Amino Acid Cost and Codon-Usage Biases in 6 Prokaryotic Genomes: A Whole-Genome Analysis

Abstract

For most prokaryotic organisms, amino acid biosynthesis represents a significant portion of their overall energy budget. The difference in the cost of synthesis between amino acids can be striking, differing by as much as 7-fold. Two prokaryotic organisms, Escherichia coli and Bacillus subtilis, have been shown to preferentially utilize less costly amino acids in highly expressed genes, indicating that parsimony in amino acid selection may confer a selective advantage for prokaryotes. This study confirms those findings and extends them to 4 additional prokaryotic organisms: Chlamydia trachomatis, Chlamydophila pneumoniae AR39, Synechocystis sp. PCC 6803, and Thermus thermophilus HB27. Adherence to codon-usage biases for each of these 6 organisms is inversely correlated with a coding region's average amino acid biosynthetic cost in a fashion that is independent of chemoheterotrophic, photoautotrophic, or thermophilic lifestyle. The obligate parasites C. trachomatis and C. pneumoniae AR39 are incapable of synthesizing many of the 20 common amino acids. Removing auxotrophic amino acids from consideration in these organisms does not alter the overall trend of preferential use of energetically inexpensive amino acids in highly expressed genes.

Chlorobaculum tepidum Modulates Amino Acid Composition in Response to Energy Availability, as Revealed by a Systematic Exploration of the Energy Landscape of Phototrophic Sulfur Oxidation

“…Efficiency in the primary sequence of proteins has been observed previously; highly expressed bacterial genes (as assessed by major codon usage) tend to code for proteins that use less energetically expensive amino acids (31)(32)(33), and extracellular proteins tend to be composed of less energetically expensive amino acids than intracellular and membrane proteins (45). Furthermore, a survey of yeast transcriptomic data found that resource limitations (e.g., salt stress, reduced exogenous amino acid availability, etc.)…”

“…C. tepidum displayed shifts in biomass amino acid composition, cell volume, and storage carbohydrate content across the energy landscape, suggesting that C. tepidum alters its physiology in response to energy availability to a greater extent than previously appreciated. While bias in the amino acid composition of highly expressed proteins toward energetically inexpensive amino acids has been inferred by bioinformatic analyses for a range of organisms (31)(32)(33), this work reports an experimentally measured shift in bulk amino acid composition for a single organism as a function of growth conditions. This observation implies that relatively simple bulk biomass measurements can be used to infer details about the energy status and adaptation of microbes.…”

A Genetic Optimization Approach for Isolating Translational Efficiency Bias

“…The gene with the smallest CAI value is assigned an index of 0 while the gene with greatest CAI value is assigned an index of N À 1, where N is the number of genes in the genome. The indexes of the reference set genes are summed (IDXs in (10)) and this number is normalized by the maximum attainable rank fitness value (11) yielding a value between 0 and 1 representing the degree to which the reference set rises to the top of a sorted (by CAI score) list of genes (12)…”

“…The degree to which a gene adheres to translational efficiency bias has been shown to correlate with the gene's expression level in many Bacteria [4], [5], [6] and some low-order Eukarya [7], [8]. This trait has been exploited in studies that have used adherence to translational efficiency bias as a proxy for expressivity [9], [10], [11], [12]. Use of translational efficiency bias as a surrogate for expressivity is useful, in part, because experimental methods for obtaining transcript or protein abundance (such as that provided by oligonucleotide microarrays or gel electrophoresis) are relatively expensive in terms of time, materials, and reagent cost.…”

The Evolution of Protein Material Costs

“…For instance, highly expressed genes tend to encode shorter proteins (Brocchieri and Karlin 2005), and to contain fewer introns (Castillo-Davis et al 2002), than less highly expressed genes. Proteins encoded by highly expressed genes also contain fewer energetically expensive or heavy amino acids (Akashi and Gojobori 2002;Heizer et al 2006;Seligmann 2003). Similarly, highly expressed genes encode proteins with relatively low material costs for nutrients that are commonly limiting.…”