Allocation of Resources to Cyanogenic Glucosides Does Not Incur a Growth Sacrifice in Sorghum bicolor (L.) Moench

In plants, the production of secondary metabolites is considered to be at the expense of primary growth. Sorghum produces a cyanogenic glycoside (dhurrin) that is believed to act as its chemical defence. Studies have shown that acyanogenic plants are smaller in size compared to the wildtype. This study aimed to investigate whether the small plant size is due to delayed germination or due to the lack of dhurrin derived nitrogen. A novel plant system consisting of totally cyanide deficient class 1 (tcd1) and adult cyanide deficient 1 (acdc1) mutant lines was employed. The data for germination, plant height and developmental stage during seedling development and final plant reproductive fitness was recorded. The possible role of phytohormones in recovering the wildtype phenotype, especially in developmentally acyanogenic acdc1 line, was also investigated. The data on plant growth have shown that the lack of dhurrin is disadvantageous to growth, but only at the early developmental stage. The tcd1 plants also took longer to mature probably due to delayed flowering. None of the tested hormones were able to recover the wildtype phenotype. We conclude that the generation of dhurrin is advantageous for plant growth, especially at critical growth stages like germinating seed by providing a ready source of reduced nitrogen.


Introduction
Secondary metabolites were originally believed to be the by-products of primary metabolism and considered not to be essential for plant survival [1]. However, recent research has shown that secondary metabolites promote plant fitness by improving plasticity under constantly changing biotic and abiotic environmental factors [2][3][4]. The most common types of secondary metabolites, or specialized metabolites as they are now known, are terpenes, phenolics and nitrogen or sulfur-containing compounds such as alkaloids, glucosinolates and cyanogenic glucosides [5]. They are generally not directly involved in primary metabolism, with some exceptions, such as the terpenes, sterols and carotenes [6]. Specialized metabolites are usually involved in a network of complex metabolic pathways, generally making it difficult to assign a specific function to them.
Cyanogenic glucosides (CNglcs) are nitrogen-based plant secondary metabolites present in more than 3000 species, representing angiosperms, gymnosperms and ferns [7]. CNglcs break down and release HCN after hydrolysis with a specific β-glucosidase [7,8]. This process, called cyanogenesis, is believed to have evolved for herbivore defence. However, there is growing evidence that cyanogenic glucosides are also involved in nitrogen turnover, transportation and in response to a variety of biotic of GA 3 application on seed germination. Germination and growth were compared with the acdc1 mutants, that have normal patterns of cyanogenesis in the first few weeks of growth, and with sibling lines that lacked the respective mutations (TCD1 and ACDC1) as well as the elite line from which the mutants were derived. The role of dhurrin in plant reproductive fitness was determined by analysing data for seed weight, days to flowering and final leaf stage at flowering. Plant hormones, in addition to GA 3 , were also applied exogenously to explore the possibility that the phenotype of the developmental mutant line (acdc1) was the result of a disruption in hormone signalling.

Role of Dhurrin in Seedling Development: Emergence, Plant Stage and Height
The first step in determining the effect of dhurrin on seedling growth and development was to measure the impact on emergence, stage and height during the first 17 days after planting (DAP) (Figure 1). Overall, during the earliest stages of development, all genotypes were at a similar developmental stage, but the two mutants were significantly shorter compared to the elite line ( Figure 1B,C). This was not the result of differences in the rate of emergence with no significant difference detected between the rate of emergence of the mutants and the elite parental line ( Figure 1A). Seeds from the nonmutated sibling line TCD1 emerged slightly earlier than any of the other genotypes but the reason for this is unclear. At 5 to 9 DAP acdc1, ACDC1 and tcd1 plants were at similar plant stage, with both TCD1 and elite lines at a significantly more advanced developmental stage. However, a shift in this trend was observed from 13 DAP onward, where the elite line was at the most advanced stage, and all other genotypes were at a similar developmental stage, with the exception of acdc1.
A similar trend was observed in plant height, where a very sharp increase in height of the elite plants was observed from 7 DAP onwards ( Figure 1C). At 3 DAP, the height of all genotypes except TCD1 was similar. TCD1 was the tallest with a height of 0.35 ± 0.04 ( Figure 1C). At 4 DAP the height of elite, TCD1 and ACDC1 was significantly higher compared with the tcd1 and acdc1 mutants, with heights of 1.06 ± 0.03, 1.26 ± 0.02 and 1.06 ± 0.04, respectively, ( Figure 1C). This is an important observation because TCD1 and ACDC1 would contain the same background mutations as the tcd1 and acdc1, respectively, suggesting that the height difference is due to the specific mutations resulting in lower dhurrin content in these plants. After 5 DAP plants of the elite line remained significantly taller than the rest of the genotypes, at 7 (2.49 ± 0.04), 12 (3.76 ± 0.06) and 16 (4.83 ± 0.06) DAP ( Figure 1C). Apparently, the lack of dhurrin in tcd1 plants was only disadvantageous to growth at very early developmental stages (up until 7 DAP). From 7 DAP onwards the height of tcd1 and acdc1 remained similar to their respective sibling lines, TCD1 and ACDC1 ( Figure 1C). At the end of this experiment, elite plants showed superior growth in both developmental stage and height, whereas the acdc1 line remained behind only in plant development from rest of the mutant and sibling lines ( Figure 1B).
Throughout the experiment the height of tcd1 plants remained lower than the rest of the genotypes but remained statistically non-significant ( Figure 1C). To check whether the slow growth of tcd1 was due to a small embryo size, the embryo weight of elite, tcd1 and TCD1 seeds were also measured ( Table 1). The average embryo weight of TCD1 (0.42 mg) and tcd1 (0.31 mg) was greater than the elite embryos (0.26 mg), suggesting the observed difference in growth was not due to an initial small embryo size. However, all the mutant and sibling lines were shorter compared to the elite line, and it is possible that this may be the result of random EMS-induced background mutations.
the growing embryo mainly relies on seed nitrogen contents, but soon after emergence, the dependence for growth shifts from seed stored nutrients to photosynthesis. Another possible reason for shorter tcd1 plants could be the lack of an auxin like compound, p-hydroxyphenylacetic acid [35]. This compound is derived from dhurrin turnover and may be acting as a weak auxin. Auxins are known to promote stem elongation in plants [36,37] and removing the dhurrin pathway may be reducing auxin-like activity [10,15,23].   Figure 1. Seedling growth of sorghum genotypes (wild type parent, adult cyanide deficient 1 (acdc1) mutant, acdc1 sibling (ACDC1), totally cyanide deficient class 1 (tcd1) mutant and tcd1 sibling (TCD1)) recorded for 17 days after sowing in soil; (A) days to 50% emergence; (B) plant stage converted to a numeric scale; (C) plant height in cm. Each data point is mean of at least 250 replicates ± 1 standard error. Bars with the same letter at one timepoint are not significantly different from each other (p > 0.05; Tukey's HSD pairwise test). Table 1. Seedling and reproductive growth data of the selected three sorghum genotypes: Elite (parent line from which mutants are derived), totally cyanide deficient class 1 (tcd1) mutant and sibling line of tcd1 mutant without the target mutation (TCD1).
Blomstedt et al. [34] and Bjarnholt et al. [10] argue that the slow growth of tcd1 plants may be the result of lack of reduced nitrogen derived from dhurrin. Data presented here are consistent with this view, but that this effect is only relevant at very early stages of development. During seed maturation dhurrin concentrations decrease until there is none present in mature seeds [20]. This decrease is due to the activity of the alternate turnover pathway as no dhurrinase transcripts were detected, suggesting that the detoxification pathway does not function in seeds. The turnover pathway would result in the storage of nitrogen in preparation for germination and growth. Therefore, the lack of dhurrin derived reduced nitrogen during seedling development could be the reason for slow growth of tcd1 plants ( Figure 1C). The data in the current study suggest that the positive effect of dhurrin on seedling growth is only visible at the early stage of plant growth and starts to disappear over time. This may be explained by the fact that at the start of germination in soil, the growing embryo mainly relies on seed nitrogen contents, but soon after emergence, the dependence for growth shifts from seed stored nutrients to photosynthesis. Another possible reason for shorter tcd1 plants could be the lack of an auxin like compound, p-hydroxyphenylacetic acid [35]. This compound is derived from dhurrin turnover and may be acting as a weak auxin. Auxins are known to promote stem elongation in plants [36,37] and removing the dhurrin pathway may be reducing auxin-like activity [10,15,23].

Effect of Gibberellic Acid (GA 3 ) on Germination and Early Plant Growth
The first experiment in this study examined the possible differences in rates of emergence of seeds from acyanogenic and cyanogenic plants ( Figure 2). In order to capture more detail of the impact on germination itself, the same genotypes were grown in vitro on vertical plates and monitored for five days ( Figure 2). A time-lapse video was made using images taken during germination of seeds in vertical plates, which can be found at <https://youtu.be/fNWzvSFqchU>. The data for plant germination were collected at 12 h intervals, up to 5 DAP ( Figure 3). In addition to germination, root and stem length was also recorded up to 72 h after planting ( Figure 4). The hormone GA 3 plays important roles in both germination and plant growth. To investigate the potential effect of GA 3 on the acyanogenic sorghum mutants, that we had already established were somewhat smaller, we grew all five genotypes under both control and GA 3 treatments ( Figure 2). While the germination indices showed high variability in the mutant and respective sibling lines, compared to the elite (Figure 3), there were several noteworthy observations. First, the final germination percentage (FGP) and mean germination time (MGT) values were not significantly different across all tested genotypes under either control or GA 3 treatment ( Figure 3A,B). The elite line showed maximum percentage of germination when grown in vitro (on MS medium in constant room temperature) ( Figure 3A). Second, the tcd1 mutant line showed lower FGP (27.78 ± 11.11) compared to the elite line (77.78 ± 5.56) under GA 3 treatment ( Figure 3A) although this was not statistically significant. In addition to low FGP, tcd1 also took longer to germinate as indicated by a higher MGT (6.67 ± 1.86) value as compared to the elite (2.42 ± 0.19) under GA 3 treatment ( Figure 3B). Third, the germination index (GI), considered by some to be a better indicator of germination rate, was significantly lower in tcd1 than the elite plants in both treatments ( Figure 3C).  Figure S1 and through the following link <https://youtu.be/fNWzvSFqchU>  Plants growing in the GA3 treatment increased the internode length but did not affect the root length (Figures 2 and 4), consistent with known functions of GA3 on growth. The difference in internode length increased over time, becoming statistically significant 60 h after shoot emergence for most of the genotypes ( Figure 4A). Internode length did not vary across tested genotypes in either  Overall stem length of all five genotypes was not significantly different in the control or GA3 treatment groups ( Figure 4A). Thus, the hypothesis that the slow growth of tcd1 might be due to lower endogenous GA3 is rejected. Another possibility is that that inability of tcd1 to complete glycosylation of the aglycone could be causing an over accumulation of free glucose in the cytosol, altering sugar signalling and the downregulation of photosynthesis [39].
The observation that 72 h after germination both tcd1 and elite lines are the same height is at odds with our observations in a previous experiment where tcd1 plants were smaller than TCD1 and elite genotype at 3 DAP. An important difference between these two experiments is the availability of light. In the present experiment, seeds were grown in transparent medium and soon after the shoot emergence the growing seed can start to photosynthesise, whereas in the previous experiment, seeds were sown into soil, and were still below ground at this stage. A germinating plant is dependent on the energy and nitrogen resources in the seed until it can photosynthesise, on the one hand, and develop roots to access nutrients from the soil, on the other. We hypothesise that dhurrin may supply both reduced nitrogen and glucose in these early stages, and the lack of dhurrin in the tcd1 mutants might be the reason for relatively slow growth of tcd1 plants in soil. Whether this is a direct result of resource limitation or due to an imbalance in the carbon to nitrogen ratio, as proposed by Paul and Driscoll [40], cannot be determined from the results presented here.  Plants growing in the GA 3 treatment increased the internode length but did not affect the root length (Figures 2 and 4), consistent with known functions of GA 3 on growth. The difference in internode length increased over time, becoming statistically significant 60 h after shoot emergence for most of the genotypes ( Figure 4A). Internode length did not vary across tested genotypes in either control or treated plants ( Figure 4A). Exogenous GA 3 application has been shown to be able to rescue the wild type phenotype in mutants lacking the ability to synthesize endogenous GA 3 [38].
Overall stem length of all five genotypes was not significantly different in the control or GA 3 treatment groups ( Figure 4A). Thus, the hypothesis that the slow growth of tcd1 might be due to lower endogenous GA 3 is rejected. Another possibility is that that inability of tcd1 to complete glycosylation of the aglycone could be causing an over accumulation of free glucose in the cytosol, altering sugar signalling and the downregulation of photosynthesis [39].
The observation that 72 h after germination both tcd1 and elite lines are the same height is at odds with our observations in a previous experiment where tcd1 plants were smaller than TCD1 and elite genotype at 3 DAP. An important difference between these two experiments is the availability of light. In the present experiment, seeds were grown in transparent medium and soon after the shoot emergence the growing seed can start to photosynthesise, whereas in the previous experiment, seeds were sown into soil, and were still below ground at this stage. A germinating plant is dependent on the energy and nitrogen resources in the seed until it can photosynthesise, on the one hand, and develop roots to access nutrients from the soil, on the other. We hypothesise that dhurrin may supply both reduced nitrogen and glucose in these early stages, and the lack of dhurrin in the tcd1 mutants might be the reason for relatively slow growth of tcd1 plants in soil. Whether this is a direct result of resource limitation or due to an imbalance in the carbon to nitrogen ratio, as proposed by Paul and Driscoll [40], cannot be determined from the results presented here.

Role of Phytohormones in Dhurrin Biosynthesis
The acdc1 line used in this study has a mutation in the promoter of CYP79A1 gene, the first step in the dhurrin biosynthesis pathway [41]. Based on the observation that this line is developmentally acyanogenic, i.e., dhurrin concentration decreased earlier in development and to a lower level compared with wild type plants, and that the mutation appears to be regulatory we hypothesized that this phenotype might be the result of altered phytohormone signalling. To investigate whether the acdc1 phenotype is associated with a disruption in plant hormone biosynthesis or a signalling pathway, six different hormones were exogenously applied to young sorghum plants. In this experiment, the tcd1 mutant line acts as a negative control of dhurrin synthesis. Overall, none of the applied hormones significantly changed the HCNp across all tested genotypes (Figures 5 and 6). We conclude, therefore, that the altered pattern of dhurrin synthesis in the acdc1 mutant is not hormonally mediated.
Drought is known to induce dhurrin synthesis in sorghum, and ABA is known for regulating the plant response to osmotic stress [42][43][44][45]. Jasmonate-responsive transcription factors are potential regulators of plant secondary metabolites in response to biotic stress [46], but relatively little is known about their effect on cyanogenesis [47,48]. In the present study, foliar application of ABA and MeJA did not result in any detectable change in the dhurrin synthesis in any of the tested genotypes, contrary to our hypothesis (Figures 5 and 6). This could be because of the age of the plants being tested. We applied the hormones to very young plants (18 DAP) compared with 45 DAP in the study by Shehab et al. [43]. Dhurrin concentrations are very high in young plants [14] and the genes are already highly expressed, thus any increase in expression may not be detectable. A more detailed experiment focusing on the effect of exogenous ABA and MeJA application on sorghum across different developmental stages might shed more light on the role of these hormones in dhurrin regulation.

Role of Dhurrin in Reproductive Fitness
Elite, tcd1 and sibling TCD1 genotypes were grown to maturity in the glasshouse and data for days to flowering, seed weight and final leaf number were recorded (Table 1). Elite plants flowered 91 ± 1 DAP, significantly faster than TCD1 (97 ± 2 DAP) and tcd1 (103 ± 1 DAP). Total leaf number at flowering was similar in both elite (12.4 ± 0.1) and tcd1 (12.4 ± 0.3) genotypes (Table 1), thus while there is a difference in time to flowering there was no significant difference in developmental stage. The lack of dhurrin in tcd1 might have caused the delay in flowering, hence requires more time to mature ( Table 1). Lack of dhurrin was not associated with reduced seed weight, with the mass of tcd1 seeds (19.8 ± 1.3) higher than the elite (18.0 ± 0.3) genotype (Table 1). At early stages of sorghum seed development dhurrin is present but begins to disappear as the seed matures, hence the mature seeds are acyanogenic [20]. The presence of dhurrin in immature seeds suggests it may play a role in reproductive development or as stored nitrogen. Other studies have also linked the presence of dhurrin at flowering with drought tolerance although not with changes in rate of development [49]. pathway, six different hormones were exogenously applied to young sorghum plants. In this experiment, the tcd1 mutant line acts as a negative control of dhurrin synthesis. Overall, none of the applied hormones significantly changed the HCNp across all tested genotypes (Figures 5 and 6). We conclude, therefore, that the altered pattern of dhurrin synthesis in the acdc1 mutant is not hormonally mediated.  Drought is known to induce dhurrin synthesis in sorghum, and ABA is known for regulating the plant response to osmotic stress [42][43][44][45]. Jasmonate-responsive transcription factors are potential regulators of plant secondary metabolites in response to biotic stress [46], but relatively little is known about their effect on cyanogenesis [47,48]. In the present study, foliar application of ABA and MeJA did not result in any detectable change in the dhurrin synthesis in any of the tested genotypes, contrary to our hypothesis (Figures 5 and 6). This could be because of the age of the plants being tested. We applied the hormones to very young plants (18 DAP) compared with 45 DAP in the study by Shehab et al [43]. Dhurrin concentrations are very high in young plants [14] and the genes are already highly expressed, thus any increase in expression may not be detectable. A more detailed experiment focusing on the effect of exogenous ABA and MeJA application on sorghum across different developmental stages might shed more light on the role of these hormones in dhurrin regulation.

Implications for Function of Cyanogenesis in Defence and Growth
Slow growth, or the smaller size, of tcd1 sorghum plants in the first few weeks of growth is not from delayed germination or sensitivity to GA 3 . It is more likely that the lack of readily available reduced nitrogen associated with rapid dhurrin synthesis after imbibition may be the reason why tcd1 plants are shorter, and hence appear to be slow growing. The totally acyanogenic tcd1 plants also take longer to mature and have delayed flowering. Dhurrin is believed to be primarily involved in plant defence and high concentration of dhurrin at germination may have evolved to prepare plants against possible herbivore attack. Resource allocation theories universally predict that diverting carbon and nitrogen to defence should result in a growth penalty. The fact that this has been hard to detect is generally attributed to the low concentrations of resources required to make these compounds, the low cost of synthesis, or that the costs of synthesis are offset by other beneficial processes. Here we demonstrate, using sorghum mutants lacking dhurrin, that dhurrin is advantageous for growth in the early stages of plant development, especially in the absence of photosynthesis, but not in older plants. Rather than being a cost, we speculate that the putative role of dhurrin as a source of nitrogen for the germinating seedlings is critically important for rapid growth of sorghum and may help it to establish in the dry tropics where rainfall may be unreliable.

Plant Material and Growth Conditions
Five Sorghum bicolor genotypes were used in this study: The parent line from which the mutants were derived (elite), the adult cyanide deficient class 1 (acdc1), the total cyanide deficient 1 (tcd1) mutants and two sibling lines from the mutated populations that are from the same initial selection but lack the respective mutation (ACDC1 and TCD1) [23]. Plants were grown in the Plant Science Complex, Monash University under natural light conditions with a mean temperature of 26.2 • C ± 7.2 day/19.5 • C ± 4.7 night, with a relative humidity of 51.7% ± 13.6/59.11% ± 11.9 day/night. Seeds were planted in Debco seed raising mix and perlite (3:1 ratio) in 135 mm pots (1.5 L capacity) pots or 30-cell seedling trays.

The Role of Dhurrin in Early Growth and Development of Sorghum
To investigate the role of dhurrin in early sorghum growth and development, 450 seeds of each genotype except tcd1, were grown in the greenhouse. For tcd1 genotype 900 plants were grown instead of 450, because previous studies had shown that tcd1 does not germinate well. The time taken for 50% of the seeds to germinate/emerge was recorded. Data for plant stage (leaf number) and stem height were measured for 17 days after planting (DAP). The data for plant stage were measured on 5, 7, 9, 13, 14, and 17 DAP. For plant height, data were recorded on 3, 4, 5, 7, 12, and 16 DAP. The data for plant stage were converted into numeric form and based on numerical rankings (Supplementary Table  S1). To determine if the difference in embryo weight is responsible for the initial slow growth of tcd1, the weight of elite, tcd1 and TCD1 embryos was determined by soaking seeds until fully imbibed (12 h), followed by dissecting the embryo from the endoplast. Isolated embryos were oven dried at 60 • C for two days then weighed. Due to the small size 20 isolated embryos were weighed together to calculate the average weight.

Role of Dhurrin on Sorghum Reproductive Fitness
To determine the effect of lack of dhurrin on sorghum reproductive fitness, elite, tcd1 and TCD1 plants at the 3-leaf stage, were transferred to 250 mm pots with 8.5 L capacity and watered daily using a drip irrigation system. Plants were grown to maturity, and days to flower, final leaf number at flowering and total seed weight were recorded to analyse plant reproductive fitness.

Effect of GibberellicAacid (GA 3 ) on Germination
To investigate the role of GA 3 on seed germination and early growth, seeds were germinated and grown in vitro on vertical plates. Seeds of each genotype were surface sterilized with 6.25% sodium hypochlorite and 0.1% Tween-20 solution for 5 min followed by 30 s washing with 70% (v/v) ethanol. The final wash included 0.1% (w/v) mercuric chloride for 6 min after which seeds were rinsed in sterile double distilled water six times. After this seeds were placed in vertical square culturing plates (10 × 10 cm) containing half strength MS medium (2.22 g L −1 of MS Salts with Gamborg vitamins, 10 g L −1 of sucrose, 1.8% phytagel and pH adjusted to 5.7) [50] solidified with 1.8% phytagel. In the treatment group, 50 µM of GA 3 (filter sterilised) was added to the medium after autoclaving, whereas the control treatment was just half strength MS medium. For each treatment there were three technical replicates for each of the five genotypes with each replicate having six biological replicates. Data for seed germination were recorded for 10 days after planting (DAP) the seeds on the medium. Plates were placed in a 28 • C constant temperature room under fluorescent light of approximately 100 µmol s −1 m −2 (16 h/day). Throughout the experiment pictures were taken every 6 h to create a 5 s time-lapse video of germinating seeds, for each vertical plate using Windows "Movie Maker 10" with frame rate of 29.97 s −1 . The final germination percentage (FGP), mean germination time (MGT), germination index (GI) and median germination time (t50) were calculated by following Aravind et al. [51].

Role of Hormones in the Biosynthesis of Dhurrin
The potential role of plant hormones in dhurrin biosynthesis was investigated. All five genotypes elite, acdc1, ACDC1, tcd1 and TCD1 were grown in 30-cell seedling trays under glasshouse conditions. A total of six different hormones consisting of salicylic acid (SA; 1 mM), indole-3-acetic acid (IAA; 100 µM), kinetin (KN; 20 µM), abscisic acid (ABA; 150 µM), methyl jasmonate (MeJa; 100 µM) and ethylene (E; 1 mM) with two controls (with and without bag) were used in this experiment. Due to the volatile nature of methyl jasmonate (MeJa) and ethylene (E), plants for both these treatments and their controls were covered with bags and kept in a separate glasshouse. Hormone solutions were sprayed 17 DAP, for the control the same carrier solution was sprayed without any hormone. To avoid the edge effect, plants from the middle of the seedling tray were chosen for harvest. To ensure that plants were at a similar ontogenetic stage individual with fully unfolded fourth leaf and height closest to the average for that particular genotype were selected for harvesting. Plants were harvested at 0, 2, 4, 8, 24, and 48 h after hormone application. Each treatment contained five genotypes with three replications per time-point. Tissue was oven dried, ground and used for cyanide assays.

Measurement of Cyanogenic Glucoside Concentrations
The dhurrin concentration was measured as cyanide potential (HCNp), i.e., the total amount of cyanide released from endogenous cyanogens using 9.5-10.5 mg of dried ground plant tissue following O'Donnell et al. [45]. Exogenous β-glucosidase (Sigma, Kawasaki, Japan, EC 3.2.1.21) in 0.1 M citrate buffer (trisodium citrate, Sigma), pH 5.6 was added in excess to ensure complete hydrolysis of dhurrin to HCN and the evolved HCN was captured in a 1 M NaOH solution and measured as NaCN in a colorimetric assay as described by Gleadow et al. [52]. Absorbance was measured at 580 nm using a microplate reader spectrophotometer (Fluostar Galaxy, BMG Labtechnologies). The cyanide concentrations were determined by comparison to the NaCN standard curve (0, 2, 5, 10, 20, 50, 100 and 150 µM) included in each microtitre plate and converted to the unit mg CN g −1 DW.

Data Measurements and Statistical Analysis
The seed emergence was recorded when at least 2 mm of shoot appeared from the soil. Seed germination was recorded when at least 2 mm of radical emerged from the seed. Stem height was measured from the base of the plant to the ligule of the youngest leaf. Qualitative data of plant stage were converted into numeric values using the ranking system given in Supplementary Table S1, prior to further data analysis. For the vertical plate experiment data for root and shoot length were measured from images using ImageJ software [53]. The significance of the data was calculated using analysis of variance (ANOVA) in IBM SPSS Statistics V25 [54]. In case of significant mean difference, post hoc comparisons were made using the Tukey's test. For all tests p < 0.05 was considered significant.