Boron-Toxicity Tolerance in Barley Arising from Efflux Transporter Amplification

See allHide authors and affiliations

Science  30 Nov 2007:
Vol. 318, Issue 5855, pp. 1446-1449
DOI: 10.1126/science.1146853


Both limiting and toxic soil concentrations of the essential micronutrient boron represent major limitations to crop production worldwide. We identified Bot1, a BOR1 ortholog, as the gene responsible for the superior boron-toxicity tolerance of the Algerian barley landrace Sahara 3771 (Sahara). Bot1 was located at the tolerance locus by high-resolution mapping. Compared to intolerant genotypes, Sahara contains about four times as many Bot1 gene copies, produces substantially more Bot1 transcript, and encodes a Bot1 protein with a higher capacity to provide tolerance in yeast. Bot1 transcript levels identified in barley tissues are consistent with a role in limiting the net entry of boron into the root and in the disposal of boron from leaves via hydathode guttation.

Of all plant nutrient elements, boron has the narrowest range between deficient and toxic soil solution concentration (1), and both boron deficiency and toxicity severely limit crop production worldwide (2, 3). Toxicity is more difficult to manage agronomically and is best dealt with by using boron-tolerant varieties. Genetic variation for boron-toxicity tolerance is known for a number of crop plant species. Tolerance is most commonly associated with the ability to maintain low boron concentrations in the shoot (46). In barley (Hordeum vulgare), the non-agronomic but highly boron-tolerant Algerian landrace Sahara 3771 (Sahara) was identified as a potential source of tolerance for variety improvement (4, 7). In a cross between Sahara and the boron-intolerant Australian malting variety Clipper, several quantitative trait loci (QTL) controlling tolerance were identified (8). The major locus on chromosome 4H affects leaf symptom expression (Fig. 1A), boron accumulation (Fig. 1B), root length response, and dry matter production under boron-toxic conditions (8). The ability of Sahara to maintain lower shoot boron accumulation is at least partially due to a mechanism of active boron efflux from the root (9).

Fig. 1.

Genetic variation for boron tolerance in barley. (A) Boron-toxicity symptoms in leaf blades of boron-intolerant (Clipper) and boron-tolerant (Sahara) barley plants. Approximately eight of the oldest leaf blades from single plants are shown 14 days after boron treatment. (B) Sahara accumulates less boron in leaf blades after growing for 14 days in a range of solution concentrations. Data are means ± SEM (n = 3 plants).

We followed a map-based approach to isolate the 4H boron-tolerance gene. Using a population representing 6720 meioses and gene colinearity with the syntenic region on rice chromosome 3 to generate markers, we delimited the tolerance locus to a 0.15-centimorgan interval between markers xBM178 and xBM162 (Fig. 2A and fig. S1) (10). The corresponding 11.2-kb interval in rice contains two intact copies and one 3′-truncated version of a gene showing similarity to a family of adenosine monophosphate (AMP)–dependent synthetases and ligases and no other predicted gene. Barley expressed sequence tags (ESTs) most closely matching one of the intact copies were used to derive the marker xBM160, which cosegregated with the tolerance locus.

Fig. 2.

Bot1 mapping and phylogenetic analysis. (A) Barley/rice comparative mapping of the boron toxicity–tolerance locus on chromosome 4H. Numbers indicate recombinants identified for each marker interval. The related rice chromosome 3 (AP008209) interval spans the position from base pairs 1,861,805 to 1,873,065. Colored arrows I to VIII denote the position and orientation of predicted rice genes Os03g0133300 (hypothetical protein), Os03g0133400 (peptidoglycan-binding LysM domain–containing protein), Os03g0133500 [AMP binding protein 1 (AMPBP1)], Os03g0133600 (AMP-dependent synthetase and ligase domain–containing protein), Os03g0133800 (AMPBP1), Os03g0133900 (serine acetyltransferase), Os03g0134000 (MuDR transposase domain–containing protein), and Os03g0134300 (adenosine triphosphate phosphoribosyl transferase), respectively. Accession numbers are: xBot1, BV723959; xBM160, BV723960; xBM165, BV723961; xBM178, BV723962; xBM181, BV723963; and xBM162, BV723964. (B) Phylogenetic tree including Bot1- and BOR1-like proteins from rice (green) and Arabidopsis (blue), constructed using PHYLIP (28). Numbers separating junctions are bootstrap values. The closest rice ortholog of Bot1 is Os01g08040. Locus accession numbers are according to Gramene (29).

In a parallel approach, we also mapped several candidate genes in barley. These were barley genes showing similarity to the Arabidopsis NIP5; 1 major intrinsic protein (11) and the Arabidopsis BOR1 efflux transporter that is related to bicarbonate transporters in animals (12, 13). Both Arabidopsis genes are required for healthy growth under conditions of low boron supply. However, in plants the genes involved in boron-toxicity tolerance may be related to those shown to function in boron efficiency. Comparisons of barley ESTs revealed four BOR1 (At2g47160.1)–related genes. Mapping localized one of the barley genes (Bot1) to the region of the boron-tolerance QTL on 4H (8). Subsequently, a marker developed from the 3′ end of Bot1 was found to cosegregate perfectly with the tolerance locus in our high-resolution mapping population (Fig. 2A), strongly suggesting that Bot1 encodes the boron tolerance from the 4H locus. Although barley/rice gene colinearity was found to be high in the region (Fig. 2A), the corresponding interval on rice chromosome 3 lacks a BOR1 ortholog, and the rice gene most closely resembling Bot1 (Os01g08040) (Fig. 2B) resides on chromosome 1.

Southern hybridization with the use of a Clipper derived Bot1 probe gave a stronger signal in Sahara than in Clipper and other boron-intolerant genotypes, indicating the occurrence of additional Bot1 copies in Sahara (fig. S2). A number of restriction enzyme digests revealed hybridizing Sahara fragments of mostly a single size (e.g., Xba I) (fig. S2), suggesting that the Bot1 copies in Sahara are highly similar. With Dra I, which distinguishes the Bot1 Clipper copy from the Sahara copies, all Bot1 genes could be mapped and were found to cosegregate with boron tolerance in the Clipper-by-Sahara F1–derived doubled-haploid population, indicating that these genes occur in a cluster. Quantitative real-time fluorescence polymerase chain reaction (QPCR) analysis using genomic DNA as the template indicated that Sahara contains about four times (3.8 ± 0.17) more copies of the gene than Clipper.

QPCR performed using cDNA as a template revealed that Bot1 transcript levels in Sahara were ∼160- and 18-fold higher in roots and leaf blades, respectively, as compared with Clipper (fig. S3). This increase in Bot1 transcript levels exceeds the ∼fourfold increase in the Bot1 copy number in Sahara, suggesting that factors additional to gene duplication may contribute to increased Bot1 transcript levels in Sahara. We performed a comparative promoter analysis between the Clipper and Sahara Bot1 alleles to search for differences that could account for the observed genotypic variation in transcript levels. Over a 1.3-kb region upstream of the mRNA transcription start site, the Clipper and Sahara Bot1 promoter regions are 96% identical. On the basis of database searches, no substantial differences were detected in known regulatory elements. More work will be required to determine the actual effect of these changes on Bot1 transcription in barley. In any case, the greater transcript levels in Sahara, relative to Clipper, offer an explanation for the functional difference between boron-tolerance and -intolerance alleles and provide additional evidence supporting Bot1 as the gene controlling boron tolerance at the 4H locus. In both roots and leaf blades, transcript levels were unaltered by exposure to a range of boron concentrations (fig. S3). Lack of transcriptional activation of a boron-tolerance mechanism is consistent with the rapid boron efflux from Sahara roots observed after the addition of either nontoxic or toxic quantities of boron (9) and the similar rank order of shoot boron accumulation in different genotypes grown over a range of boron concentrations (4).

The ability of Bot1 to function as a boron transporter was confirmed in yeast (Saccharomyces cerevisiae). Initially, we screened ∼2 million clones from a Sahara root cDNA expression library for their ability to confer boron tolerance to yeast. Three clones corresponding to Bot1 were obtained that allowed yeast to grow on high boron media. The Bot1 clones contained complete open reading frames (ORFs) and were identical in sequence. In yeast, we then compared the ability of Sahara Bot1 with that of Clipper Bot1 or Arabidopsis BOR1 to confer boron tolerance. Yeast expressing Sahara Bot1 grew better than yeast expressing Clipper Bot1 in the presence of high boron on both a solid medium (Fig. 3) and in liquid culture (fig. S4). Cells expressing either Sahara Bot1 or Clipper Bot1 also maintained ∼24 or 20% less cellular boron, respectively, than cells expressing the empty vector control (fig. S4). This is in spite of the fact that at physiological pH boron exists principally as undissociated boric acid (pKa1 to pKa3 = 9.2 to 13.8, where Ka is the acid dissociation constant), to which membranes are relatively permeable (14). Additionally, as compared with BOR1 of Arabidopsis, we could show that Sahara Bot1 has higher boron efflux transport activity and capacity to provide tolerance (fig. S5). These results confirmed that Bot1, like BOR1 of Arabidopsis, encodes a functional boron efflux transporter and that Sahara Bot1 has a higher capacity to provide boron tolerance in yeast than Clipper Bot1 or BOR1. None of the clones identified by the library screen corresponded to the tolerance cosegregating gene BM160, further supporting the notion that Bot1 and not BM160 is the tolerance gene (15).

Fig. 3.

Bot1 provides boron tolerance in yeast. Growth of S. cerevisiae on a solid medium containing (left)0 or (right) 20 mM supplemental H3BO3. Each plate shows two independent yeast clones containing either empty vector, Sahara Bot1, or Clipper Bot1 (left to right in each panel). Plating was a 10 μl aliquot of 10-fold serial dilutions down the plate (top to bottom in each panel).

The Sahara Bot1 mRNA (accession number EF660435) is predicted to encode a 666–amino acid protein with 10 to 12 putative transmembrane helices (fig. S6). Within the ORF, Clipper Bot1 (accession number EF660437) differs by 11 nucleotides, 2 of which result in differences to the translated protein: Leu305 → Ser305 and Asp592 → Gly592. Residue change 305 [Leu (hydrophobic) to Ser (polar)] in transmembrane helix eight may impart a conformation change, and residue change 592 [Asp (polar) to Gly (no side chain)] is likely to be located within the intracellular carboxyl terminus. Both changes could affect boron transport. Bot1 showed greater sequence similarity to several other Arabidopsis BOR1-related proteins than to BOR1 itself (Fig. 2B and fig. S7), consistent with Bot1 serving a different role than BOR1, which is required for boron efficiency. Overexpression of one of these BOR1-related genes in Arabidopsis improved boron tolerance (16), further supporting our hypothesis. The genomic sequence of a Bot1 gene (accession number EF660436) was obtained from a clone (accession number EU176161) of a bacterial artificial chromosome library we constructed from a barley doubled-haploid line containing the Sahara 4H boron-tolerance allele. This sequence was identical to the cDNA in the coding sequence, and it contained 13 exons and 12 introns, including a 941–base pair intron in the 5′ untranslated region (fig. S6).

Bot1 mRNA was localized by in situ hybridization to barley-root and leaf-blade tissue sections (Fig. 4, A to J). In roots, staining was strongest in all cell types of the meristematic and elongation zones at the tip (Fig. 4, I and J). Within cylindrical-sheath tissues (Fig. 4, G and H) staining was strong in all cells of the youngest leaf blades. QPCR from 10-mm segments taken along the root revealed that Bot1 mRNA level, expressed as a proportion of total RNA content, was slightly greater in more mature root segments than in end segments containing root tips (Fig. 4K). The lower levels of staining in mature roots may reflect the proportionally lower volume of cytoplasm in these cells, but the significant transcript levels nevertheless observed by QPCR in mature roots indicate a role for the transporter in both growing and mature sections of the roots. The Bot1 mRNA detected in young cells could potentially serve the dual role of limiting symplastic boron concentration (and hence toxicity) and helping to maintain a high boron supply to newly forming primary cell walls where boron forms an essential component of the pectic polysaccharide rhamnogalacturonan II (17). It is probable that Bot1 mRNA localization in mature roots helps to efflux boron from the roots. In older leaf blades, Bot1 mRNA staining was strong in mesophyll adjacent to enlarged vessels near the margins and was strongest at the tips (Fig. 4, A and B). In barley, leaf-blade tips are the site of guttation by hydathodes. Bot1 mRNA localization here suggests an additional role of Bot1 in facilitating boron removal from the leaf blade via guttation fluid and hence in reducing boron toxicity in the leaves (1820). Guttation has previously been shown to be a route by which substantial amounts of boron can be removed from leaves (18). In field conditions, rainfall decreases boron concentrations in shoots of barley (21) and wheat (22).

Fig. 4.

Localization of Bot1 mRNA in barley. In situ localization of Bot1 mRNA in transverse (A to H) and longitudinal (I and J) sections of Sahara [(A), (C), (E), (G), and (I)] and Clipper [(B), (D), (F), (H), and (J)] barley tissues grown at 1 mM H3BO3, probed with antisense Bot1. Bot1 mRNA staining is strong at the tips of leaf blades [(A) and (B)] in mesophyll surrounding the enlarged vessels [arrowheads in (B)]. Relative to the leaf-blade tip, staining is reduced 2 cm back from the tip [(C) and (D)] but remains specific to mesophyll surrounding the enlarged vessels in Sahara [arrowhead in (C)]. At the middle of the youngest fully emerged leaf blade [(E) and (F)], this pattern was not evident. Bot1 mRNA staining is also strong in the youngest tissue of the cylindrical sheath [(G) and (H)] and root tip [(I) and (J)]. In roots, staining extended into the mature stele [arrowheads in (I) and (J)]. A quantitative difference in staining between Sahara and Clipper leaf tissue sections is evident. (A to H) Transmission electron microscopy fixative. (I and J) Formalin–acetic acid–alcohol fixative. Sense hybridizations were performed for all tissue sections and produced no detectable signal. The plant at center is a 10-day-old Clipper seedling. Scale bars: 100 μm in (A) to (F); 500 μm in (G) to (J). (K) Bot1 mRNA levels in 10-mm root segments taken along the root were analyzed by QPCR. The different letters above each bar represent statistically significant differences (analysis of variance, P < 0.05, post hoc Tukey test). Data are means ± SEM (n = 3 plants).

In Australian barley crops, grain yield has been estimated to be reduced by 17% as a result of boron toxicity (23). The importance of boron tolerance throughout Central and West Asia and North Africa is based on an observation that landraces of barley and wheat sourced from these regions tend to show high levels of tolerance—for example, the line used in the current study (Sahara) and lines from Turkey (24), Iraq, and Afghanistan (25). Poor adoption of modern varieties in some of these areas (26) may be associated with boron intolerance. There is also evidence that boron toxicity will become an increasing problem on irrigated land, coincident with salinity (5).

Barley breeding programs in Australia have used linked molecular markers to introgress the Sahara Bot1 allele into commercial barley cultivars. However, lines carrying the introgression can potentially be lower-yielding than the recipient cultivars (27). We currently do not know if this potential yield reduction is the result of pleiotropic effects of Bot1 itself or is due to linked deleterious alleles from the otherwise unadapted landrace Sahara. To help break unwanted associations, lines carrying recombination events directly adjacent to Bot1 have been identified and supplied to breeders. The identification of Bot1 as the gene controlling the 4H QTL now offers new paths for dealing with boron toxicity. Diagnostic markers are being developed for use in barley breeding and selection strategies. It should also be possible to generate boron-tolerant plants by transformation. The ability of Bot1 to confer tolerance in yeast and barley suggests that, under appropriate transcriptional regulation, the gene may be of value in engineering tolerance in a wide range of plant species.

Supporting Online Material

Materials and Methods

Figs. S1 to S7

Table S1


References and Notes

Stay Connected to Science

Navigate This Article