Overexpression of a novel microRNA IamiR-4-3p from water spinach (Ipomoea aquatica Forsk.) increased Cd toxicity and accumulation in Arabidopsis thaliana

The function of IamiR-4-3p was investigated by using wild type (WT), transfected with empty vector pCambia1302 (CK) and IamiR-4-3p transgenic Arabidopsis in this study. The expression level of GST3 was reduced by 20% in the transgenic Arabidopsis (p35S::miR-4-3p Arabidopsis) when compared to WT, and both of its shoot and root were shorter than WT and CK. After 3 d Cd treatment, root Cd concentrations of p35S::miR-4-3p Arabidopsis was signicantly higher than WT and CK, while no signicant difference was found in shoot Cd concentrations. MDA and H 2 O 2 concentrations were positively correlated with the Cd concentrations in Arabidopsis. Interestingly, even though there was no signicant difference among the shoot Cd concentrations, shoot MDA and H 2 O 2 of p35S::miR-4-3p Arabidopsis were higher than those of WT and CK, and shoot T-AOC exhibited a opposite trend. These results are clearly related to the lowered expression of GST3 by the overexpression of miR-4-3p in p35S::miR-4-3p Arabidopsis. It is suggested that the function of IamiR-4-3p is able to diminish the expression level of GST3, and is responsible to the growth dwarf, higher Cd uptake and oxidative damage but not the Cd translocation from root to shoot in Arabidopsis.


Introduction
Cadmium (Cd) is a common toxicity heavy metal, which tends to deposit in human kidney tubules, guts, bones and other organs via the food chain, posing a severe threat to human health (Ishikawa et al. 2012;Rafati Rahimzadeh et al. 2017). Breeding and selecting pollution-safe cultivars (PSCs) is an effective way to alleviate Cd accumulation in crops and reduce the Cd accumulation risks in human bodies (Yu et al. 2006). Water spinach (Ipomoea aquatica Forsk.) is an important leafy vegetable in south China. In our previous studies, a a typical Cd-PSC with low-shoot-Cd (QLQ) and a high-shoot-Cd accumulation cultivar (T308) had been screened from 30 water spinach cultivars (Wang et al. 2009;Xin et al. 2010).
Furthermore, the different Cd accumulation abilities of QLQ and T308 were largely determined by the different transcriptomic pro les of Cd absorption and detoxi cation (Huang et al. 2016). The regulatory mechanism for the transcriptomic performance remains unknown and needs further stepped into discovery. Therefore, it is of great value to investigate the low-shoot-Cd accumulation mechanisms of water spinach by comparing the different molecular regulation level of Cd absorption and detoxi cation between QLQ and T308.
Glutathione transferases (GSTs) are reported to be a family of multifunctional enzymes that involve in various biotic and abiotic stresses amelioration including heavy metal stresses (Dixit et al. 2011). The biotic and abiotic stresses would bring about the boosted production of reactive oxygen species (ROS), resulting in DNA degradation and lipids damages to plants (Jalmi and Sinha 2015). Cd is known to induce ROS and could trigger oxidation damage to plant. One of the important functions of GSTs is to eliminate the oxidative damage by conjugating the generated ROS products with glutathione (GSH) in plants (Nianiou-Obeidat et al. 2017). Moreover, GSTs can also function in abiotic stress defense by improving the biosynthesis of sulfur-containing secondary metabolites, such as phenolics, glucosinolates and avonoids (Dixit et al. 2011). In most cases, the increased expressions of GSTs could enhance Cd tolerance in organisms, by promoting the formation of low-toxic glutathione (GSH)-Cd complex . However, whether the GSH biosynthesis would in uence the Cd accumulation in plant remains unclear.
A novel miRNA (IamiR-04m-3p: GATCACAGTGTATTATGGTGTCC) expressed differently between QLQ and T308 under the Cd treatment was found in our previous study. The expression level of IamiR-4-3p decreased signi cantly in T308 but kept stable in QLQ under Cd stress (Shen et al. 2017), indicating that IamiR-4-3p might play critical role in the cultivar-dependent Cd responsive regulation. GST3 was identi ed as one of the dominant targets of IamiR-4-3p according to psRNATarget server and the target area located at the 3' untranslated region (3' UTR) of GST3 indicated that IamiR-4-3p may played an important role in regulating the GST3 expression levels. Therefore, this research has important signi cance to uncover the molecular mechanisms of IamiR-4-3p in the Cd accumulation ability of water spinach.
In the present study, we developed transgenic Arabidopsis thaliana (Arabidopsis) expressing the IamiR-4-3p from water spinach to investigate the function of IamiR-4-3p in Cd accumulation. It is hypothesis that: 1) IamiR-4-3p targeted in the GST3 expression levels in transgenic Arabidopsis. 2) Oxidative damage from Cd toxicity was elevated by IamiR-4-3p in the p35S:: ia-miR-4-3p Arabidopsis, which should attribute to the retarded generation of GSH-Cd complex catalyzed by GST3. 3) IamiR-4-3p displayed signi cant role in the higher Cd uptake but not the Cd translocation from root to shoot in the transgenic Arabidopsis. It is expected that this study would be helpful to elucidate the critical role of IamiR-4-3p on Cd accumulation through GST expression in water spinach and provide effective molecular understandings for the breeding of low Cd cultivars.

Plant materials and growth condition
Water spinach (T308) and Arabidopsis thaliana (ecotype Columbia) were used in the present study. Seeds of water spinach and Arabidopsis thaliana were surface disinfected in the presence of 0.7% NaClO for 5 minutes, then thoroughly washed with deionized water for 3 times. After germination, the seedlings were cultivated in pots lled with growth media and watered with half-strength Hoagland nutrient solution every 3 days. T308 was cultivated in greenhouse with the light cycle of 12h light / 12h dark, 30 ℃ every day. After 4 weeks, T308 roots were harvested for DNA extraction. Arabidopsis thaliana grew under suitable conditions in a plant incubator, which was used for plant transformation and Cd treatment (de Felippes, Ott, and Weigel 2011).

Vector constructs and plant transformation
The genomic DNA of water spinach was extracted by CTAB method (Schiebelhut et al. 2017) and then tested by nanodrop 2000c (Thermo scienti c, USA). The IamiR-4-3p precursor sequence was ampli ed from genomic DNA by PCR using gene-speci c primer pairs designed with NcoI and Spe site (Table S1, supporting information). The ampli ed fragments were digested and cloned into the binary vector, pCambia 1302, containing a CaMV 35S promoter, kanamycin and hygromycin resistance gene. The construct was introduced into E. coli DH5α and con rmed by DNA sequencing. The con rmed constructs was named as p35S::IamiR-4-3p as shown in Figure S1 (supporting information). Also, the empty vector pCambia1302 without any modi cation was utilized as control (CK) in this research. The p35S::IamiR-4-3p and CK vector were introduced into Agrobacterium tumefaciens LBA4404 (Shanghai Weidi Biotechnology, China). Arabidopsis transformation was performed by the oral dip method (Clough and Bent 1998). T2 homozygous lines were used in all experiments presented in the current study. All transgenic plants were selected by 40mg/L hygromycin in 1/2 MS medium, then con rmed by PCR analysis of IamiR-4-3p and green uorescent protein (GFP) with leaf genomic DNA as templates (Table  S1).

RNA isolation and quantitative real-time PCR (qRT-PCR)
Total RNA was separately isolated from shoot and root samples with the aid of the RNA Easyspin Isolation System (Aidlab, China). RNA concentration and quality were tested by nanodrop 2000c (Thermo scienti c, USA). miRNA cDNA Synthesis Kit (ABM, Canada) and EvaGreen miRNA qPCR MasterMix (ABM, Canada) were used for the quantitative of IamiR-4-3p. PrimeScript™ RT regent kit (Takara, Japan) and YBR GreenII PCR Master Mix (Takara, Japan) were used for the quantitative of GST3, the target gene of IamiR-4-3p. All the primers for IamiR-4-3p, GST3 and their reference gene U6 and actin for qRT-PCR were shown in Table S2. The reactions were performed on LightCycler® 480 Real-Time PCR System (Roche, Germany). Each qRT-PCR analysis was done in triplicate and with three biological replicates. Results were analyzed with the integrated LightCycler® 480 service software. Expression levels of the tested genes were determined by CT values and calculated by the ∆∆Ct method (Schmittgen and Livak 2008).

Cd concentration measurement
Six-week-old WT, CK and transgenic Arabidopsis thaliana seedlings were treated with half strength Hoagland solution containing 5 mg/L Cd for 3 days, and 3 replicas were made for WT, CK and transgenic Arabidopsis thaliana. Shoot and root samples were harvested separately, deactivated enzyme at 105°C for 30 min, and then dried at 70°C to constant weight. The dried samples were digested (HNO 3 :H 2 O 2 , 5:1) and then measured by an atomic absorption spectroscopy (Hitachi Z-5300, Japan). To control the precision of the analytical procedures, a certi ed reference material (CRM) of plant (GBW-07603, provided by the National Research Center for CRM, China) with a Cd concentration of 0.38 mg/kg was employed.

MDA, H 2 O 2 concentrations and T-AOC measurement
Shoot and root MDA, H 2 O 2 concentrations and T-AOC in WT, CK and transgenic Arabidopsis thaliana were tested by using MDA, H 2 O 2 and T-AOC assay kits (Suzhou comin biotechnology, China), respectively.
Three repeats were conducted for each treatment.
2.6 Statistical analysis SPSS 23.0 and GraphPad Prism 8 were used for statistical analyses and gures development. One-way ANOVA with the least signi cant difference (LSD) test was performed to evaluate the signi cance of treatment effects. Values were considered to be statistically signi cantly different when p < 0.05.

Transcript analysis of IamiR-4-3p and GST3 in transgenic plants
To ensure the reliability of the transformation of Arabidopsis, the PCR of pre-IamiR-4-3p (gene of interest) and GFP (marker gene) was conducted by using the DNA of leaves samples from each plant. As shown in Figure S2, GFP was detected in both CK and p35S::IamiR-4-3p Arabidopsis plants, and pre-IamiR-4-3p could only be detected in p35S::IamiR-4-3p plants. Meanwhile, neither pre-IamiR-4-3p nor GFP could be detected in WT plants. The results indicated that both p35S::miR-4-3p and CK vectors had been successfully transfected into Arabidopsis. The results of qRT-PCR showed that p35S::miR-4-3p plants exerted signi cantly higher expression levels in IamiR-4-3p, when compared with WT and CK plants.

Effects of IamiR-4-3p on Arabidopsis growth
When compared with WT plants, CK and p35S::miR-4-3p plants did not show evident morphological difference ( Figure 3). However, when compared the growth status of these Arabidopsis plants, we noticed that sizes of p35S::miR-4-3p plants were smaller than those of WT and CK. The height of WT, CK and p35S::miR-4-3p plants were 15.6, 15.3 and 11.5 cm, respectively, and the plant height of p35S::miR-4-3p plants was signi cantly lower than those of WT and CK (p < 0.05). Also the average root length of p35S::miR-4-3p plants was 4.9 cm, which was signi cantly shorter than those of WT and CK (p < 0.05). These ndings suggested that overexpression of IamiR-4-3p inhibited the growth of Arabidopsis.

Cd concentrations in Arabidopsis
In order to investigate the effect of IamiR-4-3p overexpression on Cd accumulation ability in Arabidopsis, Cd contents of shoots and roots of WT, CK and p35S::miR-4-3p Arabidopsis plants was measured at 0 h and day 3 after Cd exposure (5 mg/L CdCl 2 ), respectively. As displayed in Table 1, the results showed that there was no signi cant difference (p > 0.05) in both the shoot and root Cd concentrations among WT, CK and p35S::miR-4-3p Arabidopsis at 0 h. After 3 days of Cd treatment, Cd concentration of the shoot in p35S::miR-4-3p was still similar to those of WT and CK Arabidopsis. However, Cd concentration of the root in p35S::miR-4-3p was 1.13 and 1.10 times of those in WT and CK Arabidopsis respectively at day 3, and the differences between p35S::miR-4-3p and WT as well as CK were signi cant (p < 0.05), indicating that p35S::miR-4-3p plants accumulated more Cd in its root.

MDA concentrations in Arabidopsis
When treated with Cd at 0 h, the shoot MDA concentrations in WT, CK and p35S::miR-4-3p plants were 5.8, 5.5 and 5.4 nmol g -1 (Figure 4), and the differences were no signi cant (p < 0.05). The root MDA concentrations in WT, CK and p35S::miR-4-3p were 8.4, 8.2 and 9.1 nmol g -1 , respectively, and the differences were also without signi cance (p < 0.05). However, after 3 days of Cd treatment, shoot MDA concentrations in WT, CK and p35S::miR-4-3p plants were 6.6, 6.7 and 9.2 nmol g -1 , and root MDA concentrations of those plants were 10.8, 10.7 and 13.5 nmol g -1 , respectively. Both the shoot and root MDA concentrations in p35S::miR-4-3p plants were signi cantly higher than those of WT and CK plants (p<0.05). These results showed that the p35S::miR-4-3p plants suffered more lipid peroxidation damage when treated with Cd.

H 2 O 2 concentrations in Arabidopsis
As shown in Figure 5 were signi cantly higher than WT and CK (p<0.05), which also indicated that p35S::miR-4-3p plant was subjected to a higher oxidative stress.

T-AOC concentrations in Arabidopsis
As shown in Figure 6, before Cd treatment (0 h), the T-AOC contents of both shoot and root in WT, CK and p35S::miR-4-3p Arabidopsis were about 15 U/mg, and there was no signi cant difference was found (p>0.05). The shoot and root T-AOC contents in WT, CK and p35S::miR-4-3p plants were signi cantly increased after 3 days of Cd treatment. Shoot T-AOC contents in WT, CK and p35S::miR-4-3p plants were 30.6, 31.0 and 25.8 U/mg, respectively, and root T-AOC contents in these plants were 40.0, 41.5 and 31.35 U/mg, respectively, after 3-day Cd exposure. Both the shoot and root T-AOC contents in p35S::miR-4-3p plant were signi cantly lower than those of WT and CK plants (p<0.05). indicating that the degree of ROS damage in p35S::miR-4-3p Arabidopsis was elevated.

The GSTs expression level was regulated by IamiR-4-3p in Arabidopsis
As a new miRNA, we found that IamiR-4-3p was able to regulate GST3 expression level in water spinach under Cd treatment. The IamiR-4-3p expression level of T308 was lower than QLQ, and accordingly higher expression level of GST3 was observed in T308 (Shen et al. 2017;Huang et al. 2016). As higher abundance of GSTs has been found in responding to Cd stress (Nianiou-Obeidat et al. 2017; Cao et al. 2017;Zhao et al. 2019), implying the role of IamiR-4-3p in regulating expression of GST3 should be also related to the Cd stress response in water spinach. The qRT-PCR results indicated that IamiR-4-3p was able to down regulate the GST3's expression level in Arabidopsis, but the performance was weaker than some other miRNAs investigated previously. Lin et al found that the miRNA160 was capable of down regulating the expression levels its targets, ARF10, ARF16 and ARF17, by 65% in Arabidopsis (Lin et al. 2018). (Li et al. 2013) discovered that OsmiRNA396c was able to down regulate the expression level of OsGRF4 by 80%. Also, the function study of miRNA in cotton suggested that ghr-miR414c had the capacity to down regulate the expression level of GhFSD1 by 75% ). In the present study, IamiR-4-3p was able to down regulate the expression level of GST3 by only 20% in Arabidopsis, which was lower than the previous reports. This low regulation e ciency might probably be ascribed to the low sequence similarity (50.15%) of GST3 between water spinach and Arabidopsis, leading to the defects in regulation ability of IamiR-4-3p in Arabidopsis. The target preference of IamiR-4-3p might be another responsible conjecture for the low regulation e ciency in GST3. Multiple target genes have been found for IamiR-4-3p in water spinach, such as RNA-capping enzyme-like and TMV-associated RING nger protein, etc (Shen et al. 2017). The target preferences were also observed between other miRNAs and their target genes. For example, the miRNA827 preferred to regulate the expression of nitrogen limitation adaptation (NLA) in Arabidopsis and phosphate transporter 5(PHT5) in Oryza sativa (Lin et al. 2018). Similarly, miRNA21 exerted target preference for different regulatory capacities which was 75% down regulation in acidic nuclear phosphoprotein and NADP but only 35% down regulation in cold shock domain-containing protein and high-density lipoprotein-binding protein1 (Schramedei et al. 2011).
Therefore, it is likely that IamiR-4-3p had the preference to regulate the expression of the other target genes other than GST3 in Arabidopsis. Also, the IamiR-4-3p overexpressed Arabidopsis (six-week-old) presented with growth retardation, which was indicated by the 20% and 15% shorter shoot and root than WT as well as CK. As Arabidopsis were not treated with Cd during the rst six weeks, the overexpression of IamiR-4-3p still in uenced the growth of Arabidopsis, which stood a good chance that IamiR-4-3p should have target preference, and the other target genes maybe related to the regulation of growth and development. For a better understanding of the function of IamiR-4-3p, therefore, the IamiR-4-3p overexpressed water spinach should be constructed, and the target preference of IamiR-4-3p should be further studied. T-AOC is an important antioxidant for plants under heavy metal stresses, and the lower the content of T-AOC in plants, the more oxidative damage will plants suffer (Sytar et al. 2013). In the present study, we also found that there was a signi cant negative correlation between T-AOC and Cd concentration in plants, which was consistent with the research of (Zhan et al. 2018), who found that a signi cant negative correlation (-0.856, p < 0.01) existed between T-AOC and Cd contents in maize. Therefore, the IamiR-4-3p overexpression reduced the antioxidation ability in Arabidopsis, which should be one of the reasons of the increases of MDA and H 2 O 2 contents in the p35S::miR-4-3p plant.

Oxidative damage from
As is well known, GSTs was considered to involve in the plant protective mechanism under various stresses, by regulating the reversible S-glutathionylation of protein thiol residues (Mieyal and Chock 2012). The overexpression of GST gene can signi cantly enhance salt, drought and heavy metals resistance of tomato tobacco and rice (Csiszár et al. 2014;Dixit et al. 2011;Kumar et al. 2013). As for water spinach, Wang et al found that the level of Cd in the NaCl-extractable form, which is related to the content of Cd-binding complex, was higher in T308 than in QLQ, indicating that the Cd chelated with proteins or peptides in T308 should be more than that in QLQ (Wang and Ren 2014;Zhang et al. 2010). In our other study, it was found that T308 possessed a higher sulfur metabolic protein and GST3 expression level, which was helpful to reduce toxicity of Cd and improve Cd tolerance by generating GSH-Cd complex (Huang et al. 2016). Because the overexpression of IamiR-m04-3p leaded to a 20% decrease in the expression level of GST3 in Arabidopsis, the detoxi cation capacity through the production of GSH-Cd complex catalyzed by GST3 would be impeded, resulting in a series of oxidative damage caused by Cd in the IamiR-m04-3p overexpressed Arabidopsis. Therefore, the presence of Cd can easily trigger ROS reaction and oxidative damage when IamiR-m04-3p was overexpressed, which explained why shoot of IamiR-m04-3p overexpressed Arabidopsis suffered a more severe oxidative damage than WT and CK Arabidopsis, although their Cd accumulation levels were almost the same.

IamiR-4-3p affects Cd uptake rather than Cd transloction in Arabidopsis
In the present study, an unusual observation has to be explained, that the shoot Cd concentrations among WT, CK and p35S::miR-4-3p plants were not signi cant, although the root Cd concentration of p35S::miR-4-3p plant was pretty higher than those of WT and CK after 3 days of Cd treatment. The results indicated that IamiR-4-3p overexpression exerted no effect on Cd translocation from root to shoot, but it increased Cd uptake in the root of Arabidopsis. It might be attributed to the higher Cd-induced oxidative damage in p35S::miR-4-3p Arabidopsis as above-mentioned. According to Han et al and Javed et al, the oxidative damage would increase the cell membrane permeability and release of organic acids of root, which could promote the Cd uptake in the root (Han et al. 2006;Javed et al. 2017). Besides, multiple target genes of IamiR-4-3p have been found in water spinach, and one of which is AWPM19-like protein (Shen et al. 2017). The function of AWPM19-like protein is to mediate abscisic acid (ABA) in ux through the plasma membrane. Low levels of ABA in plants would result in minimized deposition of apoplastic barriers and allowed maximization of Cd uptake (Tao et al. 2019). It is thus suggested that the IamiR-4-3p overexpression should also reduce the generation of AWPM19-like protein and the in ux of ABA in Arabidopsis, which could explain why Cd uptake was increased in the root of the p35S::miR-4-3p Arabidopsis. Accordingly, we proposed that the higher oxidative damage and the down-regulation of AWPM19-like protein would cause the higher Cd uptake in the root of p35S::miR-4-3p Arabidopsis. As for the shoot Cd concentration of p35S::miR-4-3p Arabidopsis similar to those of WT and CK, it might be attributed to the down regulation of the level of GST3 by IamiR-4-3p. According to He et al, GST was able to promote the produce of GSH-Cd complex, which could increase the mobility of Cd from root to shoot (He et al. 2018). It is considered that the Cd translocation from root to shoot in p35S::miR-4-3p Arabidopsis with lower GST3 expression would be in uenced by the lower generation of GSH-Cd complex, which deterred the transportation of Cd from root to shoot. Therefore, even though the root Cd concentration was higher in p35S::miR-4-3p Arabidopsis, its shoot Cd concentration showed no signi cant difference with WT and CK. Similarly, Cd accumulation in leaves of Vicia faba L. was positively correlated with the contents of PCs-Cd and GSH-Cd compound, it seemed that sulfur-containing compound facilitated Cd uptake in cell walls of leaves without retaining Cd in roots (Wu, Sagervanshi, and Mühling 2018). Matraszek-Gawron et al also stated that PCs and GSH could act as long-distance carriers of Cd in Triticum aestivum L(Matraszek-Gawron and Hawrylak-Nowak 2019). These results could also help to explain the restrained accumulation of Cd in shoots of p35S::miR-4-3p Arabidopsis, which involved in the deterred transportation of Cd with the inhibition of generation of GSH-Cd complexes by ia-miR-4-3p. For another, the function of other target genes of IamiR-4-3p in water spinach, such as RNAcapping enzyme-like and TMV-associated RING nger protein, etc (Shen et al. 2017) is still not clear, and their impacts on the Cd translocation and transportation in p35S::miR-4-3p Arabidopsis as well as water spinach are necessary to be further investigated.

Conclusion
It is concluded that IamiR-4-3p is responsible to the oxidative damage and higher Cd uptake but not the Cd translocation from root to shoot, and to the regulation of AWPM19-like protein and GST3 in Arabidopsis. Oxidative damage from Cd toxicity was elevated in p35S:: ia-miR-4-3p Arabidopsis attributing to the retarded generation of GSH-Cd complex catalyzed by GST3. These ndings should be helpful to elucidate the critical role of IamiR-4-3p on Cd accumulation and transportation in water spinach and provide effective molecular understandings for the breeding of low-Cd cultivars. Other functions of IamiR-4-3p on Cd accumulation and detoxi cation are valuable to be further veri ed.

Acknowledgment
The reviews and editors are appreciated for comments and suggestions to improve the paper.  Notes: Different lower-case letters stand for significant differences among the three types of Arabidopsis thaliana within the same Cd treatment time(p<0.05); Different capital letters stand for significant differences between the two different Cd treatment time within each of the Arabidopsis thaliana types (p < 0.05). Figure 1 The expression level of GST3 (A) and Ia-miR-4-3p (B) in six-week-old WT, CK and p35S::miR-4-3p

Figures
Arabidopsis. Note:error bars of gene expression levels represent the standard deviation among three replicates. Statistical signi cance was de ned at p < 0.05 determined using analysis of variance (ANOVA). Different letters indicate signi cant differences between WT, CK and p35S::miR-4-3p Arabidopsis.

Figure 2
Sequence comparison between IamiR-4-3p and GST3 3' UTR of Arabidopsis and water spinach. Note: At-GST3 stands for the GST3 sequence of Arabidopsis and Ip-GST3 stands for the GST3 sequence of water spinach. The letter with red color means the mismatches between Arabidopsis and IamiR-04-3p, and the letters with blue color means the mismatches between water spinach and IamiR-04-3p.

Figure 4
Effects of Cd treatment on MDA concentration in shoot (A) and root (B) of Arabidopsis thaliana at 0h and 3d. Different letters within the same cultivar indicate signi cant differences at p < 0.05 between different Cd treatment time; ns and * indicate differences between the cultivars within the same treatment is not signi cant at p > 0.05 level and signi cant at p < 0.01 level, respectively.

Figure 5
Effects of Cd treatment on H2O2 concentration in shoot (A) and root (B) of Arabidopsis thaliana at 0h and 3d. Different letters within the same cultivar indicate signi cant differences at p < 0.05 between different Cd treatment time; ns and * indicate differences between the cultivars within the same treatment is not signi cant at p > 0.05 level and signi cant at p < 0.01 level, respectively.