Soumitra S. Ghosh

Characterization of the human heart mitochondrial proteome

To gain a better understanding of the critical role of mitochondria in cell function, we have compiled an extensive catalogue of the mitochondrial proteome using highly purified mitochondria from normal human heart tissue. Sucrose gradient centrifugation was employed to partially resolve protein complexes whose individual protein components were separated by one-dimensional PAGE. Total in-gel processing and subsequent detection by mass spectrometry and rigorous bioinformatic analysis yielded a total of 615 distinct protein identifications. All protein pI values, molecular weight ranges, and hydrophobicities were represented. The coverage of the known subunits of the oxidative phosphorylation machinery within the inner mitochondrial membrane was >90%. A significant proportion of identified proteins are involved in signaling, RNA, DNA, and protein synthesis, ion transport, and lipid metabolism. The biochemical roles of 19% of the identified proteins have not been defined. This database of proteins provides a comprehensive resource for the discovery of novel mitochondrial functions and pathways.

Reduced-Median-Network Analysis of Complete Mitochondrial DNA Coding-Region Sequences for the Major African, Asian, and European Haplogroups

The evolution of the human mitochondrial genome is characterized by the emergence of ethnically distinct lineages or haplogroups. Nine European, seven Asian (including Native American), and three African mitochondrial DNA (mtDNA) haplogroups have been identified previously on the basis of the presence or absence of a relatively small number of restriction-enzyme recognition sites or on the basis of nucleotide sequences of the D-loop region. We have used reduced-median-network approaches to analyze 560 complete European, Asian, and African mtDNA coding-region sequences from unrelated individuals to develop a more complete understanding of sequence diversity both within and between haplogroups. A total of 497 haplogroup-associated polymorphisms were identified, 323 (65%) of which were associated with one haplogroup and 174 (35%) of which were associated with two or more haplogroups. Approximately one-half of these polymorphisms are reported for the first time here. Our results confirm and substantially extend the phylogenetic relationships among mitochondrial genomes described elsewhere from the major human ethnic groups. Another important result is that there were numerous instances both of parallel mutations at the same site and of reversion (i.e., homoplasy). It is likely that homoplasy in the coding region will confound evolutionary analysis of small sequence sets. By a linkage-disequilibrium approach, additional evidence for the absence of human mtDNA recombination is presented here.

Microsatellites for linkage analysis of genetic traits

Microsatellites are tandem repeats of simple sequence that occur abundantly and at random throughout most eukaryotic genomes. Since they are usually less than 100 bp long and are embedded in DNA with unique sequence, they can be amplified in vitro using the polymerase chain reaction. Microsatellites are easy to clone and characterize and display considerable polymorphism due to variation in the number of repeat units. This polymorphism is sufficiently stable to use in genetic analyses. Microsatellites are therefore ideal markers for constructing high-resolution genetic maps in order to identify susceptibility loci involved in common genetic diseases.

Oxidative Damage to Mitochondrial Complex I Due to Peroxynitrite IDENTIFICATION OF REACTIVE TYROSINES BY MASS SPECTROMETRY

There is growing evidence that oxidative phosphorylation (OXPHOS) generates reactive oxygen and nitrogen species within mitochondria as unwanted byproducts that can damage OXPHOS enzymes with subsequent enhancement of free radical production. The accumulation of this oxidative damage to mitochondria in brain is thought to lead to neuronal cell death resulting in neurodegeneration. The predominant reactive nitrogen species in mitochondria are nitric oxide and peroxynitrite. Here we show that peroxynitrite reacts with mitochondrial membranes from bovine heart to significantly inhibit the activities of complexes I, II, and V (50–80%) but with less effect upon complex IV and no significant inhibition of complex III. Because inhibition of complex I activity has been a reported feature of Parkinsons disease, we undertook a detailed analysis of peroxynitrite-induced modifications to proteins from an enriched complex I preparation. Immunological and mass spectrometric approaches coupled with two-dimensional PAGE have been used to show that peroxynitrite modification resulting in a 3-nitrotyrosine signature is predominantly associated with the complex I subunits, 49-kDa subunit (NDUFS2), TYKY (NDUFS8), B17.2 (17.2-kDa differentiation associated protein), B15 (NDUFB4), and B14 (NDUFA6). Nitration sites and estimates of modification yields were deduced from MS/MS fragmentograms and extracted ion chromatograms, respectively, for the last three of these subunits as well as for two co-purifying proteins, the β and the d subunits of the F1F0-ATP synthase. Subunits B15 (NDUFB4) and B14 (NDUFA6) contained the highest degree of nitration. The most reactive site in subunit B14 was Tyr122, while the most reactive region in B15 contained 3 closely spaced tyrosines Tyr46, Tyr50, and Tyr51. In addition, a site of oxidation of tryptophan was detected in subunit B17.2 adding to the number of post-translationally modified tryptophans we have detected in complex I subunits (Taylor, S. W., Fahy, E., Murray, J., Capaldi, R. A., and Ghosh, S. S. (2003) J. Biol. Chem. 278, 19587–19590). These sites of oxidation and nitration may be useful biomarkers for assessing oxidative stress in neurodegenerative disorders.

Global organellar proteomics

Cataloging the proteomes of single-celled microorganisms, cells, biological fluids, tissue and whole organisms is being undertaken at a rapid pace as advances are made in protein and peptide separation, detection and identification. For metazoans, subcellular organelles represent attractive targets for global proteome analysis because they represent discrete functional units, their complexity in protein composition is reduced relative to whole cells and, when abundant cytoskeletal proteins are removed, lower abundance proteins specific to the organelle are revealed. Here, we review recent literature on the global analysis of subcellular organelles and briefly discuss how that information is being used to elucidate basic biological processes that range from cellular signaling pathways through protein-protein interactions to differential expression of proteins in response to external stimuli. We assess the relative merits of the different methods used and discuss issues and future directions in the field.

Additional microsatellite markers for mouse genome mapping

Mouse sequence information from the EMBL and GenBank databases, published sequences and genomic clones have been analyzed for simple repetitive elements or microsatellites. Each microsatellite has been amplified by the polymerase chain reaction (PCR) as a single locus marker. PCR primers were designed from unique sequence flanking each repeat. Size variation of PCR products less than 750 base pairs (bp) between mouse strains has been determined using ethidium bromide-stained acrylamide or agarose gels. A further 74 newly characterized microsatellites are presented in this paper, bringing to 185 the total we have analyzed. Of these, 157/185 (85%) have more than one allele, 143/178 (80%) vary in length between C57BL/6J and Mus spretus, and 82/168 (49%) vary between DBA/2J and C57BL/6J. Microsatellites provide informative single locus probes for linkage analysis in the construction of a genetic map of the mouse genome.

Use of cytoplasmic hybrid cell lines for elucidating the role of mitochondrial dysfunction in Alzheimer's disease and Parkinson's disease.

ABSTRACT: There is substantial evidence of mitochondrial defects in neurodegenerative disorders such as Alzheimers and Parkinsons diseases (AD and PD). We have probed the molecular implications of mitochondrial dysfunction in these diseases by transferring mitochondria from platelets obtained from disease and control donors into mitochondrial DNA‐depleted recipient neuron‐based cells (ρ0 cells). This process creates cytoplasmic hybrid (cybrid) cells where the mitochondrial DNA (mtDNA) from the donor is expressed in the nuclear and cellular background of the host ρ0 cell. Differences in phenotype between disease and control groups can thus be attributed to the exogenous mitochondria and mtDNA. Key methodological issues relating to this approach were addressed by demonstrating that recipient ρ0 cells have < 1 mtDNA copy/cell, and that exclusive repopulation with donor mtDNA occurs in cybrid cells. Further, we describe that sampling of heterogeneous cell populations is a valid approach for cybrid analysis. Our studies show that the focal respiratory chain defects reported in platelets of AD and PD cybrids can be recapitulated in AD and PD cybrids. In addition, both AD and PD cybrids display increased oxidative stress and perturbations in calcium homeostasis. These data suggest that the transfer of a mtDNA defect from disease donor platelets is the likely cause of the cybrid biochemical phenotype, and highlight the potential value of these cell lines as cellular disease models.

Two-dimensional electrophoresis and mass spectrometric identification of mitochondrial proteins from an SH-SY5Y neuroblastoma cell line.

To probe the mitochondrial involvement in neurodegenerative processes, we have generated a high-resolution map of the mitochondrial proteome from a human neuroblastoma SH-SY5Y cell line that has been used for creating cytoplasmic hybrid cell systems. Two mitochondrial preparations were evaluated using two-dimensional (2D) gel electrophoresis and mass spectrometry; one obtained from differential centrifugation and the other by a multiple-step percoll/metrizamide gradient. The 2D gel maps prepared from these mitochondrial fractions separated over 300 distinct spots as visualized by colloidal Coomassie blue (CCB), or closer to 400 proteins with silver staining. The most abundant proteins identified in the mitochondrial fraction prepared by differential centrifugation were those of mitochondrial, cytoplasmic, and endoplasmic reticulum origin. Proteins obtained using the more intensive two-step gradient method were almost exclusively known to be associated with mitochondria. From this latter preparation, 84 of the most abundant gel spots were analyzed, out of which 61 proteins were identified. The absence of many membrane-associated proteins known to be associated with the mitochondrion and the limited number of total proteins observed in the 2D gel maps suggest that the majority of mitochondrial proteins are not being detected under these separation and staining conditions. An insoluble pellet obtained after solubilization of the mitochondrial fraction prepared with the percoll/metrizamide gradient was boiled in sodium dodecylsulfate (SDS) and separated by 1D sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE). This separation yielded some additional proteins, many of which are likely membrane-associated. These studies form the basis for the analysis of differential protein expression in cybrid cellular models of neurodegenerative disorders and in affected tissue from diseased states.

Longitudinal analysis of the segregation of mtDNA mutations in heteroplasmic individuals

The mutation load of the pathogenic LHON (Leber hereditary optic neuropathy) mtDNA mutation at nucleotide 3460 has been followed over time in the WBC/platelet fraction from members of a matrilineal pedigree. Longitudinal analysis over a sampling period of five to six years indicates that, in all five heteroplasmic family members, the mutation load decreases at a mean overall rate of approximately 1% per year. There was no change in mutation load in homoplasmic wildtype or in homoplasmic mutant individuals. For the purposes of comparison, a longitudinal analysis of a silent mtDNA polymorphism at nucleotide 14560 was also carried out for members of a second matrilineal pedigree. In contrast to the results for the pathogenic mtDNA mutation, there was no change in the proportion of the silent polymorphism in the WBC/platelet fraction of four family members over a period of seven years. These results indicate that the pathogenic 3460 LHON mutation segregates under negative selection in these cell populations. One possible mechanism through which selection may operate is that, in heteroplasmic individuals, the hematopoietic stem cells are generally homoplasmic, either for the wildtype or for the mutant allele. The homoplasmic mutant stem cells, because of their mitochondrial respiratory chain defect, produce fewer mature WBCs and platelets over time than do the wildtype stem cells. Alternatively, the stem cells may be heteroplasmic and selection may act to favor proliferation of mitochondria with lower levels of the pathogenic mutation in the WBC/platelet cell populations.

Linkage analysis of 84 microsatellite markers in intra- and interspecific backcrosses.

Microsatellite sequences are stretches of repetitive DNA distributed throughout the genomes of most eukaryotes (Weber and May 1989; Stallings et al. 1991). They are characterized by a short basic unit that is repeated several times in a tandem array. The core unit can be a dinucleotide, trinucleotide, or tetranucleotide. Such sequences have been shown to be highly polymorphic not only between different species of mice, but also between different inbred laboratory strains (Love et al. 1990; Dietrich et al. 1992). Microsatellite length variants are easily detected with the polymerase chain reaction (PCR) provided unique flanking sequence is available for PCR primer design and the resulting PCR products can be resolved by acrylamide/agarose gel electrophoresis. This ease of analysis, coupled with the high degree of polymorphism exhibited by microsatellites, makes them extremely useful markers in the construction of highresolution, easily accessible linkage maps (Love et al. 1990; Dietrich et al. 1992). Such genetic linkage maps have proved invaluable in the mapping of disease susceptibility genes (Todd et al. 1991). The nonobese diabetic (NOD) mouse spontaneously develops diabetes with many similarities to the human disorder type 1 (insulin-dependent) diabetes mellitus and as such is a useful tool for the mapping and identification of diabetes susceptibility loci (Todd et al. 1991). To this end we undertook the construction of a linkage map of the mouse genome, using a series of variant microsatellite markers amplified by PCR and DNA purified from the progeny of two reciprocal first backcross generations (BC1), (BIO.H-2 g7 • NOD)F 1 • NOD and NOD • (BIO.H-2 g7 • NOD)F 1. Between 37 and 471 animals were typed for 94 marker loci distributed over 16 autosomes. The use of a con-