Intercropping has been widely recognized as a valuable means to increase productivity and sustainability in agriculture. A substantial part of plant-to-plant interactions occurs in the underground where roots adopt different strategies, from niche differentiation and physical avoidance to facilitative interaction with large intermingling between neighbouring roots. These root decisions take place in the unseen and therefore they tend to be largely overlooked (Homulle et al., 2022; Xue et al., 2016).

Our work consisted of testing the phenotyping pipeline for measuring root traits typically available from rhizobox studies

Our work within CREA _Council for Agricultural Research and Economics (Italy)_ consisted of testing the phenotyping pipeline for measuring root traits typically available from rhizobox studies, such as visible root length and total root length, for lupin growing in pure stands or in intercrop with wheat. Additionally, we tested a rapid, low-cost root phenotyping protocol to visualize changes in root architecture in response to intercrop. This phenotyping pipeline will be later used to identify root architectural traits involved in plant-to-plant interactions and germplasm resources expressing root traits especially suitable for intercropping.

For the direct observation of root architectural traits during intercropping we chose to grow plants in rhizoboxes.

Rhizoboxes are special containers for plant growth provided with a transparent window for direct observation of roots growing in soil. We used sandy soil given by a 40:60 mix (w/w) between a clay loam and river sand. We used six custom-built rhizoboxes (20 × 2 × 40 cm) where four white lupin and wheat plants were grown (either alternated on a row or with two wheat plants at the edges and two lupin plants in between, always with 7 cm inter-plant spacing). The plants were grown for three weeks, during which the roots reached the bottom of the box. We tested the whole root phenotyping pipeline for measuring both visible root length and total root length. Additionally, we set up a new phenotyping step for computing a novel root interaction index based on the quantification of the area of overlap between the convex hull (the smallest area that encloses the whole root system) of neighbouring plants. This was done by using a pinboard to ‘block’ the root 2D architecture onto the rhizobox, lined with a soft black foam sheet to lift the whole root system and acquire high-resolution root images or manually tracing root features such as the convex hull, which is informative of root space filling strategy. Individual plants’ root convex hull was manually traced by placing a transparent polyethylene sheet directly on the pinboard or the foam sheets. Traced areas were scanned and imported in the freeware software ImageJ. The convex hull was digitalized manually (polygon selection), and the area was calculated automatically, also for the area of overlapping between neighbouring roots. After this step, neighbouring roots were gently disentangled and stored in ethanol solution (50:50 v/v) at 4°C for further analysis. A new index of interaction was calculated given by the ratio between the area of overlap with the neighbouring plants and the individual plant convex hull. This index, called root merge (RM), varies between 0 (total avoidance) and 1 (total overlap).

Due to its simplicity and low cost, this methodology is also very accessible to researchers even when not equipped with root phenotyping tools and technologies.

The whole phenotyping pipeline showed no methodological flaws although the great plant mortality/anomalous growth, especially for lupin, makes the use of a higher number of replicates and spare boxes mandatory. The use of the pinboard allowed to wash out of the soil without altering the root 2D architecture. The foam sheet allowed easy lifting and recovery of the whole root system that could be scanned for later image analysis or analysed in real-time for the convex hull manual digitalization, which took approximately 3-4 minutes per box. A simple, rapid and throughput index of root interaction between neighbouring plants (RM) was thus calculated. Due to its simplicity and low cost, this methodology is also very accessible to researchers even when not equipped with root phenotyping tools and technologies. The easy root 2D architecture recovery lends itself to the use of other fast phenotyping techniques such as shovelomic (Trachsel et al., 2011) and does not hinder the quantification of standard traits from rhizobox studies, such as visible root length and total root length and mass, as the total root system is recovered.

References

Trachsel, S., Kaeppler, S. M., Brown, K. M., & Lynch, J. P. (2011). Shovelomics: high throughput phenotyping of maize (Zea mays L.) root architecture in the field. Plant and soil, 341, 75-87.

Xue, Y., Xia, H., Christie, P., Zhang, Z., Li, L., & Tang, C. (2016). Crop acquisition of phosphorus, iron and zinc from soil in cereal/legume intercropping systems: a critical review. Annals of Botany, 117(3), 363-377.

Homulle, Z., George, T. S., & Karley, A. J. (2021). Root traits with team benefits: understanding belowground interactions in intercropping systems. Plant and Soil, 1-26.

This article was written by Roberta Rossi and the team at CREA (Italy)