'Root of all success': Plasticity in root architecture of invasive wild radish for adaptive benefit

Abstract

Successful plant establishment in a particular environment depends on the root architecture of the seedlings and the extent of edaphic resource utilization. However, diverse habitats often pose a predicament on the suitability of the fundamental root structure of a species that evolved over a long period. We hypothesized that the plasticity in the genetically controlled root architecture in variable habitats provides an adaptive advantage to worldwide-distributed wild radish (Raphanus raphanistrum, Rr) over its close relative (R. pugioniformis, Rp) that remained endemic to the East Mediterranean region. To test the hypothesis, we performed a reciprocal comparative analysis between the two species, growing in a common garden experiment on their native soils (Hamra/Sandy for Rr, Terra Rossa for Rp) and complementary controlled experiments mimicking the major soil compositions. Additionally, we analyzed the root growth kinetics via semi-automated digital profiling and compared the architecture between Rr and Rp. In both experiments, the primary roots of Rr were significantly longer, developed fewer lateral roots, and showed slower growth kinetics than Rp. Multivariate analyses of seven significant root architecture variables revealed that Rr could successfully adapt to different surrogate growth conditions by only modulating their main root length and number of lateral roots. In contrast, Rp needs to modify several other root parameters, which are very resource-intensive, to grow on non-native soil. Altogether the findings suggest an evo-devo adaptive advantage for Rr as it can potentially establish in various habitats with the minimal tweak of key root parameters, hence allocating resources for other developmental requirements.

Keywords: East Mediterranean; Raphanus; adaption; habitat preference; root plasticity; root system architecture (RSA); soil surrogates.

Figure 1

Distinct habit and non-overlapping distribution of Raphanus raphanistrum (a worldwide weed) and its relative, R. pugioniformis on corresponding soil types in Israel. (A) Recorded populations of R. raphanistrum (green dots) and R. pugioniformis (blue dots) were plotted on a soil map of Israel (Batjes et al., 2020). R. raphanistrum occurs mainly on sandy soils of coastal plains (=Calcisol), while R. pugioniformis is restricted to mountainous areas in north-eastern Israel on Terra Rossa (=Luvisol) and basaltic (=Vertisol) soils. (B) R. raphanistrum scattered worldwide (pale blue areas on the world map; (Holm et al., 1997)) from its native distribution area of the Mediterranean basin (highlighted by dark blue; (Ziffer-Berger et al., 2020)). R. pugioniformis is restricted to northern Israel, southern Lebanon, and Syria. (C) Representative native population of R. pugioniformis at Mt. Gilboa, Israel growing on Terra Rossa soil with high water retention. (D) Representative natural populations of R. raphanistrum growing on sandy soil (Hamra), which differ in several soil characteristics from Terra Rossa (pH value, nutrient availability, and water retention capacity), near the coastal plain at Tel Aviv, Israel. Worldwide distribution of R. raphanistrum is compiled from Holm et al. (1997) and plotted on Miller’s world projection (www.whighcharts.com, based on data from www.naturalearth.com). The soil map of Israel is prepared with the data from Soilgrids (www.soilgrids.org) with 250m resolution.

Figure 2

Characteristic difference between main and lateral root lengths of R. pugioniformis and R. raphanistrum growing on native and non-native soils. R. pugioniformis requires significant modulation of main- to lateral-root length proportion on different soils (A–C), while the general root architecture of R. raphanistrum remained unaffected (D). The two species were potted in three different soils, control (A), Hamra, native soil of Rr (B), and Terra Rossa, native soil of Rp (C). The plants were excavated at the start of bolting (78 days after germination), roots were rinsed from substrate, disentangled, photographed, and analysed via EZ-Rhizo II (see methods section for details). Next to the representative photographs of rinsed, disentangled root system, digitized root characters are presented as alpha blends, and the spread of Lateral Root Length (LRL) computed by Root-VIS II. The proportion of main- to lateral- root length (D) in all three soils were calculated from the digitized root parameters.

Figure 3

Average daily temperature and precipitation of Tulkarm over the growth period of plants used for experiment. Left y-axis shows the precipitation in millimetre, indicated by the blue bars, on the right side the temperature is given in degree Celsius, plotted in the graph as red line. The growth started at 07.12.2019 and ended with the harvest at 23.02.2020.

Figure 4

Experimental setup to mimic root growth in surrogate media for soil types and simultaneous, semi-automated root growth analysis. (A) Seedlings of R. raphanistrum and R. pugioniformis were germinated in petri dishes on moistened filter paper, until the radicles were clearly visible and cotyledons were fully developed (red arrow). (B) The seedlings were transferred to the root growth experimental set up. Three pockets of 1.5 cm depth were folded on the upper edge, with a hole at the bottom, to insert the radicle through. The distance between the seedlings was 6 cm. Two seedlings were placed on one side of the paper near the edges, and a third seedling was placed in the middle on another side (visible in the figure). The middle section of 10 cm was sprayed with soil surrogate every other day. (C) All the experimental setup were hung vertically in a climate chamber, with the lower edge touching a reservoir of distilled water for continuous capillary water uptake. Altogether 27 seedlings per species were analysed for root architecture (3 soil surrogates X 3 seedlings X 3 setups).

Figure 5

Root architecture differs between R. pugioniformis and R. raphanistrum. (A) Representative root architecture of R. pugioniformis grown on Terra Rossa soil, showing a short main root with elaborate lateral root system. (B) R. raphanistrum grown on Hamra formed a long main root and few lateral roots. The plants were harvested after bolting (approximately 70-80 days after germination) following growth in natural conditions with individual pots in Palestine (see Figure 2 ). Representative root architectures of R. pugioniformis (C) and R. raphanistrum (E) after 21 days of controlled growth on modified root growth media (Hoagland & Knop medium), mimicking native soil nutrient composition of Terra Rossa and Hamra for R. pugioniformis and R. raphanistrum, respectively. Digital profiling of the corresponding roots after analysis with EZ-Rhizo II [R. pugioniformis, (D); R. raphanistrum, (F)], which enables quantitative analysis of root architecture and growth kinetics. The main root is traced as white line, while all the lateral roots are drawn in red.

Figure 6

Different root growth kinetics in diverse surrogates of soil types indicates a faster root growth on control, than on the species natural soils. Native soil types of both species, R. pugioniformis Rp, Terra Rossa) and R. raphanistrum (Rr, Hamra), along with a control, were mimicked by specific nutrient solutions (details see M&M). Root growth was recorded after one, two, and three weeks for semi-automated root architecture analyses. The percentage difference of main root length between control and native soil of the respective species are represented as delta (δ) above the box plots. The corresponding flag plots represent summary of different root system architecture parameters computed via EZ-Rhizo II. A typical "flag plot" (upper right panel with bold outline) represents mean insertion angle of lateral roots, displayed by the upper angle from "flagpole" and "flag", length of different root zones, departed by the triangle into basal unbranched zone, length of branched zone and length of apical unbranched zone. Furthermore, lateral root density is indicated by the color intensity within the triangle, ranging from white (low) to dark turquoise (high root density) and the slope of the linear regression, displayed by the lower angle (details see M&M). Rp = R. pugioniformis, Rr = R. raphanistrum.

Figure 7

Dimensional reduction of seven root architecture parameters revealed greater modulation of lateral root length and number is required by R. pugioniformis than by R. raphanistrum to adapt to diverse soil types. PCA analysis of seven root parameters from R. pugioniformis and R. raphanistrum grown on three different soil surrogates over three weeks (126 samples in total). Soil surrogates represent nutrient solutions, mimicking the natural soils of R. raphanistrum and R. pugioniformis, i.e., Hamra and Terra Rossa, the third solution is a control. Principal component 1 separates the groups by length of different root types and the number of lateral roots, component 2 mainly by main root angle, together they include 85% of the total information content derived from the measured variables (main root length, main root vector, main root angle, total root length, lateral roots on main root, total number of lateral roots and lateral root length). The ellipses show the 95% confidence interval.