Report

AXR4 Is Required for Localization of the Auxin Influx Facilitator AUX1

See allHide authors and affiliations

Science  26 May 2006:
Vol. 312, Issue 5777, pp. 1218-1220
DOI: 10.1126/science.1122847

Abstract

The AUX1 and PIN auxin influx and efflux facilitators are key regulators of root growth and development. For root gravitropism to occur, AUX1 and PIN2 must transport auxin via the lateral root cap to elongating epidermal cells. Genetic studies suggest that AXR4 functions in the same pathway as AUX1. Here we show that AXR4 is a previously unidentified accessory protein of the endoplasmic reticulum (ER) that regulates localization of AUX1 but not of PIN proteins. Loss of AXR4 resulted in abnormal accumulation of AUX1 in the ER of epidermal cells, indicating that the axr4 agravitropic phenotype is caused by defective AUX1 trafficking in the root epidermis.

Polar auxin transport plays an important role in a wide variety of plant growth processes (1, 2). Studies indicate that the polarity of auxin transport is determined by the asymmetric localization of members of the PIN family of auxin efflux facilitators (1, 2). The polarity of PIN localization can change rapidly in response to environmental and developmental signals, resulting in localized changes in auxin concentration and distribution (35). In the case of PIN1, proper localization is associated with continuous recycling of the protein through the endomembrane system and the activity of a guanosine diphosphate/guanosine triphosphate exchange factor protein called GNOM (6, 7). Recent studies indicate that the auxin influx facilitator AUX1 is also asymmetrically localized in some root tissues (8). For example, AUX1 is preferentially localized to the upper plasma membrane of protophloem cells, where it is proposed to facilitate acropetal transport of auxin from the phloem into the root apex. However, the proteins that regulate AUX1 localization are not known.

The AXR4 gene was identified in a screen for Arabidopsis mutants that are resistant to the auxin 2,4-D (9, 10). The axr4 phenotype is markedly similar to aux1, with defects in lateral root formation and gravitropism (913). The gravitropic defects of aux1 and axr4, but not of the auxin response mutants axr2 and axr3, can be rescued by the membrane-permeable auxin 1-naphthaleneacetic acid (14, 15). Similarly, aux1 and axr4 confer selective resistance to various auxins (14, 15). Genetic analysis of aux1 axr4 double mutants suggests that these two genes function in the same pathway to regulate auxin-related developmental processes in the primary root (11). To probe this functional relationship further, we performed a principal component analysis on indole-related metabolites from root tissue of wild-type and various mutant genotypes, including axr4 and aux1. The results (Fig. 1A) indicate that aux1 and axr4 cluster together, implying that these mutations disrupt a related biological process. Neither mutant overlaps with any of the auxin signaling mutants (axr1, axr2, and axr3), supporting the view that like aux1, the axr4 mutation disrupts a process distinct from auxin signaling.

Fig. 1.

AXR4 encodes a previously unidentified protein required for auxin influx. (A) Principal component analysis of auxin (IAA, indole-3-acetic acid) and its metabolites in various genotypes. (B) Transcript profile variation between wild-type (wt) and axr4-2 roots. Quadruplicate microarray samples (24) were analyzed. Natural logarithm of the mean signal change (axr4/wt) is plotted against the significance of signal variation (natural logarithm of P-value). Arrow indicates data point for At1g54990, which showed expression in the mutant that was reduced relative to wild-type levels by a factor of 5.95 (P = 9.64 × 10–4). (C) Structure of the AXR4 gene with the positions of the axr4 mutant alleles indicated. Colored bars indicate exons and the black bar, an intron. (D) Phenotype of Col-0, axr4-2, and axr4-2 35S::AXR4 (left to right) seedlings on auxin. Four-day-old seedlings were transferred to medium containing 85 nM 2,4-D and photographed after 3 days.

To investigate the function of AXR4, we isolated the gene by positional cloning. Fine mapping localized the gene to a 750-kb region on chromosome 1. We then used ATH1 (Affymetrix) whole-genome microarrays to compare gene expression in wild-type and axr4 roots. The results (Fig. 1B) show that the expression profiles in the two genotypes are very similar. However, expression of the At1g54990 gene was decreased by a factor of ∼6 in the mutant as compared to the wild type. Because this gene lies within the interval that includes AXR4, we next characterized At1g54990 in four axr4 alleles. Sequencing of At1g54990 from the axr4-1 and axr4-2 alleles, mutants derived from T-DNA–mutagenized and γ-irradiated populations, revealed the presence of an insertion at positions 954 and 849, respectively (Fig. 1C). Sequence analysis of two new ethylmethane sulfonate–induced mutants revealed point mutations at positions 141 (axr4-3) and 412 (axr4-4), each of which result in stop codons. To confirm that At1g54990 is AXR4, we introduced the At1g54990 cDNA under control of the cauliflower mosaic virus (CaMV) 35S promoter into axr4-2 plants. When these transgenic plants were transferred to medium containing auxin, they responded like the wild-type line, with decreased root elongation (Fig. 1D). Together with our analysis of the axr4 alleles, these results indicate that At1g54990 encodes AXR4.

The expression of AXR4 was initially analyzed using two different approaches: reverse transcription–polymerase chain reaction (RT-PCR) (Fig. 2A) and promoter reporter experiments using transgenic lines expressing GUS under the control of the AXR4 promoter (Fig. 2, B to D). RT-PCR experiments indicated that AXR4 mRNA is most abundant in root tissue, with lesser amounts in rosette leaves, stems, and flowers and very little in mature siliques. Examination of the AXR4::GUS lines indicated that expression of AXR4 is highest in the root tip (Fig. 2B). We also observed GUS staining in the vascular tissue of the root and hypocotyl and at sites of lateral root initiation (Fig. 2, C and D).

Fig. 2.

Expression of AXR4 and localization of the AXR4 protein. (A) AXR4 RNA levels in various tissues determined by RT-PCR. (B to D) AXR4::GUS seedlings showing AXR4 expression in the root tip (B), stele and lateral root primorida (C), and hypocotyl (D). (E) AXR4-GFP localization in root cells using confocal imaging. (F) Expression of AXR4-GFP around nuclei and at cell margins. (G) AXR4-GFP localization in root epidermal cells. (H) AUX1-YFP localization in root epidermal cells. (I) Superimposed confocal images of AXR4-GFP and AUX1-YFP localization. (J) AXR4-GFP localization in root cells using antibodies to GFP (anti-GFP). (K) BiP localization in root cells using anti-BiP. (L) Superimposed confocal images of AXR4-GFP and BiP localization.

AXR4 appears to be unique in the Arabidopsis genome. A single copy of an AXR4-like sequence is present in all plant genomes sequenced to date (fig. S1), indicating that it is conserved among higher plants. AXR4 is 473 amino acids in length and is predicted to contain a single transmembrane domain close to the N terminus, suggesting that it is an integral membrane protein (16) (fig. S2). In addition, the protein contains an α/β hydrolase fold, indicating that it is a member of the α/β hydrolase superfamily. To investigate the subcellular localization of AXR4, we generated 35S::AXR4-GFP (green fluorescent protein) and AXR4::AXR4-GFP constructs and introduced them into axr4-1 and axr4-2 plants. The axr4-1 35S::AXR4-GFP and axr4-2 AXR4::AXR4-GFP plants displayed a wild-type phenotype, indicating that the AXR4-GFP fusion is functional (fig. S3). To investigate whether AUX1 and AXR4 are colocalized, we crossed the 35S::AXR4-GFP line with a line carrying the AUX1::AUX1-YFP (yellow fluorescent protein) transgene (17). Confocal laser microscopy of root cells of F1 seedlings (8) revealed that AXR4-GFP does not colocalize with AUX1 at the plasma membrane (Fig. 2, G to I), suggesting that AXR4 is localized to another cellular membrane. To investigate this possibility, we tested colocalization of AXR-GFP with a selection of endomembrane compartment markers (fig. S4). AXR4-GFP was detected at the periphery of the cell and surrounding the nucleus (Fig. 2, F and G), mimicking the distribution of markers for the endoplasmic reticulum (ER) such as BiP (18) (Fig. 2, J to L) rather than the punctate pattern exhibited by the golgi apparatus marker γ-COP (fig. S4). Pixel correlation analysis (fig. S5) confirmed that the AXR4-GFP signal overlaps most significantly with markers for the ER (fig. S4), strongly suggesting that AXR4 is localized to the ER. The localization of AXR4 to the ER has also been independently determined using a proteomics-based technique termed LOPIT (localization of organ-elle proteins by isotope tagging) (19).

Although AXR4 and AUX1 do not colocalize, the genetic and physiological studies described above suggest that the two proteins function together to regulate auxin transport. Given its localization to the ER, we examined whether AXR4 might regulate AUX1 trafficking. The location of AUX1 was determined in wild-type and axr4 roots. Localization of AUX1 to the plasma membrane is asymmetric (at the upper side) in wild-type protophloem cells (Fig. 3C), mainly axial (at both upper and lower sides) in the epidermis (Fig. 3A), and without clear polarity in the lateral root cap (LRC) (Fig. 3A). The asymmetric localization of AUX1 was abolished in axr4 root epidermal cells (Fig. 3B; compare to Fig. 3A), whereas AUX1 was localized to an intracellular compartment in axr4 protophloem cells (Fig. 3G; compare to Fig. 3C). AUX1 localization was not markedly altered in the LRC of axr4 roots, suggesting that AXR4 may not be required for AUX1 trafficking in these cells. However, additional studies are required to confirm this possibility. These observations prompted us to investigate whether the trafficking of other membrane proteins is also dependent on AXR4 function. The auxin efflux facilitator PIN1 is normally found on the lower side of vascular cells (5)(Fig. 3D), whereas PIN2 is localized to the lower side of cortical cells and the upper side of epidermal cells (5) (Fig. 3E). We found that both PIN1 and PIN2 are localized normally in axr4 roots (Fig. 3, H and I). The localization of other plasma membrane proteins such as proton adenosine triphosphatase (PM H+-ATPase) is also normal in the axr4 mutant (Fig. 3, F and J). The axr4 mutation therefore appears to selectively disrupt trafficking of AUX1. We next investigated where AUX1 accumulates in the axr4 background by using a selection of endomembrane compartment markers (fig. S6). Pixel correlation analysis revealed that AUX1 overlaps most significantly with markers from the ER (fig. S6) like Sec12 (Fig. 3, K to M), suggesting that loss of AXR4 causes AUX1 to accumulate in the ER.

Fig. 3.

ER-localized AXR4 is required for AUX1 plasma membrane trafficking. (A and B) Hemagglutinin (HA)–AUX1 localization in the LRC and epidermis of Col-0 (A) and axr4-2 (B). (C and G) HA-AUX1 localization in the protophloem of Col-0 (C) and axr4-2 (G). (D and H) PIN1 localization in Col-0 (D) and axr4-2 (H). (E and I) PIN2 localization in Col-0 (E) and axr4-2 (I). (F and J) Localization of plasma membrane H+-ATPase in Col-0 (F) and axr4-2 (J). (K) HA-AUX1 localization in axr4-2. (L) Sec12 localization in axr4-2. (M) Superimposed images of HA-AUX1 and Sec12 in axr4-2.

Recent work has demonstrated that for root gravitropism to occur, AUX1 must be expressed in both the LRC and expanding epidermal cells in order to transport gravity-induced lateral auxin gradients from gravisensing columella cells to graviresponsive epidermal cells (20). Localization results suggest that the gravitropic defect of axr4 seedlings is caused by the failure to traffic AUX1 to the plasma membrane in the epidermis (Fig. 3B). As a result, the axr4 mutation is expected to disrupt the AUX1-dependent transfer of the lateral auxin gradient from the LRC to expanding epidermal cells. Consistent with this model, the auxin-responsive reporter IAA2::GUS is expressed normally in the LRC but is almost undetectable in the epidermis of the axr4 mutant (compare fig. S7, A and C). In contrast, IAA2::GUS expression is not detected in either the LRC or epidermal cells in the aux1 background (fig. S7B). The pattern of IAA2::GUS expression is consistent with the fact that AUX1 remains functional in the axr4 LRC but not in the epidermal cells. The axr4 mutant is resistant to 2,4-D, and because AUX1 may be functional in the LRC in axr4 mutants, this would suggest that functional AUX1 in the LRC alone is not sufficient for 2,4-D–sensitive root growth. We tested this possibility by transactivating AUX1 in the LRC or LRC plus epidermal cells (fig. S8) using Gal4 driver lines M0013 and J0951, respectively (20). When AUX1 was expressed in LRC alone using the GAL4 line M0013, the aux1 M0013≫AUX1 line exhibited an auxin-resistant root phenotype like that of axr4 (fig. S8, E and F) (20). However, when AUX1 was also expressed in the epidermal cells using the GAL4 line J0951, normal auxin response was restored (fig. S8, B and F). Our results suggest that the axr4 phenotype, including auxin-resistant root growth and reduced gravitropism, is caused by defective AUX1 trafficking in epidermal cells.

AXR4 joins a growing list of ER accessory proteins that facilitate trafficking of plasma membrane proteins through the secretory pathway (2123). In yeast cells, the Shr3 protein acts as a molecular chaperone of amino acid permeases, preventing their inappropriate aggregation in the ER membrane (21). Other structurally unrelated accessory proteins also appear to function as molecular chaperones of their cognate substrates (21). Although unrelated to Shr3p, AXR4 may have a similar function. Alternatively, because AXR4 is a member of the α/β hydrolase superfamily, it may facilitate AUX1 polar trafficking by posttranslationally modifying AUX1, causing it to be recognized as cargo destined for a particular plasma membrane face of the plant cell. However, validation of this or other mechanisms awaits further experimentation.

Supporting Online Material

www.sciencemag.org/cgi/content/full/312/5777/1122847/DC1

Materials and Methods

SOM Text

Figs. S1 to S8

References

References and Notes

View Abstract

Navigate This Article