An Alternative Explanation

Cinderalla Solution

Metabolism Boosting Diet

Get Instant Access

One alternative hypothesis for the patterns shown in Tables 1-4 is that of mutation bias. For example, a greater mutational pressure favouring A against G in the mitochondrial genomes than in the nuclear genomes would result in a greater proportion of A-ending codons in the mitochondrial genes than in the nuclear genes. Martin (1995) has argued that organisms of high metabolic rate should experience higher mutation rate favouring A than organisms of low metabolic rate.

The mutation hypothesis can be distinguished from THCU because the two hypotheses have different predictions. Let us first focus on the consequence of mutation favouring A against G. Suppose a protein gene with equal number of A, C, G, and T distributed randomly on both template and non-template strands (i.e., the original sequence in fig. 4). When five G's are replaced by five A's through mutation on the template strand, five C's will consequently be replaced by five T's on the non-template strand. Because mutation occurs randomly on both template and non-template strand of the gene, we also expect five G's to be replaced by five A's on the non-template strand and five C's to be consequently replaced by five T's on the template strand. The net result is that on either template or non-template strand, the increment in the number of A nucleotides (five in our fictitious case) is matched by the increment in the number of T nucleotides (also five in our fictitious case). In other words, ANa = &Nt on both template and non-template strands (fig. 4), so that A-ending codons and T-ending codons will be used equally frequently, and both used more frequently than G-ending and C-ending codons.

In contrast to the mutation hypothesis, THCU predicts that, with ATP more readily available than other nucleotides, the protein gene should evolve towards maximizing the use of A in mRNA (i.e., maximizing the number of A on the non-template strand of the coding sequence, fig. 4). This will result in an increase in the number of A, and a decrease in the number of T in the non-template strand of the gene (fig. 4). In short, although both THCU and the mutation hypothesis would predict that A-ending codons should be much more frequent than G-ending codons, the two hypotheses differ in that THCU predicts A-rich and T-poor on the non-template strand, whereas the mutation hypothesis (e.g., with mutation favouring A against G) predicts that both strands should be AT-rich and GC-poor, with A's and T's distributed equally on the two strand (fig. 4).

Figure 4. Contrasting predictions from the mutation hypothesis (with mutation favouring A against G) and THCU (transcription hypothesis of codon usage). THCU predicts A-richncss and T-poorness in the non-templale strand (bottom panel), and the opposite in the template strand A designates increment Because mutation favouring A against G is expected to occurred equally frequently on both DNA strands, the mutation hypothesis expects both DNA strands to accumulate equal number of A's and, consequently, equal number of T's, so that both strands will be AT-rich and GC-pour

Figure 4. Contrasting predictions from the mutation hypothesis (with mutation favouring A against G) and THCU (transcription hypothesis of codon usage). THCU predicts A-richncss and T-poorness in the non-templale strand (bottom panel), and the opposite in the template strand A designates increment Because mutation favouring A against G is expected to occurred equally frequently on both DNA strands, the mutation hypothesis expects both DNA strands to accumulate equal number of A's and, consequently, equal number of T's, so that both strands will be AT-rich and GC-pour

The mutation hypothesis seems to explain satisfactorily the pattern of codon usage in Drosophila mitochondrial DNA. The number of codons ending with A, C, G, and T in Drosophila yakuba is 1052, 107, 45, and 1092, respectively, for protein genes on the H-strand, and is 403, 6, 31, and 428, respectively, for protein genes on the L-strand. Thus, Drosophila mtDNA is AT-rich, with A-ending and T-ending codons used roughly equally, and both used much more frequently than C-ending and G-ending codons. These fulfil the prediction based on mutation hypothesis (top panel of fig. 4). In neither strand do we observe A-richness and T-poorness expected from THCU (fig. 4).

Additional evidence confirming that codon usage in Drosophila is mainly controlled by mutations favouring A or T comes from an AT-rich region flanking the origin of replication. This region spans 1.0-5.1 kb and is homologous in various Drosophila species (Fauron and Wolstenholme 1980a; Fauron and Wolstenholme 1980b; Goddard and Wolstenholme 1980). The region exhibits extensive sequence divergence, suggesting that the nucleotide sequence is mainly under the control of mutation bias (Goddard et al. 1982). The fact that the region is made of almost entirely of AT pairs implies that the mutation spectrum in Drosophila is strongly AT-biased, and that the preponderance of A-ending and T-ending codons in Drosophila mtDNA can be explained as a consequence of the mutation bias.

Another DNA region that appears to be strongly affected by mutation bias is the D-loop of mammalian mtDNA. Goddard et al. (1982) has suggested that the D-loop is homologous to the highly variable AT-rich region in Drosophila mtDNA mentioned above. Like the AT-rich region in Drosophila, the D-loop also flanks the replication origin, is also highly variable in nucleotide sequences (Avise et al. 1994), and is not transcribed except for perhaps a few bases. Thus, the nucleotide composition of the D-loop should reflect the mutation spectrum in the mammalian mtDNA. The number of A, C, G, and T in the mouse D-loop is 258, 104, 218, and 299, respectively. This is consistent with what we would expect if the D-loop is under mutation bias favouring A or T (top panel of fig. 4).

The mutation hypothesis, however, fails in explaining the pattern of codon usage in mammalian mtDNA. The data in Table 1 shows that A-ending codons are always much more frequently used than T (or U)-ending codons in bovine mtDNA, in contrast to what we see in Drosophila mtDNA where A-ending and T-ending codons are used equally frequently, and also in contrast to the D-loop region where T is more frequent than A.

The data from mouse mtDNA further highlight the inadequacy of the mutation hypothesis. The number of codons ending with A, C, G, and T in the mouse mtDNA is 1677, 1000, 117, and 825, respectively, with A-ending codons far outnumbering not only G-ending codons, but also T-ending codons. This pattern is the same as what we see in Table 1 for the cow and is expected from THCU, but not from the mutation hypothesis. (Note that there are more NNY codons than NNR codons in mammalian mtDNA, with the difference > 300. So the observed excess of A-ending codons and deficiency of T-ending codons in mammalian mtDNA is not a consequence of protein genes made of mostly NNR codons).

Although the pattern of codon usage in mtDNA is more satisfactorily explained by THCU than by the mutation hypothesis, one can still argue that the difference in codon usage between mtDNA and nuclear DNA (Table 2) is attributable to mutations in mtDNA more biased in favour of A than mutations in nuclear genome, which could result in more A-ending codons in mtDNA than in nuclear genome. A new finding summarized below appears to favour THCU.

Zischler et al. (1995) discovered a segment (540 bp) of the human mitochondrial D-loop to have been inserted into the nuclear genome, and that the inserted sequence has presumably existed as non-functional DNA. The nucleotide frequencies of the insert for A, C, G, and T are 30.7%, 32.6%, 13.9%, and 22.8%, respectively. The equivalent values for the homologous 540 bp D-loop segment (from LOCUS HUMMTCG in GenBank) are 30.4%, 32.8%, 14.1%, and 22.8%, respectively. If mutations are more biased in favour of A in mtDNA than in nuclear genome, then we should expect a reduction of the proportion of A in the insert, which is not true.

Zischler et al. (1995) also sequenced the nuclear DNA sequences flanking the insert. The two flanking regions add up to a total of 385 bp, with the nucleotide frequencies being 41.3%, 18.2%, 13.5%, and 27.0%, respectively, for A, C, G, and T. Thus, the A-content of the non-functional DNA of nuclear origin appears to be in excess rather than in deficiency in comparison with the equivalent values in mitochondrial D-loop. This suggests that mutations in mtDNA is not more biased in favour of A than those in the nuclear genome. In short, the larger proportion of A-ending codons in mtDNA relative to nuclear DNA is not due to mutation bias favouring A in mtDNA.

It is much more difficult to distinguish THCU from the mutation hypothesis regarding the differences in codon usage among mammalian species of different metabolic rates (Tables 3-4). For example, although the proportion of A-ending codons is greater for mouse genes than for rat genes, the proportion of T-ending codons (Pt) also seems to be greater for the mouse genes than for the rat genes. This concurrent increase in both PA and PT in animals of higher metabolic rate (i.e., the mouse) is compatible with the mutation hypothesis (Martin 1995) invoking mutation bias favouring A or T in animals of higher metabolic rate (SRM). However, for the nine codon families with both A-ending and T-ending codons, PAfor the mouse genes is significantly largerthan PA for the rat genes (P = 0.017, paired T-Test, one-tailed), whereas the difference in PT between the mouse genes and the rat genes is not significant (P = 0.507). This suggests that THCU is a plausible alternative to the mutation hypothesis.

The data of introns (Table 4) are almost entirely compatible with the mutation hypothesis in that a concurrent increase in both A-content and T-content is observed in genes from mammalian species with a high metabolic rate relative to those from mammalian species with a low metabolic rate. The only exception involves comparisons between human and mouse for the p-globin gene. The mouse introns for the P-globin gene show higher A-content and lower T-content than human introns. This is expected under THCU, but not under the mutation hypothesis. However, such a single case should not be taken as a rejection of the mutation hypothesis, which to me remains a plausible hypothesis in many other cases.

I conclude that THCU is a sufficient, and perhaps unique, explanation for the biased codon usage favouring A-ending codons in mammalian mtDNA (Table 1), and the differences in codon usage between the mitochondrial genomes and the nuclear genomes (Table 2). My results further suggest that THCU is a plausible hypothesis in explaining the differences in codon usage in nuclear genomes among mammalian species of different metabolic rates (Table 3-4).

Was this article helpful?

0 0
Turbo Metabolism

Turbo Metabolism

Forget Silly Diets-They Don't Work. Weight loss has got to be the most frustrating experience for many people, young and old alike. Eating foods that are just horrible, denying yourself foods you truly love and enjoy. Exercising, even though you absolutely hate exercising, and end up stiff as a board with no results.

Get My Free Ebook


Post a comment