Low pH, Aluminum and Phosphorus Coordinately Regulate Malate Exudation through 1 GmALMT1 to Improve Soybean Adaptation to Acid Soils 2

Low pH, Al toxicity and low P often coexist and are heterogeneously distributed in acid soils. To date, the underlying mechanisms of crop adaptation to these multiple factors on acid soils 3 remain poorly understood. In this study we found that P addition to acid soils could stimulate 4 Al tolerance, especially for the P efficient genotype, HN89. Subsequent hydroponic studies 5 demonstrated that solution pH, Al and P levels coordinately altered soybean root growth and 6 malate exudation. Interestingly, HN89 released more malate under conditions mimicking acid 7 soils (low pH, +P and +Al), suggesting that root malate exudation might be critical for 8 soybean adaptation to both Al toxicity and P deficiency on acid soils. GmALMT1 , a soybean 9 malate transporter gene was cloned from the Al-treated root tips of HN89. Like root malate 10 exudation, GmALMT1 expression was also pH-dependent, being suppressed by low pH, but 11 enhanced by Al plus P addition in roots of HN89. Quantitative real time PCR, transient 12 expression of a GmALMT1-YFP chimera in Arabidopsis protoplasts, and electrophysiological analysis of Xenopus oocytes expressing GmALMT1 demonstrated that 14 GmALMT1 encodes a root-cell plasma membrane transporter that mediates malate efflux in 15 an extracellular pH-dependent and Al-independent manner. Overexpression of GmALMT1 in 16 transgenic Arabidopsis, as well as overexpression and knockdown of GmALMT1 in 17 transgenic soybean hairy roots, indicated that GmALMT1-mediated root malate efflux does 18 underlie soybean Al tolerance. Taken together, our results suggest that malate exudation is an 19 important component of soybean adaptation to acid soils, and is coordinately regulated by 20 three factors, pH, Al and P, through the regulation of GmALMT1 expression and GmALMT1 21 function. 25 26


INTRODUCTION 1
Up to half of the world's potentially arable lands are comprised of acid soils, which limit 2 crop productivity in these regions (von Uexküll and Mutert, 1995). The stresses imposed on 3 crops growing on acid soils consist of proton rhizotoxicity (low pH), nutrient deficiency 4 (primarily phosphorus, but also potassium, calcium, and other minerals) and metal toxicity 5 (aluminum, manganese). Among these constraints, low phosphorus (P) availability, toxic 6 levels of aluminum (Al) and proton rhizotoxicity (low pH stress) are considered to be the 1 0 4A). Expression of GmALMT1 in the root tip was Al-dependent, increasing with increasing 1 Al levels (Fig. 4B). To determine whether GmALMT1 expression was affected by the 2 interaction of pH and Al toxicity, temporal expression patterns of GmALMT1 were also 3 examined in the root tips. As shown in Figure 4C, expression of GmALMT1 in the root tips 4 was regulated both by pH and Al. The levels of GmALMT1 expression were high in the root 5 tips of the seedlings grown in pH 5.8 nutrient solution prior to subjecting the seedling roots to 6 low pH (Fig. 4C). The GmALMT1 expression decreased by more than 12-fold after 6 h of low 7 pH treatment, remaining low over the next 6 hours. Inclusion of Al in the low pH solution properties. Oocytes injected with GmALMT1 cRNA had significantly less negative resting control cells showed no significant currents within the range of holding potentials tested. Since, by convention GmALMT1-mediated inward currents are the product of either net al., 2008), thereby increasing the organic anion efflux driving force. Increasing {mal 2-} I 1 resulted in a concentration-dependent increase in the GmALMT1-mediated inward currents 2 in standard bath solutions at both pH 7.5 and 4.5 (Fig. 8). In contrast, although being pH 3 dependent (i.e. stimulated at lower pH as shown in Fig. 7), GmALMT1-mediated outward 4 currents were independent of changes in {mal 2-} i . Under increased {mal 2-} I conditions, we 5 were also able to establish that acidification of the bath media resulted in inhibition of the 6 GmALMT1-mediated inward (i.e. negative) currents at all {mal 2-} I tested. Analysis of the 7 slope of GmALMT1-mediated inward current (conductance) at pH 7.5 and 4.5 showed that 8 the inward conductance tended to saturate as the {mal 2-} i approach 4 mM, suggesting that the 9 GmALMT1-mediated malate transport saturates with a K m1/2 of about 1 to 2 mM internal 1 0 mal 2-, independent of the extracellular pH value (Fig. 8C). These findings also suggest that 1 1 decreasing the pH of the bath solution from pH 7.5 to 4.5 results in a change in the maximum 1 2 conductance (G max ) rather than a change in affinity for the malate substrate. The dependence pH dependence of the transporter (i.e. lower transport rates at lower pH) suggests that in 2 7 addition to changes in gene expression, the pH dependence of the whole root malate

2
The function of GmALMT1 was further investigated in two Arabidopsis transgenic abundance in the two transgenic lines by qRT-PCR indicated that GmALMT1 was highly 1 expressed in these transgenic but not in CK lines (Fig. 9A). These two overexpression lines 2 showed significantly larger root malate exudation rates and superior Al-tolerance (as 3 determined by relative root growth) compared to CK ( Fig. 9B-D). These results clearly 4 demonstrate that overexpression of GmALMT1 in Arabidopsis promotes root malate release, 5 thereby conferring a significant increase in Al tolerance. the CK. As shown in Figure 10B, the malate exudation rates from the hairy roots of these 1 5 lines was closely correlated with the level of GmALMT1 expression, being 58% lower in the 1 6 GmALMT1-KD lines and 87% greater in the GmALMT1-OX lines, relative to those 1 7 measured in the CK. The function of GmALMT1 related to Al tolerance was also quantified 1 8 in soybean transgenic hairy roots. As this hairy root system is not amendable to relative root roots of the GmALMT1-OX lines accumulated less Al in the root tips than CK roots, as 2 2 indicated by the degree of hematoxylin staining on the surface of the root tips ( Fig.10C).

3
Overall, these results strongly suggest that GmALMT1 does function as a malate efflux 2 4 transporter in soybean roots, chelating free Al at the root surface and excluding it from 2 5 entering the root system, thereby providing Al-tolerance. Highly acidic soil pH, high Al and low P availability are the three major limiting factors 2 9 in acid soils comprising up to 40% of world arable land and severely inhibit world crop  levels of low pH, Al toxicity and P availability during their growth through acidic soil 5 horizons (Table S1; Liao et al., 2006). Therefore, the present study represents, to our 6 knowledge, the first multidisciplinary study to focus on the combined effects of these three 7 major limiting factors in acid soils to plant growth, implementing field evaluations and 8 physiological, molecular and biophysical analysis in the lab, to elucidate the underlying 9 mechanisms of crop adaptation to acid soils.

0
For crops to maintain growth on acid soils, they need to be adaptive to low pH (toxic especially under conditions of P supply (Fig. 1). Further studies in hydroponics also showed 1 5 that the P efficient soybean genotype exhibited greater tolerance to Al when P was supplied supply will be discussed later in this section.

6
Previously, we found that malate exudation from the tips of the more deeply growing tap 2 7 root of the P efficient soybean genotype was specifically enhanced when the upper, more 2 8 shallowly placed lateral roots, were supplied with P while the tap root tip was challenged by 2 9 Al, suggesting that P efficiency resulting in improved P nutrition plays a role in the malate 3 0 exudation critical for Al tolerance (Liao et al., 2006). This hypothesis is also supported by the  -Camacho et al., 2004;Sasaki et al., 2004). In the present study, we used a 2 hydroponic system to analyze the relationships between malate exudation and internal malate 3 homeostasis in response to changes in solution pH, and Al, and P levels in the growth 4 solution. Our results indicate that in control solution (pH 5.8, +P, -Al), root malate exudation 5 was much higher than in the same solution adjusted to pH 4.3 (Fig. 3A). Similar responses 6 have also been observed in other plant species (Bertin et al., 2003). High root malate release 7 has been speculated to be involved in other plant responses to the environment, such as in  , 1986, 1988). Hence, greater secretion of malate from soybean roots at less acidic pH 1 2 might provide more carbon for beneficial microorganisms, such as rhizobia and mycorrhizal 1 3

Martinez
fungi. We also found that low pH significantly reduced both malate exudation and malate 1 4 content in soybean root tips (Fig. 3). Reduction of malate content as a result of low pH stress 1 5 has also been reported in the roots of maize and broad bean (Yan et al., 1992). It has been 1 6 suggested that the reduction of malate content in plant roots might be caused by the malate content of roots may constitute an adaptive change to low pH stress. Moreover, the P 2 0 efficient genotype HN89 always maintained higher root malate content at low pH regardless 2 1 of the P and Al levels in the solution (Fig. 3B), suggesting it has a greater potential to adapt 2 2 to acidic soil conditions.

3
The malate exudation in response to Pi starvation has been suggested as a mechanism 2 4 for P mobilization in the soil in several plant species grown on acid soils (Hoffland et al., had a higher root malate exudation at low pH than did plants supplied with P, these 2 7 differences were not as large as those described earlier (Fig. 3A). Increased malate exudation 2 8 under low P conditions can help to solubilize P that is fixed by Fe and Al oxides in acid soils, low pH conditions was alleviated by Al supply, particularly in the P-efficient genotype HN89 1 6 (Fig. 3A). These results strongly suggest that root malate exudation might be the critical 1 mechanism for soybean adaptation to acid soils, and that the control of the expression of the 2 gene encoding the malate transporter underlying these responses, as well as the activity of the 3 transport protein, might be regulated both by Al exposure and plant P status, as well as low 4 external pH.

5
We cloned a malate transporter (designated GmALMT1) from soybean roots treated with 6 Al using degenerate primers derived from the conserved domain of previously cloned malate 7 transporter proteins from wheat, Arabidopsis and rape. The expression of GmALMT1 was 8 found to be regulated by low pH, Al toxicity and P availability. GmALMT1 expression was The localization to the root tip will greatly reduce the carbon cost to the plant for this 1 5 tolerance mechanism. Similar Al-activated expression patterns were also observed for the present study, we have also revealed that GmALMT1 activity is Al-independent, and its 3 2 transport activity is highly dependent on extracellular pH (Fig. 8), a unique regulatory 3 3 mechanism not yet described for other ALMT-type transporters. Interestingly, not only was 1 7 GmALMT1 transport activity (Fig. 7) regulated by external pH, but also its expression in 1 planta was down-regulated by low pH (Fig. 4C and 5), which closely correlates with the 2 reduction in root malate exudation when the roots were transferred from pH 5.8 to pH 4.3 3 (Fig. 3A). The response of GmALMT1 to low pH makes GmALMT1 quite distinctive from the 4 other known ALMT genes involved in Al tolerance. In these other ALMTs, including 5 AtALMT1 and TaALMT1, their expression is not influenced by or is slightly increased in 6 response to low pH conditions (Sasaki et al., 2004;Kobayashi et al., 2007;Liu et al., 2009). 7 When subjected to the conditions that mimic the acid soil syndrome of low pH, high Al and 8 variable P supply, the expression of GmALMT1 was suppressed by low pH, but increased by 9 Al and P addition, especially for the P efficient genotype HN89 (Fig. 5).
1 0 The functionality of GmALMT1 was also investigated by overexpression in Arabidopsis, 1 1 as well as altering its expression in soybean using the soybean transgenic hairy root system.

2
Overexpression of GmALMT1 in both soybean hairy roots and Arabidopsis led to greater soybean hairy roots by RNAi led to decreased malate exudation from hairy roots (Fig. 10B). transporter that plays a critical role in soybean Al tolerance.

0
Based on field evaluations as well as physiological and molecular analysis in the 2 1 laboratory, we conclude that root malate exudation in soybean is coordinately regulated by 2 2 low pH, Al and P through GmALMT1, which might be the critical mechanism for soybean 2 3 adaptation to acid soils. Therefore, GmALMT1 could be considered as a potential candidate 2 4 gene for crop improvement in acid soils through biotechnological approaches. efficiency were used in this study. A previous study from our lab established that HN89 is a P 3 0 efficient genotype, while HN112 is a P inefficient genotype (Zhao et al., 2004).

1
In the field trial, soybean plants were grown on acidic red soils at the Boluo (E114.28°, of the top 20 cm layer were as follows: pH, 5.71; organic matter, 1.76 mg kg -1 ; available P 1 8 (Bray I method), 12.31 mg P kg -1 ; available nitrogen, 86.64 mg N kg -1 ; available potassium, 1 75.28 mg K kg -1 . There were two P treatments, including low P (no P added) and high P 2 (80kg P 2 O 5 ha -1 added as triple super phosphate to the top10 cm of soil by band application).
3 80 kg N ha -1 as urea and 80 kg K kg -1 as KCl were used to supply N and K for all plants. The 4 soil samples were separately taken from 3 layers (0-20cm, 20-40cm and 40-60cm) along the 5 soil profile before planting and their pH values, exchangeable acidity, and exchangeable P, 6 and Al 3+ content were measured. The trial was conducted using a randomized block design 7 with six replicates for each treatment. Sixty days after sowing, soybean plants were harvested 8 and total root length and dry weight were determined (Zhao et al., 2004). After being 9 carefully washed, roots were scanned and the total root length was quantified with computer 1 0 image analysis software (WinRhizo Pro, Régent Instruments, Québec, Canada).

1
For the hydroponics studies, soybean seeds were sterilized with 10 % (v/v) bleach for 1 1 2 min, and germinated in paper rolls in a beaker at 25℃. To maintain the moisture of the effects of Al toxicity on GmALMT1 expression, roots of HN89 were exposed to 0, 50 and 2 0 100µM AlCl 3 for 6 h. The first 2 cm root segments from root tips were harvested for 2 1 qRT-PCR analysis. For assaying temporal expression patterns of GmALMT1 as related to low 2 2 pH and Al toxicity, seedlings of HN89 were treated with low pH (4.3) for 12 h. For the 2 3 non-Al-treated samples, the first 2 cm root segments from the root tips were harvested at 0, 6 2 4 and 12 h. For the Al treatment samples, after 12 h low pH acclimation, the seedlings were 2 5 transferred to 50 µM AlCl 3 . Root tips were harvested at 2, 4, 6, 8, 10, 12, and 24 h of Al 2 6 treatment.

4
After 6 h of treatments, the first 2 cm tap root tips were harvested for quantifying GmALMT1 5 expression. Relative root length was measured using rulers and determined by the percentage 6 of relative net root growth as described before, using tap root growth under the treatments 7 including -Al/+P/pH5.8 and -Al/-P/pH5.8 as control (Liu et al., 2009). In order to minimize were directly transferred to 4.3 mM CaCl 2 solution including the two P levels (0 and 320 µM 1 2 KH 2 PO 4 ), two pH levels (4.3 and 5.8), and two Al levels (0 and 38 µM Al 3+ activity) as 1 3 mentioned above. All the hydroponic experiments were aerated according to Liao et al (2006).

4
All the treatments had four replicates.

6
To quantify malate exudation from soybean roots, seedlings of both soybean genotypes 1 7 were subjected to the specific treatments. For all the treatments of Al toxicity, the pH value of replicates, with each replicate containing four plants.

2
Secreted malate was processed and measured as described before (Liao et al., 2006).

3
Briefly, the root exudates were passed through a silver cartridge (OnGuardII Ag; Dionex) to 2 4 remove excess Cl -. The pass-through products were then mixed with cation-exchange resin 2 5 (100:1, v/w) for 10 min to remove cations and make the organic acids fully quantifiable.

6
Following centrifugation, the supernatants were used for malate determination. In order to 2 7 determine malate content in root tips, root tips were extracted using 18Ω water. After 2 8 centrifuging at 12,000 rpm for 10 min at 4ºC, the supernatants were collected and used for and incubated in the dark at room temperature overnight. Plasma membrane staining was 2 3 to 0.5 mMCaCl 2 -agar plates containing 0, 400 µM AlCl 3 , respectively. After 2 d of 1 treatments, plant roots were scanned, and primary root length was measured using the Image 2 J program. The relative root growth was calculated as described (Liu et al., 2009). Each 3 treatment for different lines had four replicates, which each replicate contained four plants. The transgenic soybean composite plants with wild type shoots and transgenic roots 1 5 were developed according to Kereszt et al (2007). Soybean hairy roots transformed with 1 6 empty vectors were used as control. The hairy roots from three weeks after transformation 1 7 were used for further analysis. RNA was extracted from hairy roots and qRT-PCR was used independent lines of each construct were used for the determination of malate exudation.

0
Composite plants were exposed to4.3 mM CaCl 2 (pH 4.3) for 6h and root exudates were 2 1 collected and processed for the malate measurement. To detect the function of GmALMT1 as 2 2 related to Al detoxification, soybean hairy roots were subjected to 0.5mM CaCl 2 , pH 4.3, 2 3 containing 50 µM Al. After 3 h incubation, the hairy roots were then stained with 2 4 hematoxylin as described by Delhaize et al (1993a). Briefly, the hairy root tips were first 2 5 washed for 30 min with distilled water, then stained with hematoxylin for 30 min. After staining, the root tips were washed for a further 30 min in distilled water, and then 2 7 photographed.