The axr6 mutants of Arabidopsis thaliana define a gene involved in auxin response and early development

The plant hormone auxin (indole-3-acetic acid) has been shown to regulate a wide variety of plant processes, including tropisms, vascular development and formation of lateral shoots and roots (for reviews see Aloni, 1995; Malamy and Benfey, 1997; Chen et al., 1999). At the cellular level, auxin affects elongation, division and differentiation, but the mechanisms by which it produces these cellular effects remain poorly understood. The auxin-binding protein ABP1 appears to function as an auxin receptor regulating cell elongation (Jones et al., 1998); roles as receptors for other auxin-binding proteins have not been demonstrated. The products of the AXR1 (Leyser et al., 1993) and TIR1 genes (Ruegger et al., 1998) are likely to regulate auxin response through controlling protein modification by ubiquitin or ubiquitin-like molecules (del Pozo et al., 1998). Possible targets of such ubiquitin- or ubiquitin-like modification are the proteins encoded by the Aux/IAA family of genes (Abel et al., 1994), short-lived, nuclear-localized proteins whose transcription is induced by auxin and which may encode transcriptional regulators (Abel and Theologis, 1996). The related Auxin Response Factor (ARF) genes encode proteins demonstrated to bind to and regulate transcription at auxin- responsive promoters (Ulmasov et al., 1999). The importance of these proteins in auxin response has been demonstrated by genetic approaches: three Arabidopsis mutants with defects in auxin-related processes, auxin-resistant3 (Rouse et al., 1998), short hypocotyl2 (Tian and Reed, 1999), and monopteros (Hardtke and Berleth, 1998) were found to have defects in Aux/IAA or ARF genes. There is also some evidence that the auxin signal transduction cascade may include components of a mitogen-activated protein kinase (MAPK) cascade (Kovtun et al., 1998; Mizoguchi et al., 1994), although not all experiments supports this model (Tena and Renaudin, 1998).

The factors controlling embryonic development in plants are still poorly understood, but evidence suggests that short- range or direct cell-cell interactions may play important roles (Laux and Jürgens, 1997). Putative auxin gradients established by auxin polar transport (Lomax et al., 1995) may be important for establishing bilateral symmetry and polarity during embryonic development (Schiavone and Cooke, 1987; Liu et al., 1993; Fischer and Neuhaus, 1996; Hardtke and Berleth, 1998), although the only direct evidence that auxin acts as a positional morphogen comes from studies of cambial development (Uggla et al., 1998). Auxin has been further implicated in embryonic pattern formation by characterizaton of three embryonic mutants. fass mutants (Torres-Ruiz and Jürgens, 1993) were shown to have defects in auxin homeostasis (Fisher et al., 1996). Homozygous monopteros mutants, with mutations in an Auxin Response Factor gene (Hardtke and Berleth, 1998), show severe defects in development of the basal region of the embryo, resulting in seedlings that sometimes possess only a single cotyledon after germination, and lack roots and hypocotyl; the mutants also show dramatically aberrant vascular development (Berleth and Jürgens, 1993; Przemeck et al., 1996). The recently isolated bodenlos mutant has phenotypic similarities to monopteros and a reduced response to auxin in a tissue culture assay (Hamann et al., 1999).

In the course of a series of screens for auxin-resistant mutants, we isolated two mutants in a novel gene called AXR6. The unique phenotype of these mutants, described in this paper, indicates that the gene is likely to be important for auxin response in all parts of the plant, and supports a role for auxin in pattern formation and cell determination during embryogenesis.

Arabidopsis thaliana plants were grown at 20-23°C on the soilless mixture MetroMix 360 (GraceSierra) with continuous light at a photon fluence rate of 70-120 μmol/m2/second. To score for seedling resistance to hormones and antibiotics and to perform root growth and gravitropism assays, seedlings were grown under sterile conditions on Petri plates containing nutrient solution (Wilson et al., 1990) supplemented with 7 g/l agar and 10 g/l sucrose (minimal medium). Hormones or antibiotics (Sigma) were added after the solution was autoclaved. Seeds were surface sterilized for 15-25 minutes in 30% (v/v) bleach, 0.01% Triton X-100, washed 3 times with sterile water, and spread using 0.7% agar. Seeds were cold-treated for at least 2 days at 4°C before plates were placed in a CU-32L growth chamber (Percival Scientific) at 20-22°C on a 16-hour light/ 8-hour dark cycle with a photon fluence rate of 70-100 μmol/m2/second (standard conditions). The day of placing into the growth chamber is considered day 0. Auxin sensitivity or resistance was routinely scored on day 7 of growth on minimal medium containing 8×10−8 M 2,4- dichlorophenoxyacetic acid (2,4-D). When using segregating populations of seed (as in the assays of heterozygous physiology and morphology), the genotype of viable plants (heterozygous axr6 or wild type) was determined by one of the following criteria: auxin resistant root growth of seedlings, rosette morphology, and testing progeny for auxin resistance and presence of homozygous mutants. The tetraploid plant CS3432 was obtained from the Arabidopsis Biological Resource Center.

Columbia ecotype seeds (10,000) were allowed to imbibe overnight (∼18 hours) in water, and then were treated with 10 mM diepoxybutane (DEB; Sigma) for 5 hours. Seeds were washed extensively with water and planted in two independent populations in flats at 5000 seeds/flat. M2 seeds were harvested in bulk.

Approximately 25,000 seed from each independent DEB- mutagenized populations were sterilized, spread on plates containing 1×10−7 M 2,4-D, cold-treated for 2 days, and then screened for seedlings with long roots up to the 9th day after placing in the growth chamber. 36 putative auxin-resistant mutants were isolated and their M3 progeny retested. After back-crossing and complementation tests (data not shown), nine of these mutants were found to be allelic to aux1 (Maher and Martindale, 1980; Pickett et al., 1990), two are allelic to axr1 (Lincoln et al., 1990), one is the axr3-3 allele (Leyser et al., 1996), and two are alleles of ein2 (Guzman and Ecker, 1990; Masucci and Schiefelbein, 1996). The initial isolates of the axr6 mutants were backcrossed to wild type at least 5 times (axr6-1) or 6 times (axr6-2) before genetic and physiological characterization.

Heterozygous axr6 mutants were crossed to Landsberg erecta plants and auxin-resistant F1 plants identified. DNA was isolated from individual auxin-sensitive F2 plants and scored for the set of 13 ARMS markers (Fabri and Schäffner, 1994). No linkage was found to markers other than m326 (map position 4:49.1; 11/96 recombinant F2 chromosomes) and m448a (map position 4:18.7; 4/96 recombinant F2 chromosomes; genome-www3.stanford.edu/cgi- bin/Webdriver?MIval=atdb_rimap_con). The AXR6 gene thus appears to be 4/96 =4.2 cM distal from m448a. The map distance between axr6 and prl was determined to be ∼7.3 cM by analysis of 602 F2 seedlings from a cross between axr6-2 and prl (Cesarski, 1997).

Siliques from wild-type and heterozygous mutant plants were prepared for whole-mount visualization as in Berleth and Jürgens (1993). Ovules were photographed using differential interference contrast optics on a Zeiss Axiophot microscope.

Homozygous seedlings were identified by their morphology after germination on hormone-free minimal medium, fixed with ethanol:acetic acid: water (6:3:1), and then cleared and mounted in Hoyer’s solution. Vascular patterns were photographed using dark- field optics on a Zeiss Axiolab microscope (×2.5, ×5 and ×10 objectives). Photographs were projected and seedlings measured using a map tool, calibrated with a photograph of a stage micrometer.

Root growth assays were performed essentially as described by Hobbie and Estelle (1995) and Wilson et al. (1990). Data are presented as the elongation on the hormone relative to the elongation on hormone-free medium.

Assays of hypocotyl elongation were performed by cold-treating sterilized seeds for 3-5 days and then spreading directly onto minimal medium plates without sucrose and with or without IAA, exposing to room light for 45 minutes, and then placing in a light-tight box in the growth chamber. After 7 days, seedling hypocotyl length was measured.

Lateral root growth and root elongation were measured daily under a dissecting microscope on seedlings that had been germinated on hormone-free medium in a vertical orientation and transferred on day 3 to large (150 mm) plates containing minimal medium, the root tip marked, and the plates replaced in the growth chamber in a vertical position.

Seeds were placed in 5-inch clay pots (4 plants/pot), cold-treated for 9 days, and then placed under constant illumination. Pots were watered with 1× nutrient solution (Wilson et al., 1990) at 7, 13 and 24 days, at other times as needed with tap water. Bolting time was analyzed on plants grown in a similar manner. Bolting was defined as the appearance of an inflorescence at least 5 mm high.

On day 3 of growth on hormone-free medium in a vertical orientation, seedlings were transferred to square plates of hormone- free medium and returned to the growth chamber in vertical orientation. After two additional days of growth, plates were rotated by 90 degrees, photographed, and then photographed at the indicated times thereafter. Negatives were scanned and root positions traced in Adobe Photoshop; the angle of root tip reorientation was measured by comparing the 0 time point with each subsequent photograph.

Plants containing the BA3-GUS construct (Oono et al., 1998; obtained from A. Theologis) were crossed to axr6 mutant plants and lines homozygous for the BA3-GUS construct and containing the axr6 mutant allele or AXR6 wild-type allele were identified by testing F3 families for kanamycin resistance and auxin response. GUS expression was analyzed by a modified version of the procedure of Oono et al. (1998). After the indicated growth on plates with or without auxin, seedlings were washed twice and then vacuum- infiltrated for 30 minutes with buffer containing 1 mM 5-bromo-4- chloro-3-indolyl β-D-glucuronide (LabScientific), 10 mM NaPO4, pH 7, 2 mM EDTA, 0.1% TritonX-100, then incubated at 37°C in the dark for the indicated times, mounted in 50% glycerol, and photographed.

Seedlings were grown on hormone-free minimal medium for 7 days, and then axr6 homozygotes or excised wt cotyledons were transferred to MSN medium consisting of 1× Murashige-Skoog salts (Sigma), pH 5.7, 1× Murashige-Skoog vitamins (Sigma), 0.5% agar, 0.5% sucrose, and 64.5 μM naphthaleneacetic acid and grown in standard conditions for 32 days.

Photos were taken using Kodak Ektachrome T160 color slide film or Agfa APX25 black-and-white film and scanned using a Nikon LS- 1000 35 mm film scanner. Adobe Photoshop 4.0 was used to process photographs and assemble figures.

ANOVA statistical analysis used the Analysis ToolPak of Microsoft Excel 5.0 for Macintosh.

In an effort to identify new mutants affected in auxin response, diepoxybutane-mutagenized Arabidopsis seeds were screened for auxin-resistant root growth on nutrient agar medium containing 1×10⁻⁷ M of the synthetic auxin 2,4-D (see Materials and Methods). Two mutants with similar novel phenotypes, described below, were isolated from independent populations; the mutants subsequently proved to be produced by alleles of the same gene, and were named auxin-resistant6-1 (axr6-1) and axr6-2.

Seedlings derived from self-fertilization of the auxin-resistant axr6 plants fell into three classes: wild type auxin-sensitive, mutant auxin-resistant, and rootless, in approximately a 1:2:1 ratio (Table 1). Crosses to wild-type plants demonstrated that the auxin-resistant plants are heterozygous for the axr6 mutation (Table 1); the rootless seedlings are the axr6 homozygotes. Crosses between axr6-1 and axr6-2 heterozygotes gave 1/4 rootless progeny, demonstrating allelism (Table 1). A reduced frequency of homozygotes in the progeny of axr6-1 heterozygotes was shown not to be due to reduced transmission of the mutant allele, and thus probably results from reduced viability of the homozygotes during embryogenesis or germination (Table 1). Crossing the axr6-2 mutant to a tetraploid plant produced triploid seedlings (axr6-2/AXR6/AXR6) that were auxin resistant (Table 1); this suggests that at least this allele is a gain-of-function mutation. No other mutants with the homozygous rootless morphology that also show a mutant phenotype in the heterozygote have been reported.

The location of the AXR6 gene was determined by a combination of molecular and classical mapping approaches (see Materials and Methods). The AXR6 gene maps to the interval between m448a and PRL on the short arm of chromosome 4. No other auxin-resistant or seedling-lethal mutations map to this region (genome-www3.stanford.edu/cgi- bin/AtDB/). The MONOPTEROS and BODENLOS genes, which have mutant phenotypes similar to those of axr6 homozygotes, both map to chromosome 1 (Hardtke and Berleth, 1998; Hamann et al., 1999). Therefore, the axr6 mutants define a novel gene.

Wild-type seedlings have two cotyledons (seed leaves) connected by the hypocotyl to the primary root; by 14 days, they develop true leaves and lateral roots (Fig. 1).

Homozygous axr6 seedlings display a dramatically different morphology: one or two cotyledons are attached to a small basal stub or peg of whitish tissue. The axr6-1 homozygotes generally produce only a single cotyledon, but are occasionally found with two cotyledons (Fig. 1; Table 1). The axr6-2 homozygotes generally produce two cotyledons but are very rarely found with just a single cotyledon, and may sometimes even produce a few tiny true leaves (Fig. 1; Table 1). The homozygous seedlings of either allele with two cotyledons usually show a normal 180 degree angle between the two cotyledons; an unusual axr6-1 exception is shown in Fig. 1. Rarely, homozygous seedlings with exceptional morphologies are found; these include irregularly shaped single cotyledons and two partially separated cotyledons. Homozygous seedlings also frequently develop a purplish color assumed to be anthocyanin accumulation.

Auxin’s importance for vascular development is well known (Aloni, 1995), and consistent with the abnormal vascular patterns found in the cotyledons of the mp mutants (Berleth and Jürgens, 1993). Fig. 2 compares the vascular development in wild type and in homozygous axr6 mutant seedlings at 7 days of growth. Whereas wild-type cotyledons have a single central strand and at least one semicircular strand running around the periphery of each half of their cotyledons, the homozygous axr6 mutants have a variety of incomplete vascular patterns (Fig. 2A-F). Most, but not all, of the homozygous axr6 seedlings have a complete central strand, but the peripheral strands are in general partial and incomplete, and in some cases consist of only a short segment completely detached from the central vasculature. Quantitative analysis of the vasculature in mutant and wild-type seedlings revealed the following: the frequency of the most severely defective patterns was higher in axr6-1 than in axr6-2 (Fig. 2G), both central and peripheral vein length showed significant decreases in both mutants, compared to wild type,

and the reduction in length of peripheral veins was much greater than that of the central vein in the mutants (Fig. 2H). Similar defects were seen in seedlings of different ages (data not shown), indicating that the mutant phenotypes reflect defects rather than mere delays in development. To summarize, seedlings homozygous for the axr6-1 and axr6-2 mutations displayed dramatic deficiencies in vascular development.

Stained cross sections of the basal stub of the axr6 homozygotes revealed that it was composed of large morphologically undifferentiated cells (not shown), and was very similar to the basal stub of monopteros homozygotes (Berleth and Jürgens, 1993). Occasionally root hairs were present at the very tip of the stub (not shown), as seen in bodenlos homozygotes (Hamann et al., 1999).

The variety of abnormalities found in the axr6 homozygous seedlings after germination suggested that these defects must originate during embryonic development. To determine how embryonic development of the axr6 homozygotes deviates from that of wild type, the morphology of embryos in siliques from heterozygous axr6 plants was analyzed by differential interference contrast microscopy of cleared whole embryos at different stages. Fig. 3 presents a developmental series of control wild-type and homozygous axr6 embryos, showing some of the patterns that were consistently seen in the mutants.

The wild-type zygote divides asymmetrically to give rise to a small cytoplasmically dense apical cell and an elongated basal cell. The apical cell divides in a regular pattern to produce most of the embryo proper; the basal cell divides transversely several times, yielding an elongated single-cell-wide suspensor which connects the embryo to the maternal tissue (Fig. 3A,D,G,J,M). The most striking differences in the early embryogenesis of the axr6 homozygotes are in the divisions of the suspensor, which divides periclinally one or more times before the normal anticlinal orientation of divisions resumes. The earliest abnormal division in axr6 was seen at the second cell division (Fig. 3B). However, it is possible that an abnormality at the first cell division stage would have been missed because of the small number of embryos of this stage that were examined. The periclinal divisions of the suspensor cells in the mutants result in a suspensor that is two or more cells wide along part or most of its length (Fig. 3F,H,I,K,N). In wild type, the uppermost cell of the suspensor gives rise to the hypophyseal cell, the lens- shaped cell at the base of the embryo proper which is the precursor of the root cap and the root quiescent center (Fig. 3G; Scheres et al., 1994). In the axr6 homozygotes the hypophyseal cell is not produced, presumably as a result of the abnormal suspensor development.

In the wild-type embryo, the pattern of divisions is synchronized, regular and reproducible, resulting in an embryo that at the triangular stage has an outer layer of protodermal cells and inner cells organized in three tiers (Fig. 3G). At slightly later stages, central files of cells can be seen, the precursors of the vascular tissue (Fig. 3J,M). In the homozygous mutant embryos, the divisions are frequently asynchronous and irregular in pattern, giving embryos with unusual cell numbers (Fig. 3C,E) and misaligned or disorganized cell patterns (Fig. 3I,L).

During later stages in wild-type embryos, increased cell division on both flanks of the upper tiers leads to development of the cotyledons. This was seen only on one flank in the axr6-1 homozygotes (not shown), giving rise to the characteristic one cotyledon morphology (Fig. 3L).

Thus, the axr6 homozygotes can be inferred to show abnormal orientation and timing of cell divisions starting early in development, and affecting especially those cells in the suspensor and the basal half of the embryo. There does not appear to be just a single division that is abnormal in the mutants; rather, throughout development the divisions in the embryo deviate from the wild-type pattern.

Arabidopsis produces the primary root meristem during embryogenesis, and produces other roots post-embryonically (lateral and adventitious), and from callus in tissue culture. Although the axr6 homozygous mutant seedlings lack roots, we found that axr6 homozygous callus developed morphologically normal roots when grown on 64.5 μM naphthaleneacetic acid (not shown). Wild-type tissue remained as callus under identical conditions. Thus, the axr6 mutations, when homozygous, do permit root development in tissue culture and show a difference in hormonal response from wild type.

To characterize the effects of the axr6 mutations more fully, we analyzed the morphology, auxin responses, and gravitropism of the axr6 heterozygotes.

Compared to wild type, rosette leaves of heterozygous axr6 plants are shorter and more wrinkled (Fig. 4A), and the plants are shorter and produce more inflorescences (Fig. 4B; Table 2). The siliques of the axr6 mutants lie at a more acute angle relative to the stem than those of wild-type Columbia plants (Fig. 4C).

To assess quantitatively the level of resistance to auxins and other hormones in the axr6 heterozygotes, the root elongation of seedlings grown on different concentrations of various hormones was determined. The axr6 mutant seedlings were found to be about two-fold to four-fold more resistant to the synthetic auxin 2,4-D compared to wild type (Fig. 6A). The axr6 mutants were also resistant to the synthetic auxin NAA (Fig. 6B) and the natural auxin IAA (not shown), but did not show reproducible and signficant differences from wild type in their growth responses to the ethylene precursor ACC, to abscisic acid, or to kinetin (a cytokinin) (Fig. 6B). These results indicate that the axr6 mutations produce auxin resistance but do not appear to cause cross-resistance to other tested plant hormones.

The axr6 mutations also affected hypocotyl growth in dark- grown seedlings. On hormone-free medium, the hypocotyls of dark-grown axr6-1 and axr6-2 heterozygous seedlings were slightly but significantly longer than those of wild type: wild type, 17.5±0.17 mm; axr6-1, 19.3±0.31 mm; axr6-2, 18.5±0.41 mm (n≥40 for each genotype; P for axr6-1 <<0.001, P for axr6-2 <0.005). The elongation of the heterozygous mutant hypocotyls was more resistant to inhibition by IAA in the medium than was hypocotyl elongation of wild-type seedlings (Fig. 6C). Thus the mutants are auxin-resistant in both root and hypocotyl elongation, and show small intrinsic differences in hypocotyl elongation in the dark.

Auxin is important for root development and lateral root formation, as shown both by physiological experiments and by alterations in these processes seen in other auxin-related mutants (Boerjan et al., 1995; Hobbie and Estelle, 1995; Malamy and Benfey, 1997). Seedlings heterozygous for the axr6 mutations, when grown on hormone-free medium, were found to have slightly but signficantly increased rates of root elongation (Fig. 5A) and substantially decreased lateral root growth (Fig. 5B) over a 9-day period.

Auxin is believed to be important in the differential cell growth necessary for gravitropic response; other auxin-resistant mutants have been found to be altered in their response to gravity (reviewed in Chen et al., 1999). The axr6 heterozygous mutants were tested for such defects by determining the rate of root tip curvature in seedlings grown on hormone-free medium in the light after the seedlings were reoriented from vertical to horizontal. The results show that the axr6 mutant seedlings are indeed substantially delayed in their gravitropic response (Fig. 7).

A number of other auxin-resistant mutants have been found to have changes in the expression or inducibility of auxin- regulated genes (e.g. Abel et al., 1995; Timpte et al., 1994). To test for such differences in axr6-1 heterozygotes, the β- glucuronidase (GUS) gene under the control of a synthetic auxin-regulated promoter (Oono et al., 1998) was introduced into lines carrying the mutant or wild-type alleles of AXR6 (see Materials and Methods). In the absence of hormone treatment, wild-type seedlings displayed weak expression of GUS in the cotyledons and hypocotyl (Fig. 8A and C), as well as at the hypocotyl-root junction and in the root elongation zone (not shown). axr6-1 heterozygous seedlings showed dramatically increased expression in the the first true leaves and in the cotyledons, in particular at the tips (Fig. 8B and D). axr6-2 seedlings showed an expression pattern that was overall very similar to that of the wild-type seedlings (not shown). After 2 days’ growth on plates containing 8×10⁻⁸ M 2,4-D, to induce expression of the reporter gene, wild-type seedlings showed high levels of GUS expression localized to the root elongation zone (Fig. 8E), as observed previously (Oono et al., 1998). axr6-1 heterozygous seedlings, by contrast, showed only weak and diffuse expression of GUS in the root under these conditions (Fig. 8F). The axr6-1 mutants therefore show reduced auxin inducibility of an auxin-regulated gene in their roots, along with a strikingly altered pattern of expression in the apical regions of the seedlings in the absence of auxin induction.

We have described the isolation and characterization of two mutant alleles of a novel gene, AXR6. The axr6 mutants show a unique combination of phenotypes: heterozygous mutant plants are viable but altered in a variety of auxin-related processes, whereas homozygous mutants arrest growth after germination due to dramatic defects in embryonic development. The phenotype of the heterozygous axr6 mutants directly implicates defective auxin response as the cause of the abnormal embryonic development of the axr6 homozygotes.

The axr6 homozygous embryos show altered patterns of cell division starting almost as early as could be observed, at the second cell division (Fig. 3B). These abnormal divisions in the early embryo especially affect the distinction between the cells of the embryo and the suspensor, giving rise to unusual suspensors that are several cells wide and not clearly delineated from the embryo proper. Other defects in the embryonic development of the axr6 homozygotes include failure to generate the vascular precursors (procambial cells), failure to undergo the cell proliferation on one upper flank of the embryo that normally produces one of the cotyledons (in the axr6-1 homozygotes), and failure to differentiate vasculature in the cotyledons. As much other evidence suggests that morphogenesis in plants is regulated at the regional or organ level rather than by control of cell division patterns (Smith et al., 1996; Traas et al., 1995), the aberrant timing and orientation of cell divisions are not likely to be the primary defects in axr6 homozygotes, but may instead reflect underlying defects in auxin-mediated positional signaling or cell determination.

The axr6-1 allele is stronger than the axr6-2 allele, a difference seen in both homozygotes and heterozygotes. The axr6-1 homozygotes generally have only a single cotyledon, versus two for axr6-2, and show more severe vascular defects; the axr6-1 heterozygotes are shorter, bushier, more auxin-resistant, and more altered in lateral root formation and gravity response than axr6-2 heterozygotes. The axr6 mutations are likely to be gain-of-function, as they are dominant in heterozygotes and the axr6-2 mutation is dominant in triploid seedlings of genotype axr6/AXR6/AXR6. Although both dominant (axr2: Wilson et al., 1990) and semidominant (tir1: Ruegger et al., 1998; axr3: Leyser et al., 1996; shy2: Tian and Reed, 1999) auxin-related mutations have been described, to our knowledge this is the first description of auxin-related mutants showing such dramatic differences in phenotype between the heterozygote and homozygote. Isolation and characterization of loss-of-function alleles of axr6, perhaps in screens for revertants, will give further insights into the normal physiological and developmental role of the AXR6 gene.

The importance of auxin in plant embryonic development has been previously demonstrated by the disruptions of normal developmental patterns produced by auxin-related compounds (Schiavone and Cooke, 1987; Fischer and Neuhaus, 1996; Hadfi et al., 1998) and by the alterations in auxin physiology found in several embryonic mutants of Arabidopsis, and is further supported by our characterization of axr6 mutants. The phenotypic similarities among the monopteros (Berleth and Jürgens, 1993), bodenlos (Hamann et al., 1999), and homozygous axr6 mutants (this paper) are so striking as to suggest that the three genes could act in a common pathway of auxin response during embryonic development. Eventual molecular characterization may help clarify the function of the AXR6 gene in embryogenesis and in auxin response in general.

We thank A. Wilson for DEB-mutagenized seed, T. Cesarski for the axr6 × prl mapping, A. C. Churchill, P. Springer, R. Martienssen, J.

Simorowski, J. P. Vielle-Calzada, and J. Malamy for advice; D. Fernandez, T. Short, P. Benfey and two reviewers for comments on the manuscript, A. Theologis for the BA3/GUS line, M. K. Barton for generously resharing the axr6 mutants, the Arabidopsis Biological Resource Center for molecular markers, and the staff of the Adelphi Biology Dept. (S. Greenlee, V. Briguglio, S. Komansky, F. Juarez) for help. L. H. thanks former colleagues in the Estelle lab for support, encouragement and many useful discussions. This work was begun at Indiana University with support from the NSF and the NIH, and continued at Adelphi University with support from the NSF (IBN- 9632016 to L. H.).