Human NEIL3 is mainly a monofunctional DNA glycosylase removing spiroiminohydantoin and guanidinohydantoin

Abstract

Base excision repair is the major pathway for removal of oxidative DNA base damage. This pathway is initiated by DNA glycosylases, which recognize and excise damaged bases from DNA. In this work, we have purified the glycosylase domain (GD) of human DNA glycosylase NEIL3. The substrate specificity has been characterized and we have elucidated the catalytic mechanisms. GD NEIL3 excised the hydantoin lesions spiroiminodihydantoin (Sp) and guanidinohydantoin (Gh) in single-stranded (ss) and double-stranded (ds) DNA efficiently. NEIL3 also removed 5-hydroxy-2’-deoxycytidine (5OHC) and 5-hydroxy-2’-deoxyuridine (5OHU) in ssDNA, but less efficiently than hydantoins. Unlike NEIL1 and NEIL2, which possess a β,δ-elimination activity, NEIL3 mainly incised damaged DNA by β-elimination. Further, the base excision and strand incision activities of NEIL3 exhibited a non-concerted action, indicating that NEIL3 mainly operate as a monofunctional DNA glycosylase. The site-specific NEIL3 mutant V2P, however, showed a concerted action, suggesting that the N-terminal amino group in Val2 is critical for the monofunctional modus. Finally, we demonstrated that residue Lys81 is essential for catalysis.

1. INTRODUCTION

The base excision repair pathway (BER) is the major pathway for repair of DNA damage caused by reactive oxygen species (ROS) [1, 2]. BER is initiated by a damage-specific DNA-glycosylase that excises the damaged base by cleavage of the N-glycosylic bond, creating an abasic (AP) site. The DNA glycosylases are divided into two classes, based on their mechanism; monofunctional or bifunctional. A monofunctional glycosylase recruits an AP endonuclease (APE1) to create a nick 5’ to the baseless site. A bifunctional glycosylase has an intrinsic lyase activity and incises the DNA strand 3’ to the AP site by either β-elimination or β,δ-elimination.

Several laboratories, including our own, have tried extensively to characterize the DNA glycosylase Neil3 biochemically, but hurdles like low endogenous expression, cellular toxicity upon over-expression, and instability and precipitation of the recombinant protein have made such studies particularly challenging [3-5]. For a long time, the only evidence for Neil33 being an active DNA glycosylase was a report of weak activity toward 2,6-diamino-4-hydroxy-5-formamidopyrimidine (faPy) lesions when expressed in insect cells [6]. Two groups have reported purification of enzymatically active Neil3. Takao and coworkers found a weak lyase activity for human NEIL3 on AP sites in ssDNA [7]. Liu and colleagues reported that mouse Neil3 and human NEIL3 removes oxidized purines and incises DNA by uncoupled DNA glycosylase/AP lyase activities, where base release is more efficient than strand incision [8, 9]. The structure of the glycosylase domain of mouse Neil3 was recently published and confirms an overall fold similar to the other Nei/Fpg proteins, but with some distinct features explaining the preference for lesions in single-stranded DNA [10].

Here we characterize the biochemical properties of the core glycosylase domain of human NEIL3, demonstrating that the enzyme mainly acts as a monofunctional DNA glycosylase with high affinity for the hydantoins Gh and Sp. Site-specific mutagenesis experiments show that Val2 is important for the non-concerted action of NEIL3, and Lys81 is essential for the catalytic activity.

2. MATERIALS AND METHODS

2.1 Expression vectors and mutagenesis

pETDuet-1-GD-NEIL3 expressing the core glycosylase domain (amino acids 1-301) of human NEIL3 with a C-terminal 6x His tag were generated as previously described [5].

Mutations were generated using QuickChange site-directed mutagenesis (Stratagene) with pETDuet-1-GD-NEIL3 as a template and primers containing base substitutions (V2P and K81A). Primers are listed in the supplementary information (Table S1). All constructs were sequenced.

2.2 Expression and purification of recombinant NEIL3 proteins

The expression constructs were transformed into E.coli BL21 CodonPlus (DE3)-RIL (Stratagene). Cells were grown at 37 °C in LB to an OD₆₀₀ of 0.5. The temperature was reduced to 16 °C and the culture was induced with 1 mM IPTG and grown for another 20 h. Cells were harvested by centrifugation, resuspended in buffer A (50 mM NaH₂PO₄ pH 8, 300 mM NaCl, 10 mM βME, 10 mM Imidazole) and disrupted by sonication. The protein extract was applied onto a Ni–NTA resin (Qiagen) and eluted with buffer A containing 300 mM imidazole. Fractions with GD NEIL3 were applied to a Superdex75 size-exclusion chromatography equilibrated with 100 mM Tris pH 8, 100 mM NaCl, 10 mM βME. Fractions were analyzed by SDS-PAGE and western blotting with His-tag antibody (Qiagen).

2.3 Incision and cyanoborohydride-trapping assays

Synthesis of the hydantoin-containing oligonucleotides was achieved following the method previously described [11]. DNA oligonucleotides containing a base lesion were radiolabeled with [γ-³²P] ATP (Perkin Elmer) at the 5’ end using T4 PNK (New England Biolabs) as described elsewhere [12]. The AP-substrate was generated by incubating an oligonucleotide containing a single uracil residue with uracil DNA glycosylase (UDG, New England Biolabs) at 37° C for 10 min, followed by inactivation of UDG at 80° C for 5 min.

NEIL3 incision activities were assayed in a 10 μL reaction mixture containing 10 fmol radiolabeled substrate, 50 mM MOPS pH 7.5, 1 mM EDTA, 5 % glycerol, 1mM DTT and purified enzyme as indicated at 37 °C. Standard reactions were performed for 30 min and single turnover experiments for 10 sec - 20 min. The enzymatic reactions were terminated by adding 10 μL formamide stop solution (80 % formamide, 10 mM EDTA, bromophenol blue and xylene cyanol) and analyzed on 20 % polyacrylamide/7 M urea gel followed by phosphorimaging. To analyze whether the glycosylase and AP lyase reaction steps were coupled, 10 μL reaction mixtures were quenched either by adding 2 μL 0.5 M NaOH, heating at 94 °C for 20 min followed by adding 2 μL 0.5 M HCl (for measurements of glycosylase activity) or by adding 10 μL formamide stop solution (for measurements of the coupled reaction). Incision activity was quantified using ImageQuant v2003.02 software.

Cyanoborohydride-trapping assays were performed by addition of 50 mM NaBH₃CN to the same conditions as the incision assays. Samples were separated by SDS-PAGE and analyzed by phosphorimaging. All substrate sequences used in these experiments are listed in the supplementary information (Table S2).

3. RESULTS AND DISCUSSION

3.1 NEIL3 in metazoan genomes

In independent studies, Kathe and coworkers [13] and Grin and Zharkov [14] have recently uncovered the phylogeny of the Fpg/Nei-like enzymes in animals. These enzymes appear to have evolved through two subsequent duplication events giving rise to the two subfamilies Neil1 and Neil2/Neil3. While Neil1 is found in for example cnidarians such as Nematostella vectensis and in the basal metazoan Trichoplax adhaerens in addition to chordates, Neil2/Neil3 appear to be limited to the bilaterians. Neil2 is present in chordates and in the echinoderm sea urchin Strongylocentrotus purpuratus while Neil3 has only been found in chordates in previous studies. In order to detect evolutionarily conserved residues and segments in Neil3 that are likely to be important for function we collected full-length Neil3 sequences from eleven tetrapods, three fish species and the cephalochordate amphioxus. Available sequence data suggests that Neil3 has been lost in the third chordate subphylum, the urochordates. A multiple sequence alignment of the chordate Neil3 sequences (Supplementary Fig. S1) shows absolutely conserved Val2 and Lys81 and a highly conserved N-terminus and H2TH motif. A poorly conserved insertion in mammalian Neil3 sequences corresponds approximately to residues 35-58 in human NEIL3. Following the Fpg/Nei-like DNA glycosylase domains (human NEIL3 residues 1-282) is a short structurally disordered segment, the RanBP2 type zinc finger and a conserved segment that appears to be unique to vertebrate Neil3 (residues 362-398). The C-terminus comprises a long structurally disordered segment (residues 399-506) and the two GRF type zinc fingers with unknown function.

Recently released sequence data strongly suggests that Neil3 is present in genomes in several metazoan phyla in addition to the chordates, at least in crustaceans such as lobster (GenBank [15] sequence identifier EG949310) and the zooplankton Calanus finmarchicus (FK041608) and mollusks such as abalones (GT869568), oysters (AM869093 and AM862460), and limpets (FC590695). Also in these species Val2 and Lys81 of human NEIL3 are conserved, confirming the importance of these residues for enzymatic function.

3.2 Purification of recombinant human NEIL3

The core glycosylase domain (GD) of human NEIL3 was expressed as a fusion protein with a C-terminal 6x His tag from pETDuet-1-GD-NEIL3 [5]. The cell-free protein extract of NEIL3 was subjected to Ni-NTA affinity chromatography and further purified on a Superdex75 size-exclusion column. The purity of NEIL3 was estimated to ≥ 95 % as judged by SDS-PAGE. Like the other functional H2TH family proteins, GD NEIL3 contains the N-terminal catalytic residues, the H2TH motif and the Fpg-type zinc finger (Fig. 1). We tested purified GD NEIL3 by cyanoborohydride-trapping to AP site-containing ssDNA. The DNA-protein complexes observed for the purified GD NEIL3 differed in size from the complexes formed by recombinant E. coli Fpg, Nei and Nth, excluding the possibility that contaminations of these E. coli proteins were interfering with the observed incision activity of NEIL3 (Fig. 2).

Figure 1. Cyanoborohydride-trapping assay of purified recombinant human NEIL3.

100 fmol purified human GD NEIL3 and E. coli Fpg, Nei and Nth (New England Biolabs) were incubated with 10 fmol of ssDNA containing an AP-site in a 10 μl reaction mixture containing 50 mM MOPS pH 7.5, 1 mM EDTA, 5 % glycerol, 1mM DTT and 50 mM NaCNBH₃ at 37 °C for 30 min. Protein dilution buffer was added instead of protein in the negative control. Protein-DNA complexes were visualized by SDS-PAGE and phosphorimaging.

Figure 2. Base excision and strand incision activities of human NEIL3.

100 fmol purified human NEIL1, NEIL2, GD NEIL3 and NTH1 were incubated with 10 fmol radiolabelled oligo containing a single lesion in a 10 μl reaction volume of 50 mM MOPS pH 7.5, 1 mM EDTA, 5 % glycerol and 1 mM DTT at 37 °C for 30 min. Protein dilution buffer was added instead of protein in the negative controls. The reactions were terminated with 80 mM NaOH or formamide stop solution, to measure the base excision and AP lyase activities, respectively. The cleavage products were separated from the substrate on a 20 % polyacrylamide/7 M urea gel and visualized by phosphorimaging.

Edman sequencing of purified GD NEIL3 was performed (SGS M-Scan Ltd, UK) and unraveled partial cleavage of the N-terminal methionine. Methionine removal was suggested to be essential for activity of mouse and human NEIL3 [9]. Edman sequencing is not a quantitative method and the percentage of NEIL3 containing N-terminal methionine in our purified protein was therefore not determined. However, cynoborohydride-trapping with excess substrate (ssSp) showed that at least 90 % of GD NEIL3 bound covalently to the substrate (data not shown), indicating that most of the purified protein is active.

3.4 Human NEIL3 acts mainly as a monofunctional DNA glycosylase with highest affinity for the hydantoin lesions Sp and Gh

The Fpg/Nei family of DNA glycosylases is known to remove a broad range of DNA base lesions. In order to examine the substrate range of human NEIL3, we analyzed the excision and incision activities of purified human NEIL3 towards numerous lesions, including oxidative and deaminated bases as well as mismatches, ribonucleotides and epigenetic DNA methylation products (Table 1). In accordance with the findings of Liu and coworkers [8, 9], purified human GD NEIL3 showed significant DNA glycosylase activity towards the hydantoin lesions Gh and Sp both in ssDNA and dsDNA. NEIL3 showed no cleavage of 5OHC and 5OHU in dsDNA, but a weak activity for the same lesions in ssDNA (Fig. 3). NEIL3 showed no cleavage activity towards the other substrates tested.

Table 1.

Substrates tested for GD NEIL3 activity and binding.

Lesion	Base excision or binding	Lesion	Base excision or binding
ssAP	+++	AP:C	+
ssGh	+++	Gh:C	++
ssSp	+++	Sp:C	++
ss5OHC	+	5OHC:G	−
ss5OHU	+	5OHU:A	−
ssDHT	−	DHT:A	−
ssDHU	±	DHU:A	−
ssU	−	U:A	−
hairpin	−	8oxoG:C	−
3′flap	−	5′flap	−
A:A mismatch	−	A:TT loop	−
3-Way junction	−	I:T	−
εC:G	−	εA:T	−
1 ssrA	−	1 rA:T	−
1 ssrU	−	1 rU:A	−
1 ssrC	−	1 rC:G	−
1 ssrG	−	1 rG:C	−
ssHmc	−	HmC:G	−
Ss5meC	−	5meC:G	−

Abbreviations: 3′flap, DNA duplex with single stranded 3′ overhang; 3-way junction, replication intermediate, 5′flap, DNA duplex with single stranded 5′overhang; 5meC, 5-methylcytosine; 5OHC, 5-hydroxycytosine; 5OHU, 5-hydroxyuracil; A:TT, insertion loop; εA, 1,N6-ethenoadenine; εC, 3,NA-ethenocytosine; DHT, 5,6-dihydrothymine; DHU, 5,6-dihydroxyuracil; Gh, guanidinohydantoin; HmC, 5-hydroxymethylcytosine; I, inosine; Sp, spiroiminodihydantoin;

¹DNA oligo with one ribonucleotide incorporated.

Figure 3. Single turnover kinetics for DNA base excision of human NEIL3.

100 fmol GD NEIL3 was incubated with 10 fmol ssDNA containing a Sp, Gh, 5OHC or 5OHU lesion and dsDNA containing a Sp or Gh lesion. All reactions were performed in a 10 μl volume of 50 mM MOPS pH 7.5, 1 mM EDTA, 5 % glycerol and 1 mM DTT at 37 °C for 10 sec – 20 min. Reactions were terminated by addition of 80 mM NaOH to measure the DNA base excision activity. The reaction products were separated on a 20 % polyacrylamide/7 M ureagel and visualized by phosphorimaging. DNA base excision activity was measured using ImageQuant v2003.02 software. Curves show mean values of three independent time course series. The mean values were fitted to a one phase associate model for the calculation of k_obs (min^-1).

To examine whether the glycosylase and AP lyase activities are coupled reactions, incision assays were analyzed by quenching the reactions with NaOH or formamide to detect base removal or strand incision, respectively. In these experiments we showed that the base excision reaction is more efficient than the AP lyase activity (Fig. 3), demonstrating that the two reactions are not coupled. We propose that NEIL3 in the presence of APE1 mainly acts as a monofunctional DNA glycosylase, where APE1 is performing the strand incision. Further, these experiments showed that NEIL3 mainly incised the phosphate backbone via β-elimination and not via β,δ-elimination such as NEIL1 and NEIL2. These results are in agreement with Liu et al, reporting that base removal and strand incision for mouse Neil3 is non-concerted [8] Notably, NTH1 only acts on Sp and Gh in dsDNA (Fig. 3), suggesting that only the NEIL enzymes remove hydantoins in ssDNA.

To further characterize the biochemical properties of GD NEIL3, single turnover rates for base removal were calculated (k_obs) from experiments on single- and double-stranded DNA containing Sp or Gh lesions and single-stranded DNA containing 5OHC or 5OHU (Fig. 4). The turnover numbers for Sp and Gh in single- and double-stranded DNA are approximately the same, whereas the turnover for 5OHC and 5OHU are 5-10 fold slower compared to Sp and Gh. In comparison, the turnover numbers for the glycosylase reaction of mouse Neil3 were approximately 4- and 200-fold faster on double- and single-stranded DNA, respectively, as compared to our human NEIL3 [8]. Thus, it appears that mouse Neil3 exhibit a much stronger single-strand preference than human NEIL3. Void filling residues and negatively charged residues of mouse Neil3 may create an unfavorable electrostatic environment for the complementary strand [10]. Experiments using total cell extract from neuronal stem/progenitor cells as well as thymus, testis and heart from newborn Neil3-deficient mice reveal that Neil3 is the main DNA glycosylase repairing Sp and Gh lesions in single-stranded, but not double-stranded DNA [11, 12].

Figure 4. Residues involved in DNA glycosylase and AP lyase activity of human NEIL3.

200 fmol purified WT, V2P and K81A GD NEIL3 were incubated with 10 fmol radiolabelled ssDNA containing a Sp lesion in a 10 μl reaction volume of 50 mM MOPS pH 7.5, 1 mM EDTA, 5 % glycerol and 1 mM DTT at 37 °C for 30 min. Protein dilution buffer was added instead of protein in the negative controls. The reactions were terminated by adding formamide stop solution to measure the coupled DNA glycosylase/AP lyase activity or by adding 80 mM NaOH to measure the glycosylase activity only. The cleavage products were separated from the substrate on a 20 % polyacrylamide/7 M urea gel and visualized by phosphorimaging. Representative experiments are shown. Standard deviations show uncertainties from three independent experiments.

NEIL1, NEIL2 and E. coli Fpg and Nei, incise damaged DNA by β,δ-elimination [16-19]. In contrast, our in vitro experiments showed that the uncoupled incision activity of human NEIL3 is performed by β-elimination (Fig. 3). Although NEIL3 functions as a monofunctional or a bifunctional glycosylase in vivo, it still will be involved in a different branch of the BER pathway than NEIL1 and NEIL2, which direct repair into a PNK dependent, but APE1 independent, pathway [20, 21]. Presumably, NEIL3 is initiating repair via an intact AP site (monofunctional mode) or a 3’ nicked AP site (bifunctional mode), which is subsequently processed by APE1. Furthermore, expression of NEIL3 is differently regulated during the cell cycle as compared to NEIL1 and NEIL2 [17, 22]. Neurauter and coworkers showed that release from quiescence stimulates the expression of NEIL3 under the control of the Ras dependent ERK-MAP kinase pathway [22]. In contrast, the total expression of NEIL1 was downregulated upon release from quiescence while the expression of NEIL2 was cell cycle independent. Taken together the differences between NEIL3 and NEIL1/NEIL2 in catalysis and cell cycle progression indicate that NEIL3 functions in a different cellular context and sub-pathway of BER than NEIL1 and NEIL2.

3.5 Identification of amino acids involved in glycosylase activity

The residues Val2 and Lys81 of human NEIL3 are absolutely conserved in NEIL3 in metazoa (vide supra, Fig. S1) and are likely to be involved in the catalysis and/or substrate recognition. Previously, it has been shown that the active site nucleophile Pro2, conserved in the Fpg/Nei family except the NEIL3 enzymes, acts in processing of base lesions and AP sites by a β,δ-elimination mechanism [17, 23]. It has been debated whether Val2 of NEIL3 is a nucleophile equivalent to Pro2 of the Fpg/Nei family [24, 25]. Biochemical characterization of the mutant proteins NEIL3 V2P and K81A was performed to address the role of these amino acids in the DNA glycosylase and AP lyase reactions. Each of the mutant NEIL3 proteins was expressed and purified to ≥95 % purity. The base removal versus strand incision was examined for either mutants on ssDNA containing Sp (Fig. 5) or Gh (data not shown) lesions. In contrast to the wild-type NEIL3 protein, the V2P mutant showed coupled base excision and strand incision reactions for both substrates, indicating that the N-terminal proline is required for the concerted action of the Neil enzymes. The two-nucleophile model for enzymatic catalysis of E. coli Fpg [26] showed that Lys57 act as an alternative nucleophile to Pro2. Substitution of the corresponding lysine residue (Lys81) in human NEIL3 with alanine results in no cleavage of Sp and Gh in ssDNA, demonstrating that Lys81 is essential for catalysis.

Abbreviations

5OHC

5-hydroxy-2’-deoxycytidine

5OHU

5-hydroxy-2’-deoxyuridine

AP

apurinic/apyrimidinic

BER

base excision repair

CV

column volume

faPy

2,6-diamino-4-hydroxy-5-formamidopyrimidine

Fpg

Escherichia coli formamidopyrimidine DNA glycosylase

GD

core glycosylase domain

Gh

guanidinohydantoin

H2TH

helix-two-turn-helix

MCS

multiple cloning site

Nei

E. coli endonuclease VIII

NEIL

E. coli endonuclease VIII-like

NTH

endonuclease III

ROS

reactive oxygen species

Sp

spiroiminodihydantoin

ss

single-stranded

ds

double-stranded

UDG

uracil DNA glycosylase