Genome size evolution in relation to leaf strategy and metabolic rates revisited

Abstract

Background and aims: It has been proposed that having too much DNA may carry physiological consequences for plants. The strong correlation between DNA content, cell size and cell division rate could lead to predictable morphological variation in plants, including a negative relationship with leaf mass per unit area (LMA). In addition, the possible increased demand for resources in species with high DNA content may have downstream effects on maximal metabolic efficiency, including decreased metabolic rates.

Methods: Tests were made for genome size-dependent variation in LMA and metabolic rates (mass-based photosynthetic rate and dark respiration rate) using our own measurements and data from a plant functional trait database (Glopnet). These associations were tested using two metrics of genome size: bulk DNA amount (2C DNA) and monoploid genome size (1Cx DNA). The data were analysed using an evolutionary framework that included a regression analysis and independent contrasts using a phylogenetic tree with estimates of molecular diversification times. A contribution index for the LMA data set was also calculated to determine which divergences have the greatest influence on the relationship between genome size and LMA.

Key results and conclusions: A significant negative association was found between bulk DNA amount and LMA in angiosperms. This was primarily a result of influential divergences that may represent early shifts in growth form. However, divergences in bulk DNA amount were positively associated with divergences in LMA, suggesting that the relationship may be indirect and mediated through other traits directly related to genome size. There was a significant negative association between genome size and metabolic rates that was driven by a basal divergence between angiosperms and gymnosperms; no significant independent contrast results were found. Therefore, it is concluded that genome size-dependent constraints acting on metabolic efficiency may not exist within seed plants.

Fig. 1.

The LMA ‘mega-tree’ phylogeny from Phylomatic to the family level (Webb and Donoghue, 2005). Currently, Phylocom outputs relationships within families as polytomies; therefore, published systematic data were used to resolve many of them. Phylogenetic relationships for the A_mass and R_mass data sets were pruned from the LMA ‘mega-tree’ using Phylocom (Webb et al., 2006; see Materials and Methods). Tree graphic created using TREEVIEW (Page, 1996).

Fig. 2.

The relationship between genome size and (A, B) leaf mass per unit area (LMA) and (C, D) photosynthetic rate (A_mass) without correcting for the influence of phylogeny. Data are split into gymnosperms (closed circles) and angiosperms (open circles). For 2C DNA content and LMA (A), angiosperms alone have a significant negative relationship, and gymnosperms alone have significant positive relationship. For 1Cx DNA content and LMA (B), the significant positive relationship within gymnosperms is retained; however, the relationship is no longer significant within angiosperms. For 2C DNA content and A_mass (C), for gymnosperms alone the relationship is significantly negative, whereas for angiosperms alone the slope is nearly zero. Near identical results were found when testing for a relationship between 1Cx DNA content and A_mass (D).

Fig. 3.

Contrast plots depicting the relationship between divergences in genome size and (A, B) divergences in leaf mass per unit area (LMA) and (C, D) divergences in photosynthetic rate (A_mass). The regression lines are forced through the origin. Divergences in 2C DNA are significant and positively correlated with divergences in LMA (A); this relationship is driven by divergences within angiosperms (divergences in gymnosperms were not significant). Divergences in 1Cx DNA content (B) are not correlated with divergences in LMA for all analyses. Divergences in 2C DNA content (C) and 1Cx DNA content (D) are not correlated with divergences in both A_mass and R_mass (data not shown; see Table 2).