Two new fern chloroplasts and decelerated evolution linked to the long generation time in tree ferns

Abstract

We report the chloroplast genomes of a tree fern (Dicksonia squarrosa) and a "fern ally" (Tmesipteris elongata), and show that the phylogeny of early land plants is basically as expected, and the estimates of divergence time are largely unaffected after removing the fastest evolving sites. The tree fern shows the major reduction in the rate of evolution, and there has been a major slowdown in the rate of mutation in both families of tree ferns. We suggest that this is related to a generation time effect; if there is a long time period between generations, then this is probably incompatible with a high mutation rate because otherwise nearly every propagule would probably have several lethal mutations. This effect will be especially strong in organisms that have large numbers of cell divisions between generations. This shows the necessity of going beyond phylogeny and integrating its study with other properties of organisms.

Keywords: Dicksonia; Tmesipteris; chloroplast genomes; ferns and fern allies; generation time effect; mutation rates.

Fig. 1.—

Pearson correlation results. The blue line indicates the Pearson correlation coefficient (r) of the ML distance calculated from “A” (more conserved) and “B” (less conserved) partitions. The red line indicates the r value of uncorrected p distances and ML distances for B partitions. The r values begin to increase significantly at 31,136 sites remaining and this is taken to indicate that the assumed model of nucleotide evolution is beginning to fit the data well.

Fig. 2.—

ML tree of land plants based on the original data (34,386 sites). Bootstrap support values are indicated along the branches. The two newly sequenced genomes are indicated as *. In this tree, the lycophytes are adjacent to the seed plants with weak bootstrap support (BP = 64%).

Fig. 3.—

ML tree of land plants based on the OV-sorted matrix (31,136 sites). Bootstrap support values are indicated along the branches and node numbers are marked as blue. This ML tree is the same as figure 1 except that the lycophytes are now adjacent to (seed plant + monilophytes) with 100% bootstrap support.

Fig. 4.—

Estimates of the percentage of surviving offspring (y axis) at lowered mutation rates and increasing relative generation times (x axis). The mutation rate increases when going from the black to red, purple, green, and blue lines. There are three overlapping lines at the top left, and three overlapping lines at the bottom right. As the generation time increases, particularly with higher mutation rates, the chance of successful offspring (with no lethal mutations) strongly declines. A high mutation rate and a long generation time are incompatible.