5_Isochores

percent of G and C nucleotides in the genome (G=C, A=T, but GC may not = AT)

%GC varies among species (particularly bacteria, plants, invertebrates; little variation among vertebrates)

• However, vertebrates show much more %GC heterogeneity within their genomes

The vertebrate genome can be divided into isochores

= long stretches (100s of kb) of DNA with uniform %GC

With completion of the human genome, it was found that isochores could span >10 Mb

High %GC = heavy isochores (H)

Low %GC = light isochores (L)

Often L1 and L2 are grouped together as a single L isochore

Heavy isochores are found only in warm-blooded vertebrates (mammals, birds), not in cold- blooded vertebrates (fish)

1. Selectionist hypothesis

   • GC-pairing is stronger than AT-pairing (3 vs. 2 hydrogen bonds) and may stabilize DNA at higher temperatures.

   • supported by the observation that heavy isochores are found in warm-blooded vertebrates

   • heavy isochores are gene-rich

2. Mutationist hypothesis

   • pool of available nucleotides changes over replication (which takes 8 or more hours for mammals)

   • more GC available early in replication, so mutations will be biased towards G or C

   • Over time, regions of the genome that replicate early become GC-rich

   • In general, GC-rich regions have been observed to replicate early

• defining isochores as segments >300 kb with distinct %GC and low heterogeneity

• the authors found: that isochores covered 41% of the human genome, and most had low %GC

• They suggest a four-family model with mean GC contents of 35%, 38%, 41%, and 48% and conclude:

“These findings undermine the utility of the isochore theory and seem to indicate that the theory may have reached the limits of its usefulness as a description of genomic compositional structures”

• On a much smaller scale, %GC varies among different regions of a gene

• Coding regions > introns > 5’ flanking regions > 3’ flanking regions

• However, there is a strong correlation in %GC among all regions of a gene.

• analysis of many protein-coding sequences from a species indicates that all of the synonymous codons for a particular amino acid are not used with equal frequency as would be expected at random

• This phenomenon is known as “codon bias”

• Certain codons are “preferred” and are used much more frequently than “unpreferred” codons

• For example, Leucine is an amino acid that shows very high codon bias

  • Leu can be encoded by 6 different codons, CTG, CTA, CTC, CTT, TTG, TTA

  • At random, we would expect each codon to be used about 1/6 (17%) of the time

  • However, in highly expressed E. coli genes, CTG is used ≈90% of the time

  • In yeast, TTG is used ≈90% of the time

• The preferred codons correspond to the most abundant tRNA in each species, suggesting that selection favors the use of codons that increase the level of gene expression

• Note that the preferred codons may differ from species to species

• ENC

• Fop

ENC = effective number of codons

• the average number of codons that are used to encode the 20 amino acids

• The minimum is 20 (one codon per a.a.) the maximum is 61 (all codons except the 3 stop codons)

• Low ENC = high codon bias.

ENC can be applied to any species without prior knowledge of expression or codon usage.

Fop = frequency of optimal codons

• the frequency with which the “optimal” codon is used for each amino acid

• Optimal codons are defined as those used with the highest frequency in highly expressed genes

• High Fop = high codon bias.

Fop is species-specific and requires that optimal codons are known. For this, one must have many gene sequences and expression information.

a) higher in highly-expressed genes

b) higher in short genes than in long genes

c) higher in female-expressed than in male-expressed genes

a) Selection

b) Mutation (neutral)

• Natural selection favors the use of optimal codons (those that correspond to the most abundant tRNA) to make translation faster and more accurate

• Codon bias could be used as a way to regulate gene expression post-transcriptionally.

Evidence: Highly expressed genes have higher codon bias

• Conserved protein motifs, such as DNA binding domains, have higher bias than other protein regions.

• This suggests selection for accuracy of translation.

Experimental replacement of optimal codons with non-optimal codons reduces the level of protein. Example: Drosophila Alcohol dehydrogenase (ADH)

Replacement of optimal leucine codons with non-optimal codons leads to a lower level of ADH protein in vivo and reduces ethanol tolerance in adult flies

Wa-F = wild-type; optimal leucine codons

1 leu = 1 luecine codon changed from optimal to non-optimal

6 leu = 6 leucine codons changed from optimal to non-optimal

10 leu = 10 luecine codons changed from optimal to non-optimal

In a comparison of ADH protein concentration, it was found that:

Wa-F > 1 leu > 6 leu > 10 leu

• The wild-type flies were also more tolerant to ethanol that the mutant flies

• The LD50 (the ethanol concentration at which 50% of the flies were killed within 24 hours) was 9% for wild-type flies

• It was only 7.5% for 10 leu flies.

• Replacing sub-optimal leucine codons in the Adh gene with optimal codons increases ADH enzymatic activity in larvae, but decreases it in adults

• This suggests that there may be a trade-off between optimal codon usage (or other factors) in different developmental stages

• Codon bias is highest for larval-expressed genes.

• there may be a bias in mutation

• For example, if mutations from A or T to G or C are more frequent and there is no selection on synonymous sites, then these sites should become GC-rich

• The reverse bias in mutation would lead to synonymous sites being AT-rich

• In both cases, this would lead to non-random codon usage.

• Most optimal codons end in G or C

• Thus, the mutational hypothesis would require a mutational bias towards G or C

• However, most observations indicate that the mutational bias is towards A or T, which runs counter to the prediction#

• However, there does appear to be biased mismatch repair in favor of G or C

• This tends to be strongest in regions of high recombination (where most of the highly codon-biased genes are located)

• So this process could explain at least part of the observed codon bias.

• The GC content at third codon positions is correlated with local genomic GC content, suggesting an effect of mutation of codon usage.

• The strength of selection acting on a particular codon is expected to be very weak (on the order of Ns = 1, where N is the population size and s is the selection coefficient)

• This means that it is often very difficult to distinguish between selective and neutral explanations for codon bias.

• Typically, there is evidence for selection affecting codon usage in organisms with small genomes and large population sizes (bacteria, yeast, Drosophila)

• The evidence is much weaker in humans and other vertebrates, where it appears that mutational/repair biases can explain the observed codon usage.