Bioinformatic analysis of variant sequences

To predict the effect of each variant on mRNA splicing we utilized Human Splicing Finder version 2.4 (www.umd.be/HSF) (Desmet, et al., 2009) which calculates consensus values of potential variant affected mRNA splice sites and splice regulatory sites using different algorithms, including Human Splicing Finder, MaxEnt (Yeo and Burge, 2004), ESE Rescue (Fairbrother, et al., 2002) and ESE Finder (Cartegni, et al., 2003).

Identification of splicing aberrations

For assays conducted in Brisbane, culture of LCLs were conducted with and without cycloheximide. Treated LCLs were grown in the presence of cycloheximide (100 mg/ml) for 4 hours in order to stabilise mutant mRNA (Bateman, et al., 1999). RNA was extracted from cycloheximide treated and untreated cell lines using the RNeasy Mini Kit (Qiagen, Doncaster, Victoria, Australia), according to the manufacturer’s instructions. Each RNA sample was treated with DNase to reduce DNA contamination using DNA-free Kit (Ambion, Austin, TX, USA). cDNA was synthesised using Superscript III First Strand Synthesis System (Invitrogen, Carlsbad, CA, USA). PCR was performed using Amplitaq Gold (Applied Biosystems, Scoresby, Victoria, Australia) under the following conditions: 95°C for 7 minutes followed by 35 cycles of 94°C for 30 seconds, 55°C for 30 seconds and 72°C for 1 minute and a final extension step at 72°C for 7 minutes using primers listed in Supp. Table S1. The location of primers used are illustrated in Supp. Figure S1. PCR products were purified using QIAquick PCR Purification Kit (Qiagen), and sequenced using Big-Dye Terminator version 3.1 sequencing chemistry and the ABI 377 sequencer (Applied Biosystems, Australia). Aberrant splice products from BRCA2 c.8168A>G (p.Asp2723Gly) were cloned using pGEM®-T Easy Vector (Promega Corporation, Madison, WI, USA) according to the manufacturers instructions. Recombinant clones were PCR amplified from a single colony and sequenced as above. Aberrant splice products from BRCA1 c.4868C>G p.Ala1623Gly and BRCA2 c.7988A>T (p.Glu2663Val) were excised using QIAquick Gel Extraction Kit (Qiagen) and re-amplified using PCR under the conditions outlined above and sequenced. Fourteen non-variant carrying control LCLs were used for each RT-PCR. Reactions were performed in duplicate from two separate LCL cultures.

For the UK carriers of BRCA1 c.4868C>G p.Ala1623Gly, RNA was extracted from untransformed lymphocytes by the TRIzol method (Invitrogen) and cDNA was reverse transcribed by using random primers. RT-PCR was performed on the cDNA using PCR primers located in BRCA1 exon 15 and exon 17 (Supp. Table S1). Banding patterns were compared to that of 30 normal controls by agarose gel electrophoresis, and RT-PCR products of the two bands detected in the variant carrier were extracted, purified (Qiagen), and sequenced.

Quantitation of natural splice isoforms

Real-time PCR reactions were carried out in a Lightcycler 480 (Roche, Castle Hill, NSW, Australia) using Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen). Isoform specific primer sequences are listed in Supp. Table S1 and primer design is illustrated in Supp. Figure S1. The house keeping genes GAPDH and EEF1A1 were used as internal references for normalization. Each sample was run in quadruplicate. Normalized expression values were obtained using the Lightcycler 480 Gene Scanning software. Seven to nine non-variant carrying controls were used for cycloheximide treated and untreated samples respectively.

Quantitation of the variant and wildtype allele in c.7988A>T (p.Glu2663Val) full-length transcripts

Double stranded cDNA was synthesised for the restriction enzyme digestion assay of c.7988A>T (p.Glu2663Val) using the Illumina TotalPrep RNA amplification kit (Ambion), modified by omitting the final in vitro transcription step. cDNA was digested for 2 hours using 450 ng of cDNA at 65°C and 2 units of TaaI (New England Biolabs, Ipswich, UK) per 20 ul reaction. Digestion products were used as template cDNA for real-time PCR using conditions mentioned above and primers detailed in Supp. Table S1. Data were normalized to the housekeeping gene EEF1A1. To account for the possible amplification of background single stranded (ss)-cDNA, primers were designed to target a GAPDH sequence that is digested by TaaI in the ds-cDNA. Approximately 100-fold reduction in GAPDH levels was observed after TaaI digestion indicating that most of the ss-cDNA was converted to ds-cDNA, and thus ss-cDNA was unlikely to have contributed significantly to the level of full length BRCA2 isoform measured after TaaI digestion.

Identification of the variant and wildtype allele in BRCA1 c.4868C>G p.Ala1623Gly in full-length transcripts

The full-length band of the RT-PCR product was gel purified and the DNA was cloned into the TOPO vector (Invitrogen). After transformation and growing on a LB-amp plate, sixteen colonies were picked. Miniprep DNAs were amplified and sequenced.

Functional assays of missense substitutions

Assays of protein function were carried out using methods described previously (Farrugia et al., 2008). These are summarized in brief below.

Bioinformatic predictions and in vitro analysis of splicing aberrations

All variants were screened for potential splicing aberrations using Human Splicing Finder version 2.4 (www.umd.be/HSF, summarised in Table 1), and also assessed using in vitro splicing assays shown in Figure 2. Two of three variants predicted to create a de novo splice donor showed aberrant splicing consistent with bioinformatic predictions. BRCA1 c.4868C>G (p.Ala1623Gly) with MaxEnt scores of 6.29 for the variant splice site ctgGTGGGT, compared with 5.91 for the nearest donor consensus sequence, tttGTGAGT, resulted in a 119bp deletion from exon 16 (Figure 2A). The same findings were observed for BRCA1 c.4868C>G (p.Ala1623Gly) from the UK variant carriers (data not shown). BRCA2 c.8168A>G (p.Asp2723Gly) with MaxEnt scores of 8.56 for the variant site acaGTTGGG and 8.88 for the nearest donor consensus sequence, aagGTAAAT, and resulted in a deletion of 163 bp from exon 18 (Figure 2D). Both variants represent instances where splice sites otherwise unrecognizable by the splice machinery were elevated to a comparable or more likely splice point relative to the nearest site. Both deletions result in loss of the open-reading frame and creation of a stop codon, and were not seen in controls. Each of the seven variants was predicted to create or disrupt an ESE by one or more bioinformatic splice prediction tools, but a role for ESEs was indicated for only BRCA2 c.7988A>T (p.Glu2663Val). This substitution within exon 18 is predicted to disrupt an SRp55 site, and was associated with an elevated level of the two naturally occurring splice isoforms predicted to create a loss of the open reading frame; Δexon 18 (as previously reported by (Farrugia, et al., 2008)), and Δexon 17&18 (Figure 2E).

Open in a separate window

Figure 2

RT-PCR analysis of mRNA isolated from cycloheximide treated and untreated LCLs. In all figures, cyclohexamide treated samples are denoted ‘+’ and untreated samples are denoted ‘−’. Lanes showing variant carriers are underlined and all non-variant carrying controls are not underlined (A). Lanes 1&2 represent the BRCA1 c.4868C>G (p.Ala1623Gly) carrier and lanes 3&4 represent the c.5054C>T (p.Thr1685Ile) carrier. (B) Lanes 3&4 represent the BRCA2 c.8839G>A (p.Glu2947Lys) carrier and lanes 5&6 represent the BRCA2 c.9154C>T (p.Arg3052Trp) carrier. (C) Lanes 5–6 represent the BRCA2 c.7336A>G (p.Lys2446Glu) carrier (D) Lanes 3 and 4 represent the BRCA2 c.8168A>G (p.Asp2723Gly) carrier. (E) BRCA2 c.7988A>T (p.Glu2663Val). Lanes 1 and 2 represent the variant carrier. Alternative isoforms representing ΔExon 17&18 and ΔExon 18 are indicated in (D) and (E).

Stability and quantitation analysis of mRNA isoforms for BRCA2 c.7988A>T (p.Glu2663Val)

Cycloheximide was utilized to inhibit translation in vivo to ensure that the mRNA transcripts identified by the nonsense mediated RNA decay (NMD) pathway as truncating would be stabilized for detection by semi-quantitative RT-PCR. Comparison of cycloheximide and non-cycloheximide assays for this variant suggests that the alternative transcripts undergo minimal NMD (Figure 2E), and thus could encode truncated proteins. Real-time PCR was carried out to quantify the levels of Δexon 18 and Δexon 17&18 mRNA isoforms, relative to the full-length native transcript in the c.7988A>T (p.Glu2663Val) variant and controls. Cycloheximide treated cell lines were initially used to assess the relative contribution of each isoform to overall BRCA2 expression (Figure 3A–B). The c.7988A>T (p.Glu2663Val) carriers produced at least a 3–10 fold greater proportion of the Δexon 18 product relative to the full-length isoform when compared to controls not carrying this variant (Figure 3A). A higher level of the ΔExon 17&18 isoform relative to the full-length isoform was also observed in the c.7988A>T (p.Glu2663Val) carriers compared to all but one control (Figure 3B). These results raised the possibility that a higher quantity of the truncated forms of the protein would be produced by the variant carrier compared to individuals without the variant.

Open in a separate window

Figure 3

Quantitative real-time PCR of BRCA2 full-length, Δ18 and Δ17&18 splice isoforms. (A–B) Elevated transcript production of Δexon 18 and Δexon 17&18 relative to the full-length isoform seen in BRCA2 c.7988A>T (p.Glu2663Val) compared with controls not carrying the variant. In each case RNA was analysed from cycloheximide treated samples. (C) Total full-length isoform production for BRCA2 c.7988A>T (p.Glu2663Val) carriers and controls not carrying the variant in cycloheximide treated samples showing that the variant carrier produces the full-length isoform within a normal range. (D) The BRCA2 c.7988A>T (p.Glu2663Val) carrier produces a markedly elevated level of the Δ18 isoform when compared to controls. (E) One carrier produces an elevated level of Δexon 17&18 isoform when compared to most of the controls. Error bars represent standard error of the mean.

The expression of the three isoforms (Δexon 18, Δexon 17&18 and full-length) were then normalized to expression of the reference gene GAPDH, to assess relative levels of each the three transcripts (Figure 3C–E). Figure 3(D) shows that the relative level of the Δexon 18 alternative isoform is increased in both of the variant carriers when compared to controls. The relative level of the Δexon 17&18 isoform is also increased in the variant carriers compared to some of the controls assayed (Fig 3E). However, full-length expression by the carrier was within the normal range of that expressed in the controls (Fig 3C). This suggests that, despite the increase in alternative isoform expression, the full-length isoform is expressed at levels similar to those observed in individuals who do not carry the c.7988A>T (p.Glu2663Val) variant. This is consistent with the fact that the full-length transcript predominates in the absence of NMD, comprising greater than 90% of total transcript in the c.7988A>T (p.Glu2663Val) carriers, and greater than 95% in controls (data not shown).

Results were similar for untreated cell lines when real-time PCR data was normalized to GAPDH and an alternative reference gene, EEF1A1 (Supp. Figure S2).

Variant and wildtype full-length transcript expression in BRCA1 c.4868C>G, BRCA2 c.7988A>T, and BRCA2 c.8168A>G carriers

Based on the splicing aberrations observed, these three variants might be interpreted to be pathogenic: BRCA1 c.4868C>G (p.Ala1623Gly) and BRCA2 c.8168A>G (p.Asp2723Gly) create an aberrant splice product encoding a truncated protein, and BRCA2 c.7988A>T (p.Glu2663Val) results in elevated levels of an alternative mRNA isoform encoding a truncated protein. However, we further investigated if both variant and wildtype alleles were expressed in the full-length isoform, to determine if a full-length missense protein could be encoded by variant carriers. Sequencing of the full-length band revealed the presence of both wild-type and variant allele for BRCA1 c.4868C>G (p.Ala1623Gly). Similarly, sequencing of the upper band in Figure 2D identified both alleles for c.8168A>G (p.Asp2723Gly) in the full-length transcript. Sequencing of the 895bp full-length band in Figure 2E also showed evidence that both alleles for c.7988A>T (p.Glu2663Val) were present in the full-length transcript, indicating that both variant and wildtype alleles are expressed and suggesting that both missense and normal BRCA2 protein are likely to be translated in vivo. This is consistent with the fact that full-length transcript was within the normal range in LCLs for carriers of BRCA2 c.7988A>T (p.Glu2663Val).

Assays were then carried out to quantitate the proportion of variant allele expressed as full-length transcript. A double stranded (ds) cDNA-based assay was performed for the c.7988A>T (p.Glu2663Val) variant using TaaI restriction digest of the variant allele followed by real-time PCR. Quantitation of the undigested cDNA was used to represent total full-length transcript and quantitation from the digested cDNA represented wildtype transcript only. As shown in Figure 4, the contribution of the variant allele was reduced compared to the wildtype allele, constituting 39% and 41% of the overall full-length transcript for two related carriers. Furthermore, the relative levels of full length, Δexon 18, and Δexon 17&18 isoforms shown in Figure 3(A–B) were replicated using ds-cDNA as a template, suggesting that cDNA levels were similarly represented in both the ss-cDNA and ds-cDNA (data not shown). Due to the paucity of restriction enzyme sites, it was not possible to design a quantitative variant-specific digestion assay for other variants in the study. However, results from a semi-quantitative cloning-based methodology carried out by the Wessex clinical testing laboratory indicated that the BRCA1 c.4868C>G (p.Ala1623Gly) variant had an incomplete effect on splicing, with 10/15 clones shown to have the wildtype C allele.

Open in a separate window

Figure 4

Quantitation of wildtype and variant alleles in BRCA2 c.7988A>T (p.Glu2663Val) following digestion of the variant allele in ds-cDNA by TaaI. The experiment was carried out using ds-cDNA synthesized from two related carriers of the variant.

Functional evaluation of BRCA2 missense substitutions displaying splicing aberrations

The functional consequences of the BRCA2 missense substitutions p.Glu2663Val and p.Asp2723Gly were assessed using results from homologous recombination and centrosome amplification assays, to evaluate whether the missense substitution was likely to be pathogenic. These commonly used assays of BRCA2 protein function were selected since they are considered key indicators of abrogated BRCA2 function. BRCA2 is implicated in homologous repair at sites of double stranded DNA breaks through the interaction with RAD51 (which possesses several activities required for homologous recombination) and single-stranded DNA (Davies, et al., 2001; Wu, et al., 2005), and a lack of functional BRCA2 is shown to lead to gross chromosomal rearrangements indicative of compromised homologous recombination (Patel, et al., 1998). Defective BRCA2 also leads to increased centrosome amplification (Tutt, et al., 1999) recognized as an indicator of micronucleus development resulting from genome instability (Chester, et al., 1998). Abrogation of function observed for either or both of these assays has been observed for known pathogenic variants, such as p.Asp2723His, indicating that these assays are probably independent predictors of BRCA2 protein function (Farrugia, et al., 2008; Wu, et al., 2005).

Results are summarized in Supp. Table S3, and include results from assays conducted as part of this study (for p.Glu2663Val), and assay results previously reported by some of the authors (Farrugia, et al., 2008). Full-length p.Asp2723Gly showed increased centrosome amplification (29% vs 10% for wildtype) and reduced homologous recombination (HR) activity (1.6% vs 6.1% for widltype). These results were similar to those obtained using the known pathogenic p.Asp2723His variant shown in Supp. Table S3 (Farrugia, et al., 2008; Wu, et al., 2005). Full-length p.Arg3052Trp showed reduced HR activity (1.6%) but did not induce significant centrosome amplification (13%). The p.Glu2663Val substitution showed increased centrosome amplification (29%) and reduced homologous repair (HR) activity (1.8%) relative to wildtype BRCA2. The results indicate that these missense substitutions both have the capacity to inactivate BRCA2 and predispose to cancer in the absence of dysregulated splicing.

We gratefully acknowledge the participation of the families concerned. We wish to thank Heather Thorne, Eveline Niedermayr, all the kConFab research nurses and staff, the heads and staff of the Family Cancer Clinics, and the Clinical Follow Up Study (funded by NHMRC grants 145684, 288704 and 454508) for their contributions to this resource, and the many families who contribute to kConFab. kConFab is supported by grants from the National Breast Cancer Foundation, the National Health and Medical Research Council (NHMRC) and by the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia. We thank Myriad Genetic Laboratories for information used to derive family history scores and investigate co-occurrence of variants with pathogenic. We thank Bruce Castle for providing clinical samples, and Joanne Dunlop from the West Midlands Regional Genetics Laboratory and Diana Barelle from the Wessex Regional Genetics Laboratory, for access to results from RNA analysis of BRCA1 c.4868C>G (p.Ala1623Gly).

Funding: This work was supported by a grant from The National Health and Medical Research Council (NHMRC) [ID442970]. kConFab is supported by grants from the National Breast Cancer Foundation, the NHMRC and by the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia. The kConFab Clinical Follow Up Study was funded by NHMRC grants [145684 and 288704]. PW was awarded a scholarship by the QIMR Higher Degrees Committee. ABS is an NHMRC Senior Research Fellow, LDaS was supported by a fellowship from the Ludwig Institute for Cancer Research. FC, SVT and DEG were supported in part by the INHERIT BRCAs programme from the Canadian Institute for Health Research. FJC was supported by NIH grants [CA116167] and a Breast Cancer specialized Program in research Excellence grant [P50 CA116201] and an American Cancer Society award [RSG-04-220-01-CCE].