Evolution of bacterial genomes

Bacteria have enormously diverse genomes, varying widely in size, composition, and gene copy number. We are trying to understand factors that are responsible for the generation and maintenance of this variation. Currently, we focus on testing the fitness impacts of synonymous codon variation and tRNA genes, and the evolution of genomic GC content. We use various tools such as bioinformatics, experimental evolution, phylogenetic comparative methods, and whole-genome sequencing for this work. These projects are supported by a DST INSPIRE Faculty award and a CSIR research grant.

Evolution of codon use

Earlier, we showed that altering codons in a key enzyme-coding gene (fae) in Methylobacterium extorquens decreases enzyme production and reduces fitness (Agashe et al 2013, MBE). These low-fitness mutants rapidly evolved in the lab to regain near-wild type fitness. Astonishingly, this adaptation was most often mediated by highly repeatable synonymous or non-synonymous point mutations in the 5′ coding region of fae (Agashe et al 2016, MBE). We found that none of the existing hypotheses could explain the selective benefit of these mutations. Neither mRNA structure or stability nor the presence of Shine-Dalgarno (SD)-like motifs could predict fitness consequences of point mutations (the SD sequence is typically found upstream of bacterial genes, and serves as the ribosomal binding site for translation). Nevertheless, our work clearly showed that the initial fitness defects in mutants were not due to altered codon use per se, but due to some other local sequence characteristics that could be easily rescued. Importantly, our results caution against using codon-altered viruses as an “evolution proof” strategy for making vaccines.

Previous work had suggested that internal SD-like motifs generally evolve under purifying selection in bacteria, and may explain observed patterns of codon use. However, our work with M. extorquens contradicts this idea. Hence, we further explored the frequency of internal SD-like motifs across bacteria (Diwan & Agashe 2016, GBE). We found that it is critical to account for genome GC content while analysing bacterial sequence motifs, because GC-rich motifs (such as the SD sequence) may occur more often in GC-rich genomes simply by chance. Accordingly, our results suggested that internal SD-like motifs may face stronger purifying selection in GC-rich genomes. We also found that SD-like motifs are not universally avoided, and many even evolve under positive selection in specific positions, such as at the 3′ ends of genes. Overall, our work suggested that attributing the evolution of sequence motifs in bacterial genomes to a single selective pressure is simplistic, and reality might be much more complicated.

Currently, Saurabh is using phylogenetic methods to map changes in codon use across bacteria, and test the relative impact of various factors that may influence major shifts in codon use. We are also testing whether our prior results with fae are repeatable for two other essential genes in M. extorquens mauA and mtDA.  This work is spearheaded by Nilima.

Evolution of tRNA gene pools

Codon use and tRNA genes are thought to be co-dependent due to translational selection acting on highly expressed genes. However, there are few systematic analyses of fitness consequences of changes in tRNA gene copy number or expression levels. To complicate matters further, tRNA modifying enzymes that increase the number of codons recognized by specific tRNAs are proposed to “buffer” cells against tRNA copy number changes. Thus, the strength of selection acting on tRNA gene copy number remains unclear. We are addressing this problem using E. coli, comparing fitness effects of altering tRNA pools by changing tRNA copy numbers or by altering tRNA modifying enzymes. Simultaneously, we are mapping the evolution of tRNA modifying enzymes as well as tRNA gene copy numbers across bacteria. This work is led by Gaurav and Saurabh.

Finally, Laasya and Parth are analysing the evolutionary consequences of altering levels of initiator tRNA (tRNAi, a special tRNA that incorporates the first methionine in a polypeptide, and is critical for translation initiation). tRNAi levels drop in multiple stressful conditions, and Laasya is using genetic mutants to ask whether, and why, this drop may be beneficial.

Evolution of genome GC content

For decades, biologists have debated the evolutionary causes of variation in genome GC content across bacteria. Are changes in GC content driven by selection or neutral processes, or a combination of both? What are the selective pressures responsible? We are addressing these questions using both experimental and phylogenetic methods. Mrudula has generated E. coli mutants lacking various DNA repair enzymes, which introduce mutational biases. She is testing the evolutionary consequences of altered mutation spectra under conditions proposed to generate strong positive selection for (or against) high GC content (such as nitrogen availability). These experiments will provide direct empirical evidence for the hypothesis that selection may drive evolutionary changes in genome GC content, at least in the short term.

To understand the evolution of GC content across larger timescales, Saurabh is testing hypotheses about mutation-selection balance using phylogenetic methods. He is analyzing the rate of change in GC content in various groups of closely related bacteria to test the timescale required to reach the expected equilibrium GC content. Ultimately, we hope to reconcile our findings from experimental evolution and from phylogenetic approaches to better understand the causes and consequences of changes in bacterial GC content.