Phylogeny and divergence times of gymnosperms inferred from single-copy nuclear genes

Abstract

Phylogenetic reconstruction is fundamental to study evolutionary biology and historical biogeography. However, there was not a molecular phylogeny of gymnosperms represented by extensive sampling at the genus level, and most published phylogenies of this group were constructed based on cytoplasmic DNA markers and/or the multi-copy nuclear ribosomal DNA. In this study, we use LFY and NLY, two single-copy nuclear genes that originated from an ancient gene duplication in the ancestor of seed plants, to reconstruct the phylogeny and estimate divergence times of gymnosperms based on a complete sampling of extant genera. The results indicate that the combined LFY and NLY coding sequences can resolve interfamilial relationships of gymnosperms and intergeneric relationships of most families. Moreover, the addition of intron sequences can improve the resolution in Podocarpaceae but not in cycads, although divergence times of the cycad genera are similar to or longer than those of the Podocarpaceae genera. Our study strongly supports cycads as the basal-most lineage of gymnosperms rather than sister to Ginkgoaceae, and a sister relationship between Podocarpaceae and Araucariaceae and between Cephalotaxaceae-Taxaceae and Cupressaceae. In addition, intergeneric relationships of some families that were controversial, and the relationships between Taxaceae and Cephalotaxaceae and between conifers and Gnetales are discussed based on the nuclear gene evidence. The molecular dating analysis suggests that drastic extinctions occurred in the early evolution of gymnosperms, and extant coniferous genera in the Northern Hemisphere are older than those in the Southern Hemisphere on average. This study provides an evolutionary framework for future studies on gymnosperms.

Figure 1. The ML tree of gymnosperms constructed from combined LFY and NLY CDS sequences.

Numbers associated with branches are bootstrap percentages higher than 50%. The cycads were used as functional outgroups.

Figure 2. Comparison of ML trees of gymnosperms constructed using LFY + NLY sequences.

A and C, All three codon positions were used; B and D, 1^st and 2^nd codon positions were used. A and B, Angiopteris lygodiifolia was used as outgroup; C and D, The cycads were used as functional outgroups. Numbers associated with branches are bootstrap percentages of ML higher than 50% and Bayesian posterior probabilities greater than 0.90, respectively.

Figure 3. The ML trees of cycads inferred from sequence analysis of combined LFY and NLY sequences.

A, CDS; B, CDS+Intron. Numbers associated with branches are bootstrap percentages of ML and MP higher than 50% and Bayesian posterior probabilities greater than 0.90, respectively. Ginkgo biloba was used as outgroup in Fig. 3A.

Figure 4. ML trees of Podocarpaceae constructed from sequence analysis of combined LFY and NLY sequences.

A, CDS; B, CDS+Intron. Numbers associated with branches are bootstrap percentages of ML and MP higher than 50% and Bayesian posterior probabilities greater than 0.90, respectively. Araucaria heterophylla and Agathis robusa were used as outgroups in Fig. 4A.

Figure 5. Divergence times of gymnosperms estimated from combined LFY and NLY CDS sequences using BEAST.

A time scale is shown at the bottom. A–K indicate fossil calibration points. 1–10, A, D, E and G indicate some nodes of interest. Median ages of nodes are shown, with horizontal bars indicating the 95% highest posterior density intervals.

Figure 6. The ML trees of Taxaceae+Cephalotaxaceae constructed from CDS sequences.

A, combined LFY and NLY; B, LFY; C, NLY. Numbers associated with branches are bootstrap percentages of ML and MP higher than 50% and Bayesian posterior probabilities greater than 0.90, respectively. Taiwania cryptomerioides and Cunninghamia lanceolata were used as outgroups.

Figure 7. A lineage-through-time plot showing divergence time distribution of the gymnosperm genera.

The divergence times was based on the median ages of the nodes from the BEAST analysis (see Fig. 5).

Figure 8. Comparison of divergence times of the coniferous genera between the Southern and Northern Hemispheres.

A, boxplot comparison of all genera; B, dot plot comparison of each genus. The calculation of divergence times was based on the median ages of the nodes from the BEAST analysis as shown in Fig. 5.