Transcriptome Analysis of Core Dinoflagellates Reveals a Universal Bias towards "GC" Rich Codons

Abstract

Although dinoflagellates are a potential source of pharmaceuticals and natural products, the mechanisms for regulating and producing these compounds are largely unknown because of extensive post-transcriptional control of gene expression. One well-documented mechanism for controlling gene expression during translation is codon bias, whereby specific codons slow or even terminate protein synthesis. Approximately 10,000 annotatable genes from fifteen "core" dinoflagellate transcriptomes along a range of overall guanine and cytosine (GC) content were used for codonW analysis to determine the relative synonymous codon usage (RSCU) and the GC content at each codon position. GC bias in the analyzed dataset and at the third codon position varied from 51% and 54% to 66% and 88%, respectively. Codons poor in GC were observed to be universally absent, but bias was most pronounced for codons ending in uracil followed by adenine (UA). GC bias at the third codon position was able to explain low abundance codons as well as the low effective number of codons. Thus, we propose that a bias towards codons rich in GC bases is a universal feature of core dinoflagellates, possibly relating to their unique chromosome structure, and not likely a major mechanism for controlling gene expression.

Keywords: codon bias; dinoflagellate; gene expression; toxin.

Figure 1

A guanine cytosine (GC) content versus species plot for the fifteen dinoflagellate transcriptomes used in this study. Species names for the corresponding transcriptome are given on the X-axis with strain numbers when applicable. The Y-axis shows the fraction of the overall nucleotide count of guanine and cytosine base pairs.

Figure 2

Plot of guanine cytosine (GC) content at each codon position versus the overall GC content of the transcriptome for each of the fifteen species. The X-axis shows the GC content of the transcriptome as a fraction of the total number of nucleotides. The Y-axis shows the GC content of the codon positions as a fraction of the total number of nucleotides with codon positions one, two and three colored pink, green, and blue respectively.

Figure 3

Correspondence analysis output from codonW for Amphidinium carterae and Alexandrium tamarense The X-axis is based on the eigen values from the principal component analyses based on codon bias and the Y-axis reflects synonymous third position guanine cytosine (GC) content.

Figure 4

The effective numbers of codons are shown relative to the guanine cytosine (GC) content of the transcriptome for each of the fifteen species. The X-axis is the GC content as a fraction of the total number of bases while the Y-axis is the total number of effective codons excluding stop codons. The expected number of codons shown in red is the maximum number of codons predicted for a randomly generated dataset of one thousand points while the observed values in blue are the average effective number of codons predicted by codonW given the tabulated relative synonymous codon usage for each species.

Figure 5

The relative synonymous codon usage for each codon is shown as a boxplot aggregating the data from all fifteen species. Boxplots are colored according to the “guanine cytosine (GC) weight” of each codon, i.e., the number of G or C bases occurring in the codon ranging from zero to four. Codons are arranged according to their respective amino acid given underneath in brackets.

Figure 6

Dinucleotide frequencies are given as the ratio of observed over expected frequencies. Expected frequencies were calculated from the mononucleotide frequencies. The guanine cytosine (GC) neutral species Amphidinium carterae is shown in red while the relatively GC rich species Alexandrium tamarense is shown in cyan. All sixteen possible dinucleotide combinations are shown on the X-axis while the relative ratios are on the Y-axis.

Figure 7

The number of times each codon occurred at a significantly higher than average frequency is shown for each codon. Columns are stacked and color coded for each species used in the analysis according to the legend. The X-axis shows each of the possible 64 codons, including the stop codons TGA, TAA and TAG. The Y-axis is the sum of observations across all species.

Figure 8

A phylogeny of dinoflagellates drawn using the branching order described in [33] with the guanine cytosine (GC) content for each of the fifteen species used in this study mapped onto the tree according to its own position or that of the most closely related species. GC content is shown on the X-axis as the relative proportion of GC bases pairs out of the total number of nucleotide base pairs in the transcriptome for each species plotted.