Plant genome size measurements
Plant species were first identified by a group of plant taxonomists, then all the species names were further standardized into the accepted names according to The Plant List (version 1.1; www.theplantlist.org). We recorded 286 herbaceous species in the transect (our study sites contained mainly herbaceous vegetation); of these we were able to obtain measurements of genome size from 169 species during subsequent visits to representative study sites belonging to the Chinese Grassland Long-term Research Stations (at least one site for each grassland type: alpine grassland, desert steppe, typical steppe, and meadow steppe). These stations represented the typical vegetation and species pool for each grassland type and were convenient to re-visit. Samples were collected to measure genome sizes during the growing seasons of 2017-2019 (from July to September). We sampled plant species that occurred at each grassland type of the transect, focusing primarily on (but not limited to) the more common ones (details for the common species and species richness of each study site are supplied in Supplementary Dataset 1 and 2).
For each species, three to twelve individuals were selected and leaf samples from each individual plant were measured for the genome size and the averaged value was used. For each individual, leaves from at least three fresh samples were analysed by flow cytometer (BD LSRFortessa, USA) (Doležel et al. 2007). Samples were chopped with a sharp razor blade for ca. 15-25s in a petri dish containing 1mL ice-cold nuclear LB01 isolation buffer (15 mmol/L Tris, 20 mmol/L EDTA-Na2, 0.5 mmol Spermine tetra-hydrochloride, 80 mmol/L KCl, 20 mmol/L NaCl, 0.1%volume percentage Triton X-100, 2 mmol/L DTT, pH=7.5). The homogenate was gently sucked up by pipette and filtered through a 50 μm nylon mesh. A volume of 50 μL of RNase (1 mg mL-1; Sigma, St Louis, MO, USA) and 50 μL of Propidium iodide (PI, 1 mg mL-1; CyStain PI Absolute P, Sysmex Partec GmbH Görlitz, Germany) were then added and mixed by gentle shaking. The mixture was incubated for 4 min in the dark at 4 °C and loaded onto the flow cytometer to measure genome size. The pg of DNA per nucleus in each peak is estimated by comparing measurements to a known internal reference standard. We had five species as our primary reference standard and two species as our secondary reference standard. For the primary reference standard, seeds were obtained from Centre of Plant Structural and Functional Genomics of the Institute of Experimental Botany, Czech Academy of Sciences, in 2017, including Solanum lycopersicum cv. Stupicke (1C = 0.98 pg) (Doležel et al. 1992),Glycine max cv. Polanká (1C = 1.25 pg) (Doležel et al.1994), Zea mays CE-777 (1C = 2.72 pg) (Lysak & Dolezel 1998),Pisum sativum cv. Ctirad (1C = 4.55 pg) (Doležel et al.1998), Secale cereale cv. Dankovske (1C = 8.10 pg) (Doleželet al. 1998). Based on the primary reference standard, genome size of two local species: Kerria japonica (L.) DC (1C = 0.50 pg), Hosta plantaginea (Lam.) Aschers. (1C = 11.33 pg) were further calibrated against S. lycopersicum and V. faba and chosen as a secondary reference standard. The mean, standard deviation, and individual measurements of the studied species and the used internal standards are reported in Table S1. Note that we originally measured genome size for 169 species but we excluded eight species with high coefficients of variation in peak quantification (e.g. > 5 %) given their unprecise measurement (Doležel et al. 2007).
Among the 161 species sampled for genome size measurements, 23 species occurred more than once across the resampling sites, where we could check intraspecific variation of ploidy levels (Table S2). The resampling sites refer to the sites that have been investigated during the transect study but were revisited for measuring genome sizes. Most of these species showed quite stable 1C DNA content values across the transect, except Artemisia frigida and Agropyro cristatum . Previous studies had reported the 1C value for diploid A. frigida(2.63 pg) and A. cristatum (6.49 pg) (Garcia et al. 2008; Said et al. 2018), with their polyploid species being identified in nature (Wan et al. 2011; Zhao et al. 2017). In our case, the measured 1C values indicated that A. frigida andA. cristatum occurred more often as diploid (1C = 3.25 and 6.84 pg) in meadow steppe while as tetraploid (1C = 5.26 and 13.30 pg) in the other three grassland types. We report the diploid and tetraploid 1C values for those two species according to their occurring grassland type (Table S1) and used the mean value in our data analysis to handle the intraspecific ploidy variation.
For the 161 species, we assessed their contributions to biomass and species richness of the total plant community at each site. Sites were included in further analyses if, for that specific site, the species that had genome size values contributed to more than 80% of biomass and richness. Thus, for the final analysis, we used data for plant communities containing 161 species from 52 sites along the transect. Overall, the plant 1C values varied 260-fold from the smallestAstragalus scaberrimus (0.12 pg) to the largest Allium ramosum (31.5 pg), with a median and mean of 1.59 and 3.19 pg, respectively (Fig. S1; Table S1).