Introduction

Soybean (Glycine max (L.) Merr.) is a major source of plant protein and oil worldwide. In many regions, however, soybean is often subjected to water-deficit that limits yield. To increase yields, it is essential to develop genotypes that can sustain growth and yield when subjected to water-deficit. One putative trait for enhancing soybean performance under water-limited conditions is the imposition of a maximum-transpiration rate. This trait results in restricted transpiration rate under high vapor pressure conditions so that water is conserved in the soil for use during drought periods later in the season. Sinclair et al. (2010) simulated the impact of such a trait on soybean yields across the USA and found that there was an 85% probability of a yield increase in most locations.

Fletcher et al. (2007) and Sinclair et al. (2008) demonstrated that the soybean genotype PI 416937 actually exhibited the maximum-transpiration trait. Once vapor pressure deficit was 2.1 kPa and greater, this genotype exhibited no further increase in transpiration rate. The limitation on transpiration and potential conservation of soil water was consistent with the original identification of PI 416937 in the field as a “slow-wilting” genotype (Sloane et al. 1990) in which wilting was delayed several days as compared to other genotypes. King et al. (2009) found under field conditions that slow-wilting by PI 416937 in the field was mainly related to conservation of soil water. PI 416937 is a plant introduction from Japan with unknown parentage (Pantalone et al. 1999).

Sinclair et al. (2008) demonstrated that the maximum-transpiration trait in PI 416937 was associated with a low hydraulic conductance in the leaves in the water pathway from the xylem into the stomata guard cells. They speculated that the basis for the differences in hydraulic conductance was a result of differing populations of water-transporting proteins, i.e., aquaporins. Sadok and Sinclair (2010a) subsequently fed to shoots of PI 416937 and N01-11136, which exhibited no limitation on maximum-transpiration rate, three aquaporin inhibitors (cycloheximide, mercury, and silver). Cycloheximide and mercury both resulted in substantially decreased transpiration rates for both genotypes. However, the silver inhibitor resulted in substantial decreases in transpiration rate in N01-11136, but not in PI 416937. The insensitivity of transpiration rate of PI 416937 to silver was confirmed and extended to two genotypes that were progeny of PI 416937, which also expressed the maximum-transpiration trait (Sadok and Sinclair 2010b).

Sadok and Sinclair (2010b) suggested that the three genotypes failing to exhibit a decrease in transpiration rate when fed silver had a different aquaporin population than other genotypes resulting in lower hydraulic conductance. The low hydraulic conductance in turn, was associated with expression of the maximum-transpiration trait. If there is a close correspondence between the lack of response to silver inhibitor and hydraulic conductance, there may be the possibility of genetically characterizing soybean genotypes based on their transpiration response to silver. A further possibility is that a quantitative trait locus (QTL) analysis of the silver response could help to further characterize differences in the aquaporin population in soybean.

The objectives of this study was to (i) characterize a recombinant inbred population (RIL) of soybean derived from PI 416937 for response to the silver aquaporin inhibitor, and (ii) to undertake a QTL analysis for the population using the silver phenotype results.

Materials and methods

Plant material

A RIL population was developed from the mating between cultivar Benning (PI 595645) and PI 416937 (Abdel-Haleem et al. 2011). Benning is a maturity group VII cultivar developed at the University of Georgia from a mating of cultivars Hutcheson and Coker 6738 (Boerma et al. 1997). PI 416937 is a Japanese landrace that phenotypically appears distinctly different from the ancestors of North American soybean cultivars. When compared with elite USA cultivars it has a slower-wilting leaf canopy and a more fibrous root system (Pantalone et al. 1996; King et al. 2009). In a previous study with the silver aquaporin inhibitor (Sadok and Sinclair 2010b), Benning was found to exhibit silver sensitivity, i.e., decrease in transpiration rate. One hundred forty-seven genotypes from this RIL population were selected for this silver inhibition study.

Plant response to silver

Plants for the silver test were grown in a greenhouse at North Carolina State University in Raleigh, NC (35′46°N; 78′39°W). Greenhouse air temperature was maintained at 28 ± 2°C. Seven seeds were sown in 6-L pots filled with a soil mixture of 2/3 Redi-Earth Plug & Seedling (Sun Gro Horticulture, Belleview, WA) and 1/3 pea gravel. After seedling emergence, each pot was thinned to four plants per pot. Pots were maintained in a well-watered condition by watering once a day. After approximately 4 week when the plants had reached vegetative stage V3 to V4 (three to four trifoliolate leaves), the plants were harvested for determining their response to the silver inhibitor.

On the afternoon before a silver test, four replicate plants per genotype were gently removed from the soil and de-rooted. De-rooting was done by cutting the base of the plant stem underwater. Immediately after cutting, the cut stem of the shoot was placed in a 125-ml Erlenmeyer flask containing deionized water and placed in a dark room overnight (14–16 h at 20°C). The following morning, the shoots were moved to a walk-in growth chamber with temperature of 31 ± 1.0°C and relative humidity 40 ± 2% held constant throughout the measurements of plant transpiration. Therefore, the vapor pressure deficit in the growth chamber was approximately 2.3 kPa. (A stable vapor pressure deficit at a higher level comparable to the studies of Sadok and Sinclair (2010a, b) could not be maintained in the walk-in growth chamber.) The photosynthetic photon flux density was 550 to 600 μmol m−2 s−1 at the top of the plants.

Once in the growth chamber, the de-rooted shoots were initially transferred to a new set of 125-ml flasks containing fresh deionized water. Laboratory film (Parafilm “M”, Pechiney Plastic Packaging, IL) was used to seal the stems in the flasks to avoid direct water evaporation from the flasks. The shoots were allowed to acclimatize to the growth chamber environment for 60 min when water loss rates for each shoot had reached a constant value. At this time, the flasks plus shoots were weighed. After another 40 min, the beakers were reweighed and the difference between the two weights, divided by the time interval was used to calculate initial transpiration rate in water (\({\rm{TR}}_{{{\rm{H}}_{{\rm{2}}} {\rm{O}}}}\)).

Following the second weighing of the de-rooted shoots in water, the individual shoots were quickly transferred to a dark-brown, 30-ml glass bottle containing a 200 μM AgNO3 solution. The AgNO3 solution was freshly prepared on the day before each experiment. The 200 μM AgNO3 concentration was selected because this concentration was the lowest concentration resulting in maximum-transpiration decrease (Sadok and Sinclair 2010a). The de-rooted shoots were allowed to take up the AgNO3 solution for 180 min by which time transpiration rate of the shoots had again reached a constant value (Sadok and Sinclair 2010a). After the 180 min period, all bottles were weighed to get an initial weight for the transpiration measurement following exposure to silver. After 40 min, each bottle was again weighed. The transpiration rate following exposure to silver (\({\rm{TR}}_{{{\rm{AgNO}}_{{\rm{3}}}}}\)) was calculated based on the difference between these two weights. Finally, the leaf area of each plant was measured using an area meter (Leaf meter, model LI-3100, Licor, Lincoln, NE) and transpiration rates were expressed per unit leaf area.

Experimental design

Due to limited greenhouse space for growing the plants and the logistics of weighing a large number of bottles, the silver tests had to be done in groups. Usually 48 to 80 shoots were tested in a group allowing four replicates of 12 to 20 genotypes to be included in each test. This meant that the original tests of all 147 RIL were split among 11 groups and tested during the period from September 2009 to April 2010. Even though the two parents were included in each group as references, it was concluded that the results obtained from plants grown over a wide range of growth conditions in the greenhouse were not suitable for direct QTL analysis. The results of the original tests were used to identify RILs at the extremes in sensitivity to silver for a second series of tests of silver response.

The second set of tests used 24 RILs at each extreme in the sensitivity to silver from the initial tests. The tests of the 48 extreme RILs were performed in four groups over the period from June 2010 to October 2010. Each group included six RILs previously found to have low sensitivity to silver, six RILs with high sensitivity to silver, and the two parental lines, PI 416937 and Benning. The number of replicates per genotype was increased to six in this second set to provide additional precision. Therefore, 84 plant shoots were measured in each of the four sets of tests.

Data analysis

The ratio for each plant between the transpiration rate after being fed silver (\({\rm{TR}}_{{{\rm{AgNO}}_{{\rm{3}}}}}\)) and its initial \({\rm{TR}}_{{{\rm{H}}_{{\rm{2}}} {\rm{O}}}}\) was calculated. The ratios were normalized (NTR) within each test to allow comparison of results among tests. In the initial 11 tests, all ratios were normalized within each test relative to the ratio of PI 416937. In the second set of tests, there were cases where the ratio of PI 416937 was not the absolute highest so the normalized ratio was calculated using the ratio of the genotype that expressed the highest ratio, i.e., the least sensitivity to the silver treatment. Therefore, in the second set of tests by definition the highest normalized ratio in each test was 1.0

The normalized data from the second set of tests were analyzed using SAS PROC GLM as randomized complete block experimental designs with replicates, groups and RILs considered as random effects (Statistical Analysis System, SAS Institute, 2001).

QTL analysis

Significant (P = 0.05) associations of the markers with the response in transpiration rate to silver were tested using single-factor analysis of variance (SF-ANOVA) of SAS (SAS Institute 2001). Based on RIL means, QTL analysis was performed using QTL-Cartographer 2.5 (Wang et al. 2007) using the multiple interval mapping (MIM) procedure of Kao et al. (1999) and Zeng et al. (1999). The MIM is a stepwise model adaptation procedure combined with an initial model selection of markers. The MIM was performed to test the presence of significant QTLs and QTL effects including additive and epistatic effects by simultaneous analysis of the QTLs in multiple regression models (Zeng et al. 1999). The initial model included representatives of significant markers that associated with the variation in the transpiration response to silver using SF-ANOVA. To select representatives, the significant markers were sorted by chromosomes and position within each chromosome, then one marker explaining the largest phenotypic variation from each significant marker cluster was chosen as representative of this chromosomal region. A composite interval mapping method (CIM) was employed to detect QTL and estimate the magnitude of their effects (Jansen and Stam 1994). A series of 1,000 per mutations was run to determine the experiment-wise significant level at P = 0.01 of LOD (Churchill and Doerge 1994). The genome was scanned at 2 cM intervals and the window size was set at 10 cM. The pre-selected MIM model was optimized using the “optimize QTL positions” option. The whole genome was re-scanned searching for new main effect QTL and epistatic effect between main QTL using “search for new QTL” and “QTL interaction” options, respectively. The model was then re-evaluated to fit all the significant main and epistatic QTLs in MIM model using “test existing QTLs” option, and finally, the MIM model with minimum Bayesian information criterion (BIC) was chosen, where c(n) = ln(n), and search walk speed was 1 cM. After identifying the best model, the five QTL effects, their total phenotypic variation, and the proportion of the variation explained by each QTL of the model was estimated using the “summary” option.

Results

The silver nitrate-feeding test resulted in a wide range in the response of transpiration rate among genotypes. In the initial 11 groups that included all RILs, NTR for all genotypes ranged from 0.45 to 1.45 (Fig. 1). The combined data from the first set of 11 groups results in an approximately normal frequency distribution in NTR centered on the mean value of 0.86. The average NTR value of 0.80 for Benning was significantly (P = 0.05) lower than the defined value of 1.0 for PI 416937. The RIL population was highly variable and transgressive segregation for the NTR value greater than PI 416937 and less than Benning was indicated (Fig. 1).

Fig. 1
figure 1

Frequencies in genotype number phenotyped for normalized transpiration ratio (NTR) following a silver nitrate treatment. All 147 RILs of PI 416937 × Benning included in the first set of tests are included

Full size image

The 24 RILs with the lowest NTR and 24 RILs with the highest NTR were selected for the second set of tests. Not surprisingly, the second set of tests with the extreme RILs selected on data from the first set of tests resulted in a wide range in NTR values. In this case, the NTR values for the RILs ranged from 0.68 to 1.0 with a mean across all these RILs of 0.87. The NTR value of Benning was 0.78, which was significantly (P = 0.01) less that the NTR value of PI 461937. No epistatic interactions were detected.

A linkage map with 276 informative SSR markers (Abdel-Haleem et al. 2011) was used to identify QTL controlling the variation in response to RILs fed silver nitrate. Seven SSR markers on seven soybean chromosomes were identified as being associated with the response of transpiration to silver nitrate based on SFA-NOVA (Table 1). The markers identifying these QTL were located on Gm01, Gm03, Gm04, Gm08, Gm12, Gm13, and Gm20. These QTL accounted for 11–20% of the phenotypic variation in transpiration response to silver nitrate. The PI416937 alleles at markers Satt339 (Gm03), Satt177 (Gm08), Satt581 (Gm10), and Satt419 (Gm20) were associated with the insensitivity of plants to silver nitrate, while Benning alleles at Satt294 (Gm04), Sat_175 (Gm12), and Satt649 (Gm13) enhanced silver nitrate insensitivity.

Table 1 SSR markers associated with the enhanced insensitivity to silver nitrate based on single-factor analysis of variance (SF-ANOVA) for the Benning × PI416937 RIL population
Full size table

Data on each RIL’s transpiration response to silver and its alleles at each SSR marker were then analyzed by MIM model of QTL-Cartographer. The MIM method utilizes multiple intervals marker simultaneously to identify QTLs. This method fits all QTLs into a model and has the capacity for the evaluation of QTL epistasis. The MIM model of transpiration response to silver consists of four QTL on Gm03, Gm05, Gm10, and Gm12 with LOD scores ranging from 2.9 to 5.6 (LOD threshold = 2.7) (Table 2; Fig. 2). The MIM model explained 87.5% of the variation of this trait. The qSV_Gm03, (flanked with Satt339–Sat_091) explained 16.4% of the phenotypic variation in silver sensitivity and qSV_Gm10 (flanked with Satt478–Satt581) explained 19.4% of the variation. Both of these QTL inherited their positive alleles from PI 416937 while qSV_Gm05 and qSV_Gm12 inherited their favorable alleles from Benning (Table 2). The qSV_Gm05 and qSV_Gm12 QTL explained 17.7 and 24.7% of the phenotypic variation, respectively.

Table 2 Multiple intervals mapping for QTLs associated with the response to silver nitrate in the Benning × PI 416937 RIL population
Full size table
Fig. 2
figure 2

Interval mapping of the silver response QTL in the Benning × PI 416937 RIL population. QTL nomenclature is in the form of Q-trait-chromosome-position. The length of QTL bar indicates the LOD confidence interval and the QTL line is extended to LOD confidence interval from the maximum likelihood value

Full size image

The genomic locations of the QTL that enhanced the insensitivity of plants to silver nitrate were aligned on the soybean linkage/QTL maps in Soybase (http://www.soybase.org/) by use of the SSR markers defining the genomic locations. For this analysis, QTL conditioning other traits were included if these trait QTLs were mapped within a bin that included the silver response QTL interval ±10 cM beyond the two markers that defined the silver response QTL boundary (Fig. 3). At the genomic location of qSV_Gm03 there was one QTL for seed yield (Sdyld15-3), a salt tolerance QTL (Salt1-1), two protein-related QTL (Beta_conglycinin_1-1 and Prot_21-9), an isoflavone component QTL (Glycitein 2-9) and a QTL for fibrous roots under drought stress (Q_root_Gm03). Near to qSV_Gm05 QTL, nine QTL related to seed yield (Sdyld 15-10, Pod mat 7-2, Sd abort 1-6 and Sdwt 6-4), seed quality (Oil 3-3, Oil 4-1, Prot 2-3 and Genistein 2-4), and root growth under iron deficiency (Feeffic 5-1) were located. Eleven QTL were located in the same bin as qSV_Gm12 and controlled different seed quality and seed yield traits, while at qSV_Gm10, two QTL for plant height (PlHt 17-8) and protein content (Prot 12-5) were co-localized (Fig. 3) .

Fig. 3
figure 3

Comparative QTL analysis of the response to silver nitrate QTL detected in the Benning × PI416937 population with coinciding QTL found on Soybase (http://www.soybase.org/). These QTLs were identified by alignment of both maps with common SSR markers

Full size image

Discussion

Testing genotypes for differences in response to the AQP-inhibitor silver resulted in a range of responses in transpiration rate. The results of these tests confirmed the original observations of Sadok and Sinclair (2010b) that the parents PI 416937 and Benning of the RIL population responded differently to the silver assay. In the initial tests, there were individual RIL lines that had silver responses that were outside that exhibited by the two parents. However, there may have been bias in the results of the first series of tests that they were obtained in 11 individual tests spread over a wide range of environmental growth conditions.

The second set of tests required only four trials over a shorter interval of time. The change in transpiration rate in response to the silver treatment was least, or nearly so, for PI 416937 among the tested genotypes. The NTR value of Benning was significantly lower than PI 416937. However, some RILs proved more sensitive than Benning to the exposure to silver. The results from the second series of measurements were used in the QTL analysis.

Using the MIM model, four QTL were identified as being associated with the silver nitrate response (Table 2; Fig. 2). The alleles for silver insensitivity at two QTL were inherited from PI 461937 and from Benning at the other two QTL. The marker Satt339 on Gm03 explained 16% of the phenotypic variation in the sensitivity of plants to silver nitrate (Table 2). The positive allele at this QTL was inherited from PI 416937.

The QTL for the silver response in the marker interval of Satt339 to Sat_091 is in the same interval as the QTL for fibrous root trait of PI 416937 under drought stress. This fibrous root QTL was discovered in the same RIL population (Abdel-Haleem et al. 2011), and the allele for more fibrous rooting was contributed by PI 416937. The phenotypic variation explained in fibrous root growth at this QTL was 13.5%, which was the highest among the five fibrous root QTL that were identified (Abdel-Haleem et al. 2011). The co-localization of the fibrous root trait with a putative AQP may help to explain its association with the slow-wilting phenotype, even though there is little evidence that fibrous roots would result in the slow-wilting phenotype of PI 461937 (King et al. 2009).

Another interesting co-localization in the marker interval of Satt339 to Sat_091 is the salt tolerant marker reported by Lee et al. (2004). One possibility is the involvement of an aquaporin in the transport of salt resulting in tolerance. Clearly, a search in this marker interval is warranted to track the basis for the silver response and the slow-wilting phenotype.

QTL conditioning the sensitivity of plants to silver nitrate were located in the same chromosomal regions as isoflavones QTL on Gm03, Gm05, and Gm12 (Fig. 3). Soybean plants grown under well-watered conditions tend to produce more isoflavones compare to plants under dry conditions throughout all seed developing stages (Al Tawaha et al. 2007; Bennett et al. 2004; Seguin et al. 2004). Under severe drought stress, production of seed isoflavones were altered (Gutierrez-Gonzalez et al. 2010). Co-localization of QTL for both isoflavones and the canopy wilting trait could explain the role these traits play in regulating soybean response to drought stress.

The exotic line, PI 416937 has positive alleles for drought tolerance, which can be introduced into elite soybean germplasms to improve soybean productivity under stress conditions. The present study identified two silver nitrate response QTL that inherited their positive alleles from PI 4169367. Co-localization of the silver nitrate response QTL with other morpho-physiological traits could provide the potential to enhance the soybean plant’s ability to tolerate moisture stress. Marker assisted backcrossing by introgression of these favorable alleles from PI 416937 into an elite cultivar could increase the plant ability to produce stable seed yield under drought stress conditions.