HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 279, Issue 35, 37049-37060, August 27, 2004

This Article Abstract Full Text (PDF) Supplemental data All Versions of this Article: 279/35/37049    most recent M404573200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hopkins, B. B. Articles by Reich, N. O. Search for Related Content PubMed PubMed Citation Articles by Hopkins, B. B. Articles by Reich, N. O. Social Bookmarking                      What's this?

Simultaneous DNA Binding, Bending, and Base Flipping

EVIDENCE FOR A NOVEL M.EcoRI METHYLTRANSFERASE-DNA COMPLEX^*

Ben B. Hopkins and Norbert O. Reich

From the Department of Chemistry and Biochemistry, University of California, Santa Barbara, California 93106

Received for publication, April 26, 2004 , and in revised form, June 17, 2004.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
We measured the kinetics of DNA bending by M.EcoRI using DNA labeled at both 5'-ends and observed changes in fluorescence resonance energy transfer. Although known to bend its cognate DNA site, energy transfer is decreased upon enzyme binding. This unanticipated effect is shown to be robust because we observe the identical decrease with different dye pairs, when the dye pairs are placed on the respective 3'-ends, the effect is cofactor- and protein-dependent, and the effect is observed with duplexes ranging from 14 through 17 base pairs. The same labeled DNA shows the anticipated increased energy transfer with EcoRV endonuclease, which also bends this sequence, and no change in energy transfer with EcoRI endonuclease, which leaves this sequence unbent. We interpret these results as evidence for an increased end-to-end distance resulting from M.EcoRI binding, mediated by a mechanism novel for DNA methyltransferases, combining DNA bending and an overall expansion of the DNA duplex. The M.EcoRI protein sequence is poorly accommodated into well defined classes of DNA methyltransferases, both at the level of individual motifs and overall alignment. Interestingly, M.EcoRI has an intercalation motif observed in the FPG DNA glycosylase family of repair enzymes. Enzyme-dependent changes in anisotropy and fluorescence resonance energy transfer have similar rate constants, which are similar to the previously determined rate constant for base flipping; thus, the three processes are nearly coincidental. Similar fluorescence resonance energy transfer experiments following AdoMet-dependent catalysis show that the unbending transition determines the steady state product release kinetics.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Structural transitions involving substrate-induced changes in enzyme conformation and protein-induced changes in substrate conformation occur in diverse classes of enzymes. The quantitative importance of such mechanisms toward catalysis and specificity has been demonstrated for DNA polymerases (1), phytase (2), integration host factor (3), pyridoxal kinase (4), restriction enzymes, DNA repair enzymes (5), and DNA modifying enzymes (6). Yet, conformational mechanisms are difficult to study because the reaction intermediates are largely inaccessible to classical kinetic analysis. Spectroscopic probes and, in particular, the highly distant-dependent method of fluorescence resonance energy transfer (FRET)¹ provide a viable solution-based approach to characterize conformational intermediates (7). Only through the quantitative analysis of stable conformational intermediates and the kinetic transitions involving their interconversion can a mechanistic understanding of catalysis and specificity be approached.

DNA methyltransferases provide a rich system to investigate the mechanisms and importance of conformational transitions. The enzymes are structurally characterized, carry out both DNA bending and base flipping (8), and the bacterial and human enzymes are the targets of novel antibiotic (9) and cancer drug development efforts respectively (10). These largely monomeric, AdoMet-dependent enzymes modify adenine (N⁶) or cytosine (N⁴ or C⁵) within duplex DNA by delivering a methyl moiety into the major groove. The biological consequences of this epigenetic DNA modification include the control of DNA replication and mismatch repair processes, the restriction of foreign DNA in bacteria, and gene transcriptional regulation in both prokaryotes and eukaryotes (8). M.HhaI is a structurally characterized DNA cytosine methyltransferase, and was the first protein shown to stabilize an extrahelical base as part of its catalytic mechanism (11, 12). Similar base flipping conformational transitions have now been shown for diverse DNA methyltransferases, DNA repair enzymes, and RNA-modifying enzymes (13). Some DNA methyltransferases also bend their cognate DNA sequence by 50-90° as part of their recognition mechanism, and the bending and base flipping process are energetically coupled (14, 15).

The bacterial DNA adenine methyltransferase M.EcoRI modifies the second adenine in the palindromic sequence GAATTC (15). The enzyme uses a facilitated diffusion mechanism to locate its sites within the bacterial genome, and its specificity constant (k_cat/K_m) is near diffusion-limited (16, 17). The discrimination of the enzyme against single base pair-modified non-cognate site methylation is up to 20,000-fold (18). M.EcoRI stabilizes a 50° bend upon binding its cognate site as well as extruding the target adenine into an extrahelical position to facilitate catalysis (19). The DNA binding and base-flipping transitions are nearly concerted, as determined by fluorescence two-colored experiments (6, 20, 21). Scheme I summarizes these characterized transitions and kinetic constants for M.EcoRI. Recently, we described an intriguing M.EcoRI mutant, H235N, which does not appear to bend its cognate sequence, yet shows largely wild-type steady state parameters (15). Surprisingly, this bending-deficient mutant sequence-discrimination is enhanced at least 1000-fold, suggesting a close relationship between DNA bending and specificity. Moreover, the rate of base flipping of the mutant is dramatically slowed down, to the point of determining k_cat, in contrast to the wild-type protein whose k_cat is determined by product release events.

View larger version (7K): [in this window] [in a new window]   SCHEME 1. Kinetic mechanism of M.EcoRI. (E:D)* indicates an enzyme-DNA complex in a bent conformation. (E:D) indicates a complex with the DNA in the bent conformation with the target adenine flipped out of the helix. D-CH3 is the N6-methylated DNA product. AdoMet and AdoHcy are omitted for clarity.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Enzyme Expression and Purification—M.EcoRI was purified from MM294 Escherichia coli cells harboring the pXRI overexpression plasmid to apparent homogeneity as previously described (20). Briefly, cells were grown at 37 °C in Luria Broth (LB) supplemented with 100 µl/ml ampicillin with continuous aeration. Cells were induced in early log phase (OD₆₀₀ = 0.4) with 1 mM isopropyl--D-thiogalactopyranoside for 2-4 h. Cells were harvested by centrifugation at 15,000 rpm for 10 min at 4 °C in a Sorvall RC-5B and stored at -20 °C. The protein was purified using phosphocellulose P-11, hydroxyapatite, and Biorex-70 cation exchange columns and an Amersham Biosciences FPLC system. All purification steps were performed at 4 °C. M.EcoRI purity was determined to be >95% by SDS-PAGE analysis. Aliquots of purified enzyme were flash frozen in liquid N₂ and stored at -80 °C. Protein concentration was determined using the extinction coefficient of 46,151 M^-1 cm^-1 at 280 nm calculated from the Biopolymer Calculator v.4.1.1 (paris.chem.yale.edu). Concentrations determined using the published extinction coefficient EC^1% = 10.8 gave similar values (25). Concentration was further validated by comparison to EcoRV endonuclease (gift from D. Hiller) of known concentration in a dilution SDS-PAGE analysis. Enzyme activity was determined by S-[methyl-³H]adenosylmethionine (ICN) incorporation into calf thymus DNA or synthetic 14-mer duplex DNA containing the GAATTC recognition sequence (26) and shown to be active (data not shown).

DNA Synthesis and Fluorophore Coupling—DNA oligonucleotides were synthesized at Integrated DNA Technologies, Inc. (IDT) containing either 5' or 3' C-6 primary amino modifications. 14-mer top 5'-AGACGAATTCCGAA, 14-mer bottom 5'-TTCGGAATTCGTCT, 15-mer top 5'-AGACGAATTCCGAGA, 15-mer bottom 5'-TCTCGGAATTCGTCT, 17-mer top 5'-AGCACGAATTCAGCACT, and 17-mer bottom 5'-AGTGCTGAATTCGTGCT. The M.EcoRI six base pair recognition sequence is underlined. Adenine-thymine end base pairs were used to avoid the known quenching effect of guanine bases on fluorescence (27-29). Concentrations were determined spectrophotometrically at A₂₆₀ using the calculated extinction coefficients based on nearest neighbor calculations. 5-(and-6)-Alexa Fluor 488, succinimidyl ester (A-20000) and 5-(and 6)-carboxytetramethylrhodamine, succinimidyl ester (C-1171) were purchased from Molecular Probes, Inc. Cy3 and Cy5 (RPN-5661) were purchased from Amersham Biosciences. Oligonucleotides were resuspended in dH₂0, extracted with chloroform, ethanol-precipitated, and resuspended in dH₂0 to a final concentration of 25 µg/µl (100 µg of total oligonucleotide). Coupling reactions were performed with 10-20-fold molar excess of dye to oligonucleotide. 250 µg of Alexa 488, 200 µg of TAMRA, 153 µg of Cy3, and 127 µg of Cy5 were warmed to room temperature and resuspended in 14 µl of Me₂SO immediately prior to coupling. Coupling was done in fresh 75 mM sodium tetraborate (pH 8.5) in the dark with low shaking overnight. Overnight couplings were ethanol-precipitated, and TAMRA-modified oligonucleotides were subject to a second precipitation to remove any nonspecifically interacting free dye. Coupled oligonucleotides were re-suspended in 55 µl of 0.1 M triethylammonium acetate (TEAA), pH 7.0, filtered through a B-101 stainless steel filter (Upchurch Scientific), and purified on an analytical Vydac C4 column with an in-line filter and C4 guard column on a Waters/Millipore HPLC system (616 pump, 600S controller, and 996 photodiode array detector). A linear gradient of 0-80% acetonitrile in 0.1 M TEAA, pH 7.0, from 2-40 min or from 2-60 min for Alexa 488-coupled oligonucleotides was used to remove uncoupled oligonucleotide and unreacted dye from coupled oligonucleotide. Absorbance at 260 nm (oligonucleotide) and fluorophore _max were simultaneously monitored (Alexa 488, _max = 492 nm; TAMRA, _max = 560 nm; Cy3, _max = 549 nm; Cy5, _max = 648 nm). Peaks with absorbances at both 260 nm and dye _max were manually collected. Coupled oligonucleotides were lyophilized in a Speed-Vac (Savant) and resuspended in 10 mM Tris, pH 7.8, 1 mM EDTA (TE). Concentrations of oligonucleotide and dye were determined by absorbance scan from 230 to 700 nm and determined to be 100% coupled. Fluorophore absorbance at 260 nm was subtracted from oligonucleotide/duplex absorbance at 260 nm using Equation 1,

(Eq. 1)

A_dye represents fluorophore absorbance at _max. CF represents the correction factor for dye and was determined to be 0.3 for Alexa 488 singly labeled oligonucleotides and duplexes, 0.4 for TAMRA singly labeled oligonucleotides and duplexes, 0.7 for Alexa 488-TAMRA doubly labeled duplexes, and no correction for Cy3- and Cy5-labeled oligonucleotides and duplexes. Purified oligonucleotide dye conjugates were analyzed by both 20% native PAGE (89 mM Tris, 89 mM borate, 2 mM EDTA) run at a constant 300 V for 2-3 h and by nanoelectrospray mass spectrometry in negative-ion mode (Q-STAR, Applied Biosystems/MDS Sciex) with an m/z range of 500-2000. Oligonucleotides were annealed in TE buffer supplemented with 50 mM NaCl at a 1:1.5 molar ratio with excess acceptor-labeled or unlabeled oligonucleotide, heated at 85-90 °C for 5 min, with slow cooling to room temperature over several hours. Annealed duplexes were analyzed by 20% native PAGE. DNA was visualized with a Typhoon 8600 fluorimager (Amersham Biosciences).

Dissociation Constant Determination, —To determine M.EcoRI binding affinity () for singly labeled DNA, 5'-Alexa 488-modified DNA, a gel mobility shift assay was performed (15). Briefly, 5 nM DNA was incubated with 0-50 nM M.EcoRI in the presence of 20 µM sinefungin (Sigma) in 100 mM Tris, pH 8.0, 10 mM EDTA, 200 µg/ml bovine serum albumin, and 10 mM DTT (MRB) buffer supplemented with 50 mM NaCl in a total volume of 30 µl. Binding reactions were incubated at 37 °C for 20 min, mixed with 6x gel-loading buffer (40% glycerol, 45 mM Tris borate, 1 mM EDTA), and resolved on a native 12% PAGE gel (89 mM Tris, 89 mM EDTA, 2 mM -mercaptoethanol) run at 500 V for 5 min and then 300 V for 2 h at 4 °C. Free duplex and ternary complex were visualized on a Typhoon 8600 fluorimager (Amersham Biosciences). A FRET-based determination of the DNA dissociation constant was also obtained for doubly labeled DNA (5'-Alexa 488/5'-TAMRA). Briefly, DNA (25 nM) was preincubated with 1 µM sinefungin in 50 mM Tris, pH 8.0, 5 mM EDTA, 50 mM NaCl, and 10 mM DTT. 0-100 nM M.EcoRI was added, and Alexa 488 emission was integrated and plotted against enzyme concentration. Experiments were done on a PerkinElmer LS50B luminescence spectrometer at room temperature with 485-nm excitation and 8-nm excitation and 10-nm emission slits. Data from both gel shift and fluorescence experiments were fit to a quadratic binding equation to determine . M.EcoRI was confirmed to bind all other modified DNA duplexes by a single enzyme concentration gel shift analysis. These experiments were carried out essentially as described above using DNA (5 nM) incubated with M.EcoRI (50 nM).

Stopped Flow FRET—Stopped flow fluorescence experiments was performed on an Applied Photophysics SX.18MV stopped flow reaction analyzer outfitted with a single channel emission photomultiplier tube (PMT) oriented at 90° to the excitation beam. Donor fluorophore was excited at 485 nm (_ex) through 8-mm monochromater slits with a 150 watt xenon lamp. Pressure was maintained at 100 psi with compressed N₂. Energy transfer was detected as an increase in donor (Alexa 488) emission (_em) through a 515-nm cutoff filter (Edmund Industrial Optics). All experiments were conducted by loading separate syringes with DNA alone, enzyme and cofactor, and buffer alone. The buffer contained 50 mM Tris, pH 8.0, 5 mM EDTA, 100 mM NaCl, and 1 mM DTT. Duplex DNA concentration was constant at 10 nM (final) and was rapidly mixed with enzyme concentration (50-500 nM, final). Vast excess of one reactant allows the study of a second-order reaction under pseudo first-order conditions (30). A background was obtained by mixing DNA and buffer alone. The cofactor analog sinefungin was constant at 2 µM and was preincubated with the enzyme solutions. All experiments were held constant at 22 °C with a circulating water bath. Reactions were initiated by rapidly mixing 200 µl of DNA with 20 µl of enzyme and cofactor. Data were collected for 10 s with 400 data points/sec. Data from 2 to 3 time courses were averaged to increase the signal-to-noise ratio. This averaged data set was then fit to the single exponential Equation 2,

(Eq. 2)

where y_o is the background fluorescence, a is the amplitude, b is the rate constant, and x is time.

Residual analysis of the fits was also performed. To reduce variation within a single experiment, several independent experiments were performed, and the rates calculated were averaged. To extract the forward rates (k_bend), the observed rates (k_obs) versus enzyme concentration was plotted and fit to a simple linear Equation 3 (30).

(Eq. 3)

Stopped Flow Catalysis—To observe the initial energy transfer decrease due to protein-DNA bent-complex and subsequent product release, reaction conditions were similar to those described above with the exception that 2 µM S-adenosylmethionine was substituted for sinefungin. The observation time was increased to 50 or 100 s.

EcoRI Endonuclease Challenge—The methylation state of doubly labeled DNA was determined by EcoRI endonuclease analysis. DNA and M.EcoRI taken directly from the stopped flow catalysis experiment were used. 10 nM DNA was incubated with buffer (mock), 500 nM enzyme for 50 s (stopped flow conditions), or overnight (positive control) at room temperature. The enzyme was heat-inactivated by incubating at 95 °C for 5 min with slow cooling to room temperature. Mg²⁺ was added to a final concentration of 11 mM and combined with 40 units of EcoRI endonuclease (NEB). Reactions were incubated at 37 °C for 1 h. Samples were run for 1 h at a constant 300 V on a 20% nondenaturing PAGE gel (89 mM Tris, 89 mM borate, 2 mM EDTA) at room temperature. DNA was visualized on a Typhoon 8600 fluorimager by scanning with a 532-nm laser and collecting emission through a 526-short-pass filter for Alexa 488. Images were exported into ImageQuant v5.2 (Molecular Dynamics) for analysis.

Stopped Flow Fluorescence Anisotropy—Stopped flow anisotropy was performed with the same apparatus described above with the FP.1 anisotropy hardware and software package installed (Applied Photophysics). A T-format detection scheme was used with _ex = 555 nm and horizontal and vertical emission measurements collected through two 570-nm cutoff filters (Edmund Industrial Optics). The G-factor was determined and automatically corrected for in anisotropy calculations. 50 nM TAMRA-labeled duplex DNA incubated in the above buffer was mixed with increasing enzyme (200, 300, 500 nM final) preincubated with 2 µM sinefungin. Data were collected in oversampling mode to reduce background noise. Data were collected over a 2, 5, or 10 s time course, and 2-3 runs were averaged to obtain adequate signal-to-noise. A plot of k_obs versus enzyme concentration was used to obtain the association rate (k_on). Anisotropy (Anis) was calculated according to Equation 4 (31).

(Eq. 4)

Total emission (Ems) was calculated from Equation 5.

(Eq. 5)

Steady State FRET—Equilibrium fluorescence experiments were performed on an LS50B luminescence spectrometer at room temperature. Because different fluorophores were used, excitation and emission wavelengths and excitation and emission slits varied and are indicated in the corresponding figures. Solutions contained either 10 or 25 nM DNA preincubated with 1 µM sinefungin and 28 or 50 nM enzyme was added, respectively. Quantitation of fluorescence was done by integrating _max± 5 nm. For all experiments, the lamp was warmed for 30-60 min, and scans of DNA alone were performed until consecutive scans were identical. Enzyme was added, manually mixed, and allowed to equilibrate for 2-3 min before taking emission scans. All experiments were performed in a narrow quartz fluorescence cuvette (Starna Cells). For EcoRI and EcoRV endonuclease fluorescence experiments, buffer conditions were 50 mM Tris, pH 7.5, 10 mM Ca²⁺, and 100 mM NaCl. Cleavage was initiated by supplementing with 25 mM Mg²⁺. 10 nM Alexa 488/TAMRA doubly labeled DNA was mixed with 63 nM EcoRI endonuclease or 490 nM EcoRV endonuclease.

Distance Calculations—FRET was used to calculate distance changes from free DNA to M.EcoRI-DNA complexes according to Equation 6 (31).

(Eq. 6)

The Forster radius, defined as the distance at which FRET equals 50%, is calculated from Equation 7 where units are in Å⁶ (32).

(Eq. 7)

In Equation 7, _D represents the quantum yield of donor fluorophore attached to DNA in the absence of acceptor fluorophore and can be calculated according to Equation 8 (31).

(Eq. 8)

I_d and I_rf represent the integrated fluorescence intensities (510-520 nm) of donor (Alexa 488 appended to 14-mer duplex DNA) and reference fluorescein in 0.1 N NaOH, respectively with 485-nm excitation. A_rf and A_d represent the absorbance intensity of the fluorescein reference and donor fluorophore, respectively at 485 nm. Absorbance values were 0.05 to avoid inner filter effects (31). The refractive indices of the reference and sample solution were assumed to be equal. _RF represents the quantum yield of the fluorescein reference (_RF = 0.95 for fluorescein) (33). Steady state emission was measured on a LS50B luminescence spectrometer at room temperature. Absorbance scans were performed with a Beckman Coulter DU640 spectrophotometer or a Cary 1E UV-vis spectrophotometer (Varian).

K² represents the orientation factor and is assumed to be two-thirds for a randomly oriented donor/acceptor dye pair (31). n represents the refractive index of the solution and is assumed to be 1.4 for an aqueous solution (31).

J(v) represents the overlap integral of the donor fluorescence and acceptor absorbance and is calculated according to Equation 9 where units are in M^-1 cm^-1 nm⁴ (31).

(Eq. 9)

In Equation 9, F() represents the measured fluorescence intensity of the donor dye coupled to DNA in the absence of acceptor dye at . () is the extinction coefficient of acceptor dye coupled to DNA at . A TAMRA extinction coefficient of 95,000 M^-1 cm^-1 at _max = 560 nm was used to calculate the extinction coefficient values for each .

Absorbance and fluorescence data were collected between 500 and 600 nm in 1-nm increments ( = 1). Energy transfer was calculated from the donor dye emission. The method requires donor-labeled duplex DNA and donor/acceptor-labeled DNA (28). Donor-only DNA and donor/acceptor DNA were incubated with 2-fold molar excess enzyme over DNA. Fluorescence emission was monitored in the absence and presence of enzyme. To normalize the emission signal, both donor-only and donor/acceptor duplexes were digested with EcoRI endonuclease, and the final donor emission was used to normalize the DNA-M.EcoRI complex intensity. Energy transfer (E) was calculated from Equation 10 (31),

(Eq. 10)

where DA represents the donor emission of donor/acceptor-labeled DNA and D represents the donor emission of donor-only-labeled DNA.

Dissociation Kinetics—Determination of k_off was performed on the LS50B luminescence spectrometer in time-drive mode at room temperature. 10 nM 14-mer 5'-Alexa 488/5'-TAMRA duplex was preincubated with 25 µM sinefungin, and addition of 28 nM M.EcoRI demonstrated decreased energy transfer. 1.6 µM unlabeled 15-mer duplex was then added to the mixture and manually mixed by inverting the fluorescence cuvette. Time between mixing and the beginning of monitoring fluorescence was recorded as dead-time and corrected for in the rate calculation. All other conditions were similar to those used for the stopped flow reactions (see above). Experiments were performed in triplicate, data were averaged, and standard deviation calculated. _ex = 485 nm and _em = 515 nm with 8-nm excitation and 10-nm emission slits. Data were fit to a single exponential decay equation and residuals were plotted versus time. Fluorescence scans were taken after complete decay to demonstrate the return to the free DNA state.

Sinefungin Dissociation Equilibrium Constant—The sinefungin dissociation constant () for the ternary complex was determined on a LS50B luminescence spectrometer at room temperature. 25 nM doubly labeled DNA was preincubated with 50 nM enzyme under the same solution conditions described above. Sinefungin was titrated in to a final concentration of 3.4 µM. Alexa 488 emission was integrated from 510 to 520 nm. Data from two independent experiments were normalized, averaged, and fit to a rectangular hyperbolic binding curve to obtain

Data Analysis—Data analysis for stopped flow, k_off determination, and dissociation constant determination were done with SigmaPlot 2000 v.6.10 or Microcal Origin v.5.0. Kinetic simulations to determine the limit for the bending rate constant were performed with KinTekSim v.3.20 (34-36). Data were fit to the Scheme 1 mechanism for binding and bending with k_on set to 2.7 x 10⁷ M^-1 s^-1; both k_off and k_unbend were set at 0.02 s^-1. k_bend was varied to determine the lower limit of the first-order bending constant. FRET experiments done in the presence of sinefungin prevent further steps shown in Scheme 1 (k_chem and k_release). Because we are not observing flipping, we did not include k_flip.

Protein Sequence Analysis—The sequences of 400 -class methyltransferases were obtained from the Restriction Enzyme Data base (REBASE) at NEB (rebase.neb.com/rebase). Sequence analysis was performed on only 72 of these sequences due to the presence of 330 putative methyltransferase sequences. The M.EcoRI amino acid sequence was manually included because of the NEB automatic classification program that placed M.EcoRI in the "other methyltransferase genes." The motivation to include M.EcoRI in this class is largely due to the prior alignment efforts, which included motif order to assign classification (37). Sequences were aligned in multiple alignment mode with Clustal X v.1.81 (38) using default parameters. From the alignment, a phylogenetic tree was constructed using TreeView v.1.6.6 to visualize the alignment. Sequence identity was determined using ALIGN (www2.igh.cnrs.fr/bin/align-guess.cgi) with default parameters (39).

Sequence Analysis and Consensus Visualization—Methyltransferase motif consensus sequence as determined by multiple alignment, using Malone et al. (37) as a guide, was visualized with sequence logos (40) using Weblogo v.2.5. The amino acid residue at each position in the motif is displayed according to the amount of information and by the frequency of residue x at position y. Motif polymorphisms at each position were quantified using the Shannon entropy in Equation 11 where units are in bits (41).

(Eq. 11)

H(y) indicates the uncertainty at position y, x indicates each amino acid residue, and f(x, y) indicates the frequency of residue x at position y. M is equal to 20, the total number of amino acid residues. The total information content at position y is calculated from Equation 12,

(Eq. 12)

where e(n) is a correction factor for small sample (n) size (40).

The height of each residue x at position y (H(x, y)) is proportional to its frequency at that position and is calculated according to Equation 13.

(Eq. 13)

Briefly, to show consensus residues, the H(y) value should be low or zero in a sufficiently large sample size (n > 20) (42) to decrease e(n) to zero. Therefore, the maximum information content possible is 4.3 (log₂ 20 = 4.3) from equation R(y) = log₂ 20-(H(y) + e(n)). The highest frequency residue at position y is located at the top of the stack of residues in the sequence logo.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Oligonucleotide Dye Coupling and Analysis—We utilized reversed-phase HPLC (RP-HPLC), mass spectrometry, and gel electrophoresis to obtain and confirm highly labeled oligonucleotides. RP-HPLC purification was used to separate unreacted dye and uncoupled oligonucleotide from the desired oligonucleotide-dye conjugate. In all instances, unreacted dye eluted from the column prior to the oligonucleotide species. Coupled oligonucleotide eluted 1-4 min after uncoupled oligonucleotide with baseline resolution (data not shown). PAGE analysis of oligonucleotide-dye demonstrated the existence of a single fluorescent band. Mass spectrometry analysis of uncoupled and coupled oligonucleotide demonstrated an expected increase in mass because of the addition of a fluorophore (not shown). PAGE analysis of annealed DNA duplexes demonstrated the presence of duplex and excess single-stranded acceptor-labeled oligonucleotide (not shown).

Dissociation Constant () Calculation—Gel shift analysis of M.EcoRI binding to 5'-Alexa 488 singly labeled duplex was used to determine the affinity of M.EcoRI for fluorophore-modified duplexes used for future experiments (15). A dissociation constant of 16.6 (± 5.2) nM was calculated from plots of percent duplex versus enzyme concentration (not shown). Further validation of the binding affinity was performed by integration of donor fluorescence emission of 5'-Alexa 488/5'-TAMRA doubly labeled duplex. Plots of the donor emission versus enzyme concentration were used to calculate a ( of 9.2 (± 4.7) nM (not shown). Both dissociation constants are similar to previously reported ( values of 28 (± 7) nM (15), 1.08 (± 0.10) nM (21), 4.4 (± 0.8) nM (18) for various 14-mer duplex DNA substrates. The ability of M.EcoRI to bind 5'- or 3'-modified duplexes with various fluorophores was assessed by gel shift assay using 5 nM DNA incubated with 50 nM M.EcoRI. M.EcoRI was capable of binding all modified duplexes (not shown).

M.EcoRI Binding to 5'-Alexa 488/5'-TAMRA Duplex DNA Causes a Reduced Energy Transfer—To determine how M.EcoRI deforms DNA in solution, we carried out steady state FRET studies using doubly labeled (14-mer 5'-Alexa 488/5'-TAMRA) duplex DNA. Similar analysis has been performed with EcoRV endonuclease, which bends its cognate sequence GATATC (43). Given that wild-type M.EcoRI has previously been shown to bend its cognate DNA sequence 50-54° by atomic force microscopy (AFM), comparison to intrinsically bent A-tract DNA, and circular permutation analysis (19), this reduced energy transfer was unexpected (Fig. 1). This effect is characterized by an increase (25-30%) in donor (Alexa 488) emission and a similar decrease (5%) in acceptor (TAMRA) emission; we observed a clear isosbestic point near 574 nm in the process of carrying out protein titrations (data not shown). As observed by others (44, 45), the change in TAMRA signal is not as dramatic as that of Alexa 488 because of direct excitation of TAMRA of 485 nm incident light. Changes in donor emission were monitored for all steady state and presteady state fluorescence experiments. Direct excitation of TAMRA in the absence and presence of enzyme does not demonstrate any change in emission (not shown). For all experiments with Alexa 488 and TAMRA, peaks were consistently observed at 515 ± 1 nm (Alexa) and 580 ± 1 nm (TAMRA). Taken together, the presence of an isosbestic point and lack of change in _max upon addition of protein indicates that the change in signal is due solely to a sensitization of donor and a quenching of acceptor and an absence of dye-protein interaction. To address dye-DNA sequence effects (27-29), the fluorophore attachments were switched (inset). The same effect is also seen with this switched-dye duplex, indicating that fluorophore attachment does not affect the fluorescence emission.

View larger version (13K): [in this window] [in a new window]   FIG. 1. Steady state fluorescence emission scans. A decreased FRET is observed with donor/acceptor covalently modified duplex DNA and M.EcoRI. 10 nM 14-mer 5'-Alexa 488 (Al, top oligo)/5'-TAMRA (Rh, bottom oligo) duplex DNA in the absence (black) or presence (red) of 28 nM M.EcoRI. DNA was preincubated with 1 µM cofactor analog sinefungin. Reaction conditions consisted of 50 mM Tris, pH 8.0, 5 mM EDTA, 50 mM NaCl, and 10 mM DTT at room temperature. A, excitation of Alexa 488 donor was at 485 nm and emission was monitored from 500-650 nm through 8- and 10-mm excitation and emission slits, respectively. Alexa 488 emission peak is at 515 nm and TAMRA emission peak is at 580 nm. Inset, steady state emission spectrum under similar conditions but with the fluorophore oligo-coupling switched: 14-mer 5'-Alexa 488 (bottom oligo)/5'-TAMRA (top oligo) where the FRET effect is also observed. B, close-up of the 560-610 nm range to show isosbestic point at 574 nm of ± enzyme scans.

View larger version (17K): [in this window] [in a new window]   FIG. 2. A, FRET effect is dye-independent. Steady state fluorescence spectrum of 14-mer 5'-Alexa 488 (Al, top)/5'-Cy3 (bottom)-labeled duplex DNA in the absence (black) and presence (red) of M.EcoRI. ex = 485 nm for Alexa 488, emission monitored from 500 to 650 nm where Alexa 488 peak at 515 nm and Cy3 peak at 567 nm are observed. Excitation and emission slits set at 8 and 10 mm, respectively. Inset, steady state emission spectrum of 14-mer 5'-Alexa 488 (top) singly labeled duplex DNA. No effect on emission is observed at saturating enzyme concentration. No change is also seen when 14-mer 5'-Alexa 488/5'-TAMRA duplex DNA is incubated with heat-killed M.EcoRI (not shown). B, FRET effect is independent of dye placement on oligonucleotide (5' versus 3'). Steady state fluorescence spectrum of 14-mer 5'-Cy3/5'-Cy5 duplex DNA in the absence (black) and presence (red) of M.EcoRI. ex = 515 nm for Cy3, emission monitored from 540-720 nm where Cy3 peak at 567 nm and Cy5 peak at 662 nm is observed. Excitation and emission slits were set at 15 and 15 mm, respectively. Inset, fluorescence emission of 14-mer 3'-Cy3/3'-Cy5 duplex DNA. No effect on emission is observed upon switching dye-oligo coupling from the 5'- to the 3'-end of the oligonucleotides. FRET is also observed with 3'-Alexa 488/3'-TAMRA, 3'-Alexa 488/3'-Cy3, and 3'-TAMRA/3'-Cy5 DNA duplexes (not shown).

FRET Effect Is Cofactor-dependent—To further address the relationship between protein binding and the observed FRET effect, 5'-Alexa 488/5'-TAMRA duplex were incubated with saturating enzyme in the absence of the cofactor analog, sinefungin. We previously demonstrated that tight binding to its cognate DNA requires the presence of cofactor for M.EcoRI (48). No change in FRET is observed when enzyme was mixed with DNA without sinefungin. However, once sinefungin was added, the characteristic FRET effect was observed (Fig. 3A). Gel shift analysis was also performed in the presence and absence of sinefungin. At 5-fold molar excess enzyme over DNA (5 nM DNA, 50 nM M.EcoRI), in the absence of sinefungin, no shift was observed while a complete shift was observed in the presence of sinefungin (not shown). To further validate if the FRET effect was dependent on cofactor, a sinefungin titration experiment was performed where DNA and protein were held constant and sinefungin (0-3.4 µM) concentration was varied (Fig. 3A, inset). The increase in Alexa 488 peak (integrated from 510-520 nm) was plotted against sinefungin concentration and used to calculate a dissociation constant () for the ternary complex (enzyme-DNA-sinefungin) of 103 ± 17 nM.

View larger version (25K): [in this window] [in a new window]   FIG. 3. A, FRET effect is cofactor-dependent. Steady state fluorescence emission of 14-mer 5'-Alexa 488/5'-TAMRA (black) incubated with saturating M.EcoRI (red) does not demonstrate the FRET phenomenon. Addition of 1 µM of the cofactor analog sinefungin to the DNA-enzyme solution demonstrates FRET (green). Inset, titration of increasing concentration of sinefungin to 25 nM 14-mer 5'-Alexa 488/5'-TAMRA and 50 nM M.EcoRI. Emission scans were normalized, and the Alexa 488 peak was integrated from 510-520 nm. Data were fit to a single rectangular hyperbola, and a dissociation constant for sinefungin () was calculated to be 103 ± 17 nM. Data were average of two individual experiments. Titration was done at 50 mM Tris, pH 8.0, 5 mM EDTA, 100 mM NaCl, and 1 mM DTT. Fluorescence scans under these conditions were similar to previous conditions (not shown). B, steady state fluorescence emission of 14-mer 5'-TAMRA/5'-Cy5 duplex DNA in the absence (black) and presence (red) of M.EcoRI. ex = 535 nm for TAMRA, emission monitored from 563-720 nm where TAMRA peak at 585 nm and Cy5 peak at 662 nm is observed. Excitation and emission slits were set at 10 and 20 mm, respectively. Inset, FRET effect observed in different duplex DNA lengths. 14-, 15-, and 17-mer 5'-Alexa 488/5'-TAMRA labeled DNA in the absence or presence of M.EcoRI. Emission curves were integrated: 510-520 nm for Alexa 488. Alexa 488 curves were normalized to the + M.EcoRI scans. Black bars, Alexa 488-M.EcoRI; gray bars, Alexa 488; +, M.EcoRI. FRET is characterized by an increase in donor emission concurrent with a decrease in acceptor emission. The largest increase in donor emission is seen in the shortest DNA: 14-mer. The smallest change is seen with the longest DNA. The same relationship is seen with the decrease in acceptor emission, where the largest change is seen in the shortest DNA and the smallest change is seen in the longest DNA. C, increased FRET is observed with another known DNA-bending enzyme: EcoRV endonuclease (69). Incubation of 10 nM 14-mer 5'-Alexa 488/5'-TAMRA duplex DNA (black) with 490 nM EcoRV (red) in 10 mM Ca2+ results in an increased FRET, which is characterized by a quenching of donor emission and an increase in acceptor emission. The large excess of EcoRV is necessary because the affinity of EcoRV for a GAATTC (EcoRI) site is 100 nM.4 Inset, no FRET effect is seen with a non-bending enzyme that binds the substrate DNA as determined by x-ray crystallographic studies (70). 10 nM 14-mer 5'-Alexa 488/5'-TAMRA duplex DNA (black) is incubated with 63 nM EcoRI endonuclease (red) in the presence of 10 mM Ca2+ and no effect is observed. To determine if EcoRI endonuclease is actually binding the substrate DNA, 25 mM Mg2+ was supplemented, and restriction cutting of the 14-mer duplex was monitored. Digestion was complete in 31 min (green).

Other Proteins Induce Anticipated Changes with Same DNA Substrate—To determine if the DNA used in our system shows the expected increased energy transfer when bound by a protein known to bend this sequence, 14-mer 5'-Alexa 488/5'-TAMRA duplex DNA used in previous experiments was incubated with EcoRV endonuclease in the presence of 10 mM Ca²⁺. EcoRV endonuclease is capable of bending this sequence but to a lesser extent than the cognate site.² Solution-based FRET demonstrates that EcoRV causes the expected increase in energy transfer as indicated by the quenching of the donor fluorophore (Fig. 3C). A second experiment was done with EcoRI endonuclease, a non-bending protein, which binds this 14-mer DNA sequence with tight affinity (49, 50). We observe no change in FRET (Fig. 3C, inset). To confirm that the enzyme binds this DNA, we added Mg²⁺ and observed cleavage of the duplex by virtue of the increase in donor emission and a decrease in acceptor emission (inset). These observations with a characterized bending protein, EcoRV endonuclease (43), and non-bending protein, EcoRI endonuclease, provide strong evidence that the DNA used behaves as expected and the decreased energy transfer observed with M.EcoRI is not a result of anomalous interactions between the enzyme and its substrate.

Anisotropy Measurements of M.EcoRI Binding to DNA—Association kinetics (k_on) of M.EcoRI for 14-mer duplex DNA were determined by stopped flow anisotropy. Similar to stopped flow FRET experiments, anisotropy measurements were done under pseudo first-order conditions with 50 nM singly labeled TAMRA 14-mer, 200-500 nM M.EcoRI preincubated with 2 µM sinefungin. Plots of anisotropy versus enzyme concentration were used to calculate k_on (1.7-5.0 x 10⁷ M^-1 s^-1), which is similar to previously determined values (data not shown) (6, 20, 21).

Kinetic Determination of the DNA Conformational Change Induced by M.EcoRI—Presteady state studies were performed to obtain the kinetic parameter k_bend of doubly labeled duplex DNA and M.EcoRI. Stopped flow analysis of enzyme bending DNA was performed under pseudo first-order conditions, where enzyme concentration was in vast excess of DNA concentration (30). Rapid mixing of DNA (10 nM) with enzyme (50-500 nM) and sinefungin (2 µM) solutions was done at 22 °C (Fig. 4A). The increase in donor emission (Alexa 488) detected through a 515-nm cutoff filter was used as a "bending" signal. Data were fit to a single exponential equation to determine k_obs. The singly labeled duplex DNA control demonstrated no change in signal upon mixing with M.EcoRI (not shown). The observed rate constant (k_obs) increases with increasing enzyme concentration.

View larger version (21K): [in this window] [in a new window]   FIG. 4. A, observation of bending kinetics of M.EcoRI. Stopped flow analysis by monitoring donor emission. 10 nM 14-mer 5'-Alexa 488/5'-TAMRA was rapidly mixed with 0, 50, 100, 200, 300, and 500 nM M.EcoRI preincubated with 2 µM sinefungin (final concentrations). Reaction conditions were 50 mM Tris, pH 8.0, 5 mM EDTA, 100 mM NaCl, and 1 mM DTT at a constant 22 °C. ex = 485 nm of Alexa 488. Excitation and emission slits were set at 8 and 8 mm, respectively. Donor emission was monitored through a 515 nm cutoff filter. Shown is a representative experiment. B, M.EcoRI concentration dependence on bending rate. Data from Fig. 7 were fit to a single exponential. The observed rate increases linearly with enzyme concentration, indicating the process monitored is second order. The slope of the line corresponds to the second-order forward bending constant (kbend), equal to 2.7 (± 0.2) x 107 M-1 s-1. Data from three individual experiments were averaged, and standard errors calculated.

The Unbending Transition—The dissociation rate (k_off) of M.EcoRI was determined using a solution-based competition assay by mixing prebound enzyme-DNA with excess unlabeled DNA. First, DNA was mixed with enzyme, and the decreased energy transfer effect was observed. Then a 160-fold excess of unlabeled DNA was added, manually mixed, and the emission was monitored at 515 nm (Fig. 5). Time between the addition of unlabeled DNA and beginning emission measurements was recorded and corrected for in calculating k_off. However, no significant difference was noticed in k_off with or without "dead-time" correction. Emission scans after the fluorescence signal had reached its end point overlapped with emissions scans of DNA in the absence of enzyme, indicating that the protein-induced conformational changes are reversible. Data from three independent experiments were averaged and fit to a single-exponential decay and residuals plotted against time. k_off was calculated to be 0.016 (± 0.001) s^-1 in 100 mM NaCl. This is in close agreement with previously determined k_off (0.024 s^-1) using a competition assay under similar buffer conditions with identically sized DNA (6). k_off determined by gel-shift competition assay was determined to be 0.019 (± 0.006) s^-1 with the 14-mer duplex DNA (16).When the preformed DNA-enzyme-sinefungin ternary complex was mixed with buffer alone, no change in signal was observed over 100 s (not shown).

View larger version (19K): [in this window] [in a new window]   FIG. 5. Determination of the dissociation rate, koff. 10 nM 14-mer 5'-Alexa 488/5'-TAMRA was incubated with 28 nM M.EcoRI in the presence of 25 µM sinefungin. 1.6 µM unlabeled 15-mer duplex DNA was added manually and quickly placed in the fluorometer. ex = 485 nm for Alexa 488, em = 515 nm was monitored through 8 and 10 mm excitation and emission slits, respectively. Dead-time between the addition of competitor DNA and beginning of emission monitoring was corrected for. No effect was seen in the absence of competitor DNA (not shown). Data were fit to single- and double-exponential decay equations, which fit equally well. Residual analysis for single-exponential fit is shown. Average koff from single exponential fits of three separate experiments was 0.016 ± 0.001 s-1. No significant difference was noted in fits without "dead-time" correction.

View larger version (47K): [in this window] [in a new window]   FIG. 6. Real-time observation of bending, catalysis, and product release. Similar setup as Fig. 4, but 2 µM sinefungin was replaced with 2 µM AdoMet. 10 nM 14-mer 5'-Alexa 488/5'-TAMRA was rapidly mixed with 0 (black), 50 (red), 100 (green), 200 (yellow), 300 (blue), and 500 (pink) nM M.EcoRI. The data are the average of two to three shots to increase signal-to-noise. Inset, 20% PAGE analysis of EcoRI endonuclease challenge of 10 nM 5'-Alexa 488/5'-TAMRA methylated with 500 nM M.EcoRI under stopped flow conditions and 5'-Alexa 488/5'-TAMRA unmethylated. Lane 1, DNA-endonuclease; lane 2, DNA + endonuclease; lane 3, DNA + M.EcoRI (o/n) + endonuclease; lane 4, DNA + M.EcoRI (50 s) + endonuclease.

Distance Calculation—To accurately determine distances from energy transfer, we characterized the Forster radius (R_o) of the Alexa 488-TAMRA dye pair appended to the 5'-end of duplex DNA (14-mer). The quantum yield of donor-only duplex (Alexa 488-DNA) was determined to be 0.48, which is similar to the often assumed value of 0.5 for fluorescein-labeled biomolecules. The overlap integral J(v) of Alexa 488-DNA emission and TAMRA-DNA absorption was calculated to be 3.11 x 10¹⁵M^-1 cm^-1 nm⁴, which is similar to the previously determined overlap integral for fluorescein-TAMRA (43). With the assumed K² value of two-thirds and a refractive index of 1.4, an R_o value of 53.2 Å was calculated for the Alexa 488-TAMRA dye pair, which is similar to reported fluorescein-TAMRA Forster radii of 49-54 Å (32). From donor-only and donor/acceptor-labeled duplexes, energy transfer was calculated from donor emission (28). Using the normalized signal of Alexa 488 after EcoRI digestion, energy transfer was calculated to be 45% in unbound B-DNA. This FRET efficiency is similar to that calculated for the fluorescein-rhodamine 14-mer of 44.7% (43). However, upon addition of M.EcoRI, the energy transfer was reduced to 37%. Using the equation: E = R_o⁶/R_o⁶ + r⁶, the calculated distance of uncomplexed DNA was 55.1 Å. Again, this is consistent with the length calculated with a related 14-mer labeled with fluorescein-rhodamine of 54.6 Å (43). Upon binding of M.EcoRI, the distance expanded to 58.2 Å, a change of 3 Å. However, assuming a localized bend in DNA of 50° induced by M.EcoRI results in a predicted net decrease of 4.5-5 Å. Therefore, M.EcoRI must induce an overall expansion of DNA by 8 Å.

Sequence Analysis—Multiple alignments of -class methyltransferase sequences were performed using 72 classified sequences from the NEB REBASE system. However, due to the automatic sorting system at NEB, M.EcoRI was not classified as a -class methyltransferase. The M.EcoRI sequence was manually included, resulting in 73 total sequences. Using Clustal X v.1.81, the 73 methyltransferase sequences were aligned along their entire length. From this multiple alignment, sequence logos were used to visualize motif consensus sequences. As seen in Fig. 7B, Motif IV is very highly conserved across all adenine-methyltransferases, including M.EcoRI. However, another highly conserved motif, Motif I, does not appear to be conserved in M.EcoRI. As seen in Fig. 7A, the consensus sequence for Motif I can be read as (I/V)L(D/E)PXXGXGXFLXXX. The sequence that best fits to this region on M.EcoRI is residue ⁶²VNFDNLGLKKLIASC (Fig. 7). However, Malone et al. aligned the region ⁷⁹NKEGFSSSEAAKNGF to Motif I, a shift down of 17 residues from what we identified. Wilson and co-workers (51) identified a region 30 residues upstream of what Malone et al. (37) identified as Motif I. None of these alignments in M.EcoRI appears to be significant, although our assignment of Motif I within M.EcoRI may include more conserved residues than those found in Malone. The most highly conserved motif among adenine-methyltransferases is Motif IV: NPP(Y/F). According to our multiple alignment and the sequence logo of Motif IV, the consensus sequence is FDX(I/L)(I/L)XNPP(Y/F)XX. Wilson and co-workers (51), Chandrasegaran and Smith (52) and our analysis aligned M.EcoRI region ¹³³SDIVVTNPPFSL to Motif IV. We identified the same region or an overlapping region of M.EcoRI to Malone in six (III, IV, V, VI, VII, and VIII) of the nine conserved motifs. The three motifs that we did not concur on (X, I, and II) are located at the N-terminal region of M.EcoRI. As discussed earlier, the region we determined to be Motif I in M.EcoRI appears to be a better match to the Motif I consensus sequence. As observed by this group and others, M.EcoRI does not match the majority of the motifs present in other methyltransferase sequences.

View larger version (24K): [in this window] [in a new window]   FIG. 7. Consensus sequence logos of methyltransferase motifs. A, motif I; B, motif IV. Below each logo is the M.EcoRI sequence aligned to that region with the residue number of the first position indicated to the left of the sequence. Sequence logos display residues present at each position according to its frequency and total sequence information available at each position (see "Materials and Methods"). The size of each residue is related to its frequency at each position. Maximum value is 4.3 bits as indicated on the y-axis. Residue chemical properties are indicated by color: black, hydrophobic; red, basic; blue, acidic; and green/pink, polar.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Evidence for a Novel Methyltransferase-DNA Complex—Covalently modified DNA containing various fluorescent dyes at either the 5'- or 3'-ends of 14 base pair duplexes were used as substrates in both equilibrium and stopped flow FRET studies with M.EcoRI DNA adenine methyltransferase. Previous studies of M.EcoRI interaction with its cognate site in the presence of the cofactor analog sinefungin demonstrated that the enzyme bends the DNA by 50-55° (19); thus, our initial finding with a derivative of the well established fluorescein/rhodamine dye pair, Alexa 488/TAMRA, showing a decreased energy transfer upon M.EcoRI binding was surprising (Fig. 1). This duplex interacts well with the enzyme, as the apparent steady state parameters are similar to those determined previously for short DNA duplexes (18). In the absence of complicating interactions, the observed decrease in energy transfer causing an enhanced fluorescein emission derives from an overall increase in the dye-to-dye distance (31). We carried out numerous controls to verify this unusual effect; many of these experiments used other dye pairs to determine if the initial observation was robust. A common concern in FRET studies is the direct interaction of the protein with one or both of the dyes (28, 38), which was shown not to occur for any of the dyes studied here (Figs. 2, A and B and 3B). Titration of the singly labeled duplexes with M.EcoRI results in no detectable changes in overall fluorescence, consistent with lack of any direct enzyme-dye interactions. Similarly, direct excitation of the acceptor fluorophore demonstrated no change in emission upon addition of enzyme. Switching the location of the Alexa 488/TAMRA pair so that each dye positioned onto the opposite 5'-end resulted in the identical effect upon M.EcoRI addition (Fig. 1A). These results are consistent with a net expansion of the dye-to-dye distance in the enzyme-DNA complex not resulting from any artifactual contacts with either dye.

We sought to further characterize the influence of the environment near the end of the duplex DNA on each dye fluorescence. To address effects of various terminal base pairs on dye fluorescence (27-29), we switched attachments of the donor (A T) and acceptor (T A) fluorophores (Fig. 1A). The results indicate that regardless of the base the fluorophore is coupled to, both provide the same effect. This experiment also addresses the previously described interactions between TAMRA and particular terminal base pairs. Our observation of similar effects when the dyes are attached to the 3'-ends (Fig. 2B) further shows the observed decrease in energy transfer upon protein binding to be real. The final set of experiments to characterize the M.EcoRI-DNA complex involved using the same dye pair but varying the overall length of DNA from 14 to 17 bp. The goal was to both extend the initial dye to dye distance so as to further remove any potential for direct protein-dye interactions, and to alter the orientation of the dye to the protein bound at the central hexanucleotide recognition segment. We previously demonstrated that M.EcoRI binds to two additional bases on each side of this recognition hexanucleotide (53), thus making contacts with 10 base pairs in the specific complex). Fig. 3B reveals that we observed the expected decrease in FRET by increasing the duplex from the original 14 to 17 bp. Nevertheless, in each case the addition of M.EcoRI further decreases the FRET effect, consistent with a protein-induced expansion of the dye to dye distance. The observed protein-induced FRET decreases with the 15-mer (15.6% increase in donor emission) and 17-mer (7.3%) duplexes were anticipated as compared with 14-mer DNA (21.3%), thereby further supporting the lack of any unusual effects.

An independent demonstration that the M.EcoRI-induced changes in FRET result from the formation of the sequence-specific complex is provided by the titration experiments with sinefungin (Fig. 3A). In the absence of either AdoMet or sinefungin, M.EcoRI binds DNA non-specifically at micromolar concentrations (16, 17); under these conditions no changes in energy transfer are observed (Fig. 3A). However, sinefungin addition not only induces the decrease in energy transfer, but it does so in a concentration-dependent fashion (Fig. 3A, inset). The measured dissociation constant (, 100 nM) is 430-fold tighter than the measured for the binary complex (enzyme-sinefungin) (48).

Similar binding and FRET studies with two other proteins and the same DNA sequence further demonstrate that this substrate is well behaved (Fig. 3C). The dimeric EcoRV endonuclease cleaves the recognition site GATATC between the TA dinucleotide step (43). In the presence of inactive divalent cations such as Ca²⁺ the enzyme bends its cognate sequence by 50°, as observed by x-ray crystallography (54) and FRET studies using the fluorescein/rhodamine dye pair (43). This bend occurs toward the major groove and involves no unusual protein-DNA intercalation mechanisms (54). The enzyme also bends some related sequences to the same extent. For example, crystallographic studies and steady state FRET have shown that the EcoRI site, GAATTC, is bent by EcoRV endonuclease.³ Addition of EcoRV endonuclease to the 14-bp duplex containing Alexa 488 and TAMRA dyes (Fig. 3C) results in an increased FRET consistent with a decreased dye to dye distance in the protein-DNA complex. The observed FRET enhancement is qualitatively the same as the recently reported changes with the cognate GATATC sequence (43). These results show that the DNA and dye pairs leading to the decreased FRET effects with M.EcoRI are not a characteristic of the DNA and/or dye pairs, but rather are caused by the particular protein. Furthermore, our results