Molecular bases for human complement C7 polymorphisms,C7*3 and C7*4

Complement C7 is one of the components of membrane attack complex (MAC) generated by the terminal complement cascade. C7 protein is polymorphic and most of its polymorphisms have been identified using isoelectric focusing (IEF), which detects protein charge differences. To date, the molecular bases of the polymorphisms detected by IEF have not been determined. In this paper, we describe the structural bases of two C7 IEF-detected polymorphisms, C7*3 and C7*4, both of which are common in Asian populations. C7*3 resulted from substitution of cysteine (Cys) at amino acid residue 106 by charged arginine (Arg; C106R), while charged lysine (Lys) at amino acid residue 398 was replaced by neutral glutamine (Gln; K398Q) in C7*4. As C7*3 is hypomorphic, it is important to study its possible associations with diseases such as immunological disorders and infections. We present genetic bases for this C7 polymorphism, which we determined using polymerase chain reaction (PCR)-based genotyping, a simple and accurate method suitable for large-scale studies.     

C7 polymorphism in Japanese: new allele C7*8

The polymorphism of C7 was investigated in neuraminidase-treated sera from 513 unrelated Japanese individuals using isoelectric focusing followed by an electroimmunoblotting technique. Besides the common phenotypes 5 rare variants including 2 types of new variants were detected. The family analysis suggested the hereditary occurrence of a new allele C7*8. The allele frequencies were: C7*1 = 0.8314, C7*2 = 0.0926, C7*3 = 0.0380, C7*4 = 0.0331, C7*6 = 0.0010, and C7*8 = 0.0039.     

Human C4 polymorphism: Pedigree analysis of qualitative,quantiative, and functional parameters as a basis for phenotype interpretations

Summary Ten families with 82 members were investigated for C4A- and B polymorphism in a blind trial. Phenotyping was done on neuraminidase treated sera by immunofixation and simulataneously by hemolytic overlay electrophoresis. In addition Rg, Ch, BF, C2, HLA-A, B, C, DR, and GLO were determined. After decoding the samples the reliability of blind typing was found to be 84.4% according to segregation patters. Inconsistencies occurred mostly when A 4, A 2, or A 92 were present. The detection of silent A*Q0 and B*Q0 alleles was more critical than that of difficult allotypes. The quantitation of the C4A/B ratio by densitometry of stained gels or by conventional immunochemical measurements of serum C4 level could not substantially improve the identification of A*Q0 or B*Q0. C4 dependent activity in radial diffusion hemolysis showed satisfactory correspondence with the number of expressed C4B alleles. At least three haplotypes with two C4A genes (duplicated A genes) were observed as ascertained from offspring analysis in accordance with the MHC segregation pattern. Individuals with the duplicated C4A gene (C4A*3. A*2. in the absence of any other expressed A allele or together with C4A*92) showed only partial inhibition of Rodgers antisera. Partial inhibition of Chido antisera was seen in individuals with C4B 2 (in the absence of other B allotypes). The findings support the hypothesis of at least two structural C4 loci. The also demonstrate the inconsistency of quantitative data in the recognition of silent alleles.     

Polymorphism of the fourth component of complement in Turks

An analysis of polymorphism in the fourth component of human complement (C4) was performed on EDTA-plasma from 142 unrelated, randomly selected Turks without collagen-vascular disease or recurrent infections. Plasma samples treated with neuraminidase and carboxypeptidase-B were subjected to high-voltage agarose gel electrophoresis followed by immunofixation. C4B allotypes were further detected in some samples by Western blots with monoclonal antibody 1228 (anti-C4B/Ch1 reactivity). The frequencies of C4A and C4B alleles were determined. Allele C4B*5, which has been found to be relatively common in Asian (Oriental) populations, was not detected in this study. No specific predilection could be noted among the rare variants. C4A*3-C4B*1 was the most common haplotype (n = 40/142, or 28%) but was found less frequently than in Caucasian populations. This finding may be the result of the limited number of samples examined. C4A and/or C4B null allotypes were seen in 49 of 142 (34.6%) subjects. The most frequent C4 null allotype seen was C4B null (37/142, or 26%): 28 subjects had one C4B null allele; 1 had a homozygous deficiency of C4B (C4B*QO, *QO) and 7 had C4A*QO C4B*QO, a double heterozygous haplotype. Frequencies of homozygous haplotype C4A*Q0-C4B*Q0 in the population studied were found to be 0.007. The results of this study demonstrate that the genetic composition of the Turkish population exhibits both similarities and differences with the European population, and ranges between Caucasian and Mongoloid (Asian) populations.     

Analysis of active and inactive complement C4 complotypes associated with subtypes of HLA-B17 in different racial groups.

HLA-A, B, and C antigens and the HLA-linked markers BF, C2, and C4 were determined in 1,799 unrelated Caucasians, 140 North American blacks, 140 Chinese, and 66 Japanese. One allele of the C4A locus (Rodgers), C4A*6, was found to code for a functionally inactive product in HLA-B17 (Bw57)-positive individuals, but for a functionally active product in HLA-B37- or HLA-B27-positive individuals. Further studies revealed that the functionally inactive C4A*6 gene product was found only in one subtype of HLA-B17, namely, 17.1 (Bw57, long). In addition, it was found that the frequency of the other subtype of HLA-B17, namely, 17.2(Bw58, short) varies greatly in different populations and that HLA-Bw58 is associated with either C4A*3,B*1 or C4A*3,B*Q0.     

Heterogeneity in the structural basis of the human complement C4A null allele (C4A Q0) as revealed by HindIII restriction fragment length polymorphism analysis

The highly polymorphic fourth component of human complement (C4) is usually encoded by two genes. C4A and C4B, adjacent to the 21-hydroxylase (21-OH) genes, 21-OHA and 21-OHB, and is also remarkable in the high frequency of the 'null' alleles, C4A Q0 and C4B Q0. The molecular basis for the C4A Q0 allele was studied in 26 families through restriction fragment length polymorphism (RFLP) analysis with C4 and 21-OH cDNA probes after digestion of the DNA with the endonuclease HindIII. The individuals expressing the extended haplotype HLA-A1 (of A2) Cw7 B8 C2C BfS C4AQ0B1 DR3 have a large deletion taking off the C4A and 21-OHA genes.     

Aberrant expression of HLA-B*3565Q is associated with a disrupted disulfide bond

The identification of expression variants is a challenge in HLA diagnostics. We here describe the identification of the novel allele HLA-B*3565Q. The serological HLA class I type, as determined by a lymphocytotoxicity test, was A11,24; B38; Bw4; Cw−; whereas PCR-sequence-specific primers resulted in A*11,*24, B*35,*38; Cw*12, thus suggesting the presence of a nonexpressed B*35 allele. To clarify the lack of serological HLA-B35 reactivity, exons 2 and 3 were sequenced following haplotype-specific amplification. At position 564 from the beginning of the coding region (exon 3), a transversion (C→G) was observed, which, at the amino acid level, results in a substitution from cysteine to tryptophane at position 164 of the mature polypeptide. Because this position is essential for the formation of a disulfide bond linking the cysteine residues at positions 101 and 164, which is strongly conserved in functional class I molecules of vertebrates, the disruption of this bond is very likely to be the reason for the lack of serological detectability. We later found the same novel allele in a second unrelated individual, of whom we were able to establish a lymphoblastoid cell line (B-LCL). Serological testing of this B-LCL indicated a very low aberrant expression of HLA-B*3565Q, which cannot be expected to be detected by standard serology techniques. Holger-Andreas Elsner and Peter A. Horn contributed equally to this work. The nucleotide sequence data reported in this paper have been published in the European Molecular Biology Laboratory, GenBank, and DNA Data Bank of Japan Nucleotide Sequence Database under the accession numbers AJ278746, AJ278747, and AJ879892. The name B*3565Q was officially assigned by the WHO Nomenclature Committee in December 2005. This follows the agreed policy that, subject to the conditions stated in the most recent Nomenclature Report (Marsh et al. 2005), names will be assigned to new sequences as they are identified. Lists of such new names will be published in the following WHO Nomenclature Report.     

Complement C6 and C7 polymorphisms in Japanese patients with chronic glomerulonephritis

C6 and C7 types were studied in 158 Japanese patients with different types of chronic glomerulonephritis: 75 patients with IgA nephropathy (IgA-N); 49 patients with idiopathic membranous nephropathy (IMN), and 34 patients with minimal-change nephrotic syndrome (MCNS). There were significant differences in the C6 and C7 allele and phenotype frequencies between the patient groups and controls. A strong association was found between IgA-N and C7 5 phenotype (p less than 0.001, RR = 12.71), and between MCNS and C7 5 phenotype (p less than 0.001, RR = 14.20). A significant association between MCNS and C6 B2 phenotype (p less than 0.05, RR = 2.42) was also found. In the IMN patient group, a significant association with C7 4 phenotype (p less than 0.05, RR = 2.42) was observed. Thus, C6 and C7 phenotypes may be causative factors in the development of chronic glomerulonephritis.     

Genetic control of C6 polymorphism and C6 deficiency in rabbits

The genetic control of the sixth component of complement (C6) in rabbits has been studied by quantitation of C6 functional and antigenic levels and identification of polymorphism by isoelectric focusing (IEF) in gels. Patterns of inheritance of C6 variants in families carrying a silent gene for C6 were examined, and it was found that 3 common plasma phenotypic variants, C6 A, C6 B, and C6 QO were under the genetic control of allelic genes, C6*A, C6*B, and C6*QO. In IEF patterns, C6 A could be identified by its isoelectric point that was slightly more acidic than that of C6 B. C6 QO was undetectable because it lacked functional and antigenic activity. The C6*A/C6*B genotype displayed a mixed IEF pattern with bands characteristic of both C6 A and C6 B. Functional and antigenic levels of C6 that were found in heterozygous C6*A/C6*QO and C6*B/C6*QO rabbits were approximately one-half of the C6 levels found in the corresponding homozygous animals. The phenotypic variation closely resembles that previously observed in humans and rhesus monkeys, as well as preliminary data in rabbits. The patterns of inheritance indicated that the two common C6 structural genes and the deficiency gene were allelic variants at the same genetic locus.     

Brief communication: Restricted geographic distribution for Y‐Q* paragroup in South America

We analyzed 21 paragroup Q* Y chromosomes from South American aboriginal and urban populations. Our aims were to evaluate the phylogenetic status, geographic distribution, and genetic diversity in these groups of chromosomes and compare the degree of genetic variation in relation to Q1a3a haplotypes. All Q* chromosomes from our series and five samples from North American Q* presented the derivate state for M346, that is present upstream to M3, and determined Q1a3* paragroup. We found a restrictive geographic distribution and low frequency of Q1a3* in South America. We assumed that this low frequency could be reflecting extreme drift effects. However, several estimates of gene diversity do not support the existence of a severe bottleneck. The mean haplotype diversity expected was similar to that for South American Q1a3* and Q1a3a (0.478 and 0.501, respectively). The analysis of previous reports from other research groups and this study shows the highest frequencies of Q* for the West Corner and the Grand Chaco regions of South America. At present, there is no information on whether the phylogenetic status of Q* paragoup described in previous reports is similar to that of Q1a3* paragroup though our results support this possibility. Am J Phys Anthropol, 2009. © 2009 Wiley‐Liss, Inc.     

Human C4 haplotypes with duplicated C4A or C4B

In the course of study of families for the sixth chromosome markers HLA-A, C, B, D/DR, BF, and C2, the two loci for C4, C4A, and C4B, and glyoxalase I, we encountered five examples of probable duplication of one or the other of the two loci for C4. In one of these, both parents and one sib expressed two different structural genes for C4B, one sib expressed one, and one sib expressed none, suggesting that two C4B alleles were carried on a single haplotype: HLA-A2, B7, DR3, BFS1, C2C, C4A2, C4B1, C4B2, GLO1. In a second case, two siblings inherited C4B*1 and C4B*2 from one parent and C4B*Q0 from the other. This duplication appeared on the chromosome as HLA-AW33, B14, DR1, BFS, C2C, C4A2, C4B1, C4B2, GLO2. In a third, very large family with 3 generations, a duplication of the C4B locus occurred which was followed in 2 generations. In one individual, there were three C4B alleles and two C4A alleles. One of the C4B alleles had a hemolytically active product with electrophoretic mobility near C4B2 and was designated C4B*22. It segregated with C4B1 in the family studied. The complete haplotype was HLA-A11, CW1, BW56, DR5, BFS, C2C, C4A3, C4B22, C4B1, GLO2. In another family with 12 siblings, one parent and eight children expressed two C4A alleles on the haplotype HLA-AW30, BW38, DR1, BFF, C2C, C4A3, C4A2, C4BQ0, GLO1.(ABSTRACT TRUNCATED AT 250 WORDS)     

A marked drop in the incidence of the null allele of the B gene of the fourth component of complement (C4B*Q0) in elderly subjects: C4B*Q0 as a probable negative selection factor for survival

Summary The incidence of allotypes of the genes of the fourth component (C4) and factor B of the complement system was compared in 252 persons under 45 years of fage (young group) with 482 people between 61 and 90 years of age (old group). One hundred people older than 90 years of age (nonagenarians) were also investigated. A striking difference was found between the young and old groups in the incidence (16.1% and 5.4%, respectively) of a silent gene of the C4B allele (C4BQ0). This difference was even more marked among young and old men (17.6% vs 3.4%). The incidence of the C4BQ0 allele in women dropped to the level of the men only in the nonagenarian group. The most probable explanation for this finding is that people carrying the C4BQO allele die from as yet unidentified disease(s) in their middle-age. Therefore, male (and to a lesser extent female) carriers of this allele may have a considerably shorter life expectancy than individuals without a silent gene in the C4B locus.     

Identification of a compound heterozygote for adenine phosphoribosyltransferase deficiency (APRT*J/APRT*Q0) leading to 2,8-dihydroxyadenine urolithiasis

Summary Homozygous deficiency of a purine salvage enzyme, adenine phosphoribosyltransferase (APRT), causes urolithiasis and renal failure. There are two known types of homozygous APRT deficiencies; type I patients completely lack APRT activity while type II patients only partially lack such activity. All type II patients possess at lest one APRT^*J allele with a substitution from ATG (Met) to ACG (Thr) at codon 136. Type I patients are considered to possess two alleles (APRT^*Q0) both of which code for complete deficiencies. Thus, some patients with type II APRT deficiencies may have a genotype of APRT^*J/APRT^*Q0. As no individuals with such a genotype have previously been identified, we performed extensive analysis on four members of a family by (1) the T-cell method for the identification of a homozygote, (2) the B-cell method for the identification of heterozygotes, and (3) oligonucleotide hybridization after in vitro amplification of a part of genomic APRT sequence for the identification of APRT^*J and nonAPRT^*J alleles. We report here the first evidence that 2,8-dihydroxyadenine urolithiasis developed in a boy aged 2 years with a genotype of APRT^*J/APRT^*Q0.     

Distribution of the modified nucleoside Q and its derivatives in animal and plant transfer RNA''s.

The modified nucleoside, 7-(4,5-cis-dihydroxy-1-cyclopenten-3-yl-aminomethyl)-7-deazaguanosine, designated as Q, and its derivative, Q*, were found in tRNA's from various organisms, including several mammalian tissues, other animals such as starfish, lingula and hagfish, and wheat germ. Q isolated from rat liver tRNA was found to be identical with E. coli Q by mass spectrometry and thin-layer chromatography. Thus the rare modified nucleoside Q originally isolated from E. coli tRNA, is widely distributed in various organisms. Analysis of the mass spectrum of Q* suggested that it has a different side chain from Q.     

L-nucleotides and 8-methylguanine of d(C1m8G2C3G4C5LG6LC7G8C9G10)2 act cooperatively to promote a left-handed helix under physiological salt conditions

The structure and thermal stability of a hetero chiral decaoligodeoxyribonucleotide duplex d(C1m8 G2C3G4C5LG6LC7G8C9G10)d(C11m8G12C13G14C15LG16LC17G18C19G20) (O1) with two contiguous pairs of enantiomeric 2'-deoxy-L-ribonucleotides (C5LG6L/C15LG16L) at its centre and an 8-methylguanine at position 2/12 was analysed by circular dichroism, NMR and molecular modelling. O1 resolves in a left-handed helical structure already at low salt concentration (0.1 M NaCl). The central L2-sugar portion assumes a B* left-handed conformation (mirror-image of right-handed B-DNA) while its flanking D4-sugar portions adopt the known Z left-handed conformation. The resulting Z4-B2*-Z4 structure (left-handed helix) is the reverse of that of B4-Z2*-B4 (right-handed helix) displayed by the nearly related decaoligodeoxyribonucleotide d(mC1G2mC3G4C5L G6LmC7G8mC9G10)2, at the same low salt concentration (0.1 M NaCl). In the same experimental conditions, d(C1m8G2C3G4C5G6C7G8C9G10)2 (O2), the stereoregular version of O1, resolves into a right-handed B-DNA helix. Thus, both the 8-methylguanine and the enantiomeric step CLpGL at the centre of the molecule are needed to induce left-handed helicity. Remarkably, in the various heterochiral decaoligodeoxyribonucleotides so far analysed by us, when the central CLpGL adopts the B* (respectively Z*) conformation, then the adjacent steps automatically resolves in the Z (respectively B) conformation. This allows a good optimisation of the base-base stackings and base-sugar van der Waals interactions at the ZB*/B*Z (respectively BZ*/Z*B) junctions so that the Z4-B2*-Z4 (respectively B4-Z2*-B4) helix displays a Tm (approximately 65 degrees C) that is only 5 degrees C lower than the one of its homochiral counterpart. Here we anticipate that a large variety of DNA helices can be generated at low salt concentration by manipulating internal factors such as sugar configuration, duplex length, nucleotide composition and base methylation. These helices can constitute powerful tools for structural and biological investigations, especially as they can be used in physiological conditions.     

Homozygous CYP2B6 *6 (Q172H and K262R) correlates with high plasma efavirenz concentrations in HIV-1 patients treated with standard efavirenz-containing regimens

Efavirenz (EFV) is metabolized by cytochrome P450 2B6 (CYP2B6) in the liver. We analyzed the genotypes of CYP2B6 and their contribution to plasma EFV concentrations in 35 EFV-treated patients in International Medical Center of Japan. The mean plasma EFV concentration of patients with CYP2B6 *6/*6 (Q172H and K262R) (25.4+/-7.5 microM, +/-SD, n = 2) was significantly higher than that of patients with genotypes *6 heterozygote (9.9+/-3.3 microM, n = 10) or without alleles *6 (8.0+/-2.6 microM, n = 23) (p < 0.0001). To confirm our result, we further analyzed nine patients (three with high EFV concentrations and arbitrarily selected six with normal EFV concentrations) treated in Osaka National Hospital, and it resulted that the only three patients with the high concentrations were the *6/*6 holder. EFV dose could be decreased in those patients harboring the genotype to reduce toxicity with compromising potency, representing the first step of the Tailor-Made therapy of HIV-1 infection.     

An in vivo pilot study characterizing the new CYP2A6*7, *8, and *10 alleles.

We developed genotyping assays for CYP2A6*7 (Ile471Thr) and CYP2A6*8 (Arg485Leu). We found higher allelic frequencies in Japanese and Chinese versus Caucasians and identified an allele in which both substitutions occur together (CYP2A6*10). We created a homology model for predicting the impact of allelic variants on enzymatic activity and subsequently tested this in vivo in a pilot kinetic study. Consistent with our homology model predictions, we found (i) that CYP2A6*7 produces an enzyme that has decreased (not inactive) activity for metabolizing nicotine and coumarin; (ii) that CYP2A6*8 is unlikely to affect catalytic activity in vivo; and (iii) that having both substitutions together on an allele (CYP2A6*10) dramatically reduces function and may be fully inactive for some substrates. In conclusion, this study identifies, at relatively high frequency in Asians, an allele with decreased activity (may be substrate selective), a fully functional allele, and an allele containing both substitutions in which function is dramatically reduced.     

Human Q and C chromosomal variations: distribution and incidence.

Chromosome preparations of 77 normal newborn babies from Grand Junction, Colorado, were stained first for G-band identification of each chromosome and subsequently stained for Q- and C-band localization. This approach permitted determination of the variation of the C region size in all chromosomes, and is the first such study reported. A total of 391 Q and C variants was described, an average of 5.08 +/- 0.23 per subject; 225 were Q varients, 166 were C variants. Q varients were distributed among seven chromosomes in the genome and among 76 of the 77 subjects. Chromosomes 3 and 4 had variable Q intensities at the centromere, and the acrocentric chromosomes, 13, 14, 15, 21, and 22, had variable Q intensities in their short arms and/or satellites. C variants, though fewer in number, were more widely distributed in the genome, with at least one variant described in each chromosome. C variants were identified in 65 of the 77 subjects. Most unique were the six pericentric inversions found in chromosome 9. Except for giant satellites, no correlations were found between Q and C variants. Q and C variants evaluated in 16 members of four families showed that in nearly every case each variant observed in a child could be demonstrated in one or both parents. It is evident from this study that the magnitude of chromosomal variation in human populations is far greater than heretofore believed. It has also been shown that the combination of Q- and C-banding procedures will yield much more information than either technique used alone and is therefore the preferred approach to many population and gene localization studies.     

Fertility Defects in Mice Expressing the L68Q Variant of Human Cystatin C: A ROLE FOR AMYLOID IN MALE INFERTILITY*

Hereditary cystatin C amyloid angiopathy is an autosomal dominant disorder in which a variant form of cystatin C (L68Q) readily forms amyloid deposits in cerebral arteries in affected individuals resulting in early death. L68Q protein deposits in human cystatin C amyloid angiopathy patients have also been found in tissues outside of the brain including the testis, suggesting possible effects on fertility. Heterozygous transgenic mice (L68Q) that express the human L68Q variant of cystatin C under the control of the mouse cystatin C promoter were unable to generate offspring, suggesting the presence of L68Q cystatin C amyloid affected sperm function. In vitro studies showed that epididymal spermatozoa from L68Q mice were unable to fertilize oocytes and exhibited poor sperm motility. Furthermore, spermatozoa from L68Q mice exhibited reduced cell viability compared with wild type (WT) spermatozoa and often were detected in large agglutinated clumps. Examination of the epididymal fluid and spermatozoa from L68Q mice showed increased levels and distinct forms of cystatin C amyloid that were not present in WT mice. The addition of epididymal fluid from L68Q mice to WT spermatozoa resulted in a recapitulation of the L68Q phenotype in that WT spermatozoa showed reduced cell viability and motility compared with WT spermatozoa incubated in epididymal fluid from WT mice. L68Q epididymal fluid that was depleted of cystatin C amyloids, however, did not impair the motility of WT spermatozoa. Taken together these studies suggest that amyloids in the epididymal fluid can be cytotoxic to the maturing spermatozoa resulting in male infertility.     

Immunochemical identification of coenzyme Q0-dihydrolipoamide adducts in the E2 components of the alpha-ketoglutarate and pyruvate dehydrogenase complexes partially explains the cellular toxicity of coenzyme Q0

Coenzyme Q(0) (Q(0)), a strong electrophile, is toxic to insulin-producing cells. Q(0) was incubated with rat and human pancreatic islets and INS-1 insulinoma cells, and its attachment to cellular proteins was studied with Western analysis using antiserum raised against the benzoquinone ring structure of ubiquinone (anti-Q). Q(0) covalently bonded to two proteins, one of 50 kDa and another of 70 kDa. Both proteins were found to be mitochondrial in human and rat islet cells and in many rat organs. Mitochondria were incubated with Q(0), and affinity-purified anti-Q was used to immunoprecipitate the 50-kDa protein. Amino acid sequencing identified it as dihydrolipoamide succinyltransferase, the E2 component of the alpha-ketoglutarate dehydrogenase complex (KDC). Western analysis also showed that Q bonds to the E2 components of the purified KDC and (0)the pyruvate dehydrogenase complex (PDC). Dihydrolipoamide acetyltransferase, the E2 of the PDC, has a molecular mass of 70 kDa, and the 70-kDa protein was inferred to be this enzyme. Q(0) was found to bond only to proteins containing dihydrolipoate, and in preparations of mitochondria, thiol reducing agents facilitated the attachment of Q(0), but oxidizing agents prevented it, suggesting that Q(0) bonds to thiols of dihydrolipoamide. Incubation of human or pig PDC with Q(0) followed by matrix-assisted laser desorption ionization time-of-flight and liquid chromatography/electrospray ionization mass spectrometry analyses of chymotrypsin-digested peptides of PDC E2 confirmed that Q(0) bonds to the dihydrolipoamide in these proteins. In mitochondria, coenzymes Q(1) and Q(2) did not bond to the 50-kDa protein but competed with the bonding of Q(0) to this protein. The prevention by Q(1) of characteristics the bonding of Q(0) to KDC E2, as well as other of the Q(0) effect, are reminiscent of the action of Q(0) on the mitochondrial permeability transition pore described previously (Fontaine, E., Ichas, F., and Bernardi, P. (1998) J. Biol. Chem. 273, 25734-25740).