Team:Leiden/PerchloratePlantGrowth

Plant Growth with Perchlorate

Plant growth with perchlorate

Introduction & Methods

We want to grow crops on Mars for our future astronauts to live off. Make Mars great again. Since the Martian soil contains about 1% perchlorate, we wanted to asses the ability of crops to grow in the presence of perchlorate. To see the effect of perchlorate on germination and growth, we made a timelapse video that shows the growth of common garden cress over a period of 5 days. Five different concentrations of perchlorate were added to water and 5 mL was added to each petridish (0 mg/L, 2.6 mg/L, 26 mg/L, 260 mg/L and 2600 mg/L perchlorate). A 5x5 latin square randomized study design was used to mitigate the effect of possible differences in lighting or temperature on either side of the the table. Each petridish contained 5 numbered seeds at approximately equal distance from the center. After 48h, most seeds had germinated and the covers of the petridishes were removed to prevent obstruction of growth. Since water could now evaporate from the dish, plants were watered with their respective perchlorate concentrations once per day (5mL).

Latin square design:

A latin square is any arrangement of equal rows, columns and treatments in which every treatment is assigned to every row and column exactly once. This type of randomization prevents to a great extent some of the common pitfalls in this type of study. If for example, the lighting conditions are not equal on every side of the table, this could confound the effect of perchlorate concentration. Figure 1 shows the randomly generated latin square used for this experiment.

260 mg/L (Mars)	2600 mg/L	0 mg/L (control)	2.6 mg/L	26 mg/L
26 mg/L	260 mg/L (Mars)	2.6 mg/L	2600 mg/L	0 mg/L (control)
2600 mg/L	26 mg/L	260 mg/L (Mars)	0 mg/L (control)	2.6 mg/L
2.6 mg/L	0 mg/L (control)	2600 mg/L	26 mg/L	260 mg/L (Mars)
0 mg/L (control)	2.6 mg/L	26 mg/L	260 mg/L (Mars)	2600 mg/L

Figure 1. Latin square used for perchlorate concentrations.

Timelapse

Using a DSLR and a make-shift set-up, we made photos every 30 seconds. Due to a lack of professional equipment like shutter timers and a camera that does not run on a battery, we had to improvise. We missed a few hours overnight, but this should not pose a problem for the time until event analysis, so long as we take interval censoring into account. In the video below, we marked the petridishes A, B, C, D, E, corresponding to 0 mg/L, 2.6 mg/L, 26 mg/L, 260 mg/L and 2600 mg/L perchlorate, respectively.

Results
Visual inspection

Concentrations of 0 mg/L (control) and 2.6 mg/L appeared to grow fastest. Likewise, the highest concentrations of perchlorate appeared to grow poorest, with several seeds never having sprouted at all. Plants were measured by taking photos next to a ruler and comparing their size in imageJ to the ruler. Per petridish, the biomass including root and seed shell was freeze-dried for later weighing and subsequent spectrometry analysis.

Toxicity to plants: growth after 5 days

After determining plant growth in mm using imageJ, we sorted the results per condition and visualized the effect as in figure 2. Since the seeds in a single petridish are pseudoreplicates if we respect the effect of a dish, we used the median plant length per dish. It is also possible to use the maximum length per dish or the sum of lengths per dish, but we prefer using the median, since it better takes into account the zeros from ungerminated seeds. A more ambitious approach would be to determine the mode of an empirical Bayesian estimate of the density, but we did not attempt this due to time constraints.

Figure 2. Barplots of measured plant length (from root to top) per petridish, sorted by concentration. The x-axis represents the length in mm and the y-axis shows each individual plant. White (A) corresponds to the lowest concentration (0 mg/L control) and red (E) corresponds to the highest concentration of perchlorate (2600 mg/L). A clear trend is visible towards stunted growth in higher perchlorate concentrations.

Subsequently, a linear model of the form y = log(x + 1) + eta was used to obtain estimates and p-values. is necessary since the controls have a concentration of 0. Visual inspection of the residuals showed no serious deviation from normality or homoscedasticity. And no outliers, although one petridish in the 2.6 mg/L group (B) had high leverage. Using a shapiro-wilks test for normality, we determined that normality was not violated significantly (p = 0.365). Although there was significant heteroscedasticity using a non-constant variance test (p = 0.025), this was largely due to the aforementioned petridish in the 2.6 mg/L group that had grown abnormally large. Ignoring this observation renders the residuals homoscedastic (p = 0.066) and normally distributed (p = 0.688).

Perchlorate concentration significantly affected median plant length significantly (p = 0.00157) with an estimate of beta = -1.53 (±0.88). However, since this estimate is on the log(x+1) scale, we have summarized the estimated differences from the control in table 1.

Table 1. Summary of growth of garden cress on perchlorate.

	0 mg/L (intercept)	2.6 mg/L	26 mg/L	260 mg/L	2600 mg/L
length (mm)	20.6	18.6	15.5	12.0	8.5
percent decrease	-	9.5%	24.6%	41.5%	58.6%
p-value		0.00157

Toxicity to plants: time until seed germination

Seed germination was difficult to assess from the video, although the raw photos were of sufficient resolution. We will use the metadata of the photos to determine the time of germination from the first photo at which each seeds have germinated. Due to time constraints, we aim to release the survival (time until event) data of seed germination after the competition.

Discussion

We conclude that on Mars, due to perchlorate alone, a significant decrease in growth of around 40% can be expected. A follow-up of this study will aim to determine the time until germination and the extent to which plants accumulate perchlorate in the soil.

Based on the results of this study, we determined that for the subsequent study on simulant Martian soil, it would be best to measure the plants for a slightly longer time than 5 days, e.g. 7-10 days, depending on the length of the plants by that time. Note that once the plants have a certain length, their growth will be limited by the petridish and the other seeds inside it. Ideally, the experiment should be stopped prior to this time.

Future research

In collaboration with the metabolomics group in Leiden, we will assess the concentration of perchlorate in the dryweight of the plants. This way, we can determine how much of the perchlorate in the solution ends up being accumulated by the plant. In other words, how dangerous is perchlorate really? Unfortunately, due to time constraints we could not complete this subsequent study.