Correlation versus causation: Helicobacter pylori population heterogeneity complicates the identification of mutant...

Helicobacter pylori causes cancer in approximately 1% of infected individuals. A proposed mechanism of H. pylori persistence centers on the ability of the pathogen to form biofilms, yet little is known about specific genetic requirements for this process. Our investigation revealed that during lab passage, H. pylori accumulates genetic changes that impact further phenotypic analyses. Specifically, we first sought to characterize the roles of the flagellar genes, pflA and flgS, in biofilm formation; the flgS mutant strain was biofilm deficient, but the pflA mutant strain was a hyper-biofilm former; however, the pflA mutant strain phenotype was unstable. Analysis and screening of six new pflA mutant strains revealed variable biomass phenotypes. This unexpected result led us to explore how genetic heterogeneity within an H. pylori population may complicate standard mutagenesis processes and the interpretation of downstream phenotypes. Analysis of single colony isolates from multiple wild-type strains similarly yielded different biomass phenotypes despite the expected isogenic nature of these isolates. Genomic sequencing of a subset of these isolates revealed various nucleotide changes. Analysis of some of these changes revealed that mutation of futB, babA, or babB did not affect biofilm formation, while mutation of sabA, which encodes the SabA adhesion, resulted in a significant decrease in H. pylori biofilm formation. Overall, these findings reveal that nucleotide changes that occur during a single passage of H. pylori may impact downstream phenotypic analyses. Moreover, these data emphasize the necessity of genetic confirmation redundancy and/or complementation to conclusively move from correlation to causation when analyzing phenotypes of constructed mutant strains.

H. pylori displays a high rate of genetic variability, but many studies still do not utilize independent confirmation or complementation to conclusively demonstrate that mutations of interest are responsible for identified phenotypes. Our attempts to study biofilm formation were stymied by the realization that individual colonies cultured from wild-type strains had numerous genetic changes despite their expected isogenic nature; these changes correlated with phenotypic differences for individual wild-type isolates. Analysis of a subset of these genetic changes revealed that correlation and causation were not always linked. However, constructed mutations and natural variation in sabA both dramatically decreased biofilm formation. Overall, the extensive genetic heterogeneity that exists within individual cells within an H. pylori population may affect phenotypes of interest; this serves to emphasize the necessity of redundant methods of strain construction, sequence confirmation, and/or genetic complementation to conclusively move from correlation to causation for any phenotype of interest.