Probing structure-function relationships in missense variants in the carboxy-terminal region of BRCA1

Abstract

Germline inactivating variants in BRCA1 lead to a significantly increased risk of breast and ovarian cancers in carriers. While the functional effect of many variants can be inferred from the DNA sequence, determining the effect of missense variants present a significant challenge. A series of biochemical and cell biological assays have been successfully used to explore the impact of these variants on the function of BRCA1, which contribute to assessing their likelihood of pathogenicity. It has been determined that variants that co-localize with structural or functional motifs are more likely to disrupt the stability and function of BRCA1. Here we assess the functional impact of 37 variants chosen to probe the functional impact of variants in phosphorylation sites and in the BRCT domains. In addition, we perform a meta-analysis of 170 unique variants tested by the transcription activation assays in the carboxy-terminal domain of BRCA1 using a recently developed computation model to provide assessment for functional impact and their likelihood of pathogenicity.

Figure 1. BRCA1 carboxy-terminal variants.

( A ) Natural and artificial (underlined) BRCA1 variants in the context of the analyzed region (comprising amino acids residues 1396–1863). ( B ) BRCA1 R1699 and E1836 variants in the context of the analyzed region (comprising amino acids residues 1560–1863). Structural models were obtained with Mupit tool (http://mupit.icm.jhu.edu/) using the 1jnx PDB structure (amino acid residues R1699 and E1936 are depicted in green).

Figure 2. Phylogenetic conservation of amino acid residues in the surround region of BRCA1 analyzed variants.

A multiple sequence alignment demonstrating amino acids residues evolutionary conservation are shown from H.sapiens (NP_009225.1), P.troglodytes (NP_001038958.1), M. mulatta (NP_001108421.1), C.lupus (NP_001013434.1), B.taurus (NP_848668.1), M. musculus (NP_033894.3), R. norvegicus (NP_036646.1) and G. gallus (NP_989500.1). Target amino acids residues are depicted in light grey and salt-bridge involved residues in dark grey.

Figure 3. Functional analysis of missense variants in BRCA1 C-terminal region.

Transcriptional activity of BRCA1 variants were evaluated in HEK293T cells using a GAL4-responsive firefly luciferase reporter gene (shown above the graphs) at 37°C. Cells were harvested 24h after transfections and the lysate was used to assess transcriptional activation ability by luciferase activity measurement. Activity is depicted as % of the wild-type activity. ( A ) Natural missense variants and ( B ) natural and artificial (underlined) variants located on phosphorylation sites. S1613G, M1775R and Y1853X variants were used as controls.

Figure 4. Functional analysis of BRCA1 R1699 and E1836 variants at different temperatures.

Transcriptional activity of BRCA1 variants were evaluated using a GAL4-responsive firefly luciferase reporter gene (shown above the graphs). Cells were harvested 24h after transfections and the lysate was used to assess transcriptional activation ability by luciferase activity measurement. Activity is depicted as % of the wild-type activity. ( A ) Transcriptional activity of R1699L, R1699I, R1699F, R1699E, R1699G, R1699Q and R1699H variants and the controls M1775R and Y1853X at different temperatures evaluated in HEK293T cells; ( B ) transcriptional activity of the R1699K variant and the controls M1775R and Y1853X at different temperatures evaluated in HEK293T and HCC1937 cells and ( C ) transcriptional activity of the E1836D, E1836G and E1836K variants and the control M1775R at different temperatures evaluated in HEK293T. Transcriptional activity of double variants ( D ) R1699K/E1836D, R1699K/E1836G and R1699K/E1836K ( E ) R1699E/E1836D, R1699E/E1836G and R1699E/E1836K and the control M1775R at different temperatures evaluated in HEK293T at different temperatures.

Figure 5. Estimated Variant Specific Effects (Bayesian statistical model, VarCall, graphical summary).

Variants are depicted in order of amino acid from residues 1396 to 1863. The top panel shows secondary structures in the C-terminal region. Coiled-coil region and α-helixes are depicted as pink and blue boxes, respectively. β-sheets are shown as gray arrows. The shaded area on the graph corresponds to different structures of similar color. The linker region is indicated with green shading. Each variant’s activity is represented by a boxplot summarizing the marginal posterior distribution of its random effect. A point estimate of the mixture model is plotted on the right margin. Its top component corresponds to variants with no functional impact, whereas its bottom component corresponds to variants with functional impact. The mean of the benign/damaging component is plotted as a green/red dotted horizontal line. Yellow box represents wild-type reference, green and red boxes represent low (class 1 and 2) and high (class 4 and 5) risk variants respectively, classified according to the International Agency for Research on Cancer (IARC), purple boxes represent VUS previously analyzed by our group (used to feed VarClass algorithm), and blue boxes represent the VUS analyzed by the first time in this study.