Loss of an MDR Transporter in Compact Stalks of Maize br2 and Sorghum dw3 Mutants

Plants with short stature have had a major impact on world agriculture. This notion is best exemplified by the success of the green revolution, which was made possible by the advent of dwarf varieties of rice and wheat (1). Agronomic interest in short plants derives largely from their ability to resist lodging caused by wind, rain, or higher densities, which allows them to effectively convert increased fertilizer input into higher yields (1, 2).

Elongation of plant parts is a complex phenomenon mediated by many plant hormones, including auxins, brassinosteroids, and gibberellins (GAs) (3). However, our understanding of how these hormones regulate cell elongation remains limited (3). One way to resolve mechanisms that control plant height is by characterizing mutants in which this growth component is compromised. This approach recently led to the elucidation of mechanisms that caused dwarfing of the green revolution rice and wheat cultivars, which were found to have defects in the genes involved in the biosynthesis and signaling of GAs, respectively (2).

A maize dwarf mutant of agronomic potential contains the recessive brachytic2 (br2) mutation, which results in the shortening of lower stalk internodes (4, 5). No other plant parts— including the mesocotyl, coleoptile, leaves, ear, and tassel—are notably affected in size or growth in br2 mutants (Fig. 1, A to C) (5). The br2 phenotype could not be reversed by treatment with auxins, brassinosteroids, cytokinins, or GAs, suggesting that br2 does not contain a defect in the biosynthesis of any of these growth regulators.

• Download high-res image
• Open in new tab
• Download Powerpoint

In contrast to the reduction in length, the girth of the affected stalk internodes is often enhanced (Fig. 1, B and C). We compared the cellular architecture of elongating internodes of br2 mutants with that of their wild-type siblings (6). As expected, the length of br2 stalk cells was found to be 40 to 50% of the length of their normal counterparts (Fig. 1, D and E). A similar reduction was observed in the diameter of br2 stalk cells, and it was coupled with a substantial increase in cell number (Fig. 1, F and G). This increase could be observed throughout the breadth of the br2 stalk, but it was especially evident in the hypodermal region, where a parenchyma region of up to 10 cell layers across developed immediately below the epidermis (Fig. 1G). Additional aberrations were observed in the vascular system, in which the structure, number, and distribution pattern were altered in br2 plants (Fig. 1, F and G). One consequence of these anatomical changes is that the strength of the br2 stalk is substantially enhanced. This was demonstrated genetically by combining br2 with brittle stalk2 (bk2), an easily breakable maize mutant that is normally impossible to propagate in the field (7). We were able to grow bk2 and br2 double-mutant plants to maturity in the field unassisted.

Altered polar auxin transport in br2 plants. Aberrations associated with stalk vasculature and hypodermal proliferation suggested that some aspect of auxin homeostasis was perturbed in the br2 stalk. This possibility was evaluated by comparing basipetal transport of [³H]indole-3-acetic acid (IAA) in br2 and wild-type plants (6). These assays showed a light-dependent effect on auxin transport in middle to lower stem tissues of br2 mutants. The transport of auxin in the first internodes of light-grown plants was greatly reduced in br2 mutants compared with that in the wild type (Fig. 2A). However, only slight differences were seen in br2 dark-grown coleoptile or mesocotyl tissues (Fig. 2, B and C). The noted differences were not a result of increased [³H]auxin uptake at the site of application, given that br2 consistently showed increased loading of auxin into the upper coleoptile near the site of application (Fig. 2C). These patterns of accumulation indicate that the polar transport of auxin is impaired in a light-dependent fashion in the br2 stem (8), and they are consistent with the well-documented differences between the rates of auxin transport in the upper and lower portions of maize coleoptiles (9).

• Download high-res image
• Open in new tab
• Download Powerpoint

We cloned the br2 gene by transposon tagging with Mutator (Mu) (6). Two sets of results were used to establish the identity of the cloned gene as br2. These include four independent insertions within a few kilobases of each other in four mutant alleles (Fig. 3A), and the absence of any of the insertions in the common progenitor of two of the mutant alleles, br2-3 and br2-6 (Fig. 3B). Evidence that the Br2 gene is expressed preferentially in elongating stalk internodes (Fig. 3C) lends further support to this conclusion. During this cloning endeavor, we also found that element 8 of the Mu system evolved by hijacking a part of the br2 gene (Fig. 3A).

• Download high-res image
• Open in new tab
• Download Powerpoint

BR2 is a P-glycoprotein. The Br2 gene encodes a protein similar to adenosine triphosphate (ATP)–binding cassette transporters of the multidrug resistant (MDR) class of P-glycoproteins (PGPs) (fig. S2). As with other PGPs, the predicted BR2 protein consists of two similar halves, each containing six putative transmembrane domains and an intracellular ATP nucleotide–binding domain (fig. S2A) (10). The closest Arabidopsis homolog of BR2 is AtPGP1, with 67% identity (fig. S2B). The classification of BR2 as a PGP is consistent with recent findings from Arabidopsis that suggest that PGPs modulate auxin-dependent growth. Mutations in AtPGP19 (AtMDR1), a close relative of AtPGP1, resulted in decreased auxin transport and pleiotropic auxin-related phenotypes (8). The combination of atpgp1 with the atpgp19 mutation resulted in more severe defects in growth and auxin transport (8). A function in polar auxin transport was further suggested when AtPGP1, AtPGP2, and AtPGP19 were purified by affinity chromatography with the auxin transport inhibitor N-1-naphthylphthalamic acid (11), and the putative auxin-efflux carrier PIN1 was found to be mislocalized in atpgp1 and atpgp19 mutants, resulting in enhanced lateral auxin flux and hypertropic bending responses (12). These studies, taken together with our findings with br2, strongly suggest that plant PGPs facilitate the polar movement of auxins.

Notably, the disruption of MDR genes has different consequences in maize than in Arabidopsis. Whereas mutations in maize Br2 affect primarily the growth of lower stalk internodes (Fig. 1, B and C), the loss of the equivalent gene(s) in Arabidopsis is associated with multiple morphologies (8). Conversely, the pattern of cellular proliferation, differentiation, and vasculature remains largely unaffected in the Arabidopsis stem (8), whereas all of these features are substantially altered in the br2 stalk (Fig. 1G). This disparity may simply reflect the spatial expression regimens of this subset of MDR genes, which occurs predominantly in elongating internodes of maize (Fig. 3C) but in overlapping expression patterns in meristems, shoot apices, flowers, stem nodes, and tissues associated with high auxin content in Arabidopsis (8, 13). Additionally, maize and Arabidopsis have fundamentally diverse body forms, consisting of segmented and unsegmented stems, respectively. This difference may be critical, considering that each segment (internode) of the maize stem has its own intercalary meristem. What role, if any, these meristems play in regulating auxin flow and homeostasis remains unknown, but this question can now be addressed using br2 as a tool.

Elucidation ofbr2helps demystify sorghumdw3. Agronomically, the br2 phenotype epitomizes what may be an ideal plant type for enhancing crop productivity (1). br2 has not been exploited commercially in maize, partly because of the excessively severe nature of the original mutant allele (Fig. 1A), but similar mutations have been used extensively in sorghum production since the 1950s. Sorghum is closely related to maize both in genomic organization and plant form (14). Four independent dwarfing mutations—dw1, dw2, dw3, and dw4—are available to sorghum breeders for reducing plant height (15), and typically, three mutations are variously combined to develop commercial lines. Because of its ability to improve the harvest index of sorghum, the dw3 dwarfing gene is often included in this combination (15, 16). However, the only mutant allele of dw3 available thus far is unstable and spontaneously reverts back to the tall type at a frequency of 0.1 to 0.5%, depending on the genetic background (Fig. 4A) (16, 17). The mechanism underlying dw3 instability has long puzzled sorghum geneticists, and attempts to rectify this problem have failed.

• Download high-res image
• Open in new tab
• Download Powerpoint

The dwarfing phenotype of dw3 is remarkably similar to that of br2 (Fig. 4B), and this prompted us to investigate whether the two were also identical at the molecular level. We accomplished this by hybridizing molecular probes from br2 to DNA that was isolated from isogenic dw3 mutant and revertant (Dw3) siblings (6). A single restriction band polymorphism was detected (Fig. 4C), which segregated completely with 40 mutant and 8 revertant siblings, suggesting that dw3 and br2 were identical structurally. They also share a common evolutionary origin, which was indicated by the finding that dw3 is flanked by hm1 and PIO644, two markers that are tightly linked to br2 in maize (6).

The polymorphic region of dw3 was amplified by a polymerase chain reaction (PCR) from both the mutant and revertant plants and was then sequenced (6). The maize br2 and the sorghum dw3 exhibited greater than 95% identity in the region corresponding to a part of exon 5 of br2. Subsequent cloning and sequencing of the entire dw3 gene from the mutant and revertant plants (6) indicated that it shares a common sequence and structure with the maize br2 gene (Fig. 4D and fig. S3). In addition, an 882–base pair (bp) direct duplication was found in exon 5 of the dw3 allele, which appeared responsible for the allele's loss of function (Fig. 4D). All of the eight Dw3 revertants lacked this duplication, indicating that it was also responsible for the instability in dw3.

Because direct duplications are apt to undergo unequal crossing-over (18), could this be the mechanism by which dw3 reverts back to Dw3? One way of answering this question is by identifying one or more recombinants that contain at least three copies of the duplicated region. To find such a recombinant, DNA from another 200 dwarf plants was subjected to Southern analysis (6). We identified a single plant that displayed a restriction pattern indicative of three copies of the duplicated region (Fig. 4E). Subsequent cloning and sequencing of this restriction fragment confirmed its triplicate nature, thereby demonstrating that dw3 reverts back to Dw3 by unequal crossing-over.

Interestingly, a dwarf plant with a restriction band diagnostic of wild-type revertants was also found among these 200 plants (Fig. 4E). PCR amplification and subsequent sequencing of its product indicated that unequal recombination had removed the duplicated part of the gene but introduced a number of simple nucleotide changes in the copy that was left behind (fig. S4). These changes disrupted the reading frame of DW3 and also truncated the protein by about 200 amino acids, thereby explaining the mutant nature of this new allele. Because this allele, designated dw3-sd1, lacks the duplication, it is expected to confer a stable mutant phenotype. This was determined by generating progeny that were homozygous for the dw3-sd1 allele. We screened more than 2400 such plants in the field and found that none reverted back to the tall type, confirming the stable dwarf nature of this mutant derivative. To determine whether imprecise recombination at dw3 was common enough to be practically useful, we analyzed another 500 dwarf plants by PCR (6). One plant was identified that yielded a product indicative of a loss of duplication. Its sequence revealed that it had undergone mutational changes similar to that of dw3-sd1 (19).

Concluding remarks. These findings not only resolve a long-standing puzzle in sorghum genetics and breeding but also provide a simple strategy for effectively correcting dw3 in the sorghum germplasm. Moreover, new mutant alleles of sorghum dw3 or of corresponding genes in other cereals may be generated by conventional mutagenesis approaches. There is also the prospect of inciting a renewed interest in this locus for maize breeding by generating new and improved alleles of br2. A key advantage of the dwarfing mechanism described here is its synergistic effect on stalk quality, a trait considered to be of utmost importance for enhancing crop yields beyond those that have already been achieved (1).