Identification of the gene encoding the human mitochondrial RNA polymerase (h-mtRPOL) by cyberscreening of the Expressed Sequence Tags database
Identification of the gene encoding the human mitochondrial RNA polymerase (h-mtRPOL) by cyberscreening of the Expressed Sequence Tags databaseValeriaTiranti1, AnnaSavoia2, FrancescaForti1, Maria-FeliciaD'Apolito2, MartaCentra2, MarianoRocchi3 and MassimoZeviani1,4,*
1Division of Biochemistry and Genetics, National Neurological Institute `C. Besta',Milano,Italy,2Service of Medical Genetics, CSS-IRCCS, San Giovanni Rotondo,Foggia,Italy,3Institute of Medical Genetics, School of Medicine, University of Bari,Bari,Italy and4Unit of Molecular Medicine, Children's Hospital `Bambino Gesù',Rome,Italy
Received December 21, 1996;Revised and Accepted January 31, 1997
A gene cloning strategy based on the screening of the Expressed Sequence Tags database (dbEST) using sequences of mitochondrial housekeeping proteins of yeast was employed to identify the cDNA encoding the precursor of the human mitochondrial RNA polymerase (h-mtRPOL). The 3831 bp h-mtRPOL cDNA is located on chromosome 19p13.3 and encodes a protein of 1230 amino acid residues. The protein sequence shows significant homologies with sequences corresponding to mitochondrial RNA polymerases from lower eukaryotes, and to RNA polymerases from several bacteriophages. The mitochondrial RNA polymerase carries out the central activity of mitochondrial gene expression and, by providing the RNA primers for replication-initiation, is also implicated in the maintenance and propagation of the mitochondrial genome. Genes involved in the control of mtDNA replication and gene expression are attractive candidates for human disorders due to abnormalities of nucleo-mitochondrial intergenomic signalling. The availability of the h-mtRPOL cDNA will allow us to test its role in mitochondrial pathology. In addition, we propose the `cyberscreening' of dbEST, based on yeast/human cross- species comparison, as a powerful, simple, rapid and inexpensive method, that may accelerate several-fold the molecular dissection of the human mitochondrial proteome.
Mitochondria of all organisms contain a genome that is distinct from that of the nucleus. The human mitochondrial genome (mtDNA) is a maternally transmitted minichromosome present in 2-10 copies per organelle. It contains 37 genes which encode the RNA components of the mitochondrial translational apparatus, i.e. 22 transfer RNA genes and two ribosomal RNA genes, as well as 13 polypeptide-encoding genes (1 ). All 13 polypeptides are essential components of four of the five complexes that form the mitochondrial oxidative phosphorylation (OXPHOS) pathway (complexes I, III, IV and V). However, gene expression in mitochondria relies upon numerous nuclear genes, that encode protein components required for transcription and translation of the mtDNA-encoded genes, as well as protein and RNA components required for replication of mtDNA (2 ,3 ). In addition, nuclear genes encode factors controlling the import (4 ), assembly and turnover of OXPHOS complexes, and proteins acting as general regulators of mitochondrial function (5 ,6 ). Nuclear-encoded mitochondrial proteins are synthesized by cytoplasmic ribosomes, usually as precursors containing an N-terminal extension. Import into mitochondria is carried out by a complex, ATP-dependent transport system, followed by cleavage of the leader peptide, that eventually produces a mature, functional protein (4 ).
The existence of abnormalities in the nuclear gene repertoire controlling mitochondrial biogenesis has been proposed as the cause of some human disorders. These disorders are characterized by the presence of mtDNA abnormalities transmitted as Mendelian traits (7 ). For instance, autosomal dominant (8 ,9 ) or autosomal recessive (10 ) Chronic Progressive External Ophthalmoplegia (CPEO), are neuromuscular disorders due to the accumulation of multiple deletions of mtDNA in stable tissues. Another example is tissue-specific mtDNA depletion (11 ), an autosomal-recessive disease causing severe organ-specific syndromes in early infancy. Mendelian inheritance indicates the presence of transmissible mutations in nuclear genes that can ultimately damage the structural integrity of the mtDNA molecule or its copy number. Genes involved in the control of mtDNA replication and expression are considered attractive candidates for these disorders. Thus, the characterization of the collection of human proteins related to mitochondria (i.e. the human mitochondrial proteome) is a matter of convergent interest between basic scientists that are involved in the elucidation of the fundamental mechanisms of nucleo-mitochondrial intergenomic signalling, and clinical researchers, interested in mitochondrial disorders.
Figure1 reports a map of the cDNA contig constructed to obtain the nucleotide sequence of the full-length h-mtRPOL cDNA. By probing the RPO4 protein sequence against dbEST using the TBLASTN option (see Materials and Methods), highly significant smallest sum probability values (P <10-5) were obtained for four translated human ESTs. EST_T93942 and EST_T97038 appeared as two largely overlapping sequences from distinct cDNAs corresponding to clone no. 116961 and clone no. 134269, respectively, in the Bento Soares' arrayed cDNA library from human fetal liver. EST_H03562 and EST_H03471 correspond to sequences from the 5' and 3' ends, respectively, of the same clone no. 150973, in the Bento Soares' cDNA arrayed library from human placenta.
Sequence analysis on the three cDNAs obtained from the I.M.A.G.E. Consortium Clone Distributors (http://www-bio.llnl. gov/bbrp/image/idist_add.html/), followed by comparison with the yeast analog, indicated that clone no. 150973 contained a non-overlapping cDNA to the contig formed by cDNAs of clones no. 116961 and no. 134269. The latter contig encompassed the last 766 bp on the 3' end of the gene. The 186 bp gap between clone no. 134269 and clone no. 150973 was filled-in by PCR amplification from the human liver cDNA of the 5' RACE-Ready kit®, using suitable primers. A cDNA contig encompassing the h-mtRPOL mRNA sequence was then obtained by a cDNA walking strategy, using a combination of `traditional' cDNA library hybridization and further dbEST search through the BLASTN (nucleotide vs. nucleotide database) option (EST_R73297, corresponding to clone no. 156230; EST_T16125, corresponding to clone no. IB3532). A RACE (Rapid Amplification of cDNA Ends) strategy (17 ) confirmed the 5' sequence end of the contig.
Figure2 shows the nucleotide and deduced amino acid sequence of the h-mtRPOL cDNA (GenBank U75370). The corresponding transcript is 3831 nt in length. A canonical AAUAAA polyadenylation signal is located 18 nt upstream from a poly(A) tail. The cDNA contains a potentially translatable ORF, spanning from nucleotide position (np) 33, where the first AUG codon is located, to np 3722, adjacent to an `opal' termination codon (UGA). The corresponding polypeptide is 1230 amino acids in length, with a predicted molecular mass of 138682 daltons and a calculated pI = 9.22.
Sequence analysis of six clones encompassing the 5' end of the cDNA revealed the presence of a nucleotide polymorphism at np 145. A cytosine present in clones no. 156230 (from Bento Soares' human breast cDNA library) and no. IB3235 (from Bento Soares' infant brain cDNA library) was replaced by a thymine in two independent clones from a Stratagene human colon cDNA library. The nucleotide change causes P(38) to be replaced with S(38) (P-38-S).
Figure3 a shows the results of sequence alignment between the precursors of human and yeast mtRNA polymerases. The global amino acid identity is 28.4%. Similar results were obtained by comparing h-mtRPOL with the mtRPOL sequence ofNeurospora crassa andNeurospora intermedia(18 ) (not shown). In both cases, identity is remarkably higher in the carboxy-terminal half of the protein sequences (approximately from amino acid position 820 of the human sequence), where `blocks' of colinear identical sequences are recognizable (221/410, 54% identical amino acid residues). In keeping with what has already been reported for the yeast sequence (14 ), in six portions of the caroboxy-terminus, the h-mtRPOL sequence shares significant homology with that of prokaryotic RNA polymerases, such as the T3 RNA polymerase (Fig.3 b), or the T7 and Sp6 polymerases (not shown).
Figure4 b shows the results ofin vivoimport of a chimaeric construct containing the first 330 amino-terminal residues of h-mtRPOL fused on the carboxy-terminus with a strong antigenic epitope of the Influenza virus haemoagglutinin (HA), serving as an `antibody tag'. Two bands were immunoprecipitated from a lysate of COS-7 cells transfected with the recombinant vector and radiolabeled with35S-methionine. The size of the first band was identical to that of thein vitrotranslation product from the same construct (~37.6 kDa), while the second band was of smaller size (~33 kDa). These data indicate that mitochondrial import had occurred, followed by cleavage of the amino terminal leader peptide (20 ). The low intensity of the band corresponding to the imported/cleaved protein species is likely due to either low efficiency of import or rapid intramitochondrial degradation of this artifactual polypeptide. Since mitochondrial import is energy-dependent, only the higher size band corresponding to the precursor protein was detected after immunoprecipitation of transfected COS-7 cells that were made OXPHOS-defective by incubation with CCCP, a powerful OXPHOS inhibitor (Fig.4 b). Identical results were obtained using valinomycin, a ionophor that destroys the mitochondrial electrochemical gradient (not shown). The size difference between the two bands (corresponding to ~4.6 kDa) suggests that the leader peptide encompasses ~42 amino acid residues.
A typical `mitochondrial' immunofluorescence pattern, characterized by a discrete, finely granular cytoplasmic signal, corresponding to individual mitochondria, was visualizedin situon transfected COS-7 cells, using an anti-HA antibody against the HA-tagged h-mtRPOL protein fragment (Fig.5 , left). An identical cytoplasmic pattern was obtained using an antibody against a native mitochondrial protein, mtSSB (Fig.5 , right). The nuclear signal shown in Figure5 (right) is due to spontaneous autofluorescence of nuclear structures at 485 nm, since, in contrast to the cytoplasmic signal, it was detected also in untreated cells (not shown).
Figure6 shows the results of Northern-blot analysis of h-mtRPOL on polyA+ RNA extracted from several tissues. A band corresponding to a transcript of ~3.8 knt was present in all samples but steady-state expression was several-fold higher in skeletal muscle and heart, compared to the other tissues. Densitometric analysis was carried out by normalizing the intensity of the hybridization signal corresponding to h-mtRPOL to that corresponding to actin mRNA. Considering the h-mtRPOL/actin signal ratio obtained in skeletal muscle as 100%, the percentages obtained in the other tissues were as follows: heart 58, brain 32, placenta 19, lung 21, liver 27, kidney 41 and pancreas 58%. A less defined, weaker signal in the range of 5.5-6.0 knt, was detected in the lane containing polyA+ RNA from human fetal pancreas.
Our PCR-based screening of a panel of human/rodent somatic cell hybrids showed the presence of a human-specific 587 bp DNA fragment, corresponding to a portion of the 3' untranslated region of the h-mtRPOL cDNA, in 10 out of 32 hybrids. The only human chromosome in common with the 10 `positive' hybrids was chromosome 19. Conversely, no chromosome 19 was present in the 22 `negative' hybrids.
Further screening of a second panel, containing different portions of chromosome 19, allowed us to map the h-mtRPOL gene to region 19p13.3 (Fig.7 ).
Using a combination of dbEST cyberscreening and traditional cDNA library screening we constructed a cDNA contig of 3831 nt, containing a potentially translatable continuous ORF of 3690 nt. The predicted protein, consisting of 1230 amino acid residues, was identified as the bona fide precursor of the human mitochondrial RNA polymerase (h-mtRPOL), the central component of the mitochondrial transcriptional apparatus. This conclusion is based on the following evidence.
First, the identification of the first portions of the human cDNA contig was the result of a dbEST search using the yeast mitochondrial RPOL protein sequence as a probe.
Second, the predicted human polypeptide shares significant identity and similarity with the yeast mtRPOL, especially in the carboxy-terminal half of the sequences.
Third, the amino terminus of the human protein sequence has the typical features of a mitochondrial leader peptide, suggesting that the protein is directed to and imported into mitochondria.
Fourth, energy-dependent mitochondrial targeting and cleavage of the leader peptide were demonstrated by experiments of mitochondrial importin vivoand immunofluorescencein situon COS-7 cells expressing a chimaeric recombinant protein containing the first 330 amino acid residues of the h-mtRPOL precursor.
The C145T transition polymorphism detected in 2/4 cDNAs produced a proline to serine change in the amino-terminal portion of the h-mtRPOL protein (P-38-S). However, based on sequence analysis using both MitoProt and PSORT, it seems unlikely that the P-38-S polymorphism can modify significantly the import efficiency of the protein.
Northern-blot analysis revealed the presence of a ubiquitously expressed human transcript whose size corresponded to that of our cDNA. However, different levels of expression were observed in different tissues. The signal intensity of the specific mtRPOL band, compared to that of [beta]-actin, is high in skeletal muscle, heart, and pancreas, intermediate in kidney, brain, and liver, and low in placenta and lung. Interestingly, the same or a very similar expression pattern has been observed in other mitochondrially-directed proteins, such as subunit IV of cytochromec oxidase (21 ), the mitochondrial Translation-Elongation factor (22 ), and Translation-Initiation factor 2 (23 ). The nature of a second weaker signal of higher size, detected on polyA+ RNA from human pancreas remains to be established. Its `fuzzy' appearance indeed suggests that it might represent a hybridization artifact. Alternatively, it could correspond to a precursor transcript of the mature h-mtRPOL RNA, or to a transcript expressed by a gene homologous to h-mtRPOL. However, a BLASTN search using our cDNA nucleotide sequence against the GenBank database failed to identify any homologous known gene.
Previous studies have established that mtDNA transcription in humans requires the presence of at least two components, each responsible for specific and distinct functional activities (2 ). The first component is a non-specific RNA polymerase that retains the core catalytic activity of the mitochondrial transcriptional apparatus. In the yeastS.cerevisiae, the enzyme is encoded by a single gene, RPO4, identified several years ago. However, the attempts to clone the ortholog gene in humans have consistently failed, mainly because of difficulty in protein purification. A second component, the mitochondrial transcription factor A (mtTFA), targets mtRPOL to the transcriptional promoters in the mtDNA D-loop. The gene encoding human mtTFA has been cloned (24 ), and subsequently mapped to chromosome 10q21 (25 ). An additional transcriptional control mechanism is carried out by a 34 kDa gene product, the mitochondrial termination factor, mTERF apparently absent in yeast. mTERF causes early H-strand transcription termination by binding to a promoter-independent bidirectional termination site at the 16S rRNA-tRNALeu(UUR) gene boundary (26 ). The mTERF gene has also been cloned in humans (27 ).
A number of protein products involved in mtDNA replication, transcription and translation have been identified and partially purified in vertebrates. However, the cDNAs coding for only a few proteins involved in mtDNA biogenetic functions have been cloned and characterized in humans or mammals. These include the cDNAs specifying mtTFA, mtSSB (mitochondrial Single-Stranded DNA Binding Protein) (38 ), and a new mitochondrial endonuclease, Endonuclease G, similar to the yeast Nuc1 endonuclease (25 ,39 ). Very recently, the cDNAs encoding two additional important factors, the mitochondrial transcription terminator (mTERF) and the [gamma]-DNA polymerase, have been found in humans (40 ). The identification of these genes has required a great deal of time consuming, very demanding, and costly work. We have shown that computer-based cross-species comparison between yeast and human databases may represent an effective, rapid, low-cost and simple alternative for the identification of important mitochondrial housekeeping genes. Table1 reports a number of yeast proteins involved in mitochondrial biogenesis, for which we could identify human ESTs displaying significant homology (typically,P <10-10) using the TBLASTN option. A similar approach has recently been used to identify numerous human ESTs homologous to disease-genes known inDrosophila melanogaster(41 ). One distinct advantage of this method is that it is based on protein `cyberprobes' instead of nucleotide probes used in traditional screening of cDNA libraries. Since protein sequences are less divergent than the corresponding nucleotide sequences, the chance to find significant homology is substantially increased even between evolutionarily distant organisms. In particular, many proteins of mitochondria are likely to be conserved throughout evolution, given the essential role of the mitochondrial energy metabolism in all the eukaryotic lineages.
Other advantages of the cyberscreening approach are the rapidity and reliability of the results, due to the availability of powerful computational methods. Considering the list shown in Table1 , the effectiveness of this strategy is further corroborated by the observation that we were able to identify a number of human ESTs homologous to yeast genes, whose corresponding human orthologs have already been cloned. Examples are MIP1, the gene encoding the yeast [gamma]-DNA polymerase; NUC1, the analog to human EndoG; IFM1, the gene encoding translation factor 2 (23 ); and OXA1, a factor controlling the function of both cytochromec oxidase (complex IV) and mitochondrial ATP synthase (complex V) (42 ). Needless to say, however, the success rate of this strategy is not 100%. For instance, the protein sequences of ABF2, the yeast gene analog to human mtTFA (35 ) (GenBank Q00059), and RIM1, the yeast analog to human mtSSB (GeneBank M94556), failed to recognize ESTs corresponding to the genes, or the genes themselves.
The rapid expansion of EST databases and other genetic tools linked to the human genome project are expected to increase the power and effectiveness of cloning strategies such as that outlined here, to accelerate the molecular dissection of the human mitochondrial proteome, an important goal for both basic and clinical research.
The Yeast Protein Database (YPD) world wide web site (http//:www.proteome.com/YPDhome.html) was used to retrieve both the predicted protein sequence deduced from the RPO4 ORF (yeast mitochondrial RNA polymerase gene) (14 ), as well as a number of other mitochondrion-related yeast protein sequences (see Table1 ) using a `text string' interface (e.g. keyword = `*mitochondrial*'). The TBLASTN option (protein sequence vs. 6-frame nucleotide translated database sequences) of BLAST, Basic Local Alignment Search Tool, was used for the primary dbEST searching at the TIGEMnet world wide web site (http//:www.tigem.it). Other EST components of our cDNA contig were identified by screening of dbEST with nucleotide sequences obtained from both cDNA library and previous dbEST screenings, through the BLASTN option (nucleotide sequence vs. nucleotide sequence database).
Primary library screenings were performed each time by plating ~5 × 105 p.f.u. from both an oligo(dT) cDNA [lambda]-ZAP library from the human CaCo2 (Carcinoma-Colon) cell line (kindly provided by Dr J. Rommens), and a randomly primed [lambda]-ZAP cDNA library from a human colon adenocarcinoma cell line (COLO 205, Stratagene). The first library was used to identify cDNAs on the 3' end of the cDNA contig, while the second was used to identify cDNAs in the mid-portion and on the 5' end of the cDNA contig. Hybridization of two filter replicas/plate was carried out using Genecleaned-purified probes that were radiolabeled by random priming. Duplicate plaques were picked and suspended in SM buffer overnight. Eluates were replated and plaques rescreened until isolation of single positive plaques. The corresponding cDNAs were sequenced on an ABI 737A automated DNA sequencer by the dye-terminator protocol of the DNA Sequencing kit® (Perkin-Elmer) using either vector primers or primers corresponding to specific cDNA sequences.
The 5'-RACE (Rapid Amplification of cDNA Ends) Ready kit® (Clontech) was used to obtain the 5' end of the h-mtRPOL cDNA, following the manufacturer's protocol.
The first 929 bp on the 5' side of h-mtRPOL cDNA, corresponding to the first 299 amino-terminal residues of the protein, were tagged on the 3' end with the sequence encoding a strong antigenic epitope of the Influenza virus haemoagglutinin (HA) (43 ). The predicted chimaeric protein has an expected MW of ~37 kDa. The construct was inserted into the eukaryotic plasmid vector pcDNA3 (Invitrogen). The recombinant plasmid was transfected transiently in COS-7 cells by electroporation (250 V, 125 µFarad in 250 µl of cell suspension). Forin vivomitochondrial targeting, total cellular proteins were radiolabeled with35S-methionine for 2 h, in the presence or absence of either 18 µM valinomycine or 30 µM CCCP (carbonyl cyanidem-chlorophenylhydrazone). The specific translation products were immunoprecipitated using an anti-HA specific monoclonal antibody (Boehringer-Mannheim) in the presence ofStaphylococcus aureuslysate (Staph-A) and electrophoresed through a 12% SDS-polyacrylamide gel. After fixation in 10% acetic acid, 25% isopropanol, the gel was washed for 15 min in Amplify® (Amersham), dried and autoradiographed overnight onto Hyperfilm® (Amersham).
Forin situmitochondrial localization, immunofluorescence on transfected COS-7 cells was carried out using the mouse anti HA monoclonal antibody (dilution 1:100) and a rabbit polyclonal antibody (dilution 1:200) against the human mitochondrial single-stranded DNA binding protein (mtSSB). Cells grown on coverslips were washed twice with PBS-Mg-CaCl2, fixed in 50 and 100% methanol (1 min each), and air-dried. Cells were incubated for 30 min at room temperature (r.t.) with goat serum diluted 1:6 in PBS. Cells were then incubated for 90 min at r.t. with the primary antibody diluted in goat serum-PBS and washed three times with PBS. To visualize the anti-HA immune reaction, cells were incubated for 60 min at r.t. with a 1:250 PBS dilution of a goat biotinylated anti-mouse IgG antibody (Amersham), washed thrice in PBS, and incubated again for 60 min at r.t. with a 1:250 PBS dilution of Avidin-Rhodamine (Amersham). To visualize the anti-mtSSB immune reaction coverslips were washed three times in PBS, incubated for 60 min at r.t. with a 1:50 PBS dilution of a goat fluorescein-conjugated anti-rabbit antibody. After three washings in PBS, coverslips were mounted and visualized under a Zeiss light microscope using a 546 nm filter for rhodamine and 485 nm filter for fluorescein.
A premade multiple tissue Northern blot (Clontech) containing 2 µg/lane of polyA+ RNA from eight human tissues, was purchased from Clontech (Palo Alto, CA). The filter was preincubated in RNA hybridization solution (5* SSPE, 10* Denhardt's solution, 100 mg/ml denaturated salmon sperm DNA and 2% SDS) for 6 h at 65oC and and hybridized for 20 h at 65oC with a32P-radiolabeled probe obtained by using the Random Primer kit (Boehringer-Mannheim) on a 971 bp PCR DNA fragment spanning from np 2548 to np 3519 of the h-mtRPOL. The blot was washed at 62oC with 2* SSC, 0.01% SDS; 0.5* SSC, 0.01% SDS; and 0.1* SSC, 0.5% SDS (20 min in each solution). Autoradiography was performed for 2 days at -80oC using a Kodak XAR-5 film. The same blot was re-hybridized with a [beta]-actin probe to monitor the quality and quantity of RNA samples as described above, with the exception that autoradiographic exposure was overnight. Densitometric analysis of the autoradiographic bands was carried out using a Mitsubishi Video Copy Processor (Model P68E), an IV-530 Contour Synthesizer (FOR-A Co. Limited, France) and the Bio-Profil (Vilber-Lourmat, France) software package.
A 587 bp PCR DNA fragment encompassing nt 2844-3431 in the 3' untranslated region of h-mRNAP cDNA could be directly amplified from total human genomic DNA, suggesting that this region was contained within a single exon of the corresponding gene. A pair of 25mer primers were used in a 30 cycle PCR reaction; after 3 min of denaturation at 95oC, the temperature profile was 95oC-65oC-72oC, each for 1 min. The fragment was human-specific, as it could not be PCR-amplified from hamster genomic DNA. Hence, the presence or absence of the PCR fragment was used to identify the chromosomal localization of the gene by testing a panel of human-hamster somatic cell hybrids (44 ). The chromosomal sub-localization was then obtained using the same strategy on a second panel containing different portions of chromosome 19.
Sequence comparison between yeast and human mtRPOL polypeptides, and human mtRPOL and the T7 RPOL polypeptides were carried out using the Clustal software package at the BCM web site (http://kiwi.imgen.bcm.tmc.edu). Analysis of the h-mRPOL protein sequence for the identification of mitochondrial signal peptide was performed using the MitoProt v. 1.0.1, developed by M-G Claros (claros@ccuma.sci.uma.es) and PSORT (http://psort.nibb.ac.jp/form.html) software packages.
We are indebted to Dr J. Rommens for kindly providing the cDNA library from the CaCo2 human cell line, Dr R. Barresi for useful technical advice, Dr P. Fernandez-Silva for comments and discussion, and Ms B. Geehan for revising the manuscript. We gratefully acknowledge the Telethon Institute of Genetic Medicine for computer assistance, and the I.M.A.G.E. consortium for the gift of recombinant plasmids from the Bento Soares' infant brain arrayed cDNA library. Supported by Telethon-Italy (grant n. 767 to M.Z.) and EU Human Capital and Mobility network grant to M.Z. `Mitochondrial Biogenesis in Development and Disease'.
7 Zeviani, M., Amati, P., Comi, G., Fratta, G., Mariotti, C., Tiranti, V. (1995) Biochim. Biophys. Acta, 1271, 153-158.MEDLINE Abstract
8 Zeviani, M., Servidei, S., Gellera, C., Bertini, E., DiMauro, S., DiDonato, S. (1989) Nature, 339, 309-311.MEDLINE Abstract
9 Zeviani, M., Bresolin, N., Gellera, C., Bordoni, A., Pannacci, M., Amati, P., Moggio, M., Servidei, S., Scarlato, G., DiDonato, S. (1990) Am. J. Hum. Genet., 47, 904-914.MEDLINE Abstract
10 Bohlega, S., Tanji, K., Santorelli, F.M., Hirano, M., al-Jishi, A., DiMauro, S. (1996) Neurology, 46, 1329-1334.MEDLINE Abstract
11 Moraes, C.T., Shanske, S., Tritschler, H.J., Aprille, J.R., Andreetta, F., Bonilla, E., et al. (1991) Am. J. Hum. Genet., 48, 492-501.MEDLINE Abstract
33 Morris, C.E., McGraw, N.J., Joho, K., Brown, J.E., Klement, J.F., Ling, M.L., McAllister, W.T. (1987) In W.S. Reznikoff, R.R. Burgess, J.E. Dahlberg, C.A. Gross, M.T. Record and M.P. Wickens, eds. `RNA polymerase and regulation of transcription', New York Elsevier Publisher, pp. 47-58.
34 Ikeda, R.A. and Richardson, C.C. (1987) J. Biol. Chem., 262, 3790-3798.MEDLINE Abstract
36 Soumalainen, A., Kaukonen, J., Amati, P., Timonen, R., Haltia, M., Weissenbach, J., Zeviani, M., Somer, H., Peltonen, L. (1995) Nature Genet., 9, 146-151.
37 Kaukonen, J.A., Amati, P., Soumalainen, A., Rötig, A., Piscaglia, M-G., Salvi, F., Weissenbach, J., Fratta, G., Comi, G., Peltonen, L., Zeviani, M. (1996) Am. J. Hum. Genet., 58, 763-769.MEDLINE Abstract
38 Tiranti, V., Rocchi, M., DiDonato, S., Zeviani, M. (1993) Gene, 126, 219-225.MEDLINE Abstract
39 Coté, J. and Ruiz-Carrillo, A. (1993) Science, 261, 765-769.MEDLINE Abstract
41 Banfi, S., Borsani, G., Rossi, E., Bernard, L., Guffanti, A., Rubboli, F., Marchitiello, A., Giglio, S., Coluccia, E., Zollo, M., Zuffardi, O., Ballabio, A. (1996) Nature Genet., 13, 167-174.MEDLINE Abstract
42 Bonnefoy, N., Kermorgant, M., Groudinsky, O., Minet, M., Slonimski, P.P., Dujardin, G. (1994) Proc. Natl. Acad. Sci. USA, 91, 11978-11982.MEDLINE Abstract
44 Antonacci, R., Marzella, R., Finelli, P., Lonoce, A., Forabosco, A., Archidiacono, N., Rocchi, M. (1995) Cytogenet. Cell. Genet., 68, 25-32.MEDLINE Abstract
*To whom correspondence should be addressed at: Division of Biochemistry and Genetics, Istituto Nazionale Neurologico `C. Besta' via Celoria 11, 2013 Milano, Italy
This page is maintained by OUP admin. Last updated Thu Mar 13 17:29:21 GMT 1997. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996
