ABSTRACT
The regulation of the SsoII restriction-modification system from Shigella sonnei was studied in vivo and in vitro. In lacZ fusion experiments, SsoII methyltransferase (M.SsoII) was found to repress its own synthesis but stimulate expression of the cognate restriction endonuclease (ENase). The N-terminal 72 amino acids of M.SsoII, predicted to form a helix-turn-helix (HTH) motif, was found to be responsible for the specific DNA-binding and regulatory function of M.SsoII. Similar HTH motifs are predicted in the N-terminus of a number of 5-methylcytosine methyltransferases, particularly M.EcoRII, M.dcm and M.MspI, of which the ability to regulate autogenously has been proposed. In vitro, the binding of M.SsoII to its target DNA was investigated using a mobility shift assay. M.SsoII forms a specific and stable complex with a 140 bp DNA fragment containing the promoter region of SsoII R-M system. The dissociation constant (Kd) was determined to be 1.5*10-8 M. DNaseI footprinting experiments demonstrated that M.SsoII protects a 48-52 bp region immediately upstream of the M.SsoII coding sequence which includes the predicted -10 promoter sequence of M.SsoII and the -10 and -35 sequences of R.SsoII.
The type II restriction-modification (R-M) system consists of two enzymatic components, restriction endonuclease (ENase) and methyltransferase (MTase). The ENase makes a double-strand DNA break within or near the specific recognition sequence. The cognate MTase modifies a specific base within the recognition sequence to prevent it from cleavage. Many R-M systems, e.g. EcoRII (1 ,2 ) and SsoII (3 ), are carried on self-transmissible plasmids. In the host cells, it is necessary that methylation precedes the endonuclease action. Therefore, the initial level of MTase expression should be higher compared to that of the ENase. Subsequently, the level of MTase is expected to be reduced in order to provide effective protection of the host cells against bacteriophage infection.
To date, many R-M systems have been cloned, sequenced and expressed in heterologous systems, but, in general, there has been a paucity of information on their regulation. In PvuII and BamHI R-M systems, regulation is provided by a small protein, the encoding gene being found within the respective R-M intergenic space (4 ,5 ). In EcoRII, expression of M.EcoRII is autoregulated at the transcriptional level (6 ). A similar mechanism was proposed for M.MspI (7 ). Here, we examine the regulation of the SsoII R-M system in vivo and in vitro. In addition, we propose a common regulatory mechanism for a number of related 5-methylcytosine (m5C) MTases.
Plasmid pMS2 carrying the total nucleotide sequence of SsoII R-M system (8 ) was used as a template for PCR synthesis of a 140 bp DNA fragment containing the intergenic region between the ssoIIM and ssoIIR genes. The nucleotide sequences of the forward and reverse primers were:
5'-CTTAA
BamHI
and
5'-CTGC
EcoRV BglII
respectively.
CAT in boldface is the complement of the translational start codons of R.SsoII (upper primer) and M.SsoII (lower primer), respectively. The BglII, EcoRV and BamHI sites were introduced in the primer sequences. These oligonucleotides were synthesized on an ASM 102U DNA synthesizer (Novosibirsk, Russia) using standard phosphoramidite chemistry. The conditions of PCR reactions were as follows. The first five initial cycles: 94oC for 1 min, 55oC for 1 min, 72oC for 40 s; and the following 25 cycles: 94oC for 1 min, 60oC for 1 min, 72oC for 40 s. The amplified DNA was subcloned into the pACYC184 plasmid (9 ) in both orientations. In one case, both pACYC184 plasmid and PCR fragment were treated with BamHI and EcoRV; in the other, the pACYC184 DNA was cleaved with BamHI and the PCR fragment treated with BglII and BamHI. The ligated DNA was transformed in Escherichia coli NW-22 {[Delta](lac-proAB) [F' tra[Delta]36 proA+B+ laqIq lacZ[Delta]M15] endA1 gyrA96 (NalR) hsdR2(or17) mcrA mcrB-1 mrr? relA1 supE44 thi-1}. Colonies with CmR and TcS phenotype were selected. The resulting plasmids were then treated with BglII and HindIII or BamHI and HindIII and ligated with BamHI-HindIII fragment from p10 plasmid, a derivative of pCL47 (10 ) containing the full-length lacZ gene, to generate pACYC-SsoIIM and pACYC-SsoIIR plasmids, respectively (Fig. 2 ). Thus, the first ATG codon of ssoIIM or ssoIIR gene appeared to be the start codon of lacZ gene in resulting pACYC-SsoIIM and pACYC-SsoIIR constructs, respectively. Blue colonies were selected on Luria-Bertani (LB) plates containing X-gal and IPTG (isopropyl-[beta]-thiogalactopyranoside) after transformation in E.coli NW-22.
The plasmid pMS2 (8 ) bearing the native M.SsoII gene (Fig. 2 ), and plasmids pQMSsoII, pQMNlaX, pQMSsoNlaX and pQMDSsoII carrying the genes of M.SsoII, M.NlaX, hybrid SsoII/NlaX MTase and a deletion variant of M.SsoII, respectively (11 ), were co-transformed in E.coli NW-22 cells harboring the pACYC-SsoIIM and pACYC-SsoIIR plasmids. In the listed pQM plasmids the methyltransferase genes are under the control of the artificial hybrid lac/T5 promoter (11 ), while in the case of pMS2 the ssoIIM gene is transcribed from its own M.SsoII promoter (8 ). [beta]-Galactosidase activity in these strains was calculated as described by Miller (12 ) or analysed on LB plates containing X-gal and IPTG.
Plasmid pB6490 (13 ) was used as a template to synthesize a 140 bp DNA fragment which contains the promoter region of Bovine Leucosis Virus (BLV) as a non-specific DNA fragment for mobility shift assay. The primers for PCR synthesis are: 5'-CTCTATCTCCGGTCCTCTG-3' and 5'-GAAGGAGAGAGCGCGGGCC-3'. This DNA fragment as well as the 140 bp specific DNA described above were labeled with T4 polynucleotide kinase (Fermentas, Lithuania) and [[gamma]-32P]ATP. After separation in 8% polyacrylamide gel, these fragments were purified by electroelution.
M.SsoII and M.NlaX containing the 6*His tag were purified using a two-step procedure involving heparin-Sepharose and Ni-NTA-agarose chromatography (11 ). Complexing of M.SsoII or M.NlaX with its target DNA was carried out in the buffer containing 50 mM Tris-HCl pH 7.5, 5 mM [beta]-mercaptoethanol, 150 mM NaCl and 8% glycerol in a 10-20 [mu]l reaction volume, for 10-20 min at room temperature. For the determination of equilibrium binding constant, Kd, 80 nM of M.SsoII and the 140 bp 32P-labeled DNA fragment in the concentration range from 0 to 12.5 nM were incubated under the above conditions in a 10 [mu]l volume. The free and bound DNA fractions were separated on a 6% PAGE (acrylamide:bisacrylamide = 19:1 w/v, 50 mM Tris-borate EDTA) and visualized by autoradiography.
Figure
The 140 bp specific PCR fragment was labeled at either the top or bottom strand. The binding conditions were the same as described above, except that the binding buffer contained 10 [mu]g/ml poly (dI-dC)[middot]poly (dI-dC). Fifty ng of DNA (100 000 c.p.m.) and 40 ng of M.SsoII were mixed together in binding buffer in a 20 [mu]l volume and incubated at 30oC for 30 min. Then, 10 ng of DNaseI (freshly diluted from 2.5 mg/ml stock in 50 mM Tris-HCl pH 8.0, 50 mM NaCl, 60 mM MgCl2, 15 mM CaCl2 and 10% glycerol) in a 4 [mu]l volume was added to the binding reaction. Digestion was terminated after 15 min incubation at room temperature by adding 170 [mu]l of the stop solution (200 mM NaCl, 20 mM Na2EDTA, 1% SDS and 100 [mu]g/ml tRNA). Control reactions were performed under the same conditions except for the addition of M.SsoII. The DNA was then extracted by phenol/chloroform and precipitated. The pellets were redissolved in the formamide loading buffer, heated to 90oC for 5 min and analyzed on 6% PAGE. The chemical sequencing reactions were performed as described previously (14 ).
The SsoII ENase and MTase encoding genes are divergently transcribed (8 ). The length of the intergenic space is 109 bp. To investigate the effect of M.SsoII on ssoIIM and ssoIIR gene expression, the 109 bp promoter fragment was placed in either orientation in front of the lacZ gene in plasmids pACYC-SsoIIM and pACYC-SsoIIR (Fig. 1 A). Thus, the lacZ gene is under the control of the M.SsoII promoter and the R.SsoII promoter, respectively. The activity of [beta]-galactosidase produced in the E.coli cells harboring either pACYC-SsoIIM or pACYC-SsoIIR plasmid was measured in the presence or absence of M.SsoII or its derivatives provided in trans (Table 1 ). As a result, the level of [beta]-galactosidase activity in bacterial cells harboring the pACYC-SsoIIM plasmid was found to be 540-fold higher than those harboring the pACYC-SsoIIR plasmid (Table 1 ). Such a situation probably occurs just after the R-M-bearing plasmid has been introduced into the host strain. Initially, an excessive level of MTase synthesis relative to a low level of ENase would enable the host DNA to be modified thus preventing digestion by the cognate restriction endonuclease. Upon introduction of pMS2 plasmid, which provides a constitutive M.SsoII synthesis, the level of [beta]-galactosidase increased 8-fold in the cells harboring pACYC-SsoIIR plasmid and decreased 20-fold in the cells with pACYC-SsoIIM plasmid (Table 1 ). This likely reflects a further stage of the bacterial cell cycle, i.e., having the DNA methylated and yet requiring a high level of the ENase for providing effective protection of the host cell against bacteriophage infection. Here, M.SsoII is viewed to act as a regulatory protein, it binds to the promoter region suppressing the M.SsoII synthesis while enhancing the R.SsoII production.
Intriguing results were obtained when the plasmid pQMNlaX was introduced in the cells harboring either pACYC-SsoIIM or pACYC-SsoIIR plasmid. pQMNlaX produces M.NlaX which is closely related in amino acid sequence to that of M.SsoII (11 ) and recognizes the same DNA sequence (A.K., unpublished data). Contrary to M.SsoII, expression of M.NlaX does not appear to have an appreciable effect on [beta]-galactosidase synthesis in either cell (Table 1 ). The M.NlaX lacks the N-terminal amino acid extension in comparison with M.SsoII although the remaining sequence has a high amino acid sequence identity (11 ). Thus, we deduce that the N-terminal portion of M.SsoII is responsible for promoter binding and forms the basis for SsoII R-M system regulation. Supporting evidence comes from the plasmids providing the synthesis of hybrid SsoII/NlaX MTase (consisting of the first 72 amino acids of the M.SsoII and the entire M.NlaX) and a deletion derivative of M.SsoII in which the 72 residue extraneous sequence has been deleted ([Delta]SsoII MTase) (Fig. 1 B). Expression of M.SsoII/NlaX had a similar effect as the native M.SsoII, while neither M.NlaX nor M.[Delta] SsoII affected the activity of [beta]-galactosidase.
The DNA-binding property of M.SsoII or M.NlaX was analyzed using gel-electrophoresis mobility shift assay. The concentrations of purified M.SsoII and M.NlaX refer to the monomeric forms of Mr 45 000 and 38 000, respectively. The specific and non-specific 140 bp DNA fragments represent the intergenic region of the SsoII R-M system and the promoter region of BLV, respectively. There is no sequence similarity between the two DNAs and the M.SsoII recognition sequence is absent. In concentrations of M.SsoII, from 0.13 to 0.78 mM, the main shifted band is referred to as the C2 complex (Fig. 2 A). When the concentration of DNA was higher than 1.3 mM, more than six discrete DNA bands were visualized on the autoradiograms.
Table 1
Formation of the specific C2 complex is not affected by the presence of even a 96-fold excess of competitive DNA (Fig. 3 ). The patterns of M.NlaX/specific DNA binding (Fig. 2 B) as well as M.SsoII/non-specific DNA binding (data not shown) appeared to be very similar and typical for non-specific DNA-protein interaction (15 ). In both cases, the quantity of shifted bands is a function of the added protein concentration. Moreover, the protein concentration required for multiple band formation is only slightly higher than that needed to produce the first shifted band.
To determine the constant of dissociation (Kd) of the M.SsoII- 140 bp DNA complex the dependence of the ratio of bound to free DNA ([ED]/[D]) from DNA concentration (D0) at a constant enzyme concentration (E0) was studied (Fig. 4 ). Assuming that the active enzyme specifically interacts with DNA substrate at a 1:1 ratio, then
Figure
Figure
Figure
Kd = [E][D]/[ED],
where [E], [D] and [ED] are the equilibrium concentrations of M.SsoII, DNA and DNA-protein complex, respectively. For a Scatchard plot analysis this equation has to be rearranged into
[ED]/[D] = E0/Kd - [ED]/Kd
The magnitude of E0 is determined as the length of segment cutting off on the abscissa axis by the straight line which represents the dependence of [ED]/[D] from [ED], while the Kd is determined from a tangent of the angle between the above line and the abscissa axis (1/Kd) as shown in Figure 4 A. Hence, the value of Kd was calculated as 15 +- 7 nM and E0 was determined as 4.4 +- 0.6 nM. Taking into account the dilution factor it is possible to calculate E0 in the enzyme sample being investigated. The value of E0 is equal to 285 +- 40 nM instead of 5300 nM which was obtained using the Bradford technique (16 ). To calculate the active concentration of M.SsoII (E0) more accurately the competitive binding between 32P-labeled and unlabeled specific 140 bp DNA fragment to M.SsoII was employed (Fig. 3 ). The dependence of the ratio of bound to free labeled DNA (ED*/D*) on adding unlabeled DNA (D0) concentration was analyzed. This process is described by the following schemes:
E + D* <-> *ED*
E + D <-> *ED,
where E is an active form of enzyme, D* and D represent labeled and unlabeled forms of the same 140 bp DNA fragment, respectively, and ED* and ED represent the complexes of the protein with labeled and unlabeled DNA fragments, respectively. In this experiment the concentration of labeled specific DNA fragment (D*0) was 25 nM. The following equations represent the above scheme:
D*0 = [ED*]+[D*]
D0 = [ED]+[D]
E0 = [E]+[ED]+[ED*]
Kd = [E][D*]/[ED*]
Kd = [E][D]/[ED]
Hence, [D] + [D*] = E0[D*]/[ED*] - Kd.
The concentration of the active form of enzyme (E0) can be determined as a tangent of the angle between the straight line representing the dependence of [D] + [D*] versus [D*]/[ED*] and the abscissa axis (Fig. 4 B). The value of E0 appeared to be 13 +- 1 nM. Considering the dilution factor in this experiment the value of E0 is equal to 256 +- 16 nM which approximates to the 285 +- 40 nM value obtained above. The calculated active concentration (E0) of M.SsoII is 20-fold less than the concentration determined experimentally. This may be explained either by the loss of enzyme activity during storage or by the fact that the active form of M.SsoII is dimeric or oligomeric protein. It should be noted that in this concentration range the DNA-protein complex dissociation constant Kd cannot be determined using the above method because the Kd value thus obtained is comparable with the error of the experiment.
A 140 bp DNA fragment containing the promoter region of M.SsoII was end-labeled on either the top or bottom strand, bound with a saturated amount of M.SsoII, and the resulting complex was treated with DNaseI. A control reaction was performed in the same manner except for the addition of MTase. The resulting DNA fragments were separated by electrophoresis and the positions protected by M.SsoII from DNaseI digestion were determined. Figure 5 A shows that M.SsoII protects a specific region of 48 and 52 bp in top and bottom strands, respectively. The protected region includes a portion of the putative -10 region and RBS of ssoIIM gene and both putative -10 and -35 regions of ssoIIR gene (Fig. 5 B).
All m5C-MTases have a common architecture, 10 conserved amino acid blocks, referred to as motifs I-X (17 ) are typically found within the polypeptide. Motif I forms part of the binding site for AdoMet; block IV contains a conserved Cys which plays a key role in the methylation reaction; the variable region between motifs VIII and IX is responsible for the recognition of the target DNA sequence. Some m5C-MTases have an extended N- terminal portion preceding the conserved motif I. Som and Friedman (6 ) showed that the N-terminal portion of M.EcoRIIbinds to the promoter of the ecoRIIM gene in a region containing an inverted repeat.
In an earlier study, we analyzed the N-terminal sequence of two related m5C-MTases, M.SsoII and M.ScrFI which recognize CCNGG and found the presence of a typical prokaryotic DNA binding structure known as a helix-turn-helix (HTH) motif (11 ). Figure 6 A shows that similar HTH motifs are predicted, the majority with high probability, in the N-terminal portions of eight out of the 10 m5C-MTases which possess an extended N-terminal sequence. It appears that these m5C-MTases can be divided into two groups, one recognizing the pentanucleotide DNA sequences, CC(N/W)GG (where W = A or T and N = A, C, G or T), except for the M.MspI which recognizes CCGG, the other recognizes the complement, GG(N/W)CC except for the ambiguous base in the middle of the recognition sequence. In all these systems, an inverted repeat sequence of variable length (Fig. 6 B), presumably important for protein binding, can be found in front of the respective methyltransferase sequence. These suggest a common evolution of these R-M systems and a similar way of regulation. The possible mechanism of the regulation is based on the specific interaction of the MTase with its promoter, which provides the repression of the MTase synthesis during functioning of the R-M system in the cell.
This work was done partially in frames of the RFFI-DFG project 96-04-00130. We thank Olga Sergienko and Dr Sekikawa for providing p10 and pB6490 plasmids, respectively; our colleagues from Prof. Trautner's department of Max Plank Institute of Molecular Genetics, Dahlem, Berlin for supplying E.coli NW22 strain and pACYC184 plasmid; Joan Brooks for the DNase I footprinting protocol and Dmitry Shayakhmetov for help in preparation of the figures.
Host strain
Cotransformed plasmids
[beta]-Gal activitya
Colour of colonies on
plasmids
in Miller units
(%)
the X-gal containing plates
Escherichia coli NW-22
-
269
(100)
dark blue
[pACYC-SsoIIM]
pMS2
13.5
(5)
blue
pQMSsoII
130.7
(49)
dark blue
pQMNlaX
288.6
(107)
dark blue
pQMSsoNlaX
232.7
(86)
dark blue
pQM[Delta]SsoII
254.2
(95)
dark blue
Escherichia coli NW-22
-
0.5
(100)
white
[pACYC-SsoIIR]
pMS2
4.1
(820)
blue
pQMSsoII
10.8
(2160)
blue
pQMNlaX
1.2
(240)
white
pQMSsoNlaX
12.2
(2440)
blue
pQM[Delta]SsoII
0.61
(122)
white
REFERENCES
