Combining Microarray Technology and Molecular Epidemiology to Identify Genes Associated with Invasive Group BStreptococcus

Lixin Zhang, Usha Reddi, Usha Srinivasan, Sheng Li, Stephanie M. Borchardt, Parvathy Pillai, Puja Mehta, Anne N. Styka, Joan DeBusscher, Carl F. Marrs, Betsy FoxmanInterdisciplinary 2008 · DOI

TL;DR

This paper presents a novel three-step approach combining comparative genomic hybridization (CGH) with a large-scale epidemiological screening microarray (LOS) to identify virulence genes in Streptococcus agalactiae (Group B Streptococcus, GBS). By comparing invasive and commensal GBS strains, they successfully pinpointed several candidate genes that are differentially distributed between these groups, offering a powerful new method for understanding bacterial pathogenesis.

What Problem Does This Paper Address?

Streptococcus agalactiae (GBS) is a common commensal bacterium that can cause severe invasive diseases, particularly in neonates and the elderly. While host factors play a role, bacterial virulence factors are crucial for GBS pathogenesis. However, identifying these specific genes has been challenging due to the limited understanding of GBS pathogenesis and the vast number of genes in its genome. Previous studies often relied on comparing a few invasive strains to each other or on traditional genomic subtraction, which can be labor-intensive and may miss important genes. This paper aims to develop and validate a high-throughput, molecular epidemiological approach to efficiently identify and evaluate GBS genes associated with virulence.

Methods

The study employs a three-step strategy:

  1. Comparative Genomic Hybridization (CGH) for Candidate Gene Identification:
    • They selected three representative commensal GBS isolates and one invasive isolate as "tester strains."
    • These tester strains were compared against two publicly available sequenced invasive GBS genomes (serotype III strain NEM316 and serotype V strain 2603VR) using fine-tiling oligonucleotide microarrays.
  2. This CGH step identified open reading frames (ORFs) that were present in the sequenced invasive strains but absent or divergent in the commensal tester strains. This process leverages the power of high-density tiling arrays to detect sequence differences at a genome-wide scale.
  3. Epidemiological Screening with a GBS Library on a Slide (LOS) Microarray:
    • A subset of the candidate variable ORFs identified in step 1 (23 out of 35) were then screened against a much larger collection of 949 GBS isolates using a novel "Library on a Slide" (LOS) microarray.
  4. This LOS platform prints entire bacterial genomes onto slides, allowing for rapid screening of the presence or absence of specific genetic elements across a diverse population of isolates.
  5. The goal here was to evaluate the association of these candidate genes with disease by comparing their prevalence in invasive versus commensal isolates.
  6. Statistical Association Analysis:
    • The prevalence data from the LOS microarray was analyzed to determine if specific ORFs were significantly more common in invasive GBS isolates compared to commensal ones. This statistical association is key to inferring potential virulence roles.

Datasets

  • Bacterial Strains:
    • Tester Strains: Three commensal GBS isolates (G293-061, G617-061, G654-461) and one invasive GBS isolate (H1-19).
  • Reference Strains: Two sequenced invasive GBS strains: NEM316 (serotype III) and 2603VR (serotype V).
  • Screening Collection: 949 GBS isolates, broadly categorized into invasive (n=386) and colonizing (n=563) strains, obtained from various epidemiological studies.
  • Microarrays:
    • Fine-tiling oligonucleotide microarrays: Custom-designed arrays for NEM316 and 2603VR genomes, each containing hundreds of thousands of 32-mer probes.
  • GBS Library on a Slide (LOS) microarray: A novel platform with genomic DNA from 949 GBS isolates printed in duplicate.

Key Insights and Ideas

  • A Powerful Hybrid Approach: The paper demonstrates the synergy of combining high-resolution CGH for initial candidate gene discovery with a high-throughput, population-based LOS microarray for epidemiological validation. This is more efficient than traditional methods.
  • Commensal vs. Invasive Differences: The study highlights that even within the same species, significant genetic differences exist between strains that typically colonize and those that cause invasive disease. These differences are likely key to understanding virulence.
  • Variable ORFs as Virulence Candidates: Identifying ORFs that are consistently absent or divergent in commensal strains but present in invasive strains is a strong indicator of potential virulence factors.
  • The "Unknown Function" Challenge: A significant portion of the identified variable ORFs were of unknown function, underscoring the ongoing need for functional studies to elucidate their roles in pathogenesis.
  • Genome Content Diversity: The LOS microarray screening revealed a striking diversity in the GBS genome content across the 949 isolates, with 503 distinct genotypes observed. This suggests that even strains with similar PFGE patterns can have different gene compositions.

Key Results

  • Candidate Gene Identification: CGH identified a substantial number of variable ORFs (22.7% in NEM316, 25% in 2603VR) between invasive reference genomes and commensal tester strains.
  • Differential Distribution of ORFs: Out of 23 candidate ORFs screened using the LOS microarray, five ORFs were found to be differentially distributed between invasive and commensal isolates. Specifically, four ORFs were more frequent in invasive strains, and one was more frequent in commensal strains.
  • Specific Gene Candidates: Notable candidates identified include:
  • sag2060 and sag2061: Putative glycosyl transferase genes, which are important in bacterial surface structures and can play roles in pathogenesis.
  • sag0832: A fibrinogen-binding protein, previously suggested to be an important virulence factor for invasive GBS disease.
  • sag0814 and gbs0474: Hypothetical proteins that showed differential distribution, suggesting potential roles in complex traits difficult to observe in laboratory conditions.
  • High Reproducibility: The CGH experiments demonstrated high reproducibility (correlation coefficients >90% for duplicates).
  • Sensitivity vs. Specificity: The study noted that while tiling arrays have high sensitivity for detecting conserved sequences, their specificity can be lower, requiring careful cutoff value selection for ORF classification.

Limitations and Future Work

  • Limited Tester Strain Diversity: The authors acknowledge that the small number of commensal tester strains chosen could not capture the full diversity of commensal isolates.
  • Incomplete Coverage of Variable ORFs: Not all identified variable ORFs could be synthesized into good probes for the LOS microarray due to small size or poor PCR amplification, leaving some candidates unevaluated in the large-scale screen.
  • Association vs. Causation: The study identifies associations between gene presence/absence and disease status. Further functional studies are needed to definitively establish a causal role in virulence.
  • Unsequenced Strains: The CGH approach is limited by the availability of sequenced genomes. Potential virulence genes present only in unsequenced pathogenic GBS strains would not be detected.
  • Future Directions:
    • Functional Studies: The authors emphasize the need for functional analyses (e.g., mutagenesis studies) to confirm the roles of the identified candidate genes in GBS pathogenesis.
  • Larger-Scale Screening: They plan to screen the remaining variable ORFs and analyze an additional 2000 isolates to perform a more definitive analysis.
  • Population Structure Integration: Incorporating GBS population structure could enhance the interpretation of their findings.
  • Pyrosequencing: The authors suggest that future genome comparisons might benefit from cheaper and faster direct sequencing approaches like pyrosequencing, which could overcome the limitations of CGH.

Summary updated

Abstract

Many bacterial species function as both commensals and pathogens; we used this dual nature to develop a high-throughput molecular epidemiological approach to identifying bacterial virulence genes. We applied our approach to Group B Streptococcus (GBS). Three representative commensal and one invasive GBS isolates were selected as tester strains from a population-based collection. We used microarray-based comparative genomic hybridization to identify open reading frames (ORFs) present in two sequenced invasive strains, but absent or divergent in tester strains. We screened 23 variable ORFs against 949 GBS isolates using a GBS Library on a Slide (LOS) microarray platform. Four ORFs occurred more frequently in invasive than commensal isolates, and one appeared more frequently in commensal isolates. Comparative hybridization using an oligonucleotide microarray, combined with epidemiologic screening using the LOS microarray platform, enabled rapid identification of bacterial genes potentially associated with pathogenicity.
BibTeX
@inproceedings{zhang2008combining,
  author = {Lixin Zhang and Usha Reddi and Usha Srinivasan and Sheng Li and Stephanie M. Borchardt and Parvathy Pillai and Puja Mehta and Anne N. Styka and Joan DeBusscher and Carl F. Marrs and Betsy Foxman},
  title = {{Combining Microarray Technology and Molecular Epidemiology to Identify Genes Associated with Invasive Group B<i>Streptococcus</i>}},
  booktitle = {Interdisciplinary},
  year = {2008},
  doi = {10.1155/2008/314762},
}
Figure