Progress with gene-product mapping of the Mollicutes: Mycoplasma genitalium

The comparative metabolism of the mollicutes (Mycoplasmas): the utility for taxonomic classification and the relationship of putative gene annotation and phylogeny to enzymatic function in the smallest free-living cells

Mollicutes or mycoplasmas are a class of wall-less bacteria descended from low G + C% Gram-positive bacteria. Some are exceedingly small, about 0.2 micron in diameter, and are examples of the smallest free-living cells known. Their genomes are equally small; the smallest in Mycoplasma genitalium is sequenced and is 0.58 mb with 475 ORFs, compared with 4.639 mb and 4288 ORFs for Escherichia coli. Because of their size and apparently limited metabolic potential, Mollicutes are models for describing the minimal metabolism necessary to sustain independent life. Mollicutes have no cytochromes or the TCA cycle except for malate dehydrogenase activity. Some uniquely require cholesterol for growth, some require urea and some are anaerobic. They fix CO2 in anaplerotic or replenishing reactions. Some require pyrophosphate not ATP as an energy source for reactions, including the rate-limiting step of glycolysis: 6-phosphofructokinase. They scavenge for nucleic acid precursors and apparently do not synthesize pyrimidines or purines de novo. Some genera uniquely lack dUTPase activity and some species also lack uracil-DNA glycosylase. The absence of the latter two reactions that limit the incorporation of uracil or remove it from DNA may be related to the marked mutability of the Mollicutes and their tachytelic or rapid evolution. Approximately 150 cytoplasmic activities have been identified in these organisms, 225 to 250 are presumed to be present. About 100 of the core reactions are graphically linked in a metabolic map, including glycolysis, pentose phosphate pathway, arginine dihydrolase pathway, transamination, and purine, pyrimidine, and lipid metabolism. Reaction sequences or loci of particular importance are also described: phosphofructokinases, NADH oxidase, thioredoxin complex, deoxyribose-5-phosphate aldolase, and lactate, malate, and glutamate dehydrogenases. Enzymatic activities of the Mollicutes are grouped according to metabolic similarities that are taxonomically discriminating. The arrangements attempt to follow phylogenetic relationships. The relationships of putative gene assignments and enzymatic function in My. genitalium, My. pneumoniae, and My. capricolum subsp. capricolum are specially analyzed. The data are arranged in four tables. One associates gene annotations with congruent reports of the enzymatic activity in these same Mollicutes, and hence confirms the annotations. Another associates putative annotations with reports of the enzyme activity but from different Mollicutes. A third identifies the discrepancies represented by those enzymatic activities found in Mollicutes with sequenced genomes but without any similarly annotated ORF. This suggests that the gene sequence is significantly different from those already deposited in the databanks and putatively annotated with the same function. Another comparison lists those enzymatic activities that are both undetected in Mollicutes and not associated with any ORF. Evidence is presented supporting the theory that there are relatively small gene sequences that code for functional centers of multiple enzymatic activity. This property is seemingly advantageous for an organism with a small genome and perhaps under some coding restraint. The data suggest that a concept of "remnant" or "useless genes" or "useless enzymes" should be considered when examining the relationship of gene annotation and enzymatic function. It also suggests that genes in addition to representing what cells are doing or what they may do, may also identify what they once might have done and may never do again.

In search of differentially expressed genes and proteins

A great challenge for modern cell biology is the successful examination of the co-expression of thousands of genes under physiological or pathological conditions and how the expression patterns define the different states of a single cell, tissue or a microorganism. Gene expression can be analyzed today on a large scale by advanced technical approaches for differential screening of proteins and mRNAs. The identification of differentially expressed mRNAs has been successfully applied to understand gene function and the underlying molecular mechanism(-s) of differentiation, development and disease state. Analysis of gene expression by the systematic mapping of thousands of proteins present in a cell or tissue can be achieved by the use of two-dimensional (2D) gel electrophoresis, quantitative computer image analysis, and protein identification techniques. In this article, we comment on some of these techniques and try to stress their advantages and drawbacks. We show how data from RNA/DNA mapping, sequence information from genome projects and protein pattern profiling can be linked with each other and annotated. These comprehensive approaches permit the study of differential gene and protein expressions in cells or tissues.

Unique identification of proteins from small genome organisms: theoretical feasibility of high throughput proteome analysis

We evaluate current levels of accuracy for estimation of molecular weight (Mr) and isoelectric point (pI) to proteins on two-dimensional (2-D) gels as well as the distribution and clustering of proteins in the predicted proteome of E. coli. We also examine the ability to find single candidates within the predicted proteome for matching to a protein seen on 2-D gels, based on the current level of accuracy. We discuss the levels of accuracy needed to match predicted proteins to observed proteins based solely on Mr and pI criteria obtained from genomic information, and propose methodology to achieve this level of accuracy. In addition, we will address the future goals of this work since the small genomes of bacteria provide a foundation and stepping stone to similar studies in higher organisms.

HSC-2DPAGE and the two-dimensional gel electrophoresis database of dog heart proteins

A two-dimensional gel electrophoresis database of dog (Canis familiaris) proteins is presented. The database contains 1212 protein spots which have been characterised in terms of their pI and Mr. This database has been integrated into the HSC-2DPAGE database which is accessible on the Internet via the World Wide Web with the uniform resource location (URL): (http://www.harefield.nthames.nhs.uk/nhli/ protein/index.html). Identifications for 80 of the protein spots have been obtained by visual cross-matching with the human heart protein database in HSC-2DPAGE (42 spots), N-terminal microsequence analysis (25 spots) and peptide mass fingerprinting (20 spots). This database is being used in studies of alterations in protein expression in models of heart failure and heart disease.

Mycoplasma genes: a case for reflective annotation

Although function can be assigned to genome sequence by homology at a macroscopic level, this can be misleading in the absence of data on enzyme activities. Together, such data can reveal whether open reading frames are expressed, identify multienzyme function and point to 'orphan' function. Because of their small size and small genomes, the genome sequences of some Mycoplasma spp. are very amenable to detailed analyses.

Malate/lactate dehydrogenase in mollicutes: evidence for a multienzyme protein

The malate (MDH) and lactate (LDH) dehydrogenases belong to the homologous class of 2-ketoacid dehydrogenases. The specificity for their respective substrates depends on residues differing at two or three regions within each molecule. Theoretical peptide-mass fingerprinting and PROSITE analysis of nine MDH and six LDH molecules were used to describe conserved sites related to function. A unique LDH is described which probably also confers MDH activity within the 580 kbp genome of Mycoplasma genitalium (class: Mollicutes). A single hydrophilic arginine residue was found in the active site of the M. genitalium LDH enzyme, differing from an hydrophobic residue normally present in these molecules. The effect of this residue may be to alter active site substrate specificity, allowing the enzyme to perform two closely related tasks. Evidence for a single gene affording dual enzymatic function is discussed in terms of genome size reduction in the simplest of free-living organisms. Since Mollicutes are thought to lack enzymes of the tricarboxylic acid cycle that would otherwise bind and interact with MDH in bacterial species possessing this pathway, active site modification of M. genitalium LDH is the sole requirement for MDH activity of this molecule. The closely related helical Mollicute, Spiroplasma melliferum, was shown to possess two distinct gene products for MDH/LDH activity.

'Proteomic contigs' of Mycobacterium tuberculosis and Mycobacterium bovis (BCG) using novel immobilised pH gradients

Tuberculosis remains a major health problem throughout the world and the failure of the existing bacille Calmette-Guérin (BCG) vaccine in recent trials has prompted a search for potential replacements. Recent advances in molecular and cell biology have cast doubts on the ability of genetic analysis alone to predict polygenic human diseases and other complex phenotypes and have therefore redirected our attention to proteome studies to complement information obtained from DNA sequencing initiatives. Novel acidic (pH 2.3-5) and basic (pH 6-11) IPG gel gradients were employed in conjunction with commercially available pH 4-7 gradients to significantly increase (fourfold) the number of protein spots previously resolved on two-dimensional (2-D) gels of Mycobacterium species. A total of 772 and 638 protein spots were observed for M. bovis BCG and M. tuberculosis H37Rv, respectively, the latter corresponding to only the pH regions 4-7 and 6-11. Of interest was the bimodal distribution observed for proteins separated from M. bovis BCG across both M(r) and pH ranges. Some differences in protein expression were observed between these two organisms, contrary to what may have been expected considering the high degree of conservation in gene order and sequence similarity between homologous genes. Further work will be directed towards a more detailed analysis of these differences, so as to allow more accurate diagnosis between vaccination and active tuberculosis. The latter is of major importance to epidemiological studies and for patient management.

Proteomic 'contigs' of Ochrobactrum anthropi, application of extensive pH gradients

The most extensive linear pH gradients yet employed in combination with two-dimensional gel electrophoresis are described, along with their application in proteome analysis. A significant proportion of the protein compliment of bacterial species is believed to be accessible using an extended linear pH gradient of 2.3 to 11.0. Protein standards with predicted isoelectric points (pI) ranging from 3.24 to 9.56 were used to confirm focusing positions with respect to the immobilised pH gradients (IPG) prior to mapping studies of Ochrobactrum anthropi. Multiple gel images were used to construct contiguous windows of protein expression ('proteomic contigs') within 18 cm pH gradients 2.3-5, 4-7, and 6-11 in conjunction with 15% T and 7.5% T acrylamide gels, the latter being used to resolve higher molecular weight (M(r)) proteins. Each IPG had a 5 cm region of similar pH gradient overlap at pH 4-5 and pH 6-7 that was used to construct an image of protein expression characteristic of whole cell lysates. This is reminiscent of genomic sequencing initiatives whereby portions are combined to form a contiguous picture of the whole. The protein maps obtained demonstrated a means of resolving the many tens of thousands of cellular proteins likely to occur in eukaryotic systems, but also highlighted the need to further optimise protein extraction, equilibration buffers, and separation conditions of higher M(r) proteins occurring at extreme pI. Theoretical 2-D protein maps were constructed for five organisms for which the total DNA sequence is now available. In all cases, higher M(r) acidic and basic proteins were shown to be common.

Proteome research: complementarity and limitations with respect to the RNA and DNA worlds

A methodological overview of proteome analysis is provided along with details of efforts to achieve high-throughput screening (HTS) of protein samples derived from two-dimensional electrophoresis gels. For both previously sequenced organisms and those lacking significant DNA sequence information, mass spectrometry has a key role to play in achieving HTS. Prototype robotics designed to conduct appropriate chemistries and deliver 700-1000 protein (genes) per day to batteries of mass spectrometers or liquid chromatography (LC)-based analyses are well advanced, as are efforts to produce high density gridded arrays containing > 1000 proteins on a single matrix assisted laser desorption ionisation/time-of-flight (MALDI-TOF) sample stage. High sensitivity HTS of proteins is proposed by employing principally mass spectrometry in an hierarchical manner: (i) MALDI-TOF-mass spectrometry (MS) on at least 1000 proteins per day; (ii) electrospray ionisation (ESI)/MS/MS for analysis of peptides with respect to predicted fragmentation patterns or by sequence tagging; and (iii) ESI/MS/MS for peptide sequencing. Genomic sequences when complemented with information derived from hybridisation assays and proteome analysis may herald in a new era of holistic cellular biology. The current preoccupation with the absolute quantity of gene-product (RNA and/or protein) should move backstage with respect to more molecularly relevant parameters, such as: molecular half-life; synthesis rate; functional competence (presence or absence of mutations); reaction kinetics; the influence of individual gene-products on biochemical flux; the influence of the environment, cell-cycle, stress and disease on gene-products; and the collective roles of multigenic and epigenetic phenomena governing cellular processes. Proteome analysis is demonstrated as being capable of proceeding independently of DNA sequence information and aiding in genomic annotation. Its ability to confirm the existence of gene-products predicted from DNA sequence is a major contribution to genomic science. The workings of software engines necessary to achieve large-scale proteome analysis are outlined, along with trends towards miniaturisation, analyte concentration and protein detection independent of staining technologies. A challenge for proteome analysis into the future will be to reduce its dependence on two-dimensional (2-D) gel electrophoresis as the preferred method of separating complex mixtures of cellular proteins. Nonetheless, proteome analysis already represents a means of efficiently complementing differential display, high density expression arrays, expressed sequence tags, direct or subtractive hybridisation, chromosomal linkage studies and nucleic acid sequencing as a problem solving tool in molecular biology.

Proteome studies of Saccharomyces cerevisiae: identification and characterization of abundant proteins

Two-dimensional (2-D) gel electrophoresis can now be coupled with protein identification techniques and genome sequence information for direct detection, identification, and characterization of large numbers of proteins from microbial organisms. 2-D electrophoresis, and new protein identification techniques such as amino acid composition, are proteome research techniques in that they allow direct characterization of many proteins at the same time. Another new tool important for yeast proteome research is the Yeast Protein Database (YPD), which provides the sequence-derived protein properties needed for spot identification and tabulations of the currently known properties of the yeast proteins. Studies presented here extend the yeast 2-D protein map to 169 identified spots based upon the recent completion of the yeast genome sequence, and they show that methods of spot identification based on predicted isoelectric point, predicted molecular mass, and determination of partial amino acid composition from radiolabeled gels are powerful enough for the identification of at least 80% of the spots representing abundant proteins. Comparison of proteins predicted by YPD to be detectable on 2-D gels based on calculated molecular mass, isoelectric point and codon bias (a predictor of abundance) with proteins identified in this study suggests that many glycoproteins and integral membrane proteins are missing from the 2-D gel patterns. Using the 2-D gel map and the information available in YDP, 2-D gel experiments were analyzed to characterize the yeast proteins associated with: (i) an environmental change (heat shock), (ii) a temperature-sensitive mutation (the prp2 mRNA splicing mutant), (iii) a mutation affecting post-translational modification (N-terminal acetylation), and (iv) a purified subcellular fraction (the ribosomal proteins). The methods used here should allow future extension of these studies to many more proteins of the yeast proteome.

Proteome analysis of Spiroplasma melliferum (A56) and protein characterisation across species boundaries

Spiroplasma melliferum (Class: Mollicutes) is a wall-less, helical bacterium with a genome of approximately 1460 kbp encoding 800-1000 gene-products. A two-dimensional electrophoresis gel reference map of S. melliferum was produced by Phoretix 2-D gel software analysis of eight high quality gels. The reference map showed 456 silver-stained and replicated protein spots. 156 proteins (34% of visible protein spots) from S. melliferum were further characterised by one, or a combination, of the following: amino acid analysis, peptide-mass fingerprinting via matrix assisted laser desorption ionisation-time of flight (MALDI-TOF) mass spectrometry, and N-terminal protein microsequencing. Proteins with close relationship to those previously determined from other species were identified across species barriers. Thus, this study represents the first larger-scale analysis of a proteome based upon the attribution of predominantly 'unique numerical parameters' for protein characterisation across species boundaries, as opposed to a sequence-based approach. This approach allowed all database entries to be screened for homology, as is currently the case for studies based on nucleic acid or protein sequence information. Several proteins studied from this organism were identified as hypothetical, or having no close homolog already present in the databases. Gene-products from major families such as glycolysis, translation, transcription, cellular processes, energy metabolism and protein synthesis were identified. Several gene-products characterised in S. melliferum were not previously found in studies of the entire Mycoplasma genitalium and Mycoplasma pneumoniae (both closely related Mollicutes) genomes. The presence of such gene-products in S. melliferum is discussed in terms of genome size as compared with the smallest known free-living organisms. Finally, the levels of expression of S. melliferum gene-products were determined with respect to total optical intensity associated with all visible proteins expressed in exponentially grown cells.

Peptide-mass fingerprinting and the ideal covering set for protein characterisation

The rules that govern the dynamics of protein characterisation by peptide-mass fingerprinting (PMF) were investigated through multiple interrogations of a nonredundant protein database. This was achieved by analysing the efficiency of identifying each entry in the entire database via perfect in silico digestion with a series of 20 pseudo-endoproteinases cutting at the carboxy terminal of each amino acid residue, and the multiple cutters: trypsin, chymotrypsin and Glu-C. The distribution of peptide fragment masses generated by endoproteinase digestion was examined with a view to designing better approaches to protein characterisation by PMF. On average, and for both common and rare cutters, the combination of approximately two fragments was sufficient to identify most database entries. However, the rare cutters left more entries unidentified in the database. Total coverage of the entire database could not be achieved with one enzymatic cutter alone, nor when all 23 cutters were used together. Peptide fragments of > 5000 Da had little effect on the outcome of PMF to correctly characterise database entries, while those with low mass (near to 350 Da in the case of trypsin) were found to be of most utility. The most frequently occurring fragments were also found in this lower mass region. The maximum size of uncut database entries (those not containing a specific amino acid residue) ranged from 52,908 Da to 258,314 Da, while the failure rate for a single cutter in identifying database entries varied from 10,865 (8.4%) to 23,290 (18.1%). PMF is likely to be a mainstay of any high-throughput protein screening strategy for large-scale proteome analysis. A better understanding of the merits and limitations of this technique will allow researchers to optimise their protein characterisation procedures.

The proteome of Salmonella enterica serovar typhimurium: current progress on its determination and some applications

Salmonella typhimurium (official designation Salmonella enterica serovar Typhimurium) is an enteric pathogen and a principal cause of gastroenteritis in humans. A comprehensive description of the proteins of Salmonella and their patterns of expression under different environmental conditions would greatly increase our understanding of the virulence of this organism at the molecular level and provide insights into many other aspects of Salmonella biology. While a variety of two-dimensional studies of Salmonella have been previously carried out to address specific questions, little systematic information is available at the protein level on the numbers of Salmonella polypeptides that have homologues in other organisms, their abundance, and the frequency of post-translational modifications. To test the feasibility of determining the proteome of Salmonella, the identities of 53 randomly sequenced cell envelope proteins have been determined by N-terminal sequencing of spots from two-dimensional gels. In addition to confirming the existence of previously hypothetical proteins predicted from genomic sequencing projects, we found that approximately 20% of the proteins had no matches in sequence databases. The results suggest that proteome analysis is an efficient way to identify novel proteins from prokaryotes and that the analysis provides a useful approach to the study of Salmonella virulence.

Comparing the predicted and observed properties of proteins encoded in the genome of Escherichia coli K-12

Mining the emerging abundance of microbial genome sequences for hypotheses is an exciting prospect of "functional genomics". At the forefront of this effort, we compared the predictions of the complete Escherichia coli genomic sequence with the observed gene products by assessing 381 proteins for their mature N-termini, in vivo abundances, isoelectric points, molecular masses, and cellular locations. Two-dimensional gel electrophoresis (2-DE) and Edman sequencing were combined to sequence Coomassie-stained 2-DE spots representing the abundant proteins of wild-type E. coli K-12 strains. Greater than 90% of the abundant proteins in the E. coli proteome lie in a small isoelectric point and molecular mass window of 4-7 and 10-100 kDa, respectively. We identified several highly abundant proteins, YjbJ, YjbP, YggX, HdeA, and AhpC, which would not have been predicted from the genomic sequence alone. Of the 223 uniquely identified loci, 60% of the encoded proteins are proteolytically processed. As previously reported, the initiator methionine was efficiently cleaved when the penultimate amino acid was serine or alanine. In contrast, when the penultimate amino acid was threonine, glycine, or proline, cleavage was variable, and valine did not signal cleavage. Although signal peptide cleavage sites tended to follow predicted rules, the length of the putative signal sequence was occassionally greater than the consensus. For proteins predicted to be in the cytoplasm or inner membrane, the N-terminal amino acids were highly constrained compared to proteins localized to the periplasm or outer membrane. Although cytoplasmic proteins follow the N-end rule for protein stability, proteins in the periplasm or outer membrane do not follow this rule; several have N-terminal amino acids predicted to destabilize the proteins. Surprisingly, 18% of the identified 2-DE spots represent isoforms in which protein products of the same gene have different observed pI and M(r), suggesting they are post-translationally processed. Although most of the predicted and observed values for isoelectric point and molecular mass show reasonable concordance, for several proteins the observed values significantly deviate from the expected values. Such discrepancies may represent either highly processed proteins or misinterpretations of the genomic sequence. Our data suggest that AhpC, CspC, and HdeA exist as covalent homomultimers, and that IcdA exists as at least three isoforms even under conditions in which covalent modification is not predicted. We enriched for proteins based on subcellular location and found several proteins in unexpected subcellular locations.

Characterisation of basic proteins from Spiroplasma melliferum using novel immobilised pH gradients

Two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) has become the method of choice for efficient separation of complex protein mixtures. Previously, analysis of the Spiroplasma melliferum proteome (protein complement of a genome) has been performed with pH 3-10 and narrow range pH 4-7 IPG gel strips. We report here on the use of novel 18 cm basic (pH 6-11) immobilised pH gradients (IPG) to increase the resolution of protein spots visible within 2-D gels. These gradients were synthesised to emulate the gradient of commercially available IPG gel strips in a 5 cm region of overlap so as to attempt construction of a more complete map of cellular protein expression. Approximately 50 additional gene products were detected from S. melliferum that were not previously well-resolved or visible using wide-range pH 3-10 IPG gel strips. Twenty-seven of these were electrotransferred to polyvinylidene difluoride (PVDF) membrane and analysed by N-terminal protein microsequencing. Protein spots with an initial peak yield of as little as 100 femtomoles (fm) were sequenced to 5-10 amino acid residues, demonstrating the importance of improved sample handling procedures and analytical technologies. Many essential metabolic enzymes were shown to have basic pI, including: glyceraldehyde-3-phosphate dehydrogenase, pyruvate kinase, carbamate kinase and lactate dehydrogenase. A very basic protein (pI approximately 11.0) was identified as uridylate kinase, an enzyme indirectly associated with pyrimidine biosynthesis and thought be absent in some members of the bacterial class Mollicutes. The advent of novel basic (pH 6-11) IPGs has allowed the visualisation of a significantly greater percentage of the 'functional proteome', that portion of the total protein complement of a genome actively translated within a specific time frame, on 2-D electrophoresis gels. This will aid in the characterisation of translated gene products in conjunction with genome sequencing initiatives.

Evaluation of algorithms used for cross-species proteome characterisation

The ability to effectively search databases for the identification of protein spots from two-dimensional electrophoresis gels has become an essential step in the study of microbial proteomes. A variety of analytical techniques are currently being employed during protein characterisation. A number of algorithms used to search databases, accessible via the World Wide Web, depend upon information concerning N- and C-terminal microsequence, amino acid composition, and peptide-mass fingerprinting. The effectiveness of nine such algorithms, as well as COMBINED (software developed in this laboratory for identifying proteins across species boundaries) was examined. Fifty-four ribosomal proteins from the Mycoplasma genitalium genome, and 72 amino acyl tRNA synthetases from the Haemophilus influenzae, M. genitalium and Methanococcus jannaschii genomes were chosen for study. These proteins were selected because they represent a wide range of sequence identities across species boundaries (22.7-100% identity), as detected by standard sequence alignment tools. Such sequence variation allowed for a statistical comparison of algorithm success measured against published sequence identity. The ability of analytical techniques used in protein characterisation and associated database query programs to detect identity at the functional group level was examined for proteins with low levels of homology at the gene/protein sequence level. The significance of these theoretical data manipulations provided the means to predict the utility of data acquired experimentally for non-sequence-dependent software in proteome analysis. The data obtained also predicted that 'sequence tagging' of peptide fingerprints would need to be accompanied by at least 11-20 residues of amino acid sequence for it to be widely used for protein characterisation across species boundaries.

Development of a Haemophilus two-dimensional protein database

Members of the Haemophilus genus are responsible for various human infections including respiratory infections and meningitis. The complete nucleotide sequence of the Rd strain of Haemophilus influenzae has been reported and represents a valuable resource to investigate gene expression within this bacterial group. We described previously the application of two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) to characterise the proteins of Haemophilus influenzae (Cash et al., Electrophoresis 1995, 16, 135-148). We have extended these data with comparative studies of the proteins from other members of the Haemophilus genus (specifically H. parainfluenzae, H. haemolyticus and H. parahaemolyticus) to identify homologous proteins and, by extension, the genes encoding them, among these bacteria. The proteins extracted from each of these bacterial isolates were compared by coelectrophoresis to the 2-D protein profile of the reference nontypable strain of H. influenzae (HI-64443) used as the basis for the 2-D protein database. A composite reference 2-D protein profile of HI-64443 was derived from three independent analyses of the soluble bacterial proteins. Between 21% and 37% of the HI-64443 proteins from the reference 2-D protein profile comigrated with proteins in the other isolates from the Haemophilus genus. This compared with 62% and 64% comigration when HI-64443 was compared with the Eagan and Rd strains of H. influenzae, respectively. The 2-D protein profile of the Rd strain of H. influenzae was compared to that of HI-64443 by coelectrophoresis; 64% of the proteins detected for the Rd strain comigrated with proteins found for HI-64443 when analysed in parallel. The capacity of 2-D PAGE to investigate global interactions of gene expression was applied to the analysis of superoxide dismutase (SOD) expression in H. influenzae strain Eagan. A "knock-out" mutant in the sodA gene which encodes [Mn]-SOD was characterised with respect to protein synthesis compared to the parental isolate. From these analyses, the primary product of sodA was provisionally identified as a protein with a molecular mass of 25500 Da and an estimated pI of 6.55. Quantitative changes in the expression of two other proteins in the SOD mutant were detected by comparison with the parental isolate. These data are discussed in relation to the development of a 2-D protein database for H. influenzae and related bacteria to investigate genome homologies and gene expression.

Proteome analysis: genomics via the output rather than the input code

A knowledge of the 'proteome,' total protein output encoded by a genome, provides information on (1) if and when predicted gene products are translated, (2) the relative concentrations of gene products, and (3) the extent of posttranslational modification, none of which can be accurately predicted from the nucleic acid sequence alone. The current status of proteome analysis is reviewed with respect to some of the techniques employed, automation, relevance to genomic studies, mass spectrometry and bioinformatics, limitations, and recent improvements in resolution and sensitivity for the detection of protein expression in whole cells, tissues, or organisms. The concept of 'proteomic contigs' is introduced for the first time. Traditional approaches to genomic analysis call upon a number of strategies to produce contiguous DNA sequence information, while 'proteomic contigs' are derived from multiple molecular mass and isoelectric point windows in order to construct a picture of the total protein expression within living cells. In higher eukaryotes, the latter may require several dozen image subsets of protein spots to be stitched together using advanced image analysis. The utility of both experimental and theoretical peptide-mass fingerprinting (PMF) and associated bioinformatics is outlined. A previously unknown motif within the peptide sequence of Elongation Factor Tu from Thermus aquaticus was discovered using PMF. This motif was shown to possess potential significance in maintaining structural integrity of the entire molecule.

New Rhizobium leguminosarum flavonoid-induced proteins revealed by proteome analysis of differentially displayed proteins

Proteome analysis was used to establish the first two-dimensional protein map of Rhizobium. R. leguminosarum bv. trifolii strain ANU843 was grown in defined medium in the presence and absence of the flavonoid 7,4'-dihydroxyflavone. Over 1,700 constitutive proteins were resolved, representing about 30% of the estimated genomic output. Proteome analysis of flavonoid-treated cells was done to reveal differentially displayed proteins. The results showed that while the global expression pattern of proteins was largely unaltered by the treatment, four inducible proteins were observed. The four inducible proteins and 20 constitutively expressed proteins were subjected to sequence analysis to provide internal standards for the construction of a two-dimensional Rhizobium protein data base. The identity of 12 proteins, including NodE and NodB, was established. NodE was present throughout the growth of the cells but was diminished in amount in stationary phase cells whereas NodB was not detected in the later stages of growth. Two of the induced proteins sequenced did not match any known nodulation gene product, with one of these being present in mid-late log and stationary phase cells and possessing four consecutive His residues at the N-terminal sequencing was successful with 100 to 200 fmol of protein. Proteome analysis provides a sensitive new tool to examine plant-microbe interactions.

Large-scale protein modelling and integration with the SWISS-PROT and SWISS-2DPAGE databases: the example of Escherichia coli

Knowledge-based molecular modelling of proteins has proven useful in many instances, including the rational design of mutagenesis experiments, but it has generally been limited by the availability of expensive computer hardware and software. To overcome these limitations, we developed the SWISS-MODEL server for automated knowledge-based protein modelling. The SWISS-MODEL server uses the Brookhaven Protein Data Bank as a source of structural information and automatically generates protein models for sequences which share significant similarities with at least one protein of known three-dimensional structure. We have now used the software framework of the server to generate large collections of protein models, and established the SWISS-MODEL Repository, a new database for automatically generated and theoretical protein models. This repository is directly integrated with the SWISS-PROT and SWISS-2DPAGE databases through the ExPASy World Wide Web server (URL is http://expasy.hcuge.ch). Here we present an illustration of this process by an application to the Escherichia coli sequences.

The Dictyostelium discoideum proteome--the SWISS-2DPAGE database of the multicellular aggregate (slug)

The cellular slime mold Dictyostelium discoideum is a eukaryotic microorganism which has developmental life stages attractive to the cell and molecular biologist. By displaying the two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) protein map of different developmental stages, the key molecules can be identified and characterised, allowing a detailed understanding of the D. discoideum proteome. Here we describe the preparation of reference gel of the D. discoideum multicellular aggregate, the slug. Proteins were separated by 2-D PAGE with immobilised pH gradients (pH 3.5-10) in the first dimension and sodium dodecyl sulfate (SDS)-PAGE in the second dimension. Micropreparative gels were electroblotted onto polyvinylidene difluoride (PVDF) membranes and 150 spots were visualised by amido black staining. Protein spots were excised and 31 were putatively identified by matching their amino acid composition, estimated isoelectric point (pI) and molecular weight (M(r)) against the SWISS-PROT database with the ExPASy AAcompID tool (http:// expasy.hcuge.ch/ch2d/aacompi.html). A total of 25 proteins were identified by matching against database entries for D. discoideum, and another six by cross-species matching against database entries for Saccharomyces cerevisiae proteins. This map will be available in the SWISS-2DPAGE database.

Large-scale amino-acid analysis for proteome studies

Amino-acid analysis is a relatively new method for identification of proteins separated by two-dimensional gel electrophoresis and blotted onto polyvinylidene difluoride (PVDF) membranes. This article describes modified amino-acid analysis methods for this purpose. Streamlined sample handling is a key feature of the process. To minimise sample manipulation, a single vial is used for hydrolysis and the protein hydrolysate on PVDF membrane is extracted by a one-step procedure. The hydrolysate should not be stored for long periods before analysis. Applications of the technique are presented to demonstrate the identification procedure. This approach is the most cost-effective and time-effective first step in mass protein screening for a large-scale proteome project.

Application of combined mass spectrometry and partial amino acid sequence to the identification of gel-separated proteins

The combined use of peptide mass information with amino acid sequence information derived by chemical sequencing or mass spectrometry (MS)-based approaches provides a powerful means of protein identification. We have used a two-part strategy to identify proteins from nerve growth factor (NGF)-stimulated rat adrenal pheochromocytoma cell line PC-12 cell lysates that associate with the adaptor protein Shc (Shc homologous and collagen protein). Initial experiments with metabolically radiolabeled cell extracts separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) revealed a number of proteins that coimmunoprecipitated with anti-Shc antibody compared with control (unstimulated) cell extracts. The experiment was scaled up and cell lysate from NGF-stimulated PC-12 cells was applied to a glutathione-S-transferase (GST)-Shc affinity column, eluted, separated by SDS-PAGE and blotted to Immobilon-CD. The blotted proteins were proteolytically digested in situ, and the masses obtained from the extracted peptides were used in a peptide-mass search program in an attempt to identify the protein. Even if a strong candidate was found using this search, an additional step was performed to confirm the identification. The mixtures were fractionated by reversed-phase high-performance liquid chromatography (RP-HPLC) and subjected to chemical sequencing to obtain (partial) sequence information, or post-source decay (PSD-) matrix-assisted laser-desorption ionization (MALDI)-MS to obtain sequence-specific fragment ions. This data was used in a peptide-sequence tag search to confirm the identity of the proteins. This combined approach allowed identification of four proteins of M(r) 43,000 to 200,000. In one case the identified protein clearly did not correspond to the radiolabeled band, but to a protein contaminant from the column. The advantages and pitfalls of the approach are discussed.

Identification of proteins by their amino acid composition: an evaluation of the method

Expression of different genomes can be studied by high-resolution two-dimensional electrophoresis (2-D PAGE). To help these studies, two-dimensional reference maps of different biological tissues and fluids have been built and can be found in the SWISS-2DPAGE database, accessible via the World Wide Web network on the ExPASy molecular biology server. Different techniques were used to identify the polypeptides. At the present time, the method considered to be the fastest and the most cost-effective is amino acid composition analysis (AAC). Proteins, transferred onto polyvinylidene (PVDF) membranes, were submitted to vapor-phase hydrolysis, derivatized with 9-fluorenylmethyl chloroformate (FMOC) and separated on an ODS-Hypersil column. Identification was obtained by using the program 'AACompIdent' available from ExPASy. In this work, different experimental parameters, such as contamination, reproducibility and accuracy, have been assessed. First, it has been found that a major source of contamination was human keratin. Next, amino acids have been classified into 'reliable' and 'nonreliable'. Accordingly, 'bias' and 'weights' were defined for each amino acid, which could be set in the 'AACompIdent' program. Finally, examples of identification, including the use of Edman degradation sequence tagging, are described.

Characterization of human plasma glycoproteins separated by two-dimensional gel electrophoresis

Purification of protein isoforms for the characterization of post-translational modifications, such as glycosylation, can be laborious and demanding. We report a means of determining monosaccharide composition and the identity of glycoproteins from a single spot on a two-dimensional (2-D) gel. The sensitivity of the method depends on the degree of glycosylation of the protein. We show that bovine fetuin can be analyzed and identified at the level of 100 pmol. 2-D reference maps enable quick identification of glycoprotein isoforms, and the nature of glycosylation differences. Human sera glycoforms were isolated by micropreparative 2-D PAGE using a narrow-range immobilized pH gradient. Single spots excised from one polyvinylidene difluoride blot of a 2-D gel were used sequentially for sialic acid analysis, neutral and amino sugar analysis, and finally amino acid analysis. The glycosylation variations in isoforms of human fetuin and alpha-1-antitrypsin were determined. The amino acid composition, in conjunction with protein pI and MW, successfully identified the glycoproteins.

Different electrophoretic techniques produce conflicting data in the analysis of myocardial samples from dilated cardiomyopathy patients: protein levels do not necessarily reflect mRNA levels

A variety of electrophoretic techniques were used to search for potential causes of human dilated cardiomyopathy (DCM). Northern blots were used to quantify alpha-cardiac and alpha-skeletal muscle actins, and beta-myosin heavy chain mRNAs which are the predominant expressed isoform species. We found a wide range of mRNA levels expressed in both DCM and nondiseased (ND) samples of left ventricles. However, sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gels of the same heart samples revealed a stable and constant ratio of actin and myosin. Dystrophin deficiency might account for the DCM symptoms and so dystrophin levels of DCM and ND samples were evaluated using Western blots probed with monoclonal antibodies for the N-, C- and mid-rod portions of this protein. We found that dystrophin levels were constant in all 29 DCM and 5 ND samples suggesting that dystrophin deficiency is probably not a contributing cause. We explored the possibility that terminal failure may be due to an apoptotic-like event in the cardiomyocytes. Zymograms of DCM and ND samples revealed a significant increase in DNase I activity in the DCM group compared to the ND samples. These data raise the possibility that end-stage failure may be associated with apoptosis.

From proteins to proteomes: large scale protein identification by two-dimensional electrophoresis and amino acid analysis

Separation and identification of proteins by two-dimensional (2-D) electrophoresis can be used for protein-based gene expression analysis. In this report single protein spots, from polyvinylidene difluoride blots of micropreparative E. coli 2-D gels, were rapidly and economically identified by matching their amino acid composition, estimated pI and molecular weight against all E. coli entries in the SWISS-PROT database. Thirty proteins from an E. coli 2-D map were analyzed and identities assigned. Three of the proteins were unknown. By protein sequencing analysis, 20 of the 27 proteins were correctly identified. Importantly, correct identifications showed unambiguous "correct" score patterns. While incorrect protein identifications also showed distinctive score patterns, indicating that protein must be identified by other means. These techniques allow large-scale screening of the protein complement of simple organisms, or tissues in normal and disease states. The computer program described here is accessible via the World Wide Web at URL address (http:@expasy.hcuge.ch/).