Cloning of the human NADH: ubiquinone oxidoreductase subunit B13: localization to Chromosome 7q32 and identification of a pseudogene on 11p15

Mitochondrial dysfunction is becoming increasingly recognized asa cause of cardiovascular and neuromuscular diseases. Defects ofthe mitochondrial respiratory chain represent a heterogeneousgroup of disorders that range in presentation from severe multi-system organ failure in the neonatal period to exaggerated fatiguewith exercise (Shoffner and Wallace 1995). To date, many of thedescribed mitochondrial defects responsible for human diseaseshave involved the mitochondrial NADH:ubiquinoneoxidoreduc-tase enzyme (Complex I), which catalyzes the transfer of electronsfrom NADH to ubiquinone, couples with ADP phosphorylation(Schapira et al. 1988). This complex is composed of at least 40proteins, seven of which are encoded by mitochondrial DNA andthe remainder by nuclear genes (Walker et al. 1992). Complex Idefects have been noted in patients with a variety of neuromuscu-lar disorders, including patients with Parkinson’s disease (Parker etal. 1989). Recently, one component of this complex, the B22 sub-unit, was proposed as a candidate gene for the Branchio-oto-renalsyndrome, an autosomal dominant disorder characterized by hear-ing loss, renal abnormalities, and cervical fistulae (Gu et al. 1996).Some patients with severe Complex I deficiency have been dem-onstrated to have reduced amounts of the nuclear-encoded B13subunit as well as a marked diminution of the 24-kDa subunit(Moreadith et al. 1987; Morgan-Hughes et al. 1988). Although thefunction of the B13 subunit is unknown, it demonstrates significanthomology to a fungal protein, the 29.9-kDa subunit from the Com-plex I of Neurospora crassa (Walker et al. 1992). The evolutionaryconservation of this protein and its reduced expression in patientswith severe Complex I deficiency suggest that it may have func-tional domains that are important for the understanding of Com-plex I disorders.During the course of a physical mapping project on Chromo-some (Chr) 11p15.5, we identified sequence from the end of cos-mid clone cCI-253 (gift from Yusuke Nakamura) with a highdegree of similarity to the bovine NADH:ubiquinone oxidoreduc-tase subunit B13. From the genomic sequence, we constructed asequence tag site (STS) corresponding to the putative carboxyl and38 untranslated sequence, using the bovine gene as a reference.These STSs were amplified from Chr 11 cosmids (Chr 11-specificcosmid library, LA11NC01, from L. Deaven, Los Alamos Na-tional Laboratory) that spanned the region. the PCR product wasradiolabeled and used to screen an arrayed infant brain cDNAlibrary (Adams et al. 1993). Four cDNA clones of varied sizeswere isolated, all with identical sequences in the overlapping in-tervals. Two apparently identical 1.7-kb clones contained a Kozakconsensus start sequence (Kozak 1987). A single clone was se-quenced (Genbank accession: U64028) on both strands with fluo-rescent cycle sequencing (fmol cycle sequencing kit, Promega Inc.,Madison, Wis.). This clone contains a single 348-nucleotide openreading frame (116 amino acids) and a relatively large 38 untrans-lated region (1,092 nucleotides). During the course of this work,Pata and colleagues have deposited a similar sequence into Gen-bank (accession number: U53468). Except for some differences inthe 38 untranslated region, our sequence is identical to that depos-ited by Pata and associates. It is unknown whether the differencesbetween the sequence reported here and that from Pata and col-leagues are naturally occurring variants or sequencing errors.The nucleotide sequence of this clone was 88% and 83% iden-tical over the predicted open reading frame compared with bovine(accession number: X63218) and rat B13 (accession number:D86215) subunit genes respectively. The position of the initiationand termination codons was conserved. Sequence immediately fol-lowing the translation start site encodes a series of positivelycharged amino acids, which are necessary for the transit of theprotein into the mitochondrion (Neupert et al. 1990). The predictedhuman B13 polypeptide sequence is 88% and 83% identical com-pared with the bovine and rat proteins (Fig. 1). Interestingly, thehuman protein differs from the rat and bovine protein at ten resi-dues where the rat and cow also differ from each other. Theseregions of difference may indicate residues not evolutionarily con-strained by function.To determine the chromosomal localization of the B13 subunitgene, we designed an STS from the 38 untranslated sequence of thecDNA where the cDNA significantly diverges from the Chr 11sequence. A monochromosome somatic cell hybrid (SCH) panelwas screened with this STS, and only the SCH containing humanChr 7 was positive (data not shown). This STS was then used toisolate a P1 clone (DMPC-HFF #1, clone 56A9; Shepard et al.1994) with a PCR-based strategy. Fluorescent in situ hybridizationwith this P1 localized the B13 gene to Chr 7q32 (Fig. 2). Extensivesequencing of the genomic DNA on 11p15.5 revealed that therewas high similarity to the true gene, but that there was a one basepair deletion in the ‘‘coding sequence’’ of the pseudogene and itdid not contain a Kozak consensus start sequence. This similaritywas observed in both the coding and 38 untranslated sequence, butthe nucleotide differences were not concentrated at the third or‘‘wobble’’ base of the coding sequence. Furthermore, there was noevidence of a splice acceptor site at the 58 extent of the pseudo-gene, indicating that it was not likely to be an exon from a largertranscript that had incorporated a B13 functional domain. Thesequence was searched against the Genbank and dbEST databaseswith the BLAST program (Altschul et al. 1990). In addition to