Kinetic characterization of single strand break. ligation in duplex DNA by T4 DNA Ligase

Transcription

1 Kinetic characterization of single strand break ligation in duplex DNA by T4 DNA Ligase Gregory J. S. Lohman, Lixin Chen, and Thomas C. Evans, Jr.* DNA Enzymes Division, New England Biolabs, Ipswich, MA, SUPPORTING INFORMATION 1

2 Figure S1. Determination of k cat and k cat /K M for T4 DNA Ligase and nicked substrates. Reaction of (A-C) 1 nm T4 DNA ligase with 1 nm ( ), 2 nm ( ), 5 nm ( ), 10 nm ( ), 20 nm ( ) and 50 nm ( ) substrate 1 in standard assay buffer at 16 C and (D-F) 1 nm T4 DNA ligase (<5% adenylylated) with 1 nm ( ), 2 nm ( ), 5 nm ( ), 10 nm ( ), 20 nm ( ) and 50 nm ( ) substrate 1A in standard assay buffer without ATP at 16 C. The plotted points are corrected to account for non-reactive starting material as described in the experimental. The dashed lines are the fit by simulation of all six concentrations to a single Michaelis- Menten model in KinTek Global Kinetic Explorer. Fitting for the AppDNA substrate, plots D, E and F, was limited to the first five minutes due to the apparent loss of activity in the 50 nm and 20 nm plots seen at longer times. The displayed plots correspond to kinetic parameters of (A) k cat = 0.34 ± 0.02 s -1, k cat /K M = 210 ± 30 µm -1 s -1, (B) k cat = 0.51 ± 0.03 s -1, k cat /K M = 190 ± 20 µm -1 s -1, (C) k cat = 0.33 ± 0.02 s -1, k cat /K M = 90 ± 7 µm -1 s -1, (D) k cat = 0.62 ± 0.04 s -1, k cat /K M = 250 ± 30 µm -1 s -1, (E) k cat = 0.53 ± 0.04 s -1, k cat /K M = 225 ± 30 µm -1 s -1, and (F) k cat = 0.46 ± 0.04 s -1, k cat /K M = 330 ± 70 µm -1 s -1. B = CE peak area of nicked dsdna; X = CE peak area of AppDNA; P = CE peak area of ligated product. 2

3 Figure S2. Fit for single turnover of 100 nm substrate 1 with 500 nm T4 DNA ligase (<5% adenylylated) and 1 mm ATP. The data points shown are pdna ( ), nicked AppDNA ( ), and ligated product ( ). The data points are the average of three repeats with the error bars one standard deviation. The dashed lines represent a fit setting k -2 = k -3 = 0, DNA binding to a K D of 1.5 nm with diffusion limited on rate, not including product release, and fixing the rates of Step 1 binding and chemistry to literature values (E + A EA F + PPi; k +ATP = 1 µm -1 s -1, k-atp = 0.1 s -1, k 1 = 13 s -1, k -1 = 0.4 µm -1 s -1 ) and constraining k 2 and k 3 by simultaneous fitting with the single turnover data from Figure 4A. Determined constants are k 2 = 5.4 ± 0.25 s -1, k 3 = 34.8 ± 3.5 s -1 and the residuals show the difference between the simulation and the data for the amount of ligated product at each time point. The dotted lines show fits leaving all constants as they were but setting k 1 = 0.5 s -1, a rate slow enough to account for steady-state turnover rates. These curves show the expected single turnover results if the rate of enzyme readenylylation was the ratelimiting step under turnover conditions. 3

4 Figure S3. RQF assay of enzyme self-adenylylation with α- 32 P-ATP. The reaction contained 2.5 µm T4 DNA ligase (<5% adenylylated) and 1 mm ATP with specific activity of 200 µci 32 P/µmol ATP. The data points are the average of three repeats with the error bars one standard deviation. The dashed lines represent a fit through simulation to the model E EA F where E is apo-enzyme, F is adenylylated enzyme, and A is ATP. The step 1 constant (k 1 ) was allowed to vary, k -1 was set to 0, and the binding rates set to 1 µm -1 s -1 with a reverse rate of 0.2 s -1 based on literature values. The best fit simulation gave k 1 = 6 ± 1 s -1. The residual plot shows the difference between the fit and the data at each time point. 4

5 Figure S4. Fits allowing reverse reactions for the ligation of substrate 1 under single turnover conditions. Reaction of 100 nm substrate 1 with 500 nm T4 DNA ligase (>95% adenylylated). The data points are the same as those shown in Figure 4A for nicked pdna ( ), nicked AppDNA ( ), and ligated product ( ). The data points are the average of three repeats with the error bars one standard deviation. The dashed lines represent (A) a fit allowing the reverse of Step 2 (k -2 ) and Step 3 (k -3 ) to vary (determined constants are k 2 = 9.4 ± 2 s -1, k -2 = 30 ± 15 s -1, k 3 = 46 ± 6 s -1, k -3 = 1.7 ± 0.8 s -1 ) and (B) a fit allowing the reverse of Step 3 to vary but locking k -2 = 0 (determined constants are k 2 = 5.4 ± 0.4 s -1, k 3 = 42 ± 7 s -1, k -3 = 0.8 ± 0.9 s -1 ). The residual plots show the difference between the fit and the data for the ligated product at each time point. 5

6 Figure S5. Simultaneous fitting of the single turnover ligation of substrate 1 and substrate 1A. Reactions were of 100 nm substrate with 500 nm T4 DNA ligase, >95% adenylylated for substrate 1 (left) and <5% adenylylated for substrate 1A (right). The data points are the same as those shown in Figure 4 for nicked pdna ( ), nicked AppDNA ( ), and ligated product ( ), and are the average of three repeats with the error bars one standard deviation. The dashed lines represent simulations assuming fast and tight initial binding of the substrate (k +DNA = 1000 µm -1 s -1, k -DNA = 2 s -1 ) for binding both pdna and AppDNA, irreversible chemistry (k -2 = k -3 = 0), and with one additional step not included in the model presented in the Experimental: (A) A two-step binding model, with initial complexes (GB and GX) irreversibly transitioning to the active complexes (FA and EX); (B) The possible initial, reversible formation of a nonproductive bound complex (GB and GX), which must disassociate completely before the enzyme can bind to form the productive complexes (FB and EX). E = deadenylylated enzyme; F = adenylylated enzyme; G = enzyme in a nonproductive conformation; B = nicked dsdna phosphorylated at the nick; X = nicked dsdna adenylylated at the nick. The constants used in the simulations are: (A) k 2 >300 s -1, k 3 = 40 ± 5 s - 1, GB FB = 5.4 ± 0.4 s -1, GX EX = 4.9 ± 0.6 s -1 ; (B) k 2 = 9 ± 2 s -1, k 3 = 46 ± 2 s -1, GB on/off = 330 ± 220 µm -1 s -1 /2.5 ± 2.1 s -1, GX on/off = 3800 ± 700 µm -1 s -1 /17 ± 4 s -1. Note that the binding constants in B are not well constrained by the data, and represent only one possible fit with this model. 6