Robust Transcriptional Regulatory Response Upon Blocking NHEJ

: 25 Double strand breaks are one of the most lethal forms of DNA lesions that, if left unrepaired can 26 lead to genomic instability, cellular transformation, and cell death. However, cells have two main 27 machineries namely error prone Non homologous end joining repair (NHEJ) or an accurate 28 homology dependent repair to repair the double strand breaks. NHEJ is the preferred mechanism 29 for DNA repair and basically consists of two forms: Canonical (C-NHEJ) and Alternative (A- 30 NHEJ) NHEJ. Our study examined the cellular repair outcome when NHEJ is blocked by targeting 31 two key DNA repair proteins: XRCC4 and MRE-11. We developed an extrachromosomal NHEJ 32 fluorescent reporter assay that uses Transcription activator-like effector nucleases (TALEN) to 33 introduce double strand breaks and detect the NHEJ editing by the presence of GFP fluorescence. 34 We demonstrated the presence of NHEJ editing in the XRCC4 (-/-) cells treated with Mirin (a 35 pharmacological inhibitor of MRE-11), albeit with a ~52% efficiency of the normal cells. The 36 transcriptional profiles of the Mirin treated HeLa XRCC4 (-/-) cells had 307 uniquely differentially 37 expressed genes that was far greater than HeLa XRCC4 (-/-) sample (83 genes) and Mirin treated 38 HeLa cells (30 genes). Pathway analysis unique to the XRCC4 (-/-) +Mirin group included 39 differential expression of p53 downstream pathways, and metabolic pathways indicating cell 40 adaptation for energy regulation and stress response. In conclusion, our study showed that the 41 double strand DNA repair can be sustained even in absence of key DNA repair proteins XRCC4 42 and MRE-11. 43 44 45

3 Introduction: 49 DNA carries genetic instructions for the development and function of all known living organisms; 50 therefore, it is important to preserve the integrity of the DNA. However, DNA is not inert, but 51 susceptible to multiple types of damages. The common sources of DNA damage include 52 environmental agents such as UV light, ionizing radiation, and chemical mutagens. Additionally, 53 endogenous biological processes such as cellular metabolism including oxidative damage, DNA 54 alkylation or hydrolysis, and double-strand breaks (DSBs) from collapsed replication forks 55 contributes to DNA damage 1 . In fact every day, DNA in normal cells has approximately 10,000 56 DNA aberrations and thus require an efficient repair of DNA damage to maintain its integrity 2,3 . 57 Failure to repair such damages can lead to genomic instability, cellular transformation, and cell 58 death. England Biolabs) and cloned into the BsmBI digested LenticrisprV2 plasmid (Addgene # 52961)). 140 The plasmid was delivered into HeLa cells by lentiviral transduction. The cells were selected in 141 complete media with puromycin (1.5 µg/ml) for 1 week followed by clonal selection. Clonal cells 142 were screened for biallelic XRCC4 knockout by Western Blotting with a XRCC4 antibody (Santa 143 Cruz Biotechnology sc-271087) on Nitrocellulose membrane (GE Amersham) and by targeted 7 sequencing of gDNA. One of the clones 2G3 (HeLa XRCC4(-/-)

Quantitative Real-time Polymerase Chain Reaction (qRT-PCR) 217
For validation of RNA-seq results, cells were transfected with a NHEJ reported plasmid, TALEN 218 expression constructs, and treated with Mirin or vehicle control as described above. Cells were 219 harvested 48 hrs after transfection, Total RNA was extracted using the Zymogen RNA prep kit 220 (Zymogen) and cDNA was synthesized with SuperScript™ IV VILO™ Master Mix (Invitrogen). 221 Gene expression of select DEGs were quantified by qRT-PCR with gene-specific primers (Table  222   S1

Construction and validation of a NHEJ reporter assay 233
In order to test and detect NHEJ repair when TALEN introduces DSBs, we first constructed and 234 tested a plasmid encoding an extrachromosomal NHEJ reporter assay system. The reporter plasmid 235 has a CMV promoter for constitutive expression of the mCherry coding sequence. The mCherry 236 coding sequence is flanked on either side with TBSs (Fig. 1A). Under normal conditions, the 237 mCherry reading frame terminates with a stop codon. Consequently, the downstream GFP coding 238 sequence lacks a promoter; hence not expressed. When the cells are co-transfected with the 239 11 reporter plasmid and the pair of TALEN expressing plasmids, that targets the encoded TBSs, 240 double strand breaks are introduced at both TBSs excising the mCherry coding region. Upon 241 subsequent NHEJ repair, the CMV promotor is ligated in proximity to the GFP coding region 242 driving its expression. Therefore, GFP is expressed upon NHEJ repair. A flow chart depicting the 243 relationships between DNA editing and fluorescence output is shown (Fig. 1A). 244 245 First, we tested the NHEJ reporter assay system in HeLa cells. The reporter plasmid was co-246 transfected with TALEN expression constructs T256 and T278. GFP positive (GFP + ) cells were 247 detected indicating editing by NHEJ (Fig. 1B). Co-transfection of the NHEJ reporter plasmid with 248 empty vectors, or with either one of the TALEN pair did not express any GFP + cells as expected 249 for cells that are not edited (Fig. 1B). This experiment confirms that our NHEJ assay system is 250 functional, specific, sensitive, and detects DNA repair. 251 252

Inhibition of XRCC4 and MRE-11 in Hela cells 253
To determine the effect of XRCC4 and MRE-11 on C-NHEJ and A-NHEJ repair pathways, we 254 needed to abolish expression of these proteins. XRCC4 was knocked out in Hela cells using 255 targeted CRISPR-Cas9 editing within the coding region. The XRCC4(-/-) biallelic knockout in 256 clone 2G3 was confirmed by the loss of XRCC4 protein expression as detected by Western blot 257 analysis (Fig. 1C). A complete absence of XRCC4 protein expression was observed, and this clone 258 was selected for further experiments. The knockout in clone 2G3 was confirmed by Sanger 259 sequencing, which revealed a 2 bp deletion and 10 bp deletion at both the alleles in the XRCC4 260 gene (Fig. 1D). MRE-11 expression was indispensable for cell survival thus, MRE-11 was 261 inhibited by Mirin, a well-established inhibitor of MRN complex and MRE-11 exonuclease 262 activity. We also analyzed the cell survival when XRCC4(-/-) cells were treated with Mirin ( Fig. 12   1E). Blocking both XRCC4 and MRE-11 indicate that that the cells are viable as detected in an 264 MTT assay (Fig. 1E). 265 266 NHEJ repair is sustained when XRCC4 and MRE-11 are blocked 267 Since cells with both XRCC4(-/-) and MRE-11 blocked survived and were metabolically active, 268 we next assessed the impact of blocking NHEJ on DNA repair activity. XRCC4(-/-) and control 269 cells HeLa-NT (Non-targeting) transfected with reporter plasmid and empty TALEN vectors did 270 not show any GFP + cells as expected for these negative controls ( Fig. 2A & E). Mirin treatment of 271 these cells also had no effect on GFP expression (Fig. 2B & F). mCherry expression indicated that 272 the cells were expressing the reporter construct (Fig. 2). HeLa-NT and XRCC4(-/-) cells co-273 transfected with the NHEJ reporter and TALEN constructs (T270 and T278) showed GFP 274 expression, indicating NHEJ repair (Fig. 2C & G). 275

276
In co-transfected HeLa-NT cells treated with Mirin, GFP + cells were present indicating NHEJ 277 repair when MRE11 is inhibited (Fig. 2D). Surprisingly, co-transfected Mirin treated XRCC4(-/-) 278 also showed NHEJ repair as indicated by the presence of GFP expression (Fig 2 H). This suggests 279 that NHEJ repair is sustained even when XRCC4 and MRE-11 are inhibited. 280

NHEJ efficiency was lowered when both XRCC4 and MRE-11 are blocked 282
To quantify the NHEJ repair efficiency, cells expressing the fluorescent reporters were quantified 283 by flow cytometry. Cells were first gated based on mCherry + expression, selecting the population 284 expressing the reporter system. Next, we determined percentage of cells expressing GFP within 285 the mCherry + cell populations (Fig. 3A, B). The mean percentage of GFP + expressing cells within 286 the mCherry + cells for each sample is compared in Fig. 3B. The efficiency of NHEJ repair was 287 13 (25.5% ± 1.3) in HeLa-NT cells and (15.4% ± 0.4) for HeLa-NT cells treated with Mirin (Fig. 3B), 288 indicating that blocking MRE11 activity reduced NHEJ editing efficiency. Similar inhibition of 289 NHEJ editing was observed in XRCC4(-/-) cells (15.9% ± 0.2) or those treated with Mirin (13.26% 290 ± 0.2). These results indicated that NHEJ repair was prevalent even when either or both XRCC4 291 and MRE-11 are blocked, however; the efficiency is reduced by approximately 40-48%. 292 293

Altered expression of genes when NHEJ is blocked 294
The reporter assay demonstrates that NHEJ repair is partially inhibited, but still present in cells 295 when XRCC4 and MRE-11 expression is blocked alone, as well as in combination. To assess the 296 molecule basis for the robustness in preservation of NHEJ activity, gene expression of 297 transcriptomes was compared for cells with inhibition of XRCC4 and MRE-11. To determine 298 differences in transcriptomes, cell populations selected as above were analyzed by RNA-seq. A 299 principle component analysis showed a clear separation between each sample category, but 300 clustering for duplicate samples (Fig. 4A) differentially expressed genes (Fig. 4B). However, the Mirin treated XRCC4(-/-) cells had 307 306 uniquely differentially expressed genes, far greater than other samples reflecting a more impactful 307 transcriptional response. This differential transcriptional response was further supported when the 308 top differentially expressed genes (DEGs) were plotted as heatmap ( Fig.4C and Supplementary 309 Table 2). The gene expression for XRCC4(-/-) cells treated with Mirin were most different from 310 14 expressed genes identified from the RNA-seq analysis (CA9, CDKN1A, ENO2, DUSP5 and 312 ZMAT3) were assessed by real time PCR. The top differentially expressed genes plotted for 313 heatmap (Fig. 4) and the p53 downstream pathway were selected for gene expression quantitation 314 by real time PCR (Fig. 5A). The PCR data and RNA-seq gene expression measurements were 315 consistent with each other, thereby validating the RNA-seq results (Supplementary Fig. 1) nodes are displayed as pies (Fig. 6A). Seven significant network modules were identified with the 340 MCODE algorithm (Fig. 6B), which included 71 proteins from which ENO3, CACNG6, ITGB8, 341 PDE10A, COL12A1 and FSTL3 served as seed proteins. Gene Ontology (GO) terms associated 342 with each module are depicted in Fig. 6B  We considered three possible explanations for the presence of NHEJ repair despite inhibiting 359 XRCC4 and MRE-11: 1) There could be a yet to be discovered NHEJ pathway. However, this 360 possibility is unlikely considering that the comparative transcriptomic profiles did not reveal any 361 differential expression of DNA repair proteins or components of DNA repair complexes. We 362 recognize that this negative result is not conclusive, as is the case for any negative result. However, 363 analysis of whole transcriptome is a global measurement, but there could be genes with repair 364 functions that are not yet identified or annotated; 2) Another possible explanation is that either has a critical role in DSB recognition and protein complex recruitment in both the NHEJ and HR 386 pathways. Another protein, BRCA-1, is known to regulate the nuclease activity of MRE-11. 387 BRCA1 phosphorylation mediated by Checkpoint kinase-2-enhances NHEJ fidelity but can also 388 deplete 53BP in S/G2 phase to favor HR, suggesting that post-translation modification of BRCA1 389 regulates pathway preference 25,26 . Thus, it is possible that both C-NHEJ and A-NHEJ could also 390 have signaling interactions that rescue blocking both NHEJ pathways 9,28 . 391

392
One theme that emerged from blocking NHEJ was that blocking of NHEJ pathways by either 393 XRCC4 knock out or Mirin treatment alone or in combination lead to pathways like transcriptional 394 misregulation in cancer, alcoholism, defense response to viruses, and response to oxygen levels.   Measurement of NHEJ e ciency using ow cytometry. A) Representative images of cells positive for mCherry expression followed by GFP+ expression. B) Mean GFP+ cells were plotted. Experiments were done in triplicates and data was analysed using FLowJo 10.7.1. Statistical signi cance was determined by ANOVA where indicates p <0.01 Figure 4 Principal component analysis (PCA), Venn Diagram and Heatmap from RNA-seq pro le. A) A PCA plot for HeLa-NT, XRCC4(-/-), Mirin and XRCC4(-/-)+ Mirin from whole transcriptome RNA-seq data using ClusVis 20. B) A Venn diagram representing shared and unique DEGs across 3 categories (i) XRCC4(-/-) (blue) (iii) Mirin treated (green) (iii) XRCC4(-/-) with Mirin treated (red) when compared to HeLa-NT cells. C) A heatmap of top DEGs across each sample categories.

Figure 5
Functional enrichment analysis A) A heatmap with embedded dendogram showing relationships between enriched GO/KEGG terms and canonical pathways. 0.3 kappa score was applied as the threshold to cast the tree into term clusters. B) A network graph with enrichment ontology clusters colored by cluster ID. Each term is represented by a node, where its size is proportional to the number of genes for each term.
Terms with a similarity score > 0.3 are linked by an edge (the thickness of the edge represents the similarity score). The network was created with Cytoscape (v3.1.2).

Figure 6
Protein-protein interaction network analysis. A) PPI network of DEGs across all three NHEJ categories. Nodes are displayed as pies to indicate NHEJ sample. B) MCODE components were identi ed from merged network for all samples. Each MCODE network is assigned a unique color and the network was generated with Cytoscape (v3.