Functional and mutational landscapes of BRCA1 for homology-directed repair and therapy resistance

Rachel W Anantha; Srilatha Simhadri; Tzeh Keong Foo; Susanna Miao; Jingmei Liu; Zhiyuan Shen; Shridar Ganesan; Bing Xia

doi:10.7554/eLife.21350

. 2017 Apr 11;6:e21350. doi: 10.7554/eLife.21350

Functional and mutational landscapes of BRCA1 for homology-directed repair and therapy resistance

Rachel W Anantha ^1,^2,^†, Srilatha Simhadri ^1,^2,^3,^†, Tzeh Keong Foo ^1,^2,^†, Susanna Miao ⁴, Jingmei Liu ^1,², Zhiyuan Shen ^1,², Shridar Ganesan ^1,^2,³, Bing Xia ^1,^*

Editor: Charles L Sawyers⁵

PMCID: PMC5432210 PMID: 28398198

Abstract

BRCA1 plays a critical role in homology-directed repair (HDR) of DNA double strand breaks, and the repair defect of BRCA1-mutant cancer cells is being targeted with platinum drugs and poly (ADP-ribose) polymerase (PARP) inhibitors. We have employed relatively simple and sensitive assays to determine the function of BRCA1 variants or mutants in two HDR mechanisms, homologous recombination (HR) and single strand annealing (SSA), and in conferring resistance to cisplatin and olaparib in human cancer cells. Our results define the functionality of the top 22 patient-derived BRCA1 missense variants and the contribution of different domains of BRCA1 and its E3 ubiquitin ligase activity to HDR and drug resistance. Importantly, our results also demonstrate that the BRCA1-PALB2 interaction dictates the choice between HR and SSA. These studies establish functional and mutational landscapes of BRCA1 for HDR and therapy resistance, while revealing novel insights into BRCA1 regulatory mechanisms and HDR pathway choice.

DOI: http://dx.doi.org/10.7554/eLife.21350.001

Research Organism: Human

eLife digest

Genes are the instruction manuals of life and contain the information needed to build the proteins that keep cells alive. Over time, genes can accumulate errors or mutations and eventually become faulty, which can lead to diseases like cancer. Sometimes mutations can be passed on through generations and increase the chances of getting cancer. The BRCA1 gene, for example, provides instructions for making a protein that helps to repair or remove damaged DNA and stops cells from growing uncontrollably. When the BRCA1 gene becomes faulty, cells could continue to grow with damaged DNA. This makes it more likely for cancer to develop, especially breast cancer and ovarian cancer.

However, not all changes in BRCA1 gene cause the protein to become faulty or lead to cancer. In fact, about 30% of BRCA1 gene changes identified by genetic tests are referred to as ‘variants of uncertain clinical significance’, meaning that it is not clear if these variants are indeed mutations that could affect the clinical outcome of the people that carry them. Software predictions based largely on patient data have categorized many of these variants as not cancer-causing, but the majority still need to be experimentally tested and confirmed. Many studies have tried to determine the effect of selected variants on the BRCA1 protein, but a complete picture remains lacking.

Now, Anantha et al. have tested the top 22 common variants in the BRCA1 gene, some of which had known effects and some did not. The study tested how these variants affect the ability of the protein to repair damaged DNA and the efficacy of chemotherapies targeting cancer cells with a DNA repair defect. The experiments revealed that three specific parts of the protein must remain intact in order for the protein to carry out this activity, i.e. mutations that affect these three areas are likely to cause cancer and also make cancer cells vulnerable to these chemotherapies. Anantha et al. also generated a series of 10 artificially shortened BRCA1 proteins, each missing a specific part, to determine the possible effects of other variants in those missing parts.

Together the findings reveal previously unknown effects of certain variants that are commonly seen in cancer patients as well new insights into how the BRCA1 protein repairs DNA. The next step will be to assess rarer variants where little data is available. A better understanding of how these variants affect DNA repair and drug response will help to improve the genetic counseling and treatment of patients with breast cancer and ovarian cancer.

DOI: http://dx.doi.org/10.7554/eLife.21350.002

Introduction

Germline, heterozygous mutations in BRCA1 confer high risk of breast and ovarian cancer development in an autosomal dominant fashion (Couch et al., 2014; Fackenthal and Olopade, 2007). BRCA1 has been implicated in numerous cellular processes including DNA repair, cell cycle checkpoints, centrosome duplication, and transcriptional regulation, etc. (Deng, 2006; Mullan et al., 2006; Roy et al., 2012). Ever since BRCA1 was found to localize to discrete nuclear foci and colocalize with the recombination enzyme RAD51 (Scully et al., 1997), its function in homologous recombination (HR)-based repair of DNA double strand breaks (DSBs) has been a subject of intense study (Moynahan and Jasin, 2010). Tumors arising from BRCA1 mutation carriers usually show loss of the wild-type (wt) allele, which renders tumor cells biallelically null for the gene. It is generally believed that genome instability resulting from the DNA repair defect following the loss of BRCA1 is a driver of tumor development (Li and Greenberg, 2012; Venkitaraman, 2014). Importantly, the very DNA repair defect that leads to tumor development is also an ‘Achilles’ Heel’ of the resulting tumor cells, which can be selectively killed by suitable DNA-damaging agents that target HR defect, such as platinum drugs and poly (ADP-ribose) polymerase (PARP) inhibitors (Lord and Ashworth, 2016).

The human BRCA1 gene consists of 24 exons encoding a large polypeptide of 1863 amino acid residues. BRCA1 contains a RING domain at the N terminus and a tandem BRCT domain at the C terminus (Figure 1A). Two nuclear localization signals (NLSs) facilitate the localization of BRCA1 primarily to the nucleus (Chen et al., 1996), whereas a nuclear export signal (NES) can mediate its cytoplasmic export (Rodríguez and Henderson, 2000). The RING domain of BRCA1 binds to a similar domain in its close partner BARD1 (Wu et al., 1996), leading to the formation of a stoichiometric complex that possesses substantial ubiquitin E3 ligase activity (Hashizume et al., 2001; Ruffner et al., 2001). At the same time, binding of BARD1 to BRCA1 shields the NES and helps retain BRCA1 in the nucleus (Fabbro et al., 2002). The BRCT domain directly binds, in a phosphorylation-dependent manner, to at least three other proteins, namely BRIP1, CtIP and Abraxas, all of which have function in DNA repair (Huen et al., 2010; Jiang and Greenberg, 2015). Additionally, BRCA1 contains a highly conserved coiled-coil (CC) motif that directly binds PALB2, the partner and localizer of BRCA2 (Xia et al., 2006), which links BRCA1 and BRCA2 in the HR pathway (Sy et al., 2009; Zhang et al., 2009a, 2009b).

Figure 1. — (A) Domain structure of BRCA1 and the binding sites for its interacting partners. NES, nuclear export signal; NLS, nuclear localization signal. (B) Effects of the BRCA1 variants on the binding of BARD1, PALB2, BRIP1 and Abraxas. The 3xMyc-tagged BRCA1 proteins were transiently expressed in 293T cells and IPed with an anti-Myc antibody. WCE, whole cell extract. (C) Schematic of the HR reporter assay. The DR-GFP reporter contains two defective copies of the GFP gene, one disrupted by an I-SceI site and the other lacking a promoter. I-SceI cutting of the first copy generates a DSB, and repair by HR with the second copy as a template leads to restoration of a functional GFP gene. (D) HR activities of the variants relative to the wt BRCA1 protein. Data shown are the means from two to seven independent experiments for each variant or mutant. Error bars represent standard deviations (SDs). The grey bars indicate variants that are among the top 20 missense variants but are already present in BRCA1 cDNAs obtained from three independent sources. The calculated cutoff threshold is indicated by horizontal lines. **p<0.01. See Figure 1—source data 1 for details. (E) Levels of BRCA1 protein following knockdown and re-expression. Cells treated with a control siRNA (NSC1) were used as a control for the endogenous protein abundance.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.003

Figure 1—source data 1. HR activities of the BRCA1 variants and mutants analzyed in this study.
See texts and materials and methods for details.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.004

elife-21350-fig1-data1.xlsx^{(23.7KB, xlsx)}

DOI: 10.7554/eLife.21350.004

Figure 1—figure supplement 1. — (A) Domain structure of BRCA1 and the binding sites for its interacting partners. NES, nuclear export signal; NLS, nuclear localization signal. (B) Effects of the BRCA1 variants on the binding of BARD1, PALB2, BRIP1 and Abraxas. The 3xMyc-tagged BRCA1 proteins were transiently expressed in 293T cells and IPed with an anti-Myc antibody. WCE, whole cell extract. (C) Schematic of the HR reporter assay. The DR-GFP reporter contains two defective copies of the GFP gene, one disrupted by an I-SceI site and the other lacking a promoter. I-SceI cutting of the first copy generates a DSB, and repair by HR with the second copy as a template leads to restoration of a functional GFP gene. (D) HR activities of the variants relative to the wt BRCA1 protein. Data shown are the means from two to seven independent experiments for each variant or mutant. Error bars represent standard deviations (SDs). The grey bars indicate variants that are among the top 20 missense variants but are already present in BRCA1 cDNAs obtained from three independent sources. The calculated cutoff threshold is indicated by horizontal lines. **p<0.01. See Figure 1—source data 1 for details. (E) Levels of BRCA1 protein following knockdown and re-expression. Cells treated with a control siRNA (NSC1) were used as a control for the endogenous protein abundance.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.003

Figure 1—source data 1. HR activities of the BRCA1 variants and mutants analzyed in this study.
See texts and materials and methods for details.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.004

elife-21350-fig1-data1.xlsx^{(23.7KB, xlsx)}

DOI: 10.7554/eLife.21350.004

The Breast Cancer Information Core (BIC) database was the first major public database into which patient-derived mutations and variants were deposited. It contains a total of 15,311 entries of BRCA1 sequence alterations, among which 6133 (40%) are frameshift or protein-truncating mutations, which can generally be classified as pathogenic. Importantly, 4577 (30%) are missense variants, most of which are considered to be of ‘unknown clinical importance’, or commonly known as variants of unknown (or uncertain) clinical significance (VUSs), the interpretation of which remains a challenge (Couch et al., 2014). Recently, more data have been deposited into the ClinVar database, which now includes all BIC cases along with ones from multiple other sources. Based largely on genetic and epidemiological evidence, a large number of VUSs has been designated by ClinVar as ‘benign’. However, the majority of the variants have not been functionally characterized.

A number of studies have been conducted to assess the functions of select BRCA1 VUS in transcription, DNA repair, and in supporting the viability of mouse embryonic stem (ES) cells (Bouwman et al., 2013; Chang et al., 2009; Millot et al., 2012; Ransburgh et al., 2010; Towler et al., 2013). However, data on the impact of VUS on drug sensitivity have only begun to emerge. In particular, a recent study used a recombinase-mediated cassette exchange (RMCE) approach to profile 74 patient-derived missense variants for their activities in conferring resistance to cisplatin and a PARP inhibitor in mouse ES cells, and it found that all functionally deleterious variants were confined to the RING and BRCT domains (Bouwman et al., 2013).

Results

The HR activity of BRCA1 requires intact RING, CC and BRCT domains

To establish an initial functional landscape of BRCA1 mutations for HR-mediated DSB repair, we constructed 32-patient-derived missense variants along the entire length of the protein (Table 1). The list included 11 variants with known or partially known functional consequences: two in the RING domain (C61G and C64R) that disrupt BARD1 binding, three in the CC motif (M1400V, L1407P and M1411T) that affect PALB2 binding and six in the BRCT domain (S1655F, C1697R, R1699Q, A1708E, S1715R and M1775R) that abrogate BRCA1 interaction with Abraxas, BRIP1 and CtIP (Figure 1B and Figure 1—figure supplement 1). Also included were the top 20-patient-derived missense variants in the BIC database (21 in total due to a tie at the 20th place) and the top 22 in the ClinVar database (Table 1). Additionally, we included four other VUSs, L668F, S1101N, S1140G and T1561I, which have now been recorded between 16 and 40 times in ClinVar. As a control for the BRCT missense variants, we also generated a patient-derived frameshift mutation (5055delG or p.Val1646Serfs) that truncates the protein immediately before the BRCT domain.

Table 1.

BIC database and ClinVar reports of patient-derived BRCA1 missense variants characterized in this study.

DOI: http://dx.doi.org/10.7554/eLife.21350.007

HGVS cDNA	BIC designation	BIC entry count	BIC clinical importance	ClinVar individuals	ClinVar designation	Align-GVGD grade	IARC classification
c.4837A>G	S1613G	248	No	1319	Benign	C0	1
c.181T>G	C61G	239	Yes	323	Pathogenic	C65	5
c.2612C>T	P871L	211	No	1348	Benign	C0	1
c.3113A>G	E1038G	182	No	1237	Benign	C0	1
c.3548A>G	K1183R	164	No	1215	Benign	C0	1
c.4039A>G	R1347G	161	Unknown	712	Benign	C0	1
c.1067A>G	Q356R	155	Unknown	315	Benign	C0	1
c.3024G>A	M1008I	139	Unknown	146	Benign	C0	1
c.2521C>T	R841W	119	Unknown	128	Benign	C15	1
c.4883T>C	M1628T	96	Unknown	491	Benign	C0	1
c.1487G>A	R496H	86	Unknown	90	Benign	C0	1
c.736T>G	L246V	80	Unknown	83	Benign	C0	1
c.3119G>A	S1040N	68	Unknown	141	Benign	C0	1
c.4910C>T	P1637L	67	Unknown	76	Benign	C0	1
c.2077G>A	D693N	65	No	229	Benign	C0	1
c.4956G>A	M1652I	61	Unknown	94	Benign	C0	1
c.2315T>C	V772A	60	Unknown	62	Benign	C0	1
c.4535G>T	S1512I	58	No	72	Benign	C0	1
c.1456T>C	F486L	56	Unknown	58	Benign	C0	1
c.536A>G	Y179C	54	Unknown	59	Benign	C35	1
c.1648A>C	N550H	54	Unknown	59	Benign	C0	1
c.5123C>A	A1708E	46	Yes	80	Pathogenic	C65	5
c.5324T>G	M1775R	31	Unknown	43	Pathogenic	C45	5
c.3418A>G	S1140G	29	Unknown	40	Benign	C0	1
c.4682C>T	T1561I	26	Unknown	33	Conflicting	C0	N/A
c.2002C>T	L668F	25	Unknown	28	Benign	C0	1
c.3302G>A	S1101N	14	Unknown	16	Benign	C0	1
c.190T>C	C64R	12	Unknown	15	Conflicting	C65	N/A
c.5096G>A	R1699Q	11	Unknown	22	Conflicting	C35	5
c.5145C>G	S1715R^#	5	Unknown	5	Pathogenic	C65	5
c.4964C>T	S1655F	3	Unknown	4	Conflicting	C25	N/A
c.5098T>C	C1697R	3	Unknown	7	Conflicting	C65	N/A
c.4198A>G	M1400V	1	Unknown	1	Uncertain	C0	N/A
c.4220T>C	L1407P	1	Unknown	1	Uncertain	C65	N/A
c.4232T>C	M1411T	1	Unknown	1	Uncertain	C65	N/A
5055delG	V1646Sfs	5	Yes	6	Pathogenic	-	-

Open in a new tab

Align-GVGD grade: C0 to C65 denote increasing likelihood of a variant to cause damage (to protein function). IARC (ENIGMA) classification: 5, Definitely pathogenic; 4, Likely pathogenic; 3, Uncertain, 2, Likely not pathogenic or of little clinical significance; 1, Not pathogenic or of no clinical significance. # c.5145C>G, c.5145C>A and c.5143A>C have all been reported to generate BRCA1-S1715R. Data shown are up to date as of March 21, 2017.

Notably, our sequencing analysis revealed P871L, E1038G, K1183R and S1613G in the ‘wild-type’ BRCA1 cDNAs from multiple independent sources. Indeed, they are considered as neutral/benign by both BIC and ClinVar. Among the rest of the top 22 variants in ClinVar, C61G and A1708E are considered pathogenic, and the rest are all considered benign largely based on genetic or epidemiological evidence. Similarly, L668F, S1101N, S1140G and T1561I are now also considered as benign. Note that this study was initiated in 2010 largely based on BIC, in which most of the above are now still listed as VUS.

To measure the activity of BRCA1 variants in HR, we re-developed a previously described HR assay based on a ‘protein replacement’ strategy (Ransburgh et al., 2010; Sy et al., 2009; Xia et al., 2006) (Figure 1C). In this assay, the endogenous BRCA1 was first depleted in U2OS/DR-GFP cells (Nakanishi et al., 2005; Xia et al., 2006) using an siRNA targeting the 3’-UTR of the gene, and various BRCA1 proteins were then expressed from cDNAs (lacking the 3’-UTR) co-transfected with a plasmid expressing I-SceI, which cleaves the reporter to induce DSB formation and subsequent repair. To ensure uniformity of BRCA1 knockdown, siRNA transfections were performed in 10 cm plates and the cells were harvested, mixed and reseeded into six-well plates before the second transfection. The averaged expression levels of exogenous BRCA1 was three to five times that of the endogenous protein (Figure 1E), suggesting a significant degree of overexpression. However, the variants were expressed at levels comparable to that of the wt BRCA1 (not shown), with the exception being the BRCT point mutants, among which all but R1699Q showed much lower expression levels (Figure 1—figure supplement 2).

Under the setting used, C61G and C64R reduced BRCA1 HR activity by ~5-fold to a level slightly above the vector control (Figure 1D and Table 2). Similarly, a profound impact on HR was observed with mutations in the BRCT domain. Consistent with a previous report (Sy et al., 2009), variants L1407P and M1411T, which abrogate or substantially impair PALB2 binding, reduced HR by ~4.5-fold, whereas M1400V, which moderately impairs PALB2 binding (Figure 1B, lane 15), caused a ~ 25% decrease in the apparent HR activity. In contrast, all variants that are neutral for protein-protein interaction at the RING, coiled-coil and BRCT domains showed greater than 85% of the HR activity of the wt protein. Using variants that have been functionally characterized previously as controls, our assay was found to be both sensitive and specific, and the cutoff threshold of the assay was calculated to be 54.3–54.7%, which would define M1400V and all the protein-protein interaction neutral variants as being functionally benign. However, on an individual basis, the modest decrease of HR activity in the case of M1400V is statistically significant.

Table 2.

Comparison of results on HR activity and drug response of the BRCA1 variants analyzed in this study obtained from previous and this studies.

DOI: http://dx.doi.org/10.7554/eLife.21350.008

	Domain	HR activity (%)						Cisplatin response		Olaparib response
	Domain	Sy et al., 2009	Ransburgh et al. (2010)	Towler et al. (2013)	Bouwman et al. (2013)	Lu et al., 2015	This study	Bouwman et al. (2013)	This study	Bouwman et al. (2013)	This study
Vector		N/A	~10	~9	~20	~18	17.3	S	S	S	S
WT		98	100	100	100	100	100	R	R	R	R
C61G	RING	-	~17	-	-	23.6	21.2	S	S	-	S
C64R	RING	-	-	-	-	-	22.3	-	S	-	S
Y179C		-	-	~95	-	157.1	92.7	-	R	-	R
L246V		-	-	-	-	-	91.5	R	R	-	R
Q356R		-	-	-	-	-	95.5	-	R	-	R
F486L		-	-	-	-	160	95.8	-	R	-	R
R496H		-	-	-	-	-	95.9	-	R	-	R
N550H		-	-	-	-	90.8	88.4	-	R	-	R
L668F		-	-	-	-	96.8	93.6	R	R	-	R
D693N		-	-	-	-	-	111.5	R	R	-	R
V772A		-	-	-	-	110.4	84.5	-	R	-	R
R841W		-	-	-	-	-	98.2	-	R	-	R
P871L		-	-	-	-	-	wt	-	wt	-	wt
M1008I		-	-	-	-	-	99.2	R	R	-	R
E1038G		-	-	-	-	-	wt	-	wt	-	wt
S1040N		-	-	-	-	-	98.0	-	R	-	R
S1101N		-	-	-	-	-	97.3	R	R	-	R
S1140G		-	-	-	-	-	106.2	R	R	-	R
K1183R		-	-	-	-	-	wt	-	wt	-	wt
R1347G		-	-	-	-	-	106.5	-	R	-	R
M1400V	CC	56	-	-	-	-	74.8	R	S	R	S
L1407P	CC	24	-	-	-	-	24	S	S	S	S
M1411T	CC	25	-	-	-	-	26	R	S	R	S
S1512I		-	-	-	-	-	95.3	-	R	-	R
T1561I		-	-	-	-	133.2	102.3	-	R	-	R
S1613G		-	-	-	-	-	wt	-	wt	-	wt
M1628T		-	-	-	-	107	107.6	R	R	-	R
P1637L		-	-	-	-	98.8	99.5	-	R	-	R
5055△G	BRCT	-	-	-	-	-	22.5	-	S	-	S
M1652I	BRCT	-	-	-	-	-	106.2	R	R	-	R
S1655F	BRCT	-	-	-	~30	-	8.4	S	S	-	S
C1697R	BRCT	-	-	-	-	-	8.0	-	S	-	S
R1699Q	BRCT	-	-	-	~45	-	16.9	S	S	S	S
A1708E	BRCT	-	-	-	-	-	10.7	-	S	-	S
S1715R	BRCT	-	-	-	-	-	9.3	-	S	-	S
M1775R	BRCT	36	-	~6	-	-	10.2	-	S	-	S

Open in a new tab

R, resistant; S, sensitive. See Figure 1—source data 1, Figure 2—source data 1 and Figure 3—source data 1 for details.

The BRCA1-PALB2-BRCA2 interactions prevent RAD52-dependent SSA

Given the proposed role of BRCA1 on DSB end resection (Bunting et al., 2010), the activities of all the above variants in SSA, a deletion-causing repair pathway that also utilizes resected DNA ends, were measured. For this purpose, we used a similar protein replacement strategy with another U2OS-based reporter cell line containing an integrated SA-GFP reporter (Stark et al., 2004). Consistent with a previous report that Brca1-deficient mouse cells show reduced SSA (Stark et al., 2004), depletion of BRCA1 in the human cells resulted in a reduction of the (already low) basal SSA activity (Figure 2A), and re-expression of wt BRCA1 restored the SSA activity (Figure 2B). As reported previously that pathogenic missense mutations in the RING domain can cause defects in both HR and SSA (Towler et al., 2013), we found that the BARD1-binding mutants (C61G and C64R) showed greatly reduced SSA activities (Figure 2B). At the same time, all the BRCT mutants were completely defective. In contrast, the three mutations that impact PALB2 binding all resulted in increased SSA activity. In particular, the L1407P and M1411T mutations elevated SSA by ~2.5-fold. This finding suggests that BRCA1 interactions with BARD1 and BRCT-binding partners are both required for its function in the resection step of the DSB repair, whereas the BRCA1-PALB2 interaction functions downstream to promote HR and reduce SSA.

Figure 2. — (A) Effect of BRCA1 depletion on SSA. Data shown are the means of the four data points obtained from two independent experiments each performed in duplicates. Error bars represent SDs. ***p<0.001. (B) SSA activities of the BRCA1 variants relative to the wt protein. A schematic of the SA-GFP reporter assay is shown at the upper left corner. Data shown are the means ± SDs from two to five independent experiments for each variant or mutant. *p<0.05. See Figure 2—source data 1 for details. (C) Effects of BRCA2 and PALB2 depletion on SSA. Data shown are the means ± standard errors of mean (SEMs) from four independent experiments. *p<0.05; **p<0.01. (D) Requirement of RAD52 for SSA upregulation following BRCA2 and PALB2 depletion. Data shown are the means ± SEMs from three independent experiments. *p<0.05; **p<0.01. (E) BRCA1 and BRCA2-binding defects of PALB2-L21A, L35A and A1025R mutants. The FLAG-HA-tagged mutants were transiently expressed in 293T cells and IPed with anti-FLAG M2 agarose beads. (**F–G**) HR (F) and SSA (G) activities of PALB2-L21A, L35A and A1025R mutants. Data shown are the means ± SDs from three to five independent experiments for each mutant. ***p<0.001. (G) Levels of PALB2 protein following knockdown and re-expression in U2OS/DR-GFP cells. Cells untreated with any siRNA were used as a control for the endogenous protein abundance.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.009

Figure 2—source data 1. SSA activities of the BRCA1 variants and mutants analyzed in this study.
See texts and materials and methods for details.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.010

elife-21350-fig2-data1.xlsx^{(23.1KB, xlsx)}

DOI: 10.7554/eLife.21350.010

It has been reported that loss of BRCA2 increases SSA in mouse cells (Stark et al., 2004; Tutt et al., 2001), and a recent report showed that some PALB2 heterozygous human lymphoblastoid cell lines have increased SSA activity (Obermeier et al., 2016). To better understand the role of PALB2 and BRCA2 in this process, we depleted the two proteins in parallel in the U2OS/SA-GFP reporter cells and measured the effect. As shown in Figure 2C, depletion of either protein caused substantial upregulation of SSA, with BRCA2 siRNAs stimulating SSA by 1.8–5.4 fold and PALB2 siRNAs 5.6–7.2 fold. Initially, two siRNAs for each gene were used (#1949 and #9025 for BRCA2 and #1493 and #2693 for PALB2). The number of BRCA2 siRNAs were later increased to six due to the observation that siRNAs #1949 and #9025 showed a large difference in the fold change of SSA they induced. With the exception of siRNA #11170, all BRCA2 siRNAs showed similar and effective knockdown. Although we cannot explain why siRNA #1949 stood out as the sole outlier in terms of SSA induction, we did notice that this siRNA had the least negative effect on cell morphology and growth rate (data not shown). The level of RAD51 was reduced by all BRCA2 siRNAs, which could be explained by potentially reduced stability in the absence of BRCA2, a ‘carrier’ of RAD51 that can simultaneously bind six molecules of the latter (Jensen et al., 2010; Liu et al., 2010) and thus could potentially serve as a stabilizer as RAD51. However, there is no correlation between the levels of RAD51 reduction and SSA increase among all siRNAs used (Figure 2C), suggesting that the partial loss of RAD51 is not a significant cause of SSA upregulation.

RAD52 has been shown to possess strand annealing activity in vitro (Mortensen et al., 1996) and to promote SSA in vivo (Stark et al., 2004). To test if RAD52 is required for the increased SSA in the absence of BRCA2 and PALB2, we silenced its expression, either alone or in combination with BRCA2 and PALB2, and measured the effect on SSA efficiency. As shown in Figure 2D, strong depletion of RAD52 (by siRNA #1972) reduced basal SSA, whereas relatively mild depletion of RAD52 (by siRNA #2569) showed little effect. Strong RAD52 depletion greatly impeded the increase of SSA elicited by both BRCA2 and PALB2 loss, while the relatively weaker depletion of RAD52 also reduced the upregulation of SSA upon loss of either BRCA2 or PALB2, albeit to lesser extents.

To further elucidate the role of PALB2 in the DSB repair pathway choice, we employed point mutants that are defective for BRCA1 binding (L21A and L35A) or BRCA2 binding (A1025R) (Figure 2D) (Oliver et al., 2009; Sy et al., 2009). As expected, and in contrast to wt PALB2, which effectively restored HR in reporter cells depleted of the endogenous protein, the BRCA1- and BRCA2-binding mutants all showed dramatically reduced HR activity (Figure 2E). Re-expression of wt PALB2 in the SA-GFP reporter cells depleted of the endogenous PALB2 significantly suppressed the greatly enhanced SSA activity, whereas neither the BRCA1-binding mutants nor the BRCA2-binding mutant showed such ability (Figure 2F). Very recently, we reported the first patient-derived missense pathogenic mutation in PALB2, L35P, that disrupts the binding of BRCA1 (Foo et al., 2017). In a way virtually identical to L35A, this mutant also failed to suppress SSA in the above setting (data not shown). The suppression by wt PALB2 was moderate, which could be due to that the depletion of PALB2 might have caused certain secondary effects that cannot be immediately restored upon sudden re-expression of the protein. Taken together, our results establish that PALB2 controls the HR/SSA pathway choice following resection and that the interactions between PALB2 with both BRCA1 and BRCA2 are required for this function.

Effects of BRCA1 missense variants on PARPi and cisplatin resistance

To profile the functional impact and therapeutic relevance of BRCA1 variants, we developed a relatively simple drug sensitivity assay (Figure 3A–B) using the BRCA1-deficient, triple negative MDA-MB-436 breast cancer cells (Elstrodt et al., 2006). Briefly, MDA-MB-436 cells were transfected with cDNA constructs expressing wt BRCA1 or the variants and then subjected to selection with either Geneticin (G418) alone (for transfection efficiency) or a combination of G418 and either olaparib, a potent PARP inhibitor (PARPi), or cisplatin. Colonies were allowed to grow for a period of 3–4 weeks. Under the condition used, we never obtained a single olaparib-resistant colony in untransfected or vector transfected cells; in only one (out of about ten) occasion, four cisplatin-resistant colonies were obtained from vector transfected cells. These results allow us to rule out any meaningful contribution of the endogenous truncated BRCA1 to our readout. Thus, MDA-MB-436 is a suitable cell line for BRCA1 functional analysis using our protocol.

Resistance to PARPi was found to largely correlate with the HR activity of each variant (Figure 3C). Specifically, the BARD1-binding mutants (C61G and C64R), the PALB2-binding mutants (M1400V, M1407P and M1411T) and the BRCT mutants all demonstrated virtually no growth under PARP inhibition, whereas other VUSs were fully or largely resistant. Notably, BRCA1-M1400V, which showed only moderately impaired binding to PALB2 and 73% of the HR activity of the wt protein (Figure 1B,D), failed to confer PARPi resistance. The pattern of cisplatin resistance also largely reflected the HR activities of the variants, as the RING, CC and BRCT mutants all showed hypersensitivity, while other variants conferred greater than ~90% colony-forming ability of the wt protein. These data clearly demonstrate the correlation between HR activity, but not SSA activity, and drug resistance. They also underscore the importance of the RING, CC and BRCT domains and their corresponding binding partners for BRCA1 function in DNA repair.

Localization defect of BRCT missense mutants

To understand the cause of the functional deficiency of the mutants more fully, we analyzed their subcellular localization in U2OS cells. Although the endogenous BRCA1 is almost exclusively localized in the nucleus, transiently expressed exogenous BRCA1 can be found in the cytoplasm in some cells, with the extent depending on the transfection condition. This is commonly explained by the insufficient amount of the endogenous BARD1, which is thought to shield the NES of BRCA1 thereby retaining BRCA1 in the nucleus (Fabbro et al., 2002). Indeed, BRCA1-C61G, which does not bind BARD1, showed a more diffuse localization pattern than the wt protein (Figure 4A). The PALB2-binding mutant M1411T behaved like the wt protein, as reported before (Sy et al., 2009). Interestingly, patient-derived mutations that truncate BRCA1 at the BRCT domain have been reported to cause BRCA1 to localize primarily in the cytoplasm (Rodriguez et al., 2004), even though the mutants retain intact NLSs and can still bind BARD1. The localization defect of the BRCT mutants cannot be rescued by Leptomycin B (LMB), which blocks CRM1-dependent export but can instead be rescued by co-expression of BARD1 (Rodriguez et al., 2004). In this vein, we tested the above aspects for the BRCT missense mutations and obtained the same results (Figure 4B–E). We consider the BRCT point mutants to be different from truncating mutants for two reasons: first, the majority of the these mutants were expressed at much lower levels than the wt protein presumably due to destabilization, whereas the truncating mutants were not (Figure 1—figure supplement 2), although frameshift mutations often cause nonsense-mediated decay (NMD) of mRNA in vivo; second, the missense mutants may still bind known interaction partners with low affinity or even certain unknown factors that may affect their localization. Collectively, the data lend strong support to the notion that the BRCT mutants may be unable to enter the nucleus and further suggest that the BRCT domain may bind a heretofore unidentified factor that promotes BRCA1 nuclear entry, with BARD1 directly or indirectly involved in the process.

Figure 4. — (A) Immunofluorescence staining of wt BRCA1 and representative RING, CC and BRCT mutants. 3xMyc-tagged BRCA1 expression constructs were transiently transfected into U2OS/DR-GFP cells first depleted of endogenous BRCA1, cells were fixed 48 hr after transfection, and the proteins were stained with an anti-Myc antibody. (B) Quantification of the subcellular distribution of wt BRCA1 and a panel of BRCT mutants. cDNA constructs were transfected into U2OS cells and the tagged BRCA1 proteins were stained as in (A). (C) Effect of Leptomycin B (LMB) on the localization of wt BRCA1 and BRCA1-M1775R. The BRCA1 proteins were expressed and stained as in (A). LMB (50 ng/ml) was added 48 hr after transfection, and cells were incubated with LMB for 12 hr prior to fixation. Staining of endogenous cyclin B in untransfected U2OS cells was used as a positive control. (**D–E**) Rescue of the localization defect of BRCA1 BRCT mutants by BARD1. 3xMyc-tagged BRCA1 and FLAG-HA-tagged BARD1 were transiently co-expressed in U2OS cells, cells were fixed 48 hr after transfection, and the proteins were stained with Myc and HA antibodies, respectively. Panel D shows the staining patterns of wt and two representative mutants of BRCA1 and the co-expressed BARD1. Panel E shows the quantification of the results. (F) No rescue of the HR defect of the BRCA1 BRCT mutants by BARD1 co-expression. U2OS/DR-GFP cells were depleted of the endogenous BRCA1 for 48 hr and then co-transfected with BRCA1 and BARD1 expression constructs. GFP-positive cells were scored another 52 hr later. The values were normalized against that of wt BRCA1, which was set as 1. Data shown are means ± SDs from three independent experiments.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.013

To learn to what extent the lack of HR activity of the BRCT mutants is due to mislocalization, we tested their activities when co-expressed with exogenous BARD1. Co-expression of BARD1 slightly increased the HR activity of some of the mutants, but the overall effect was modest at best (Figure 4F). For example, mutant S1655F showed comparable, if not better, localization than the wt protein (Figure 4E), but still had a threefold lower HR activity. These data suggest that the lack of HDR activity of the BRCT mutants is due to both their localization defect and, perhaps more importantly, their inability to interact with one or more of the three binding partners, namely Abraxas, CtIP and BRIP1, or with another binding partner yet to be identified.

Role of the BRCA1 I26A mutation and BRCA1 sumoylation in HDR and drug resistance

The BRCA1/BARD1 heterodimer exhibits robust ubiquitin E3 ligase activity in vitro (Hashizume et al., 2001; Ruffner et al., 2001). An artificial I26A mutation was isolated in a screen for residues in the RING domain that abrogate E3 ligase activity without affecting BARD1 binding (Brzovic et al., 2003). It was first thought that the E3 ligase activity may play an important role in DNA repair and/or tumor suppression; however, mouse cells carrying a I26A knockin allele did not show any significant defect in HR activity (Reid et al., 2008), and the mutant protein was still able to suppress mammary tumor development in Brca1 conditional knockout models (Shakya et al., 2011). To our surprise, I26A substantially reduced, by about three fold, the coIP of endogenous BARD1 with ectopically expressed BRCA1 (Figure 5A), indicating that this mutation significantly weakens the BRCA1-BARD1 interaction. Thus, I26A is a complex mutation affecting at least two different BRCA1 properties. Functionally, I26A reduced the HR activity of BRCA1 by ~30% (Figure 5B); although the apparent HR defect caused by I26A was mild, the mutation caused a 67% reduction in colony formation in the presence of cisplatin and virtually a complete loss of resistance to olaparib (Figure 5C).

Figure 5. — (A) Effect of I26A and K119R mutations on BRCA1 binding to BARD1 and other interacting partners. The proteins were transiently expressed in 293T cells and IPed with anti-Myc. (B) Quantification of the BARD1-binding capacity of the BRCA1 mutants. Data shown are the means ± SDs of the ratios of BARD1 and BRCA1 band intensities from four independent experiments. **p<0.01. (C) HR and SSA activities of the BRCA1 mutants relative to that of the wt protein. Data shown are the means ± SDs from two to six independent experiments for each mutant. ***p<0.001. See Figure 1—source data 1 and Figure 2—source data 1 for details. (D) Abilities of the BRCA1 mutants to confer resistance to cisplatin and olaparib. Data shown are the means ± SDs from three independent experiments for both mutants. **p<001. See Figure 3—source data 1 for details.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.014

It has been reported that BRCA1 can be sumoylated and this modification promotes the E3 ligase activity of the BRCA1/BARD1 complex (Morris et al., 2009). The same study identified lysine 119 (K119) as a critical residue for BRCA1-SUMO conjugation and the promotion of its E3 ligase activity. Thus, we generated and analyzed a K119R mutation in parallel with the I26A mutant. K119R had no effect on BRCA1 HR, nor did it affect cisplatin resistance (Figure 5B–C); however, colony formation in the presence of olaparib was reduced by ~30%, suggesting a possible HR-independent role of BRCA1 sumoylation for PARPi resistance.

Deletion analyses of BRCA1 for HDR and drug resistance

To systematically assess the contribution of various BRCA1 structural elements in HDR and drug resistance, we generated a series of 10 overlapping deletions (Figure 6A). Interestingly, BD1, which lacks both the RING domain and the NES, was expressed at a much higher level, and despite the complete lack of the RING domain, it still associated with a small but significant amount of BARD1 (Figure 6B); it showed ~50% activity for both HR and SSA (Figure 6C), which is significantly higher than the activities of either C61G or C64R mutants (Figure 1D). As expected, BD8, which lacks the CC motif, was unable to bind PALB2, and BD10 lacking the BRCT domain failed to bind Abraxas and BRIP1. In line with observations made with the point mutants (Figures 1D and 2A), BD8 showed no HR activity and a threefold increase in SSA, while BD10 displayed similar HR and SSA activities as the vector. Notably, BD4, lacking both NLSs, still retained ~75% activity in both HR and SSA assays, indicating that a significant amount of BRCA1 can still be recruited into the nucleus. In contrast, BD7, which retains all known functional domains, showed only ~50% of HR and SSA activities, suggestive of the possible existence of a novel element in the deleted region that is important for DNA repair.

Consistent with their lack of HR activity, BD8 and BD10 were completely or nearly completely unable to confer resistance to either drug (Figure 6D). Although BD1, BD4 and BD7 all displayed 50–75% apparent HR activity in the reporter assay, they supported no more than 20% colony formation. Also, BD9 only supported ~50% of colony formation despite having no defect in the aforementioned protein interactions and possessing greater than 80% apparent HR and SSA activities. These observations demonstrate a positive but non-linear correlation between apparent HR activity and drug resistance. The non-linearity could be due to different effects of the deletions on protein folding, stability, conformation, posttranslational modifications and dynamics in protein-protein interactions; it may also reflect the differences between the two cell lines (U2OS/DR-GFP vs MDA-MB-436), time frames used or the different nature of I-SceI and drug-induced DNA breaks. Surprisingly, BD6 was twice as effective as the wt protein in supporting colony formation in the presence of olaparib but less effective in the presence of cisplatin, indicating a significant difference in the mechanisms of resistance to the two drugs despite a common requirement for HR.

Discussion

In this study, we systematically assessed HR and SSA activities of 32 natural variants, including 11-patient-derived missense mutations that affect known protein-protein interactions and the top 22-patient-associated missense variants, as well as 10 overlapping deletion mutations that span the entire length of the protein. Our data clearly demonstrate that the HR function of BRCA1 requires the RING, CC and BRCT domains. Interestingly, deletion of residues 1056–1357 (BD7) caused ~50% reduction in both HR and SSA activity without affecting the binding of any of the above known partners. This segment of BRCA1 does not contain any recognizable domains and has not been implicated in HR before, suggesting the existence of a novel binding partner or regulatory mechanism. Together, these results establish a functional landscape of BRCA1 for its HDR function.

The RING domain of BRCA1 binds to the RING domain of BARD1, and the BRCA1/BARD1 heterodimer formation is critical for their stability, their E3 ubiquitin ligase activity and the nuclear retention of BRCA1 (Fabbro et al., 2002; Hashizume et al., 2001; Ruffner et al., 2001). Conditional knock out of either Brca1 or Bard1 in murine mammary epithelial cells led to the development of mammary carcinomas that are indistinguishable from each other (Shakya et al., 2008), and the BRCA1 C61G mutant that abrogates BARD1 binding failed to suppress mammary tumor development in mice (Drost et al., 2011). Moreover, the I26A mutation that abolishes its E3 ligase activity did not affect HR activity or mitomycin C (MMC) sensitivity (Reid et al., 2008), nor did it affect tumor suppression in mice (Shakya et al., 2011). These findings suggested that the BARD1-binding function of the RING domain, rather than its E3 ligase activity, plays a key role in BRCA1 HR and tumor suppression activity. In this regard, our findings on the I26A mutation are worth noting. First, it caused a significant impairment in BARD1 binding (Figure 5A); second, it moderately reduced HR efficiency (Figure 5B); and third, it caused sensitivity to both cisplatin and olaparib under the condition used (Figure 5C). The reduced BARD1 binding is consistent with the fact that I26A mutant mouse cells had substantially reduced amount of both BRCA1 and BARD1 (Reid et al., 2008). Also, a recent report showed that HeLa cells expressing only BRCA1-I26A were similarly sensitive to olaparib as BRCA1 knockdown cells (Densham et al., 2016). Thus, it is safe to conclude that the I26A mutation affects HR and drug resistance in human cells. The discrepancy in observed HR activities and drug sensitivities in mouse ES cells and human cancer cells may be explained by the different tolerance for partial loss of BARD1 binding in the two systems or the presence of other mutations in cancer cells. At the same time, although I26A mutant mouse cells showed no defect in apparent HR activity or MMC resistance, reduced gene targeting efficiency and increased chromosomal abnormalities after MMC treatment were noted (Reid et al., 2008), indicative of a possible defect in HR in mouse cells as well. Despite the discrepancy, it should be noted that our results do not contradict the argument that the E3 ligase activity of BRCA1 is dispensable for tumor suppression (Shakya et al., 2011).

Mutations of the BRCT domain dramatically abolish HR activity, and the mechanism is also complex. First, this domain directly interacts with at least three proteins, Abraxas, a scaffold protein (Wang et al., 2007), BRIP1, a DNA helicase (Cantor et al., 2001), and CtIP, an endonuclease involved in resection (Sartori et al., 2007). It has recently been shown that CtIP binding to BRCA1 is dispensable for CtIP-mediated DNA resection (Polato et al., 2014; Reczek et al., 2013), yet to what extent BRCA1 HR activity depends on Abraxas and BRIP1 remains unclear. Second, mutation of the BRCT domain causes a major defect in the nuclear accumulation of BRCA1 (Figure 4A–B), supporting a key role of the BRCT domain for BRCA1 nuclear entry. Our finding that BARD1 can significantly rescue nuclear localization of BRCT mutants but not their HR defects (Figure 4F) also supports an important and direct role of the BRCT domain for BRCA1 DNA repair activity. Collectively, our data demonstrate the dual role of the BRCT domain and underscore a complex regulation of BRCA1 localization and repair function by the RING and BRCT domains and their binding partners (Figure 7A).

Figure 7. — (A) Roles of BRCA1 structural elements and binding partners on its nuclear localization and HDR (HR and SSA) activities. In brief, the NLSs and BRCT domain both promote BRCA1 nuclear entry, whereas the NES mediates BRCA1 export from the nucleus. BARD1 bound to the RING domain shields the NES of BRCA1 thereby promoting its nuclear retention. The RING and BRCT domains are required for both HR and SSA, whereas the CC domain promotes HR but inhibits or prevents SSA through its binding to PALB2. (B) A model of the BRCA1-PALB2-BRCA2 axis in the regulation of HR and SSA. Following DSB formation, the BRCA1/BARD1 complex was recruited to DNA damage sites by Abraxas and perhaps another yet to be defined factor(s). The presence of BRCA1 at damage sites promotes resection and inhibits non-homologous end joining (NHEJ), at least in part, by counteracting the resection-inhibiting activity of the 53BP1-RIF1 complex. At the same time, BRCA1 helps recruit the PALB2/BRCA2 complex, which displaces RPA from the resected ssDNA ends and loads RAD51 to initiate HR. When PALB2 or BRCA2 is lost or when the direct interactions in the BRCA1-PALB2-BRCA2 axis are disrupted, RAD52 binds to resected ends and mediates SSA when homologous sequences are available, leading to genomic deletions. It is unclear whether PALB2 or BRCA2 can directly suppress the binding of RAD52 to ssDNA or inhibiting its strand annealing activity.

**DOI:** http://dx.doi.org/10.7554/eLife.21350.016

Consistent with early observations made in mouse cells (Stark et al., 2004), BRCA1 depletion reduced the efficiency of both HR and SSA, whereas depletion of BRCA2 led to severe loss of HR (data not shown) but increased SSA (Figure 2B). As efficient resection is necessary for both repair mechanisms, the data is consistent with the notion that BRCA1 plays a key role in resection (Bunting et al., 2010). Importantly, we found that depletion of PALB2 showed similar, if not more dramatic, induction of SSA as did BRCA2 depletion (Figure 2B). Moreover, all three point mutations in BRCA1 that affect PALB2 binding led to increased SSA usage (Figure 2A), and two different PALB2 mutants with greatly impaired BRCA1-binding capacity both failed to suppress the upregulation of SSA in PALB2-depleted cells (Figure 2F). These results clearly demonstrate that the direct interaction between BRCA1 and PALB2 is required for suppressing, or at least preventing, SSA, while promoting HR (Figure 7A–B). As such, our results support two critical roles of BRCA1 function in DSB repair, promoting resection and recruiting/stimulating PALB2, which then channels the ssDNA down the HR path (Figure 7B). Based on the available data, PALB2 is the key partner of BRCA1 specifically for its HR function and the primary switch point, acting upstream of BRCA2, for HDR pathways following resection.

VUSs are commonly found during clinical genetics tests; however, their biological and clinical significance is often difficult to ascribe with confidence, and this uncertainty poses significant challenges for both clinicians and patients. Our results demonstrate that relatively common VUSs outside the RING and BRCT domains do not significantly affect the HDR activity of BRCA1. Our findings are mostly consistent with several previous studies (Table 2), including the above noted mouse ES cell-based study that covered 6 of the top 22 variants (Bouwman et al., 2013); on the rare PALB2-binding mutants L1407P and M1411T, however, our results showed that they are severely defective in HR (Figure 1D) and conferring drug resistance (Figure 3C), whereas their effects in mouse cells were less pronounced. Moreover, M1400V, which was found to be neutral in the mouse cell system, shows clear sensitivity to olaparib treatment in the human cell line. This discrepancy could again be due to a difference in the thresholds of human and mouse cells to tolerate partial loss of BRCA1-PALB2 interaction or the presence of other mutations in the human cancer cells.

Platinum-based therapeutics have been commonly used for treatment of ovarian cancer, and olaparib has recently been approved for the treatment of BRCA1/2 mutant ovarian cancer, with multiple trials on breast and other cancers ongoing or recently completed. These agents generate lesions that block or collapse DNA replication forks, leading to DSBs that require HR and therefore BRCA1/2 for repair (Lord and Ashworth, 2016). Under the condition used, olaparib appeared to have a better selectivity in killing MDA-MB-436 cells expressing mutant BRCA1 proteins. Although this could be due to a slightly lower than optimal concentration of cisplatin used, the possibility that cisplatin is intrinsically less selective than olaparib cannot be ruled out. Based on data from this study and other recent studies, patients with relatively common missense mutations in the RING and BRCT domains, such as C61G and A1708E, respectively, or rare mutations in the CC motif, such as L1407P or M1411T, are likely to respond to platinum- or PARPi-based therapies, unless resistance mechanisms have already occurred prior to treatment or are induced by the treatment (Lord and Ashworth, 2013). However, except for C61G and A1708E, all other top 22 missense variants are fully or mostly functional both in HR and in conferring cisplatin and olaparib resistance (Figure 3C), which corroborates with clinical genetics evidence to further cast doubts on their pathogenicity and diminish the likelihood of a positive response of the patients to the above regimens. Based on the results of our deletion analyses and other available data, it is unlikely for any VUS outside the RING, CC and BRCT domains to be strongly pathogenic or to elicit high sensitivity to platinum or PARPi therapies (unless it destabilizes the transcript or protein). However, the region deleted in BD7 (residues 1056–1,307) may harbor VUS that have intermediate effects.

Materials and methods

Cell culture

U2OS/DR-GFP HR reporter cells were described before (Nakanishi et al., 2005; Xia et al., 2006). U2OS/SA-GFP SSA reporter cells (Gunn and Stark, 2012) were a gift from Dr. Jeremy Stark. U2OS and 293T cell lines were purchased from American Type Culture Collection (ATCC). These four cell lines were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1x Penicillin-Streptomycin. MDA-MD-436 cells was cultured in DMEM/F12 (1:1) supplemented with 10% FBS and 1x Penicillin-Streptomycin. All cells were cultured in a humidified chamber with 5% CO₂ at 37°C. Mycoplasma was not tested, but the cells have been cultured in the presence of Plasmocin (ant-mpt, InvivoGen) to eliminate potential mycoplasma contamination. No commonly misidentified cell lines were used.

Plasmids and mutagenesis

All BRCA1 expression constructs were based on pcDNA-3xMyc-BRCA1 (Chen et al., 1998). The pOZ-FH-C1-PALB2 vector expressing FLAG-HA-tagged PALB2 was described before (Ma et al., 2012). The pOZ-FH-N-BARD1 expressing FLAG-HA-tagged BARD1 was described before (Greenberg et al., 2006). All mutations or deletions were generated through site-directed mutagenesis following the QuikChange protocol (Agilent Technologies).

cDNA transfection, immunoprecipitation (IP) and western blotting

For testing protein-protein interaction, BRCA1 expression constructs were transfected into 293T cells plated at a density of 5 × 10⁵ cells per well in six-well plates and transfected with 2 μg of plasmid per well using FuGENE HD or X-tremeGENE 9 XP. Cells were collected 30 hr after transfection and lysed with 350 µl of NETNG-300 (300 mM NaCl, 1 mM EDTA, 20 mM Tris-HCl, 0.5% Nonidet P-40% and 10% Glycerol) containing Complete protease inhibitor cocktail (Roche). The 3xMyc-tagged BRCA1 proteins were IPed for 3–4 hr with anti-Myc (9E10, Covance) and protein A-agarose beads (Roche). For western blotting analyses, proteins were resolved on 4–12% Tris-Glycine SDS gels, transferred onto nitrocellulose membranes and probed with the following antibodies- Myc (9E10, Covance), PALB2 (M11) (Xia et al., 2006), BARD1 (H300, Santa Cruz), BRIP1 (a gift from Dr. Sharon Cantor, University of Massachusetts Medical School), Abraxas (ab139191, AbCam). The secondary antibodies used were Horseradish peroxidase (HRP)-conjugated sheep anti-mouse IgG (NA931V, GE Healthcare) and donkey anti-rabbit IgG (NA9340V, GE Healthcare). Immobilon Western Chemiluminescent HRP Substrate (Millipore) was used to develop the blots.

Measurement of HR and SSA efficiency

To measure the HR and SSA activities of the BRCA1 and PALB2 variants, U2OS/DR-GFP and U2OS/SA-GFP cells were plated in 10 cm dish at 1.2 × 10⁶ cells per dish and allowed to adapt overnight. Cells were transfected with an siRNA targeting BRCA1 3’-UTR using Lipofectamine RNAiMax (Life Technologies) (10 nM final concentration of siRNA and 12 µl of RNAiMax per plate). Cells were split into six-well plates (200,000 cells per well) 30 hr after transfection. After another 18 hr, cells were co-transfected with BRCA1 or PALB2 expression constructs (1 µg) and the I-SceI expression plasmid pCBAsce (1.5 µg) using 6 µl of X-tremeGene 9 (Roche). Cells were harvested 54 hr post the second transfection, and GFP-positive cells were counted by fluorescence-assisted cell sorting (FACS). The sense strand sequence of the siRNA is GGAUCGAUUAUGUGACUUAdTdT.

To measure SSA activity after BRCA2, PALB2 or RAD52 knockdown with different siRNAs, U2OS/SA-GFP cells were seeded into six-well plates at density of 175,000 cells per well and transfected with siRNAs ~18 hr later using using Lipofectamine RNAiMax (Life Technologies) (10 nM final concentration of each siRNA and 3 µl of RNAiMax per well). Media were refreshed 24 hr post transfection, and cells were transfected again with pCBAcse with X-tremeGene 9 another 24 hr later (48 hr after the first transfection). Cells were collected and subjected to FACS analysis ~52 hr after the second transfection. See Supplementary file 1 for the sequences of the siRNAs used.

Immunofluorescence

Cells were seeded onto coverslips in 12-well plates the day before transfection. Cells were transfected with 1 µg BRCA1 expression constructs using 2.5 µl of FuGENE 6 or X-tremeGENE 9 (Roche). At 48 hr after transfection, cells were fixed with 3% (w/v) paraformaldehyde (in PBS with 300 mM sucrose) for 10 min at room temperature, permeabilized with 0.5% Triton X-100 (in PBS) and then sequentially incubated with primary and secondary antibodies (diluted in PBS containing 5% goat serum) for 1 hr each at 37°C. Each of the above steps was followed by three PBS washes. After staining, coverslips were mounted onto glass slides with VECTASHIELD Mounting Medium with DAPI (Vector Labs) and observed using a fluorescent microscope. The primary antibodies used were anti-Myc (9E10, Covance) and anti-BRCA1 (#07–434, Millipore).

Assay for PARPi and cisplatin resistance

MDA-MB-436 cells were seeded into six-well plates at a density of 1 × 10⁶ cells per well the day before transfection. Cells were transfected with 1 µg of empty vector or 2 µg of BRCA1 expression vectors using 6 µl of X-tremeGENE 9 (Roche). Cells were trypsinized 36 hr after transfection and reseeded into 10 cm plates at 100,000 cells per plate in a volume of 10 ml. Another 16 hr later, 1 ml of 10X G418 (6 mg/ml dissolved in the same culture medium) or 10X G418 containing 2.2 μM cisplatin or olaparib was added to each plate. The final concentrations of G418, cisplatin and olaparib were 550 µg/ml, 200 nM and 200 nM, respectively. All experiments were performed in duplicate plates for each construct. For cells selected with G418 alone, 15 days after selection, one plate was trypsinized and cells counted with a Vi-CELL Cell Counter (Beckman Coulter) to determine viable cell number, and the other plate was stained with Crystal Violet (0.5% w/v in 95% ethanol, 5% acetic acid) to count the number of colonies. For cells selected with both G418 and cisplatin, both plates were stained 21 days after selection. For cells subjected to G418 and olaparib double selection, both plates were stained 28 days after selection. Colonies with approximately 50 cells or more were counted, and the numbers were normalized against the number of viable cells on corresponding plates with G418 alone. The vast majority of constructs were measured by three or more independent experiments (each with duplicate sets of plates), with the only exceptions being when the first two experiments produced nearly identical results or were carried out using two independent plasmids.

Sensitivity and specificity analyses

We assessed the sensitivity and specificity of our assays using the Receiver Operating Characteristics (ROC) analysis to simultaneously capture the maximum sensitivity and specificity of the assays with 95% confidence interval as previously described (Guidugli et al., 2013). The sensitivity of each assay was defined as the ability to detect variants retaining wt activity of BRCA1 in the respective assays and specificity the ability to detect variants showing loss of BRCA1 function. For the calculation of specificity and sensitivity of our assays, we compared the values of benign variants (Align-GVD grade of C0 and IARC classification of 1) and pathogenic variants (Align-GVD grade of C35-C65 and IARC classification of 5) that have been confirmed by previous functional studies (Tables 1 and 2). We also included values of the wt cDNA in the benign group and the vector as well as a bona fide pathogenic mutation 5055delG (p.Val1646Serfs or L1657STOP) in the pathogenic group. Threshold of each assays were defined individually. To rule out variability in the assays, we also calculated the sensitivity and specificity using the highest observed values for the pathogenic variants and the lowest observed values for the benign variants with the cut off used to define either the upper or lower limit of the threshold. For each of the assays, dotted lines were drawn to represent the cut off threshold that captures the specificity and sensitivity of the assays at 100%, respectively. Analysis was not performed for the SSA assay as elevated SSA activity is associated with PALB2-binding mutants, which has yet to be defined, while loss of SSA is associated with BRCT and RING domain mutations. All analyses were conducted using GraphPad Prism 5.0 (GraphPad Software).

Statistical analysis

Raw data were normalized against values of wt BRCA1, wt PALB2, control siRNA (NSC1) or the empty vector, where applicable. Statistical significance was calculated using Students’ t test. p-Values smaller than 0.05 were considered significant.

Acknowledgements

We thank Drs. Jeremy Stark (City of Hope) for providing the U2OS/SA-GFP reporter cells, Sharon Cantor (Univ. of Massachusetts) for the BRIP1 antibody and Roger Greenberg (Univ. of Pennsylvania) for the pOZ-FH-N-BARD1 plasmid. We also thank Dr. Jean-Yves Masson (Laval University, Quebec) for sharing unpublished results. This work was supported by the National Cancer Institute (R01CA138804 and R01CA188096 to BX, R01CA169182 to SG and R01CA195612 to ZS) and by the Flow Cytometry Core Facility of Rutgers Robert Wood Johnson Medical School, a Shared Resource of Rutgers Cancer Institute of New Jersey (P30CA072720) and NIH Shared Instrumentation Grant (1 S10 RR025468). RWA was a recipient of a postdoctoral fellowship from the Department of Defense Breast Cancer Research Program (W81XWH-10-1-0486).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

National Cancer Institute R01CA138804 to Bing Xia.
Congressionally Directed Medical Research Programs W81XWH-10-1-0486 to Rachel W Anantha.
National Cancer Institute R01CA188096 to Bing Xia.
National Cancer Institute R01CA169182 to Shridar Ganesan.
National Cancer Institute R01CA195612 to Zhiyuan Shen.

Additional information

Competing interests

The authors declare that no competing interests exist.

Author contributions

RWA, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing.

SS, Data curation, Formal analysis, Investigation, Methodology, Writing—review and editing.

TKF, Data curation, Formal analysis, Investigation, Methodology, Writing—review and editing.

SM, Data curation, Investigation, Writing—review and editing.

JL, Resources, Investigation.

ZS, Resources, Funding acquisition, Writing—review and editing.

SG, Supervision, Funding acquisition, Writing—review and editing.

BX, Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing.

Additional files

Supplementary file 1. Sequences of siRNAs used in this study.

DOI: http://dx.doi.org/10.7554/eLife.21350.017

elife-21350-supp1.docx^{(13.5KB, docx)}

DOI: 10.7554/eLife.21350.017

References

Bouwman P, van der Gulden H, van der Heijden I, Drost R, Klijn CN, Prasetyanti P, Pieterse M, Wientjens E, Seibler J, Hogervorst FB, Jonkers J. A high-throughput functional complementation assay for classification of BRCA1 missense variants. Cancer Discovery. 2013;3:1142–1155. doi: 10.1158/2159-8290.CD-13-0094. [DOI] [PubMed] [Google Scholar]
Brzovic PS, Keeffe JR, Nishikawa H, Miyamoto K, Fox D, Fukuda M, Ohta T, Klevit R. Binding and recognition in the assembly of an active BRCA1/BARD1 ubiquitin-ligase complex. PNAS. 2003;100:5646–5651. doi: 10.1073/pnas.0836054100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bunting SF, Callén E, Wong N, Chen HT, Polato F, Gunn A, Bothmer A, Feldhahn N, Fernandez-Capetillo O, Cao L, Xu X, Deng CX, Finkel T, Nussenzweig M, Stark JM, Nussenzweig A. 53bp1 inhibits homologous recombination in Brca1-deficient cells by blocking resection of DNA breaks. Cell. 2010;141:243–254. doi: 10.1016/j.cell.2010.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cantor SB, Bell DW, Ganesan S, Kass EM, Drapkin R, Grossman S, Wahrer DC, Sgroi DC, Lane WS, Haber DA, Livingston DM. BACH1, a novel helicase-like protein, interacts directly with BRCA1 and contributes to its DNA repair function. Cell. 2001;105:149–160. doi: 10.1016/S0092-8674(01)00304-X. [DOI] [PubMed] [Google Scholar]
Chang S, Biswas K, Martin BK, Stauffer S, Sharan SK. Expression of human BRCA1 variants in mouse ES cells allows functional analysis of BRCA1 mutations. Journal of Clinical Investigation. 2009;119:3160–3171. doi: 10.1172/JCI39836. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chen CF, Li S, Chen Y, Chen PL, Sharp ZD, Lee WH. The nuclear localization sequences of the BRCA1 protein interact with the importin-alpha subunit of the nuclear transport signal receptor. Journal of Biological Chemistry. 1996;271:32863–32868. doi: 10.1074/jbc.271.51.32863. [DOI] [PubMed] [Google Scholar]
Chen J, Silver DP, Walpita D, Cantor SB, Gazdar AF, Tomlinson G, Couch FJ, Weber BL, Ashley T, Livingston DM, Scully R. Stable interaction between the products of the BRCA1 and BRCA2 tumor suppressor genes in mitotic and meiotic cells. Molecular Cell. 1998;2:317–328. doi: 10.1016/S1097-2765(00)80276-2. [DOI] [PubMed] [Google Scholar]
Couch FJ, Nathanson KL, Offit K. Two decades after BRCA: setting paradigms in personalized Cancer care and prevention. Science. 2014;343:1466–1470. doi: 10.1126/science.1251827. [DOI] [PMC free article] [PubMed] [Google Scholar]
Deng CX. BRCA1: cell cycle checkpoint, genetic instability, DNA damage response and Cancer evolution. Nucleic Acids Research. 2006;34:1416–1426. doi: 10.1093/nar/gkl010. [DOI] [PMC free article] [PubMed] [Google Scholar]
Densham RM, Garvin AJ, Stone HR, Strachan J, Baldock RA, Daza-Martin M, Fletcher A, Blair-Reid S, Beesley J, Johal B, Pearl LH, Neely R, Keep NH, Watts FZ, Morris JR. Human BRCA1-BARD1 ubiquitin ligase activity counteracts chromatin barriers to DNA resection. Nature Structural & Molecular Biology. 2016;23:647–655. doi: 10.1038/nsmb.3236. [DOI] [PMC free article] [PubMed] [Google Scholar]
Drost R, Bouwman P, Rottenberg S, Boon U, Schut E, Klarenbeek S, Klijn C, van der Heijden I, van der Gulden H, Wientjens E, Pieterse M, Catteau A, Green P, Solomon E, Morris JR, Jonkers J. BRCA1 RING function is essential for tumor suppression but dispensable for therapy resistance. Cancer Cell. 2011;20:797–809. doi: 10.1016/j.ccr.2011.11.014. [DOI] [PubMed] [Google Scholar]
Elstrodt F, Hollestelle A, Nagel JH, Gorin M, Wasielewski M, van den Ouweland A, Merajver SD, Ethier SP, Schutte M. BRCA1 mutation analysis of 41 human breast Cancer cell lines reveals three new deleterious mutants. Cancer Research. 2006;66:41–45. doi: 10.1158/0008-5472.CAN-05-2853. [DOI] [PubMed] [Google Scholar]
Fabbro M, Rodriguez JA, Baer R, Henderson BR. BARD1 induces BRCA1 intranuclear foci formation by increasing RING-dependent BRCA1 nuclear import and inhibiting BRCA1 nuclear export. Journal of Biological Chemistry. 2002;277:21315–21324. doi: 10.1074/jbc.M200769200. [DOI] [PubMed] [Google Scholar]
Fackenthal JD, Olopade OI. Breast Cancer risk associated with BRCA1 and BRCA2 in diverse populations. Nature Reviews Cancer. 2007;7:937–948. doi: 10.1038/nrc2054. [DOI] [PubMed] [Google Scholar]
Foo TK, Tischkowitz M, Simhadri S, Boshari T, Zayed N, Burke KA, Berman SH, Blecua P, Riaz N, Huo Y, Ding YC, Neuhausen SL, Weigelt B, Reis-Filho JS, Foulkes WD, Xia B. Compromised BRCA1-PALB2 interaction is associated with breast Cancer risk. Oncogene. 2017 doi: 10.1038/onc.2017.46. [DOI] [PMC free article] [PubMed] [Google Scholar]
Greenberg RA, Sobhian B, Pathania S, Cantor SB, Nakatani Y, Livingston DM. Multifactorial contributions to an acute DNA damage response by BRCA1/BARD1-containing complexes. Genes & Development. 2006;20:34–46. doi: 10.1101/gad.1381306. [DOI] [PMC free article] [PubMed] [Google Scholar]
Guidugli L, Pankratz VS, Singh N, Thompson J, Erding CA, Engel C, Schmutzler R, Domchek S, Nathanson K, Radice P, Singer C, Tonin PN, Lindor NM, Goldgar DE, Couch FJ. A classification model for BRCA2 DNA binding domain missense variants based on homology-directed repair activity. Cancer Research. 2013;73:265–275. doi: 10.1158/0008-5472.CAN-12-2081. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gunn A, Stark JM. I-SceI-based assays to examine distinct repair outcomes of mammalian chromosomal double strand breaks. Methods in Molecular Biology. 2012;920:379–391. doi: 10.1007/978-1-61779-998-3_27. [DOI] [PubMed] [Google Scholar]
Hashizume R, Fukuda M, Maeda I, Nishikawa H, Oyake D, Yabuki Y, Ogata H, Ohta T. The RING heterodimer BRCA1-BARD1 is a ubiquitin ligase inactivated by a breast cancer-derived mutation. Journal of Biological Chemistry. 2001;276:14537–14540. doi: 10.1074/jbc.C000881200. [DOI] [PubMed] [Google Scholar]
Huen MS, Sy SM, Chen J. BRCA1 and its toolbox for the maintenance of genome integrity. Nature Reviews Molecular Cell Biology. 2010;11:138–148. doi: 10.1038/nrm2831. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jensen RB, Carreira A, Kowalczykowski SC. Purified human BRCA2 stimulates RAD51-mediated recombination. Nature. 2010;467:678–683. doi: 10.1038/nature09399. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jiang Q, Greenberg RA. Deciphering the BRCA1 tumor Suppressor Network. Journal of Biological Chemistry. 2015;290:17724–17732. doi: 10.1074/jbc.R115.667931. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li ML, Greenberg RA. Links between genome integrity and BRCA1 tumor suppression. Trends in Biochemical Sciences. 2012;37:418–424. doi: 10.1016/j.tibs.2012.06.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
Liu J, Doty T, Gibson B, Heyer WD. Human BRCA2 protein promotes RAD51 filament formation on RPA-covered single-stranded DNA. Nature Structural & Molecular Biology. 2010;17:1260–1262. doi: 10.1038/nsmb.1904. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lord CJ, Ashworth A. Mechanisms of resistance to therapies targeting BRCA-mutant cancers. Nature Medicine. 2013;19:1381–1388. doi: 10.1038/nm.3369. [DOI] [PubMed] [Google Scholar]
Lord CJ, Ashworth A. BRCAness revisited. Nature Reviews Cancer. 2016;16:110–120. doi: 10.1038/nrc.2015.21. [DOI] [PubMed] [Google Scholar]
Lu C, Xie M, Wendl MC, Wang J, McLellan MD, Leiserson MD, Huang KL, Wyczalkowski MA, Jayasinghe R, Banerjee T, Ning J, Tripathi P, Zhang Q, Niu B, Ye K, Schmidt HK, Fulton RS, McMichael JF, Batra P, Kandoth C, Bharadwaj M, Koboldt DC, Miller CA, Kanchi KL, Eldred JM, Larson DE, Welch JS, You M, Ozenberger BA, Govindan R, Walter MJ, Ellis MJ, Mardis ER, Graubert TA, Dipersio JF, Ley TJ, Wilson RK, Goodfellow PJ, Raphael BJ, Chen F, Johnson KJ, Parvin JD, Ding L. Patterns and functional implications of rare germline variants across 12 cancer types. Nature Communications. 2015;6:10086. doi: 10.1038/ncomms10086. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ma J, Cai H, Wu T, Sobhian B, Huo Y, Alcivar A, Mehta M, Cheung KL, Ganesan S, Kong AN, Zhang DD, Xia B. PALB2 interacts with KEAP1 to promote NRF2 nuclear accumulation and function. Molecular and Cellular Biology. 2012;32:1506–1517. doi: 10.1128/MCB.06271-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
Millot GA, Carvalho MA, Caputo SM, Vreeswijk MP, Brown MA, Webb M, Rouleau E, Neuhausen SL, Hansen T, Galli A, Brandão RD, Blok MJ, Velkova A, Couch FJ, Monteiro AN, ENIGMA Consortium Functional Assay Working Group A guide for functional analysis of BRCA1 variants of uncertain significance. Human Mutation. 2012;33:1526–1537. doi: 10.1002/humu.22150. [DOI] [PMC free article] [PubMed] [Google Scholar]
Morris JR, Boutell C, Keppler M, Densham R, Weekes D, Alamshah A, Butler L, Galanty Y, Pangon L, Kiuchi T, Ng T, Solomon E. The SUMO modification pathway is involved in the BRCA1 response to genotoxic stress. Nature. 2009;462:886–890. doi: 10.1038/nature08593. [DOI] [PubMed] [Google Scholar]
Mortensen UH, Bendixen C, Sunjevaric I, Rothstein R. DNA strand annealing is promoted by the yeast Rad52 protein. PNAS. 1996;93:10729–10734. doi: 10.1073/pnas.93.20.10729. [DOI] [PMC free article] [PubMed] [Google Scholar]
Moynahan ME, Jasin M. Mitotic homologous recombination maintains genomic stability and suppresses tumorigenesis. Nature Reviews Molecular Cell Biology. 2010;11:196–207. doi: 10.1038/nrm2851. [DOI] [PMC free article] [PubMed] [Google Scholar]
Mullan PB, Quinn JE, Harkin DP. The role of BRCA1 in transcriptional regulation and cell cycle control. Oncogene. 2006;25:5854–5863. doi: 10.1038/sj.onc.1209872. [DOI] [PubMed] [Google Scholar]
Nakanishi K, Yang YG, Pierce AJ, Taniguchi T, Digweed M, D'Andrea AD, Wang ZQ, Jasin M. Human fanconi Anemia monoubiquitination pathway promotes homologous DNA repair. PNAS. 2005;102:1110–1115. doi: 10.1073/pnas.0407796102. [DOI] [PMC free article] [PubMed] [Google Scholar]
Obermeier K, Sachsenweger J, Friedl TW, Pospiech H, Winqvist R, Wiesmüller L. Heterozygous PALB2 c.1592delt mutation channels DNA double-strand break repair into error-prone pathways in breast Cancer patients. Oncogene. 2016;35:3796–3806. doi: 10.1038/onc.2015.448. [DOI] [PMC free article] [PubMed] [Google Scholar]
Oliver AW, Swift S, Lord CJ, Ashworth A, Pearl LH. Structural basis for recruitment of BRCA2 by PALB2. EMBO Reports. 2009;10:990–996. doi: 10.1038/embor.2009.126. [DOI] [PMC free article] [PubMed] [Google Scholar]
Polato F, Callen E, Wong N, Faryabi R, Bunting S, Chen HT, Kozak M, Kruhlak MJ, Reczek CR, Lee WH, Ludwig T, Baer R, Feigenbaum L, Jackson S, Nussenzweig A. CtIP-mediated resection is essential for viability and can operate independently of BRCA1. The Journal of Experimental Medicine. 2014;211:1027–1036. doi: 10.1084/jem.20131939. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ransburgh DJ, Chiba N, Ishioka C, Toland AE, Parvin JD. Identification of breast tumor mutations in BRCA1 that abolish its function in homologous DNA recombination. Cancer Research. 2010;70:988–995. doi: 10.1158/0008-5472.CAN-09-2850. [DOI] [PMC free article] [PubMed] [Google Scholar]
Reczek CR, Szabolcs M, Stark JM, Ludwig T, Baer R. The interaction between CtIP and BRCA1 is not essential for resection-mediated DNA repair or tumor suppression. The Journal of Cell Biology. 2013;201:693–707. doi: 10.1083/jcb.201302145. [DOI] [PMC free article] [PubMed] [Google Scholar]
Reid LJ, Shakya R, Modi AP, Lokshin M, Cheng JT, Jasin M, Baer R, Ludwig T. E3 ligase activity of BRCA1 is not essential for mammalian cell viability or homology-directed repair of double-strand DNA breaks. PNAS. 2008;105:20876–20881. doi: 10.1073/pnas.0811203106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rodriguez JA, Au WW, Henderson BR. Cytoplasmic mislocalization of BRCA1 caused by cancer-associated mutations in the BRCT domain. Experimental Cell Research. 2004;293:14–21. doi: 10.1016/j.yexcr.2003.09.027. [DOI] [PubMed] [Google Scholar]
Rodríguez JA, Henderson BR. Identification of a functional nuclear export sequence in BRCA1. Journal of Biological Chemistry. 2000;275:38589–38596. doi: 10.1074/jbc.M003851200. [DOI] [PubMed] [Google Scholar]
Roy R, Chun J, Powell SN. BRCA1 and BRCA2: different roles in a common pathway of genome protection. Nature Reviews Cancer. 2012;12:68–78. doi: 10.1038/nrc3181. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ruffner H, Joazeiro CA, Hemmati D, Hunter T, Verma IM. Cancer-predisposing mutations within the RING domain of BRCA1: loss of ubiquitin protein ligase activity and protection from radiation hypersensitivity. PNAS. 2001;98:5134–5139. doi: 10.1073/pnas.081068398. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sartori AA, Lukas C, Coates J, Mistrik M, Fu S, Bartek J, Baer R, Lukas J, Jackson SP. Human CtIP promotes DNA end resection. Nature. 2007;450:509–514. doi: 10.1038/nature06337. [DOI] [PMC free article] [PubMed] [Google Scholar]
Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J, Ashley T, Livingston DM. Association of BRCA1 with Rad51 in mitotic and meiotic cells. Cell. 1997;88:265–275. doi: 10.1016/S0092-8674(00)81847-4. [DOI] [PubMed] [Google Scholar]
Shakya R, Reid LJ, Reczek CR, Cole F, Egli D, Lin CS, deRooij DG, Hirsch S, Ravi K, Hicks JB, Szabolcs M, Jasin M, Baer R, Ludwig T. BRCA1 tumor suppression depends on BRCT phosphoprotein binding, but not its E3 ligase activity. Science. 2011;334:525–528. doi: 10.1126/science.1209909. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shakya R, Szabolcs M, McCarthy E, Ospina E, Basso K, Nandula S, Murty V, Baer R, Ludwig T. The basal-like mammary carcinomas induced by Brca1 or Bard1 inactivation implicate the BRCA1/BARD1 heterodimer in tumor suppression. PNAS. 2008;105:7040–7045. doi: 10.1073/pnas.0711032105. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stark JM, Pierce AJ, Oh J, Pastink A, Jasin M. Genetic steps of mammalian homologous repair with distinct mutagenic consequences. Molecular and Cellular Biology. 2004;24:9305–9316. doi: 10.1128/MCB.24.21.9305-9316.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sy SM, Huen MS, Chen J. PALB2 is an integral component of the BRCA complex required for homologous recombination repair. PNAS. 2009;106:7155–7160. doi: 10.1073/pnas.0811159106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Towler WI, Zhang J, Ransburgh DJ, Toland AE, Ishioka C, Chiba N, Parvin JD. Analysis of BRCA1 variants in double-strand break repair by homologous recombination and single-strand annealing. Human Mutation. 2013;34:439–445. doi: 10.1002/humu.22251. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tutt A, Bertwistle D, Valentine J, Gabriel A, Swift S, Ross G, Griffin C, Thacker J, Ashworth A. Mutation in Brca2 stimulates error-prone homology-directed repair of DNA double-strand breaks occurring between repeated sequences. The EMBO Journal. 2001;20:4704–4716. doi: 10.1093/emboj/20.17.4704. [DOI] [PMC free article] [PubMed] [Google Scholar]
Venkitaraman AR. Cancer suppression by the chromosome custodians, BRCA1 and BRCA2. Science. 2014;343:1470–1475. doi: 10.1126/science.1252230. [DOI] [PubMed] [Google Scholar]
Wang B, Matsuoka S, Ballif BA, Zhang D, Smogorzewska A, Gygi SP, Elledge SJ. Abraxas and RAP80 form a BRCA1 protein complex required for the DNA damage response. Science. 2007;316:1194–1198. doi: 10.1126/science.1139476. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wu LC, Wang ZW, Tsan JT, Spillman MA, Phung A, Xu XL, Yang MC, Hwang LY, Bowcock AM, Baer R. Identification of a RING protein that can interact in vivo with the BRCA1 gene product. Nature Genetics. 1996;14:430–440. doi: 10.1038/ng1296-430. [DOI] [PubMed] [Google Scholar]
Xia B, Sheng Q, Nakanishi K, Ohashi A, Wu J, Christ N, Liu X, Jasin M, Couch FJ, Livingston DM. Control of BRCA2 cellular and clinical functions by a nuclear partner, PALB2. Molecular Cell. 2006;22:719–729. doi: 10.1016/j.molcel.2006.05.022. [DOI] [PubMed] [Google Scholar]
Zhang F, Fan Q, Ren K, Andreassen PR. PALB2 functionally connects the breast Cancer susceptibility proteins BRCA1 and BRCA2. Molecular Cancer Research. 2009a;7:1110–1118. doi: 10.1158/1541-7786.MCR-09-0123. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zhang F, Ma J, Wu J, Ye L, Cai H, Xia B, Yu X. PALB2 links BRCA1 and BRCA2 in the DNA-damage response. Current Biology. 2009b;19:524–529. doi: 10.1016/j.cub.2009.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]

eLife. 2017 Apr 11;6:e21350. doi: 10.7554/eLife.21350.018

Decision letter

Editor: Charles L Sawyers¹

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Functional and mutational landscapes of BRCA1 for homology-directed repair and therapy resistance" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Charles Sawyers as the Senior Editor and Reviewing Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision. We are willing to consider a revised submission but only if it addresses the major concerns listed below under Essential Revisions.

Summary:

The authors have performed a comprehensive analysis of BRCA1 patient-derived mutations (VUS) for their effects on two DNA double-strand break (DSB) repair pathways, homologous recombination (HR) and single-strand annealing (SSA), and sensitivity to cisplatin and the PARP inhibitor olaparib. Mutations were examined along the length of the protein, including those falling in three established domains, the RING which interacts with partner protein BARD1, the coiled-coil domain which interacts with PALB2, and the BRCT which interacts with phosphoproteins, including Abraxis and BRIP1. The authors confirm that interactions with partner proteins are disrupted with mutations in these various domains. A total of 32 mutations were analyzed, including 11 with known functional consequences.

The authors make good use of established GFP reporters for HR and SSA. The duality of these reporters is notable in that proteins like BRCA1 that impair the first step of HR and SSA, end resection, are defective in both DSB repair pathways, while those like BRCA2 that impair downstream steps are decreased for HR and increased for SSA (Stark et al. 2004). The authors demonstrate that mutations that abrogate binding to key interacting proteins reduce HR and increase sensitivity to olaparib and cisplatin. Similarly, mutations that disrupt the RING or BRCT show the typical BRCA1 phenotype of a concomitant reduction in SSA, but those that disrupt PALB2 binding show a typical BRCA2-like phenotype of increased SSA (as does loss of PALB2 or binding to BRCA1). Thus, one could infer from these results that the first group is defective in end resection, while the second group is proficient at end resection but deficient in downstream steps (expected to be recruitment of BRCA2 based on previous reports). Deletion analyses largely support these conclusions, although there were some unexplained differences.

Overall this manuscript provides a comprehensive analysis of BRCA1 functional domains and patient-derived missense mutations for their role in DSB repair and response to chemotherapeutics, and differentiates the role of BRCA1 binding of PALB2 from that of other interacting proteins in promoting HR, while suppressing mutagenic SSA.

Essential revisions:

1) Novelty concerns and use of screening assays with appropriate citations

This manuscript presents a comprehensive mutation analysis, and so it is good to include mutations that have been analyzed previously. However, it is difficult to parse what information is new from that which is confirmatory of previous reports. The authors should therefore indicate which VUS have been analyzed before in functional assays.

The HR/SSA assay system is not entirely novel. Previously, Ransburgh et al. (Cancer Research 2010) have already measured the effects of BRCA1 sequence variants on HR in human cancer cells using siRNAi of endogenous BRCA1 and complementation by transiently expressed cDNA. Although the integration of the SSA reporter assay is novel, Ransburgh et al. should be cited to acknowledge the fact that they developed and used a similar assay system.

Increased SSA (and MMEJ) has recently been shown in lymphoblastoid cells from heterozygous PALB2 mutation carriers (Obermeier et al., Oncogene 2016).

Also the analysis of the effects of a series of BRCA1 deletion mutants on HR, SSA and the response to olaparib or cisplatin is an extension of previous work (Ransburgh et al., Cancer Research 2010).

The data in Figure 4 confirm previous findings by Rodriguez et al. (Exp Cell Res 2004) and nicely show that the defect of BRCT mutants in HR is not primarily caused by mislocalization of overexpressed mutant BRCA1. It is interesting to speculate why BRCT mutants do not seem to enter the nucleus efficiently. However, as this has been published before, Figure 4 could be moved to supplemental data.

2) Specificity and sensitivity of the assays

Functional analysis of BRCA1 VUS is important to aid clinical decision-making. RNAi of endogenous BRCA1 and transient overexpression of cDNA constructs may not capture all aspects of BRCA1 biology, but when the specificity and sensitivity of the assays are high, the functional data are relevant for genetic counseling. In addition to the controls mentioned by the authors, there are a number of variants known to be pathogenic or neutral (the BIC database needs to be updated). The ENIGMA consortium has published lists of these variants in Millot et al. (Hum Mut 2012). The authors should indicate all known controls and use these to calculate the specificity and sensitivity of their assays.

3) Truncated BRCA1 in MDA-MB 436 cells

The authors used MDA-MB 436 cells to analyze the abilities of BRCA1 variants to confer cisplatin or olaparib resistance. It should be noted that MDA-MB 436 cells have a BRCT truncating mutation and express a near full-length BRCA1 protein that can be stabilized and confer cisplatin and PARPi resistance (Johnson et al., PNAS 2013). It is therefore not a BRCA1-null cell line and results have to be evaluated with caution. The authors should include a paragraph on the limitations of their assays systems in the discussion, and cite the PNAS paper from Johnson et al.

4) BRCA1 and PALPB2 protein expression levels

The assays rely on artificial overexpression of BRCA1 variants, which may affect the results. How does the expressed BRCA1 protein level compare with the endogenous level? Preferably, BRCA1 needs to be expressed at equal and physiologically relevant levels in the relevant cell lines and assays. If it is heavily overexpressed it could be swamping the system, e.g. diluting interacting proteins, which does become relevant later in the paper. To address this issue, the authors should show expression levels of endogenous and overexpressed BRCA1 proteins.

How does the expressed PALB2 protein level compare with the endogenous level (Figure 2D)?

5) Variable knockdown efficiencies and potential off-target effects of siRNAs

For the assays in U2OS cells, it is essential that there is efficient knockdown of endogenous BRCA1 in all cases. Knockdown efficiency should therefore be shown.

Figure 2B shows that knockdown of BRCA2 with 2 different siRNAs gives very different quantitative results: a 30-fold increase in SSA for siRNA 1949 and an 8-fold increase for the 9025. Even the PALB2 siRNAs give very high SSA increases. RAD51 protein levels are frequently affected by siRNAs (as shown by Elledge and colleagues), so it is important to show that RAD51 levels are not affected by the various siRNAs, especially 1949, which appears to be especially effective at increasing SSA while not being demonstrably better at knocking down BRCA2. Related to this, the less effective siRNA for PALB2 seems to be more effective at increasing SSA.

Although PALB2 knockdown results in more than 10-fold increase in SSA activity (Figure 2B, C), re-expression of PALB2 only suppresses SSA less than 50% (Figure 2F). What is the reason for this discrepancy?

6) Statistical analysis

At present, it is unclear which of the observed differences are significant. Statistical analysis is lacking as are source data files and both should be submitted.

eLife. 2017 Apr 11;6:e21350. doi: 10.7554/eLife.21350.019

Author response

Essential revisions:

1) Novelty concerns and use of screening assays with appropriate citations

We acknowledge the reviewers’ concerns. We have added a Table 2 to include functional results of previous studies on any of the VUSs that are analyzed in our study with respect to HR activity and drug sensitivity. Table 1 has also been updated incorporating information from ClinVar and other databases.

We apologize for not including Ransburgh et al. in our citation but we did cite Towler et al. (2013) that utilized a HR/SSA system developed by the same group. We have now added a citation for Ransburgh et al. as well. As far as we know, we might have been the first in the BRCA field to employ such transient protein replacement strategy to measure HR activity (Xia, et al., Mol Cell, 2006), although it was used to measure PALB2 activity, and Junjie Chen’s group appears to be the first to use the strategy for BRCA1 variants (Sy et al., PNAS, 2009). Also, we now use a newly designed siRNA targeting the 3’-UTR of BRCA1, and the protocol used in the current study is also different. Still, to avoid potential and unnecessary conflicts, we have changed the wording in the abstract from “developed” to “employed”.

Increased SSA (and MMEJ) has recently been shown in lymphoblastoid cells from heterozygous PALB2 mutation carriers (Obermeier et al., Oncogene 2016).

Thanks for pointing this out. We are aware of the report, although it appeared to us that only some of the cell lines showed increased SSA. We have now added a mention and a citation acknowledging this study in our revised manuscript. At the same time, we would like to note the much more profound increase in SSA following PALB2 knockdown and the emphasis of our study on the BRCA1-PALB2 and PALB2-BRCA2 interactions and complex formations in suppressing SSA.

Also the analysis of the effects of a series of BRCA1 deletion mutants on HR, SSA and the response to olaparib or cisplatin is an extension of previous work (Ransburgh et al., Cancer Research 2010).

While Ransburgh et al. did generate 4 large BRCA1 deletion mutants and measured their effects on HR, the study did not measure any effects of the deletions on SSA nor drug sensitivity. Our study provides a detailed mapping of regions of BRCA1 involved in HR, SSA and therapy response. As expected, deletion of the N-terminal RING domain, the PALB2-binding coiled-coil domain and the C-terminal BRCT domain, sensitizes cells to both cisplatin and olaparib. Moreover, our series also includes Δ486-627, which contains the NLS sequence of BRCA1 but is unable to effectively support drug resistance despite retaining ~70% HR activity, as well as Δ1056-1307 with substantially reduced HR activity despite the region having no known domains or interacting proteins identified to date. More interestingly, deletion of amino acid 887-1061 does not significantly affect HR or SSA, yet it leads to a greatly enhanced olaparib resistance. We believe that the findings from our deletion series will spawn new ideas and thinking that may significantly advance the BRCA1 field in the future.

We understand the concern. However, we still wish to keep the figure as a formal one due to the following reasons. First, all mutations studied in the Rodriguez paper are BRCT truncating mutations, whereas the present manuscript is focused on point mutations, which in our view are different. A strong indication is that several point mutants (A1708E and M1775R, etc.) appears to be grossly destabilized, whereas truncating mutants, such as 5055delG, are not (although frameshift mutations often lead to nonsense mediated decay (NMD) of mRNA in vivo). Also, the missense mutants may still bind known interaction partners with low affinity or even certain unknown factors that may affect their localization. Second, our Figure 4A includes not only the BRCT mutants but also mutants that affect BARD1 and PALB2 binding (C61G and M1411T, respectively); therefore, the figure serves a broader purpose and emphasizes the comparison between the different class of mutants. Third, Rodriguez et al. did not test whether BARD1 can rescue the HR defect of the BRCT mutants. Fourth, we believe we have already properly acknowledged their findings in our manuscript.

2) Specificity and sensitivity of the assays

Thanks for raising this issue and providing valuable advice. We have assessed the sensitivity and specificity of our HR and drug sensitivity assays using the Receiver Operating Characteristics (ROC) analysis to simultaneously capture the maximum sensitivity and specificity of the assays with 95% confidence interval as previously described (Guidugli et al., 2013), using all known controls as suggested. The result confirms that our assays are both highly sensitive and highly specific to distinguish between neutral and pathogenic variants. Threshold lines have added to the figures and the Results and Methods sections have been updated as well.

3) Truncated BRCA1 in MDA-MB 436 cells

Work by Johnson et al. is indeed helpful in the understanding of drug resistance mechanisms of BRCA1-deficient cells, and this citation has been added. At the same time, we would like to note the major differences between the two studies. First, the BRCA1 stabilization and PARPi resistance observed by Johnson et al. was after repeated selection of MDA-MB-436 cells with rucaparib over 4-6 months, whereas in our case we used only one round of selection over 3-4 weeks. Moreover, we used double selection with both G418 (to select for transfected cells) and olaparib or cisplatin (for repair function). Under the condition used, we never had a single olaparib-resistant colony in vector transfected cells or untransfected cells, which allow us to rule out any contribution of endogenous BRCA1 to our readout. On one occasion we did get 4 cisplatin-resistant colonies from vector transfected cells (as compared with roughly 100 colonies we got from wt BRCA1 cDNA), which we think could be due to the effective dose of cisplatin in the particular plate being slightly lower than the threshold, as cisplatin is not a very stable compound. Thus, we believe that MDA-MB-436 is a suitable cell line for BRCA1 functional analysis under the condition used. A brief discussion of these has been added.

4) BRCA1 and PALPB2 protein expression levels

Thanks for the suggestion. We have added a small panel showing the expression levels of endogenous and exogenous BRCA1 in the HR reporter cells (new Figure 1E). The averaged expression level of BRCA1 in the well of cells first depleted of the endogenous protein and then transfected with wt BRCA1 cDNA was typically 3-5 times that of the endogenous level. However, the real situation is more complex, as all the cells were not transfected and the cells that were transfected would express BRCA1 at different levels, as can be seen in Figure 4A. Based on our experience gained from the 6-plus years working on this project, we would say that some of transfected cells express BRCA1 very highly and these cells have a large contribution to the averaged BRCA1 level; however, there are always a few percent of (the total population of) cells that express BRCA1 at levels comparable to the endogenous one, and these are likely the cells that turn green. With these said, more important is that all VUSs but the BRCT point mutants showed expression levels similar to that of the wt protein as assayed by western, and all but the RING and BRCT mutants showed similar expression patterns when analyzed by immunofluorescence. Therefore, despite the overexpression of exogenous BRCA1, we believe that the results of our analyses are reliable. A brief mention of the above has been added to the Results section. The expression levels of the BRCT mutants are now shown in Figure 1—figure supplement 2.

How does the expressed PALB2 protein level compare with the endogenous level (Figure 2D)?

We have added a panel of western blots to address this concern (new Figure 2H). The apparent overall level of exogenous PALB2 expression is about 2-5 times that of the endogenous level, and the same caveat as noted above also applies here. For reasons unclear, the BRCA1 binding mutant L21A showed substantially higher expression than the wt protein (yet it still failed to support HR and suppress SSA).

5) Variable knockdown efficiencies and potential off-target effects of siRNAs

For the assays in U2OS cells, it is essential that there is efficient knockdown of endogenous BRCA1 in all cases. Knockdown efficiency should therefore be shown.

The knockdown efficiencies of BRCA1 and PALB2 are now shown in the above-noted new panels addressing the expression levels of endogenous and exogenous BRCA1 and PALB2 (Figure 1E and 2H, respectively). Please note that for our HR and SSA assays involving “protein replacement”, the knockdown was conducted in 10 cm dishes and the cells were then mixed and reseeded into 6-well plates for the second transfection, therefore the wells have uniform levels of knockdown.

Thanks for raising this issue. We have now repeated the SSA experiment using 4 more (6 total) BRCA2 siRNAs (updated Figure 2C). Out of the 6 BRCA2 siRNAs, 5 of them increased SSA by 2-3 fold, whereas siRNA 1949 remained an outlier, elevating SSA by about 5 fold. The 2 PALB2 siRNAs behaved more similarly, stimulating SSA by about 6 fold. Our old results were obtained in 2013-2014 and the much larger increases in SSA following BRCA2 and PALB2 knockdown were reproducible at the time (at least 3 independent experiments). Although we cannot explain exactly why our new results show substantially less dramatic increase, we did notice time and times again that the performance of the reporter cell line, in terms of the number of green cells yielded, tend to deteriorate over time and can fluctuate greatly between consecutive passages. We also noticed that BRCA2 in the new stock of cells used for revision is not degraded as quickly upon PALB2 loss as in the old cells. The exact mechanism as to how BRCA2 and PALB2 suppress SSA awaits further investigation.

The western blotting results in the original Figure 2B have been replaced with new results that now include RAD51 (new Figure 2C). All 6 BRCA2 siRNAs were found to reduce RAD51 protein abundance; however, there was no correlation between the reductions in RAD51 amount with the increases in SSA activity. Moreover, siRNA 1949 was among the ones that had the mildest effect on RAD51 level. Therefore, the reduction in RAD51 amount is unlikely to be a major cause of increased SSA. Given that one molecule of BRCA2 can bind up to 6 molecules of RAD51 (Jensen et al., 2010), the reduction of RAD51 could be explained by the co-degradation of RAD51 with their carrier BRCA2 during the siRNA treatment period; or, it could be caused by an indirect effect resulting from altered cell cycle due to loss of BRCA2 combined with potential off-target effect on other genes. In the case of PALB2 siRNAs, neither of the 2 used significantly reduced RAD51 amount; in fact, RAD51 amount even increased slightly in the case of siRNA 1493, which caused a 7 fold increase in SSA. It is therefore safe to say that changes in RAD51 abundance has limited, if any, impact on SSA in the case of PALB2 knockdown.

As to the last point, we have used the 2 PALB2 siRNAs in multiple cell lines and numerous experiments, and their knockdown efficiencies were always similar. In fact, the levels of PALB2 knockdown were also similar in our original figure (the difference was in BRCA2, for reasons unknown). Due to the use of more BRCA2 siRNAs as described above, the western blots in the original figure have been replaced with new ones, in which BRCA2 amounts were similar, as mentioned above.

This is indeed puzzling. We have repeated the experiment several more times and the suppression was even slightly less than before (but still highly significant) (new Figure 2G). In our protocol, PALB2 siRNA was first used to deplete the endogenous protein, cDNAs for PALB2 and I-SceI were delivered 48 hr later, and cells were harvested and green cells counted another ~52 hr later. It is conceivable that the depletion of PALB2 not only caused loss of the PALB2 itself but also led to certain secondary effects as a result of PALB2 loss, such as potential changes in chromatin state, that cannot not be immediately restored upon sudden re-expression of PALB2. It is also possible that in order for PALB2 to function in SSA suppression, it needs to be post-translationally modified or form functional complexes with other proteins, which cannot be fully achieved immediately after re-expression.

6) Statistical analysis

At present, it is unclear which of the observed differences are significant. Statistical analysis is lacking as are source data files and both should be submitted.

We have added p values for relevant parts of Figure 2 and Figure 5. For the larger sets of HR and drug resistance data, providing a p value to each variant is unnecessary or even counterproductive, as some of the small differences will be significant due to the number of repeats we have conducted, even though the difference could in fact be simply due to a slight variation of plasmid quality or other technical issues. For these data, a cutoff threshold is provided, with p values added only to M1400V. Source data are now submitted.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure 1—source data 1. HR activities of the BRCA1 variants and mutants analzyed in this study.

See texts and materials and methods for details.

DOI: http://dx.doi.org/10.7554/eLife.21350.004

elife-21350-fig1-data1.xlsx^{(23.7KB, xlsx)}

DOI: 10.7554/eLife.21350.004

Figure 2—source data 1. SSA activities of the BRCA1 variants and mutants analyzed in this study.

See texts and materials and methods for details.

DOI: http://dx.doi.org/10.7554/eLife.21350.010

elife-21350-fig2-data1.xlsx^{(23.1KB, xlsx)}

DOI: 10.7554/eLife.21350.010

Figure 3—source data 1. Abilities of the BRCA1 variants and mutants to support olaparib and cisplatin resistance in MDA-MB-436 cells.

See texts and materials and methods for details.

DOI: http://dx.doi.org/10.7554/eLife.21350.012

elife-21350-fig3-data1.xlsx^{(55.7KB, xlsx)}

DOI: 10.7554/eLife.21350.012

Supplementary file 1. Sequences of siRNAs used in this study.

DOI: http://dx.doi.org/10.7554/eLife.21350.017

elife-21350-supp1.docx^{(13.5KB, docx)}

DOI: 10.7554/eLife.21350.017

[bib1] Bouwman P, van der Gulden H, van der Heijden I, Drost R, Klijn CN, Prasetyanti P, Pieterse M, Wientjens E, Seibler J, Hogervorst FB, Jonkers J. A high-throughput functional complementation assay for classification of BRCA1 missense variants. Cancer Discovery. 2013;3:1142–1155. doi: 10.1158/2159-8290.CD-13-0094. [DOI] [PubMed] [Google Scholar]

[bib2] Brzovic PS, Keeffe JR, Nishikawa H, Miyamoto K, Fox D, Fukuda M, Ohta T, Klevit R. Binding and recognition in the assembly of an active BRCA1/BARD1 ubiquitin-ligase complex. PNAS. 2003;100:5646–5651. doi: 10.1073/pnas.0836054100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] Bunting SF, Callén E, Wong N, Chen HT, Polato F, Gunn A, Bothmer A, Feldhahn N, Fernandez-Capetillo O, Cao L, Xu X, Deng CX, Finkel T, Nussenzweig M, Stark JM, Nussenzweig A. 53bp1 inhibits homologous recombination in Brca1-deficient cells by blocking resection of DNA breaks. Cell. 2010;141:243–254. doi: 10.1016/j.cell.2010.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] Cantor SB, Bell DW, Ganesan S, Kass EM, Drapkin R, Grossman S, Wahrer DC, Sgroi DC, Lane WS, Haber DA, Livingston DM. BACH1, a novel helicase-like protein, interacts directly with BRCA1 and contributes to its DNA repair function. Cell. 2001;105:149–160. doi: 10.1016/S0092-8674(01)00304-X. [DOI] [PubMed] [Google Scholar]

[bib5] Chang S, Biswas K, Martin BK, Stauffer S, Sharan SK. Expression of human BRCA1 variants in mouse ES cells allows functional analysis of BRCA1 mutations. Journal of Clinical Investigation. 2009;119:3160–3171. doi: 10.1172/JCI39836. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] Chen CF, Li S, Chen Y, Chen PL, Sharp ZD, Lee WH. The nuclear localization sequences of the BRCA1 protein interact with the importin-alpha subunit of the nuclear transport signal receptor. Journal of Biological Chemistry. 1996;271:32863–32868. doi: 10.1074/jbc.271.51.32863. [DOI] [PubMed] [Google Scholar]

[bib7] Chen J, Silver DP, Walpita D, Cantor SB, Gazdar AF, Tomlinson G, Couch FJ, Weber BL, Ashley T, Livingston DM, Scully R. Stable interaction between the products of the BRCA1 and BRCA2 tumor suppressor genes in mitotic and meiotic cells. Molecular Cell. 1998;2:317–328. doi: 10.1016/S1097-2765(00)80276-2. [DOI] [PubMed] [Google Scholar]

[bib8] Couch FJ, Nathanson KL, Offit K. Two decades after BRCA: setting paradigms in personalized Cancer care and prevention. Science. 2014;343:1466–1470. doi: 10.1126/science.1251827. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] Deng CX. BRCA1: cell cycle checkpoint, genetic instability, DNA damage response and Cancer evolution. Nucleic Acids Research. 2006;34:1416–1426. doi: 10.1093/nar/gkl010. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] Densham RM, Garvin AJ, Stone HR, Strachan J, Baldock RA, Daza-Martin M, Fletcher A, Blair-Reid S, Beesley J, Johal B, Pearl LH, Neely R, Keep NH, Watts FZ, Morris JR. Human BRCA1-BARD1 ubiquitin ligase activity counteracts chromatin barriers to DNA resection. Nature Structural & Molecular Biology. 2016;23:647–655. doi: 10.1038/nsmb.3236. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] Drost R, Bouwman P, Rottenberg S, Boon U, Schut E, Klarenbeek S, Klijn C, van der Heijden I, van der Gulden H, Wientjens E, Pieterse M, Catteau A, Green P, Solomon E, Morris JR, Jonkers J. BRCA1 RING function is essential for tumor suppression but dispensable for therapy resistance. Cancer Cell. 2011;20:797–809. doi: 10.1016/j.ccr.2011.11.014. [DOI] [PubMed] [Google Scholar]

[bib12] Elstrodt F, Hollestelle A, Nagel JH, Gorin M, Wasielewski M, van den Ouweland A, Merajver SD, Ethier SP, Schutte M. BRCA1 mutation analysis of 41 human breast Cancer cell lines reveals three new deleterious mutants. Cancer Research. 2006;66:41–45. doi: 10.1158/0008-5472.CAN-05-2853. [DOI] [PubMed] [Google Scholar]

[bib13] Fabbro M, Rodriguez JA, Baer R, Henderson BR. BARD1 induces BRCA1 intranuclear foci formation by increasing RING-dependent BRCA1 nuclear import and inhibiting BRCA1 nuclear export. Journal of Biological Chemistry. 2002;277:21315–21324. doi: 10.1074/jbc.M200769200. [DOI] [PubMed] [Google Scholar]

[bib14] Fackenthal JD, Olopade OI. Breast Cancer risk associated with BRCA1 and BRCA2 in diverse populations. Nature Reviews Cancer. 2007;7:937–948. doi: 10.1038/nrc2054. [DOI] [PubMed] [Google Scholar]

[bib15] Foo TK, Tischkowitz M, Simhadri S, Boshari T, Zayed N, Burke KA, Berman SH, Blecua P, Riaz N, Huo Y, Ding YC, Neuhausen SL, Weigelt B, Reis-Filho JS, Foulkes WD, Xia B. Compromised BRCA1-PALB2 interaction is associated with breast Cancer risk. Oncogene. 2017 doi: 10.1038/onc.2017.46. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] Greenberg RA, Sobhian B, Pathania S, Cantor SB, Nakatani Y, Livingston DM. Multifactorial contributions to an acute DNA damage response by BRCA1/BARD1-containing complexes. Genes & Development. 2006;20:34–46. doi: 10.1101/gad.1381306. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] Guidugli L, Pankratz VS, Singh N, Thompson J, Erding CA, Engel C, Schmutzler R, Domchek S, Nathanson K, Radice P, Singer C, Tonin PN, Lindor NM, Goldgar DE, Couch FJ. A classification model for BRCA2 DNA binding domain missense variants based on homology-directed repair activity. Cancer Research. 2013;73:265–275. doi: 10.1158/0008-5472.CAN-12-2081. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] Gunn A, Stark JM. I-SceI-based assays to examine distinct repair outcomes of mammalian chromosomal double strand breaks. Methods in Molecular Biology. 2012;920:379–391. doi: 10.1007/978-1-61779-998-3_27. [DOI] [PubMed] [Google Scholar]

[bib19] Hashizume R, Fukuda M, Maeda I, Nishikawa H, Oyake D, Yabuki Y, Ogata H, Ohta T. The RING heterodimer BRCA1-BARD1 is a ubiquitin ligase inactivated by a breast cancer-derived mutation. Journal of Biological Chemistry. 2001;276:14537–14540. doi: 10.1074/jbc.C000881200. [DOI] [PubMed] [Google Scholar]

[bib20] Huen MS, Sy SM, Chen J. BRCA1 and its toolbox for the maintenance of genome integrity. Nature Reviews Molecular Cell Biology. 2010;11:138–148. doi: 10.1038/nrm2831. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] Jensen RB, Carreira A, Kowalczykowski SC. Purified human BRCA2 stimulates RAD51-mediated recombination. Nature. 2010;467:678–683. doi: 10.1038/nature09399. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] Jiang Q, Greenberg RA. Deciphering the BRCA1 tumor Suppressor Network. Journal of Biological Chemistry. 2015;290:17724–17732. doi: 10.1074/jbc.R115.667931. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] Li ML, Greenberg RA. Links between genome integrity and BRCA1 tumor suppression. Trends in Biochemical Sciences. 2012;37:418–424. doi: 10.1016/j.tibs.2012.06.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] Liu J, Doty T, Gibson B, Heyer WD. Human BRCA2 protein promotes RAD51 filament formation on RPA-covered single-stranded DNA. Nature Structural & Molecular Biology. 2010;17:1260–1262. doi: 10.1038/nsmb.1904. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] Lord CJ, Ashworth A. Mechanisms of resistance to therapies targeting BRCA-mutant cancers. Nature Medicine. 2013;19:1381–1388. doi: 10.1038/nm.3369. [DOI] [PubMed] [Google Scholar]

[bib26] Lord CJ, Ashworth A. BRCAness revisited. Nature Reviews Cancer. 2016;16:110–120. doi: 10.1038/nrc.2015.21. [DOI] [PubMed] [Google Scholar]

[bib27] Lu C, Xie M, Wendl MC, Wang J, McLellan MD, Leiserson MD, Huang KL, Wyczalkowski MA, Jayasinghe R, Banerjee T, Ning J, Tripathi P, Zhang Q, Niu B, Ye K, Schmidt HK, Fulton RS, McMichael JF, Batra P, Kandoth C, Bharadwaj M, Koboldt DC, Miller CA, Kanchi KL, Eldred JM, Larson DE, Welch JS, You M, Ozenberger BA, Govindan R, Walter MJ, Ellis MJ, Mardis ER, Graubert TA, Dipersio JF, Ley TJ, Wilson RK, Goodfellow PJ, Raphael BJ, Chen F, Johnson KJ, Parvin JD, Ding L. Patterns and functional implications of rare germline variants across 12 cancer types. Nature Communications. 2015;6:10086. doi: 10.1038/ncomms10086. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] Ma J, Cai H, Wu T, Sobhian B, Huo Y, Alcivar A, Mehta M, Cheung KL, Ganesan S, Kong AN, Zhang DD, Xia B. PALB2 interacts with KEAP1 to promote NRF2 nuclear accumulation and function. Molecular and Cellular Biology. 2012;32:1506–1517. doi: 10.1128/MCB.06271-11. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] Millot GA, Carvalho MA, Caputo SM, Vreeswijk MP, Brown MA, Webb M, Rouleau E, Neuhausen SL, Hansen T, Galli A, Brandão RD, Blok MJ, Velkova A, Couch FJ, Monteiro AN, ENIGMA Consortium Functional Assay Working Group A guide for functional analysis of BRCA1 variants of uncertain significance. Human Mutation. 2012;33:1526–1537. doi: 10.1002/humu.22150. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] Morris JR, Boutell C, Keppler M, Densham R, Weekes D, Alamshah A, Butler L, Galanty Y, Pangon L, Kiuchi T, Ng T, Solomon E. The SUMO modification pathway is involved in the BRCA1 response to genotoxic stress. Nature. 2009;462:886–890. doi: 10.1038/nature08593. [DOI] [PubMed] [Google Scholar]

[bib31] Mortensen UH, Bendixen C, Sunjevaric I, Rothstein R. DNA strand annealing is promoted by the yeast Rad52 protein. PNAS. 1996;93:10729–10734. doi: 10.1073/pnas.93.20.10729. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] Moynahan ME, Jasin M. Mitotic homologous recombination maintains genomic stability and suppresses tumorigenesis. Nature Reviews Molecular Cell Biology. 2010;11:196–207. doi: 10.1038/nrm2851. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] Mullan PB, Quinn JE, Harkin DP. The role of BRCA1 in transcriptional regulation and cell cycle control. Oncogene. 2006;25:5854–5863. doi: 10.1038/sj.onc.1209872. [DOI] [PubMed] [Google Scholar]

[bib34] Nakanishi K, Yang YG, Pierce AJ, Taniguchi T, Digweed M, D'Andrea AD, Wang ZQ, Jasin M. Human fanconi Anemia monoubiquitination pathway promotes homologous DNA repair. PNAS. 2005;102:1110–1115. doi: 10.1073/pnas.0407796102. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] Obermeier K, Sachsenweger J, Friedl TW, Pospiech H, Winqvist R, Wiesmüller L. Heterozygous PALB2 c.1592delt mutation channels DNA double-strand break repair into error-prone pathways in breast Cancer patients. Oncogene. 2016;35:3796–3806. doi: 10.1038/onc.2015.448. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib36] Oliver AW, Swift S, Lord CJ, Ashworth A, Pearl LH. Structural basis for recruitment of BRCA2 by PALB2. EMBO Reports. 2009;10:990–996. doi: 10.1038/embor.2009.126. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib37] Polato F, Callen E, Wong N, Faryabi R, Bunting S, Chen HT, Kozak M, Kruhlak MJ, Reczek CR, Lee WH, Ludwig T, Baer R, Feigenbaum L, Jackson S, Nussenzweig A. CtIP-mediated resection is essential for viability and can operate independently of BRCA1. The Journal of Experimental Medicine. 2014;211:1027–1036. doi: 10.1084/jem.20131939. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] Ransburgh DJ, Chiba N, Ishioka C, Toland AE, Parvin JD. Identification of breast tumor mutations in BRCA1 that abolish its function in homologous DNA recombination. Cancer Research. 2010;70:988–995. doi: 10.1158/0008-5472.CAN-09-2850. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] Reczek CR, Szabolcs M, Stark JM, Ludwig T, Baer R. The interaction between CtIP and BRCA1 is not essential for resection-mediated DNA repair or tumor suppression. The Journal of Cell Biology. 2013;201:693–707. doi: 10.1083/jcb.201302145. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] Reid LJ, Shakya R, Modi AP, Lokshin M, Cheng JT, Jasin M, Baer R, Ludwig T. E3 ligase activity of BRCA1 is not essential for mammalian cell viability or homology-directed repair of double-strand DNA breaks. PNAS. 2008;105:20876–20881. doi: 10.1073/pnas.0811203106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib41] Rodriguez JA, Au WW, Henderson BR. Cytoplasmic mislocalization of BRCA1 caused by cancer-associated mutations in the BRCT domain. Experimental Cell Research. 2004;293:14–21. doi: 10.1016/j.yexcr.2003.09.027. [DOI] [PubMed] [Google Scholar]

[bib42] Rodríguez JA, Henderson BR. Identification of a functional nuclear export sequence in BRCA1. Journal of Biological Chemistry. 2000;275:38589–38596. doi: 10.1074/jbc.M003851200. [DOI] [PubMed] [Google Scholar]

[bib43] Roy R, Chun J, Powell SN. BRCA1 and BRCA2: different roles in a common pathway of genome protection. Nature Reviews Cancer. 2012;12:68–78. doi: 10.1038/nrc3181. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib44] Ruffner H, Joazeiro CA, Hemmati D, Hunter T, Verma IM. Cancer-predisposing mutations within the RING domain of BRCA1: loss of ubiquitin protein ligase activity and protection from radiation hypersensitivity. PNAS. 2001;98:5134–5139. doi: 10.1073/pnas.081068398. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] Sartori AA, Lukas C, Coates J, Mistrik M, Fu S, Bartek J, Baer R, Lukas J, Jackson SP. Human CtIP promotes DNA end resection. Nature. 2007;450:509–514. doi: 10.1038/nature06337. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib46] Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J, Ashley T, Livingston DM. Association of BRCA1 with Rad51 in mitotic and meiotic cells. Cell. 1997;88:265–275. doi: 10.1016/S0092-8674(00)81847-4. [DOI] [PubMed] [Google Scholar]

[bib47] Shakya R, Reid LJ, Reczek CR, Cole F, Egli D, Lin CS, deRooij DG, Hirsch S, Ravi K, Hicks JB, Szabolcs M, Jasin M, Baer R, Ludwig T. BRCA1 tumor suppression depends on BRCT phosphoprotein binding, but not its E3 ligase activity. Science. 2011;334:525–528. doi: 10.1126/science.1209909. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] Shakya R, Szabolcs M, McCarthy E, Ospina E, Basso K, Nandula S, Murty V, Baer R, Ludwig T. The basal-like mammary carcinomas induced by Brca1 or Bard1 inactivation implicate the BRCA1/BARD1 heterodimer in tumor suppression. PNAS. 2008;105:7040–7045. doi: 10.1073/pnas.0711032105. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] Stark JM, Pierce AJ, Oh J, Pastink A, Jasin M. Genetic steps of mammalian homologous repair with distinct mutagenic consequences. Molecular and Cellular Biology. 2004;24:9305–9316. doi: 10.1128/MCB.24.21.9305-9316.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib50] Sy SM, Huen MS, Chen J. PALB2 is an integral component of the BRCA complex required for homologous recombination repair. PNAS. 2009;106:7155–7160. doi: 10.1073/pnas.0811159106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] Towler WI, Zhang J, Ransburgh DJ, Toland AE, Ishioka C, Chiba N, Parvin JD. Analysis of BRCA1 variants in double-strand break repair by homologous recombination and single-strand annealing. Human Mutation. 2013;34:439–445. doi: 10.1002/humu.22251. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib52] Tutt A, Bertwistle D, Valentine J, Gabriel A, Swift S, Ross G, Griffin C, Thacker J, Ashworth A. Mutation in Brca2 stimulates error-prone homology-directed repair of DNA double-strand breaks occurring between repeated sequences. The EMBO Journal. 2001;20:4704–4716. doi: 10.1093/emboj/20.17.4704. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib53] Venkitaraman AR. Cancer suppression by the chromosome custodians, BRCA1 and BRCA2. Science. 2014;343:1470–1475. doi: 10.1126/science.1252230. [DOI] [PubMed] [Google Scholar]

[bib54] Wang B, Matsuoka S, Ballif BA, Zhang D, Smogorzewska A, Gygi SP, Elledge SJ. Abraxas and RAP80 form a BRCA1 protein complex required for the DNA damage response. Science. 2007;316:1194–1198. doi: 10.1126/science.1139476. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib55] Wu LC, Wang ZW, Tsan JT, Spillman MA, Phung A, Xu XL, Yang MC, Hwang LY, Bowcock AM, Baer R. Identification of a RING protein that can interact in vivo with the BRCA1 gene product. Nature Genetics. 1996;14:430–440. doi: 10.1038/ng1296-430. [DOI] [PubMed] [Google Scholar]

[bib56] Xia B, Sheng Q, Nakanishi K, Ohashi A, Wu J, Christ N, Liu X, Jasin M, Couch FJ, Livingston DM. Control of BRCA2 cellular and clinical functions by a nuclear partner, PALB2. Molecular Cell. 2006;22:719–729. doi: 10.1016/j.molcel.2006.05.022. [DOI] [PubMed] [Google Scholar]

[bib57] Zhang F, Fan Q, Ren K, Andreassen PR. PALB2 functionally connects the breast Cancer susceptibility proteins BRCA1 and BRCA2. Molecular Cancer Research. 2009a;7:1110–1118. doi: 10.1158/1541-7786.MCR-09-0123. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] Zhang F, Ma J, Wu J, Ye L, Cai H, Xia B, Yu X. PALB2 links BRCA1 and BRCA2 in the DNA-damage response. Current Biology. 2009b;19:524–529. doi: 10.1016/j.cub.2009.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Functional and mutational landscapes of BRCA1 for homology-directed repair and therapy resistance

Rachel W Anantha

Srilatha Simhadri

Tzeh Keong Foo

Susanna Miao

Jingmei Liu

Zhiyuan Shen

Shridar Ganesan

Bing Xia

Roles

Abstract

eLife digest

Introduction

Figure 1. Sequence alterations generated in BRCA1 and their effects on protein-protein interactions and HR.

Figure 1—figure supplement 1. Capacity of the BRCA1 variants to bind BARD1, PALB2, CtIP and BRIP1.

Figure 1—figure supplement 2. Expression levels of BRCA1 BRCT missense and truncating mutants in U2OS/DR-GFP cells first depleted of the endogenous BRCA1 and then transfected with cDNA expressing the mutants.

Results

The HR activity of BRCA1 requires intact RING, CC and BRCT domains

Table 1.

Table 2.

The BRCA1-PALB2-BRCA2 interactions prevent RAD52-dependent SSA

Figure 2. The BRCA1-PALB2 interaction suppresses SSA.

Effects of BRCA1 missense variants on PARPi and cisplatin resistance

Figure 3. Abilities of the wt and mutant or variant BRCA1 proteins to confer cisplatin and olaparib resistance.

Localization defect of BRCT missense mutants

Figure 4. Rescue of the localization defect but not HR activity of BRCA1 BRCT missense mutants by BARD1.

Role of the BRCA1 I26A mutation and BRCA1 sumoylation in HDR and drug resistance

Figure 5. Effects of BRCA1 sumoylation and E3 ligase activity on HR and drug resistance.

Deletion analyses of BRCA1 for HDR and drug resistance

Figure 6. Functionalities of BRCA1 deletion mutants in HR, SSA and drug resistance.

Discussion

Figure 7. Summary models of the regulation and function of BRCA1 in HR and SSA.

Materials and methods

Cell culture

Plasmids and mutagenesis

cDNA transfection, immunoprecipitation (IP) and western blotting

Measurement of HR and SSA efficiency

Immunofluorescence

Assay for PARPi and cisplatin resistance

Sensitivity and specificity analyses

Statistical analysis

Acknowledgements

Funding Statement

Funding Information

Additional information

Competing interests

Author contributions

Additional files

References

Decision letter

Roles

Author response

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases