Article pubs.acs.org/biochemistry
Caulobacter crescentus Cell Cycle-Regulated DNA Methyltransferase Uses a Novel Mechanism for Substrate Recognition Clayton B. Woodcock, Aziz B. Yakubov, and Norbert O. Reich* Department of Chemistry and Biochemistry, University of California, Santa Barbara, California 93106, United States S Supporting Information *
ABSTRACT: Caulobacter crescentus relies on DNA methylation by the cell cycle-regulated methyltransferase (CcrM) in addition to key transcription factors to control the cell cycle and direct cellular differentiation. CcrM is shown here to efficiently methylate its cognate recognition site 5′-GANTC-3′ in single-stranded and hemimethylated double-stranded DNA. We report the Km, kcat, kmethylation, and Kd for single-stranded and hemimethylated substrates, revealing discrimination of 107-fold for noncognate sequences. The enzyme also shows a similar discrimination against single-stranded RNA. Two independent assays clearly show that CcrM is highly processive with single-stranded and hemimethylated DNA. Collectively, the data provide evidence that CcrM and other DNA-modifying enzymes may use a new mechanism to recognize DNA in a key epigenetic process.
T
the end of replication and because of rapid, Lon proteasemediated proteolysis is present for a short ∼10−20 min window,5,12 during which the 4515 GANTC sites are rapidly remethylated. Recent work using single-molecule real-time DNA sequencing has provided an unprecedented level of information about the dynamics of global methylation during this cell cycle.1,4,7 Nearly 100 transcriptional start sites have GANTC sites within 50 bp upstream of the TSS and are regulated by methylation at specific stages of the cell cycle. Interestingly, of the 4515 GANTC sites, 27 GANTC sites remain unmethylated at all stages of the cell cycle.1,4,7 Our motivation for studying CcrM was to understand its mechanism of DNA recognition and methylation; the relative lack of such information1,6,13−15 is surprising given the depth of understanding of another bacterial “orphan” DNA adenine N6methyltransferase, Dam. Also, we sought to clarify two highly conflicting reports about the ability of CcrM to act processively, one of which claimed highly unusual results for a DNA MTase.1,13 We obtained highly pure and active CcrM that provided the basis for reporting several novel findings, as well as reconciling the prior controversies. Importantly, our finding that CcrM displays unprecedented levels of sequence discrimination with both single- and double-stranded DNA suggests the reliance on an entirely novel mechanism of DNA recognition by CcrM and related enzymes.
he discovery of diverse epigenetic mechanisms―including RNA silencing, histone modification, and new forms of DNA modification―has resulted from the study of a wide range of organisms. The Gram-negative, aquatic bacterium Caulobacter crescentus is particularly amenable to the study of prokaryotic cell cycle control because each cell replicates its chromosome only once per cell division.1 The cell cycle-regulated methyltransferase (CcrM) is an orphan Sadenosylmethionine-dependent DNA adenine N6-methyltransferase with homologues that are widespread throughout αproteobacteria, including the human pathogen Brucella abortus.1−4 CcrM, in concert with three transcription factors (DnaA, GcrA, and CtrA), orchestrates the cell cycle-regulated asymmetric cell division in which a single genome gives rise to two distinct and heritable cell types (flagellated swarmer cell and a stalked cell), a clear example of epigenetics.1,5 Together, the four global regulators control 21% of C. crescentus’s transcriptional start sites (TSS). DNA methylation by CcrM directly affects the master regulators ctrA and dnaA and two genes required for cell division, f tsZ and mipZ.5−9 CcrM was thought to be essential for viability in C. crescentus, Agrobacterium tumefaciens, B. abortus, and Rhizobium meliloti, but recent results under low-nutrient conditions indicate it to be dispensable in C. crescentus, presumably because of slow growing conditions, allowing for a complete cell division. In contrast, growths in rich media cause the cells to deform and die.3,4,8−11 At the beginning of the cell cycle in nonreplicating C. crescentus swarmer cells, the chromosome is fully methylated at CcrM recognition sites (5′-GANTC-3′). Following replication initiation, the two replication forks proceed bidirectionally, generating hemimethylated DNA. CcrM is expressed only near © 2017 American Chemical Society
Received: April 24, 2017 Revised: May 30, 2017 Published: June 29, 2017 3913
DOI: 10.1021/acs.biochem.7b00378 Biochemistry 2017, 56, 3913−3922
Article
Biochemistry
■
MATERIALS AND METHODS Cloning and Purification of CcrM. The ccrm (UniProt accession number B8GZ33) gene was cloned from plasmid pLS2468 (kindly provided by L. Shapiro) and into plasmid pET28a using EcoRI and NcoI sites, producing a C-terminal His tag. pET28a-CcrM was then transformed into NEB NiCo21(DE3) Escherichia coli cells; 8 L of cells was grown to an OD of 1.0 at 37 °C, cooled to 23 °C on ice, induced with 1 mM isopropyl β-D-1-thiogalactopyranoside, and grown further at 23 °C for 3 h. Cultures allowed to grow into the fourth and fifth hour will generate more protein, but with a decrease in units of activity. Cells were harvested using a Beckman centrifuge at 5000 rpm (JA-10 rotor) at 4 °C for 20 min. Cell paste was resuspended in lysis buffer [50 mM HEPES (pH 8.0), 300 mM NaCl, 10% glycerol, and 70 mM imidazole] to a final volume of 40 mL and sonicated using a Branson digital sonifier in a water/ice bath. The cell lysate was clarified using a Beckman centrifuge at 11000 rpm (JA-21 rotor) for 2.5 h. The clarified lysate was purified on an AKTA Start FPLC system using a GE 5 mL HisTrap column. Samples were loaded and then washed using the lysis buffer (above) for 9.5 column volumes, and then an isocratic elution step using 50 mM HEPES (pH 8.0), 300 mM NaCl, 10% glycerol, and 160 mM imidazole was performed. Elution fractions were then concentrated using Amicon Ultra 0.5 mL centrifugal filters with a 10 kDa cutoff. Four buffer exchanges were performed to remove the imidazole and place the protein in storage buffer [100 mM HEPES with 300 mM NaCl and ≥50% glycerol, 1 mM DTT, and 1 mM EDTA (pH 8.0)], and aliquots were made and stored at −80 °C. The purification yield and purity were assessed by 12% sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE), and densitometry was performed in ImageJ using a known concentration range of BSA (Thermo Scientific). An 8 L growth produced 140 mg with a purity of >95%. CcrM maintains robust activity while stored at −80 °C, and when it is stored at −20 °C, a decrease in activity was observed over time. We observed a weak dependence of activity on the reaction buffers used, with a decrease in activity in a Tris-HCl buffer and greater activity in a HEPES buffer. CcrM activity depends greatly on the purification protocol used and the storage conditions. We have achieved reproducible units of activity for each purification by inducing the cell cultures for a maximum of 3 h. CcrM does not form inclusion bodies under any known growth conditions. The greatest activity of a preparation can be achieved by growing the cells and purifying on the same day without storing the cell paste at −80 °C. Radiochemical Assays for Determining Km, kcat, and kmethylation. Km, kcat, and single-turnover experiments were conducted in methylation reaction buffer (MRB) consisting of 100 mM HEPES (pH 8.0) with 20 mM NaCl, 1 mM DTT, 1 mM EDTA, 2 mg/mL BSA, and saturating cofactor AdoMet (15 μM), in a 1:10 ratio (radiolabeled:unlabeled) in triplicate at 23 °C. AdoMet stocks were made using 32 mM AdoMet supplied by NEB in 10 mM H2SO4 and [[3H]CH3 − 1 mCi (82.7 mCi/mmol)] AdoMet supplied by PerkinElmer at a final concentration of 50 μM in a 1:10 ratio. DNA substrates labeled with a 5′ 6-fluorescein tag (FAM) were supplied by IDT, and N6-methyladenine substrates were purchased from the Keck Oligo facility at Yale University (New Haven, CT). Activity was measured by the incorporation of tritiated methyl groups from AdoMet onto the DNA product. Reactions were initiated by
the addition of substrate into a mix of MRB and enzyme, and time points were taken by spotting 5 μL of reaction mix onto DE81 anion exchange filter paper (supplied by GE) [activity was observed to be independent of the order of addition of substrates (data not shown)]. Samples were allowed to dry at room temperature and then washed with three consecutive 400 mL volumes of 50 mM KH2PO4 followed by one 400 mL 70% EtOH wash and an additional 400 mL 100% EtOH wash step, and a final dehydration step using 400 mL of anhydrous ethyl ether (all washes were conducted at room temperature on a benchtop shaker at 90 rpm). Up to 50 filters could be processed in this fashion simultaneously. The filter paper was then allowed to dry at room temperature before being added to scintillation vials containing 3 mL of BioSafe II scintillation fluid. Data were collected using a Beckman Coulter LS-6500 scintillation counter with units of disintegrations per minute (dpm) and then converted to molarity. Background readings were subtracted from all points, and the data were fit using the GraphPad Prism 5 standard Michaelis−Menten equation and one-phase decay for single-turnover reactions. EMSA for Obtaining Thermodynamic Constants. An EMSA (electrophoretic mobility shift assay) was used to determine the dissociation constants (Kd) for substrates, conducted in MRB on ice with FAM-tagged DNA supplied by IDT. The assay also included 50 μM sinefungin, purchased from Enzo Life Sciences. The binding assay was allowed to incubate at 4 °C for 30 min. Five microliters of the assay was mixed with 5 μL of 50% glycerol and 5 μL loaded onto a 12% (1:75) Native PAGE gel. The running buffer contained Trisborate EDTA and was run for 90 min in the cold room. Gels were scanned using a GE Typhoon Trio instrument, and densitometry was conducted using the ImageQuant TL program. The disappearance of the band was analyzed using GraphPad Prism 5 and fit to a quadratic equation for a one-site binding model. Processivity Assays Using Radiochemical and Endonuclease Challenge Assays. CcrM’s processivity (the ability for the enzyme to methylate multiple sites before dissociation) was assessed by two distinct assays. The first assay (Figure 3B,C) used HincII endonuclease (NEB) to digest the methylated product, as previously described.13 Methylation reactions by CcrM were conducted in MRB with saturating unlabeled AdoMet (15 μM) and initiated with the addition of enzyme. Five microliters of sample at each time point was removed and added to a new tube of 5 μL of Millipore (nanopure) water and heat-inactivated at 95 °C for 2 min. The protocol for ssDNA is identical except the secondary tube contained 5 μL of the complementary sequence at concentrations identical to that of the top for annealing and then heattreated at 95 °C for 2 min and allowed to cool to 23 °C. Both double-stranded and single-stranded DNA (dsDNA and ssDNA, respectively) products were treated with 1 μL of HincII per 10 μL reaction along with the HincII restriction buffer supplied by NEB. Five microliters of the subsequent reaction mixtures was mixed with 5 μL of 50% glycerol, loaded on a 12% Native PAGE (29:1) gel, and allowed to run for 45 min at 4 °C in Tris-Borate EDTA. Rate constants for product formation under initial velocity conditions were obtained using the ImageQuant TL program to perform densitometry on bands associated with DM (doubly methylated, fully protected from HincII digestion), SM (singly methylated, only one site protected), and UM (both sites unmethylated) forms. Equation 1 was used to approximate the Fp.14 3914
DOI: 10.1021/acs.biochem.7b00378 Biochemistry 2017, 56, 3913−3922
Article
Biochemistry
Figure 1. CcrM kinetic parameters for single- and double-stranded DNA reveal robust dual activity. (A) SDS−PAGE of CcrM purified to >95% determined by densitometry. Numeric values with in of kilodaltons (CcrM, 38 kDa). (B) Michaelis−Menten kinetics on a single-stranded substrate [ssDNA, GACTC, 60 nucleotides in length (see Table 1A for the sequence); blue circles] and a double-stranded substrate [dsDNA, GACTC, 60 bp in length (see Table 1A for the sequence); red squares]: CcrM (1.0 nM), substrates ranging from 3 to 60 nM, and saturating AdoMet at 15 μM. (C) Single-turnover reaction on ssDNA (blue circles) and dsDNA (red squares): 150 nM CcrM and 100 nM substrate. (D) Initial velocity studies of ssDNA (blue circles) and dsDNA (red squares): 100 nM CcrM, 3 μM substrate, and saturating AdoMet at 15 μM. (E) Single turnover on the DNA hairpin (control). The hairpin is 50 nucleotides in length; the stem is 11 bp (composed of nine GC and two AT base pairs), and the GACTC site (purple circles) is located in the middle of the hairpin loop: 150 nM CcrM and 100 nM substrate. Hairpin recognition site specificity was tested by replacing the canonical site with GACTA (red squares): 1.5 μM CcrM and 1 μM substrate. All reactions were conducted in triplicate at 23 °C in 100 mM HEPES (pH 8.0) with 20 mM NaCl and 15 μM saturating cofactor AdoMet.
Fp = DM/(2SM + DM) = Vdm/(2Vsm + Vdm)
1D).15,16 Because of prior work demonstrating CcrM can dimerize (Kd = 400 nM),17 we characterized CcrM’s kinetics from 50 to 2500 nM (SI Figure 1A) on our primary substrate (above) and found a linear relationship between velocity and enzyme concentration. Direct measurement of the methylation constant for dsDNA (Figure 1C; kmethylation, 0.62 ± 0.08 min−1) confirms that methylation or a conformational step prior to methylation limits the turnover of the enzyme. The kmethylation constant is particularly slow when compared to those of other DNA adenine N6 MTases such as EcoRI (2460 min−1) and T4 Dam (36 min−1).15,16 CcrM shows a surprising level of sequence discrimination (Figure 2C and Table 1E, dsDNA).18,19 Single-base pair changes in the recognition sequence [e.g., 5′-GACTC-3′ to 5′-AACTC-3′, which does not include the target A (italicized)] result in ≤106-fold decreases in the apparent kmethylation and a 1.9 × 107-fold change in specificity (kmethylation/Kd). This level of sequence discrimination is at least 103−104-fold greater than what is typically observed with other DNA MTases.15 We were intrigued by the recent report that some 27 sites remain largely hemimethylated by CcrM in vivo; these sites have a consensus sequence of 5′-AG(G/C)gaGtcATT-3′.20 We therefore used hemimethylated substrates and systematically varied the flanking sequence. Our results show that the enzyme exhibits an only 8-fold discrimination against these sites based on kmethylation/KdDNA (Table 1B,C). This suggests that other factors (e.g., regulatory DNA binding proteins and higher-order nucleic acid structures) must contribute to these in vivo
(1)
The second assay (Figure 3D) used a radiochemical approach to assess processivity on ssDNA. Reactions were conducted as described in the radiochemical assay section. Four DNA substrates identical in length with varying numbers of GANTC sites1−4 were methylated with CcrM under initial velocity conditions (3 μM substrate, 100 nM CcrM, and
