DNA Repair 24 (2014) 63–72

Contents lists available at ScienceDirect

DNA Repair journal homepage: www.elsevier.com/locate/dnarepair

Analysis of SHPRH functions in DNA repair and immunoglobulin diversification Nils-Sebastian Tomi a , Kathrin Davari a , David Grotzky b , Friedemann Loos b,1 , Katrin Böttcher a , Samantha Frankenberger b , Berit Jungnickel a,∗ a Department of Cell Biology, Institute of Biochemistry and Biophysics, Center for Molecular Biomedicine, Friedrich-Schiller University Jena, Hans-Knoell-Strasse 2, 07745 Jena, Germany b Institute of Clinical and Molecular Biology, Helmholtz Center Munich, Marchioninistrasse 25, 81377 Munich, Germany

a r t i c l e

i n f o

Article history: Received 10 April 2014 Received in revised form 29 August 2014 Accepted 23 September 2014 Available online 11 October 2014 Keywords: SHPRH PCNA Ubiquitin DNA repair DT40

a b s t r a c t During replication, bypass of DNA lesions is orchestrated by the Rad6 pathway. Monoubiquitination of proliferating cell nuclear antigen (PCNA) by Rad6/Rad18 leads to recruitment of translesion polymerases for direct and potentially mutagenic damage bypass. An error-free bypass pathway may be initiated via K63-linked PCNA polyubiquitination by Ubc13/Mms2 and the E3 ligase Rad5 in yeast, or HLTF/SHPRH in vertebrates. For the latter two enzymes, redundancy with a third E3 ligase and alternative functions have been reported. We have previously shown that the Rad6 pathway is involved in somatic hypermutation of immunoglobulin genes in B lymphocytes. Here, we have used knockout strategies targeting expression of the entire SHPRH protein or functionally significant domains in chicken DT40 cells that do not harbor a HLTF ortholog. We show that SHPRH is apparently redundant with another E3 ligase during DNA damage-induced PCNA modification. SHPRH plays no substantial role in cellular resistance to drugs initiating excision repair and the Rad6 pathway, but is important in survival of topoisomerase II inhibitor treatment. Removal of only the C-terminal RING domain does not interfere with this SHPRH function. SHPRH inactivation does not substantially impact on the overall efficacy of Ig diversification. Redundancy of E3 ligases in the Rad6 pathway may be linked to its different functions in genome maintenance and genetic plasticity. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Genome maintenance is based on several complementary DNA repair pathways as well as DNA damage signaling. In combination, these processes ensure that DNA is largely damage-free before being replicated. If a replication fork does encounter DNA damage, its bypass is coordinated by the Rad6 pathway [1]. Monoubiquitination of the sliding clamp proliferating cell nuclear antigen (PCNA) facilitates efficient binding of translesion polymerases harboring a flexible catalytic site, allowing direct bypass even of bulky lesions [2,3]. Depending on the type of lesion and the polymerase used, this pathway may be rather error-prone. Alternatively, PCNA may be polyubiquitinated via K63-linkage in ubiquitin, employing the Ubc13/Mms2 dimer and the E3 ligase Rad5 in yeast, or SNF2

∗ Corresponding author. Tel.: +49 3641 949 960; fax: +49 3641 949 962. E-mail address: [email protected] (B. Jungnickel). 1 Present address: Department of Reproduction and Development, Erasmus MC, University Medical Center, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. http://dx.doi.org/10.1016/j.dnarep.2014.09.010 1568-7864/© 2014 Elsevier B.V. All rights reserved.

Histone Linker PHD RING helicase (SHPRH) and/or helicase like transcription factor (HLTF) in vertebrates [4–6]. Then, error-free bypass of the damage is triggered by a barely understood pathway that involves recruitment of the ZRANB3 translocase for fork restart [1,7]. While Rad18 plays a substantial role in PCNA monoubiquitination [8], the existence of at least one alternative E3 ligase for this process has been postulated [9], and two candidates have been proposed [10,11]. Importantly, PCNA monoubiquitination may enhance the efficacy of other repair pathways such as the Fanconi anemia pathway [12] and non-conservative mismatch repair, the latter also occurring outside of S phase [13,14]. So far, no differential contribution of the different E3 ligases to these different effects of PCNA monoubiquitination has been reported. Interestingly, though, Rad18 also has other functions in the regulation of DNA double strand break repair [15–17] and e.g. viral infection [18]. For the E3 ligases SHPRH and HLTF, involved in PCNA polyubiquitination, different reports have also pointed at redundancy and cooperation as well as other functions [4,5,19,20]. These ligases belong to the RING family, share a unique domain

64

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

architecture with Rad5 [5,21,22], and may promote PCNA polyubiquitination in in vitro assays or when overexpressed in mammalian cells [5,6,23,24]. While one study reported that both E3 ligases need to cooperate for efficient PCNA polyubiquitination [5], another report detected PCNA diubiquitination even in the absence of both enzymes, hinting at a third responsible E3 ligase for PCNA polyubiquitination [19]. In addition, HLTF was found to enhance PCNA monoubiquitination and recruitment of Polymerase ␩ while inhibiting SHPRH functions upon UV treatment. On the other hand, SHPRH was shown to increase polymerase ␬ recruitment upon HLTF degradation after MMS treatment [4]. Clearly, the redundancy of E3 ligases claimed to have a function in PCNA polyubiquitination complicates mechanistic studies on their impact on DNA repair. To exacerbate this issue, E3 ligases are rather large enzymes containing several functional modules that are encoded by multiple exons. While in RNAi studies, off-target effects may never be ruled out completely, knockout approaches may thus lead to alternative splicing events and production of partially functional or dominant negative proteins. Potential DNA damage specific functions or effects restricted to certain cell types or organs make it even more complicated to study E3 ligases involved in PCNA ubiquitination. We and others have previously shown that Rad18 and PCNA ubiquitination are involved in diversification of immunoglobulin genes by somatic hypermutation [25–28]. In the present study, we have asked whether PCNA polyubiquitination contributes to or counteracts Ig diversification. To this end, we have inactivated SHPRH by 3 different strategies in chicken DT40 B cells that do not harbor a HLTF ortholog. We show that in this system, SHPRH may have some effect on spontaneous PCNA ubiquitination in the absence of exogenous DNA damage, but does not change the levels of DNA damage-induced PCNA mono- or polyubiquitination. We show that SHPRH plays no role in survival of MMS- or cisplatininduced damage processed by base or nucleotide excision repair, but is required for survival of topoisomerase II inhibitors triggering DNA double strand breaks. Deletion of only the RING domain of SHPRH is not sufficient to cause this phenotype. Consistent with differential roles in processing of single strand lesions or double strand breaks, SHPRH does not affect somatic hypermutation of Ig genes, but may have some effect on mismatch tolerance during immunoglobulin gene conversion. Our findings support the notion of at least one additional E3 ligase involved in PCNA polyubiquitination, and highlight the complexity of studying the functions of these enzymes in different processes linked to genome maintenance.

2. Results

exon 22 and in addition deleted exons 23–26 of the SHPRH gene, aiming at direct inactivation of the crucial RING domain of this E3 ligase. All three strategies were attempted in the DT40V− cell line lacking the pseudogenes required for Ig gene conversion, which allows measurement of both DNA damage survival and somatic hypermutation. While gene inactivation was successful using strategies PHD and RING (Suppl. Tables 1 and 2), targeting efficiencies for strategy TSS in DT40V− were too low to obtain reliable complete knockout clones for subsequent analyses (Suppl. Table 3). For this knockout strategy, we therefore used DT40Cre1 cells still containing the pseudogenes required for Ig gene conversion, as this cell line shows somewhat higher targeting efficiencies for unknown reasons (Suppl. Table 4 and data not shown). Southern blot analyses confirmed complete removal of the respective parts of the SHPRH gene (Fig. 1E–G). RT-PCR also confirmed lack of transcription of the targeted exons (Fig. 1H–J). However, for strategy TSS and PHD, residual transcription downstream of the targeted exons was detected, which may lead to production of at least partially functional truncated proteins. qRTPCR analyses revealed a substantial decrease of transcription in case of the promoter knockout in strategy TSS (Suppl. Fig. 1). We therefore directly assessed effects of the knockout strategies on SHPRH protein expression by western blot analyses. Reactivity of the antibody used was demonstrated by its ability to detect an HA-tagged chicken SHPRH protein expressed in DT40 cells (Suppl. Fig. 2), however crossreactivity of the antibody with other proteins may obscure some residual smaller SHPRH fragments, in particular as endogenous SHPRH expression in DT40 appears to be low (Fig. 2A). In case of strategy TSS, we detected disappearance of the full length protein in the knockout clones (Fig. 2A) yet no appearance of truncated forms. The same was true for strategy PHD (Fig. 2B). In case of strategy RING (Fig. 2C), in addition to disappearance of the full length form a truncated protein was detected in the knockout cells at the expected molecular weight for SHPRH lacking a RING domain. Accordingly, while knockout of the promoter of SHPRH in strategy TSS bears the risk of residual downstream transcription, potentially leading to low level expression of functional protein(s) that might contain most important domains of SHPRH, introduction of a premature stop codon in strategy PHD and downstream transcription may lead to residual SHPRH fragments with partial functions. In case of strategy RING, a protein lacking the RING domain but encompassing other functional domains of SHPRH is definitively produced, and a partial positive or dominant negative function of this truncated SHPRH can not be excluded. We therefore decided to use all three knockout sets for functional analyses.

2.1. Generation of SHPRH knockout cells by 3 different strategies

2.2. Effects of SHPRH inactivation on PCNA ubiquitination

SHPRH is a RING type E3 ligase encoded by a gene with 29 exons, spanning 47 kb. While no functional domain could be identified by homology search within the region encoded by exons 1–5, exons 6–13 code for SNF2 and PHD domains, while exons 24 and 25 code for the RING domain and exons 26–29 for a HELICc domain (Fig. 1A). In order to make sure that our SHPRH knockout impairs protein function(s), we decided to use three different strategies. In the strategy henceforth called TSS (transcriptional start site) (Fig. 1A and B), we removed a region starting 1000 bp upstream of the transcriptional start site and ending at the end of exon 4, aiming at inactivation of SHPRH transcription. In the strategy called PHD (Fig. 1A and C), we inserted a premature stop codon into exon 6, at the start of the sequence coding for the first SNF2 domain, and in addition deleted the subsequent exons 7, 8 and 9, aiming at inactivation of all functional domains of the protein. In the strategy RING (Fig. 1A and D), we introduced a premature stop codon into

SHPRH has been identified as an E3 ligase involved in PCNA polyubiquitination, which can be detected in ubiquitination reactions using purified proteins or cells overexpressing SHPRH [5,6,23,24]. However, reproducible detection of PCNA polyubiquitination in normal cells has proven exceedingly difficult [4]. It was recently reported, though, that treatment of cells with H2 O2 leads to rapid and proficient PCNA monoubiquitination [14]. Using this approach, we could also reproducibly detect substantial di- and triubiquitination of PCNA in DT40 cells (Suppl. Fig. 3). Applying this approach to the knockout sets generated before revealed two different issues. In undamaged cells, knockout via strategies TSS and PHD may potentially lead to some increase in basal PCNA monoubiquitination (Fig. 3A and B). This was not observed for strategy RING, a first warning that that the truncated protein produced here could have retained some residual function (Fig. 3C and data mentioned below). In cells treated with H2 O2 ,

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

65

Fig. 1. Inactivation of SHPRH in DT40 cells. (A) Protein and gene structure of SHPRH and indication of the deleted region for each gene targeting strategy. The stop codons introduced in frame for strategy PHD and RING are displayed. Exons distribution is to scale. (B–D) Gene targeting strategy for deletion of the promoter region (B), the PHD domain (C) and the RING domain (D). The wild type locus is depicted in the upper drawing and the lower one indicates the targeted integration of the resistance cassette of the targeting vector shown in the middle. (E) Southern blot of DT40Cre1 SHPRH wild type, two heterozygous and three knockout clones, respectively, using BclI and the probe shown in (B). The wild type allele generates a 5.6 kb fragment and the mutated allele a 3.9 kb fragment. (F) Southern blot of DT40V− SHPRH wild type, two heterozygous and seven knockout clones, respectively, using BclI and the probe shown in (C). The probe hybridizes to a 5.6 kb and 6.7 kb genomic fragment derived from the wild type and mutated alleles, respectivley. (G) Southern blot analysis of DT40V− SHPRH wild type cells, two heterozygous and three knockout clones, using PvuII and the probe shown in (D). The wild type allele generates a 6.4 kb fragment and the mutated alleles have a size of 7.4 kb and 10.6 kb, depending on the resistance cassette used. (H–J) Loss of the targeted exons in the mRNA of the different clones analyzed in (E–G) was confirmed by RT-PCRs. Primer pairs outside of the targeted region indicate residual transcription of the untargeted region. SNF2: sucrose non fermentable; H15: histone 1 and 5; PHD: plant homeodomain; RING: really interesting new gene; HELICc: helicase superfamily c-terminal domain; UTR: untranslated region, arrows indicate primer binding sites.

we observed some variability in between clones of one knockout set (e.g. Fig. 3C) as well as in different experiments for the same cells (Suppl. Fig. 4). Overall, the sensitivity of the assay system only allows to conclude that the function of SHPRH in DNA damage-induced PCNA polyubiquitination can clearly be replaced by another E3 ligase even in cells lacking HLTF.

2.3. Effects of SHPRH inactivation on DNA repair To assess the function of SHPRH in DNA damage tolerance, we performed colony survival assays in methylcellulose medium. We could not confirm the recently reported function of SHPRH in MMS survival [6] in any of our 3 knockout sets, even though an effect

66

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

Fig. 2. Loss of SHPRH protein expression in SHPRH−/− cells. Western blot analyses of SHPRH expression in the clones derived from the knockout strategies TSS (A), PHD (B) and RING (C). For strategy TSS, comparison of endogenous expression levels with those obtained by the expression vector are shown. For strategy PHD, the clones are shown as two sets according to the two heterozygous clones from which they were generated. The estimated band of SHPRH (192 kDa) is marked with an arrow and the asterisk indicates an unspecific band. A truncated SHPRH protein (∼159 kDa) appears in strategy RING.

of inactivation of the Rad6 pathway, using Rad18−/− or PCNAK164R mutants, was clearly seen (Fig. 4A–C). We additionally used cisplatin treatment, as this drug shows a stronger effect in mutants in the Rad6 pathway [29], but once again no effect of SHPRH inactivation was detectable (Fig. 4D–F). Finally, we used etoposide treatment, as inactivation of Rad5 in yeast has been shown to affect double strand break repair [30]. Intriguingly, a reproducible effect was detected here in different knockout clones and experiments in case of strategy TSS (Fig. 4G) and to a lesser extend in case of PHD (Fig. 4H). Deletion of the RING domain alone did not

result in etoposide sensitivity (Fig. 4I), once again indicating that this truncated protein retains at least partial functionality. Interestingly, overexpression of SHPRH in wild type (Fig. 2A) as well as TSS knockout cells also interferes with etoposide-, but not MMSresistence of the cells (Suppl. Fig. 5), implying that adequate SHPRH levels are required for this function of the protein. We thus conclude that inactivation of SHPRH does not affect sensitivity to drugs requiring other components of the Rad6 pathway, which is potentially due to redundancy of SHPRH with another E3 ligase mediating DNA damage-induced PCNA polyubiquitination.

Fig. 3. Effects of SHPRH inactivation on PCNA ubiquitination. Induction of PCNA ubiquitination upon treatment with H2 O2 . (A) For targeting strategy TSS, three different exposures are shown. (B) For targeting strategy PHD, clones were divided into two sets according to the heterozygous clone from which they were generated. Two different blot exposures are shown for each set. (C) Two different blot exposures are shown for targeting strategy RING. ∅: untreated. Ub1: mono-ubiquitination. Ub2: di-ubiquitination. Ub3: tri-ubiquitination. exp. = exposure.

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

67

Fig. 4. Effects of SHPRH inactivation on drug survival. Clonogenic survival assays of the SHPRH targeted clones upon treatment with methyl methanesulfonate (MMS) (A–C), cisplatin (D–F) and etoposide (G–I). In case of MMS and cisplatin treatment, Cre1 Rad18−/− or V− PCNAK164R cells were used as positive controls. Each point represents the mean of 2 independent experiments (or 3 assays for etoposide treatment in TSS and PHD cells). For the etoposide assay using PHD cells, only 4 representative clones are shown for clarity. The error bar represents the standard deviation, shown for clarity only for the positive error.

However, etoposide survival is clearly affected by SHPRH expression. In agreement with this, SHPRH knockout cells were also sensitive to other double strand break inducing drugs, such as Doxorubicin (Suppl. Fig. 6A), but not to the UV-mimetic 4Nitroquinoline 1-oxide (4-NQO, Suppl. Fig. 6B). 2.4. Influence of SHPRH on immunoglobulin diversification To assess a potential function of SHPRH in DNA double strand break repair in a physiological context, we used the knockout strategy TSS in DT40Cre1 cells performing Ig gene conversion. These cells harbor a frameshift mutation in the rearranged Ig␭ light chain gene which may be reverted by gene conversion with upstream pseudogenes, thus allowing a measurement of surface IgM expression gain as a surrogate marker for Ig gene conversion activity [31]. Analysis of Ig restoration in multiple individual subclones generated from the cells with the different genotypes did not reveal a substantial change in overall Ig gene conversion activity (Fig. 5A and

Suppl. Fig. 7A). Sequencing of different representative subclones also revealed no difference in the number of overall Ig gene conversion events (Fig. 5B and C). However, the total number of nucleotide changes due to Ig gene conversion was significantly increased in the SHPRH knockout cells in a sequence analysis for two clones of each genotype (Fig. 5D and E), which may be linked to a moderate increase in Ig gene conversion tract length (Fig. 5F and G and Suppl. Fig. 8). SHPRH might thus moderately increase the number of tracts containing more mismatches during Ig gene conversion events, leading to more base changes even though overall event number is not altered. To investigate whether SHPRH might also play a role in somatic hypermutation, we used the DT40 V− cells carrying an SHPRH inactivation via strategy PHD. In these cells, deletion of the pseudogenes required for Ig gene conversion leads to processing of AID-induced lesions by mutagenesis. The cells carry a functional Ig␭ gene and hence surface IgM receptor, and AID-triggered deleterious mutations lead to sIgM loss as a surrogate marker of somatic

68

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

Fig. 5. Influence of SHPRH on Ig gene conversion. (A) Ig gene conversion measured by sIgM restoration in subclones of the indicated genotype after 2 weeks. P-values (Student’s t-test) are given only in case of significance. (B) Schematic illustration of the variable region of the Ig light chain. The sequenced region is indicated by an arrow for the sequencing primer and a subsequent stretch of 698 bp. The complementary determining regions (CDR1–3) are indicated by dark gray boxes. L: leader. V: variable. J: joining. (C) The total number of gene conversion events of representative subclones of the gene conversion assay shown in (A). The amount of analyzed sequences is indicated below. Significance analysis: Fisher’s Exact Test. (D) Absolute number of total substitutions in the sequences shown in (C). Significance analysis: Fisher’s Exact Test. (E) The total number of substitutions shown in (D) is classified into gene conversion mutations (GCM), ambiguous mutations (AM) and point mutations (PM). Significance analysis: Fisher’s Exact Test. (F) The minimum (min.) length of gene conversion events as defined from the first base change to the last base change. (G) The maximum (max.) length of Ig gene conversion events as characterized by the maximum length of homology between the analyzed sequence and the most likely pseudogene with which recombination occurred. P-values are given (Student’s t-test) for (F and G).

hypermutation activity [32]. Analysis of surface Ig loss in multiple subclones generated from cells of the different genotypes did not reveal substantial effects of SHPRH (Fig. 6A and Suppl. Fig. 7B). Likewise, sequencing of representative subclones did not indicate significant changes in either mutation frequency or mutagenesis pattern in SHPRH knockout cells (Fig. 6B, C and D). We thus conclude that SHPRH does not play a role in somatic hypermutation even in cells lacking HLTF, consistent with a redundant role in DNA damage-induced PCNA ubiquitination and base damage survival in these cells.

3. Discussion In the present study, we used three different strategies to inactivate SHPRH, an E3 ligase implicated in PCNA polyubiquitination. As parental cells for the knockout, we used chicken DT40 B cells that do not harbor a HLTF ortholog. The results of our study are summarized in Table S6. We show that deletion of only the RING domain does not cause the same phenotype as inactivation of larger regions of the SHPRH gene. The latter form of SHPRH inactivation causes no difference in H2 O2 -induced PCNA polyubiquitination. Survival

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

69

Fig. 6. Influence of SHPRH on Ig somatic hypermutation. (A) sIgM loss in subclones of the indicated genotypes after 2 weeks. P-values (Student’s t-test) are given in case of significance and the number of analyzed subclones is indicated below. (B) Schematic illustration of the variable region of the Ig light chain, as given in Fig. 5B, indicating deletion of the pseudogenes. (C) Sequence analysis of representative subclones of the Ig loss assay shown in (A). The number of point mutations per DNA sequence is given in the periphery of the chart segments and the total number of analyzed sequences is given in the middle of the pie chart diagram. Significance analysis: Student’s t-test. (D) The pattern of total nucleotide substitutions for each genotype is shown on the left side and the percentages of total mutations is given on the right.

in the presence of MMS and cisplatin, two drugs that greatly challenge cells lacking other components of the Rad6 pathway [29], is not impaired. In contrast, drugs causing DNA double strand breaks lead to decreased survival of SHPRH knockout cells. While the overall frequency of Ig gene conversion is not affected, base exchanges during this process, which is at least in part based on DNA double strand break repair [33], appear to be moderately increased, while mutagenesis during somatic hypermutation (which does not require double strand break repair [34]) is not affected. Our study highlights the complexity of E3 ligase function in the Rad6 pathway, and their contributions to other mechanisms of DNA repair. E3 ligases belong to two main functional families, RING and HECT [35–37]. While in the HECT family, the ubiquitin is covalently bound to the E3 ligase for transfer to the target protein, RING type E3 ligases employ the RING domain to foster proximity of the ubiquitinating E2 enzyme and its target [38–40]. As RING E3 ligases may contain other binding sites for either E2 proteins or the target protein, and are often multidomain proteins encoded by several exons, their inactivation is indeed a challenge. Truncated proteins formed upon alternative transcription initiation or splicing may be either partially functional or dominant negative. In the present study, for example, we could not detect any phenotype in an SHPRH mutant lacking only the supposedly crucial C-terminal RING domain, while in a previous study of our lab an analogous Rad18 protein lacking the N-terminal RING domain showed phenotypes despite in-frame splicing over the inactivated exons [25]. Such differential effects of the knockout strategy must be considered when interpreting data on RING type E3 ligase inactivation.

Irrespective of this caveat, our data reveal new functions of SHPRH in DNA repair, and add to findings supporting a high redundancy of E3 ligases involved in PCNA polyubiquitination [19]. The cells we used do not contain HLTF, i.e. no such gene was found in the chicken genome despite existent synteny at the respective chromosomal position. This is interesting as such, suggesting that even HLTF functions not related to DNA repair [41–43] became redundant during avian evolution. Despite the lack of HLTF and different strategies for SHPRH inactivation, we can detect efficient DNA damage-induced PCNA polyubiquitination in SHPRH knockout cells. Certainly, other E3 ligases may support this reaction in this cellular setting, strengthening the conclusion of an earlier report in mouse cells [19]. To date, the identity of the E3 ligase that apparently mediates most PCNA polyubiquitination upon treatment with genotoxic drugs remains unclear, despite a considerable and long term quest for functional Rad5 orthologs in vertebrates. Its identification will certainly be a challenge, given that most cell systems studied do contain SHPRH and HLTF, and it is unknown to which extend the ligases may really replace each other. As another obstacle in such studies, reproducible detection of PCNA polyubiquitination in normal cells is exceedingly difficult [4]. We induced highly efficient PCNA ubiquitination by H2 O2 [14], as we were not able to obtain equivalent or detectable levels of PCNA polyubiquitination with other drugs (data not shown). Given that the E3 ligases involved in PCNA ubiquitination may perform DNA damage-specific functions [4], this limitation may preclude their adequate identification. Moreover, western blotting does not

70

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

detect PCNA polyubiquitination in its strict sense, but rather di- and triubiquitination at best. The minimum ubiquitin chain length on PCNA required for activation of the error-free damage bypass is not known. However, en bloc transfer of the ubiquitins [24] may lead to high molecular weight modified PCNA, precluding detection of polyubiquitination. In this context, it may interesting that we detect a potential slight increase in spontaneous PCNA monoubiquitination in cells lacking SHPRH function, despite being unable to detect polyubiqutination bands in their SHPRH-proficient parental cells in the steady state situation. While it is possible that SHPRH directly affects spontaneous PCNA monoubiquitination, this finding could also suggest that a certain portion of PCNA is polyubiquitinated in a high molecular weight form in normal culture cells depending on SHPRH function, e.g. due to replication intermediates [44] or other types of spontaneous DNA damage. As both, the consequences of PCNA polyubiquitination as well as clear phenotypes of its inactivation are barely known, further studies of this phenomenon will be challenging, though. In contrast to others, we do not detect a clear effect of SHPRH inactivation on MMS survival [6]. Also, cisplatin survival that clearly requires the Rad6 pathway is not affected. As a note of caution, one should consider that redundancy in E3 ligase function in the Rad6 pathway may be mediated by other proteins in chicken than in mammals, as HLTF is lacking. As HLTF has been shown to also support PCNA monoubiquitination in one study with mammalian cells [4], such considerations require careful choice of the right organism/cell system in the study of Rad6 pathway functions. Adding to these thoughts, expression of proteins involved in the Rad6 pathway may be cell type- or organ-specific [45], presumably because not all cells should be predestined to initiate potentially mutagenic bypass of DNA damage during replication. In any case, though, our finding of defective etoposide survival in SHPRH knockout cells is interesting, as it adds to previous findings implicating Rad5 function in double strand break repair [30]. In this case, excessive PCNA monoubiquitination was found to interfere with homologous recombination. We also detected defective etoposide survival in the mutants apparently showing a moderately higher spontaneous PCNA monoubiquitination. As a potential mechanistic approach to decipher this phenomenon, we see higher tolerance for mismatches during Ig gene conversion in SHPRH knockout cells. PCNA ubiquitination has not been implicated in general mismatch repair, but PCNA monoubiquitination was shown to induce non-canonical mismatch repair functions that may contribute to mutagenesis [13]. It is tempting to speculate that mismatch recognition or exclusion during homologous recombination could be impaired in SHPRH knockout cells. One must consider, though, that intrinsic mismatch tolerance is quite high in DT40 cells, as several mismatches are tolerated in even short conversion tracts as well as in gene targeting approaches [46,47]. It will thus be interesting to investigate in a mammalian cell system whether the fidelity of homologous recombination is affected upon inactivation of SHPRH or rather PCNA polyubiquitination, once the technical obstacles described above are overcome. Our study reveals a considerable complexity in the analysis of E3 ligases involved in PCNA ubiquitination and other functions. However, it is worthwhile considering how and why redundancy in these enzymes may have evolved, given that our genome encodes for over 600 E3 ligases [48] to ubiquitinate most cellular proteins. Similarly to phosphorylation, targeted ubiquitination is a versatile tool to change protein function. Hence, triggering modification of the same lysine residue by more than one pathway, employing more than one E3 ligase, may facilitate to adequately exploit its full potential in cellular regulation.

4. Materials and methods Antibodies. Antibodies used for immunoblotting were: anti-AID (EK2 5G9, Cell Signaling), anti-HA (3F10, Roche) anti-PCNA (ab-29, Abcam), anti-SHPRH (HPA034854, Sigma–Aldrich), anti-vinculin (BZL03106, Biozol), anti-tubulin (ab4047, abcam) and anti-actin (A2066, Sigma–Aldrich). Cell culture. The DT40 cell lines were cultured at 41 ◦ C and 5% CO2 . The RPMI 1640 medium (Invitrogen; 21875091) was supplemented with 10% FCS (Biochrom AG), 100 ␮g/␮l penicillin/streptomycin, 2 mM glutamine, 1 mM sodium pyruvate (all GIBCO), 1% chicken serum (Sigma) and 0,1 ␮M 2-mercaptoethanol (Sigma). Transfections were performed with a Gene Pulser Xcell (BioRad) set at 50 ␮F and 800 V. Inactivation and overexpression of SHPRH in DT40 cells. The homologous arms of targeting vectors for the different knockout strategies of SHPRH were amplified from genomic DT40 DNA (primers: Suppl. Table 5). The PCR products were restricted using the enzymes annotated in Suppl. Table 5 and cloned into a pBluescript II KS (+/−) vector (Fermentas). Three different loxP-flanked resistance cassettes (from ploxpuro, ploxbsr and ploxgpt) [49] were inserted between the arms using BamHI (strategy TSS and RING) or BclI (strategy PHD) restriction sites. In DT40V− , the first SHPRH allele was disrupted with a blasticidin based and the second allele with a mycophenolic acid (gpt) based targeting vector. In DT40Cre1, the first SHPRH allele was disrupted with a puromycin based and the second allele with a blasticidin based targeting vector. Drug resistant single cell clones were selected with 0,8 ␮g/ml puromycin (Sigma–Aldrich), 5 ␮g/ml blasticidin S HCl (Mobitec GmbH), 30 ␮g/ml gpt (VWR), respectively, or with 2 of the appropriate selections simultaneously, if necessary. Drug resistant subclones were screened for targeted integration by PCR (Expand Long Template PCR Sytem, Roche) using a primer binding upstream of the homologous arm in combination with primers binding in the respective resistance cassette (Suppl. Table 5). Gene disruption was also confirmed by southern blot analysis using enriched cellular DNA cleaved with BclI for strategy TSS and PHD or PvuII for strategy RING. The different probes were amplified from genomic DT40 DNA (primers: Suppl. Table 5) and the membrane was incubated with the particular probe at 65 ◦ C. Recycling of the loxP-flanked selectable marker gene was achieved by overnight culture in 1 ␮M 4-hydroxy tamoxifen (H7904, SIGMA), followed by limiting dilution subcloning [49]. For overexpression, the coding sequence of DT40 SHPRH was amplified and tagged with HA on the N-terminus (primers: Suppl. Table 5) using Phusion polymerase (Finnzymes) and cut with EcoRV at the 5 - and 3 end for cloning into pExpress [49]. Subsequently, the cDNA expression cassette was excised with SpeI and cloned into the NheI site of ploxpuro. For RT-PCR and quantitative RT-PCR (qRT-PCR), RNA was isolated (RNeasy Mini Kit, Qiagen) and reversely transcribed using the First Strand cDNA Synthesis Kit (Roche). For RT-PCR, primers amplifying from exon 4 to 8 and exons 25 to 28 were used (Suppl. Table 5). For qRT-PCR, primers amplifying from exons 25 to 28 and amplifying the glyceraldehyde 3-phosphate dehydrogenase gene (GAPDH) were used (Suppl. Table 5). The expression levels of the housekeeping gene GAPDH serve as control to determine relative SHPRH expression levels. Analysis of drug survival and immunoblot analysis. A colonyforming assay was used to analyze the survival in the presence of etoposide (Sigma, USA), doxorubicin (Santa Cruz, Germany), cisplatin (Ribosepharm, Gräfelfing) or methyl methanesulfonate (MMS) (Sigma, USA) as previously described [50]. In short, sensitivity to the drugs was measured plating cells onto methylcellulose containing the drug, and colonies were counted after 10–14 days of culture. PCNA mono- and polyubiquitination was induced by

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

incubation of 6 × 105 –8 × 105 cells with 1 mM hydrogen peroxide for 40 minutes. Cell lysates for immunoblotting were prepared by washing the cells once with PBS, resuspending in lysis buffer (20 mM Hepes, 350 mM NaCl, 20% Glycerin, 1 mM MgCl2 , 0.5 mM EDTA, 0.1 mM EGTA, 1% NP-40) supplemented with Complete Protease Inhibitor (Roche) and Phosphatase Inhibitor (Roche). Incubation for 15 min on ice was followed by sonification and centrifugation at 14.000 rpm for 10 min. Analysis of Ig diversification in DT40 cells. For analysis of somatic hypermutation in DT40V− and Ig gene conversion in DT40Cre1, the different cell lines were subcloned by limiting dilution and 24–48 clones per cell line were cultured for about 2 to 4 weeks. For FACS analysis, cells were stained with anti-chickenIgM-PE (8310-09, Southern Biotech). For sequence analyses of somatic hypermutation (in DT40V− ) and Ig gene conversion (in DT40Cre1), genomic DNA was isolated from representative subclones after 7 weeks and 4 weeks of culture, respectively. To amplify the rearranged light chain ␭ locus, Phusion polymerase (Finnzymes) and ␭-primer (Suppl. Table 5) were used. PCR products were cloned into the pGEM® -T vector (Promega) and sequenced with the primer 5 -GAG CGC AGG GAG TTA TTT GCA TAG-3 . Sequence alignment was performed with Geneious software in order to identify changes from the respective parental sequences in each clone. The SHMTool was used to acquire SHM frequencies and pattern analyses in DT40V− [51] (http://scb.aecom.yu.edu/cgi-bin/p1). Sequence analysis of Ig gene conversion was performed as previously described [50,52]. Conflict of interest statement None. Acknowledgements We thank Angelika Schmidt for expert technical assistance, C. Kosan and M. Godmann for stimulating discussion and all members of the Jungnickel lab for critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (JU2690/1-2 and a Heisenberg fellowship to B.J.). Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.dnarep. 2014.09.010. References [1] K.Y. Lee, K. Myung, PCNA modifications for regulation of post-replication repair pathways, Mol. Cells 26 (1) (2008) 5–11. [2] K. Watanabe, et al., Rad18 guides poleta to replication stalling sites through physical interaction and PCNA monoubiquitination, EMBO J. 23 (19) (2004) 3886–3896. [3] P.L. Andersen, F. Xu, W. Xiao, Eukaryotic DNA damage tolerance and translesion synthesis through covalent modifications of PCNA, Cell Res. 18 (1) (2008) 162–173. [4] J.R. Lin, et al., SHPRH and HLTF act in a damage-specific manner to coordinate different forms of postreplication repair and prevent mutagenesis, Mol. Cell 42 (2) (2011) 237–249. [5] A. Motegi, et al., Polyubiquitination of proliferating cell nuclear antigen by HLTF and SHPRH prevents genomic instability from stalled replication forks, Proc. Natl. Acad. Sci. U.S.A. 105 (34) (2008) 12411–12416. [6] A. Motegi, et al., Human SHPRH suppresses genomic instability through proliferating cell nuclear antigen polyubiquitination, J. Cell Biol. 175 (5) (2006) 703–708. [7] A. Ciccia, et al., Polyubiquitinated PCNA recruits the ZRANB3 translocase to maintain genomic integrity after replication stress, Mol. Cell 47 (3) (2012) 396–409. [8] C. Hoege, et al., RAD6-dependent DNA repair is linked to modification of PCNA by ubiquitin and SUMO, Nature 419 (6903) (2002) 135–141.

71

[9] L.J. Simpson, et al., RAD18-independent ubiquitination of proliferating-cell nuclear antigen in the avian cell line DT40, EMBO Rep. 7 (9) (2006) 927–932. [10] K. Terai, et al., CRL4(Cdt2) E3 ubiquitin ligase monoubiquitinates PCNA to promote translesion DNA synthesis, Mol. Cell 37 (1) (2010) 143–149. [11] S. Zhang, et al., PCNA is ubiquitinated by RNF8, Cell Cycle 7 (21) (2008) 3399–3404. [12] L. Geng, C.J. Huntoon, L.M. Karnitz, RAD18-mediated ubiquitination of PCNA activates the Fanconi anemia DNA repair network, J. Cell Biol. 191 (2) (2010) 249–257. [13] J. Pena-Diaz, et al., Noncanonical mismatch repair as a source of genomic instability in human cells, Mol. Cell 47 (5) (2012) 669–680. [14] A. Zlatanou, et al., The hMsh2-hMsh6 complex acts in concert with monoubiquitinated PCNA and Pol eta in response to oxidative DNA damage in human cells, Mol. Cell 43 (4) (2011) 649–662. [15] D. Szuts, et al., Role for RAD18 in homologous recombination in DT40 cells, Mol. Cell. Biol. 26 (21) (2006) 8032–8041. [16] K. Watanabe, et al., RAD18 promotes DNA double-strand break repair during G1 phase through chromatin retention of 53BP1, Nucleic Acids Res. 37 (7) (2009) 2176–2193. [17] L. Ting, H. Jun, C. Junjie, RAD18 lives a double life: Its implication in DNA doublestrand break repair, DNA Repair (Amst) 9 (12) (2010) 1241–1248. [18] A.G. Lloyd, et al., Effect of DNA repair protein Rad18 on viral infection, PLoS Pathog. 2 (5) (2006) pe40. [19] P.H. Krijger, et al., HLTF and SHPRH are not essential for PCNA polyubiquitination, survival and somatic hypermutation: existence of an alternative E3 ligase, DNA Repair (Amst) 10 (4) (2011) 438–444. [20] I. Unk, et al., Role of yeast Rad5 and its human orthologs, HLTF and SHPRH in DNA damage tolerance, DNA Repair (Amst) 9 (3) (2010) 257–267. [21] R. Sood, et al., Cloning and characterization of a novel gene, SHPRH, encoding a conserved putative protein with SNF2/helicase and PHD-finger domains from the 6q24 region, Genomics 82 (2) (2003) 153–161. [22] I. Unk, et al., Human HLTF functions as a ubiquitin ligase for proliferating cell nuclear antigen polyubiquitination, Proc. Natl. Acad. Sci. U.S.A. 105 (10) (2008) 3768–3773. [23] I. Unk, et al., Human SHPRH is a ubiquitin ligase for Mms2-Ubc13-dependent polyubiquitylation of proliferating cell nuclear antigen, Proc. Natl. Acad. Sci. U.S.A. 103 (48) (2006) 18107–18112. [24] Y. Masuda, et al., En bloc transfer of polyubiquitin chains to PCNA in vitro is mediated by two different human E2–E3 pairs, Nucleic Acids Res. 40 (20) (2012) 10394–10407. [25] J. Bachl, I. Ertongur, B. Jungnickel, Involvement of Rad18 in somatic hypermutation, Proc. Natl. Acad. Sci. U.S.A. 103 (32) (2006) 12081–12086. [26] H. Arakawa, et al., A role for PCNA ubiquitination in immunoglobulin hypermutation, PLoS Biol. 4 (11) (2006) pe366. [27] P. Langerak, et al., Somatic hypermutation of immunoglobulin genes: lessons from proliferating cell nuclear antigenK164R mutant mice, Philos. Trans. R. Soc. Lond. B: Biol. Sci. 364 (1517) (2009) 621–629. [28] S. Roa, et al., Ubiquitylated PCNA plays a role in somatic hypermutation and class-switch recombination and is required for meiotic progression, Proc. Natl. Acad. Sci. U.S.A. 105 (42) (2008) 16248–16253. [29] K. Nojima, et al., Multiple repair pathways mediate tolerance to chemotherapeutic cross-linking agents in vertebrate cells, Cancer Res. 65 (24) (2005) 11704–11711. [30] S. Moertl, et al., Regulation of double-stranded DNA gap repair by the RAD6 pathway, DNA Repair (Amst) 7 (11) (2008) 1893–1906. [31] H. Arakawa, J. Hauschild, J.M. Buerstedde, Requirement of the activationinduced deaminase (AID) gene for immunoglobulin gene conversion, Science 295 (5558) (2002) 1301–1306. [32] H. Arakawa, Immunoglobulin gene conversion and hypermutation assay by FACs, Subcell Biochem. 40 (2006) 351–352. [33] H. Arakawa, J.M. Buerstedde, Immunoglobulin gene conversion: insights from bursal B cells and the DT40 cell line, Dev. Dyn. 229 (3) (2004) 458–464. [34] C. Rada, J.M. Di Noia, M.S. Neuberger, Mismatch recognition and uracil excision provide complementary paths to both Ig switching and the A/T-focused phase of somatic mutation, Mol. Cell 16 (2) (2004) 163–171. [35] M.B. Metzger, V.A. Hristova, A.M. Weissman, HECT and RING finger families of E3 ubiquitin ligases at a glance, J. Cell Sci. 125 (Pt 3) (2012) 531–537. [36] M. Scheffner, S. Kumar, Mammalian HECT ubiquitin-protein ligases: biological and pathophysiological aspects, Biochim. Biophys. Acta 1843 (1) (2014) 61–74. [37] M.B. Metzger, et al., RING-type E3 ligases: master manipulators of E2 ubiquitinconjugating enzymes and ubiquitination, Biochim. Biophys. Acta 1843 (1) (2014) 47–60. [38] H.C. Ardley, P.A. Robinson, E3 ubiquitin ligases, Essays Biochem. 41 (2005) 15–30. [39] A. Plechanovova, et al., Structure of a RING E3 ligase and ubiquitin-loaded E2 primed for catalysis, Nature 489 (7414) (2012) 115–120. [40] E. Maspero, et al., Structure of a ubiquitin-loaded HECT ligase reveals the molecular basis for catalytic priming, Nat. Struct. Mol. Biol. 20 (6) (2013) 696–701. [41] R.A. Helmer, et al., Role of helicase-like transcription factor (hltf) in the G2/m transition and apoptosis in brain, PLoS ONE 8 (6) (2013) pe66799. [42] R.A. Helmer, et al., Helicase-like transcription factor (hltf) regulates g2/m transition, wt1/gata4/hif-1a cardiac transcription networks, and collagen biogenesis, PLoS ONE 8 (11) (2013) pe80461. [43] G. Debauve, et al., The helicase-like transcription factor and its implication in cancer progression, Cell. Mol. Life Sci. 65 (4) (2008) 591–604.

72

N.-S. Tomi et al. / DNA Repair 24 (2014) 63–72

[44] P. Stelter, H.D. Ulrich, Control of spontaneous and damage-induced mutagenesis by SUMO and ubiquitin conjugation, Nature 425 (6954) (2003) 188–191. [45] S. Velasco-Miguel, et al., Constitutive and regulated expression of the mouse Dinb (Polkappa) gene encoding DNA polymerase kappa, DNA Repair (Amst) 2 (1) (2003) 91–106. [46] J.M. Buerstedde, et al., Light chain gene conversion continues at high rate in an ALV-induced cell line, EMBO J. 9 (3) (1990) 921–927. [47] J.M. Buerstedde, S. Takeda, Increased ratio of targeted to random integration after transfection of chicken B cell lines, Cell 67 (1) (1991) 179–188.

[48] J.G. Marblestone, et al., Comprehensive ubiquitin E2 profiling of ten ubiquitin E3 ligases, Cell Biochem. Biophys. 67 (1) (2013) 161–167. [49] H. Arakawa, D. Lodygin, J.M. Buerstedde, Mutant loxP vectors for selectable marker recycle and conditional knock-outs, BMC Biotechnol. 1 (2001) p7. [50] S. Frankenberger, et al., Checkpoint kinase 1 negatively regulates somatic hypermutation, Nucleic Acids Res. 42 (6) (2014) 3666–3674. [51] T. Maccarthy, et al., SHMTool: a webserver for comparative analysis of somatic hypermutation datasets, DNA Repair (Amst) 8 (1) (2009) 137–141. [52] J.E. Sale, et al., Ablation of XRCC2/3 transforms immunoglobulin V gene conversion into somatic hypermutation, Nature 412 (6850) (2001) 921–926.