Rapid Time Scale Analysis of T4 DNA Ligase–DNA Binding

Subscriber access provided by University of Newcastle, Australia

Article

Rapid Timescale Binding Analysis of T4 DNA Ligase-DNA Binding Robert John Bauer, Thomas J. Jurkiw, Thomas C. Evans, and Gregory J. S. Lohman Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.6b01261 • Publication Date (Web): 06 Feb 2017 Downloaded from http://pubs.acs.org on February 12, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Rapid Timescale Analysis of T4 DNA Ligase-DNA Binding 1

2

1

1†

Robert J. Bauer , Thomas J. Jurkiw , Thomas C. Evans, Jr. *, and Gregory J. S. Lohman 1 2

DNA Enzymes Division, New England BioLabs, Inc., Ipswich, MA, 01938-2723, USA University of Michigan Medical School, Ann Arbor, Michigan 48109-0600, USA

* To whom correspondence should be addressed. DNA Enzymes Division, New England BioLabs, Inc., Ipswich, MA, 01938-2723, USA. Tel: 978-927-5054; Fax: 978-921-1350; Email: [email protected] †

To whom correspondence should be addressed. DNA Enzymes Division, New England BioLabs, Inc., Ipswich, MA, 01938-2723, USA. Tel: 978-998-7916; Fax: 978-921-1350; Email: [email protected] · Key Words: DNA ligase, enzyme kinetics, DNA binding, nick-sealing, ligation, DNA replication

Abbreviations: FAM – 6-Carboxyfluorescein, nDNA – Nicked Substrate DNA, PBCV-1 – Paramecium bursaria chlorella virus, RDS- Rate-Determining Step, RFU – Relative Fluorescence Units, RQF- Rapid Quench Flow

Abstract DNA ligases, essential to both in vivo genome integrity and in vitro molecular biology, catalyze phosphodiester bond formation between adjacent 3’-OH and 5’-phosphorylated termini in dsDNA. This reaction requires enzyme self-adenylylation, using ATP or NAD+ as a cofactor, and AMP release concomitant with phosphodiester bond formation. In this study, we present the first fast-timescale binding kinetics of T4 DNA ligase to both nicked substrate DNA (nDNA) and product-equivalent non-nicked dsDNA, as well as binding and release kinetics of AMP. The described assays utilized a fluorescein-dT labeled DNA substrate as a reporter for ligase—DNA interactions via stopped-flow fluorescence spectroscopy. The analysis revealed that binding to nDNA by the active adenylylated ligase occurs in two steps, an initial rapid association equilibrium followed by a transition to a second bound state prior to catalysis. Furthermore, the ligase binds and dissociates from nicked and non-substrate dsDNA rapidly with initial association affinities on the order of 100 nM regardless of enzyme adenylylation state. DNA

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

binding occurs through a two-step mechanism in all cases, confirming prior proposals of transient binding followed by a transition to a productive ligase—nDNA (Lig.nDNA) conformation but suggesting that weaker nonproductive “closed” complexes are formed as well. These observations demonstrate the mechanistic underpinnings of competitive inhibition by rapid binding of non-substrate DNA, and of substrate inhibition by blocking of the self-adenylylation reaction through nick binding by deadenylylated ligase. Our analysis further reveals that product release is not the rate-determining step in turnover.

DNA ligases are enzymes responsible for sealing breaks in the phosphodiester backbone of DNA, essential for the in vivo maintenance of genome integrity, and have found widespread use in in vitro biotechnology applications such as cloning, gene assembly and DNA library assembly. These enzymes catalyze the formation of a phosphodiester bond between adjacent 3′-hydroxyl and 5′-phosphate termini at the site of a single strand break (nick).1, 2 Some DNA ligases, most notably the DNA ligase from the T4 bacteriophage, can also join two dsDNA fragments with short complementary ssDNA overhangs or blunt-ended termini.1-4 DNA ligases are broadly divided into two classes: those dependent on ATP for self-adenylylation, found in eukaryotes, viruses, and some bacteria and those dependent on NAD+, found in bacteria and some eukaryotic viruses.5

Nick-sealing by DNA ligases proceeds through a ping-pong mechanism involving two substrates and three highly conserved nucleotidyl-transfer chemical steps.6-10 The reaction pathway begins with nucleophilic attack by an active site lysine residue on the α-phosphate group of either ATP or NAD+, forming an adenylylated ligase intermediate and releasing PPi or β-NMN, respectively. The adenylylated ligase then binds to a 5′-phosphorylated nick site in dsDNA, whereupon a proposed structural rearrangement, based on crystal structure data, transitions the ligase from

ACS Paragon Plus Environment

Page 2 of 42

Page 3 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

an open form to a stably bound closed form.11-15 Upon closure of the enzyme on the nick, the adenylyl group is transferred from the enzyme’s active site lysine to the 5′-phosphate, producing a 5′ to 5′ adenosine diphosphate (AppDNA) intermediate.11 Nick-sealing then occurs through a third nucleophilic attack event by the 3′-OH on the α-phosphate of the AppDNA, resulting in the formation of a phosphodiester bond and AMP. Dissociation of both products from the ligase is presumably required before reaction with another ATP/NAD+ cofactor can occur.1 The kinetics of the ligation chemistry steps (self-adenylylation, adenylyl transfer, and phosphodiester bond formation) are fairly well understood and have been described for ligases from multiple organisms.8, 16-21 The rates of adenylyl transfer and phosphodiester bond formation have been measured previously for various DNA ligases including: Paramecium bursaria chlorella virus (PBCV-1) where kpA transfer = 2.35 ± 0.14 s−1 and kseal = 24.5 ± 1.1 s−1, human DNA ligase I where kpA transfer = 2.6 ± 0.6 s-1 and kseal = 12 ± 2 s-1 and T4 DNA ligase where kpA

transfer

=

5.3 ± 0.4 s-1 and kseal = 38 ± 5 s-1.8, 17, 18 For PBCV-1 DNA ligase the rate-limiting chemistry step was proposed to be adenylyl transfer, while for human DNA ligase I self-adenylylation was proposed to be rate limiting at saturating Mg2+ concentrations.8, 18 While no direct measure of self-adenylylation has been performed on PBCV-1 DNA ligase or human DNA ligase I, it has been characterized for T4 DNA ligase by Cherepanov et al., where an ATP on-rate of 0.87 ± 0.011 µM-1 s-1 and off-rate < 1 s-1 in the presence of 5 mM Mg2+ were observed, indicating a KD for ATP binding to T4 DNA ligase of ≤ 1 µM

22, 23

Under these same conditions the rate of

conversion of ligase—ATP to adenylylated ligase was determined to be 12.8 ± 0.1 s-1.22 For each of these ligases the measured rates of chemistry are significantly faster than their steady state turnover rates (PBCV-1 DNA ligase: kcat = 0.73 ± 0.02 s−1, human DNA ligase I: kcat = 0.74 ± 0.09 s−1, and T4 ligase kcat = 0.4 ± 0.1 s−1); for T4 DNA ligase, all three chemical steps have been directly measured and none are slow enough to explain the observed turnover rates.8, 16, 17, 24

Further, a burst phase was observed in pre-steady state nick-ligation experiments using T4

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

DNA ligase, leading to the proposal that release of sealed product, or a post-product-release conformational change, was the overall rate-determining step (RDS).17

In our prior study, it was shown that DNA lacking a nick can compete with nicked dsDNA for binding to T4 DNA ligase, and serves as a competitive inhibitor for both the self-adenylylation and nick-sealing steps of the ligation pathway.16 This inhibition was proposed as evidence for the first step in a two-step binding mechanism utilized by DNA ligases in order to quickly search large stretches of dsDNA and identify their nicked target sites. Two-step binding by T4 DNA ligase was first discussed in the mechanistic depiction of substrate recognition by Rossi et al, where the first step in their model was nick-scanning via a transient association of the adenylylated form of the ligase with duplex DNA. A later model based on Human DNA ligase I and PBCV-1 DNA ligase crystal structures provided evidence that nick recognition involved encircling and bending of the dsDNA at the nick junction by the ligase, accompanied by a transition of standard B-form DNA to A-form for the two nucleotides on either side of the nick.11, 12, 14, 15

A nick site possesses the necessary flexibility allowing for the bending transition and

stable formation of the ligase—AMP—nDNA (LigpA.nDNA) complex, while dsDNA will not deform as readily.25 Both of these models imply an initial transient interaction with dsDNA by the ligase in order to allow for the interrogation of the DNA for breaks in the phosphodiester backbone, followed by transition to a closed form resulting in stable binding only in the presence of a nick.

In the current study, we utilize rapid time-scale analysis techniques including rapid quench and stopped-flow fluorescence spectroscopy to probe the rates of the interactions between T4 DNA ligase with all its substrates and products: nicked dsDNA (nDNA), non-nicked dsDNA, ATP and AMP. Through fitting of binding and dissociation experiments by simulation, it is shown that T4 DNA ligase binds both nDNA and non-nicked dsDNA through a two-step mechanism, likely the association/closing mechanism previously proposed by Rossi and Shuman. However, closing

ACS Paragon Plus Environment

Page 4 of 42

Page 5 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

occurs regardless of enzyme adenylylation state or the presence of a nick, albeit with different net binding constant (KD,net) and different equilibrium constants for the conformational transition. Additionally, the kinetics of binding and release of AMP was also explored, and the rates of the chemistry steps were re-confirmed. Based on this analysis, we propose that the rate limiting step is not a chemistry step or product release, but rather a conformational change not directly observable by current experimental methods.

Materials and Methods Materials T4 DNA ligase, 1 M DTT, and 10 mM ATP were obtained from New England Biolabs (Ipswich, MA). Deadenylylated T4 DNA ligase ( 100 µM ATP, < 100 nM AMP depending on substrate and Mg2+ concentration).

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 42

Figure 6: Global Fit of T4 DNA Ligase AMP Binding and Release Global fitting of the kinetics of T4 DNA AMP binding and release were generated with KinTek Global Kinetic Explorer using χ2 threshold of 1.05 and the equation above. All reported kinetic parameters are reported as ranges determined through FitSpace error analysis of the respective simulation.29 (A) Stopped-flow analysis of T4 DNA ligase AMP binding through monitoring of fluorescence of an active site tryptophan (B) Stopped-flow analysis of T4 DNA ligase AMP release by dilution. A 30 µM pre mixed Lig—AMP complex was shot against buffer only in a 1:1 mixing reaction. The fluorescence of the active site tryptophan was measured as AMP dissociated. The global fit determined the following rates konAMP = 0.37 - 0.4 µM-1 s-1 and koffAMP = 17.6 - 19.2 s-1 corresponding to KD = 44 - 53 µM.

Discussion In this study, we have developed a new fluorescence-based assay system to probe the fast timescale binding of DNA by T4 DNA ligase. This assay has been applied to measure the rates and mechanism of DNA binding between the ligase and both substrate and product analog DNA. Fitting multiple experiments by simulation, shows that DNA binding by T4 DNA ligase appears to minimally require a two-step binding mechanism to account for the data. Through these experiments, along with product detected RQF and ATP/AMP binding experiments, we have been able to assign rates to many steps in the ligation pathway, both in single turnover ligation and in the formation of many nonproductive complexes between ligase and DNA (Figure 7).

ACS Paragon Plus Environment

Page 27 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 7: T4 DNA Ligation Pathway Full ligation pathway for T4 DNA ligase. The catalytic nick sealing pathway (black) begins with the self-adenylylation of deadenylylated T4 DNA ligase in the open conformation (Ligaseo) and includes rate parameters for all chemistry and DNA binding steps including rates of transition to the closed form of the enzyme (Ligasec).The experiments performed here were unable to determine whether release of the AMP and sealed dsDNA product occur in a random or ordered process, these steps are indicated by the red box. Nonproductive DNA binding pathways (shown in red text) include the rates derived from the two-step binding models applied here. Any rates which were un-constrained by the models and set to be non-reversible are indicated by N.D. 0 s-1.

Binding of nDNA by T4 DNA ligase appears to follow a two-step mechanism, with initial rapid equilibrium association followed by transition to a second, more stable form. The biochemical observation of two-step binding in the nick sealing reaction is consistent with prior hypothesis for DNA ligase substrate association. The observation of two-step binding is also supported by crystal structure data showing that DNA ligases undergo a structural transition from an open to closed form upon complex formation, completely encircling their bound nDNA substrates.11, 12, 14, 16

Here, we suggest that the structural transition during substrate binding observed by x-ray

crystallography is the best explanation for the observed two-step binding, with initial rapid equilibrium association followed by a kinetically relevant transition to this closed form. It is known from crystal structures of human DNA ligase I and chlorella virus DNA ligase that stable

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 42

binding of the ligase on its nicked substrate induces in a 13° bent A-form like DNA conformation. The unstacking of the bases at the ligation junction and interaction of the ligase with the base at the 3’-end of the upstream fragment would be expected to produce a dramatic change in environment for the fluorophore in this region, and is seen in this experiment as a drop of ~ 4 relative fluorescence units (RFU). In the binding of nonproductive complexes, including both deadenylylated ligase binding nDNA and LigpA binding dsDNA, a much smaller change is seen, ~ 0.75 RFU, despite the kinetic analysis indicating these interactions are also two-step. This smaller fluorescence drop is likely a result of binding in a different “closed” conformation, without the kinking and partial unwinding of the DNA that occurs upon productive binding, and consequently only a minor effect on fluorophore environment vs the “open” conformation.

In single turnover ligation, the steep fluorescence decrease upon productive binding is followed by a rapid increase in fluorescence, returning fluorescence intensity to near baseline for lower ligase concentrations (Figure 2). This large fluorescence recovery observed upon sealing of the substrate would be expected as the A-form bent DNA structure transitions back to standard Bform, unbent duplex DNA. Notably the fluorescence does not return fully to the initial reading, and clearly plateaus at a lower RFU value as ligase concentration increases. At saturation, this difference from baseline is ~ 0.8 RFU, quite similar to the drop upon binding of non-nicked dsDNA by LigpA.

Further, this drop relative to baseline increases with increasing ligase

concentration, indicating this final state does not represent free DNA released after turnover, but rather free DNA in equilibrium with a LigpA·DNA bound state resulting from rapid rebinding of the product DNA by excess adenylylated ligase in solution. During turnover the slowest chemical step (self-adenylylation, 7.0 - 8.6 s-1) and the off-rate of AMP (17.6-19.2 s-1) are much faster than the kcat seen during turnover (~0.6 s-1). However, we

ACS Paragon Plus Environment

Page 29 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

observed that the off-rate of deadenylyated ligase from dsDNA (analogous to post chemistry ligase plus ligated DNA, but lacking non-covalently bound AMP) was slow compared to all chemistry steps (koff ~ 2.24 s-1). Similarly, upon closing of LigpA on nDNA there appears to be a high commitment to catalysis; the global fit of the stopped-flow nDNA binding/sealing reaction (Figure 2C) suggests reopening is very slow (k-2 ≤ 2.62) relative to the forward rate of chemistry. Post chemistry, we would thus expect the ligase to be in a “closed” conformation encircling the DNA, and that transition of the ligase to an open form post chemistry would be required before dissociation of the AMP and dsDNA products, and that this step could be slow enough to be rate-determining, were it occurring with a similar rate. However, single-turnover stopped-flow nick-sealing experiments lacked any clear phase after chemistry suggestive of slow product release.17 We considered the possibility that the endpoint final ligase-DNA binding equilibrium caused by binding of released product by excess enzyme may be obscuring any slow release phase, particularly at high ligase concentration. Any observable product release fluorescence change would be small (