Identification of BC005512 as a DNA Damage Responsive Murine Endogenous Retrovirus of GLN Family Involved in Cell Growth Regulation

Genotoxicity assessment is of great significance in drug safety evaluation, and microarray is a useful tool widely used to identify genotoxic stress responsive genes. In the present work, by using oligonucleotide microarray in an in vivo model, we identified an unknown gene BC005512 (abbreviated as BC, official full name: cDNA sequence BC005512), whose expression in mouse liver was specifically induced by seven well-known genotoxins (GTXs), but not by non-genotoxins (NGTXs). Bioinformatics revealed that BC was a member of the GLN family of murine endogenous retrovirus (ERV). However, the relationship to genotoxicity and the cellular function of GLN are largely unknown. Using NIH/3T3 cells as an in vitro model system and quantitative real-time PCR, BC expression was specifically induced by another seven GTXs, covering diverse genotoxicity mechanisms. Additionally, dose-response and linear regression analysis showed that expression level of BC in NIH/3T3 cells strongly correlated with DNA damage, measured using the alkaline comet assay,. While in p53 deficient L5178Y cells, GTXs could not induce BC expression. Further functional studies using RNA interference revealed that down-regulation of BC expression induced G1/S phase arrest, inhibited cell proliferation and thus suppressed cell growth in NIH/3T3 cells. Together, our results provide the first evidence that BC005512, a member from GLN family of murine ERV, was responsive to DNA damage and involved in cell growth regulation. These findings could be of great value in genotoxicity predictions and contribute to a deeper understanding of GLN biological functions.


Introduction
Genotoxicity assessment plays an important role in both toxicity screening during early drug discovery and regulatory drug safety evaluation in the preclinical stage [1]. Although a great number of genotoxicity assays have been developed, there is still a requirement for tests with both high specificity and sensitivity [2]. The use of microarray technology in toxicology, known as toxicogenomics, can potentially identify novel genotoxicity biomarkers and provide mechanistic insights into the mode of action of genotoxic compounds [3,4,5,6,7,8]. We identified an unknown gene BC005512 (official full name: cDNA sequence BC005512), whose expression was specifically induced by genotoxins (GTXs) but not by non-genotoxins (NGTXs) in an in vivo microarray study. Elevated expression of BC005512 has been reported previously in thymocytes of Parp-2 deficient mice [9], suggesting that it is relevant to DNA damage. Further analysis of this gene uncovered that it is a member of the GLN family of murine endogenous retrovirus (ERV).
ERV sequences, most probably originating from infections of germ-line cells by ancient exogenous retroviruses during evolution [10], account for approximately 8% of the human genome [11] and 10% of the mouse genome [12]. ERVs were once thought to be junk DNA, but a number of studies have shown that some have important physiological roles [13,14,15] or are implicated in certain diseases [16,17]. Several studies have reported elevated expression of ERVrelated sequences in hepatocarcinogen treated rodents [18,19].
The GLN family, designated due to an unusual primer-binding site sequence corresponding to tRNA Gln , is one of a number of murine ERV families. It was first identified over two decades ago [20], but remains little-studied [21,22]. The relationship between GLN and genotoxic stress and the biological function of GLN family members are largely unknown. Here we report that BC005512, a member of the GLN family of murine ERV, was responsive to DNA damage and involved in regulation of cell growth.

Selection of specific and sensitive genotoxic stress responsive genes using microarray
Microarray is a powerful way of examining genomic scale gene expression changes. To identify specific and sensitive genotoxic stress inducible genes, we carried out an in vivo microarray study specifically investigating liver tissue in B6C3F1 mice administered with seven well-characterized genotoxins (GTXs) and three nongenotoxins (NGTXs). Compounds with all negative data in regulatory genotoxicity assays (including Ames test, in vitro chromosome aberration test, mouse lymphoma assay and in vivo micronucleus test) were chosen as non-genotoxins. The dosage used for GTXs was selected based on data from in vivo transgenic mouse mutation assays, where significantly higher mutant frequencies were observed in liver tissue. The mutant frequency was determined as described previously [23]. While the dosage used for NGTXs was 1/2 LD 50 (Table 1). To study both early and late or sustained genotoxic stress responses, time points at 4 h, 20 h, 2 weeks and 4 weeks after treatment were chosen. To select genotoxic stress responsive genes, we adopted a self-defined weight scoring approach. Candidate genes were scored based on their specificity, sensitivity (including average ratio, positive condition, positive chemical and reverse change), statistical P value, basal expression level, and coefficient of variation (CV). A total score, considering all the above parameters, was finally calculated ( Table 2). Further analysis of the top ranked 50 genes by hierarchical clustering showed clear gene sets, whose expression could distinguish GTXs from NGTXs (Fig. 1A). These included some well-known DNA damage inducible genes e.g. p21 WAF1/Cip1 [24] and ccng1 [25]. The highest scoring gene was an unknown gene BC005512 (identified by probe set 1426936_at, Gene symbol: BC005512, official full name: cDNA sequence BC005512). Its expression was specifically induced by GTXs, but not by NGTXs, which was further confirmed by quantitative real-time PCR (Fig. 1B and 1C).

BC005512 is a member of the GLN family of murine endogenous retrovirus
To characterize BC005512 in the mouse database, we performed bioinformatics analysis. A nucleotide BLAST search using the genechip probe sequence (Fig. S1) as the query identified two cDNAs (GenBank accession no. BC005512 and BC062922). Analysis of both sequences using ORF finder (http://www.ncbi. nlm.nih.gov/gorf/gorf.html) and protein BLAST revealed that the putative proteins encoded by these sequences have great similarity to retrovirus related proteins encoded by env genes of mouse ERV (data not shown) [10]. To determine the relationship of BC005512 with mouse ERVs, we analyzed BC005512 and BC062922 sequences using an online CENSOR program (http://www. girinst.org/censor/index.php), which screens query sequences for interspersed repeats [26]. The sequence analysis classified them into ERV-Class I, and revealed strong similarities to MMERGLN_I, a sequence submitted as one copy of GLN family in Repbase (Fig. 2). These findings suggested that the sequence identified in the above microarray study, i.e. BC005512 (abbreviated as BC in the following text), was a member of the GLN family of murine ERV.

Expression of BC was specifically induced by genotoxins in NIH/3T3 cells
The microarray study showed that BC was induced by genotoxic stress in vivo. To further characterize the responsiveness of BC to GTXs in vitro, we tested another seven GTXs with different genotoxic mechanisms and two NGTXs (Table 3). For each GTX, 24 h IC 50 was determined and used to study the effects of treatment on BC expression. NIH/3T3, a well-characterized mouse embryonic fibroblast cell line, was used as the in vitro model system. In accordance with results obtained in vivo, all seven GTXs up-regulated expression of BC to varying degrees, while all three NGTXs tested had no obvious effects on BC expression in NIH/ 3T3 cells (Fig. 3). Similar results were also obtained in a mouse hepatoma cell line, Hepa 1-6 cells (Fig. S2).

Induced expression level of BC correlated with DNA damage in NIH/3T3 cells
Inducing DNA damage is one of many genotoxicity mechanisms. To further study the relationship between BC expression and DNA damage, we compared transcriptional expression level of BC and the extent of DNA damage using methyl methanesulfonate (MMS) as a DNA-reactive model compound [27]. MMS was chosen as it gave the strongest response in the in vitro BC induction assay (Fig. 3) and has been used extensively as a DNA damaging model agent. MMS modifies both guanine (to 7methylguanine) and adenine (to 3-methlyladenine) causing base mispairing and replication blocks, respectively [27]. DNA damage was indicated by olive tail moment in the alkaline comet assay [28]. As shown in Fig. 4A and 4B, MMS induced a concentrationdependent increase in both BC expression and DNA damage.
In addition to DNA-reactive chemicals, aneugens that directly target spindles but not DNA during chromosome segregation were another class of GTXs. To determine whether BC was responsive to aneuploidy, we examined the effects of two aneugens, colchicine [29] and paclitaxel [30], on BC expression. Chromosome abnormality was determined using the micronucleus test [31]. The highest concentration of paclitaxel or colchicine was limited to 48 h IC 50 . As expected, both colchicine and paclitaxel induced a dose-dependent increase in micronucleus formation ( Fig. 4F and 4I) but not in DNA damage except at very high concentrations, possibly due to nonspecific effects under cytotoxic conditions (bar graph in Fig. 4D and 4G, 4E and 4H). In accordance with MMS treatment, quantitative PCR analysis showed that expression level of BC was in parallel with DNA damage i.e., BC induction only occurred when DNA damage was observed, regardless of whether aneuploidy was induced (line graph in Fig. 4D and 4G). Furthermore, the linear regression analysis of data obtained from MMS, colchicine and paclitaxel revealed a strong correlation between expression level of BC and the extent of DNA damage (Fig. 4C).

Expression of BC could not be induced by GTXs and did not correlate with DNA damage in L5178Y cells
In addition to NIH/3T3 cells, L5178Y, a widely used mouse lymphoma cell line in in vitro genotoxicity assays was also adopted to investigate the effects of GTXs on BC expression and the relationship between BC expression and DNA damage. For each GTX, 24 h IC 50 was determined and used in L5178Y cells. Surprisingly, GTXs that substantially induce BC expression in NIH/3T3 cells did not show similar effects in L5178Y cells ( Fig. 5A and Fig. S3). Concordantly, expression level of BC did not correlate with DNA damage induced by MMS (Fig. 5B, 5C and 5D). Together, these results suggested that expression of BC could not be induced by GTXs and did not correlate with DNA damage in L5178Y cells.
6. Down-regulating BC expression suppressed cell growth in several mouse cell lines DNA damage triggers a variety of biological responses including the transcriptional activation of genes regulating DNA repair, cell cycle checkpoint and cell death [32]. The observation that BC was responsive to DNA damage raised the question whether BC was involved in these processes.   We first examined the effects of knocking-down BC expression on cell growth by RNA interference. Two BC siRNAs were designed; one nonsense siRNA and two corresponding scrambled siRNAs were used as negative controls. Quantitative PCR showed that at 48 h after siRNA transfection, the level of BC mRNA was markedly reduced in cells transfected with BC siRNAs (siBC.1 and siBC.2) compared with cells transfected with nonsense ( Fig. 6A) or scrambled siRNAs (Fig. S4A). Due to the lack of appropriate antibody against BC, to confirm the knock-down efficiency of BC siRNAs at the protein level, we generated a myc-tagged BC clone. Following co-transfection of the clone and BC siRNA into NIH/ 3T3 cells, the protein expression level of BC was monitored using  a myc-tag antibody. Meanwhile, we introduced two point mutations into the siBC.2 target region to generate a mutant clone. By co-transfecting this mutant clone and siBC.2 into NIH/ 3T3 cells, we further examined the specificity of siBC.2. As shown in Fig. S5, siBC.2 decreased the wild-type BC, but not the mutant BC protein levels, indicating that siBC.2 worked specifically and effectively at the protein level.
Cell growth analysis of NIH/3T3 cells transfected with BC siRNAs showed that the total cell number was markedly reduced compared with cells transfected with nonsense ( Fig. 6B) or scrambled siRNAs (Fig. S4B). Similar results were also obtained in a mouse hepatoma cell line Hepa 1-6 ( Fig. 6C and 6D), and a mouse glomerular mesangial cell line SV40 MES 13 ( Fig. 6E and  6F). These results suggested that knock-down of BC expression suppressed cell growth in several mouse cell lines.

Knock-down of BC induced G1/S phase arrest and inhibited cell proliferation in NIH/3T3 cells
To investigate whether BC affected cell growth due to its effect on cell proliferation or cell survival, we performed an EdU incorporation assay, which is similar to BrdU incorporation assay. As shown in Fig. 7A S4C).
To determine whether down-regulation of BC expression affected cell death, we examined apoptosis in NIH/3T3 cells by flow cytometry using Annexin V and propidium iodide double staining. Knock-down of BC did not induce apoptosis at 48 h after siRNA transfection (Fig. S6A). Consistently, no obvious sub-G1 peak was observed (Fig. 7C), indicating that down-regulation of BC did not induce apoptosis directly in NIH/3T3 cells. Nevertheless, the obvious apoptosis observed at 96 h after siRNA transfection (Fig. S6B) might be caused by cell cycle arrest, occurring at 48 h after transfection.
Taken together, these results suggested that down-regulation of BC induced G1/S arrest, and thus inhibited cell proliferation in NIH/3T3 cells.

Discussion
In this work, we have shown that BC005512, a member of the GLN family of murine ERV, was responsive to DNA damage. This conclusion was supported by several observations. Multiple well-characterized GTXs with a diversity of genotoxicity mechanisms induced elevated expression of BC both in vivo (mouse liver) and in vitro (NIH/3T3 cells). More specifically, a strong correlation   was found between expression level of BC and the extent of DNA damage. Besides the above findings, we provide the first evidence that BC was involved in cell growth regulation, suggesting that BC was biologically relevant to the DNA damage response. Recent studies have reported elevated expression of ERVs in many different types of human cancers or tumor cell lines [17]. UVB irradiation was reported to induce transcriptional activation of ERV sequences in human epidermal keratinocytes [33]. However, the relationship between GLN and DNA damage was not clear. This is the first report that demonstrates the involvement of a GLN family member in genotoxic stress, particularly in the DNA damage response. This quantitative DNA damage-responsive property of BC could be potentially useful in genotoxicity prediction.
In addition to BC, we identified another ERV with a high score (1455892_x_at, BB794742, mouse endogenous murine leukemia virus mRNA), whose expression was induced by GTXs (Table 2). A great number of human ERVs have been reported to have perfect p53 binding sites that not only regulate adjacent gene expression, but may also give ERVs the advantage of exiting the host cell for their own survival under stress conditions [14]. As discussed later, we also found that the long terminal repeat (LTR) of GLN has a putative p53 binding site. Combined with our findings, it would be of great interest to study the relationship between ERVs and stress responses, and determine whether ERVs have intrinsic advantages in becoming potential genotoxicity biomarkers.
In the microarray study, BC was initially identified using selfdefined weight scoring. In our scoring system, well-known DNA damage responsive genes such as p21 [24] and ccng1 [25] were also assigned high scores, indicating that this method worked effectively in identifying genotoxic stress inducible genes. Although the basal expression level was low, BC had priority over other candidate genes in its high specificity and sensitivity (Table 2). However, it should be noted that this would not exclude the possibility that other genes could become potential genotoxicity biomarkers.
Transcriptional expression of BC was induced by GTXs in NIH/3T3 cells, but not in L5178Y cells. We inferred that this discrepancy might be due to a lack of functional p53 in L5178Y cells [34]. A search for transcriptional factor binding sites revealed that GLN LTR has a putative p53 binding site (Fig. S7A). In a preliminary study on the mechanisms of BC expression regulation in NIH/3T3 cells, we found that the wild-type LTR had great promoter/enhancer activity, even stronger than the positive control using luciferase reporter gene assay (Fig. S7B). However, when the putative p53 binding site was deleted, the promoter/ enhancer activity dropped significantly (Fig. S7B). To address the role of p53 in BC induction, we examined the effects of treatment with pifithrin-alpha (PFT-a), a widely used p53 inhibitor [35], on BC expression in NIH/3T3 cells. Consistent with our expectations, treatment with PFT-a decreased ETOP or MMS induced BC expression (Fig. S8). These results indicated that p53 might play a role in BC induction following GTXs treatment.
There are many genotoxicity mechanisms and the model compounds selected in this work could not be expected to cover all of them. More model compounds with diverse genotoxicity mechanisms should be tested to further examine the sensitivity and specificity of BC expression. An interesting finding in this study was that BC correlated with DNA damage but not chromosomal lesion, indicating that BC might not be appropriate in aneugen genotoxicity prediction. Given both the in vivo ( Fig. 1 and Table 2) and in vitro data (Fig. 3), BC has the intrinsic ability to become a sensitive biomarker of DNA damage. Although BC induction was not obvious in response to colchicine and paclitaxel treatment, this would not compromise the sensitivity of BC as a potential biomarker of DNA damage, since colchicine and paclitaxel mainly target microtubules rather than DNA. Additionally, it should be noted that compounds that require S9 activation were excluded from the in vitro study, in order to avoid potential S9 effects on BC expression. It would be of great interest to examine the effects of indirect genotoxic chemicals on BC expression in future studies. Moreover, we are currently developing a GLN LTR driven luciferase reporter gene system in NIH/3T3 cells, to facilitate toxicity screening in early drug development [36,37,38,39].
The GLN family was recently reported to be potentially active in the mouse genome [21], thus uncovering its cellular functions would be of great significance. To the best of our knowledge, this is the first report showing the involvement of a GLN family member in cell growth regulation, thus providing basic information towards a deeper understanding of its biological roles. Several studies have reported the mitogenic effects of exogenous retroviruses (XRVs) on certain cell lines [40,41]. One suggested mechanism was the interaction of glycoproteins encoded by retroviral env genes with cellular cytokine receptors, such as interleukin-2 receptor or erythropoietin receptor. One study also suggested a role for env of mink cell focus-forming (MCF) ERV in regulating pluripotent hemopoietic progenitor proliferation. Given previous reports that GLN was potentially active and might have an extracellular life cycle similar to XRVs, it would be of great interest to study whether GLN regulates cell growth through similar mechanisms. Such studies will not only assist in a deeper understanding of its biological function, but also provide theoretical support for its application in genotoxicity prediction.
In summary, this is the first study to investigate the relationship between ERVs and genotoxicity. We showed that expression level of BC005512, a member from GLN family of murine ERV, was elevated by multiple GTXs both in vivo and in vitro, and correlated with the extent of DNA damage. It is thus possible that BC might be useful as a potential molecular biomarker for assessment of DNA damaging effects. Additionally, we first demonstrated that BC was involved in cell growth regulation. This contributes to a deeper understanding of the biological functions of BC005512 and GLN, and provides fundamental support for their application in genotoxicity predictions.

Animal treatment in microarray study
Two-month-old male B6C3F1 mice were given a single intraperitoneal injection of model compounds or solvent control (corn oil or saline) as listed in Table 1. O-AAT, DBP, DMBA and

Microarray and data analysis
Five mg pooled total RNA of liver from 4 animals per group was used as starting material. For each time point of each compound, only one sample pooled from 4 animals was tested. cDNA synthesis, cRNA labeling, and cRNA fragmentation were conducted according to the manufacturer's instructions (Affymetrix Inc., Santa Clara, CA, USA). The hybridization mixture was hybridized to an Affymetrix Mouse Genome 430 2.0 array. Hybridized arrays were washed and stained, and fluorescence signals were detected using the Affymetrix GeneChip Scanner 3000. The image files were converted into expression data by the Microarray Suite Software (Affymetrix) and the data was imported into GeneSpring software (Silicon Genetics, Redwood City, CA, USA). Signal intensity was normalized by per-gene and per-chip. Ratio was calculated by normalizing treatment samples to solvent control samples. The procedure was in compliance to MIAME guidelines. The raw data has been deposited in GEO database (GSE33248).
We used a step-wise selection criterion (set different significance levels according to the intensity) in selecting differentially expressed genes. For each time point of each chemical, genes that met with either criterion listed in Table S1 were considered as up-regulated genes and the corresponding cells in the Table S3 were marked with pink. While genes that met with either criterion listed in Table S2 were considered as down-regulated genes and the corresponding cells in the Table S3 were marked with blue.
Those selected up-regulated genes were further analyzed by selfdefined weight scoring based on several important parameters in identifying genotoxic stress responsive genes (including specificity, sensitivity, statistical P value, basal expression level, and coefficient of variation) Detailed scoring rules are described in supporting information (Text S1). Top 50 genes were further analyzed by hierarchical clustering (average linkage clustering) by using Gene Cluster and TreeView programs [42].

Chemicals treatment in in vitro BC expression assay
Compounds that require metabolic activation were excluded to avoid potential effects of S9 fraction on BC expression. The seven selected genotoxins represent different mechanisms of action. NIH/3T3 or L5178Y cells were treated with indicated concentration of model compounds or irradiated with c-ray by Gammacell 3000 Elan (MDS Nordion, Ottawa, ON, Canada) as listed in Table 3. After 4, 8, 12 and 24 h treatment or irradiation for NIH/3T3, 4 h for L5178Y, cells were harvested and subjected to RNA isolation and qPCR. MMC was purchase from Kyowa (Tokyo, Japan), ACN from Merck (Darmstadt, Germany), SUC from Sinopharm, and other chemicals were from Sigma.

Alkaline comet assay
The alkaline comet assay was performed as previous described with slight modifications [28]. Briefly, NIH/3T3 cells were exposed to MMS for 8 h or to paclitaxel (Sigma) or colchicine (Sigma) for 24 h. Then cells were trypsinized, resuspended in PBS, mixed with 0.5% low-melting agarose and applied to glass slides pre-coated with 1% normal-melting agarose. Slides mounted with cells were immersed in cold lysing solution (2.5 M NaCl, 100 mM EDTA, 10 mM Tris, 1% Triton X-100, 10% DMSO, pH = 10; the last two compounds were added fresh) for 2.5 h. After lysis, slides were immersed in alkaline buffer (300 mM NaOH, 1 mM EDTA, pH = 13) for 20 min to allow DNA unwinding and DNA breakage at alkali-labile sites. Then electrophoresis was performed at 300 mA for 20 min. After electrophoresis, slides were neutralized in 0.4 M Tris-HCl, dipped in ethanol and air-dried. Cells were stained with SYBR Green and observed by a fluorescent microscope (Olympus BX 51). At lease 100 cells were pictured for each group, and olive tail moment (tail length6percentage of DNA in tail) was determined by image analysis software Komet 5.5 (Kinetic Imaging, Liverpool, UK).

Micronucleus test
The micronucleus test was performed as described previously with some modifications [44] [31]. Briefly, NIH/3T3 cells were exposed to paclitaxel or colchicine for 24 h. Then the attached cells were trypsinized, incubated in 0.075 M KCl hypotonic solution, fixed in methanol-acetic acid (3:1) and finally suspended in methanol containing 1% acetic acid. A drop of cell suspension was placed on a clean glass slide and air-dried. Cells were stained by mounting with acridine orange and immediately observed by a fluorescent microscope (Olympus BX 51). The micronucleus frequency was determined in 1000 total cells according to published criteria [44], and cells with irregular shape of nucleus were also counted as micronucleated cells. Cytochalasin B was not used since it has been reported that the alteration of two cytoskeletal elements, microtubules and microfilaments, concomitantly could influence the formation of micronucleated cells [45].

EdU incorporation assay
EdU incorporation assay was performed by using Cell-Light EdU DNA imaging kit (RiboBio, Guangzhou, China) according to the manufacture's instructions. Briefly, at 48 h after siRNA transfection, NIH/3T3 cells were cultured in medium containing 50 mM EdU for 3.5 h. After EdU incorporation, cells were fixed in 4% paraformaldehyde and permeabilized by 0.5% Triton X-100. After a click reaction between Apollo dye and ethynyl group in EdU, EdU positive cells were visualized with a green light excitation (550 nm), while total cells were visualized by Hoechst 33342 staining with UV excitation. At least a total of 4000 cells were pictured for each group. EdU positive and total cells were counted by using software Image-Pro Plus (Media Cybernetics, Bethesda, MD, USA).

Cell cycle analysis
At 48 h after siRNA transfection, NIH/3T3 cells were trypsinized, fixed in 70% ethanol, incubated with RNaseA, stained with propidium iodide (Sigma) and analyzed by a FACSCalibur (BD, Franklin Lakes, NJ, USA) instrument. Analysis of cell cycle phase distribution was performed by using the Modfit software (Verity Software House, Topsham, ME, USA).

Statistical analysis
Data were analyzed by Student's unpaired t-test in Excel. Differences were considered significant at *P,0.05, **P,0.01. For BC expression study, the raw data was divided by that of control, generating the normalized BC expression (folds of control). For linear regression analysis, each dot in Fig. 4C represents the mean of data shown in Fig. 4A, 4D and 4G. The linear regression analysis was performed by using GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA).

Supporting Information
Text S1 Supporting materials and methods.  Table S1 Step-wise criterion in selecting up-regulated genes.