	Qiagen
pDB001	pQE60 derivative containing scrFIAM	This study
pAK80	Cmr, contains promoterless β-Gal gene	9
pCI921	pAK80 derivative containing promoter region upstream of scrFIAM	This study
pCI922	pAK80 derivative containing promoter region upstream of orfX	This study
pCI923	pCI372 derivative containing scrFIAM	This study
pCI925	pCI372 derivative containing a deleted version of scrFIAM and a deleted version of scrFIR	This study
pCI926	pCI372 derivative containing a deleted version of scrFIR and the complete scrFIAM gene	This study
pCI934	pCI372 derivative containing orfX, scrFIBM, scrFIR, and scrFIAM	24
pCI941	pCI372 derivative containing the entire ScrFI locus with a deleted version of scrFIBM	24
pCI945	pCI372 derivative containing orfX and scrFIBM	24

^aIPTG, isopropyl-β-d-thiogalactopyranoside. 

FIG. 1.

Schematic representation of the molecular organization of the ScrFI R/M locus, indicating the position of the restriction endonuclease gene (scrFIR) flanked by two 5-methylcytosine methylase genes. An ORF of unknown function (orfX) and two ribosomal protein genes (rpmF and rpmG) are situated upstream of scrFIBM. Promoter sequences are indicated by P. (Modified from reference 6.)

The mechanism by which the genes involved in the ScrFI system are regulated was investigated in this study. This is of particular interest since the difficulties encountered during attempted cloning of the complete active locus may have been due to interference with the system's natural regulation (24). Northern and 5′ extension analyses were performed to determine the number and sizes of the ScrFI transcripts. In addition, the activities of the two promoter motifs were analyzed by performing expression analysis.

Transcriptional analysis of the ScrFI locus.

Previously, analysis of the ScrFI locus revealed that there are two putative promoter sequences present, one upstream of orfX and the other upstream of scrFIAM. Secondary structures in the form of a stem-loop situated between rpmG and orfX and another secondary structure just downstream of scrFIAM were also noted (Fig. 1) (6).

Northern analysis was performed to elucidate the manner in which the locus was transcribed. Aliquots of total RNA were extracted from lactococcal cultures in the early exponential phase as outlined by Keilhauer et al. (13) (the procedure was modified by the addition of lysozyme to aid cell lysis), and the aliquots were resuspended in diethyl pyrocarbonate-treated water (17). RNA samples were treated with DNase and with RNase inhibitor (Roche Diagnostics, Lewes, East Sussex, United Kingdom), denatured at 70°C for 10 min, and loaded with formamide-containing dye onto a 1.2% formaldehyde gel (1). RNA size standards from Promega (Madison, Wis.) were included to enable estimation of the sizes of the observed transcripts. Capillary blotting to Hybond-N⁺ nylon membranes (Amersham, Little Chalfont, Buckinghamshire, United Kingdom) was performed as described by Sambrook et al. (17). PCR products, used as probes, were labeled with ³²P by using a Prime-a-Gene kit (Promega). Overnight hybridization was done in a 0.5 M sodium phosphate (pH 7.0)–5% sodium dodecyl sulfate (SDS) buffer at 50°C, after which the blots were washed with 2× SSC–0.1% SDS and then with 0.2× SSC–0.1% SDS and 0.1× SSC–0.1% SDS, at temperatures ranging from 55 to 65°C depending on the required stringency (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate).

A blot containing total RNA from L. lactis subsp. cremoris UC503 was probed with DNA corresponding to each of the ORFs on the locus. In this manner it was deduced that two transcripts were produced. scrFIAM was transcribed individually on a 1.3-kb transcript, while orfX, scrFIBM, and scrFIR were transcribed together on a single 2.6-kb polycistronic mRNA molecule (Fig. 2).

FIG. 2.

Northern analysis of the ScrFI locus using RNA from L. lactis UC503. The probes used were orfX (lane orfX), scrFIBM (lane BM), scrFIR (lane R), and scrFIAM (lane AM). The sizes of the transcripts are indicated on the left and right.

5′ extension analysis was also performed using primers positioned 100 bp downstream of each of the putative promoters. This analysis indicated that the transcriptional start site for the promoter just upstream of scrFIAM was a thymine 4 bp downstream of the −10 motif (Fig. 3). The transcriptional start site for the orfX promoter was a thymine that is more atypically located 24 bp downstream of the −10 motif proximal to the start codon (Fig. 4). The increased distance may be attributed to the stem-loop structure located between the −10 motif and the position determined for the start site as RNA hairpins can cause premature termination by reverse transcriptase (25). The smearing directly upstream of the position of the stem-loop structure could represent transcripts initiating at adjacent positions; these positions are located a more typical distance from the −10 hexamer.

FIG. 3.

(a) 5′ extension analysis of the promoter region upstream of scrFIAM of the ScrFI locus of L. lactis subsp. cremoris UC503. The arrow indicates the position of the extension product. (b) Sequence of an extended region between scrFIR and scrFIAM, showing the regulatory signals. The transcriptional start site is indicated by a solid triangle. RBS, ribosome-binding site.

FIG. 4.

(a) 5′ extension analysis of the orfX promoter region of the ScrFI locus of L. lactis subsp. cremoris UC503. The arrow indicates the position of the extension product. A deviation from the correct sequence published by Twomey et al. (28) is indicated by the underlined base. (b) Sequence of an extended region upstream of orfX, showing the regulatory signals. The transitional start site is indicated by a solid triangle. A direct repeat upstream of the −35 hexamer is indicated by boldface type. RBS, ribosome-binding site.

No hybridization to the 2.6-kb transcript was observed with the scrFIAM probe, and it was therefore assumed that transcription of this mRNA molecule terminated within the intergenic region between scrFIR and scrFIAM. Termination was possibly via a Rho-dependent event since no stem-loop structures with strong ΔG values were identified within this region. The possibility that the 1.3-kb RNA molecule is a processed transcript arising from cleavage of a larger RNA molecule originating at the orfX promoter cannot be discounted. Nevertheless, assuming that the locus is transcribed as two separate mRNA molecules, this result provides for independent regulation of both the scrFIR and scrFIAM genes.

M.ScrFIA binding assays.

To date, the mechanisms by which expression of type II R/M systems is controlled fall into two distinct groups, and it has been proposed that these systems may actually be characterized in accordance with their mechanisms of regulation (10, 11). Two of the best-studied examples, the EcoRII and BamHI systems, illustrate the two modes of regulation. The BamHI system is regulated by the product of a third ORF found close to the normal restriction and modification genes. The encoded protein, designated C.BamHI, was found to control expression of the other two genes in the locus (18, 19). It was determined that C.BamHI acts as a transcriptional regulator by way of an H-T-H motif, a feature common to all C proteins identified to date (10, 19). The second transcriptional regulatory mechanism which has been reported is typified by the EcoRII system (5). In this case an H-T-H motif identified at the N terminus of M.EcoRII has been shown to bind specifically to the promoter region between the methylase and endonuclease genes and to subsequently modulate expression of these genes (20).

The presence of both a small ORF, which could potentially code for a protein equivalent to a C protein, and a highly probable H-T-H motif (90% possibility of being a functional DNA-binding domain according to the algorithm of Dodd and Egan [3]) at the N terminus of M.ScrFIA suggested the possibility that both mechanisms of control of the ScrFI locus are present. The autoregulatory promoter-binding methylases M.MspI (21) and, more significantly, M.SsoII (11, 12), which shows up to 70% similarity with scrFIAM, also have similarly positioned H-T-H motifs. To establish if the M.ScrFIA H-T-H had a DNA-binding function, the protein was overexpressed and purified by using the QIAexpress system (Qiagen Ltd., Surrey, United Kingdom). The intragenic region between scrFIR and scrFIAM was amplified, and the 300-bp product was labeled with polynucleotide kinase by using [γ-³²P]ATP (Amersham) and used in retardation assays with the purified M.ScrFIA protein. DNA-binding assays were performed in 20-μl reaction mixtures containing 50 mM Tris (pH 8.0), 10% (vol/vol) glycerol, 1 mM EDTA, 5 mM MgCl₂, 500 mM KCl, 2 mM dithiothreitol, 50 μg of bovine serum albumin/ml, 75 μg of poly(dI-dC)/ml, probe (approximately 0.3 ng), and up to 4 pmol of purified M.ScrFIA. Incubation was for 15 min at room temperature; this was followed by addition of 5 μl of 50% glycerol, and then the samples were loaded onto a 4% polyacrylamide gel containing 2.5% glycerol. Gels were electrophoresed in TAE buffer (0.04 M Tris-acetate [pH 7.5], 2 mM EDTA) at 120 V for 3 h, dried, and exposed overnight at −70°C to X-Omat film (Kodak, Rochester, N.Y.).

From these assays (Fig. 5) it can be deduced that M.ScrFIA does bind to the intragenic region upstream of scrFIAM, which contains the promoter motif. Titration of the quantity of protein added to the binding assays with respect to the DNA quantity present revealed up to four DNA-methylase complexes, which probably indicates that four M.ScrFIA molecules bound to the DNA segment (three bands are visible on the autoradiogram in Fig. 5; overexposure of the retardation gel revealed an additional fourth bound band). This implies that more than one molecule of the methylase binds to the promoter region, which correlates with an analogous observation for M.SsoII, for which a similar band pattern was observed (11). In order to determine more precisely the location of the M.ScrFIA binding site(s) within the scrFIAM promoter region, a series of subfragments was generated by PCR (Fig. 6). The results of binding assays in which these subfragments were used indicated that the M.ScrFIA binding site is located on a 30-bp segment (Fig. 6). This region contains a putative symmetrical target sequence.

FIG. 5.

Gel retardation assay results, showing binding between the M.ScrFIA protein and the scrFIAM promoter region. A 0.3-ng portion of ³²P-labeled fragment harboring scrFIAM promoter DNA (300-bp region) was used as the probe. Lanes 1 to 7, 0, 0.1, 0.2, 0.5, 1.0, 2.0, and 4.0 pmol of M.ScrFIA, respectively; lane 8, negative control containing a ³²P-labeled 300-bp lactococcal DNA fragment from the origin of replication region of pCI372 plus 4 pmol of M.ScrFIA.

FIG. 6.

Schematic diagram of the scrFIAM promoter region. The transcriptional signals of the promoter are indicated. A to G represent the ³²P-labeled PCR-generated fragments used in the gel retardation assays. The positions of these fragments as found in the GenBank sequence are indicated.

When this assay was repeated by using the same purified protein but DNA from the intergenic region upstream of orfX (including the promoter), no binding was observed despite the fact that numerous binding reactions were performed under various binding conditions. The intergenic region upstream of orfX does not contain a similar putative symmetrical target sequence, although a 10-nucleotide direct repeat is present upstream of the −35 motif (Fig. 4b). Nevertheless, it is still possible that M.ScrFIA is involved in regulating this promoter but that additional external factors not present in the assay mixtures are required for binding to occur in vitro. These results imply that M.ScrFIA plays a role in regulation of the promoter upstream of scrFIAM only.

Analysis of promoter strength.

Two promoters were identified on the ScrFI locus by comparison with other lactococcal promoter sequences and also by transcriptional analysis. In order to determine the strength of the promoter upstream of scrFIAM, a 300-bp fragment containing this promoter region was cloned into the vector pAK80 immediately upstream of a promoterless lacZ gene. This fragment included the entire intergenic region between scrFIR and scrFIAM, and the resultant construct was designated pCI921. Introduction of pCI921 into MG1363 resulted in a blue colony phenotype indicative of β-galactosidase (β-Gal) activity, confirming that the intergenic region contains an active promoter. The resultant culture was designated L. lactis subsp. cremoris DB002. β-Gal activity was measured as described by Miller (14), with the following modifications. Cultures were grown in M17 and harvested in the mid-log phase. Cells were harvested by centrifugation and concentrated up to 10-fold in Z buffer. One-half milliliter of bacterial suspension was mixed with 12.5 μl of 0.1% SDS and 25 μl of chloroform in a vortex mixer for 10 s. After 5 min of incubation in a 30°C water bath, 100 μl of o-nitrophenyl-β-d-galactopyranoside (4 mg/ml of A medium [14]) was added, and the suspension was vortexed for 2 s and incubated further at 30°C. Reactions were stopped by adding 250 μl of 1 M sodium carbonate. After centrifugation, A₄₂₀ and A₅₅₀ values were measured for the supernatant. If the A₅₅₀ exceeded 0.050, the sample was centrifuged again and remeasured. β-Gal activity (in Miller units) was calculated as follows: (522 × A₄₂₀)/(t × v × OD₆₀₀), where t is time (in minutes), v is the volume of culture used in the assay (in milliliters), and OD₆₀₀ is the optical density of the culture at 600 nm. The β-Gal assays gave an activity of 138.5 Miller units for this promoter (Table 2).

TABLE 2.

β-Gal activities of constructs

Construct	β-Gal activity (Miller units)
	DB002	DB003
MG1363 (pAK80)	3	3
pCI372	138.5	132
pCI934	15	20
pCI941	20	23
pCI945	130	138
pCI923	20	30
pCI925	146	130
pCI926	16.6	26

To determine the strength of the promoter identified just upstream of orfX, a 252-bp fragment containing this promoter region was introduced into pAK80 to create a construct designated pCI922. Introduction of pCI922 into MG1363 resulted in a culture which was designated L. lactis subsp. cremoris DB003. β-Gal assays performed with MG1363 containing pCI922 gave an activity of 132 Miller units (Table 2).

Investigation of the in trans effect of segments of the ScrFI locus on promoter activity.

To further delineate the extent of the involvement of the ORFs present in the ScrFI locus in regulation of the scrFIAM promoter, the construct pCI934 (containing the entire ScrFI locus) was transformed into DB002, which resulted in a significant loss of β-Gal activity (the activity decreased from 138.5 to 20 Miller units) (Table 2). Subclones of the ScrFI locus were subsequently transformed into DB002 to determine the smallest fragment of the locus sufficient to decrease the β-Gal activity. A construct designated pCI925 containing the scrFIAM promoter region and the scrFIAM gene with an internal deletion which removed part of the H-T-H motif was found to have no effect on the β-Gal activity, while a similar construct with the complete scrFIAM gene, pCI926, gave decreased β-Gal activity (Table 2). This is consistent with the gel retardation assay results which indicated that there is a binding interaction between M.ScrFIA and the promoter region preceding scrFIAM.

Introduction of pCI934 into DB003 (which harbors pCI922 [i.e., the orfX promoter cloned in pAK80]) resulted in downregulation of the promoter's activity (the activity decreased from 132 to 30 Miller units) (Table 2). Interestingly, introduction of pCI945 (containing orfX and scrFIBM) into DB003 resulted in no change in β-Gal activity, indicating that orfX itself does not have any noticeable effect on the promoter (Table 2). This result implies that the product of orfX is not involved in regulating transcription of the ScrFI locus. As with DB002, any construct containing a complete scrFIAM gene resulted in a decrease in β-Gal activity.

Transformation of pCI921 and pCI922 into L. lactis subsp. cremoris UC503 was performed, and transformants containing each of these constructs were assayed for β-Gal activity. A decrease in β-Gal activity was observed when either pCI921 or pCI922 was introduced into UC503; the β-Gal activity level decreased to about 80% of that observed when either construct was in a MG1363 background.

Somewhat unexpectedly considering the results of the binding experiments in which M.ScrFIA downregulated only its own promoter region, it was found in these experiments that the protein downregulated expression of both promoters. The reason for this apparent discrepancy is not obvious. As indicated above, it is possible that the failure to show M.ScrFIA binding to the orfX promoter may have been due to some deficiency in the assay conditions and that M.ScrFIA requires some external factor not present under the reaction conditions to bind the orfX promoter. However, it is also noteworthy that the β-Gal assays were performed with M.ScrFIA expressed from a multicopy gene cloned on the vector pCI372. Thus, the concentration of this protein within the cell was significantly higher than would be the case in the wild-type situation. Therefore, this may have favored a level of interaction between M.ScrFIA and the orfX promoter that was not truly reflective of the situation in the wild-type cell. The results of the β-Gal assays performed when pCI921 and pCI922 were introduced into a UC503 background support this theory. A single copy of scrFIAM is present in the UC503 genome, which probably results in a lower concentration of M.ScrFIA than would be the case if multiple copies of the gene were present. The less dramatic drop in β-Gal activity (only 20%) observed may be attributed to the lower concentration of M.ScrFIA.

Conclusion.

In conclusion, transcriptional analysis revealed that orfX, scrFIBM, and scrFIR are cotranscribed as a single 2.6-kb polygenic mRNA molecule, while scrFIAM is transcribed independently on a 1.3-kb transcript. In addition, the results obtained with respect to the promoter upstream of orfX are ambiguous as M.ScrFIA does appear to downregulate its activity but cannot be shown to interact directly with the promoter in binding assays. However, binding of M.ScrFIA to the promoter region upstream of scrFIAM and the subsequent downregulation of promoter activity show that M.ScrFIA is similar to M.EcoRII-like methylases from a regulatory viewpoint.