ABSTRACT
To create new, effective reagents for affinity modification of restriction-modification (R-M) enzymes, a regioselective method for reactive dialdehyde group incorporation into oligonucleotides, based on insertion of a 1-[beta]-D-galactopyranosylthymine residue, has been developed. We synthesized DNA duplex analogs of the substrates of the EcoRII and MvaI R-M enzymes that contained a galactose or periodate-oxidized galactose residue as single substituents either in the center of the EcoRII (MvaI) recognition site or in the flanking nucleotide sequence. The dependence of binding, cleavage and methylation of these substrate analogs on the modified sugar location in the duplex was determined. Cross-linking of the reagents to the enzymes under different conditions was examined. M.EcoRII covalent attachment to periodate-oxidized substrate analogs proceeded in a specific way and to a large extent depended on the location of the reactive dialdehyde group in the substrate. The yield of covalent attachment to a DNA duplex with a dialdehyde group in the flanking sequence with EcoRII or MvaI methylases was 9-20% and did not exceed 4% for R.EcoRII.
Type II restriction-modification (R-M) enzymes are attractive models for studying specific DNA-protein recognition. Recently the structures of several complexes of restriction endonucleases and DNA methyltransferases (Mtases) with substrates were solved and DNA-protein contacts were determined (1 -3 ).
Affinity modification of the enzymes by modified substrates followed by proteolysis of a cross-linked product is an alternative promising way to determine the amino acid residues responsible both for specific recognition of the target DNA and for catalytic action. Previously, 5-bromouridine-containing substrates were photocross-linked to EcoRI and EcoRV restriction endonucleases (4 ). Thio analogs of thymidine and deoxyguanosine were introduced into synthetic substrates of the EcoRI and EcoRV R-M enzymes and these DNA duplexes were converted to photoaffinity reagents (5 ). DNA duplexes with an active monosubstituted pyrophosphate bond were used for affinity modification of several restriction endonucleases (6 -8 ) and two DNA methyltransferases (6 ). However, these reagents are supposed to interact with any nucleophilic group of the proteins depending on its proximity to the modified internucleotide bond and formation of an unstable covalent enzyme-DNA bond is likely.
Periodate-oxidized nucleotides have been used extensively for affinity modification of enzymes involved in nucleic acid metabolism (9 -10 ). This labeling technique involves formation of an unstable dihydroxymorpholine derivative, which may be reduced to stable morpholine compounds. Using a similar methodology and starting from 3'-modified oligonucleotides (ONs), conjugates with poly(lysine) were prepared (11 ).
We describe here a novel type of cross-linking reagent for R-M enzymes, DNA duplexes with reactive dialdehyde groups introduced at single defined positions of the ON strand. The advantages of such reagents are their specificity, mainly toward proximal lysine residues, and their ability to form stable covalent complexes upon reduction of the cross-linked products. To create such reagents, we developed a methodology for regioselective modification of ONs that enables introduction of reactive groups into the middle of an ON chain. Previously the reactive groups were introduced only at the 5'- (12 ) or 3'-ends (13 ). Introduction of additional cis diol groups may be achieved in several different ways, the simplest based on insertion of hexopyranosylnucleoside residues into an ON chain. These analogs have four hydroxyl groups, two of which may be used for incorporation into the ON chain and the other two for further modification. It should be noted that several ONs containing 2',3'-dideoxyhexopyranosylnucleosides have been prepared (14 -16 ). It was shown that an ON with one or two such residues may form a stable duplex (14 ,16 ). It is noteworthy that such DNA duplexes with dialdehyde groups in the middle of the ON strand were not used for covalent attachment to DNA binding proteins. Ebralidse et al. (17 ) have reported mildly depurinated DNA cross-linking to histones.
In this work efficient synthesis of DNA duplexes with reactive dialdehyde groups is developed, the properties of such duplexes are described and cross-linking to EcoRII and MvaI R-M enzymes is investigated.
Liquid secondary ion mass spectra (LSIMS) were obtained using a Kratos Concept 1H mass spectrometer. Column chromatography was performed on silica gel (0.06-0.20 mm), TLC was carried out on Kieselgel 260 F (Merck) with detection by UV light using: A, CHCl3; B, 98:2 CHCl3:EtOH; C, 95:5 CHCl3:EtOH. NMR spectra were recorded on Gemini 200 NMR and Bruker AMX 400 spectrometers at 22oC. Chemical shifts were measured relative to the solvent signals.
Buffers: A, 40 mM Tris-HCl, pH 7.0, 15 mM MgCl2; B, 1 M aniline acetate, pH 4.5; C, 40 mM Tris-HCl, pH 7.6, 50 mM NaCl, 7 mM dithiothreitol (DTT); D, 40 mM Tris-HCl, pH 7.6, 50 mM NaCl, 5 mM MgCl2, 7 mM DTT; E, 40 mM Tris-HCl, pH 7.9, 5 mM DTT, 1 mM EDTA; F, 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 15 mM MgCl2, 0.1 mM DTT; G, 10 mM Tris-HCl, pH 8.5, 150 mM NaCl, 1 mM DTT, 100 [mu]g/ml BSA; H, 50 mM Tris-HCl, pH 9.0, 20 mM NaCl, 1 mM DTT, 100 [mu]g/ml BSA.
Restriction endonuclease MvaI (44 [mu]g/ml, 25 U/[mu]l), DNA methyltransferase MvaI (450 [mu]g/ml, 20 U/[mu]l) and T4 polynucleotide kinase were purchased from MBI Fermentas (Lithuania). Restriction endonuclease EcoRII (640 [mu]g/ml, 80 U/[mu]l) was purified by N.I.Matvienko (Institute of Protein, Russian Academy of Sciences). DNA methyltransferase EcoRII (840 [mu]g/ml, 5-8 U/[mu]l) was purified according to Buryanov et al. (18 ). One unit of restriction endonuclease activity was defined as the amount of enzyme that completely digested 1 [mu]g [lambda] DNA at 37oC in 1 h. One unit of Mtase activity was the amount of enzyme required to protect 1 [mu]g [lambda] DNA from cleavage by a cognate endonuclease at 37oC for 1 h.
A suspension of dry thymine (9.45 g, 75 mmol) in hexamethyldisilazane (50 ml) and ammonium sulfate (0.5 g) was boiled under reflux in the absence of moisture until complete dissolution (10 h). The mixture was concentrated in vacuo to dryness and dry 1,2-dichloroethane (30 ml) was evaporated from the residue. A solution of 1,2,3,4,6-penta-O-acetyl-[alpha]-D-galactopyranose (27.4 g, 70.26 mmol) in dry 1,2-dichloroethane (200 ml) and SnCl4 (9.6 ml, 80 mmol) was added to the residue and the mixture was stored for 16 h at 20oC. Chloroform (100 ml) and saturated aqueous sodium hydrogen carbonate (100 ml) were added, the mixture was stirred for 20 min at 20oC and then filtered through Hyflo Super Cel. The organic layer was separated, washed with aqueous sodium hydrogen carbonate (30 ml) and water, dried and evaporated. The products were purified by column chromatography on silica gel using solvent B to give a foam-like product (28.1 g, 88%). 1H NMR (CDCl3) (400.13 MHz): 8.79 brs (1H, NH), 7.12 q (1H, J6,5 = 1.2 Hz, H-6), 5.80 d (1H, J1',2' = 9.3 Hz, H-1'), 5.47 d (1H, J4',3' = 3.3 Hz, H-4'), 5.27 dd (1H, J2',3' = 10.1 Hz, H-2'), 5.17 dd (1H, H-3'), 4.16-4.06 m (3H, H-5', 6'a, 6'b), 2.18 s (3H, Ac), 2.02 s (3H, Ac), 1.97 s (3H, Ac), 1.96 s (3H, Ac), 1.93 d (3H, Me-5). 13C NMR (CDCl3): 170.36, 169.85, 169.77 and 169.67 (C=O), 163.31 (C-4), 150.39 (C-2), 134.77 (C-6), 112.10 (C-5), 80.56 (C-1'), 73.72 (C-5'), 70.87 (C-3'), 67.01 (C-2'), 66.95 (C-4'), 61.23 (C-6'), 20.65, 20.45 and 20.43 (Me, Ac), 12.59 (Me-5).
The solution of 1-(2,3,4,6-O-tetraacetyl-[beta]-D-galactopyranosyl)- thymine (4.56 g, 10 mmol) in methanol (40 ml) semi-saturated with ammonia at 0oC was kept for 20 h at 20oC and then concentrated in vacuo to dryness. The residue was washed with acetone to yield 1 as a syrup (2.6 g, 90%). UV (pH 1-7): [lambda]max 267 nm ([epsilon] 9500) (pH 13): [lambda]max 267 nm ([epsilon] 7000). 1H NMR (D2O) (400.13 MHz): 7.68 q (1H, J6,5 = 1.2 Hz, H-6), 5.53 d (1H, J1',2' = 9.1 Hz, H-1'), 4.02 dd (1H, J4',3' = 3.3 Hz, J4',5' = 1.0 Hz, H-4'), 3.88 dd (1H, J2',3' = 9.6 Hz, H-2'), 3.87 ddd (1H, J5',6'a = 6.5 Hz, J5',6'b = 5.6 Hz, H-5'), 3.79 dd (1H, H-3'), 3.74 m (2H, H-6'a, 6'b), 1.87 d (3H, Me-5). 13C NMR (D2O): 165.16 (C-4), 151.08 (C-2), 136.40 (C-6), 110.95 (C-5), 81.97 (C-1'), 76.98 (C-5'), 72.10 (C-3'), 67.75 (C-2'), 67.60 (C-4'), 59.88 (C-6'), 10.44 (Me-5).
A suspension of nucleoside 1 (2.9 g, 10 mmol) in dry acetone (10 ml) and 2,2-dimethoxypropane (30 ml) in the presence of p-toluenesulfonic acid monohydrate (0.57 g, 3 mmol) was stirred at 20oC for 3 days. The mixture was neutralized with saturated aqueous sodium hydrogen carbonate, filtered and filtrates were evaporated to dryness. The products were purified by column chromatography on silica gel using solvents A and C to give 2 as a foam (2.5 g, 76%). m.p. 145-146oC (chloroform). Rf 0.08 (solvent C). LSIMS: (M+H) 329. 1H NMR (CDCl3-CD3OD) (400.13 MHz): 7.12 q (1H, J6,5 = 1.2 Hz, H-6), 5.30 d (1H, J1',2' = 8.7 Hz, H-1'), 4.09 dd (1H, J4',3' = 6.6 Hz, J4',5' = 2.1 Hz, H-4'), 4.03 t (1H, J2',3' = 8.7 Hz, H-2'), 3.86 ddd (1H, J5',6'a = 6.9 Hz, J5',6'b = 5.1 Hz, H-5'), 3.65 dd (1H, J6'a,6'b = -11.8 Hz, H-6'a), 3.60 dd (1H, H-6'b), 3.51 dd (1H, H-3'), 1.71 d (3H, Me-5), 1.36 s (3H, Me), 1.18 s (3H, Me). 13C NMR (CDCl3-CD3OD): 164.27 (C-4), 151.27 (C-2), 135.75 (C-6), 111.07 (CMe2), 110.06 (C-5), 82.32 (C-1'), 78.92 (C-5'), 76.04 (C-3'), 73.39 (C-2'), 70.61 (C-4'), 61.35 (C-6'), 27.62 and 25.55 (CMe2), 12.00 (Me-5).
Dinucleoside monophosphate 3 was prepared by condensation of 1-(3,4-O-isopropylidene-6-O-monomethoxytrityl-[beta]-D-galactopyranosyl)thymine (0.71 mmol) and N2,3'-O-diacetyl-2'-deoxyadenosine 5'-phosphate (1 mmol) in dry pyridine (5 ml) in the presence of N,N'-dicyclohexylcarbodiimide (8 mmol) for 7 days at 20oC followed by removal of acyl blocking groups (5 M ammonia in methanol for 48 h at 20oC) and the isopropylidene group (80% acetic acid for 16 h at 20oC) and separation on DEAE-cellulose using a NH4HCO3 concentration gradient (0.0-0.1 M). The pooled fractions (eluted at 0.05 M) were evaporated to dryness, co-evaporated with water (5 * 10 ml) and freeze dried. Yield 25%. Rf 0.56 (TLC, silica gel, 2-propanol:NH4OH:H2O 7:1:2). 1H NMR (400.13 MHz) (D2O): 8.37 s (1H, H-8 A), 8.23 s (1H, H-2 A), 7.63 q (1H, J6,5 =1.2 Hz, H-6 T), 6.45 t (1H, J1',2'a = J1',2'b = 6.5 Hz, H-1' A), 5.65 d (1H, J1',2' = 8.0 Hz, H-1' T), 4.59 ddd (1H, J3',2'a = 6.3 Hz, J3',2'b = 4.3 Hz, J3',4 = 3.6 Hz, H-3' A), 4.37 ddd (1H, J2',3' = 9.3 Hz, J2',P = 9.0 Hz, H-2' T), 4.13 ddt (1H, J4',5'a = 4.0 Hz, J4',5'b = 3.6 Hz, J4',P = 1.5 Hz, H-4' A), 4.00 dd (1H, J4',5' = 0.9 Hz, J4',3' = 3.3 Hz, H-4' T), 3.98 ddd (1H, J5'a,5'b = -11.4 Hz, J5'a,P = 6.1 Hz, H-5'a A), 3.91 ddd (1H, J5'b,P = 5.1 Hz, H-5'b A), 3.88 dt (1H, J5',6'a = J5',6'b = 6.1 Hz, H-5' T), 3.81 dd (1H, H-3' T), 3.76 d (2H, H-6'a and H-6'b T), 2.70 ddd (1H, J2'a,2'b = -14.0 Hz, H-2'a A), 2.57 ddd (1H, H-2'b A), 1.71 d (3H, 5-Me T). The signals were assigned by the double resonance technique and two dimensional COSY method. 31P NMR (161.98 MHz) (D2O) chemical shift in p.p.m. from 80% phosphoric acid: 2.15.
A standard procedure was used to prepare the title compound starting from 2. Yield 86% (foam). LSIMS (Thgly-NaOAc) m/z 653 (M+Na+, 5), 303 (DMTr, 100). 1H NMR (200 MHz) (CDCl3): 8.63 brs (1H, NH), 7.50-7.20 m (10H, arom-H, H-6), 6.84 d (4H, J = 8.8 Hz, arom-H), 5.65 d (1H, J1',2' = 8.5Hz, H-1'), 4.41-3.70 m (4H, H-2', 3', 4', 5'), 3.81 s (6H, 2*MeO), 3.40 m (2H, J = 6.1 Hz, H-6'a, 6'b), 1.90 s (3H, Me-5), 1.57 s (3H, Me), 1.44 s (3H, Me). 13C NMR (CDCl3): 158.42, 149.66, 144.69, 135.98, 135.86, 135.46, 130.08, 128.15, 127.68, 126.70, 123.71 and 112.99 (Ph), 163.86 (C-4), 151.16 (C-2), 135.98 (C-6), 111.11 (Me2C), 109.87 (C-5), 86.17 (Ph3C), 82.67 (C-1'), 78.65 (C-5'), 74.82 (C-3'), 73.41 (C-4'), 71.62 (C-2'), 62.22 (C-6'), 55.14 (2*MeO), 28.03 and 25.82 (CMe2), 12.37 (Me-5).
This was prepared by a standard procedure. Yield 79% (white powder). Rf (hexane:acetone:TEA 49:49:2): 0.58. LSIMS (NBA) m/z 831 (M+H+, 1), 303 (DMTr, 100).1H NMR (200 MHz) (CDCl3): 8.85 brs (1H, NH), 7.50-7.15 m (10H, arom-H, H-6), 6.83 d (4H, J = 8.6 Hz, arom-H), 5.64 d (1H, J1',2' = 8.4Hz, H-1'), 4.41-3.15 m (10H, H-2', 3', 4', 5', 6'a, 6'b, CH in iPr, POCH2), 3.80 s (6H, 2*MeO), 2.81-2.50 m (2H, J = 6.2 Hz, CH2CN), 1.98 s and 1.95 s (3H, Me-5), 1.58 s and 1.56 s (3H, Me), 1.43 s and 1.42 s (3H, Me), 1.29-1.00 m (12H, J = 6.8 Hz, Me in iPr). 13C NMR (CDCl3): 158.32, 144.54, 135.72, 129.95, 128.02, 127.59, 126.63 and 112.89 (Ph), 163.27 and 163.10 (C-4), 150.42 and 150.25 (C-2), 135.05 (C-6), 117.57 (CN), 111.11 and 110.74 (Me2C), 109.72 and 109.63 (C-5), 86.10 (Ph3C), 81.64 and 81.48 (C-1'), 78.65 and 78.43 (C-3'), 74.78 (C-5'), 74.33 and 73.28 (d, J = 15Hz, C-2'), 73.17 (C-4'), 61.95 (C-6'), 58.09 and 57.72 (d, J = 18.0 Hz, POCH2), 55.09 (2*MeO), 43.27 and 42.90 (d, J = 12.7 Hz, NCH), 27.98 and 25.87 (Me2C), 24.7-24.0 (NCMe), 20.10, 19.98.1 (CH2CN), 12.29 (Me-5). 31P NMR (80.99 MHz) (CDCl3) (external reference = H3PO4 capil.): 152.6, 151.7.
Oligonucleotide synthesis was performed on an ABI 392 at 1 [mu]mol scale using commercial 2-cyanoethylphosphoroamidites and standard methodology, except for a longer coupling time (80 s) and a higher concentration (0.15 M) for the amidite 4 to ensure high coupling yields. The products were deprotected and removed from the solid support by concentrated ammonia (55oC, 16 h) and purified by ion exchange on a MonoQ column using a NaCl gradient (0.3-0.7 M) in 10 mM NaOH and reverse phase HPLC on a Nucleosil 100 C18, 5 [mu] column (4 * 250 mm) using a MeCN concentration gradient (10-20% MeCN over 25 min) in 0.1 M triethylammonium acetate, pH 6.9, with subsequent gel filtration of the products on a Toyopearl HW-40 column in water.
A solution of ON (3 OD260) in 80% acetic acid (0.1 ml) was kept at 20oC for 2 h and the mixture separated by HPLC under the above-mentioned conditions. Yields of ON containing Tgal IIa, IIIa and IVa (the upper strands of duplexes II-IV; Table 1 ) were in the range 40-50% (the retention times were 17.3, 17.3 and 18.1 min respectively). The recovered corresponding ONs with isopropylidene groups had retention times of 18.5, 18.7 and 21.3 min. 5'-32P-Labeling of the ONs were carried out using T4 polynucleotide kinase and [[gamma]-32P]ATP.
Table 1
An oligonucleotide solution (0.2 OD) in 400 [mu]l 50 mM Tris-HCI buffer, pH 8.6, 50 mM sodium chloride, 7 mM MgCl2) was digested with 1.0 U snake venom phosphodiesterase (Pharmacia) at 37oC for 16 h. This digestion was followed by treatment with 1.0 U alkaline phosphatase (Boehringer Mannheim) for 4 h at 37oC. Aliquots of this solution (50 [mu]l) were analyzed by HPLC on a Nucleosil 100 C18, 5 [mu] column (4 * 250 mm). Separation of the isopropylidene containing the ON digest was carried out in 0.1 M triethylammonium acetate, pH 6.9, 8% MeCN (flow rate 1 ml/min). The retention time was 7.6 min for 2 and 3.0, 3.9, 4.3, 6.5 min for dC, dG, T and dA.
Separation of unprotected ON digest was performed on the same column using a MeCN concentration gradient (0-10% of MeCN over 25 min, flow rate 1 ml/min) in 0.1 M triethylammonium acetate, pH 6.9. The retention time was 4.4 min for 1 and 7.2, 13.5, 14.3, 19.7 min for dC, dG, T and dA. All oligonucleotides showed a correct ratio of unmodified nucleosides over the modified one. The ON structure was also proved by Maxam-Gilbert sequencing (cleavage at Tgal residues proceeds more slowly than at thymidine).
Each [32P]ON (20-50 pmol) was dissolved in 40-100 [mu]l 50 mM NaIO4 and incubated at 37oC for 1.5-2.5 h. To stop the reaction, ON was precipitated by adding 100 [mu]l 2 M LiClO4 and 1 ml acetone (excess NaIO4 and the products of its reduction were removed). The yield of oxidation was 65-75% as estimated by [beta]-elimination in buffer B at 95oC for 15 min and subsequent electrophoresis on a 20% polyacrylamide gel containing 7 M urea.
The thermal melting curves were obtained in buffer A using a Hitachi 150-20 spectrophotometer (Japan); concentration per duplex (cd) 1 [mu]M. To determine the Tm values, the first derivative was calculated.
M.EcoRII (3.16 [mu]M), M.MvaI (2.47 [mu]M), R.EcoRII (1.45 [mu]M) or R.MvaI (0.5 [mu]M) were incubated with 32P-labeled duplexes (cd 0.35 [mu]M) in 10 [mu]l buffers E, H, C or G respectively, containing 8% glycerol and 0.1 mM AdoHcy in the case of Mtases, at room temperature for 5 min and at 0oC for 15 min. The reactions were run for 1.5-2 h at 120 V on native 8% polyacrylamide gels (the gels were pre-run for 1 h at 100 V). For autoradiography of the electrophoretic pattern, Kodak-XOMAT-S film was exposed with an intensifier screen at -20oC overnight. The efficiency of non-covalent complex formation (R) was determined as the ratio [c.p.m. bound DNA duplex/c.p.m. free DNA duplex]. To normalize data from different experiments, the ratios of R for modified substrates II*-IV* to R for canonical substrate I in each experimental data set were defined. The dependences of the non-covalent complex yields on the enzyme concentration were obtained using 0.23 [mu]M DNA duplexes and 0.35-4.74 [mu]M M.EcoRII, 0.35-4.95 [mu]M M.MvaI or 0.36-4.35 [mu]M R.EcoRII. These data were normalized by dividing the yields of non-covalent complexes of canonical substrate I by the appropriate enzyme in each dataset [cd(I) 0.23 [mu]M; c(M.EcoRII) 4.95 [mu]M; c(M.MvaI) 4.35 [mu]M).
The efficiency of methylation was monitored by incorporation of radioactivity (C[3H]3) into DNA duplexes I-IV*. Methylation reactions were carried out at 20oC for 30 min in 10 [mu]l reaction mixtures containing buffer E (for M.EcoRII) or H (for M.MvaI), 1 [mu]Ci [methyl-3H]AdoMet (15 Ci/mmol; Amersham) and 0.35 [mu]M DNA duplexes I-IV*. Final concentrations of EcoRII and MvaI Mtases were 1.58 and 0.82 [mu]M respectively. Samples were spotted on DE 81 filters (2.5 cm; Whatman), washed with 50 mM KH2PO4 (5 * 150 ml), dried and counted in a liquid scintillation spectrometer. For each series blank values (without enzyme) were treated like the samples (with the complete washing procedure).
Enzymatic cleavage was performed by incubating 0.35 [mu]M 32P-labeled substrate analogs with R.EcoRII (1.45 [mu]M) in 10 [mu]l buffer D or with R.MvaI (5.45 [mu]M) enzyme in 10 [mu]l buffer F at 20oC for 1 h. Enzymatic reactions were stopped by heating at 95oC for 2-4 min. The products of cleavage of 32P-labeled DNA duplexes were analyzed in 20% polyacrylamide gels containing 7 M urea as described (19 ).
Cross-linking of M.EcoRII (0.35-4.74 [mu]M), M.MvaI (0.35-4.95 [mu]M) or R.EcoRII (0.36-4.35 [mu]M) to duplexes II*-IV* (0.23 [mu]M) was performed in 10 [mu]l buffers E, H and C respectively, containing 0.1 mM AdoHcy at room temperature for 5 min and at 0oC for 15 min. Aliquots of 6 [mu]l 10 mM NaBH4 was added and reaction mixtures were kept at 0oC for 40 min. Reactions were followed by 0.1% SDS-8% PAGE after heating the samples in 0.1% SDS at 95oC. The gels were analyzed by autoradiography and staining with Coomassie blue. Coexistence of DNA and protein in the same band proved formation of the DNA-enzyme covalent complex. The cross-linking yield was determined as the ratio of the covalent conjugate radioactivity to total radioactivity of the conjugate and unbound DNA. Competitive inhibition of M.EcoRII (3.16 [mu]M) and M.MvaI (3.3 [mu]M) cross-linking to DNA duplexes II* and IV* (0.23 [mu]M) was studied in the presence of increasing amounts of unlabeled DNA duplexes V (0, 0.46 and 4.6 [mu]M) and VI (0 and 2.3 [mu]M). To determine the influence of Mg2+ ions on R.EcoRII cross-linking to duplex IV*, MgCl2 was added to final concentrations of 5, 15 and 25 mM.
EcoRII and MvaI R-M enzymes recognize the DNA sequence 5'... <=> CC[dArr]A/TGG... (20 ). EcoRII and MvaI restriction endonucleases (R.EcoRII and R.MvaI) cleave DNA as shown by the arrows ( <=> ,R.EcoRII; [dArr], R.MvaI). The cognate DNA Mtases (M.EcoRII and M.MvaI) methylate the `inner' cytosine residues at carbon C5 (M.EcoRII) or at nitrogen N4 (M.MvaI). The enzymes require different cofactors. For Mtases it is S-adenosyl-L-methionine (AdoMet), which is converted to S-adenosyl-L-homocysteine (AdoHcy), whereas Mg2+ ions are necessary for the endonucleases.
To regiospecifically introduce the reactive dialdehyde groups into the EcoRII (MvaI) recognition site, an ON thymidine residue was replaced by 1-([beta]-D-galactopyranosyl)thymine (Tgal) as proposed in Scheme 1. This analog retains all the functions of a natural nucleoside crucial for binding to the enzymes. It also contains a cis diol group. Thus, after periodate oxidation one can obtain ONs with Tgal* (Scheme 1) where dialdehyde groups are in any desired position on the ON strand. 14mer DNA duplex I (Table 1 ) is a canonical substrate of EcoRII and MvaI R-M enzymes (19 ). 14mer DNA duplexes with Tgal (II-IV) and their periodate- oxidized analogs with Tgal* (II*-IV*) are suggested as substrate analogs of the enzymes (Table 1 ). A modified sugar residue is introduced into the center of the EcoRII (MvaI) recognition site (II and II*) or into the 5'-end flanking nucleotide sequence adjacent to the recognition site (III and III*) or separated from its 3'-end by one nucleotide residue (IV and IV*).
An efficient preparative four step synthesis of 1-(3,4-O-isopropylidene-[beta]-D-galactopyranosyl)thymine was developed starting from D-galactose. Fully acetylated D-galactopyranose was condensed with bis-trimethylsilylthymine under Vorbruggen's conditions (21 ) to give, after deacetylation, 1-([beta]-D-galactopyranosyl)thymine (1) (Scheme 1) in high overall yield. We have examined several possibilities for the synthesis of 3,4-O-blocked 1-([beta]-D-galactopyranosyl)thymines and found the simplest and most effective to be that using the isopropylidene group.
Melting temperatures (Tm) of DNA duplexes II-IV with a single galactopyranose residue were determined in buffer A (Table 1 ). Insertion of the modified sugar moiety resulted in a destabilization of 6-12oC compared with the regular duplex I. We could not determine the Tm of DNA duplexes II*-IV* due to their rapid decomposition. Incubation of duplexes with Tgal* under melting conditions (a temperature change from 15 to 80oC, buffer A) resulted in ~50% breaking of the upper strands at the modified sugar residue. It was shown previously that in the case of a 14mer DNA duplex with the dT residue of the EcoRII site replaced by the `open sugar' analog 9-[1'-hydroxy-2'-hydroxymethyl)ethoxy] methylguanine (23 ) the Tm was 14oC below that of a regular duplex . Hence, one can expect a similar destabilization effect for duplexes II*-IV*. Also, drastic double helix distortions are unlikely, since modified DNAs are still substrates of highly sequence-specific enzymes (see below). All cross-linking experiments were done under conditions favorable for thermodynamic stability of the periodate-oxidized duplexes.
Specific binding. The use of DNA duplexes with reactive dialdehyde groups as reagents for cross-linking to R-M enzymes requires evaluation of both their ability to specifically bind to the enzymes and their substrate properties. To obtain correct complexes and to exclude methylation, EcoRII and MvaI binding to the reagents was studied in the presence of the cofactor analog AdoHcy. In addition, we took into account that complex formation between M.EcoRII and the 14mer regular substrate was substantially reduced if no cofactor was added (24 ).
Ternary complex (Mtase-reagent-AdoHcy) formation was monitored in gel mobility shift assays. An increase in binding is observed when the enzyme concentration is increased (data not shown). For both Mtases very poor binding occurred with reagent III* (Table 1 ). M.EcoRII and M.MvaI had high and approximately equal affinity for DNA duplexes II* (Tgal* is in the center of the recognition site) and IV* (Tgal* is in the flanking nucleotide sequence) (Table 1 ).Methylation. Insertion of Tgal into DNA duplex I reduced or even eliminated the activity of EcoRII and MvaI Mtases. For both enzymes the substrate properties of Tgal-containing DNA duplexes strongly depended on the location of the galactopyranose residue (Table 1 ). DNA duplexes with the modified sugar residue in the 5'-end flanking nucleotide sequence adjacent to the recognition site (III) or separated from its 3'-end by one nucleotide residue (IV) retain their ability to be methylated. DNA duplex II with Tgal in the center of the recognition site was not methylated.
Methylation of DNA duplexes with periodate-oxidized galactose residues (II*-IV*) proceeded similarly to galactose-containing duplexes II-IV (Table 1 ).We used all three reagents II*-IV* for affinity modification of EcoRII and MvaI Mtases, taking into account that duplex II* with Tgal in the center of the recognition site was able to bind to the enzymes but was not methylated.
Periodate-oxidized DNA duplex analogs of substrates of type II restriction endonucleases and DNA methyltransferases that contain an active dialdehyde group located as a single substituent in a definite position on the oligonucleotide strand were obtained for the first time. A striking dependence of binding, cleavage and methylation of these duplexes both on the active group location in the duplex and the enzyme was observed. The reagents obtained proved to be efficient for affinity modification of these enzymes, especially in the case of the Mtases. The efficiency of cross-linking was dependent on the active group location and correlated with enzyme binding affinity for the reagents. An advantage of the reagents was their ability to specifically react, mainly with lysine residues of the enzymes, under conditions of lysine proximity to the reactive dialdehyde groups to form a stable covalent bond between the ON and the protein. To understand the functional meaning of the cross-linking dependence on the reactive group location, it is necessary to determine the amino acid residues responsible for binding to the reagent. Such investigations are now in progress. These data will be important for the determination of the active site structure and the mechanism of catalysis of EcoRII and MvaI R-M enzymes.
This research was supported by an award from the Howard Hughes Medical Institute (grant HHMI 75195-545501), INTAS (project 93-1500), NATO (grant HTECH.LG 961324) and the Russian Foundation for Basic Research (projects 96-04-49178 and 95-04-12283A). A.V.A. is a research associate of the Belgian NFWO.
*To whom correspondence should be addressed. Tel: +7 095 939 3144; Fax: +7 095 939 3181; Email: gromova@biorg.chem.msu.su
DNA duplexa
Tm (oC)
Relative methylation (%)b
Relative cleavage (%)c
Relative affinitiesd
M.EcoRII
M.MvaI
R.EcoRII
R.MvaI
M.EcoRII
M.MvaI
R.EcoRII
R.MvaI
I5'-GCCAACCTGGCTCT
64
100
100
100
100
1
1
1
1
3'-CGGTTGGACCGAGA
100
100
II 5'-GCCAACCXGGCTCT
52
0
4
0
0
3'-CGGTTGGACCGAGA
0
30
III 5'-GCCAXCCTGGCTCT
55
15
30
0
0
3'-CGGTAGGACCGAGA
0
90
IV 5'-GCCAACCTGGCXCT
58
75
70
80
100
3'-CGGTTGGACCGAGA
12
70
II* 5'-GCCAACCYGGCTCT
0
1
0
0
0.7
1
0
0
3'-CGGTTGGACCGAGA
0
30
III* 5'-GCCAYCCTGGCTCT
22
23
0
0
0.1
0.1
0.5
0
3'-CGGTAGGACCGAGA
0
80
IV*5'-GCCAACCTGGCYCT
52
41
0
100
1
1
0.8
1
3'-CGGTTGGACCGAGA
0
60
Scheme 1.
REFERENCES