Specific Localization of Missense Mutations in the VHL Gene in Clear Cell Renal Cell Carcinoma

Missense mutations in the VHL gene during sporadic clear cell renal cell carcinoma were studied to evaluate their localization in relation to functionally important motifs of the VHL protein. Somatic mutations were identified in 124 of 307 samples. All missense mutations in the α-domain were localized in the binding site for elongin C. Substitutions in the β-domain (77%) were found in the HIF-binding site. Five missense mutations were absent in these sites, which illustrates their role in VHL protein formation or suppressor function of other protein cofactors. Mutation c.392A→T (p.N131I) was identified for the first time. Our results hold much promise to estimate the boundaries of functionally important sites in the VHL suppressor gene and contribute to the interpretation of a pathogenic role of mutations in direct DNA diagnostics.

Renal cell carcinoma is an urgent problem of modern oncourology. The annual worldwide incidence of this disease is more than 200,000 cases [3]. Kidney tumors include a variety of pathomorphological types. Clear cell renal cell carcinoma (CCRCC) is the most abundant type of these tumors (80% cases) [10]. The major candidate gene, whose mutations play a role in CCRCC carcinogenesis, was identified in studying the Von Hippel-Lindau disease (VHL disease, OMIM 193300). This state is accompanied by CCRCC, CNS hemangioblastoma, and adrenal pheochromocytoma. VHL disease results from mutations in the VHL gene located in the 3p25 region [7,13]. Germline VHL mutations are characterized by genotypic-phenotypic correlations. VHL disease is classified to type 1 (absence of pheochromocytoma and high risk for CCRCC) and type 2 (presence of pheochromocytoma). Type 1 VHL disease is charac-terized by frameshift mutations, nonsense mutations and, more rarely, by missense mutations (preventing VHL polypeptide folding). Missense mutations are typical of type 2 VHL disease [2,15]. Somatic VHL mutations are the most abundant point mutations during sporadic CCRCC and therefore attract much attention as the main driver mutations and possible predictors of the response to the target agents. VHL mutations are the early specific event in CCRCC. However, there are contradictory data on their association with the stage, degree of differentiations, and survival rate [1,5]. Sporadic CCRCC is usually accompanied by gene-inactivating deletions/insertions with frameshift mutations or missense mutations. They play a crucial role in the interaction with HIF (hypoxia-inducible factor) or elongin C [8]. At the same time, some missense mutations produce an adverse effect on other functional motifs of VHL and play various pathogenetic roles in CCRCC carcinogenesis [12].
In the present work, missense mutations in the VHL gene during sporadic CCRCC were studied to evaluate their specific localization in relation to binding sites for the proteins interacting with VHL.

MATERIALS AND METHODS
The study was performed on 307 frozen specimens of primary CCRCC tissues. Genomic DNA was isolated with AmpliPrime DNA-sorb B kit (InterLabService).
Encoding sequence regions of the VHL gene were amplified with primers flanking exons 1-3 [9]. PCR products were treated with alkaline phosphatase (1 U) and E. coli exonuclease I (4 U) to remove the nonreacting primers and deoxynucleotide triphosphates.
PCR products were subjected to Sanger sequencing with BigDye Terminator v. 3.1 Cycle Sequencing kit. Fluorescent-labeled products of the reaction were detected on a 24-capillary genetic analyzer 3500xl (Applied Biosystems).

RESULTS
VHL mutations were found in 41% cases of CCRCC (124/307). These mutations were somatic (absent in the adjacent renal parenchyma). Our results on the incidence of mutations are consistent with the COS-MIC database and published data [1,11,12,14]. Point mutations in VHL were previously revealed in 40-60% cases of sporadic CCRCC. These mutations appeared as follows: 58% deletions (72/124), 14% insertions (17/124), 5% complex mutations (7/124), and 23% single-nucleotide replacements (28/124). Single-nucleotide replacements were presented by nonsense mutations (18%, 5/28), 7% splicing mutations (7%, 2/28) and, most often, by missense mutations (75%, 21/28). In 75% mutant samples (93/124), DNA changes were followed by a shift of the reading frame and/or formation of new stop codons and impairment of splicing. Point deletions (insertions) not affecting the reading frames are of interest in relation to critical domains of VHL. These mutations were revealed in 10 cases (Table 1). Only 3 deletions were shown to impair the HIF-binding site. However, the removal of several amino acids from α-or β-domains near the binding sites for elongin C or HIF is followed by conformational changes in VHL. Moreover, missense mutations illustrate functionally important motifs of VHL ( Table  2). Among missense mutations, all replacements in the α-domain were localized exactly in the binding site for elongin C (codons 157-171). Most replacements (77%) were found in the β-domain, exactly in the binding sites for main isoforms of HIF (HIF-1α and HIF-2α, codons 67-117). Our results are consistent with previous data that HIF and elongin C serve as major cofactors in the formation of a ubiquitin-ligase complex for HIF degradation. The loss of binding to any of these proteins makes impossible the function of VHL as a tumor suppressor [2]. However, 5 missense mutations were absent in the binding sites for HIF or elongin C. This fact is associated with their role in VHL protein folding or suppressor function of other proteins, which bind to VHL in the sites of mutations (Table 2) [5,12]. According to the HGMD database, 43% missense mutations (9/21) were previously described as germline variants in patients with VHL disease (which confirms high pathogenicity of these mutations). According to the COSMIC database, 20 of 21 missense mutations were found in sporadic CCRCC. One mutation c.392A→T (p.N131I) was identified for the first time.  New data on somatic missense mutations in VHL during CCRCC allow us not only to map the bounda ries of critical motifs in this tumor suppressor, but also to facilitate the diagnostics of VHL disease (in the absence of similar mutations in databases) [1,12]. There are programs for in silico identification of single-nucleotide replacements in VHL as mutations of low and high risk for carcinogenesis and for the differentiation of passenger mutations from driver mutations. These features are important for the next-generation sequencing (NGS) assay [6]. The use of NGS will allow us to perform deep profiling of somatic driver mutations not only in VHL, but also in other candidate genes. The most effective target agents can be evaluated [4,5].