CiApex 1 has AP endonuclease activity and abrogated AP site repair disrupts early embryonic development in Ciona intestinalis

Kento Igarashi, Masafumi Funakoshi, Seiji Kato, Takahito Moriwaki, Yuichi Kato and Qiu-Mei Zhang-Akiyama* Laboratory of Stress Response Biology, Department of Biological Sciences, Graduate School of Science, Kyoto University, Kitashirakawa-Oiwakecho, Sakyo-ku, Kyoto 606-8502, Japan Department of Applied Pharmacology, Graduate School of Medical and Dental Sciences, Kagoshima University, 8-35-1 Sakuragaoka, Kagoshima, Kagoshima 890-8544, Japan Department of Stem Cell Biology, Atomic Bomb Disease Institute, Nagasaki University, Sakamoto 1-12-4, Nagasaki 852-8523, Japan Engineering Biology Research Center, Kobe University, 1-1 Rokkodai, Nada-ku, Kobe 657-8501, Japan


INTRODUCTION
Apurinic/apyrimidinic (AP) sites, which are a frequently generated form of DNA damage, are cytotoxic.Approximately 9,000 AP sites are spontaneously generated per cell per day in mammalian cells (Nakamura et al., 1998).In addition to spontaneous hydrolysis, AP sites are produced by DNA glycosylases, which remove damaged bases from DNA (Krokan et al., 1997;Guillet and Boiteux, 2003).If not repaired, AP sites interfere with DNA replication and RNA transcription, resulting in hazardous effects such as cellular dysfunction, cell death or mutations (Kunkel, 1984;Boiteux and Guillet, 2004;Fung and Demple, 2005).AP endonuclease recognizes AP sites and cleaves a phosphodiester bond from the 5′ side of the AP site, thereby generating a 3′-OH end (Baute and Depicker, 2008).This reaction is followed by the synthesis of DNA to produce proper base pairs (Baute and Depicker, 2008).
AP endonuclease 1 (APEX1) is the major AP endonuclease in mammals (Kane and Linn, 1981;Demple et al., 1991).Previous studies reported that APEX1 deficiency leads to embryonic lethality in mouse and zebraf-2 K. IGARASHI et al. ish (Xanthoudakis et al., 1996;Ludwig et al., 1998;Wang et al., 2006).However, it is not known whether APEX1 deficiency truly abrogates AP site repair.In fact, there is a report that due to a backup enzyme, AP endonuclease activity does not decrease in APEX1 knockdown embryos compared with control embryos (Fortier et al., 2009).Therefore, for estimating the effects of disturbed AP site repair on early embryonic development, it is also necessary to employ a method other than gene knockout or knockdown.
The ascidian Ciona intestinalis is a close relative of vertebrates and is used as a model animal to study the evolutionary developmental biology of chordates (Satoh et al., 2003).Ciona intestinalis embryos undergo a larval stage in which they have a primitive body plan with a notochord and neural tube, similar to other chordates (Di Gregorio and Levine, 1998).The criteria for detecting developmental stages in C. intestinalis are simpler than those in other chordates (Hotta et al., 2007).Ciona intestinalis has been used in previous studies focusing on base excision repair.A previous study suggested that CiApex1, a C. intestinalis homolog of human APEX1 (hAPEX1), was expressed during early development of C. intestinalis (El-Mouatassim et al., 2007).Furthermore, homologs of the DNA glycosylases Ogg1 and Nth have already been identified and characterized in C. intestinalis (Jin et al., 2006;Kato et al., 2012), and were suggested to participate in base excision repair.
In this study, we characterized CiApex1 and found that CiApex1 functions as an AP endonuclease, exhibiting enzymatic properties similar to those of hAPEX1.To investigate the effects of abrogated AP site repair on early embryonic development, we employed three types of AP site repair-disturbing methods: methyl methanesulfonate (MMS) treatment, which induces AP site accumulation; methoxyamine (MXA) treatment, which inhibits AP site repair by binding to AP sites; and treatment with dominant-negative CiApex1 (DN-CiApex1), which inhibits AP endonuclease directly.We found that AP site repair is important for early embryonic development, and that CiApex1 contributes to AP site repair during this developmental stage.

Bacterial strains
The Escherichia coli strains AB1157 and RPC501 (an xth nfo-deficient mutant of AB1157) used in this study were described previously (Cunningham et al., 1986), and were grown in Luria-Bertani (LB) medium with appropriate antibiotics.

Plasmid construction
We searched the Ghost database (http://ghost.zool.kyoto-u.ac.jp/cgi-bin/gb2/gbrowse/ kh/) for AP endonuclease genes with homology to hAPEX1.The C. intestinalis EST clone ciem810c12 was identified and we named the corresponding gene CiApex1.The CiApex1 gene amplified from ascidian genomic DNA possessed two nucleotide acid substitutions, resulting in two amino acid substitutions, compared to the sequence in the NCBI database (Supplementary Fig. S1A).Ciona intestinalis has not been established as a stable laboratory line and displays a high polymorphism rate (Kim et al., 2007).Therefore, we changed the CiApex1 nucleotide sequence to the sequence published in the NCBI database using overlap extension PCR prior to the experiment.This full-length CiApex1 gene was amplified to construct a plasmid for recombinant protein expression (pGEX-CiApex1) using a forward primer with a BamHI site, 5′-ATCGGGATCCATGATA-AAAGTTTTTCAAGTTAAAAT-3′, and a reverse primer with an XhoI site, 5′-ATATCTCGAGTTATGGCACAG-CAAGTAA-3′, and ligated with the vector pGEX-4T-1.The expression vectors used in Fig. 2C were constructed by overlap extension PCR to substitute the E116, D301 or D326 codons of the CiApex1 gene in pGEX-CiApex1 with alanine codons.The expression vector for DN-CiApex1 was constructed as follows: the nucleotide sequence of the CiApex1 gene in pGEX-CiApex1 was changed to encode a protein that carried two amino acid substitutions (E116Q and D229N).The DN-CiApex1 sequence was subcloned into the pSP72 plasmid, which carries a Ci-FoxA-a minimal promoter region (Di Gregorio et al., 2001;Oda-Ishii and Di Gregorio, 2007), and the Venus sequence was inserted into the 3′-terminus of DN-CiApex1.The PCR primers used were a forward primer with a KpnI site, 5′-CTTCTTGGTACCCATGATAAAAAGTTTTTCAA-3′, and a reverse primer with an EcoRI site and a BamHI site, 5′-CGGAATTCCGCCGCGGATCCTGGCACAGCAAGTA-ACAAAGT-3′.Plasmids were verified by sequencing.
Active site structure of CiApex1 protein The conserved structure of CiApex1 protein was predicted using the PHYRE2 program (Kelley et al., 2015), and the predicted structure was merged with the structure of hAPEX1 (Protein Data Bank entry 1DE9) using PyMOL software (The PyMOL Molecular Graphics System, Version 1.8, Schrödinger, NY, USA).
Expression and purification of CiApex1 and DN-CiApex1 protein Expression and protein purification were performed according to previously described protocols with some modifications (Jin et al., 2006;Kato et al., 2012).To avoid endogenous AP endonuclease contamination, AP endonuclease-deficient E. coli RPC501 (xth − nfo − ) was used as a host strain for purifying GST-CiApex1 and GST-DN-CiApex1 protein.The E. coli RPC501 strain carrying pGEX-CiApex1 or pGEX-DN-CiApex1 was incubated at 37 °C overnight in 15 ml of LB medium containing 100 μg/ml of ampicillin.Fifteen milliliters of the overnight culture was added to 1.5 l of fresh LB and Abrogated AP site repair affects embryogenesis grown at 16 °C until the optical density at 600 nm reached approximately 0.6.The culture was further incubated at 15 °C overnight in the presence of 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG).Cells were harvested, washed and resuspended in buffer A (274 mM NaCl, 5.4 mM KCl, 20 mM Na 2 HPO 4 , 3.52 mM KH 2 PO 4 (pH 7.4), 1 mM PMSF (phenylmethylsulfonyl fluoride) and 1 mM DTT (dithiothreitol)).The cell suspension was sonicated and the cell lysate was centrifuged at 20,000 × g at 4 °C for 30 min.The supernatant was applied to a GSH-Sepharose 4B column (Pharmacia Biotech, Uppsala, Sweden) that had been equilibrated with buffer A. GST-CiApex1 or GST-DN-CiApex1 fusion protein was eluted from the column with buffer B (50 mM Tris-HCl (pH 9.6), 15 mM GSH), dialyzed at 4 °C overnight against buffer C (20 mM Tris-HCl (pH 8.0), 300 mM NaCl, 5% v/v glycerol, 1 mM DTT), and stored at − 80 °C until used.The purified protein was analyzed by SDS-PAGE, and protein concentrations were determined by the Bradford method.
Detection of AP endonuclease activity AP endonuclease activity was detected according to previously described protocols with some modifications (Jin et al., 2006;Kato et al., 2012).The sequences of the oligonucleotides used in this study are shown in Fig. 1A.The oligonucleotide 5′-CCTGCCCTGFGCAGCTGTGGG-3′    were then heated at 95 °C for 5 min, immediately cooled on ice, and then loaded onto a denaturing gel (20% polyacrylamide, 89 mM Tris-borate (pH 8.3), 7 M urea, 2 mM EDTA).After electrophoresis at 1,300 V for 90 min, the gel was autoradiographed using BAS-1800II (Fujifilm, Tokyo, Japan).Quantitative analyses of band intensities were performed using ImageJ software.
Escherichia coli survival assay Overnight-cultured E. coli were serially diluted.To estimate survival, appropriate dilutions of the overnight culture were plated or spotted on LB plates containing agents indicated in each figure, and incubated at 28 °C or 25 °C (in experiments using DN-CiApex1).After incubation for one day at 28 °C or two days at 25 °C, the number of colonies was counted or the plates were photographed.Statistical analyses were conducted using Student's t-test.
Strain and culture conditions of C. intestinalis Ciona intestinalis adults were obtained from the National BioResource Project for C. intestinalis.Eggs and sperm were surgically obtained from gonoducts and the eggs were fertilized in vitro.Embryos were raised as described previously (Imai et al., 2012), with the exception that they were allowed to develop at 18 °C in filtered seawater after they were dechorionated with thioglycolic acid (Nacalai Tesque, Kyoto, Japan) and sodium hydroxide.
Evaluation of embryonic development To assay the drug sensitivities of embryos, MMS (M0369, Tokyo Chemical Industry, Tokyo, Japan) or MXA, also called O-methylhydroxylamine (06993-14, Nacalai Tesque) was added to filtered seawater and embryos were fixed in phosphate-buffered saline containing 4% paraformaldehyde (Nacalai Tesque).Images of embryos were obtained with an Olympus IX70 (Tokyo, Japan) or Zeiss LSM510 (Tokyo, Japan), and analyzed using ImageJ software and the Zeiss LSM Image Browser.Statistical analyses were conducted using Student's t-test (*, P < 0.05).
Electroporation Electroporation was performed according to previously described protocols with some modifications (Oda-Ishii et al., 2010;Vierra and Irvine, 2012).Fertilized eggs in 14 μl of seawater were mixed with 25 μl of 0.96 M mannitol and 4.75 μl of 1 μg/μl plasmid DNA, transferred to a cuvette with a 2-mm electrode gap, and electroporated using the Gene Pulser Xcell System (Bio-Rad, CA, USA) using the square pulse protocol (35 V and 16 ms per pulse).
Measurement of AP sites Genomic DNA was extracted from embryos of C. intestinalis at 6 hours post fertilization (hpf) using a GenElute Mammalian Genomic DNA Miniprep kit (Sigma-Aldrich, MO, USA) at 4 °C.AP sites were detected using an aldehyde reactive probe (Dojindo, Kumamoto, Japan) as described previously (Mohsin Ali et al., 2004;Kato et al., 2015).
Semi-quantitative RT-PCR Total RNA was extracted from C. intestinalis embryos at 13 hpf with TriPure isolation reagent (Roche, Basel, Switzerland) as recommended by the manufacturer.To eliminate contaminating genomic DNA, the extracted RNA was treated with DNase I (Takara Bio, Shiga, Japan).Using this purified RNA, cDNA was synthesized with oligo-dT primer and ReverTra Ace (Toyobo).Synthesized cDNA was amplified using Go-taq (Promega, WI, USA) and the following specific primers described previously (El-Mouatassim et al., 2007): GAAATGGAGAAAAGGCCACA and TTCACCCACTTTACCCCATC for CiApex1, and AATCCACCCTTCACCTTGTG and GGGAGATCTTGC-CATTTTCA for CiRps27.The PCR cycler conditions were as follows: an initial denaturation step at 94 °C for 5 min, followed by 25 or 28 cycles of 94 °C for 1 min, 60 °C for 1 min and 72 °C for 1 min.Amplified products were separated by electrophoresis on a 2% (w/v) agarose gel and visualized using ethidium bromide staining.Quantitative analyses of band intensities were performed using ImageJ software.), but they did not confirm that CiApex1 is a functional APEX1 homolog.To do this, we first characterized CiApex1 in vitro.For cloning CiApex1, a search for hAPEX1 was done using the Ghost database and the full-length amino acid sequence of hAPEX1.The EST clone ciem810c12 was detected from the 5′-full-length EST database.The ciem810c12 clone had two nonsynonymous single-nucleotide polymorphisms (SNPs) compared to the CiApex1 sequence previously reported (NCBI accession number XP_002123983.3)(Supplementary Fig. S1A).These two SNPs of the ciem810c12 clone were changed to encode the same amino acid sequence as XP_002123983.3(Supplementary Fig. S1B).The amended clone encoded a 337-amino acid protein (38.7 kDa) (Kato et al., 2012).The amino acid similarity between hAPEX1 and CiApex1 was 64% (Supplementary Fig. S1B).In addition to AP site repair function, mammalian homologs of APEX1 are considered to have regulatory functions as Ref-1 (Xanthoudakis and Curran, 1992;Bhakat et al., 2009).Previous studies identified the redox catalytic site residues as cysteine 65 and 93 in human Ref-1 (Xanthoudakis and Curran, 1992).As shown in Fig. S1B, we found that the critical residues for redox function are not conserved in CiApex1.This is consistent with a previous study which reported that the Ref-1 domain is only conserved among mammals (Georgiadis et al., 2008).The candidate CiApex1 was subcloned into pGEX-4T-1 plasmid.The plasmid obtained was named pGEX-CiApex1.

Identification
Purified GST-CiApex1 exhibits AP endonuclease activity To investigate whether CiApex1 has AP endonuclease activity, purified CiApex1 fused with a GST tag (GST-CiApex1) was used (Supplementary Fig. S2).GST-CiApex1 was incubated with a double-stranded DNA substrate containing a THF-AP site.As shown in Fig. 1A, like hAPEX1 (lane 3) and XthA, a major AP endonuclease in E. coli (lane 4), GST-CiApex1 cleaved the AP site substrate and produced a 9-mer oligonucleotide (lane 5).This result indicated that CiApex1 has AP endonuclease activity.The specific activity of CiApex1 was calculated as 42 nmol/min/mg.As shown in Fig. 1A, XthA exhibited strong exonuclease activity toward the AP site-containing DNA substrate, while the exonuclease activity of hAPEX1 and GST-CiApex1 was barely detectable.This result indicated that CiApex1 and hAPEX1 are similar types of AP endonuclease.hAPEX1 is a Mg 2 + -dependent enzyme, while Nfo-type AP endonuclease is a Zn 2 + -dependent enzyme (Barzilay et al., 1995;Hosfield et al., 1999).The effects of various metals were therefore investigated.As shown in Fig. 1B, the metal specificity of CiApex1 was in the order Mg 2 + , Ca 2 + > Mn 2 + , K + , Ni 2 + and Zn 2 + , and CiApex1 exhibited strong Mg 2 + -dependent activity.These experiments indicated that purified GST-CiApex1 protein exhibits similar enzymatic properties to hAPEX1.
Cleavage assays were also performed at different temperatures (Fig. 1C).The GST-CiApex1 protein exhibited the strongest activity at 28 °C, which was close to the optimal temperature of 24 °C for CiOgg1 DNA glycosylase (Jin et al., 2006).
CiApex1 rescues the sensitivity of AP endonuclease-deficient E. coli to DNA damaging agents To investigate whether CiApex1 can function as an AP endonuclease in vivo, a complementation assay was performed using AP endonuclease-deficient E. coli RPC501 (xth − nfo − ).Relative to RPC501 transformed with the empty vector, RPC501 transformed with pGEX-CiApex1 or pGEX-hAPEX1 vector showed restoration of resistance to MMS and partial restoration of resistance to H 2 O 2 (Fig. 2A and Supplementary Fig. S3A, S3B).MMS and H 2 O 2 treatment generate damaged DNA bases that are subsequently processed by DNA glycosylase to yield substrates for AP endonuclease (Blakely et al., 1990;Shrivastav et al., 2009).Therefore, these results suggested that CiApex1 functions as an AP endonuclease and that its in vivo activity is similar to that of hAPEX1.
Aspartate and glutamate residues are critical for AP endonuclease activity and are conserved between CiApex1 and hAPEX1 Our finding that the purified GST-CiApex1 protein required Mg 2 + ions to cleave substrate DNA (Fig. 1B) suggested that the structure of the active site of CiApex1 is similar to that of hAPEX1.The predicted structure of the active site of CiApex1 was indeed similar to that of hAPEX1 (Fig. 2B).E96, D283 and D308 are regarded as critical for the reaction mechanism of AP endonuclease because hAPEX1 proteins with mutations in these residues exhibit lower AP endonuclease activity (Barzilay et al., 1995).These residues are also conserved in CiApex1 (Supplementary Fig. S1B), in which E116, D301 and D326 align with E96, D283 and D308, respectively.To investigate whether the active site structure is functionally conserved, we constructed CiApex1 vectors, each with a single mutation (E116A, D301A or D326A) at the predicted active site, and determined whether mutant CiApex1 could rescue the resistance of the E. coli RPC501 strain to MMS.We found that none of the CiApex1 mutants rescued the resistance to MMS (Fig. 2C and Supplementary Fig. S3C).These results showed that all three CiApex1 mutants lost AP endonuclease activity and that the active site structure of AP endonuclease 1 is functionally conserved between CiApex1 and hAPEX1.

Embryos of C. intestinalis are sensitive to MMS and MXA
To examine the effects of AP site accumulation on early embryonic development, the effects of MMS treatment on embryos were investigated.MMS alkylates DNA bases, which are then converted to AP sites by spontaneous depurination or cleavage by specific glycosylases (Singer and Brent, 1981;Drabløs et al., 2004;Shrivastav et al., 2009).The accumulation of AP sites was confirmed in MMS-treated C. intestinalis embryos by an aldehyde reactive probe (ARP) assay.Treatment with 0.8 mM MMS increased AP site accumulation to 130% of the control level at 6 hpf (Fig. 3A).This result indicated that MMS treatment also induces AP site accumulation in C. intestinalis.This assay also showed that the number of AP sites increased 2.4-fold from 0 to 6 hpf (Fig. 3A), indicating that AP sites are produced spontaneously during early embryonic development in C. intestinalis.We found that 0.8 mM MMS treatment led to obvious abnormalities in embryos, especially at 18 hpf (Fig. 3B).We also found that 0.4 mM MMS treatment induced growth delay compared with the growth of untreated embryos.Hotta et al. previously showed that the tail/head ratio indicated the developmental stage of C. intestinalis embryos during the tail-bud stage (Hotta et al., 2007).Accordingly, to evaluate the effects of MMS treatment quantitatively, especially at low MMS concentration, the tail/ head ratio was calculated.Compared with the tail/head ratio in untreated embryos, 0.4 mM MMS treatment led Abrogated AP site repair affects embryogenesis to a 28 (10 hpf) and 21% (18 hpf) decrease in tail/head ratio (Fig. 3C).MMS concentration-dependent effects were also investigated.The results showed that even 0.2 mM MMS treatment led to a significant tail/head ratio decrease (32%, P < 0.05) compared with the ratio in control embryos (Fig. 3D).Therefore, AP site generation may lead to abnormal development in C. intestinalis.
MXA blocks AP endonuclease activity by reacting with the aldehyde-sugar group of the AP site, causing a stable MXA-AP-intermediate adduct (Rosa et al., 1991;Laev et al., 2017).It is also reported that MXA treatment enhances the effects of MMS treatment (Gerson and Liu, 2003).To examine the effects of AP site repair inhibition on early embryonic development, the effects of MXA treatment on embryos were investigated (Fig. 3E).The tail/ head ratio of embryos raised in 1 mM MXA was significantly decreased (by 41%, P < 0.05) compared with that of embryos raised in seawater without MXA.Embryos raised in 2 mM MXA had severe abnormalities.We were unable to determine the tail/head ratio of embryos in 2 mM MXA.Embryos raised in 3 mM MXA became like a cell cluster.When raised in 3 mM MXA, embryos were not detected at the tail-bud embryo stage.These results suggested that AP site repair inhibition leads to abnormal development.
A dominant-negative form of GST-CiApex1 efficiently inhibits THF-AP site cleavage activity and causes cytotoxic effects As a method for inhibiting the AP endonuclease directly, a dominant-negative form of hAPEX1 (named DN-hAPEX1 or ED protein) was previously described (McNeill and Wilson, 2007).In this method, hAPEX1 harboring mutations at two active residues was used.Although DN-hAPEX1 lost its enzymatic activity, it retained binding activity and stayed at the damaged DNA site, leading to inhibition of the endogenous hAPEX1.This AP site repair inhibition has been suggested to lead to cytotoxic effects (McNeill et al., 2009).To create a dominant-negative form of CiApex1 (DN-CiApex1), two amino acids (E116Q and D229N) at the predicted active site were changed (Fig. 4A).When substrate DNA containing THF-AP sites was reacted with purified DN-CiApex1 fused with a GST tag (GST-DN-CiApex1), 9-mer product DNA was not detected, indicating that GST-DN-CiApex1 possesses no AP endonuclease activity (Supplementary Fig. S4).GST-DN-CiApex1 was also assessed by an in vitro competition assay to examine whether it can inhibit the AP site repair performed by GST-CiApex1 in vitro.Substrate DNA was preincubated with the GST-DN-CiApex1 protein or bovine serum albumin (BSA), followed by the addition of GST-CiApex1 protein.The addition of GST-DN-CiApex1 resulted in an approximately 80% decrease of GST-CiApex1 AP endonuclease activity, while the addition of BSA did not (Fig. 4A, 4B).These results indicated that GST-DN-CiApex1 protein could inhibit the incision activity of CiApex1 in vitro.Next, to test whether the expression of DN-CiApex1 led to decreased cell survival, wild-type E. coli AB1157 was transformed with pGEX-DN-CiApex1, pGEX-CiAPEX1 or pGEX-4T-1 plasmid.Overnight cultures of each transformant were diluted serially and spotted on LB agar plates.As shown in Fig. 4C, although IPTG-induced CiApex1 had no effect on survival rate, DN-CiApex1 expression led to decreased AB1157 survival rate (Fig. 4C).The MMS sensitivity of each transformant was also assessed by a colony formation assay.Even without IPTG induction, DN-CiApex1 expression increased the sensitivity to 5 mM MMS treatment 1,000-fold in AB1157 (Fig. 4D).Again, CiApex1 expression itself did not decrease the AB1157 survival rate (Fig. 4D).These cytotoxic effects of DN-CiApex1 are consistent with those of DN-hAPEX1 (McNeill and Wilson, 2007;McNeill et al., 2009).These results suggested that DN-CiApex1 inhibits AP site repair in vivo, and this AP site repair inhibition leads to cytotoxic effects.

Abnormal development of C. intestinalis caused by transfection of fertilized eggs with DN-CiApex1 expression plasmid
To assess the contribution of AP endonuclease to AP site repair during early embryonic development, DN-CiApex1 was expressed in C. intestinalis embryos.DN-CiApex1 is expected to inhibit all the AP endonucleases in C. intestinalis, but previously we confirmed that in C. intestinalis, no other AP endonuclease homologs except CiApex1 exhibited AP endonuclease activity (Funakoshi et al., 2017).Therefore, we consider that the effects of DN-CiAPEX1 are derived from specific CiApex1 inhibition.The FoxA-a promoter was previously shown to be activated from the 16-cell stage (2.35-2.65 hpf) in C. intestinalis embryos (Imai et al., 2004) and is expected to be highly active throughout embryonic development (Supplementary Fig. S5).To express sufficient DN-CiApex1 protein in C. intestinalis embryos, DN-CiApex1 DNA was subcloned into the pSP72 plasmid downstream of the promoter region of Ci-FoxA-a (Di Gregorio et al., 2001;Oda-Ishii and Di Gregorio, 2007).We confirmed that DN-CiApex1 mRNA is overexpressed compared with endogenous CiApex1 mRNA at 13 hpf (Fig. 5A).We also confirmed that CiApex1 mRNA was actually expressed during early embryonic development from the results of non-electroporated embryos (Fig. 5A).The in vivo expression of DN-CiApex1 was confirmed using Venus protein fused to the C-terminus of the DN-CiApex1 protein (Fig. 5B).Eggs were transfected with DN-CiApex1-Venus plasmid by electroporation soon after fertilization, and with the Venus plasmid for comparison.We found that the expression of DN-CiApex1-Venus protein led to a significant decrease (16%, P < 0.05) in tail/ head ratio compared with the ratio in embryos expressing the Venus protein at 18 hpf (Fig. 5C).A similar result Abrogated AP site repair affects embryogenesis  was also obtained with embryos raised in 0.6 mM MXA (Fig. 5C).These results supported the possibility that CiApex1 inhibition leads to abnormal development and that CiApex1 contributes to AP site repair during early embryonic development.

DISCUSSION
AP sites are a frequently generated form of DNA dam-age under physiological conditions (Nakamura et al., 1998).They inhibit transcription and replication and induce DNA mutation (Boiteux and Guillet, 2004;Kunkel, 1984).AP sites are the substrate for AP endonuclease in the base excision repair pathway (Baute and Depicker, 2008).Although APEX1 deficiency leads to a lethal phenotype in mice and zebrafish (Xanthoudakis et al., 1996;Ludwig et al., 1998;Wang et al., 2006), it remains unknown whether this severe phenotype is truly derived Abrogated AP site repair affects embryogenesis from AP site repair deficiency (Fortier et al., 2009).In this study, to clarify the significance of AP site repair, we conducted experiments focusing on AP site repair during early embryonic development.
A previous study that focused on DNA repair and reproduction detected CiApex1 transcripts in oocytes and embryos until the larval stage (El-Mouatassim et al., 2007).However, AP endonuclease activity of CiApex1 had not yet been demonstrated.In the present study, we searched the database again, and found a candidate clone of CiApex1, which was 69 bases longer at the 5′-terminus than the sequence previously reported (El-Mouatassim et al., 2007) (Supplementary Fig. S1A).In vitro enzymatic assays revealed that CiApex1 exhibited AP endonuclease activity in a Mg 2 + -dependent manner, and exonuclease activity of CiApex1 was hardly detected (Fig. 1A, 1B).These enzymatic properties are consistent with those of hAPEX1 (Barzilay et al., 1995;Wilson et al., 1995).Like hAPEX1, CiApex1 restored the resistance of AP endonuclease-deficient E. coli to MMS but did not completely restore its resistance to H 2 O 2 (Fig. 2A and Supplementary Fig. S3A, S3B).These results showed that CiApex1 exhibits similar characteristics to hAPEX1.Previously, APEX2 and P0 homologs were reported to possess AP endonuclease activity and predicted to participate in AP site repair (Yacoub et al., 1996;Hadi and Wilson, 2000).Our group identified the corresponding homologs in C. intestinalis, CiApex2 and CiP0, and investigated their enzymatic activities in vitro using purified proteins, but found that these proteins did not exhibit AP endonuclease activity (Funakoshi et al., 2017).This indicated that CiApex1 is the dominant AP endonuclease in C. intestinalis.
To evaluate the significance of AP site repair during early embryonic development, the effects of abnormal AP site repair were assessed by two different methods.First, MMS was used as an AP site-generating agent, and we confirmed that the number of AP sites was increased by MMS treatment (Fig. 3A).As shown in Fig. 3B and 3D, MMS treatment interfered with embryogenesis in a concentration-dependent manner.Next, the effects of AP site repair inhibition on embryogenesis were also investigated.As shown in Fig. 3A, the number of AP sites was significantly higher at 6 hpf than at 0 hpf.This result suggested that AP sites accumulate during early embryonic development.MXA specifically binds to AP sites and inhibits AP site cleavage by AP endonuclease.Like MMS treatment, MXA treatment led to abnormal embryonic development in a concentration-dependent manner (Fig. 3E).From the MMS and MXA treatment results, it was concluded that abnormal AP site repair, which is caused by an increase in the number of AP sites or by AP site repair inhibition, leads to abnormal embryonic development in C. intestinalis.These results suggested that AP site repair is indispensable for early embryonic development.
Finally, we investigated the contribution of CiApex1 to AP site repair during early embryonic development.Initially, we planned to carry out knockdown using a morpholino antisense oligonucleotide (MO) for CiApex1.However, from the Ghost database (http://ghost.zool.kyoto-u.ac.jp/cgi-bin/gb2/gbrowse/kh/), it was found that CiApex1 is expressed as a polycistronic mRNA (Supplementary Fig. S6).Since the use of an MO could lead to inhibition of trans-splicing, we adopted another approach to inhibit CiApex1 using a DN-CiApex1 protein.This method enables us to inhibit the AP endonuclease directly, thus making it possible to evaluate the effect of AP endonuclease inhibition on early embryonic development.Although DN-CiApex1 should be able to inhibit all of the AP endonucleases in C. intestinalis, there are no enzymes exhibiting AP endonuclease activity other than CiApex1 (Funakoshi et al., 2017).Therefore, it is expected that the effects of DN-CiApex1 are derived from specific inhibition of CiApex1.In addition, mammalian homologs of APEX1 are considered to have regulatory function as Ref-1 in addition to AP endonuclease activity (Bhakat et al., 2009), but the critical residues for regulatory function as Ref-1 are not conserved in CiApex1 (Supplementary Fig. S1B).This result is consistent with a previous report, which proposed that the critical residues for Ref-1 function are only conserved among mammals (Georgiadis et al., 2008).Therefore, the present experimental results also suggest that the effect of DN-CiApex1 is derived from only AP site repair inhibition.Consistent with the findings of previous studies, DN-CiApex1 inhibited AP endonuclease activity of CiApex1 in vitro (Fig. 4A, 4B), and this AP endonuclease inhibition led to cytotoxic effects in E. coli (Fig. 4C, 4D).These results indicated that DN-CiApex1 works effectively as an AP endonuclease inhibitor.Under the control of the Ci-FoxA-a promoter, DN-CiApex1 was overexpressed compared with endogenous CiApex1 (Fig. 5A).After electroporation, Venus-positive embryos (i.e., embryos into which the plasmid DNA was successfully introduced) were assessed (Fig. 5B).As a result, we found that the expression of DN-CiApex1 also led to embryonic growth impairment (Fig. 5C).This result suggested that CiApex1 inhibition induced abnormal embryonic development, and also that CiApex1 contributes to AP site repair during normal early embryonic development.
In this study, CiApex1 was found to be a functional AP endonuclease in C. intestinalis, which belongs to the subphylum Urochordata.The present identification of AP endonuclease activity is the first such report in Urochordata (Funakoshi et al., 2017) and this is the first study demonstrating that AP site repair is important for early embryogenesis in an invertebrate.It was also demonstrated here that expressing DN-CiApex1 is a good method to inhibit AP site repair in C. intestinalis.However, at present, it is not clear whether CiApex1 deficiency also halts embryonic development.Recently, genome editing technology using TALEN or CRISPR has been established in C. intestinalis (Sasaki et al., 2014;Treen et al., 2014).In the future, these powerful methods will provide further comprehensive understanding about how mechanisms to repair not only AP sites but also other types of DNA damage, such as base damage, are controlled during embryonic development by using C. intestinalis embryos.
of full-length CiApex1 from the Ghost database In a previous study, El-Mouatassim et al. detected CiApex1 mRNA (El-Mouatassim et al., 2007

Fig. 3 .
Fig. 3. Abnormal AP site repair leads to abnormal development in C. intestinalis.(A) Average number of AP sites detected with ARP.Fertilized C. intestinalis eggs were obtained from at least two adult ascidians.Embryos were raised at 18 °C for 6 hpf, and then genomic DNA was extracted at 4 °C.Experiments were repeated three times.Error bars represent the standard deviation of the means.Statistical analyses were conducted using Student's t-test (* P < 0.05).(B) Representative images of C. intestinalis embryos in MMS treatment experiment.Fertilized C. intestinalis eggs were raised in normal seawater (without MMS) or in seawater containing MMS at 18 °C.Embryos were then fixed at 2, 10 and 18 hpf.Scale bars, 150 μm.(C) Time course of effects of MMS treatment.The tail/head ratio during the tail-bud stage was calculated and plotted (n > 3).Circles represent the tail/head ratio of embryos incubated in normal seawater, while triangles represent the ratio of embryos in MMS.Error bars represent standard errors.(D) Dose-dependence of MMS effect.MMS-treated embryos were fixed at 10 hpf, and the tail/head ratio of embryos was calculated and plotted (n > 10).Statistical analyses were conducted using Student's t-test (* P < 0.05).Photographs are representative images of C. intestinalis embryos in the MMS treatment experiment.Scale bars, 100 μm.(E) Developmental defects caused by MXA treatment.Eggs were fertilized and transferred to seawater containing MXA at 2 hpf.Embryos were fixed at 12 hpf and the tail/head ratio of embryos was calculated and plotted (n > 10).Statistical analyses were conducted using Student's t-test (* P < 0.05).Photographs are representative images of C. intestinalis embryos in the MXA treatment experiment.Scale bars, 100 μm.