Development of Active Root System Architecture of Upland Cotton for Improved Sub-surface Water Uptake During Drought Conditions

In the first year, the objectives were to determine the cotton varieties for this project;660 cotton accessions were ordered from USDA-GRIN. Those were planted in the greenhouse to grow for seed accession. Those were screened using phenotypes and their origins to reduce to 200 accessions to move to the next objective. The elite commercial cultivars grown in West Texas were collected from various seed industry partners, including BASF, Bayer, Armor, Phytogen, and NexGen. These companies selected varieties they were interested in testing for drought tolerance in this project. Some varieties already have some drought tolerance characteristics. The main goal of this objective is to get a diverse set of genetic variation to compare root systems for the best and the worst RSA, as well as to compare water efficiency for sustainable water use. Traits related to yield, disease resistance, and fiber quality will also be considered.

For the second objective, all the selected lines were screened for root morphology at the seedling stage, 2 weeks after planting. They were screened using a visual scoring and high-throughput phenotyping platform, WinRhizo, Inc. WinRHIZO is an image analysis system specifically designed for measuring root morphological traits. We focused on total root length, diameter, and surface area among other characteristics. According to Kadam et al. (2015), this process is done in Dr. Jagadish’s lab, which has extensive experience working with root morphology and anatomy of various crops. These experiments will help us narrow the lines for in-depth analysis under well-watered and water-deficient conditions.

The next phase of objective two is to identify lines with differential response to water deficit conditions regarding an altered RSA. To achieve this goal, we plan on growing a subset of contrasting lines in controlled-environmental greenhouse conditions under well-watered 100% field capacity and water-deficit conditions around 40%-60% field capacity (Kadam et al., 2015). Water deficit stress will be imposed after seedling emergence. Plants under the control treatment will be maintained at 100% field capacity throughout the experiments. After 30 days of stress, plants will be harvested around 50 days after sowing, and traits associated with the RSA, such as shoot biomass, will be measured. Physiology will be collected as detailed in Kadam et al., 2015. The anatomical studies will consider the root diameter, stele diameter, and metaxylem number.

After evaluating RSA and anatomical traits from objective 2, cotton cultivars showing differential or contrasting phenotypes were further evaluated under field conditions. Uprooting cotton plants in the field would have been challenging, so large woven bags filled with field soil were used to simulate the field conditions to overcome this issue. Watering was done manually by adding the recommended amount determined by the University of Georgia.

During the cotton growth, the morphological (shoot) traits were measured. These accessions were grown and intended to be harvested at the boll formation stage (i.e., the peak root biomass accumulation phase). However, a hailstorm destroyed some of the bags, and some were damaged. The time it took to wash these manually was longer than expected. Therefore, this experiment had to be repeated over the winter in the greenhouse with the same field soil and the same bags.

The repeated experiment was done in the greenhouse because of the winter conditions. With the limited space in the greenhouse and the amount of soil needed to transfer to the greenhouse from the field, the number of genotypes was reduced to 12 instead of the original 24. This was intended to have three replications of 12 rather than fewer replications. Watering was done similarly to manually measuring the field with the graduated cylinder. The volume and weight were calculated for the soil density, and the soil was set to fit within the same mass per volume to get the ideal density and compaction.

A comprehensive analysis of diverse germplasm and elite commercial cultivars grown in West Texas will allow us to identify the functional role of root tissue plasticity in adapting to water-deficit stress. It will be interesting to evaluate whether current commercial cultivars, in addition to deep root systems, possess active shallow roots and whether this helps conserve soil moisture during the vegetative and reproductive stages.

This information will help investigate, understand, and optimize root plasticity in upland cotton and translate this knowledge to develop next-generation cotton cultivars with better RSA under water-deficit conditions to improve sustainable agriculture. Commercial producers can apply this research to their ongoing projects to improve the future elite varieties, enabling farmers to reduce their water consumption and save money while increasing their yields. 

At the end of December 2023, the top 0-12 cm depth of topsoil was moved from the Quaker Research Farm to the Horticulture Greenhouse so field soil could be used for the repeated aboveground experiment to get the root architecture traits for the selected accessions. Accessions were previously selected based on their differences in root phenotypic traits. At the end of the 2023 summer season, many of our replications were destroyed in a hailstorm, and the timing for extraction prevented getting good, replicated data to move forward. The bags were filled with field soil by measuring mass and volume to have as similar a density and compaction as manually allowed (Fig. 1 A&B). 12 Genotypes were used three elite cultivars, TM1, and Coker 312 were used for the reference genotypes, also 3 of the lower and 4 of the higher performing root traits. They were randomly placed for three replications (Fig. 1C).

A) B)

Figure 1: Bag Set up in the Horticultural Greenhouse with three bags across and 12 bags down.  A 5-gallon bucket was used to measure the mass of soil that went into each bag. And the height of the cylinder was measured by marking the inside of the bag. Similar soil density and compaction must be improved to get the same amount of soil mass within a given volume for each bag. B) Representing the three bags across and how the bags look when filled. C) Random design of 3 replications 

 Each replication was planted two days apart to allow time for extraction for 2 days for each replication. Trying to allow them to be at the same growth stage when extracted. Plants were grown for six weeks and were extracted before they started the flowering stage. Many varieties differ in their flowering stages, and it was intended to see their root traits before allocating resources to the reproductive system. For manually watering the plants, a graduated cylinder was used to give them their recommended inches once per week. Above-ground traits were documented at two growth points before extraction. Plant height and health, number of nodes, level of pubescence, anthocyanin present, and number of leaves are some documented traits (Fig. 2). 

When extracting the samples, some difficulties arose with the stickiness of the soil, combined with how delicate the roots were. This made collecting the complete root system intact impossible. Nearly all the roots broke off from the primary root. They were manually collected from each bag until no more roots could be found (Fig. 3). The roots were washed by placing them into a bin of water with a mesh over the top, allowing the fine soil to fall away from the root and keeping the roots moist until time for imaging. 

Each root piece was manually placed on a painted black metal sheet (Fig. 4). They were carefully placed so as not to touch each other to obtain the total root length, surface area, and average diameter. Unfortunately, the correct root angle, number of tips and forks, and tap root depth could not be obtained. 

These accessions were intended to be placed in this year’s field experiment to see how their performance was under drought conditions; however, attempts to get enough seeds from some of the landrace varieties (105, 107, 099, 141) in time for this year’s field season were unsuccessful. Only a couple of seeds were collected because of their late flowering due to the assumption of photoperiod sensitivity traits. Therefore, accessions 105, 107, 099, and 141 were planted earlier than the previous year, hopeful that more seeds would be collected for future experimentation. 

Shortly after removing the tall bags from the greenhouse, another experiment was conducted to collect the RSA phenotypes for the GWAS analysis. Two hundred twelve accessions, 12 elite cultivars, 200 various accessions from obsolete cultivars, landraces, breeding lines, and TM1 and Coker 312, were selected to represent cotton germplasm from around the world and West Texas (Fig. 5). These accessions were selected based on their ability to flower in this area and their similarities and differences in visual phenotypes. 

The bags used were 2 gallons filled with the Quaker Research Farm soil. Three replications of each accession were randomly distributed throughout the four tables in the same greenhouse room. Each day, 64 bags were completed. Five seeds were placed in each of the 636 bags. After germination, the weakest seedlings were thinned out, leaving only one plant per bag. This experiment attempted to keep the conditions optimal for well-watered conditions for all accessions because there was not enough room or time for another 636 bags for drought conditions. Each bag was saturated with ~ 600 ml and left to dry for 24 hours to reach field capacity before planting the seeds. Each set of 64 bags was one day offset from the previous 64 bags to allow them to grow the same amount of time and be extracted on the same days after planting (DAP). Within 11 days, all bags were planted (Fig. 6A). Each bag received 200 ml of water on 3, 8, and 15 DAP as the greenhouse temperature would reach 110 degrees during the day. Water was added using a graduated cylinder to ensure the same amount was added to each bag. Plants were extracted on 17 DAP (Fig. 6B). Washing the roots was done by first soaking the bags, allowing them to become overly saturated to reduce the stickiness of the soil to the roots. The bags were cut and ripped away as one held onto the plant's stem. Then, working the way down the tap root, try to loosen as many roots as possible from the soil and gather any roots that broke off. Samples were stored in Ziplock baggies with 50% ethanol and placed into a 4° C room until they could be scanned (Fig. 6C). 

Samples were scanned the same way as the previous year's sample with the WinRHIZO scanner and software by placing an acrylic tray with some water to ease the roots' separation. Plastic tweezers were used to pry apart the roots to prevent all bunching up in one section. However, some roots were so prolific that they made separation challenging. Each sample took around 7 minutes to scan, and while one was being scanned, another sample was being prepped and separated.  

The same 212 accessions grown in the greenhouse were also planted in the Quaker Research Farm in a 16-row plot with sub-surface drip irrigation that I could manually turn on or off for eight rows at a time (Fig. 7). For the first eight rows, I left the irrigation off for the entire season and only received water from the rain. In the second 8 rows, I supplemented the weeks with no rain with irrigation. Aimed for about 1” of water per week. The irrigated plot received 5 weeks of irrigation more than the rainfed plants.

Figure 8 shows the difference in canopy growth between the two plots in September. Irrigation was stopped at the beginning of September. A sensor from AquaSpy was placed in the rainfed and irrigated fields to take measurements until the end of the season. It measured every 15-minute interval with moisture levels, salinity, and temperature. Figure 8 focuses on the moisture levels collected by both sensors. The sensors were placed at the beginning of July; irrigation began four weeks after planting. 

By the peaks at 4 inches depth, it is possible to determine when it rained during the months of July through October. There was no rain during the hottest months of the year, from August through September, contributing to the drought stress for the cotton during this time. The difference in water between the treatments through August and September is a clear separation at most depths. Looking at the 8 to 12-inch depth where the drip irrigation emitters are located, there are two steady points in the month of July. But a slow decrease in August, which could suggest that the 1 inch of irrigation was not enough to bring it back up to saturation levels, but most likely helped to prevent the plants from being too stressed during August. A repeated experiment next summer, with time to do physiological measurements, would be helpful to know the stress of the leaves between the two treatments.

Leaf samples were collected from all the accessions and sent to a company, SGS Technologies, to do the genotype sequencing by running an 18K SNParray, we received SNPs for around 17K markers with around 10K that were homologous with unique markers and the other markers were either heterozygous and/or the marker matched multiple locations in the genome—using R version 4.4.2 and RStudio package called GAPIT (Wang and Zhang, 2021).