Association genetics, geography and ecophysiology link stomatal patterning in Populus trichocarpa with carbon gain and disease resistance trade‐offs

Abstract

Stomata are essential for diffusive entry of gases to support photosynthesis, but may also expose internal leaf tissues to pathogens. To uncover trade-offs in range-wide adaptation relating to stomata, we investigated the underlying genetics of stomatal traits and linked variability in these traits with geoclimate, ecophysiology, condensed foliar tannins and pathogen susceptibility in black cottonwood (Populus trichocarpa). Upper (adaxial) and lower (abaxial) leaf stomatal traits were measured from 454 accessions collected throughout much of the species range. We calculated broad-sense heritability (H(2) ) of stomatal traits and, using SNP data from a 34K Populus SNP array, performed a genome-wide association studies (GWAS) to uncover genes underlying stomatal trait variation. H(2) values for stomatal traits were moderate (average H(2) = 0.33). GWAS identified genes associated primarily with adaxial stomata, including polarity genes (PHABULOSA), stomatal development genes (BRASSINOSTEROID-INSENSITIVE 2) and disease/wound-response genes (GLUTAMATE-CYSTEINE LIGASE). Stomatal traits correlated with latitude, gas exchange, condensed tannins and leaf rust (Melampsora) infection. Latitudinal trends of greater adaxial stomata numbers and guard cell pore size corresponded with higher stomatal conductance (gs ) and photosynthesis (Amax ), faster shoot elongation, lower foliar tannins and greater Melampsora susceptibility. This suggests an evolutionary trade-off related to differing selection pressures across the species range. In northern environments, more adaxial stomata and larger pore sizes reflect selection for rapid carbon gain and growth. By contrast, southern genotypes have fewer adaxial stomata, smaller pore sizes and higher levels of condensed tannins, possibly linked to greater pressure from natural leaf pathogens, which are less significant in northern ecosystems.

Exploiting Natural Variation to Uncover an Alkene Biosynthetic Enzyme in Poplar

“…These differences in quantity and composition could affect other tree traits. To test this, several previously published traits were compared as a function of the alkene phenotype (McKown et al, 2014a(McKown et al, , 2014bLa Mantia et al, 2013). Alkene absence did not appear to affect photosynthetic rate, but reduced transpiration: AM trees showed lower stomatal conductance and higher instantaneous water-use efficiency (WUE; Figure 7B; Supplemental Data Set 4).…”

Evolutionary quantitative genomics ofPopulus trichocarpa

“…Phenotyping of genotype accessions within the common gardens and climate of origin data were obtained from previously published work (for full phenotyping details, see [38,37], [45]). In brief, phenology, ecophysiology, biomass [45], leaf stomatal anatomy [44] and leaf rust (Melampsora xcolumbiana) resistance traits [38] were repeatedly measured from accessions planted at the University of British Columbia's research field through replication in space (clonal ramets) and in time (measurements across years). Wood chemistry and ultrastructure traits were measured from wood cores of the nine-year-old ortets representing the same genotypes and growing in Surrey [37].…”

“…Among four GWAS studies in P. trichocarpa, a total of 619 SNPs had been identified with significant trait associations (at α = 0.05): 410 with biomass, ecophysiology and phenology [39], 141 with wood property traits [43], 40 with Melampsora xcolumbiana resistance [38], and 28 SNPs related to leaf stomata variation [44].…”

“…S2 Table. Comprehensive summary table of all SNP detection results from GWAS [ecology [39]; rust [38]; stomata [44]; wood [43]] and outlier analysis (geographic F ST [36], this study: climate F ST , unsupervised SPA, climate SPA) for the black cottonwood population (presented in Fig 1) and using the 34k SNP chip [35]; adaptive traits (significant Q ST ) are in bold. In red and dark blue are 1% cutoffs (spa = 2.78025 and spa = 1.50795), in orange and light blue are 5% cutoffs (spa = 2.12467 and spa = 1.08868) in unsupervised SPA and climate SPA analyses, respectively.…”

“…We also performed F ST outlier Note: P-value obtained by comparison of the observed Q ST -F ST to the quantile of the simulated Q ST -F ST distribution for a neutral trait [96]. a biomass trait [45] b ecophysiology trait [45] c leaf rust resistance trait [38] d phenology trait [45] e wood trait [37] f leaf stomata traits [44] *spatially adjusted trait [45] Evolutionary Quantitative Genomics analysis on climate clusters. While the mean F ST value for the complete dataset (29,354 SNPs) was 0.0108, we obtained a mean neutral F ST value (0.0078) after removing loci identified to be potentially under selection [41].…”