DNA Microarray Analysis of Bacterial Pathogens

A DNA microarray is a high-density array of nucleic acid targets immobilized on a glass slide or a silicon chip. The nucleic acids are either denatured complementary DNA or genomic products amplified by PCR for spotted DNA arrays. For oligonucleotides arrays, oligonucleotides complementary to specific gene sequences are generally synthesized in situ. DNA microarray was first used in parallel detection and the analysis of the expression patterns of thousands of genes in plant tissues (7) and human cells exposed to different conditions (8). Since complete bacterial genome sequences have been available, DNA microarrays have been used to analyze genome-wide gene expression of many bacteria and also to compare the whole genetic content of the sequenced bacterial strains with those of closely related species or other isolates of the same species (9,10). As discussed in Subheadings 3. and 4., microarray has facilitated the identification of putative virulence determinants and genes that determine host specificity.

The genus Brucella consists of six species, B. melitensis (primarily in goats and sheep), B. abortus (cattle), B. suis (pig), B. canis (dogs), B. ovis (sheep), and B. neotomae (wood rats). B. abortus, B. melitensis, and B. suis are also pathogenic to humans; all causing similar serious disease consequences. The complete genome sequence of B. melitensis

16M was determined and a whole-genome high-density oligonucleotide DNA micro-array was developed and used to compare genomes of other Brucella species (11). Hybridization results showed that the great majority of the ORFs in B. melitensis 16M are present in all five Brucella genomes examined, supporting the suggestion that Brucella is a monospecific genus with limited genetic diversity (12). Many of the putative deleted genes are clustered into nine GEIs. Comparison of B. melitensis 16M with two Brucella species that are nonpathogenic for humans, B. ovis REO198 and B. neotomae 5K33, is of great interest because it may identify genes that are responsible for pathogenicity in humans. Microarray hybridization, followed by PCR cloning and sequencing, identified 84 ORFs that are found in the 16M genome, but are absent or partially deleted in the genome of B. ovis REO198. Eighty of the 84 ORFs were clustered in five regions corresponding to GEIs (GI-1, GI-2, GI-5, GI-7, and GI-9) of the 16M genome (11). The genome of B. neotomae 5K33 is very similar to that of 16M; only 17 ORFs were identified as deleted or altered, including a region containing 10 contiguous ORFs corresponding to about 7.5 kb in the GI-6 of 16M. The GEIs (GI-1, GI-2, GI-5, GI-7, and GI-9) missing in B. ovis, are present in Brucella species that are pathogenic to humans. However, B. neotomae, a species not pathogenic to humans, also has these islands. It is possible that some of these genes in B. neotomae are expressed at a much lower level or are inactivated, because microarray hybridization cannot identify minor base changes and thus will not detect most pseudogenes. Further studies involving proteomic analysis and mutagenesis of specific genes are needed to understand the role of these GEIs in host adaptation and virulence of the Brucella species.

Porphyromonas gingivalis is a Gram-negative oral anarobe associated with periodontal disease in humans. The genome of a virulent strain of P. gingivalis W50 was recently sequenced (13). A DNA microarray was prepared from PCR amplicons derived from the predicted ORFs of this virulent strain. This microarray was used as a reference to compare the genome of W50 with that of a nonvirulent strain, ATCC33277, with the hope of identifying genes that are required for virulence. Hybridization results showed that 154 ORFs (7% of the total) that are present in W50 are highly divergent or absent in the nonvirulent strain ATCC33277. Interestingly, many of these divergent genes are clustered into three regions of the W50 genome with a lower guanine and cytosine content than the rest of the genome, corresponding to GEIs which are clusters of genes that were likely acquired through horizontal gene transfer. These regions contain putative virulence genes, including genes for capsular polysaccharide synthesis, a lipoprotein gene (rag B), and many species-specific hypothetical genes. These GEIs should be the focus for a future pathogenesis study of this pathogen.

Chlamydia trachomatis is an obligate intracellular pathogen of humans. Isolates are differentiated into biovars based on their in vitro infection properties and type of disease caused. They are grouped into 15 serovars based on antigenic variation of the major outer membrane protein, OmpA. Serovars A to C are the etiological agents of trachoma; serovars D to K and L1 to L3 cause cervicitis and urethritis or lymphogranuloma venereum , respectively (14). The genomes ofthree Chlamydia species, C. trachomatis, C. pneumoniae, and C. psittaci have been sequenced (15-17). Comparative analysis of these genomes (1.04-1.3 Mb) showed a high degree of conservation in terms of gene content and gene order with the exception of one polymorphic region named the plasticity zone (PZ), which showed a significant amount of variation among the different species. A DNA microarray was developed based on the genomic sequence of C. trachomatis serovar D and used in analyzing the genomic diversity of the 15 serovars. The hybridization results showed that the genomes of the 15 serovars are highly conserved (>99%). In contrast, similar studies done with isolates of Helicobacter pylori (18) and Staphylococcus aureus (19) identified major intraspecies diversity. Variation in the OmpA gene was observed as expected, and was confirmed by gene cloning and sequencing. Serovar B showed the greatest diversity with a maximum of eight deleted genes. Deletions were also observed with serovars A, Ba, C, L1, L2, and L3. No gene deletion was observed in serovars E, F, G, H, I, J, and K. Without exception, all the deleted genes observed localized to the PZ region (20). To precisely define the PZ region of C. trachomatis, PCR primers were designed to amplify the region as four overlapping fragments and sequenced. PCR amplification results showed that most of the ocular and genital serovars have a large internal deletion compared with the intact cytotoxin gene of C. muridarum, a mouse-adapted strain. Deletions in the ocular serovars (A, Ba, and C) are smaller than those observed in the genital serovars (D, E, F, G, I, and K).

The DNA sequences obtained from all the serovars were aligned with the mouse-adapted C. muridarum cytotoxin (TC0438) gene. Analysis of the sequence showed that each serovar possessed a unique combination of DNA sequences and ORFs, but none encoded a full-length cytotoxin. It appears that the genital serovars have a large ORF that could encode both the uridine diphosphate (UDP)-glucose binding and the gly-cosyltransferase domains, while the ORF of the ocular serovars could encode only the UDP-glucose binding domain. Deletion in the genome of the lymphogranuloma venereum serovars removed the encoding sequence of both domains. These results suggest the UDP-glucose binding and the glycosyltransferase domains of the chlamydia cyto-toxin, a homolog of the clostridial toxin, may have an important role in urogenital infection (20).

Microarrays have also been used to analyze intraspecies diversity and to identify strain-specific genes of many bacterial pathogens, including H. pylori, Streptococcus pneu-moniae, Salmonella enterica serovar Typhimurium, Campylobacter jejuni (21,22), Vibrio cholerae (23), E. coli, and many others. Microarray analysis of 15 H. pylori strains showed great genetic diversity, 362 genes (22% of all H. pylori genes) are not conserved among all 15 strains (9).

An Affymetrix high-density oligonucleotide array based on the genomic sequence of S. pneumoniae serotype 4 strain was used to analyze genetic diversity of 20 clinical S. pneumoniae isolates and 9 oral streptococcal isolates. Most of the S. pneumoniae strains differed from the sequenced strain by 8-11%. In contrast, the nine oral streptococci strongly diverged from the sequenced strain, as only 15-61% of their DNA hybridized with the reference array. This study identified 470 S. pneumoniae strain-specific genes, which were not detected in at least one of the strains examined (24). A desired future goal is to be able to associate a specific gene(s) lost to a specific disease phenotype.

S. pneumoniae infection can lead to bacteremia and meningitis. Invasive diseases are the result of spread of the bacteria from the nasopharynx, the site of colonization, to the lung and bloodstream with possible sequelae of septicemia or meningitis. With the aim of identifying virulence determinants of S. pneumoniae, a whole-genome microarray was used to analyze the bacterial genes that are expressed in vivo. The study involved isolating total RNA from S. pneumoniae isolated from infected blood, infected cere-

brospinal fluid (CSF), and bacteria attached to a pharyngeal epithelial cell (ECC) line in vitro. Gene expression levels at these three sites were compared with levels when S. pneumoniae was grown in semisynthetic casein liquid medium. Such in vivo studies are limited by the difficulties of obtaining sufficient quantities of pure and relatively intact bacterial RNA from infected tissues. Interestingly, the majority of the genes (92% in the blood, 85% in CSF, and 90% after ECC) were expressed in a similar fashion as growth in culture medium. However, distinct patterns of gene expression in each anatomical site can be identified. Amazingly, only eight genes showed similar alterations in gene expression during bacterial growth in blood, in CSF, or during ECC (25). Two of the eight genes encode pspA (26) and prtA (27) previously characterized as virulence factors, three involved in manganese acquisition and transport (psa operon), two in energy metabolism and one transporter. Orihuela et al. (25) postulated that these eight genes may be part of the core set of genes required for virulence and, therefore, deserve further investigation.

DNA microarrays also were successfully used in analyzing whole-genome gene expression (transcriptome) of uropathogenic E. coli strain CFT073 during urinary tract infection of CBA/J mice (28). Total RNA was isolated from CFT073 bacteria obtained directly from the urine of infected mice. The in vivo transcription profiles were compared with those of CFT073 grown statically to exponential phase in rich medium. Overall, transcription of 313 genes was found elevated, whereas that of 207 genes was reduced. Of the 313 CFT073 genes that were to be elevated, only 45 genes were unique to the uropathogenic strain and not found in nonpathogenic E. coli K12. The author proposed that these 45 are candidate virulence genes for urinary tract infection. Twenty-five of these genes have previously been implicated in virulence. These include genes involved in iron acquisition, capsule synthesis, and synthesis of microcin secretion proteins. Thirteen new candidate virulence genes encoding hypothetical proteins were also identified.

0 0

Post a comment