Transcription profiling-based identification of Staphylococcus aureus genes regulated by the agr and/or sarA loci.

The advent of transcription profiling technologies has provided researchers with an unprecedented ability to study biological processes. Accordingly, a custom-made Affymetrix GeneChip, constituting >86% of the Staphylococcus aureus genome, was used to identify open reading frames that are regulated by agr and/or SarA, the two best-studied regulators of the organism's virulence response. RNA extracted from wild-type cells and agr, sarA, and agr sarA mutant cells in the early-, mid-, and late-log and stationary phases of growth was analyzed. Open reading frames with transcription patterns expected of genes either up- or downregulated in an agr- and/or SarA-dependent manner were identified. Oligonucleotide microarray and Northern blot analyses confirmed that the transcription of several known virulence genes, including hla (alpha-toxin) and spa (protein A), is regulated by each effector and provided insights about the regulatory cascades involved in both alpha-hemolysin and protein A expression. Several putative virulence factors were also identified as regulated by agr and/or SarA. In addition, genes that are involved in several biological processes but which are difficult to reconcile as playing a direct role in the organism's pathogenesis also appeared to be regulated by each effector, suggesting that products of both the agr and the sarA locus are more-global transcription regulators than previously realized.

Bioinformatics: use in bacterial vaccine discovery.

Bioinformatics has now become a common laboratory name for groups studying genomic sequences. It is composed of many different, yet interrelated scientific fields such as genomics, proteomics, and transcriptional profiling. The availability of complete genomic sequences, especially prokaryotic organisms, allows one to rapidly identify, analyze, and clone genes of interest. For bacterial vaccine discovery, one can "mine" the genomic sequence for potential surface targets using various algorithms, characterize these gene targets, and produce primers for cloning, all before one enters the wet laboratory. This review will focus on various genomic mining tools/algorithms available for predicting open reading frames and their associated annotation (if known), physical and functional characterization, and cellular localization. Finally, examples are given of how all of this is being used for the identification of potential bacterial vaccine candidates.

Application of genomics and proteomics for identification of bacterial gene products as potential vaccine candidates.

The ability of bioinformatics to characterize genomic sequences from pathogenic bacteria for prediction of genes that may encode vaccine candidates, e.g. surface localized proteins, has been evaluated. By applying appropriate tools for genomic mining to the published sequence of Haemophilus influenzae Rd genome, it was possible to identify a putative vaccine candidate, the outer membrane lipoprotein, P6. Proteomics complements genomics by offering abilities to rapidly identify the products of predicted genes, e.g. proteins in outer membrane preparations. The ability to identify the P6 protein uniquely from entries in a sequence database from the expected peptide-mass fingerprint of P6 demonstrates the power of proteomics. The application of proteomics for identification of vaccine candidates for another pathogenic bacterium, Helicobacter pylori using two different approaches is described. The first involves rapid identification of a series of monoclonal antibody reactive proteins from N-terminal sequence tags. The other approach involves identification of proteins in outer membrane preparations by 2-D electrophoresis followed by trypsin digestion and peptide mass map analysis. Our combined studies demonstrate that utilization of genome sequences by application of bioinformatics through genomics and proteomics can expedite the vaccine discovery process by rapidly providing a set of potential candidates for further testing.

Identification of a Haemophilus influenzae 5'-nucleotidase protein: cloning of the nucA gene and immunogenicity and characterization of the NucA protein.

We report on the identification of a surface-exposed, highly conserved, immunogenic nontypeable Haemophilus influenzae (NTHi) protein, which elicits cross-reactive bactericidal antibodies against NTHi. The protein was extracted from NTHi strain P860295 with KSCN and purified; it migrated as a single band on a sodium dodecyl sulfate-polyacrylamide gel with an apparent molecular mass of 63 kDa. Mouse antiserum generated against the purified protein was reactive on whole-cell enzyme-linked immunosorbent assay (ELISA) with seven NTHi strains and type b Eagan and Whittier strains and exhibited bactericidal activity to homologous and heterologous NTHi strains. However, the protein is made in small amounts in NTHi as corroborated by immunoelectron microscopy. To further study this protein, we cloned, sequenced, and expressed it recombinantly in Escherichia coli. The recombinant protein is localized in the periplasm of E. coli and has been purified to homogeneity. Both the recombinant and native proteins possess 5'-nucleotidase activity; hence, the protein has been called NucA. Mouse antiserum directed against the recombinant NucA protein was reactive on Western immunoblots and whole-cell ELISA with all H. influenzae strains tested including Eagan and was bactericidal for two heterologous strains tested. The antiserum also resulted in a log reduction in bacteremia, in an infant-rat protection study with H. influenzae type b as the challenge strain. These features suggest that NucA is a potential subunit vaccine candidate against NTHi disease.

Mining genomes and mapping proteomes: identification and characterization of protein subunit vaccines.

Currently, there is an extensive and unprecedented effort to obtain the complete nucleotide sequence of the complex genomes of many micro-organisms. In this post-genomic era, based on the availability of the entire genome sequence of an organism, three new disciplines of molecular biology have emerged: genomics, transcriptional profiling and proteomics. All these technologies have the potential to accelerate the process of identifying protective protein antigens as subunit vaccine targets as well as validating and extending the range of available candidate antigens. The progress of these technologies has led to the origination of the science of bioinformatics for management and critical evaluation of the large amount of information generated. Although genomics, transcriptional profiling and proteomics are each based on different principles, there is considerable synergy between them. Appropriate application of any one, or a combination of two or more of these approaches, coupled with bioinformatics, would allow identification of a short-list of vaccine candidates from the entire list of several hundreds to thousands of proteins encoded by the genome. These candidates would then require usual channelling through the subsequent process involving recombinant expression, purification and testing for immunogenicity and protective efficacy.

Quantitation of cellular receptors by a new immunocytochemical flow cytometry technique.

Flow cytometry provides a rapid qualitative method for analyzing the binding of a fluorescent probe to a cell. To quantitate the binding of a probe using flow cytometry, one must be able to first calibrate the fluorescent signal with some known standard. We have compared a new one-step method with a previous two-step method for determining the number of binding sites (receptors) on the surface of cells using immunofluorescent staining of the cells and analysis by flow cytometry. Experimentally, recombinant chinese hamster ovary cells, expressing cell surface glycoprotein receptors, IIb/IIIa or Mac-1, were assayed using specific mouse monoclonal antibodies against these receptors. The two methods yielded comparable results and, depending on the cell type, detected anywhere from 100,000 to 300,000 antibody-binding sites per cell, respectively.

Construction of infectious molecular clones of HIV-1 containing defined mutations in the protease gene.

A DNA clone of HIV-1 containing the full-length infectious viral sequence was cleaved at a unique Nco I restriction site within the viral genome, and DNA fragments containing the 5' and 3' portions of the HIV genome were subcloned into separate plasmid vectors. The 5' 'half-virus' construct was further modified by incorporating a class IIS restriction site, Esp3I, near the 3' end of the protease gene of HIV. This site, in combination with a natural ApaI site near the 5' end of the protease gene, creates a convenient cassette shuttle vector in which the protease coding region can be easily replaced. Recombinant viruses containing protease genes either altered by site-directed mutagenesis or amplified from clinical or laboratory isolates can be reconstructed. The DNA fragment containing the protease gene is first subcloned into the 5' half-virus shuttle vector plasmid. Infectious recombinant virus is subsequently recovered by cotransfecting 5' and 3' half-virus plasmids linearized at their common Nco I sites into mammalian cells. This method was successfully applied to constructing viruses containing various substitutions in protease.

Detection of single base differences using biotinylated nucleotides with very long linker arms.

A simple primer extension method for detecting nucleotide differences is based on the substitution of mobility-shifting analogs for natural nucleotides (1). This technique can detect any single-base difference that might occur including previously unknown mutations or polymorphisms. Two technical limitations of the original procedure have now been addressed. First, switching to Thermococcus litoralis DNA polymerase has eliminated variability believed to be due to the addition of an extra, non-templated base to the 3' end of DNA by Taq DNA polymerase. Second, with the analogs used in the original study, the mobility shift induced by a single base change can usually be resolved only in DNA segments 200 nt or smaller. This size limitation has been overcome by synthesizing biotinylated nucleotides with extraordinarily long linker arms (36 atom backbone). Using these new analogs and conventional sequencing gels (0.4 mm thick), mutations in the human beta-hexosaminidase alpha and CYP2D6 genes have been detected in DNA segments up to 300 nt in length. By using very thin (0.15 mm) gels, single-base polymorphisms in the human APOE gene have been detected in 500-nt segments.

Use of 33P for Sanger DNA sequencing.

The use of [alpha-33P]deoxyadenosine 5'-triphosphate ([alpha-33P]dATP) in DNA sequencing has been described. 33P has a maximum beta-emission energy that is 50% stronger than 35S, but fivefold weaker than 32P. As a result, sequences generated using [alpha-33P]dATP have short exposure times like 32P, yet they maintain band resolution similar to 35S. Handling of [alpha-33P]dATP is straightforward because no special lead or Plexiglas shielding is necessary.

In vitro production of large single-stranded templates for DNA sequencing.

A method has been developed for the preparation of large single-stranded DNA sequencing templates from primary cloning plasmids or cosmids. The method utilizes the separate action of T7 Gene 6 exonuclease and exonuclease III to generate large quantities of single-stranded template for each strand of a large-cloned fragment. Since the procedure is intended for use on primary clones, it avoids the time-consuming subcloning steps associated with most sequencing programs. The procedure also has the advantage of avoiding clone instability problems associated with subcloning in M13.

Improvements in the chain-termination method of DNA sequencing through the use of 7-deaza-2'-deoxyadenosine.

Significant improvements in the quality of DNA sequencing data have been shown when deoxyadenosine triphosphate (dATP) is replaced by 7-deaza-2'-deoxyadenosine triphosphate (c7dATP). The use of c7dATP in conjunction with 7-deaza-2'-deoxyguanosine triphosphate (c7dGTP) further decreases anomalies in electrophoretic mobility which are caused by compressions involving G and/or A residues. This effect is observed for both isotope-based and fluorescence-based sequencing approaches. Replacing dATP with c7dATP also results in a higher degree of uniformity in the frequency of chain termination reactions, when such terminations involve the incorporation of fluorescence-labeled dideoxynucleotides by T7 polymerase. These improvements in the gel-resolution and distribution of chain-terminated DNA products result in higher accuracy in both manual and automated base assignment.

DNA sequencing separations in capillary gels on a modified commercial DNA sequencing instrument.

DNA sequencing separations of standard DNA fragments of known sequence have been achieved in small diameter capillary gels electrophoresed and analyzed in parallel in a modified commercial DNA sequencer instrument. DNA sequencing in terms of base-calling accuracy is comparable to conventional slab gels; however, the separations in the capillary were performed somewhat faster and required less sample than those in the slab gel. Advantages of this approach vs. separations on conventional slab gels are discussed.

Mutational analysis of the lambda int gene: DNA sequence of dominant mutations.

We have combined techniques of genetic and physical mapping with rapid DNA sequence analysis to identify the nucleotide change in lambda int mutations. These mutations define two dominant phenotypic classes: (i) recombination that is partially independent of accessory factors, and (ii) inhibition of wild-type Int by missense or nonsense proteins, i.e., negative complementation.

A system for rapid DNA sequencing with fluorescent chain-terminating dideoxynucleotides.

A DNA sequencing system based on the use of a novel set of four chain-terminating dideoxynucleotides, each carrying a different chemically tuned succinylfluorescein dye distinguished by its fluorescent emission is described. Avian myeloblastosis virus reverse transcriptase is used in a modified dideoxy DNA sequencing protocol to produce a complete set of fluorescence-tagged fragments in one reaction mixture. These DNA fragments are resolved by polyacrylamide gel electrophoresis in one sequencing lane and are identified by a fluorescence detection system specifically matched to the emission characteristics of this dye set. A scanning system allows multiple samples to be run simultaneously and computer-based automatic base sequence identifications to be made. The sequence analysis of M13 phage DNA made with this system is described.

A shuttle vector plasmid for studying carcinogen-induced point mutations in mammalian cells.

We have constructed a shuttle vector plasmid for studying mutagenesis in mammalian cells. The plasmid replicates in cell lines permissive for SV40 virus as well as in the bacterium Escherichia coli and carries a bacterial suppressor tRNA gene (supF) that can serve as a mutagenesis marker. The plasmid replicates as efficiently as SV40 virus in African Green Monkey kidney CV1 cells, indicating that all traces of the inhibitory sequences normally found in pBR322 and its derivatives have been removed. The design of the plasmid and the small size of the mutagenesis target gene decrease the probability of recovering spontaneous deletion mutations that have been shown to occur at high frequency during passage in mammalian cells. The frequency of spontaneous-mutant plasmids recovered after passage in CV1 cells is substantially lower than with other vectors described previously. When the plasmid DNA is treated with UV radiation before passage in CV1 cells, mutants are observed at a frequency about 20-fold above the spontaneous background.

Cloning vectors that yield high levels of single-stranded DNA for rapid DNA sequencing.

We have constructed chimeric plasmid vectors with the origin and intergenic region from M13 phage cloned into the PvuII ( pZ145 ) and AhaIII ( pZ150 , pZ152 ) sites of pBR322. In the absence of M13 phage, these plasmids replicate like any other ColE1-derived plasmid and confer both ampicillin and tetracycline resistance (Amp, Tet). Upon infection with M13 phage, the viral origin present on the plasmids permits phage-directed plasmid replication and results in high yields of single-stranded (ss) plasmid DNA in M13-like particles. This ssDNA, which represents only one of the plasmid strands, is useful as a substrate for rapid DNA sequence determination by the dideoxy sequencing method described by Sanger et al. (1977). Since these plasmids contain an intact pBR322, the intergenic region can be transferred onto most pBR322 derivatives to yield ss plasmid DNA without affecting the recipient plasmid for further studies. We also constructed a deletion derivative of pZ145 , plasmid pZ146 , that does not exhibit interference with the growth of the M13 helper, although this plasmid is encapsidated into phage particles. This result confirms the theory that the intergenic region consists of two domains: one domain being a segment involved in phage morphogenesis and the other being a region of functional origin which interferes with M13 replication.

Nonideal statistics and positive correlation in phage recombination: studies with lambda tandem duplication phages.

The question of nonideality in phage recombination, that is, the extent to which recombinant frequencies differ from those expected from the proportions of the two parental types in the mass culture, was addressed by experiments with lambda tandem duplication phages. Isolation and genotypic analysis of triplication-phage progeny, all of which must be the result of intermolecular recombination, yielded a value of about 0.5 for the nonideality parameter h, i.e., the frequency of unlike-parent matings was only about 1/2 the "ideal" value. This value was independent of multiplicity and about the same for the Rec or Red recombination systems. Similar analysis of single-copy phage progeny yielded estimates of k, the ratio of intramolecular to intermolecular recombination of about 1/6 for the Rec system; no intramolecular events were detected in Red-mediated crosses. Consideration of known nonideality factors (finite input, limited number of intracellular sites for phage growth) suggests that the observed h values correspond to intracellular mixing efficiencies of 55 to 100%, depending on the number of intracellular phage growth sites assumed. Analysis of long-range positive correlation (negative interference) indicates that statistical effects caused unlike-parent double crossovers to be three to four times as frequent as an independent-event calculation would predict. In addition, Rec-mediated crosses showed a 1.3-fold positive correlation for unlike-parent crossovers (in a second interval) among the progeny of like-parent recombinations.

Expression of the phage lambda recombination genes exo and bet under lacPO control on a multi-copy plasmid.

The bacteriophage lambda genes exo and bet, whose products (lambda exonuclease and beta protein, respectively; Red phenotype) mediate homologous recombination of lambda phages, have been cloned under lacPO lacIq control on multi-copy plasmids. Induction of recA3 cells harboring these plasmids with isopropylthiogalactoside (IPTG) resulted in lambda exonuclease levels (assayed in vitro) that were proportional to the time of induction (for at least 4 h); recombination of lambda Red- phages in vivo was similarly inducible. Only one out of 25 bet delta plasmids (constructed by a variety of in vitro techniques) expressed lambda exonuclease, a result consistent with the polarity of several known phage bet mutations. A general method for transferring phage exo and bet mutations to plasmids was devised and plasmids bearing polar (bet3) and nonpolar (bet113) mutations were constructed. Mutant derivatives of the plasmid showed the same complementation pattern as analogous phage red mutants. When lambda bet3 phages (Exo-Bet-) infected IPTG-induced recA3 bacteria containing exo+bet+ plasmids, recombination frequencies were not more than twice those typical for infection of plasmid-free recA3 cells with exo+bet+ phages, even in the case of IPTG induction sufficient to elevate the production of lambda exonuclease about 100-fold. Even when plasmid induction was delayed till as late as 50 min after infection, recombination was significant. Preliminary experiments suggest that these plasmids encode a polypeptide with Gam activity that corresponds to the 98-amino acid "shorter" open reading frame assigned to gam by Sanger et al. (J. Mol. Biol. 162 (1982) 729-773).