Atp11p and Atp12p are assembly factors for the F(1)-ATPase in human mitochondria.

Atp11p and Atp12p were first described as proteins required for assembly of the F(1) component of the mitochondrial ATP synthase in Saccharomyces cerevisiae (Ackerman, S. H., and Tzagoloff, A. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 4986-4990). Here we report the isolation of the cDNAs and the characterization of the human genes for Atp11p and Atp12p and show that the human proteins function like their yeast counterparts. Human ATP11 spans 24 kilobase pairs in 9 exons and maps to 1p32.3-p33, while ATP12 contains > or =8 exons and localizes to 17p11.2. Both genes are broadly conserved in eukaryotes and are expressed in a wide range of tissues, which suggests that Atp11p and Atp12p are essential housekeeping proteins of human cells. The information reported herein will be useful in the evaluation of patients with ascertained deficiencies in the ATP synthase, in which the underlying biochemical defect is unknown and may reside in a protein that influences the assembly of the enzyme.

The energy needs of aerobic organisms are met principally through the process of oxidative phosphorylation in which ATP production by the ATP synthase is coupled to a transmembrane proton gradient (1). The ATP synthase is comprised of an integral membrane component called F 0 and a peripheral moiety called F 1 . Different organisms show variation with respect to the subunit composition of the F 0 sector, while the ␣ 3 ␤ 3 ␥␦⑀ oligomeric structure of the catalytic F 1 unit has been highly conserved (1). All five F 1 subunits and most F 0 subunits of the mammalian ATP synthase are nuclear gene products; only two F 0 proteins (subunits 6 and 8) are coded for in the mitochondrial DNA (2). As obligate aerobes, mammals are completely dependent on the ATP synthase for life. However, while a genetic defect that leads to a complete loss of the enzyme function is lethal, partial loss of ATP synthase function can be tolerated to some extent. For example, the T8993G mutation of human mitochondrial DNA, which results in an amino acid substitution in F 0 subunit 6, produces either the NARP syndrome (neurogenic muscle weakness, ataxia, retinitis pigmentosa) or the more severe Leigh's syndrome depending on the mutant load (3,4).
Nuclear mutations that might be expected to lower the amount of functional ATP synthase in human cells, without depleting the enzyme completely, are those that affect proteins involved in its biogenesis. Genetic screens of respiratory-defective mutants of Saccharomyces cerevisiae have identified several nuclear genes (ATP10, ATP11, ATP12) whose products are required for ATP synthase assembly (5,6). Notably, none of these "assembly factors" are structural components of the yeast ATP synthase, which is now reported to contain at least 17 different types of subunits (7,8). Instead, Atp10p is an assembly factor that is required for proper formation of the F 0 component (5), and Atp11p and Atp12p are assembly factors for the F 1 oligomer (6). Work in our laboratory has provided evidence that Atp11p interacts specifically with the F 1 ␤ subunit (9), while Atp12p binds selectively to the F 1 ␣ subunit (10). It is through this binding mechanism that Atp11p and Atp12p are proposed to protect the F 1 subunits from forming nonproductive (␣) n and (␤) n complexes during assembly of the enzyme oligomer (10).
To facilitate the analysis of F 1 assembly in human cells, efforts were made to determine whether functional human orthologs of the yeast Atp11p and Atp12p assembly factors exist. To this end, we have isolated a partial human cDNA for Atp11p and have shown that the encoded product interacts with the F 1 ␤ subunit. A full-length human cDNA for Atp12p was also isolated, and we report here that human Atp12p interacts with the F 1 ␣ subunit in a two-hybrid screen and complements a yeast atp12 mutant. Genomic analysis provides evidence that supports the chromosomal assignment of human ATP11 to 1p32.3-p33 and of human ATP12 to 17p11.2.

EXPERIMENTAL PROCEDURES
Cloning Human cDNAs for Atp11p and for Atp12p-Reverse transcription of cDNA from human fetal liver mRNA (CLONTECH, Palo Alto, CA) and reactions to rapidly extend the cDNA end (RACE) 1 at the 5Ј terminus followed procedures from Life Technologies, Inc. manual number 18374-058. The primers used for first-strand cDNA synthesis to clone ATP11 were designed to be complementary to nucleotides 3Ј to the coding region, using a contig sequence assembled from 70 expressed sequence tags (ESTs) belonging to the UniGene Cluster Hs.10964. The cDNA was dC-tailed at the 5Ј end, which enabled Life Technologies' Abridged Anchor Primer to be annealed at this terminus for 5Ј RACE reactions; 3Ј primers for such reactions employed sequences complementary to portions of one of the longer ATP11 ESTs (GenBank TM accession number AA150795). To clone the cDNA for human ATP12, a dC-tailed first-strand cDNA was synthesized from human fetal liver mRNA using a poly(dT) primer. This product was then amplified using the 5Ј RACE Abridged Anchor Primer (Life Technologies, Inc.) and a 3Ј primer that is complementary to nucleotides ϩ943 to ϩ959 in human mRNA 24418 2 (GenBank TM accession number AF052185, see text for details). This fragment served as the template for a second PCR with an upstream primer (5Ј-CCAGAATTCGCGTCTCGCATC-3Ј) corresponding to nucleotides Ϫ41 to Ϫ22 of mRNA 24418, except for two nucleotides (underlined) altered to create an EcoRI site, and a downstream primer (5Ј-TGCCTCGAGCCTGCTGAGTGT-3Ј) complementary to nucleotides ϩ885 to ϩ905 of mRNA 24418 except for two nucleotides (underlined) altered to create an XhoI site. This product was ligated to the EcoRI and XhoI sites of pCUP2A5/CEN316 (11) to create pCUPHUATP12/CEN316.
Genomic Analysis-Sequence alignments, genomic, mRNA, and EST sequence homology searches and genomic structure determinations were performed using the on-line BLAST algorithm package at the National Center for Biotechnology Information web site (www.ncbi. nlm.nih.gov/BLAST/). EST data was accessed from the UniGene website (www.ncbi.nlm.nih.gov/UniGene/Hs.Home.html) and aligned using AssemblyLIGN (Oxford Molecular, Cambridge, United Kingdom). SAGE analysis data were accessed from the National Cancer Institute Cancer Genome Anatomy Project's SAGEmap web resource (www.ncbi.nlm.nih.gov/SAGE/). Mapping information was accessed from GeneMap '99 (18), CompView (19), the Unified Data base (bioinformatics.weizmann.ac.il/udb/), and the Whitehead Institute Center for Genome Research radiation hybrid (RH) map (20).
Miscellaneous Methods-The yeast two-hybrid screen (21) employed yeast vectors pACT2 and pAS2-1 and host strain Y190 (described above) as supplied in the MATCHMAKER Two-Hybrid System 2 from CLONTECH. Yeast were grown in SD medium omitted for leucine, tryptophan, and histidine. Qualitative assessment of expression from the lacZ reporter gene was made using 5-bromo-4-chloro-3-indoyl ␤-D-galactopyranoside (X-gal) as a chromogenic substrate for ␤-galactosidase in a colony-lift filter assay, as described in the CLON-TECH manual. Standard techniques were used for restriction endonuclease analysis of DNA, purification and ligation of DNA fragments, and transformations of and recovery of plasmid DNA from Escherichia coli (22). Yeast transformations employed the LiAc procedure (23). The preparation of yeast mitochondria and assay of oligomycin-sensitive ATPase activity were done as described previously (24). Protein concentrations were estimated by the method of Lowry et al. (25).

RESULTS
Cloning of a Human cDNA for Atp11p-At the time this work was initiated, a TBLASTN search of the human EST data base identified the 444-base pair sequence AA150795, which upon translation in the 5Ј 3 3Ј direction in frame 1 yields a putative peptide that shares significant sequence homology with yeast Atp11p (BLAST score: p ϭ 3 ϫ 10 Ϫ5 ). AA150795 is a member of UniGene Cluster Hs.10964. A contig for human ATP11 assembled with ESTs from this cluster includes a termination codon, but is incomplete at the 5Ј terminus. We isolated a human ATP11 cDNA clone bearing more sequence at the 5Ј end than any of the 5Ј-read ESTs from RACE reactions that employed human fetal liver mRNA as the template (see "Experimental Procedures"). However, despite numerous attempts, the best product achieved was found to harbor 270 nucleotides proximal to the 5Ј end of EST AA150795, which codes for an additional 90 amino acids but still lacks the initiator methionine. We subsequently identified the mRNA HRC06325 (GenBank TM accession number AK026004) as the complete transcript for the ATP11 gene of Homo sapiens (BLAST score: p ϭ 3 ϫ 10 Ϫ10 versus yeast Atp11p). Efforts to clone the complete cDNA for Atp11p from human fetal liver mRNA by PCR were not successful.
An alignment of the primary sequences for Atp11p deduced from the human mRNA and S. cerevisiae gene is given in Fig.  1. The sequences show 44% similarity (including identical and physico-chemically similar amino acids) and 21% identity. In comparison, the human protein shows 46% similarity and 22.3% identity with Atp11p of Schizosaccharomyces pombe and 58.6% similarity and 31% identity with Atp11p of Drosophila yakuba (11). The black bar in Fig. 1 highlights a stretch of nine amino acids, PXFXXXLPR (termed the "flipper sequence"), which can be considered a signature sequence for Atp11p proteins (see Ref. 11). The arrow indicates the first amino acid (Lys-83) that is encoded by the partial human cDNA for Atp11p isolated in our laboratory. The primary translation product of yeast ATP11 includes a leader sequence at the amino terminus that targets the protein to mitochondria (16); the first amino acid of the mature protein is predicted to be Glu-40 (27)  The amino acid sequence of Atp11p from S. cerevisiae (SCER) and H. sapiens (HSAP) was aligned using the CLUSTALW program (26). Shading indicates identical and physico-chemically similar amino acids; identical amino acids are further highlighted in bold. The boxed amino acids are the approximated start sites for the mature proteins. The following symbols are used to indicate specific elements that are discussed in the text: arrow, human Atp11p Lys-83; black bar, Atp11p consensus sequence. residue in S. cerevisiae Atp11p sequence, Fig. 1). As mitochondrial leader sequences are known to be rich in basic and hydroxylated residues and deficient in acidic amino acids (28), we predict that the amino terminus of the mature human Atp11p protein begins in the vicinity of Glu-57 (boxed residue in H. sapiens Atp11p sequence, Fig. 1). By this analysis our cDNA clone encodes a human Atp11p protein that lacks ϳ25 amino acids from the mature amino terminus.
Functional Studies with Truncated Human Atp11p Proteins-Previous work has indicated that yeast Atp11p retains function following the deletion of either 72 amino acids from the amino terminus or 135 amino acids from the carboxyl terminus of the mature protein (24). Hence, it was deemed worthwhile to pursue functional studies with incomplete cDNAs coding for partial human Atp11p products. In one set of studies, the yeast two-hybrid screen was used to test the putative human Atp11p assembly factor for binding interactions with the human F 1 ␣ and ␤ subunits. The ATP11 plasmid used in this study, pAS2-1/HuATP11(83-300), bears most of the sequence for mature human Atp11p (Lys-83 through Glu-300, see Fig. 1) fused to the DNA binding domain (BD) of Gal4p. Plasmids coding for the mature portions of the human F 1 ␣ subunit (Gln-44 through Ala-553; plasmid pACT2/ HuATP1(44 -553)) and of the human F 1 ␤ subunit (Arg-54 through Ser-539, plasmid pACT2/HuATP2(54 -539)) harbor fusions to the transcription activation domain (AD) of Gal4p. Yeast colonies that co-produced Gal4pBD-human Atp11p and Gal4pBD-human F 1 ␤ subunit fusion proteins were stained dark blue following exposure to X-gal on filter paper (see "Experimental Procedures"). Such results, which indicate a functional reconstitution of the Gal4p transcription factor in the lacZ reporter strain (Y190), provide evidence for direct binding interactions between the human Atp11p and ␤ subunit proteins. In contrast, there was no evidence from this assay of human Atp11p binding to the human F 1 ␣ subunit, as yeast cells that co-produced Gal4pBD-human Atp11p in combination with Gal4pAD-human ␣ subunit did not stain blue with X-gal. Of particular note is that also yeast Atp11p shows substrate specificity for the F 1 ␤ subunit (9).
The truncated human Atp11p also interacts with the ␤ subunit of S. cerevisiae in the two-hybrid assay, on which basis we predicted that the human cDNA clone for Atp11p might com-plement a yeast atp11 mutant. For these experiments, a multicopy yeast plasmid was constructed (pG13L/HuATP11(83-328)) that harbors most of the sequence for mature human Atp11p (Arg-83 through Thr-328) fused, at the 5Ј end, in frame with the yeast Atp11p leader peptide sequence. By such means it was anticipated that the plasmid-borne protein would be targeted correctly to mitochondria in yeast. The chimeric gene failed to confer respiratory competence to the atp11 mutant, aW303⌬G13, following selection on nonfermentable substrates (ethanol-glycerol plates) at three different temperatures (25, 30, and 37°C). This result may reflect species-specific features of the Atp11p mechanism that preempt the human protein from substituting effectively for the yeast assembly factor in vivo (see "Discussion").
Cloning of a Human cDNA for Atp12p and Alignment of Its Product with Other Homologous Proteins-The TBLASTN program identified the human mRNA 24418 (GenBank TM accession number AF052185) of UniGene cluster Hs.13434 as the transcript for human Atp12p (BLAST score: p ϭ 2 ϫ 10 Ϫ7 versus yeast Atp12p). A full-length cDNA, putatively identified as coding for human Atp12p, was cloned by PCR using a template that was reverse transcribed from human fetal liver mRNA (see "Experimental Procedures").
In addition to the human ortholog, data base searches with S. cerevisiae Atp12p as the query have identified four additional complete amino acid sequences that share significant sequence similarity with the protein of budding yeast. These sequences are deduced from genomic information available for Drosophila melanogaster (GenBank TM accession number AE003669, clone CG8674, 4 BLAST score: p ϭ 5.9 ϫ 10 Ϫ3 ), Rhodobacter capsulatus (GenBank TM accession number RU37407, 5  number AL121764, clone SPAC9, 7 p ϭ 3 ϫ 10 Ϫ32 ). A multiple alignment of these four sequences, along with Atp12p from S. cerevisiae and H. sapiens, is given in Fig. 2. Overall, the extent of similarity (12.7%) and identity (4.6%) between all six Atp12p sequences is low. However, pairwise analysis yields similarity values higher than 40% and identities of at least 20% in most cases ( Table I). The mitochondrial leader peptide of S. cerevisiae Atp12p is estimated to span the first 30 amino acids such that mature protein starts with Gly-31 (14) (boxed residue in S. cerevisiae Atp12p sequence, Fig. 2). The other four eukaryotic Atp12p proteins are likewise predicted to harbor a mitochondrial targeting sequence, while the protein of the photosynthetic bacterium should not and is accordingly much shorter at the amino terminus. The fact that Glu-289 of S. cerevisiae Atp12p aligns with an acidic amino acid in the other five sequences (see arrow, Fig. 2) is of particular significance, since work in yeast has shown that an acidic side chain is required in this position for maximal Atp12p activity (14).
Functional Studies with Human Atp12p-The product encoded by the putative human cDNA for Atp12p was evaluated for function in yeast two-hybrid and genetic complementation assays. The two-hybrid screen employed plasmid pAS2-1/ HUATP12 , which codes for the Gal4p DNA binding domain fused to what is approximated to be the mature portion of the human Atp12p protein (Met-27 through Glu-289), and pACT2 plasmids bearing fusions between the Gal4p transcription activation domain and the sequences for the human F 1 ␣ and ␤ subunits (see above). Yeast cells, which co-produced the Gal4p fusion proteins that paired human Atp12p with the human F 1 ␣ subunit stained dark blue with X-gal, while there was only a weak blue signal for cells in which Gal4pBD-human Atp12p was produced in combination with the Gal4pAD-human F 1 ␤ subunit. Such results are in accord with the fact that yeast Atp12p shows high selectivity for the yeast ␣ subunit over the yeast ␤ subunit in a two-hybrid screen (10).
The cDNA cloned for human Atp12p is anticipated to encode a precursor protein that is complete with respect to mitochondrial targeting function. Hence, for complementation studies, the entire cDNA was subcloned into a multicopy yeast vector 3Ј of the yeast CUP1 promoter (plasmid pCUPHUATP12/YEp), and this plasmid was used to transform the respiratory-deficient atp12 strain aW303⌬ATP12. Such transformants failed to grow on ethanol-glycerol plates. On the premise that the human targeting peptide might not function correctly in yeast, a plasmid was constructed to produce a protein chimera in which the yeast Atp11p leader is fused to an approximated form of the mature human Atp12p protein (Ile-33 through Glu-289, plasmid pG13L/HUATP12 ). The yeast atp12 mutant harboring this plasmid grows moderately well on ethanol-glycerol plates (Fig. 3). The level of oligomycin-sensitive mitochondrial ATPase activity measured for this transformant is 2.49 units/mg, which can be compared with 0.12 unit/mg measured for the aW303⌬ATP12 mutant and 4.71 units/mg measured for the respiratory competent parent, W303-1A. In contrast, pG13L/HUATP12(33-289) does not confer respiratory competence to the atp11 mutant, aW303⌬G13.
Genomic Characterization of Human ATP11 and ATP12-The ESTs belonging to the ATP11 UniGene cluster Hs.10964 (176 homologous sequences) and those of the ATP12 UniGene cluster Hs.13434 (76 homologous sequences) were derived from a wide distribution of tissues, including brain, breast, colon, heart, kidney, lung, muscle, stomach, and whole embryo. SAGE data also indicate a wide tissue distribution for these genes, although a light tag frequency within most tissues (less than 62 tags/million tags surveyed for each gene) suggests low to moderate abundance of each transcript. Several other mammalian ESTs show strong homology to ATP11 (14 mouse, 3 rat, and 1 cow) and to ATP12 (11 mouse, 11 cow, 5 rat, 5 pig, 3 Xenopus, and 1 chicken), which suggests that Atp11p and Atp12p orthologs are present also in these species.
The ATP11 cDNA sequence is collectively contained within three draft human genomic sequences (AL357392, AL136373, and AC013593). Alignment of the cDNA and genomic sequences yields nine regions of homology, which suggests a genomic structure of 9 exons. The precise exon/intron boundaries of all 9 putative exons, each of which contains consensus gt/ag splice donor/acceptor sequences, and the length of all but one intron, could be inferred from the alignment (Table II) 3. Complementation of a yeast atp12::LEU2 mutation with the human ATP12 cDNA. The following three S. cerevisiae strains were streaked on a YEPG plate and incubated for 2 days at 30°C: W303-1A, respiratory competent wild type yeast; W303⌬ATP12, respiratory-deficient mutant that harbors an atp12::LEU2 allele; W303⌬ATP12/pG13L/HUATP12 , ⌬ATP12 transformed with a multicopy plasmid that carries the human cDNA for Atp12p. YEPG plates: 1% yeast extract, 2% ethanol, 2% bactopeptone, 3% glycerol, 2% agar. genomic sequence. EST-derived marker D1S3533 (SHGC-11813) representing the 3Ј-untranslated region of ATP11 has been localized on the GeneMap '99 RH map between D1S2843 and D1S417 (48.8 -81.6 centiMorgans), and on the CompView chromosome 1 RH map between RH67834 and D1S197 (872.3-907 centiRays). The CompView-inferred cytogenetic position for marker D1S3533 is 1p32-p34, and the Unified data base estimates the marker to be 49.018 megabases proximal of the 1p telomere. In addition, BAC clone bA49P4, from which the ATP11-containing genomic sequence AL136373 is derived, has been mapped by fluorescence in situ hybridization to 1p32.2-p33. Therefore, ATP11 can be assigned to 1p and most likely localizes within 1p32.2-p33; this assignment is consistent with the localization of several other markers and genes also found within the three ATP11-containing genomic sequence tracts.
At the time of writing, no draft or finished human sequence including human ATP12 had been submitted to GenBank TM ; however, a draft mouse sequence (GenBank TM accession number AC026681) that spans most of the coding region of human ATP12 was identified by BLASTN. Seven blocks of homology between the mouse genomic sequence and nucleotides 130 -874 of human ATP12 were identified, which range in size from 51 to 149 base pairs and show 86 to 93% identity. On this basis we predict that the mouse gene for Atp12p contains at least 8 exons. An ATP12 EST-derived marker (WI-9504) has previously been localized to chromosome 17 on the GeneMap '99 and Whitehead Institute RH maps, between marker pairs D17S922 and D17S798, and AFMa126yd5 and D17S798, respectively. Several genes mapping to these intervals have been cytogenetically localized to 17p11.2, suggesting that ATP12 also maps to this region. The mouse genomic sequence AC026681 homologous to human ATP12 has been assigned to mouse chromosome 11, a portion of which is syntenic to 17p11.2. Furthermore, mouse AC026681 also contains the gene Drg2, the human ortholog of which has been localized to 17p11.2 (29). Taken together, this evidence strongly indicates that ATP12 maps to human chromosome 17p11.2.

DISCUSSION
Proteins involved in the assembly pathway for the ATP synthase are prime suspects for harboring genetic lesions responsible for deficiencies in this enzyme. This article presents the first report of assembly proteins that have been identified for the human ATP synthase. Atp11p and Atp12p were recognized originally as proteins required for correct assembly of the F 1 component of the ATP synthase in S. cerevisiae (6). Several lines of evidence indicate that we have assigned correctly the human mRNAs HRC06325 (GenBank TM accession number AK026004) and 24418 (GenBank TM accession number AF052185) as ATP11 and ATP12, respectively. First, the results of amino acid sequence alignments are excellent in both cases, with regions of homology that extend throughout the protein sequences (Figs. 1 and 2 and Table I). Of added note are specific sequence elements, such as the PXFXXXLPR consensus sequence in human Atp11p and a highly conserved acidic residue near the carboxyl terminus of human Atp12p, which strengthen the interpretation of the alignment data. Second, results from yeast two-hybrid screens support the claim that the human cDNAs for ATP11 and ATP12 isolated in our laboratory encode assembly factors for human F 1 that are fundamentally similar to their yeast counterparts. The yeast Atp11p and Atp12p proteins are known to be substrate-specific; Atp11p binds to the F 1 ␤ subunit (9), and Atp12p binds to the F 1 ␣ subunit (10). Comparable observations were made in twohybrid assays with the human proteins in that human Atp11p showed specificity for binding the human F 1 ␤ subunit, and human Atp12p was selective for binding the human F 1 ␣ subunit. Finally, the human ATP12 cDNA rescues the respiratory defect of a yeast atp12 mutant (Fig. 3), which is compelling evidence for equivalence of function.
A curious observation made during the genetic complementation experiments was that the atp12 strain is not rescued with the full-length cDNA for human Atp12p, but rather with a genetic hybrid coding for a chimera bearing the mitochondrial targeting sequence of the yeast Atp11p protein fused to the mature portion of the human Atp12p protein. The limitations that prevent the substitution of the complete human ATP12 gene for the yeast gene in our experiments are not known. Nor can we comment with certainty about the fact that the respiratory-deficient phenotype of a yeast atp11 mutant is not rescued by a plasmid that produces most of the mature human Atp11p protein fused to the yeast Atp11p targeting sequence. While it is true that the human protein encoded by our ATP11 cDNA clone is estimated to be deficient for 25-30 amino acids from the mature amino terminus, this was not considered to be of particular consequence, first because yeast Atp11p retains function following the removal of up to 72 amino acids from its mature amino terminus (24), and second because the truncated human Atp11p protein shows evidence of binding to the yeast F 1 ␤ subunit in a two-hybrid screen. On this point it is noteworthy that Atp11p of Drosophila binds the yeast F 1 ␤ subunit in a two-hybrid assay, yet the foreign protein confers only a minimum of respiratory competence to an atp11 mutant (11). Cumulatively, such results suggest there may be an aspect of Atp11p mechanism, peripheral to F 1 ␤ subunit binding, that limits the ability of non-yeast Atp11p proteins to function properly in the context of a yeast cell. For example, it is not yet known if Atp11p is regulated in some way, nor is it known if Atp11p action is dependent on some other mitochondrial protein(s).
In recent years there have been reports of ATP synthase deficiencies that are ascribed to nuclear mutations and which correlate with disease symptoms of severe lactic acidosis and cardiomyopathy (30,31). The amount of ATP synthase is also reported as decreased, relative to the amounts of the respiratory complexes of the oxidation phosphorylation pathway, in samples from patients with Alzheimer's disease (32). Of further interest is that the chromosomal regions to which human ATP11 (1p32.2-p33) and ATP12 (17p11.2) map have both been implicated in human disease. For example, a putative locus for hereditary congenital ptosis, a muscle-specific disorder that is characterized by unilateral or bilateral drooping of the upper eyelids, has been localized to 1p32-p34.1 (33). Muscle-Eye-Brain disease, which includes severe early-onset muscle weakness, congenital myopia, and mental retardation also maps to this region (1p32-p34) (34). In addition, deletion of 1p32-p33 is frequently observed in a number of malignancies, including meningioma and carcinomas of the lung, colon, and stomach (35,36). Disease loci that map to 17p11.2 include the neuropathies Charcot-Marie-Tooth type IA and Hereditary Neuropathy with Liability to Pressure Palsies (37,38), the chromosome microdeletion disorder of the Smith-Magenis syndrome (39), and frequent deletion breakpoints in medulloblastomas (40). The determination of whether Atp11p or Atp12p contribute to any of these disease phenotypes awaits further investigation.