Logo of plntphysLink to Publisher's site
Plant Physiol. Aug 2010; 153(4): 1706–1715.
Published online Jun 25, 2010. doi:  10.1104/pp.110.158832
PMCID: PMC2923888

Boron Toxicity Tolerance in Barley through Reduced Expression of the Multifunctional Aquaporin HvNIP2;11,[W]


Boron (B) toxicity is a significant limitation to cereal crop production in a number of regions worldwide. Here we describe the cloning of a gene from barley (Hordeum vulgare), underlying the chromosome 6H B toxicity tolerance quantitative trait locus. It is the second B toxicity tolerance gene identified in barley. Previously, we identified the gene Bot1 that functions as an efflux transporter in B toxicity-tolerant barley to move B out of the plant. The gene identified in this work encodes HvNIP2;1, an aquaporin from the nodulin-26-like intrinsic protein (NIP) subfamily that was recently described as a silicon influx transporter in barley and rice (Oryza sativa). Here we show that a rice mutant for this gene also shows reduced B accumulation in leaf blades compared to wild type and that the mutant protein alters growth of yeast (Saccharomyces cerevisiae) under high B. HvNIP2;1 facilitates significant transport of B when expressed in Xenopus oocytes compared to controls and to another NIP (NOD26), and also in yeast plasma membranes that appear to have relatively high B permeability. We propose that tolerance to high soil B is mediated by reduced expression of HvNIP2;1 to limit B uptake, as well as by increased expression of Bot1 to remove B from roots and sensitive tissues. Together with Bot1, the multifunctional aquaporin HvNIP2;1 is an important determinant of B toxicity tolerance in barley.

Barley (Hordeum vulgare) is an annual grain, ranked fourth among the cereals in terms of total world production (FAOSTAT; http://faostat.fao.org). Cultivated barleys are not well adapted to high soil boron (B), although toxicity to B has been known for some time (Christensen, 1934; Cartwright et al., 1986). B is an essential micronutrient for higher plants (Sommer and Lipman, 1929), and there is only a very small difference between deficient and toxic concentrations of B in plant tissues (Eaton, 1944). Early genetic studies of B toxicity tolerance in a barley mapping population identified favorable alleles at four genetic loci on chromosomes 2H, 3H, 4H, and 6H (Jefferies et al., 1999). Tolerance to B toxicity is associated primarily with reduced B accumulation, but also with increased root growth and dry matter production and reduced leaf symptom expression. In all cases, the favorable alleles are derived from the highly B toxicity-tolerant Algerian landrace barley Sahara 3771 (Sahara). We recently identified a gene (Bot1) underlying the tolerance quantitative trait locus (QTL) on chromosome 4H of barley, that is a putative integral trans-membrane B transporter with similarity to bicarbonate transporters in animals (Sutton et al., 2007). In roots of Sahara, active efflux of B and high Bot1 expression at least partially control lower B accumulation in roots and shoots of barley plants (Hayes and Reid, 2004; Reid, 2007; Sutton et al., 2007).

A second class of proteins capable of B transport belongs to the superfamily of major intrinsic proteins (MIPs) or aquaporins. Aquaporins have previously been suggested to be involved in B transport in higher plants (Dordas et al., 2000; Fitzpatrick and Reid, 2009) and are required for normal growth of Arabidopsis (Arabidopsis thaliana) under B deficiency conditions (Takano et al., 2006; Tanaka et al., 2008). However, aquaporins have not been shown to be involved in B toxicity tolerance. Five major subgroups of aquaporins in plants are now recognized (Danielson and Johanson, 2008; Sade et al., 2009; Shelden et al., 2009). In particular, members of one of these subgroups, the nodulin-26-like intrinsic proteins (NIPs), were recently described as functional equivalents of aquaglyceroporins (Bhattacharjee et al., 2008). These channel proteins usually enable water transport but recent studies have demonstrated that they also facilitate transport of other small, uncharged solutes under physiological conditions including glycerol and lactic, boric (B), silicic (Si), arsenious (As), antimonious (Sb), and germanic (Ge) acids (Ma et al., 2006, 2008; Takano et al., 2006; Choi and Roberts, 2007; Bienert et al., 2008; Chiba et al., 2008; Isayenkov and Maathuis, 2008; Tanaka et al., 2008; Kamiya et al., 2009; Mitani et al., 2009).

Here we report that HvNIP2;1, a member of the NIP subfamily, underlies a B toxicity tolerance QTL on chromosome 6H of barley associated with reduced B uptake (Jefferies et al., 1999). This gene was recently identified and described in barley as the Si influx transporter HvLsi1, largely on the basis of the observed characteristics of the putative orthologous protein OsNIP2;1 (OsLsi1) in rice (Oryza sativa; Chiba et al., 2008). HvLsi1 showed transport activity for Si when expressed heterologously and immunohistological staining in barley revealed that the protein is expressed on the distal side of epidermal and cortical cells in the basal part of seminal roots, embedded in the plasma membrane (Chiba et al., 2008). However, expression was poorly correlated with Si uptake in barley (Chiba et al., 2008). In this work, our molecular models suggested that HvNIP2;1 (HvLsi1) could facilitate the transport of a range of small, uncharged solutes including Si and also B. We expressed HvNIP2;1 protein in yeast (Saccharomyces cerevisiae) and Xenopus oocytes and functionally characterized its B transport properties. Furthermore, we observed lower shoot B accumulation in a rice mutant displaying a point mutation Ala-132Thr in OsNIP2;1 (lsi1; Ma et al., 2006; Mitani et al., 2008). Higher tolerance to B in Sahara appears to be mediated through lower transcript levels of HvNIP2;1 in basal root segments of barley, possibly owing to a repeat insertion in the promoter region of HvNIP2;1, approximately 2 kb upstream of the start codon. Based on these results we propose that reduced expression of HvNIP2;1 lowers the passive influx of B into roots under high soil B concentration, contributing to lower shoot B accumulation and thus, higher B tolerance in Sahara.


Genetic Mapping of HvNIP2;1

Previous QTL mapping in a barley doubled haploid population identified a B toxicity tolerance locus on chromosome 6H, with the favorable allele derived from the B toxicity-tolerant landrace Sahara (Jefferies et al., 1999). Prior to F3 population development for fine mapping of the 6H locus gene, we assessed the effect of alternative chromosome 4H-Bot1 alleles on leaf blade B concentration, determined by segregation of 6H gene alleles (Fig. 1A). In both the 4H Clipper (Bot1:C) and 4H Sahara (Bot1:S) allele backgrounds, the 6H Sahara allele contributed positively to reduced leaf blade B concentration. In the presence of 4H-Bot1:C, plants carrying the 6H Sahara allele (HvNIP2;1:S) accumulated 25% less B in leaf blades than those carrying the 6H Clipper allele (HvNIP2;1:C). Similarly, in the presence of 4H-Bot1:S, plants carrying the 6H Sahara allele (HvNIP2;1:S) accumulated 40% less B in leaf blades than those carrying the 6H Clipper allele (HvNIP2;1:C). Furthermore, this analysis also demonstrated that a mapping population fixed for the 4H Sahara allele (Bot1:S) provided the phenotypic discrimination (i.e. largely nonoverlapping distribution) required for accurate mapping of the 6H QTL (Fig. 1A). We subsequently developed an F3 mapping population consisting of 192 plants derived from a cross between two Clipper × Sahara doubled haploid lines that differed for alleles at the B tolerance locus on chromosome 6H, but not for alleles at the other known B toxicity tolerance loci (Jefferies et al., 1999). The population was fixed for the Sahara allele of the chromosome 4H tolerance gene Bot1.

Figure 1.
HvNIP2;1 genetic analysis. A, Phenotypic effect of chromosome 4H-Bot1 and chromosome 6H-HvNIP2;1 Clipper and Sahara allele combinations on leaf blade B concentration in barley. Individual F2 plants derived from a cross of two Clipper × Sahara ...

Gene-derived PCR markers were developed using synteny to rice and were mapped in this F3 population. We were able to narrow the 6H QTL interval by characterizing F4 progeny individuals of F3 lines recombinant in the region, both for molecular markers and leaf blade B accumulation in hydroponics (Supplemental Fig. S1). This analysis positioned the 6H tolerance locus between markers xCSM-1 and xRAT-3 (Fig. 1B). The corresponding interval in rice was approximately 474 kb, between rice genes Os02g0740300 at position 30,977,102 bp and Os02g0748100 at position 31,451,191 bp in chromosome 2 (GenBank accession no. NC_008395). We considered one of the rice genes in the mapped syntenic rice interval (OsNIP2;1; GenBank accession no. AB222272) to be a strong candidate based on its known transport properties (Ma et al., 2006). We isolated the Clipper and Sahara barley orthologs of OsNIP2;1 and developed the cleaved amplified polymorphic sequence marker xHvNIP2;1 from this sequence, which cosegregated with B toxicity tolerance in our F3 mapping population. This confirmed HvNIP2;1 as a candidate for the B toxicity tolerance gene in barley. While we considered an ortholog of OsNIP2;1 to be the likely candidate tolerance gene in barley, we are unable to completely rule out the possibility of a role of other tolerance cosegregating genes acting either independently or in conjunction with HvNIP2;1 in B tolerance in barley.

Characteristics of HvNIP2;1 in Clipper and Sahara Barley

The open reading frame sequences for HvNIP2;1 from Clipper and Sahara barley contain a single nucleotide polymorphism (C to G), which does not alter the encoded amino acid residue (Val-283; Supplemental Fig. S2). HvNIP2;1 shares 82% sequence identity with OsNIP2;1 at the amino acid level (Supplemental Fig. S2; Chiba et al., 2008).

Leaf blade and root northern hybridization analyses in Clipper and Sahara detected expression of HvNIP2;1 in roots only, which was unresponsive to an increase in B concentration (Fig. 2A). Higher transcript levels of HvNIP2;1 were evident in Clipper roots compared to Sahara. We also observed significant varietal transcript differences when we analyzed cDNAs from root sections using quantitative (Q)-PCR (Fig. 2B). Transcript levels of HvNIP2;1 increased for both lines along the length of the root from the apex toward the basal root regions, with up to 15-fold higher transcript levels observed in roots of the B-sensitive barley variety Clipper (Fig. 2B).

Figure 2.
Expression of HvNIP2;1 mRNA in the B-tolerant variety Sahara and B-intolerant variety Clipper. A, HvNIP2;1 mRNA levels in whole roots (lanes 2–11) and leaf blades (lanes 12–21) of Clipper and Sahara barley grown in hydroponic solution ...

We determined the nucleic acid sequences of HvNIP2;1 from Clipper and Sahara barley genotypes upstream of the ATG start codon and compared them using the Web-based promoter analysis tool PlantPAN (Chang et al., 2008). The first 1,377 bp upstream of the ATG start codon are identical between Clipper and Sahara. Further upstream we found six single nucleotide polymorphisms between 1,378 and 2,128 bp in the putative promoter region. Beyond this region the two sequences became dissimilar and we found an insertion in Sahara relative to Clipper that contained a repeat sequence beginning at position 2,311 bp upstream from the start ATG codon. We also found a non-long terminal repeat transposon classified as an EMMA repeat, inserted in the promoter region of Sahara, between 3,097 and 3,151 bp, upstream of the start ATG codon (Supplemental Fig. S3). Further sequencing upstream of the currently obtained promoter regions in Clipper and Sahara should reveal the extent and significance of these changes in alleles of HvNIP2;1.

Molecular Modeling of HvNIP2;1

To investigate structural determinants in HvNIP2;1, and how transport of B and other metalloids is mediated across a membrane, molecular modeling studies were performed. An archeal aquaporin from Methanobacterium thermoautotrophicum (AqpM; Lee et al., 2005) was chosen as a structural template for modeling, since it showed the highest amino acid sequence identity/similarity (36%/72%) with HvNIP2;1. Evaluation of molecular models (parameters are specified in “Materials and Methods”) indicated that AqpM was a suitable template for modeling and that the constructed models were of satisfactory quality. The molecular model of HvNIP2;1 revealed up-down bundle protein folding, housing six membrane helices and two reentrant helices (Fig. 3). The latter structural elements generated an additional membrane span, resembling an hourglass arrangement (Engel et al., 2000; Borgnia and Agre, 2001; Lee et al., 2005; Krissinel and Henrick, 2007). Analysis of the buriedness of the pore of HvNIP2;1 showed that the hourglass-shaped channel spanned the entire length of the membrane and that buriedness increased progressively toward the center of the pore. Calculations indicated that the pore volume of AqpM was approximately 1,113 cubic Å, while that of HvNIP2;1 varied between 1,149 and 1,392 cubic Å. The likely homotetrameric organization of HvNIP2;1 (Krissinel and Henrick, 2007) could be simulated, and it is expected that this higher structural organization could permeate metalloids and solutes more effectively (Engel et al., 2000; Borgnia and Agre, 2001; Krissinel and Henrick, 2007).

Figure 3.
Cross-sectional views of 3D molecular models of HvNIP2;1. A, The clip shows an up-down bundle fold (rainbow colors) of HvNIP2;1. Accessibility of B acid (atomic colors, arrow) and four water molecules (dark-blue spheres) are illustrated in the central ...

Molecular models of HvNIP2;1 allowed the disposition of interacting amino acid residues to be defined in the pore with respect to various metalloids. The positions of two highly conserved NPA (Asn-Pro-Ala) motifs in HvNIP2;1 and AqpM were almost identical. Superposition of residues in the channels of both porins indicated that four of the six key residues in the constriction of AqpM were also conserved in HvNIP2;1. One B, As, Si, or Ge acid with four water molecules were modeled in the pore of HvNIP2;1 (Fig. 3). All solutes were positioned favorably in the channel and did not generate steric clashes with neighboring residues. The models showed that Asn side chains in the two NPA motifs were critical for mediating interactions of all four solutes near the hourglass constriction. Water molecules in the HvNIP2;1 pore could form hydrogen bonds of 3.2 to 3.5 Å with all metalloids. The aromatic/Arg (ar/R) selectivity filter in aquaporins is considered to be a major determinant for solute permeability (Beitz et al., 2006). We found that while the ar/R filter in AqpM was up to 6.5 Å wide, it was up to 8 Å wide in HvNIP2;1, primarily due to the substitution of a bulky Phe in AqpM with Gly in HvNIP2;1 on the α-helix H2. This substitution could lead to a more open and polar character of the HvNIP2;1 pore.

Similar pore diameter values were also observed in a model of OsNIP2;1 in complex with B. When comparing amino acid residues that line the pore of OsNIP2;1 with those in HvNIP2;1, the only residue that was found to protrude into the OsNIP2;1 pore was Ser-74. At this position in OsNIP2;1, the Gly-Ser-Asp motif is substituted with an Ala-His-Asp motif; these components of the aquaporin sequences form highly variable regions. Thus, while the residue Ser-74 could potentially protrude into the cavity of the pore of OsNIP2;1, the overall transport properties of OsNIP2;1 and HvNIP2;1 are not expected to differ greatly.

Heterologous HvNIP2;1 Expression Studies

Heterologous HvNIP2;1 protein expression increased the sensitivity of yeast to high B concentrations in the media (Fig. 4A; Supplemental Fig. S4). Growth of HvNIP2;1-expressing yeast was inhibited slightly even without addition of B to the media (Supplemental Fig. S4), suggesting either that HvNIP2;1 expression altered the membrane permeability of yeast to a range of solutes affecting growth, or that sufficient B was present in the basal medium to restrict growth. To confirm that the observed growth sensitivity under high and low B was due to expression of HvNIP2;1, Gal was replaced with Glc in minimal media to repress transcription that is under the control of the GAL1 promoter (West et al., 1984). Under these conditions, phenotypic variation was not observed between the HvNIP2;1-transformed and control yeast cells (Supplemental Fig. S5), clearly indicating that the sensitivity of yeast to B grown in the presence of Gal is due to expression of HvNIP2;1. Si added to agar media at soluble concentrations did not affect the growth of yeast and we were therefore unable to determine whether HvNIP2;1 has Si transport activity, as has recently been shown in Xenopus and rice (Chiba et al., 2008). However, Ge is a toxic metalloid, which is often used as a chemical analog for comparatively nontoxic Si in plant studies of Si transport (Nikolic et al., 2007). We found that HvNIP2;1 expression increased the sensitivity of yeast to both Ge (Fig. 4B) and the metalloid As (Fig. 4C).

Figure 4.
Heterologous expression of HvNIP2;1 in yeast. A to C, Growth of yeast expressing HvNIP2;1 in the presence of B (A), Ge (B), and As (C) on Gal minimal media. On each plate are three replicates transformed with empty vector (top), together with three replicates ...

The ability of the HvNIP2;1 protein to facilitate B transport was demonstrated by heterologous expression in both yeast and Xenopus oocytes. In yeast we expressed HvNIP2;1 in the SY1 mutant strain (Ruetz and Gros, 1987; Nakamoto et al., 1991) and used stopped-flow analysis of isolated vesicles to measure permeability to B (Dordas and Brown, 2000). The rate constant of vesicle reswelling induced by an inwardly directed B concentration gradient was significantly faster for vesicles derived from HvNIP2;1-expressing yeast compared to empty vector controls (t test P < 0.05, n = 9 independent yeast cultures; Fig. 4D). Particle size analysis (dynamic light scattering) demonstrated that vesicles from all yeast preparations were of the same diameter (99.7 nm ± 10.9 nm, six independent vesicle preparations), and we could therefore calculate the B permeability (PB = k × r/3, where r is vesicle radius and k is the rate constant for vesicle reswelling) of the two isolates. For HvNIP2;1 vesicles, PB was 1.6 × 10−6 cm s−1, while for empty vector controls PB was 0.98 × 10−6 cm s−1.

In Xenopus, we compared the transport properties of HvNIP2;1 with GmNOD26 from soybean (Glycine max), the archetype of the NIP subfamily of aquaporins with well-characterized water and glycerol transport properties (Rivers et al., 1997). Oocytes were injected with either HvNIP2;1 or GmNOD26 capped RNA (cRNA). A further set of oocytes were injected with water to act as controls. We measured oocyte swelling in response to an inwardly directed B concentration gradient (160 mm) in a solution of the same osmotic pressure as the original bath solution (see “Materials and Methods”). The influx of B down its concentration gradient induces swelling due to osmotically induced water influx. The rate of swelling can be used as a measure of B permeability since water influx is not rate limiting. Water transport in the oocytes was also measured after replacing the incubation solution with a 5-fold diluted solution. Swelling measurements showed that expression of both HvNIP2;1 and GmNOD26 increased the permeability of Xenopus oocytes to water (Fig. 5A), whereas only expression of HvNIP2;1 increased the permeability of the oocytes to B (Fig. 5B). Water-injected oocytes showed limited permeability to water and were not permeable to B (Fig. 5).

Figure 5.
Heterologous expression of HvNIP2;1 in Xenopus oocytes. Increase in relative volume (V/V0) of Xenopus oocytes injected with HvNIP2;1 cRNA, GmNod26 cRNA, or water (negative control), after replacing the incubation solutions with A, a 5-fold diluted solution, ...

OsNIP2;1 Facilitates B Transport

We measured B accumulation in leaf blades of wild-type rice (cv Oochikara) and an OsNIP2;1 (lsi1) rice mutant line with an Ala-132Thr mutation that results in loss of function as determined by measurements of Si transport (Ma et al., 2006). Hydroponically grown leaf blades of OsNIP2;1 mutant rice had significantly lower levels of leaf B compared to wild-type Oochikara rice, both when supplied with low and excessive concentrations of B in the growth solution (Fig. 6A). We also expressed OsNIP2;1 mutant protein in yeast to analyze the effect of the Ala-132Thr mutation on B transport in yeast. While it is known that the Ala-132Thr mutation disrupts Si transport through OsNIP2;1 (Ma et al., 2006), it is unknown if this mutation similarly affects B transport. Yeast expressing Ala-132Thr OsNIP2;1 mutant protein were more sensitive to high concentrations of B than cells with an empty vector (control cells), but grew better than yeast expressing the OsNIP2;1 wild-type protein (Fig. 6B). These data demonstrate that the mutation in OsNIP2;1 partially affects B transport through OsNIP2;1 and support the observed reduction in leaf B content of OsNIP2;1 mutant plants compared to wild-type Oochikara rice grown hydroponically.

Figure 6.
Phenotypic effect of the OsNIP2;1 Ala-132Thr mutation. A, Leaf blade B concentration of wild-type (cv Oochikara) rice plants and OsNIP2;1 (lsi1) mutant plants after 14 d growth in hydroponics solution containing 0 mm B or supplemental B to 1 mm. Data ...


The Algerian barley landrace Sahara carries at least four known loci for B toxicity tolerance, of which two are related to lower leaf blade B concentrations. These QTL are located on chromosomes 4H and 6H (Jefferies et al., 1999). We recently identified the gene underlying the 4H QTL as an efflux-type borate anion transporter. Here we have identified a gene underlying the chromosome 6H QTL, the other locus contributing to the reduced shoot B concentration in barley. We show that the gene, HvNIP2;1, facilitates B transport and that differential expression between Sahara and Clipper barley may contribute to the reduced levels of shoot B in Sahara. HvNIP2;1 is a member of the MIP, or aquaporin superfamily and has recently been described as a Si influx transporter (HvLsi1; Chiba et al., 2008). Interestingly, Sahara barley is also reported to have reduced levels of shoot Si compared to Clipper barley (Nable et al., 1990).

HvNIP2;1 belongs to the NIP subfamily of plant MIPs, members of which have been identified to transport numerous small, uncharged solutes including B (Takano et al., 2006; Tanaka et al., 2008), Si (Ma et al., 2006; Chiba et al., 2008; Mitani et al., 2009), As (Bienert et al., 2008; Ma et al., 2008), Sb (Bienert et al., 2008), and lactic (Choi and Roberts, 2007) acids, as well as glycerol (Wallace and Roberts, 2005) and water. Recent phylogenetic studies have classified plant NIPs into three broad subgroups (Mitani et al., 2008; Rougé and Barre, 2008), based on key amino acids forming the ar/R selectivity filter. The archetype of the NIP subfamily, GmNOD26, is unable to transport B (Fig. 5) and falls into NIP subgroup I, while the Arabidopsis B transport proteins, AtNIP5;1 and AtNIP6;1, cluster in subgroup II and have a wider filter (Rougé and Barre, 2008). HvNIP2;1 and OsNIP2;1 have an even larger, more open pore and are members of a third subgroup. In fact, in a recent modeling study, several NIP subfamily members had the largest ar/R filter among 70 maize (Zea mays) and rice aquaporins (Bansal and Sankararamankrishnan, 2007). Our molecular modeling indicates that the pore formed by HvNIP2;1 when embedded in a lipid bilayer may be up to 8 Å wide at the constriction, and is sufficiently large to allow for the passage of B, As, Si, and Ge acids (Fig. 3).

Furthermore, we functionally characterized the protein and used heterologous protein expression in yeast and in Xenopus oocytes to show that HvNIP2;1 can facilitate the transport of B, Ge, As, and water (Figs. 4 and and5),5), while Si transport has also been demonstrated (Chiba et al., 2008). The B permeability we measured in yeast vesicles not expressing HvNIP2;1 (1.0 × 10−6 cm s−1) is at the higher end of the range of values reported in the literature for plant membranes and lipid vesicles (Dordas and Brown, 2000; Dordas et al., 2000; Stangoulis et al., 2001) and it is likely that a component of this permeability could be attributed to other B transporters in the native yeast membrane (Nozawa et al., 2006) in addition to the lipid permeability that can be high depending on lipid composition (Dordas and Brown, 2000). Despite this intrinsically high permeability, HvNIP2;1 expression afforded a substantial increase in permeability to 1.6 × 10−6 cm s−1. HvNIP2;1 appears to be able to facilitate the transport of a broad range of small, near apolar solutes, and we are currently introducing site-specific mutations into HvNIP2;1 to investigate the relationship between pore size and discrimination between B and the more toxic As and Ge acid metalloids. It has also been reported that the NIP aquaporins function in the acquisition of B under B-limiting conditions (Takano et al., 2006), which raises the question of whether the tolerance allele of HvNIP2;1 may confer a disadvantage to Sahara barley under conditions of low B supply. This hypothesis could be tested in material genetically fixed for other loci in Sahara known to be involved in B uptake (Jefferies et al., 1999). While this may be of low importance in areas of high soil B, such as North Africa where the Sahara landrace originated, it may be of more relevance in the target environments of modern barley breeding programs where localized B deficiency may be more common.

For in planta analyses we measured B accumulation in a rice mutant (lsi1) defective in OsNIP2;1 and found lower levels of shoot B in the mutant relative to wild-type Oochikara rice (Fig. 6A). We also demonstrated that the Ala-132Thr mutation in OsNIP2;1 affects growth of yeast differently to cells expressing the wild-type OsNIP2;1 protein. The growth phenotype we observed in yeast expressing mutant OsNIP2;1 was intermediate between yeast expressing wild-type OsNIP2;1 and the empty vector control, suggesting that the Ala-132Thr mutation did not completely abolish transport.

Northern and Q-PCR analyses of HvNIP2;1 RNA levels in barley leaf blade and root tissues demonstrated that HvNIP2;1 is specifically expressed in roots, and that there are higher levels of RNA in roots of the B-sensitive barley variety, Clipper (Fig. 2). Significantly, we found higher expression in basal root segments, which was particularly evident in Clipper. Our transcript data support the hypotheses that higher HvNIP2;1 expression in Clipper increases the permeability of the roots to B, and that significant expression is only found in mature regions of the root, where apoplastic transport of solutes is inhibited by the development of suberized/lignified Casparian bands (Steudle and Peterson, 1998). Our observations of increased transcript levels of HvNIP2;1 in mature roots is consistent with those of RNA levels of OsNIP2;1 in roots of rice (Yamaji and Ma, 2007), and subsequent observations of barley HvNIP2;1 expression (Chiba et al., 2008). Immunohistological investigations showed that expression of HvNIP2;1 protein in barley was limited to the distal sides of cortical, epidermal, and hypodermal cells, in regions of the root where a suberized layer was present (Chiba et al., 2008).

The observed differences in transcript levels of HvNIP2;1 in Clipper and Sahara barley may be explained by a repeat insertion into the cis-regulatory region, approximately 2 kb upstream of the translational start site (Supplemental Fig. S3). Repeat insertions in noncoding regions of genes have the ability to alter gene expression in plants and animals, and result in subtle or dramatic phenotypic changes. They are considered to be a rich source of functional diversification (Kidwell and Lisch, 1997). Recently, a study of rice genes containing insertions in putative promoter regions found a positive relationship between the distance of the repeat sequence from the start codon and the likelihood of gene expression (Krom et al., 2008). We cannot rule out the possibility that the repeat insertion in the promoter region of Sahara barley might cause altered tissue- or cell-specific expression of HvNIP2;1 in the root. Together these findings indicate that reduced HvNIP2;1 expression should contribute to lower leaf blade B accumulation and greater tolerance of barley to excessive soil B.

B toxicity is a significant agronomic problem in a number of cereal growing regions across the world, including West Asia, Northern Africa, and Southern Australia. The discovery of key genes responsible for B toxicity tolerance in cereals presents opportunities for the development of elite lines adapted for growth in high B soils, through either genetic engineering or conventional breeding. HvNIP2;1 is the second B toxicity tolerance gene to be identified in barley. Tolerance to high B may be achieved by a loss of function of this gene and, therefore, there is significant potential for the development of B-tolerant germplasm derived from identification of individuals with mutations in HvNIP2;1. We are currently undertaking a search for suitable mutants for this purpose.


Plant Growth and Analyses

We used the parent barley (Hordeum vulgare) lines Sahara and Clipper, and lines derived from Clipper × Sahara F1 doubled haploids. The rice (Oryza sativa) cv Oochikara and mutant line lsi1 were obtained from Prof. J.F. Ma (Okayama University, Japan). Seeds of barley or rice were germinated on filter paper and grown hydroponically in a base nutrient solution as described previously (Sutton et al., 2007). The seedlings were grown in controlled conditions (15°C night/23°C day with a 14-h photoperiod for barley, and 24°C night/28°C day with a 16-h photoperiod for rice). Plant material was oven dried and analyzed by inductively coupled plasma optical emission spectrometry.

Genetic Mapping of HvNIP2;1

The F3 mapping population was derived from a cross between two Clipper × Sahara F1-derived doubled haploids, which differed for alleles at the B tolerance locus on 6H but not for alleles at other known B tolerance loci. The population was fixed for the Sahara allele of Bot1 on chromosome 4H. F2 plants selected as heterozygous for the 6H B toxicity tolerance locus region were used to generate F3 seed for recombinant screening. Cleaved amplified polymorphic sequence markers were developed using barley ESTs related to genes from the corresponding interval on rice chromosome 2. Markers are listed in Supplemental Table S1. F3 recombinants for the region were marker selected and their tolerance phenotype was determined by measuring leaf blade B accumulation in F4 progeny individuals. To enable precise scoring of the QTL, the same progeny plants were scored for markers to follow the inheritance of recombinant and nonrecombinant 6H chromosomes, to confirm when the observed variation in B accumulation was controlled by segregation at the 6H locus.

Q-PCR and Northern Hybridization

Q-PCR assays were performed as described previously (Sutton et al., 2007). Primers (Supplemental Table S1) contained no mismatches to the corresponding HvNIP2;1 sequences. Northern hybridization was performed using standard methods.

Construction of a Three-Dimensional Molecular Model of HvNIP2;1

We used an archaeal aquaporin, AqpM, from Methanobacterium thermoautotrophicum (Protein Data Bank accession no. 2evu; Lee et al., 2005) as a structural template for modeling. The template structure was aligned with HvNIP2;1 using MUSTER (Wu and Zhang, 2008), after truncation of the HvNIP2;1 sequence at the NH2- and COOH-termini to eliminate nonstructural sequence. The alignment was checked manually to maintain the integrity of secondary structural elements, using Robetta (Kim et al., 2004) and hydrophobic cluster analysis (Callebaut et al., 1997). The structurally aligned sequences were used as input parameters to generate three-dimensional (3D) models of HvNIP2;1 in complex with B, As, Si, and Ge acids and four water molecules, using modeller 9v6 (Sali and Blundell, 1993). A 3D model of OsNIP2;1 in complex with B acid was also constructed. The coordinates of B and As acids were taken from the Protein Data Bank, while tetragonally coordinated Si and Ge acids were generated from Si acid diester and dimethylammonium gallotrigermanate, respectively, deposited in the Cambridge Crystallographic Data Centre. Models with the lowest value of the modeller 9v5 objective function were chosen for evaluation. The overall G-factors (estimates of stereochemical parameters) as evaluated by PROCHECK (Laskowski et al., 1993), were 0.45 for AqpM and between 0.03 and 0.08 for HvNIP2;1 in complex with metalloids. The z-score values (Sippl, 1993), reflecting combined statistical potential energy, for AqpM and the HvNIP2;1 models were −7.98 and between −5.23 and −5.47, respectively. The super algorithm in PyMol (DeLano, 2009) was used to determine the root-mean-square-deviation values in the Cα positions between the models and their template. The root-mean-square-deviation values, excluding three deletions of modeled HvNIP2;1, were 0.31 to 0.36 Å over 208 superposed residues. Buriedness of HvNIP2;1 was calculated by PocketPicker plug-in (Weisel et al., 2007) in PyMol. Molecular graphics were generated with PyMol (DeLano, 2009).

Heterologous Expression

NIP2;1 cDNA sequences were subcloned into the Gateway vector pCR8 and transferred via recombination to pYES-DEST52 (Invitrogen) for Gal-inducible expression in yeast (Saccharomyces cerevisiae). We used the yeast strains INVSc2 (MATαhis3-Δ1ura3-52) and SY1 (MATa,ura3-52,leu2-3,112,his4-619,sec6-4,GAL) in this study, and lithium-acetate transformation (Gietz and Woods, 2002). Yeast was routinely cultured in liquid or solid minimal media containing 2% (w/v) Glc or 2% (w/v) Gal as described previously (Sutton et al., 2007). For assessment of phenotypes, 10-fold dilutions of precultured yeast cells were deposited onto solid media containing B (pH 5.5), GeO2 (Aldrich), or As2O3 (Sigma). Alternatively, precultured cells were added to liquid medium supplemented with B, with growth assessed by measuring A600 of samples using a spectrophotometer. For measurement of permeability to B, we used SY1, a temperature-sensitive sec6-4 mutant yeast strain. Galactose cultures of SY1, transformed with or without HvNIP2;1, were grown at 22°C to an A600 of approximately 1, harvested by centrifugation, resuspended in fresh medium, and incubated for 2 h at 37°C to induce the accumulation of vesicles. Vesicles were isolated as described previously (Coury et al., 1999) and resuspended in a buffer of 0.3 m Suc and 25 mm Tris-HCl (pH 7.0) with added protease inhibitor (Sigma P8125; 3.8 μL mL−1). Permeability of the vesicle preparations to B was measured using a stopped flow spectrophotometer (DX.17MV, Applied Photophysics), where reswelling of vesicles (at 21.5°C) was measured by 90° light scattering (500 nm) upon rapidly mixing with a solution that created an inwardly directed concentration gradient for B in 25 mm Tris-HCl buffer (pH 7.0), containing 0.3 m Suc and 0.4 m H3BO3. The reswelling kinetics was fitted to a single exponential function, from which the rate constant was obtained. The diameter of the vesicles was determined using dynamic light scattering (NICOMP 380, Particle Sizing Systems) operating in a vesicle mode and was weighted on numbers of vesicles.

We performed cRNA transcription for Xenopus oocyte expression, using a T7 RNA polymerase kit (mMessage mMachine, Ambion). Xenopus oocytes were surgically removed and defolliculated, then injected with 30 ng cRNA or water, using a microinjector (Drummond Nanoject II automatic nanolitre injector, Drummond Scientific). The oocytes were allowed to recover for 3 d at 18°C in ND96 in 5 mm HEPES-NaOH buffer, pH 7.4 containing 96 mm NaCl, 2 mm KCl, 1.8 mm CaCl2, and 1 mm MgCl2 (osmolarity 199 mOsmol kg−1), with added antibiotics (tetracycline at 5 mg mL−1, penicillin, and streptomycin at 10 mg mL−1, all from Sigma).

To measure water permeability, oocytes were transferred to a 5-fold diluted solution of ND96 (osmolarity 47 mOsmol kg−1). To assess permeability to B, oocytes were transferred to a 5-fold diluted solution of ND96 supplemented with 160 mm B (osmolarity 201 mOsmol kg−1). Oocyte volume change was derived from images captured at 5 s intervals for 6 min under a dissecting microscope, and data were analyzed using GLOBAL LAB Image/2 software (Data Translation Inc.).

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers GQ496519 to GQ496522.

Supplemental Data

The following materials are available in the online version of this article.

  • Supplemental Figure S1. Determination of chromosome 6H B toxicity tolerance locus genotypes.
  • Supplemental Figure S2. Protein sequence alignment of HvNIP2;1 and OsNIP2;1.
  • Supplemental Figure S3. Alignment of Clipper and Sahara HvNIP2;1 nucleotide sequences immediately upstream of the 5′ ATG start site.
  • Supplemental Figure S4. Growth of yeast containing empty vector (black circles) compared to yeast expressing HvNIP2;1 (white circles) in liquid Gal minimal media, containing A, no added B, or B, 5 mm B.
  • Supplemental Figure S5. Growth of yeast containing empty vector and those transformed with vector containing HvNIP2;1 on solid Glc media (to repress transcription from GAL1 promoter) containing A, no added B, B, 5 mm B, and C, 10 mm B.
  • Supplemental Table S1. Primers used in this work.

Supplementary Material

[Supplemental Data]


We thank M. Pallotta, A. Hay, A. Okada, T. Oz, and N. Shirley for technical assistance; N. Collins for propagation of mapping lines; and J.F. Ma for provision of the rice mutant lsi1.


  • Bansal A, Sankararamankrishnan R. (2007) Homology modeling of major intrinsic proteins in rice, maize and Arabidopsis: comparative analysis of transmembrane helix association and aromatic/arginine selectivity filters. BMC Struct Biol 7: 27. [PMC free article] [PubMed]
  • Beitz E, Wu B, Holm LM, Schultz JE, Zeuthen T. (2006) Point mutations in the aromatic/arginine region in aquaporin I allow passage of urea, glycerol, ammonia and protons. Proc Natl Acad Sci USA 103: 269–274 [PMC free article] [PubMed]
  • Bhattacharjee H, Mukhopadhyay R, Thiyagarajan S, Rosen BP. (2008) Aquaglyceroporins: ancient channels for metalloids. J Biol 7: 33. [PMC free article] [PubMed]
  • Bienert GP, Thorsen M, Schüssler MD, Nilsson HR, Wagner A, Tamás MJ, Jahn TP. (2008) A subgroup of plant aquaporins facilitate the bi-directional diffusion of As(OH)3 and Sb(OH)3 across membranes. BMC Biol 6: 26. [PMC free article] [PubMed]
  • Borgnia M, Agre P. (2001) Reconstitution and functional comparison of purified GlpF and AqpZ, the glycerol and water channels from Escherichia coli. Proc Natl Acad Sci USA 98: 2888–2893 [PMC free article] [PubMed]
  • Callebaut I, Labesse G, Durand P, Poupon A, Canard L, Chomilier J, Henrissat B, Mornon JP. (1997) Deciphering protein sequence information through hydrophobic cluster analysis (HCA): current status and perspectives. Cell Mol Life Sci 53: 621–645 [PubMed]
  • Cartwright B, Zarcinas BA, Spouncer LR. (1986) Boron toxicity in South Australian barley crops. Aust J Agric Res 37: 351–359
  • Chang WC, Lee TY, Huang HD, Huang HY, Pan RL. (2008) PlantPAN: Plant Promoter Analysis Navigator, for identifying combinatorial cis-regulatory elements with distance constraint in plant gene groups. BMC Genomics 9: 561. [PMC free article] [PubMed]
  • Chiba Y, Mitani N, Yamaji N, Ma JF. (2008) HvLsi1 is a silicon influx transporter in barley. Plant J 57: 810–818 [PubMed]
  • Choi WG, Roberts DM. (2007) Arabidopsis NIP2;1, a major intrinsic protein transporter of lactic acid induced by anoxic stress. J Biol Chem 282: 24209–24218 [PubMed]
  • Christensen JJ. (1934) Non-parasitic leaf spots of barley. Phytopathology 24: 726–742
  • Coury LA, Zeidel ML, Brodsky JL. (1999) Use of yeast sec6 mutant for purification of vesicles containing recombinant membrane proteins. Methods Enzymol 306: 169–186 [PubMed]
  • Danielson JAH, Johanson U. (2008) Unexpected complexity of the aquaporin gene family in the moss Physcomitrella patens. BMC Plant Biol 8: 45. [PMC free article] [PubMed]
  • DeLano W. (2009) The PyMOL Molecular Graphics System. DeLano Scientific LLC, San Carlos, CA
  • Dordas C, Brown PH. (2000) Permeability of boric acid across lipid bilayers and factors affecting it. J Membr Biol 175: 95–105 [PubMed]
  • Dordas C, Chrispeels MJ, Brown PH. (2000) Permeability and channel-mediated transport of boric acid across membrane vesicles isolated from squash roots. Plant Physiol 124: 1349–1361 [PMC free article] [PubMed]
  • Eaton FM. (1944) Deficiency, toxicity and accumulation of boron in plants. J Agric Res 69: 237–277
  • Engel A, Fujiyoshi Y, Agre P. (2000) The importance of aquaporin channel protein structures. EMBO J 19: 800–806 [PMC free article] [PubMed]
  • Fitzpatrick KL, Reid RJ. (2009) The involvement of aquaglyceroporins in transport of boron in barley roots. Plant Cell Environ 32: 1357–1365 [PubMed]
  • Gietz RD, Woods RA. (2002) Transformation of yeast by the Liac/SS carrier DNA/PEG method. Methods Enzymol 350: 87–96 [PubMed]
  • Hayes JE, Reid RJ. (2004) Boron tolerance in barley is mediated by efflux of boron from the roots. Plant Physiol 136: 3376–3382 [PMC free article] [PubMed]
  • Isayenkov SV, Maathuis FJM. (2008) The Arabidopsis thaliana aquaglyceroporin AtNIP7;1 is a pathway for arsenite uptake. FEBS Lett 582: 1625–1628 [PubMed]
  • Jefferies SP, Barr AR, Karakousis A, Kretschmer JM, Manning S, Chalmers KJ, Nelson JC, Islam A, Langridge P. (1999) Mapping of chromosome regions conferring boron toxicity tolerance in barley (Hordeum vulgare L.). Theor Appl Genet 98: 1293–1303
  • Kamiya T, Tanaka M, Mitani N, Ma JF, Maeshima M, Fujiwara T. (2009) NIP1;1, an aquaporin homolog, determines the arsenite sensitivity of Arabidopsis thaliana. J Biol Chem 284: 2114–2120 [PubMed]
  • Kidwell MG, Lisch D. (1997) Transposable elements as sources of variation in animals and plants. Proc Natl Acad Sci USA 94: 7704–7711 [PMC free article] [PubMed]
  • Kim DE, Chivian D, Baker D. (2004) Protein structure prediction and analysis using the Robetta server. Nucleic Acids Res (Suppl 2) 32: W526– 231 [PMC free article] [PubMed]
  • Krissinel E, Henrick K. (2007) Inference of macromolecular assemblies from crystalline state. J Mol Biol 372: 774–797 [PubMed]
  • Krom N, Recla J, Ramakrishna W. (2008) Analysis of genes associated with retrotransposons in the rice genome. Genetica 134: 297–310 [PubMed]
  • Laskowski RA, MacArthur MW, Moss DS, Thornton JM. (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Cryst 26: 283–291
  • Lee JK, Kozono D, Remis J, Kitagawa Y, Agre P, Stroud RM. (2005) Structural basis for conductance by the archeal aquaporin AqpM at 1.68 Å. Proc Natl Acad Sci USA 102: 18932–18937 [PMC free article] [PubMed]
  • Ma JF, Tamai K, Yamaji N, Mitani N, Konishi S, Katsuhara M, Ishiguro M, Murata Y, Yano M. (2006) A silicon transporter in rice. Nature 440: 688–691 [PubMed]
  • Ma JF, Yamaji N, Mitani N, Xu XY, Su YH, McGrath SP, Zhao FJ. (2008) Transporters of arsenite in rice and their role in arsenic accumulation in rice grain. Proc Natl Acad Sci USA 105: 9931–9935 [PMC free article] [PubMed]
  • Mitani N, Yamaji N, Ma JF. (2008) Characterization of substrate specificity of a rice silicon transporter, Lsi1. Pflugers Arch 456: 679–686 [PubMed]
  • Mitani N, Yamaji N, Ma JF. (2009) Identification of maize silicon influx transporters. Plant Cell Physiol 50: 5–12 [PMC free article] [PubMed]
  • Nable RO, Lance RCM, Cartwright B. (1990) Uptake of boron and silicon by barley genotypes with differing susceptibilities to boron toxicity. Ann Bot (Lond) 66: 83–90
  • Nakamoto RK, Rao R, Slayman CW. (1991) Expression of the yeast plasma-membrane [H+]ATPase in secretory vesicles: a new strategy for directed mutagenesis. J Biol Chem 266: 7940–7949 [PubMed]
  • Nikolic M, Nikolic N, Liang Y, Kirkby EA, Römheld V. (2007) Germanium-68 as an adequate tracer for silicon transport in plants: characterization of silicon uptake in different crop species. Plant Physiol 143: 495–503 [PMC free article] [PubMed]
  • Nozawa A, Takano J, Kobayashi M, von Wiren N, Fujiwara T. (2006) Roles of BOR1, DUR3, and FPS1 in boron transport and tolerance in Saccharomyces cerevisiae. FEMS Microbiol Lett 262: 216–222 [PubMed]
  • Reid R. (2007) Identification of boron transporter genes likely to be responsible for tolerance to boron toxicity in wheat and barley. Plant Cell Physiol 48: 1673–1678 [PubMed]
  • Rivers RL, Dean RM, Chandy G, Hall JE, Roberts DM, Zeidel ML. (1997) Functional analysis of nodulin 26, an aquaporin in soybean root nodule symbiosomes. J Biol Chem 272: 16256–16261 [PubMed]
  • Rougé P, Barre A. (2008) A molecular modeling approach defines a new group of Nodulin 26-like aquaporins in plants. Biochem Biophys Res Commun 367: 60–66 [PubMed]
  • Ruetz S, Gros P. (1987) Functional expression of P-glycoproteins in secretory vesicles. J Biol Chem 289: 12277–12284 [PubMed]
  • Sade N, Vinocur BJ, Diber A, Shatil A, Ronen G, Nissan H, Wallach R, Karchi H, Moshelion M. (2009) Improving plant stress tolerance and yield production: is the tonoplast aquaporin SlTIP2;2 a key to isohydric to anisohydric conversion? New Phytol 181: 651–661 [PubMed]
  • Sali A, Blundell TL. (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234: 779–815 [PubMed]
  • Shelden MC, Howitt SM, Kaiser BN, Tyerman SD. (2009) Identification and functional characterisation of aquaporins in the grapevine, Vitis vinifera. Funct Plant Biol 36: 1065–1078
  • Sippl MJ. (1993) Recognition of errors in three-dimensional structures of proteins. Proteins 17: 355–362 [PubMed]
  • Sommer AL, Lipman CB. (1929) Evidence on the indispensible nature of zinc and boron for higher green plants. Plant Physiol 1: 231–249 [PMC free article] [PubMed]
  • Stangoulis JCR, Reid RJ, Brown PH, Graham RD. (2001) Kinetic analysis of boron transport in Chara. Planta 213: 142–146 [PubMed]
  • Steudle E, Peterson CA. (1998) How does water get through roots? J Exp Bot 49: 775–788
  • Sutton T, Baumann U, Hayes J, Collins NC, Shi BJ, Schnurbusch T, Hay A, Mayo G, Pallotta M, Tester M, et al. (2007) Boron-toxicity tolerance in barley arising from efflux transporter amplification. Science 318: 1446–1449 [PubMed]
  • Takano J, Wada M, Ludewig U, Schaaf G, von Wiren N, Fujiwara T. (2006) The Arabidopsis major intrinsic protein NIP5;1 is essential for efficient boron uptake and plant development under boron limitation. Plant Cell 18: 1498–1509 [PMC free article] [PubMed]
  • Tanaka M, Wallace IS, Takano J, Roberts DM, Fujiwara T. (2008) NIP6;1 is a boric acid channel for preferential transport of boron to growing shoot tissues in Arabidopsis. Plant Cell 20: 2860–2875 [PMC free article] [PubMed]
  • Wallace IS, Roberts DM. (2005) Distinct transport selectivity of two structural subclasses of the nodulin-like intrinsic protein family of plant aquaglyceroporin channels. Biochemistry 44: 16826–16834 [PubMed]
  • Weisel M, Proschak E, Schneider G. (2007) PocketPicker: analysis of ligand-binding sites with shape descriptors. Chem Cent J 1: 7–24 [PMC free article] [PubMed]
  • West RW, Jr, Yocum RR, Ptashne M. (1984) Saccharomyces cerevisiae GAL1-GAL10 divergent promoter region: location and function of the upstream activating sequence UASG. Mol Cell Biol 4: 2467–2478 [PMC free article] [PubMed]
  • Wu S, Zhang Y. (2008) MUSTER: improving protein sequence profile-profile alignments by using multiple sources of structure information. Proteins 72: 547–556 [PMC free article] [PubMed]
  • Yamaji N, Ma JF. (2007) Spatial distribution and temporal variation of the rice silicon transporter Lsi1. Plant Physiol 143: 1306–1313 [PMC free article] [PubMed]

Articles from Plant Physiology are provided here courtesy of American Society of Plant Biologists
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...