Role of the Exocyst Complex Component Sec6/8 in Genomic Stability

Abstract

The exocyst is a heterooctomeric complex well appreciated for its role in the dynamic assembly of specialized membrane domains. Accumulating evidence indicates that this macromolecular machine also serves as a physical platform that coordinates regulatory cascades supporting biological systems such as host defense signaling, cell fate, and energy homeostasis. The isolation of multiple components of the DNA damage response (DDR) as exocyst-interacting proteins, together with the identification of Sec8 as a suppressor of the p53 response, suggested functional interactions between the exocyst and the DDR. We found that exocyst perturbation resulted in resistance to ionizing radiation (IR) and accelerated resolution of DNA damage. This occurred at the expense of genomic integrity, as enhanced recombination frequencies correlated with the accumulation of aberrant chromatid exchanges. Sec8 perturbation resulted in the accumulation of ATF2 and RNF20 and the promiscuous accumulation of DDR-associated chromatin marks and Rad51 repairosomes. Thus, the exocyst supports DNA repair fidelity by limiting the formation of repair chromatin in the absence of DNA damage.

INTRODUCTION

The faithful repair of DNA damage is integral to the maintenance of the genome and suppression of oncogenesis (1). This relationship has motivated intense efforts to elaborate the composition and mechanism of action of core DNA repair machinery as well as peripheral molecular systems that modulate this machinery to suppress genomic instability (2–5). With respect to the latter, emerging evidence implicates multiple regulatory layers that link activation of the DNA damage response (DDR), repair pathway choice, and resolution of the DDR to chromatin organization (6–9), RNA metabolism (10, 11), and autophagy (12–14). By extrapolation, the coordinated response of cellular processes to DNA damage is necessary for efficient DNA repair, and perturbations of these pathways can lead to genomic instability and development of neoplastic disease.

The exocyst (also known as the Sec6/8 complex) is a conserved heterooctomeric protein complex, which includes Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo84, and Exo70. The holocomplex is well appreciated for its role in the dynamic trafficking of secretory vesicles to specialized membrane domains such as the basolateral membrane of polarized epithelial cells (15) and abscission planes in dividing cells (16) and to lamellipodia and growth cones of migrating cells and differentiating neurons (17, 18). Accumulating evidence indicates that exocyst subcomplexes, and their regulation by Ras and Rho family GTPases, also selectively participate in the assembly and activation of signal transduction events that mediate host defense, autophagy, cell growth, and oncogene signaling (19–23). An overarching implication is that the exocyst and its subcomplexes serve as physical platforms that coordinate organellar assembly with the activation of attendant regulatory cascades required for the execution of distinct cell biological programs.

Here, we describe the identification of the exocyst as a modulator of DNA repair. Through a combination of genome-wide pairwise protein interaction analysis and mass spectrometry of immunoisolated endogenous Sec8, we identify 33 exocyst-associated proteins involved in the cellular response to DNA damage. Consistent with a functional role in DNA repair, we find that Sec8 depletion results in genomic instability while conferring radioresistance. This is a consequence, in part, of the upregulation of histone-modifying proteins, ATF2 and RNF20, and the concomitant acceleration of DDR resolution. Our cumulative observations suggest that the exocyst contributes to genomic stability through spatial and temporal restraint of chromatin modifications that specify DNA repair pathway choice.

MATERIALS AND METHODS

Cell culture.

U2OS cells (from the ATCC) were cultured in McCoy's 5A medium supplemented with 10% fetal bovine serum (FBS). HBEC3 KT cells were cultured in keratinocyte serum-free medium (KSFM) (Invitrogen). MCF7A DR-GFP cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and 10 ng/ml puromycin. U2OS GFP-LC3 cells were maintained in DMEM supplemented with 10% FBS, 1 mg/ml G418, and 5 μg/ml blasticidin.

p53 siRNA screen.

Small interfering RNA (siRNA) pools (four siRNAs) targeting a single colorectal cancer (CRC) candidate gene (24) were obtained from the Qiagen human whole-genome siRNA library (version 1.0). HCT116 or RKO cells were transiently transfected with the pp53-TA-Luc reporter (Clontech), the SV40-RL reporter, and siRNAs by using Effectene (Qiagen). A final concentration of 33 nM siRNA was used to transfect 10,000 cells plated in 96-well plates. Experiments were performed in triplicate. Firefly and Renilla luciferase activities were measured after 36 h by using the Dual Luciferase reporter assay system (Promega). Normalized p53 activity was calculated as the ratio of firefly to Renilla luciferase.

Yeast two-hybrid screen and mass spectrometry analysis.

The coding sequence for full-length human Sec3, Sec5, Sec6, Sec8, Sec10, Exo84, and Exo70 was cloned into pB27 as a C-terminal fusion to LexA and used as a bait to screen at saturation a high-complexity random-primed human placenta cDNA library, as previously described (25). Using the raw data, an interaction map was generated from all potential interactions with a confidence score of “D” or higher in the screen (the confidence score is detailed in reference 26). Proteins were manually assigned a function in the DNA damage response by a curated literature search.

Immunoprecipitation and immunoblotting.

U2OS cells and HBEC 3KT cells were seeded onto 10-cm dishes and allowed to grow to 100% confluence for immunoprecipitation (IP) or seeded onto 35-mm dishes for transfection and immunoblot analysis. For IP, cell cultures were washed with cold phosphate-buffered saline (PBS) and lysed on ice for 10 min in a solution containing 20 mM Tris-HCl (pH 7.4), 137 mM NaCl, 10% glycerol, 1% Triton X-100, 0.5% sodium deoxycholate, 10 mM MgCl₂, and 2 mM EGTA plus protease and phosphatase inhibitors (Roche EDTA-free protease inhibitor cocktail [1 mM phenylmethylsulfonyl fluoride {PMSF}, 50 mM NaF, 1 mM NaVO₄, 80 mM β-glycerophosphate]) and then Dounce homogenized 15 times. Lysates were then centrifuged at 20,000 × g for 10 min at 4°C. Soluble fractions were collected, and the protein concentration was assessed by a Precision Red Advanced protein assay (Cytoskeleton, Inc.). Protein A/G beads and mouse hemagglutinin (HA) antibody (Ab) (Santa Cruz) were used to clear nonspecific interactions for 1 h at 4°C. Cleared lysates were incubated with the indicated antibodies (1 μg Ab per 200 mg lysate) coupled to protein A/G beads for 1 h at 4°C. Beads were then washed with lysis buffer three times, followed by elution of the protein in 2× SDS sample buffer.

For overexpression/coimmunoprecipitation, HEK 293T cells were seeded at 2 × 10⁵ cells onto 35-mm dishes 24 h before transfection. Two micrograms of total DNA was transfected with Fugene 6 at a ratio of 1:3 in Opti-MEM. At 72 h posttransfection, cells were lysed and processed as described above.

Immunofluorescence.

For colocalization studies, cells were cultured on coverslips, fixed and permeabilized with methanol at −20°C for 1 min, blocked with 5% bovine serum albumin (BSA) in 0.1% Tris-buffered saline–Triton X-100 (TBS-T) for 20 min at 37°C, washed in 0.1% TBS-T, and sequentially incubated with Sec8 (10C2) primary antibody and donkey anti-mouse rhodamine-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc.) and with 53bp1 primary antibody (Bethyl) and Alexa Fluor 488–donkey anti-rabbit secondary antibody (Invitrogen) for 1 h each at 37°C. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Vectashield mounting medium; Vector Laboratories). Slides were imaged by using a Leica TCS SP5 confocal microscope (Leica Micro-Systems, CMS GmbH) with custom software (Leica Micro-Systems LAS AF), using a sequential 3-channel scan. All images were captured by using the same electronic settings. Images were imported to ImageJ (http://rsb.info.nih.gov/) by using the LOCI Bio-formats plug-in (University of Wisconsin, Madison).

For analysis of DDR foci, siRNA-transfected cells were cultured in chamber slides, fixed in 2% paraformaldehyde for 15 min, washed in PBS, permeabilized for 5 min on ice in 0.2% Triton X-100, and blocked in PBS with 1% bovine serum albumin. Immunostaining protocols were previously described (27, 28). Epifluorescence images were captured by using a Zeiss Axioskop 2 Mot epifluorescence microscope equipped with a charge-coupled-device camera and ISIS software (Metasystems, Altlussheim, Germany). All results shown are from three independent experiments.

For ATF2 nuclear analysis, cells were seeded onto 2-well glass chamber slides and subsequently transfected as described above with control or Sec8 siRNA. At 72 h posttransfection, cells were fixed and stained as described above, and epifluorescence images were captured by using a Zeiss Axioskop 2 Mot epifluorescence microscope equipped with a charge-coupled-device camera and ISIS software (Metasystems, Altlussheim, Germany). Images were scored in a blind manner by calculating the mean nuclear intensity, using DAPI as a mask, in ImageJ. Data represent results from at least 2 independent experiments (n > 75).

GFP-LC3 quantitation.

The total green fluorescent protein (GFP) fluorescence per cell was measured by flow cytometry for at least 10,000 events per condition on a BD FACSCalibur instrument using CellQuest and FlowJo software. For visualization of GFP-LC3-positive punctae, cells were cultured on glass coverslips, rinsed in PBS, fixed with 3.7% paraformaldehyde, permeabilized with cold acetone, and mounted with Vectashield medium containing DAPI. Images were acquired by using a Zeiss Axioplan 2E microscope and a Hamamatsu monochrome digital camera with OpenLab software.

Cell cycle analysis.

U2OS cells were harvested after irradiation (Cs¹³⁷ Mark 1 irradiator; JL Shepherd & Associates) and fixed in 80% ethanol at 4°C. Cells were centrifuged, washed in a solution containing 1% BSA and 0.1% Tween 20 in PBS, and either labeled with mouse anti-phospho-MPM2 for 3 h at room temperature (RT), pelleted, and labeled with goat anti-mouse antibody–Alexa Fluor 488 for 30 min or resuspended in propidium iodide (PI)-RNase staining buffer (BD Pharmingen) for 30 min at 37°C. A total of 20,000 events were analyzed on a BD FACSCalibur instrument by using CellQuest and FlowJo software.

DNA fragmentation analysis.

Neutral comet assays were performed with the CometAssay kit (Trevigen) according to the manufacturer's instructions. At 72 h posttransfection, cells were irradiated with the indicated doses and incubated for 30 min. Cultures were then trypsinized, diluted to 1 × 10⁵ cells/ml in PBS, mixed with molten low-melting-point agarose (LMAgarose) at a 1:10 ratio, and spotted onto glass slides. After solidification, slides were immersed in lysis solution at 4°C for 30 min and equilibrated in chilled neutral electrophoresis buffer for 30 min. Electrophoresis was performed in neutral electrophoresis buffer for 20 min with an electric field of 1 V/cm. Slides were further treated with DNA precipitation solution followed by 70% ethanol for 30 min each at room temperature. After air-drying, cells were stained with SYBR green (1 mg/ml). Comet images were captured with an epifluorescence microscope (Zeiss Axioplan 2E). Comet tails were measured according to pixel length with ImageJ.

DR-GFP cell assay.

MCF7 DR-GFP stable cells were derived as previously described (29), transfected with the indicated siRNAs, and assayed at 72 h posttransfection. To measure the repair of I-SceI-generated double-strand breaks (DSBs), the I-SceI plasmid was mixed with Lipofectamine 2000 at a ratio of 1:2. The percentage of cells that were GFP positive was quantitated by flow cytometric analysis on a Becton Dickinson FACScan instrument at 48 h post-I-SceI transfection. Detection of BRCA1, Rad51, and KU80 proteins at I-SceI-induced DSB sites by chromatin immunoprecipitation (ChIP) was done by using DR95 cells with and without depletion of Sec8, as described previously (30, 31).

DNA fiber assay.

Fiber assays were performed as previously described (32). Briefly, exponentially growing cells were pulsed with 50 mM 5-iododeoxyuridine (IdU) for 20 min, washed three times with PBS, treated with 2 mM hydroxyurea (HU) for the indicated intervals, washed three times with PBS, incubated in fresh medium containing 50 mM 5-chlorodeoxyuridine (CldU) for 20 min, and then washed three times in PBS. DNA fiber spreads were produced with a modified procedure described previously (33). IdU- and CldU-labeled cells were mixed with unlabeled cells at a ratio of 1:10. Two-microliter cell suspensions were then dropped onto a glass slide and mixed with 20 μl hypotonic lysis solution (10 mM Tris-HCl [pH 7.4], 2.5 mM MgCl₂, 1 mM PMSF, and 0.5% Nonidet P-40) for 8 min. Air-dried slides were fixed, washed with 1× PBS, blocked with 5% BSA for 15 min, and incubated with primary antibodies against IdU and CldU (rat anti-IdU monoclonal antibody [MAb] [1:150 dilution; Abcam] and mouse anti-CldU MAb [1:150 dilution; BD]) and secondary antibodies (anti-rat antibody–Alexa Fluor 488 [1:150 dilution] and anti-mouse antibody–Alexa Fluor 568 [1:200 dilution]) for 1 h each. Slides were washed with 1× PBS containing 0.1% Triton X-100 and mounted with Vectashield mounting medium without DAPI.

Metaphase spread analysis.

Control or Sec8 siRNA-transfected U2OS cells were irradiated with 1 Gy of ionizing radiation (IR), incubated for 2 h, and then exposed to colcemid (Sigma-Aldrich) (0.1 mg per ml). After 3 h of colcemid treatment, cells were harvested. Metaphase spreads were prepared and analyzed as described previously (34, 35). Fifty metaphases were scored per sample, and each sample was analyzed three times. Means and standard deviations were derived from three independent experiments.

Clonogenic survival assay.

Control or Sec8 siRNA-transfected U2OS cells were replated into 60-mm dishes in 5.0 ml of medium, incubated for 7 h, and subsequently exposed to graded doses (0, 2, 4, 6, and 10 Gy) of ionizing radiation (Cs¹³⁷ at a dose rate of 1 Gy per min). Cells were incubated for 12 days and fixed in methanol-acetic acid (3:1) prior to staining with crystal violet. Only colonies containing >50 cells were counted. Survival fractions are reported as means of data from three independent experiments.

Reverse transcription-quantitative PCR (qRT-PCR).

Control or Sec8 siRNA-transfected cells were collected, and RNA was extracted by using a Qiagen RNeasy kit (catalog number 74104). cDNA was made by using 100 ng of RNA for all samples, using iScript reagents from Bio-Rad. ATF2 or RNF20 TaqMan probes (Applied Biosystems) were used for real-time analysis on a Roche LightCycler instrument. Expression levels were calculated by comparing the threshold cycle (C_T) values of target genes to the value for β-actin (used as an internal control), and the fold change was calculated by comparing the C_T values of untreated samples to those of treated samples. The experiment was done in triplicate.

siRNA sequences.

siGenome pools were purchased from Dharmacon for the following genes: Sec8 (ExoC4), Sec3, Sec5, Sec6, Sec10, Sec15, Exo70, 53BP1, TP53, ULK1, BECN1, Huwe1, and Ubr5. The control siRNA was nontargeting pool 2 siRNA (catalog number D-001206-14-05). Exo84 was custom designed with the following sequence: 5′-GCCACUAAACAUCGCAACUdTdT-3′.

Plasmids and reagents.

Myc-Sec8 (15), Flag-53bp1 (36), and pCMV-I-SceI (37) were described previously. Aphidicolin was diluted in dimethyl sulfoxide (DMSO) at a stock concentration of 2 mg/ml. Hydroxyurea was diluted in DMSO. Camptothecin (CPT) was diluted in PBS.

Antibodies.

Mouse anti-HA (catalog number sc-7392) and rabbit anti-Myc (catalog number sc-789) were purchased from Santa Cruz Biotechnology, Inc. Rabbit anti-Flag (catalog number F2555) and mouse anti-Flag (catalog number F1804) were purchased from Sigma-Aldrich, Inc. Rabbit anti-53bp1 (catalog number A300-272A) was purchased from Bethyl Laboratories. Mouse anti-pS1981 ATM (catalog number 200-301-400) was purchased from Rockland, Inc. Mouse anti-γH2AX (catalog number 16-193), mouse anti-phospho-MPM2 (catalog number 05-368), and rabbit anti-Tip60 (catalog number 07-038) were purchased from Millipore. Mouse Sec8 was provided by Charles Yeaman (38–40). Rabbit anti-ATF2 (catalog number 9226), rabbit anti-H2B (catalog number 8135), rabbit anti-H2BK120ub (catalog number 5546), rabbit anti-H3K4me3 (catalog number 9751), rabbit anti-H3K79me3 (catalog number 4260), and rabbit anti-ATM (catalog number 2873) were purchased from Cell Signaling Technology. Rad51 antibody was purchased from Abcam, and antibodies for BRCA1 and KU80 were obtained from Cell Signaling. Rabbit anti-phospho-ATF2 (S490/498) (catalog number PAB9605) was purchased from Abnova. Mouse anti-Rad51 (catalog number ab213) was purchased from Abcam.

Sec3 network representation.

Network visualization of synthetic genetic array (SGA) genetic correlations with Sec3 was downloaded from the DRYGIN database (http://drygin.ccbr.utoronto.ca/).

Image editing.

Images were edited and enhanced for visual purposes by using ImageJ, Photoshop, and Microsoft PowerPoint. Quantification was performed on unaltered images.

Statistical analysis.

P values were calculated by a two-tailed t test. All data were analyzed by using GraphPad Prism 5 and 6.

RESULTS

Physical and functional association of Sec8 with DNA damage response proteins.

TP53 loss-of-function mutations occur in 50% of colorectal cancers, suggesting that suppression of a p53 response is a common feature leading to the initiation and/or progression of this disease (41). In tumors that retain wild-type p53, somatic alterations have been identified within the p53 regulatory network, which are functionally equivalent to p53 loss of function, either through enhanced destabilization of p53 due to posttranslational modifications or through inactivation of p53 effectors (42, 43). To potentially identify previously unrecognized proteins that are functionally linked to the p53 tumor suppressor network, we screened for the effect of siRNA-mediated depletion of 140 candidate colorectal cancer genes (CAN genes [24]) on p53 promoter activity in two independent p53-positive colorectal cancer cell lines (HCT116 and RKO) (Fig. 1A). Among the top 10 enhancers of the pp53-TA-Luc reporter in both cell lines were siRNAs targeting the exocyst subunit Exoc4/Sec8L1 (here referred to as Sec8). Depletion of Sec8 also induced the expression of the endogenous p53 target genes Cdkn1a and Bbc3 (Fig. 1B), suggesting that Sec8 depletion either induces or derepresses a canonical p53 response. Notably, 33 proteins implicated as being participants in p53-associated DNA damage responses were identified within an exocyst protein-protein interaction (PPI) network (Fig. 1C). This exocyst PPI was constructed with a combination of yeast two-hybrid screens and mass spectrometry of immunoisolated endogenous Sec8 (20, 26) and contained elements of the FancD2-FancI complex that participate in interstrand cross-link (ICL) repair (44), Prpf19, which has been independently validated to interact with members of the exocyst and is implicated in DNA repair (10, 45), members of the CAK complex involved in transcription-coupled nucleotide excision repair (TC-NER) (46), members of the nonhomologous-end-joining (NHEJ) pathway, members of the Bre complex (RNF20/RNF40) involved in DNA repair-associated chromatin remodeling (47), and 53BP1, a key adaptor protein involved in regulating DNA DSB repair pathway choice (2, 48, 49). The Sec8/53BP1 interaction was recapitulated by endogenous coimmunoprecipitation (co-IP) from HBEC 3KT cells using anti-Sec8 antibodies (Fig. 1D) and by coexpression/co-IP of epitope-tagged proteins expressed in HEK 293 cells (Fig. 1E). The endogenous complex was sensitive to IR, suggesting that the Sec8/53BP1 interaction may be disrupted upon detection of DNA damage (Fig. 1D). While predominantly a cytosolic complex, nuclear roles for components of the exocyst have been described (45). Consistent with this, we found a subpopulation of endogenous Sec8 localized with 53BP1 in proliferating cell cultures that was irradiation sensitive (Fig. 1F).

FIG 1.

The exocyst interacts with DNA damage response proteins. (A) The consequences of siRNAs targeting candidate colorectal cancer genes on p53-dependent luciferase reporter activity were assessed in HCT116 and RKO cells. siRNA pools were rank ordered by effect size. Values indicate the means and standard deviations of the ratios of p53-dependent firefly luciferase activity to constitutively expressed Renilla luciferase activity (n = 3). (B) Relative Cdkn1a, Bbc3 (left), and Sec8 (right) mRNA concentrations upon exposure to the indicated siRNAs. Error bars indicate standard errors of the means (n = 3). **** indicates a P value of <0.0001 by Student's t test. (C) The exocyst/DNA damage response protein-protein interaction network derived from whole-genome yeast two-hybrid screens and IP/mass spectroscopy. Edge colors indicate literature-based interactions (yellow) or two-hybrid screen confidence scores (confidence scores were detailed previously [26], where red indicates a score of A, blue indicates B, green indicates C, and orange indicates D). Unconnected nodes indicate proteins detected exclusively by IP/mass spectrometry. (D) Endogenous Sec8 was immunoprecipitated from HBEC 3KT cells by using an anti-Sec8 monoclonal antibody. Immunoprecipitates were probed for endogenous 53bp1, as indicated. Anti-HA monoclonal antibodies were used as a specificity control. Ten micrograms of the soluble fraction was loaded as an input control (IN). (E) HEK 293T cells were transfected with the indicated constructs. At 72 h posttransfection, Flag-53bp1 was immunoprecipitated by using anti-Flag antibody-bead conjugates, and Myc-Sec8 was probed with anti-Myc antibody. WCL, whole-cell lysate. (F) Colocalization was assessed by confocal microscopy with the indicated antibodies using a TCS SP5 confocal microscope with a sequential 3-channel scan. siRNA for Sec8 was used as a control for the specificity of colocalization. Representative images from 2 independent experiments are shown. Arrowheads in the enlarged boxes (left) indicate spontaneous 53BP1- and Sec8-positive nuclear punctae. Image scaling is indicated by the bars (22 μm). MM, mismatch mutant.

53BP1 depletion induces autophagy that is beclin1 dependent but exocyst independent.

The interaction of exocyst and DDR proteins led us to consider functional cross talk between these regulatory systems. Given the recently described mechanistic interactions of both exocyst and DDR proteins with the core machinery required for autophagosome formation and maturation (12, 21, 50, 51), we first examined the contribution of 53BP1 to exocyst-dependent autophagic flux. A U2OS cell line with stable expression of GFP-LC3 was used for these studies. Lipidated LC3 coats the membranes of nascent autophagosomes and is proteolytically degraded in mature autophagolysosomes, thus serving as an indicator of autophagy dynamics (52). Depletion of 53BP1 in these cells reduced the accumulation of total GFP-LC3 (Fig. 2A) under nutrient-replete culture conditions and enhanced the accumulation of GFP-LC3 punctae in the presence of chloroquine (Fig. 2B). In addition, GFP-LC3 accumulation in 53BP1-depleted cells was rescued by codepletion of the core autophagy protein beclin1, suggesting that the observed alterations in GFP-LC3 accumulation are autophagosome dependent (Fig. 2C). Together, these observations indicate an induction of autophagic flux upon depletion of 53BP1. We previously demonstrated that the exocyst complex supports starvation-induced autophagosome biogenesis through Exo84 subunit interactions with ULK1 and beclin1 (21). As expected, cells depleted of Exo84 showed significant accumulation of GFP-LC3, as would result from the inhibition of autophagic flux (Fig. 2D). However, the enhanced autophagic flux observed upon 53BP1 depletion was independent of Exo84 expression (Fig. 2D). Likewise, the Sec8/53BP1 interaction was insensitive to alterations of nutrients, suggesting an independent contribution of 53BP1 to the regulation of autophagy (Fig. 2E). p53 itself has been identified as an inhibitor of autophagosome biogenesis independent of its role as a transcription factor (53). The consequence of 53BP1 depletion on total LC3-GFP accumulation was of a magnitude similar to that observed upon depletion of p53, and codepletion of both proteins had no further detectable consequences (Fig. 2F), consistent with the participation of 53BP1 in p53 inhibition of autophagic flux. Given the seemingly independent contributions of 53BP1 and the exocyst to the regulation of autophagy, we instead considered the possibility that exocyst proteins participate together with 53BP1 in the DNA damage response.

FIG 2.

53bp1 suppresses exocyst-independent autophagy. (A) Single-cell fluorescence intensity distributions for U2OS GFP-LC3 cells 72 h after transfection with the indicated siRNAs. The histogram is representative of data from 3 independent experiments. (B) Image-based quantification of the number of GFP-LC3 punctae per cell after treatment with 50 μM chloroquine (CQ) in U2OS cells 72 h after transfection with the indicated siRNAs. Bars indicate the means and standard errors of the means from 3 independent experiments (MM, 36.71 ± 1.945 [n = 106]; Sec8, 60.23 ± 2.761 [n = 103]). **** indicates a P value of <0.001. (C, top) Population means from flow cytometry analysis (as described above for panel A) of GFP-LC3 in U2OS cells 72 h after transfection with the indicated siRNAs. Bars indicate the log₂ mean fluorescence intensities ± standard errors of the means normalized to values for the control siRNA (MM) (n = 3). * indicates a P value of 0.0121, and ** indicates a P value of 0.0060. (Bottom) Immunoblot confirmation of target knockdown. Results shown are representative of data from 3 replicates. (D, top) Population means from flow cytometry analysis (as described above for panel A) of GFP-LC3 in U2OS cells 72 h after transfection with the indicated siRNAs. Bars indicate the log₂ mean fluorescence intensities ± standard errors of the means normalized to values for the control siRNA (MM) (n = 3). *** indicates a P value of 0.0002, and * indicates a P value of 0.0356. (Bottom) Immunoblot confirmation of target knockdown. Results shown are representative of data from 3 replicates. (E) Endogenous Sec8 was immunoprecipitated from HBEC 3KT cells following a 4-h incubation in nutrient-replete medium (RPMI medium; FM) or Earle's balanced salt solution (EBSS), as indicated. Immunoprecipitates were probed for endogenous 53bp1, as indicated. Anti-HA monoclonal antibodies were used as a specificity control. (F) Single-cell fluorescence intensity distributions for U2OS GFP-LC3 cells 72 h after transfection with the indicated siRNAs. The histogram is representative of data from 2 independent experiments.

Sec8 supports recombination-based DNA DSB repair.

Phosphorylation of H2AX on serine 139 (γH2AX) is one of the earliest detectable steps in the recruitment of repair factors to sites of DNA damage, particularly DNA double-stranded breaks, and occurs in response to the activation of most DNA repair pathways (54). Therefore, we first asked if the γH2AX response was perturbed by Sec8 depletion. Upon exposure of HBEC 3KT cells to IR, we found that γH2AX was equivalently detectable by 10 min in both Sec8-depleted cells and controls (Fig. 3A). However, Sec8-depleted cells had markedly decreased levels of residual γH2AX at later time points, suggesting abbreviated DNA damage processing compared to that of controls (Fig. 3A). This phenotype was recapitulated in U2OS cells with both 1 Gy and 5 Gy IR exposure, was reproducible by using four independent siRNA oligonucleotides targeting Sec8 in a dose-dependent manner, and was recapitulated by depletion of the exocyst subunits Sec6, Sec5, Sec10, and Sec3, collectively suggesting on-target effects of the knockdown of Sec8 and the exocyst complex (Fig. 3B to D). Sec8 depletion also significantly reduced the accumulation of 53BP1 foci (Fig. 3E), which marks cells under replication stress (55), or camptothecin (CPT) (Fig. 3F), which leads to stalled replication forks, indicating that the phenotype is not specific to the DNA DSB-inducing agent employed. At single-cell resolution, we were surprised to find that Sec8 depletion resulted in a significant increase in the number of IR-induced γH2AX foci per nucleus by 10 min postexposure relative to the control (Fig. 3G). This was followed by a significant decrease in the number of γH2AX foci per nucleus in Sec8-depleted cells relative to the control by 1 h postirradiation (Fig. 3G). Consistent with the increased γH2AX levels, we observed increased phospho-ATM levels under conditions of Sec8 depletion (Fig. 3H and I). The altered γH2AX kinetics were independent of any gross differences in cell cycle kinetics, as Sec8 depletion did not alter G₁-S-G₂/M distributions in proliferating cultures (Fig. 3J) or DNA damage-induced cell cycle arrest (Fig. 3K). Together, these observations suggest that Sec8 depletion sensitizes cells to DNA break formation or detection in a manner that also alters the persistence of the DNA damage response or resolution of the breaks themselves.

FIG 3.

Sec8 modulates the γH2AX response to DNA damage. (A) Immunoblots from whole-cell lysates with the indicated antibodies for HBEC 3KT cells after exposure to 5 Gy irradiation and incubation for the indicated time points 72 h after transfection with the indicated siRNAs. Results shown represent data from >3 independent experiments. (B) Immunoblots from whole-cell lysates with the indicated antibodies for U2OS cells after exposure to 1 and 5 Gy irradiation and 30 min of incubation 72 h after transfection with the indicated siRNAs. (C) Immunoblots from whole-cell lysates with the indicated antibodies for U2OS cells after exposure to 5 Gy irradiation and 1 h of incubation 72 h after transfection with the individual Sec8 siRNAs. (D) Immunoblots from whole-cell lysates with the indicated antibodies for HBEC 3KT cells after exposure to 10 Gy irradiation and 1 h of incubation 72 h after transfection with the individual exocyst subunit siRNAs. (E) Image-based quantification of aphidicolin (APH)-induced 53bp1 foci for HBEC 3KT cells 72 h after transfection with the indicated siRNAs. Cells with >3 53bp1 foci, as observed by confocal microscopy, were scored as positive. Bars indicate means and standard errors of the means from 2 independent experiments (MM, 86.09 ± 1.444 [n = 407]; Sec8, 27.48 ± 2.900 [n = 354]). *** indicates a P value of <0.0001. (F) Immunoblot of whole-cell lysates from HBEC 3KT cells following exposure to 100 nM CPT followed by the addition of fresh medium without drug and incubation for 2 h. Results are representative of data from 2 independent experiments. (G, top) Representative images of γH2AX foci in U2OS cells after irradiation 72 h after transfection with the indicated siRNAs. (Bottom) Image-based quantification of γH2AX foci in U2OS cells following 1 Gy irradiation and incubation for the indicated time points 72 h after transfection with the indicated siRNAs. The numbers of foci per cell are indicated by box-and-whisker plots (minimum to maximum) (MM at 10 min, 23.16 ± 0.6484 [n = 96]; Sec8 at 10 min, 28.43 ± 0.7831 [n = 77]; MM at 1 h, 23.48 ± 0.6752 [n = 81]; Sec8 at 1 h, 20.65 ± 0.5598 [n = 81]). **** indicates a P value of <0.0001, and ** indicates a P value of <0.0015. (H, top) Immunoblot of nuclear extracts with the indicated antibodies for U2OS cells 72 h after transfection with the indicated siRNAs following exposure to irradiation. (Bottom) Quantitation of the relative intensity of pATM. *** indicates a P value of <0.0001, and ** indicates a P value of <0.01. Results shown represent data from 3 independent experiments. (I) Histogram of DNA content as indicated by the propidium iodide (PI) (FL3-H) fluorescence intensity of U2OS cells 72 h after transfection with the indicated siRNA. Results shown are representative of data from 2 independent experiments. (J) Seventy-two hours after transfection with the indicated siRNAs, U2OS cells were fixed and stained with PI and anti-phospho-MPM2 1 h after exposure to the indicated doses of IR. (K) Bar graph indicating the percentage of diploid phospho-MPM2-positive cells as detected by fluorescence-activated cell sorter analysis. Error bars indicate standard errors of the means.

We next directly examined the extent and persistence of IR-induced DNA damage upon exocyst perturbation. Analysis of DNA fragmentation in neutral comet assays revealed that Sec8-depleted cells displayed a significant decrease in tail length compared to that of control cells 30 min after exposure to 10 Gy IR (Fig. 4A). Furthermore, IR-induced cell killing, as determined by clonogenic survival post-IR exposure, was reduced in Sec8-depleted cells (Fig. 4B). Together, these observations indicate that Sec8 depletion enhances DSB resolution or increases resistance to DSB formation. Notably, analysis of metaphase spreads from irradiated cells revealed a reduced frequency of persistent DNA breaks in Sec8-depleted cells (Fig. 4C, blue arrows). However, the frequency of chromatid exchanges (triradials and quadriradials) was significantly increased in these same cells (Fig. 4C, red arrows). This phenotype is reminiscent of chromosomal aberrations that occur as a consequence of altered NHEJ and the associated increased frequency of HR (homologous recombination) (56, 57). We therefore evaluated HR frequencies using a GFP reporter system to monitor the resolution of a site-specific break induced by the rare-cutting endonuclease I-SceI (58). Within this system, depletion of Sec8 significantly increased the percentage of cells expressing GFP relative to the controls (Fig. 4D), consistent with increased HR.

FIG 4.

Sec8 depletion accelerates low-fidelity DNA repair. (A, left) Representative images of DNA damage (neutral comet tail) following irradiation of U2OS cells 72 h after transfection with the indicated siRNAs. (Right) Quantification of comet tail length per cell in pixels (MM, 28,805 ± 927.0 [n = 91]; Sec8, 18,866 ± 841.3 [n = 96]). *** indicates a P value of <0.0001. (B, top) Ionizing radiation response in U2OS cells seeded and irradiated at the indicated doses at 72 h posttransfection. At 10 days post-IR exposure, colonies with >50 cells were counted, and survival fractions were calculated. Results shown represent data from 3 independent experiments performed in triplicate. * indicates a P value of <0.05 for 4 Gy and 8 Gy. (Bottom) Western blot confirmation of knockdown. (C, top) Representative images of metaphase spreads in colcemid-treated (3 h) U2OS cells following exposure to 1 Gy irradiation and 30 min of incubation with the indicated doses at 72 h posttransfection. Blue arrows indicate breaks, and red arrows indicate exchanges. (Bottom) Quantification of data. Results shown represent data from 2 independent experiments. * indicates a P value of <0.05. UT, untransfected. (D) HR frequency shown as percent accumulation of GFP-positive MCF7 cells following transfection with the indicated siRNAs and the I-SceI plasmid. * indicates a P value of <0.01. (E) Expression of Sec8 and knockdown of Sec8 with three different specific siRNAs. (F) Detection of repair proteins at I-SceI-induced DSB sites by ChIP with overexpression of Sec8 and knockdown of Sec8 with specific siRNAs. I-SceI-induced DSB and ChIP analyses were done according to procedures described previously (30). The closest PCR product to the DSB site is located between 94 and 378 nucleotides from the DSB site.

As depletion of Sec8 enhances HR-mediated DNA DSB repair, we wanted to determine whether depletion of Sec8 influences the recruitment of HR-related proteins. To determine the effect of Sec8 depletion on recruitment at the I-SceI site, cleavage was analyzed by ChIP, using RAD51-, KU80-, and BRCA1-specific antibodies in combination with site-specific PCR primer pairs. Levels of RAD51, BRCA1, and KU80 appeared elevated in closest proximity to the DNA DSB and declined as the distance from the DSB increased. In contrast, the levels of Rad51 and BRCA1 at the DSB site were elevated in cells depleted for Sec8 (Fig. 4E and F), whereas levels of KU80 were unaltered.

Finally, we assayed fluorescent nucleotide analog incorporation into individual DNA fibers to monitor replication fork stalling and origin-of-replication firing in response to replicative DNA damage (33). To measure stalled replication kinetics, cells were pulse-labeled with 5-iododeoxyuridine (IdU), exposed to hydroxyurea (HU) for 1, 4, or 24 h to deplete the nucleotide pool, and then allowed to recover in the presence of 5-chlorodeoxyuridine (CldU). To measure new replication origins, cells were pulse-labeled with IdU, exposed to HU, and then pulse-labeled with CldU for 1, 4, and 24 h, as previously described (59, 60). Within this system, stalled replication forks are indicated by IdU-positive/CldU-negative fibers, and new origins of replication are indicated by IdU-negative/CldU-positive fibers. Equivalent accumulations of stalled replication forks (IdU-only tracks) were observed in both control and Sec8 siRNA-treated cells (Fig. 5), suggesting that Sec8 depletion does not deflect damage-induced replicative arrest. However, Sec8-depleted cells displayed an increased frequency of new replication origins (CldU-only tracks) by 4 h and 24 h after exposure to HU (Fig. 5). The latter observation is consistent with faster resolution of DSB-induced cell cycle arrest in Sec8-depleted cells. These cumulative observations indicate that Sec8 depletion alters both the kinetics and fidelity of DNA repair.

FIG 5.

Effect of Sec8 depletion on stalled replication and new origins as determined by a chromatin fiber assay. (Left) Representative images of DNA fibers at 24 h with or without hydroxyurea in U2OS cells 72 h after transfection with the indicated siRNAs. (Right, top) Quantification of IdU-only-labeled fibers, representing stalled replication, after the indicated times of HU exposure 72 h after transfection with the indicated siRNAs. (Bottom) Quantification of CldU-only-labeled fibers, representing new origins of replication, after the indicated times of HU exposure 72 h after transfection with the indicated siRNAs. Bars indicate means and standard errors of the means from 3 independent experiments. * indicates a P value of <0.05, and ** indicates a P value of <0.01.

Sec8 regulates histone-modifying proteins ATF2 and RNF20.

In an effort to parse key effectors driving the altered DDR kinetics and genomic instability seen upon Sec8 depletion, we queried available synthetic genetic array (SGA) data from yeast for genetic interactions with exocyst mutants. Analysis of the SGA resource using pairwise correlation coefficients allows for clustering of significant (r = >0.1) biological relationships (61). Among the exocyst mutants that were not single-gene lethal (62, 63), we identified a functional cluster centered on Sec3 that contained additional subunits of the exocyst and the histone acetyltransferase HPA2, which suggests that the exocyst complex participates in the regulation of histone modification (Fig. 6A) (r = 0.282 for Sec15; r = 0.278 for Sec10; r = 0.23 for Exo70; r = 0.221 for HPA2). Notably, yeast HPA2 and the human exocyst-interacting protein ATF2 have overlapping biochemical specificities for histone H4 lysine acetylation. Together with Tip60 (KAT5), ATF2 histone acetyltransferase activity supports homologous recombination-mediated DNA damage repair (6, 64). We found that Sec8 depletion resulted in a marked accumulation of ATF2 protein in the absence of altered mRNA transcript concentrations (Fig. 6B and C). We noted that the exocyst PPI domain contains the Huwe1 and Ubr5 E3 ligases (Fig. 1C). Evidence suggests that both Huwe1 and Ubr5 participate in the negative regulation of BRCA1-mediated homologous recombination (65–67). Like Sec8, depletion of Huwe1 resulted in the accumulation of ATF2 (Fig. 6D). The elevated ATF2 protein concentration corresponded to enhanced nuclear accumulation of ATF2 and enhanced IR-induced phosphorylation of ATF2 at its ATM kinase substrate sites (S490/S498), suggesting increased ATF2 activity (Fig. 6E) (68, 69). This was mirrored by the accumulation of Tip60 (Fig. 6F). Importantly, RNF20 protein also accumulated upon Sec8 depletion independently of altered transcript levels (Fig. 6G). As previously reported, aberrant accumulation of this E3 ligase is associated with genomic instability (70). We found that RNF20 accumulation correlated with the ubiquitination of its target protein H2B on K120, together with the accumulation of H2BK120ub-dependent trimethylation of H3K4 (Fig. 6H). These histone modifications are required for proper chromatin remodeling and subsequent recruitment of DNA repair factors in both the NHEJ and HR pathways (7, 47). The significant increase in the number of HR-associated Rad51 foci upon Sec8 depletion indicated that these altered histone modifications affect repair pathway specification (Fig. 6I).

FIG 6.

Sec8 modulates the accumulation and activity of DNA damage-associated histone modifiers. (A) SGA-derived correlation network centered on the yeast exocyst component Sec3, derived from the DRYGIN similarity matrix. Node length corresponds to the r value (r = 0.282 for Sec15; r = 0.278 for Sec10; r = 0.23 for Exo70; r = 0.221 for HPA2). (B) Immunoblots from whole-cell lysates with the indicated antibodies for U2OS cells after exposure to 5 Gy irradiation and incubation for the indicated time points 72 h after transfection with the indicated siRNAs. Results shown are representative of data from 3 independent experiments. (C) qRT-PCR analysis of relative ATF2 mRNA expression levels in U2OS cells 72 h after transfection with the indicated siRNAs. Experiments were performed in triplicate. (D) Immunoblots from whole-cell lysates with the indicated antibodies for U2OS cells 72 h after transfection with the indicated siRNAs. (E, left) Representative images of nuclear ATF2 in U2OS cells transfected with the indicated siRNAs. (Right) Image-based quantification of ATF2 nuclear intensity with ImageJ using a DAPI mask and mean intensity. Results shown represent the means of data from 2 independent experiments (MM, 34.57 ± 0.9943 [n = 84]; Sec8, 49.68 ± 1.718 [n = 77]). ***, P < 0.0001. (F and G) Accumulation of the indicated proteins was evaluated as described above for panel B. Results shown are representative of data from 3 independent experiments. (H) qRT-PCR analysis of relative RNF20 expression levels in U2OS cells 72 h after transfection with the indicated siRNAs. Experiments were performed in triplicate. (I, left) Representative images of Rad51 foci in U2OS cells with 0 Gy or 4 h after 5 Gy irradiation 72 h after transfection with the indicated siRNAs. (Right) Image-based quantification of Rad51 foci. Cells with >8 foci were scored as positive (MM, 16.10 ± 0.4403 [n = 124]; Sec8, 19.84 ± 0.6516 [n = 116]). **** indicates a P value of <0.0001.

DISCUSSION

Collectively, the observations described here suggest that the exocyst complex enhances DNA damage repair, at least in part, by restraining the activity of chromatin remodeling factors in the absence of appropriate DNA damage signaling (Fig. 7). Upon depletion of Sec8, the promiscuous accumulation of chromatin modifiers appears to serve as a priming event for the generation and resolution of DNA damage repair foci that favors HR and resolution of stalled replication forks by upregulation of histone posttranslational modifiers such as ATF2 and RNF20. While conferring a modicum of resistance to IR, the resulting enhanced DNA repair activity upon Sec8 depletion is of low fidelity with a high-level recombination that generates chromosomal aberrations. Although HR is typically considered to support higher-fidelity repair than NHEJ, evidence indicates that altered HR is a key initiator of tumorigenesis (71, 72). Thus, exocyst perturbation can be considered to result in a genomic instability phenotype.

FIG 7.

Schematic depicting suggested spatial and temporal regulation of histone-modifying proteins driven by Sec8 and the exocyst complex. Some preexisting modifications like H4K16ac have an impact on the DNA damage response and the influence of Sec8 on the choice of DNA DSB repair by HR, whereas in the absence of these modifications, this HR can become hyperrecombinogenic, which could lead to genomic instability. HATs, histone acetyltransferases; ME, methylation; UB, ubiquitination; AC, acetylation.

Of note, somatic mutations in Sec8 have been identified in colorectal cancers, and Sec8 deletions are significantly overrepresented in colorectal tumors of African American patients (73). If associated with defective DDR activity, these alterations could both confer resistance to radiotherapy and enhance genomic instability within the tumor harboring them. A key focus for future investigations is the nature of the selective pressure leading to the physical and functional association of the exocyst and canonical DNA repair machinery. By extrapolation from the recently described mechanistic participation of the exocyst in host defense signaling and autophagy, we suspect that this macromolecular machine serves as a spatially constrained platform from which to effectively coordinate localized enzyme/substrate interactions (19–21). Thus our data indicate that, in addition to having a major role in membrane function, Sec8 as part of the exocyst complex interacts with DNA repair and chromatin modification proteins. The importance of these interactions in maintaining genomic stability is indicated by the increased number of chromosomal aberrations observed in cells with perturbed Sec8 levels, which correlated with higher homologous recombination activity.