Gravity plays a major role in plant morphogenesis by determining

Gravity-regulated differential auxin transport from columella to lateral root cap cells Iris Ottenschläger*, Patricia Wolff*, Chris Wolverton, Rishikesh P. Bhalerao,Göran Sandberg, Hideo Ishikawa, Mike Evans, and Klaus Palme* *Institut für Biologie II, Universität Freiburg, D-79014 Freiburg, Germany; Department of Plant Biology, Ohio State University, Columbus, OH 43210; and Department of Forest Genetics and Plant Physiology, Umeå Plant Science Center, Swedish University of Agricultural Sciences, S-901 83 Umeå, Sweden Communicated by Josef S. Schell, Max Planck Institute for Plant Breeding Research, Cologne, Germany, December 26, 2002 (received for review November 10, 2002) Gravity-induced root curvature has long been considered to be regulated by differential distribution of the plant hormone auxin. However, the cells establishing these gradients, and the transport mechanisms involved, remain to be identified. Here, we describe a GFP-based auxin biosensor to monitor auxin during Arabidopsis root gravitropism at cellular resolution. We identify elevated auxin levels at the root apex in columella cells, the site of gravity perception, and an asymmetric auxin flux from these cells to the lateral root cap (LRC) and toward the elongation zone after gravistimulation. We differentiate between an efflux-dependent lateral auxin transport from columella to LRC cells, and an effluxand influx-dependent basipetal transport from the LRC to the elongation zone. We further demonstrate that endogenous gravitropic auxin gradients develop even in the presence of an exogenous source of auxin. Live-cell auxin imaging provides unprecedented insights into gravity-regulated auxin flux at cellular resolution, and strongly suggests that this flux is a prerequisite for root gravitropism. gravitropic root curvature polar auxin transport auxin carrier proteins Gravity plays a major role in plant morphogenesis by determining the directional growth of plant organs (gravitropism). Roots orient at a preferred angle with respect to gravity [their gravitropic set-point angle (GSA); ref. 1], allowing efficient exploration of the soil (root gravitropism). Main roots of Arabidopsis seedlings, for instance, have a GSA of 0 and grow parallel to the gravity vector. Changes in gravity vector orientation (gravistimulation) induce root curvature, resulting in realignment of the root tip to the GSA. Root curvature is a consequence of gravity signal perception, involving amyloplast sedimentation in the columella cells of the root cap (2), and differential growth induced on opposite flanks in the elongation zone (EZ). In the 1920s, the Cholodny Went hypothesis and various interpretations of it ever since have proposed that this differential growth within the EZ is mediated by an asymmetric distribution of the plant hormone auxin (3). Supportive evidence for an auxin asymmetry in the EZ after gravistimulation has come from the analyses of radio-labeled auxin distribution, or differential induction of auxin-response promoters (4). It has been questioned, however, whether auxin gradients are necessary or sufficient to cause root gravitropism (3, 5). Furthermore, it is not clear as to how the gravisensing events in the columella cells can give rise to changes in auxin concentration in the EZ. Recently, the gravity-dependent relocation of an auxin efflux carrier protein in columella cells suggested gravity-regulated changes of auxin transport right at the site of gravity perception in the root cap (6). However, differential auxin fluxes through the cap cells and their contribution to gravitropic root curvature remain to be demonstrated. In the work presented here, we applied a GFP-based auxin biosensor to study gravity-induced auxin fluxes and their transport mechanisms in vivo and on a cellular level. Methods The DR5-GFP Construct. Specific primers were used to amplify the synthetic auxin-response promoter DR5 [kindly provided by T. Ulmasov and T. Guilfoyle (Department of Biochemistry, University of Missouri, Columbia) as a DR5-GUS fusion in a pck vector background]. The auxin-response promoter DR5 consists of 9 inverted repeats of the 11-bp sequence 5 -CCT- TTTGTCTC-3, a 46-bp CaMV35S minimal promoter element, and a TMV leader sequence (7). GFPm was generated by fusing DNA sequences encoding for the endoplasmatic reticulum basic chitinase target signal and HDEL retention signal derived from mgfp5-er (8) to the GFP-LT coding region (kindly provided by G. Jach, Max Planck Institute for Plant Breeding, Cologne, Germany). The amino acid sequence of GFP-LT corresponds to the amino acid sequence of enhanced GFP, commercially available from CLONTECH (G. Jach, unpublished results). Plant Material. Arabidopsis thaliana Columbia-0 plants were transformed with the DR5-GFPm construct. Single-locus insertion lines were selected in T2. Homozygous T3 was used for all experiments described. A. thaliana Columbia-0 were also used for curvature kinetic measurements. Eir1-1 plants were transformed with the DR5-GFPm construct. Single-locus insertion lines were selected in T2 and analyzed. Plant Growth Conditions. Seeds were surface-sterilized as described (9) and sown on solid AM medium (2.3 g/liter MS salts 1% sucrose 1.6% agar agar (ph 6.0) with KOH). After vernalization in the dark for 3 days at 4 C, seeds were germinated as described (9). For microscopic analyses 12 h before imaging, seedlings were transferred to microscope slides covered with a thin layer (1 mm) of AM medium containing 0.8% agarose and supplemented with auxins and auxin transport inhibitors, respectively. For each treatment, 20 40 seedlings were examined in independent experiments. Seedlings on microscope slides were gravistimulated by rotating the stage to 135. Analysis of Indole-3-Acetic Acid (IAA) Contents. Analysis was performed as described (10). Imaging. For better resolution, root tissue was stained with 10 M propidium iodide before microscopy. Fluorescent signal detection was performed by using a confocal laser scanning (CLS) microscope (Leica DMIRBE, TCS 4D with digital imaging Abbreviations: EZ, elongation zone; IAA, indole-3-acetic acid; CI, columella initial; QC, quiescent center; 1-NAA, 1-naphthyl-acetic acid; 2,4-D, 2,4-dichloro-phenoxyacetic acid; NPA, 1-naphthylphthalamic acid; 1-NOA, 1-naphthoxyacetic acid; BFA, brefeldin A; LRC, lateral root cap; dlrc, distal LRC; plrc, proximal LRC. I.O. and P.W. contributed equally to this work. Present address: Department of Botany Microbiology, Ohio Wesleyan University, Delaware, OH 43015. To whom correspondence should be addressed. E-mail: klaus.palme@biologie.unifreiburg.de. www.pnas.org cgi doi 10.1073 pnas.0437936100 PNAS March 4, 2003 vol. 100 no. 5 2987 2991

Fig. 1. Changes in gravity vector orientation induce asymmetric expansion of DR5-GFPm signal in LRC cells. (A) Schematic representation of the Arabidopsis root apex (after refs. 28 and 29). The color code labels different tissues (top to bottom: stele, pericycle, endodermis, cortex endodermis initial, cortex, epidermis, plrc, dlrc, QC, CI, and columella). (B) DR5-GFPm expression in QC, CI, and columella of vertically grown roots. (C) Faint GFPm signals appear in dlrc cells neighboring columellas S2 and S3 (indicated by arrows) in roots after 1.5 h of gravistimulation. (D) After 3hofgravistimulation, the DR5-GFPm signal expands from columella to complete lower half of the LRC. (E) A 15-min gravistimulus is sufficient to induce DR5-GFPm signal expansion from columella to LRC cells. Note that the lower half of gravistimulated roots is at the right side (C E). (Bar 20 m.) processing) using a 530 15-nm band pass filter for GFPm detection and a 580 15-nm band pass filter for detection of propidium iodide and tissue autofluorescence. For histological signal localization both images were electronically overlaid and further processed with PHOTOSHOP (Adobe Systems, Mountain View, CA). Curvature Measurements. Kinetic measurements of root gravitropic curvature were done by using automated root image analysis software as described (11). (for up to 24 h) by further rotation led to an increase in signal intensity, but the staining pattern did not expand to other tissues. No signals were observed in the EZ; this obviously deviates from the model of an auxin-dependent differential elongation in the EZ as proposed by the Cholodny Went hypothesis. Whether auxin gradients did not extend to the EZ or whether they were below the threshold required for detection of DR5-GFPm ac- Results and Discussion DR5-GFPm Expression Identifies Elevated Auxin Levels in Columella Cells of the Root Cap. We developed a fluorescent biosensor to monitor relative auxin contents in root tips of living Arabidopsis seedlings. Local auxin accumulation was inferred from the expression of an endoplasmatic reticulum-targeted GFP (GFPm) driven by the synthetic auxin-response promoter DR5 (7). DR5-GFPm reflects relative auxin levels exceeding a certain threshold and allows monitoring of auxin responses at cellular resolution by the use of CLS microscopy. Changes in auxininduced GFPm expression can be detected with a time lag of 1.5 h (data not shown), the time required for GFPm maturation (12). Transgenic plants carrying DR5-GFPm displayed wild-type phenotype, normal phototropic, and gravitropic responses, and expressed GFPm in equivalent tissues as was reported for DR5-GUS (13) (data not shown). In 6-day-old seedlings with vertically grown roots, images of root caps typically showed a single story of columella initial (CI), three horizontal stories (S1 S3), and four vertical files of columella cells (Fig. 1A). GFPm appeared in specific stele cell files and was highly localized in the quiescent center (QC), CI, and mature columella (Fig. 1B), indicating elevated auxin levels in these cells. Gravistimulation Rapidly Causes Changes in the DR5-GFPm Expression Pattern. To analyze the effect of gravity on auxin distribution, we gravistimulated roots by a 135 rotation from the vertical and imaged samples at different time points. Changes in DR5-GFPm expression pattern were first observed after 1.5 h. Weak fluorescent signals appeared on lower halves of gravistimulated roots, in distal lateral root cap (dlrc) cells next to columella S2 (Fig. 1C). Slightly later, signals were also detected in dlrc cells adjacent to columellas S1 and S3, resulting in staining of the entire LRC along the lower root half of the roots after 3 h (Fig. 1D). Concomitant chronological and spatial development of asymmetric DR5-GFPm signal implies an auxin flux from columella cells to the LRC. Maintaining the lateral gravistimulus Fig. 2. Effect of exogenous auxin application on vertically grown and gravistimulated roots. Auxin biosensor signal in roots grown on 1 M IAA (A C), 1 M 1-NAA (D F), 1 M 2,4-D (G I) in vertically grown roots (A, D, and G), and after 5h(B, E, and H) and 24 h (C, F, and I) of gravistimulation, respectively. Images are aligned with root tips to the vertical for better comparison. Note that lower half of gravistimulated roots is at the right side (B, C, E, F, H, and I). (Bar 20 m.) 2988 www.pnas.org cgi doi 10.1073 pnas.0437936100 Ottenschläger et al.

Fig. 3. Root curvature correlates with asymmetric auxin biosensor distribution. (A) Total curvature 5 and 20 h (black and gray bars, respectively) after gravistimulation at 135. Kinetic measurements of root gravitropic curvature were done by using automated root image analysis software as described (11). Values represent means SE; n 7 10 for all treatments. (B) DR5-GFPm signal asymmetry in LRC and epidermal cells of 1 M 1-NAA-treated roots, gravistimulated for 24 h. (Bar 20 m.) tivity could not be deduced. When root orientation toward the gravitropic set-point angle was complete, lateral GFPm signal faded within 4 h, indicating a decay of auxin levels. Gravitropic curvature can be induced by short periods of gravistimulation. To investigate the role of auxin as a signal linking gravity perception and root curvature, we determined the ability of short-term gravistimulation [in the range of 5 min, the time when amyloplasts sediment (14), and 20 min, the latent period between gravistimulation and detectable root curvature (15)] to ultimately induce auxin asymmetry. We took into account the time lag for the detection of signal changes due to GFPm maturation, and took advantage of the GFPm protein stability (16). Roots were therefore exposed to short-term gravistimulation by rotating the stage 135 from the vertical, holding it in that position for 15 min, and then rotating it back to the vertical for 3 h. This procedure allowed delayed detection of earlier changes in DR5-GFPm activity by permitting GFPm maturation over a 3-hr development period. With this approach, gravity-induced GFPm signal asymmetry within the root cap was detected as early as 15 min after short-term gravistimulation (Fig. 1E). Exogenous Auxin Application Does Not Mask Gravity-Induced DR5- GFPm Signal Asymmetry. One of the greatest weaknesses of the proposition that auxin gradients cause root gravitropism is the fact that root curvature can be observed in roots exposed to high concentrations of exogenous auxin. It was assumed that exogenous auxin treatments would mask endogenous auxin gradients and would in turn prevent the graviresponse (5). To reinvestigate this point, we tested DR5-GFPm expression in roots exposed to 1 M of various exogenous auxins before and then during gravistimulation. Besides the natural auxin IAA, a substrate for both efflux and influx carriers, we chose 1-naphthyl-acetic acid (1-NAA) and 2,4-dichloro-phenoxyacetic acid (2,4-D), substrates for efflux and influx carriers, respectively (17). Roots on 2,4-D showed DR5-GFPm signal in virtually all tissues of the root tip (Fig. 2G) and no signal alteration was detected after gravistimulation (Fig. 2 H and I). With IAA, we observed strong fluorescence in the epidermis and the entire LRC (Fig. 2A). Gravistimulation led to a shift of signal distribution within the LRC, with strong asymmetry in the dlrc and only weak signal asymmetry in the proximal LRC (plrc) (Fig. 2 B and C). 1-NAA caused staining essentially similar to IAA but less pronounced in the epidermis and the plrc (Fig. 2D). In contrast to the IAA-treated roots, gravistimulation of 1-NAAtreated roots resulted in strong signal asymmetry in the entire LRC with significantly increased fluorescence on the lower and decreased fluorescence on the upper half (Fig. 2 E and F). Our data therefore indicate that a continuous supply of exogenous auxin does not prevent formation of gravitropic auxin gradients. Moreover, in 1-NAA-treated roots, signal asymmetry between lower and upper halves continued in the EZ. Epidermal cells on the upper half showed considerably weaker fluorescence than on the lower half (Fig. 3B). We conclude that endogenous auxin levels in this region are normally below the threshold for DR5-GFPm detection but become visible when the overall auxin content in the tissue is increased by 1-NAA application. To analyze gravitropic curvature of auxin-treated roots, we investigated the kinetics of the graviresponse (Fig. 3A). Strongest downward curvature was observed in the presence of 1-NAA. Gravitropic curvature of IAA-treated roots was close to zero, despite a slight but significant downward curvature. Roots on Fig. 4. DR5-GFPm signal patterns induced by exogenous auxin application reflect differential auxin accumulation. In combination with either influx inhibitor 1-NOA or efflux inhibitor NPA, IAA-induced fluorescence patterns correspond to those induced by 1-NAA and 2,4-D, respectively (compare with Fig. 2). DR5-GFPm signals in roots grown on 1 M IAA plus 50 M 1-NOA (A), and 1 M IAA plus 10 M NPA (B), respectively. (Bar 20 m.) Ottenschläger et al. PNAS March 4, 2003 vol. 100 no. 5 2989

Fig. 6. Eir1-1 mutants have substantially elevated auxin levels in the root tip. (A) Free IAA levels of young seedling root tips determined by mass spectrometry. For each line, data were sampled from 10 measurements in two different experiments and are represented as means and SD. (B) Auxin biosensor signal reveals increased auxin levels in the plrc of eir1-1 mutant root tip. (Bar 20 m.) with the influx inhibitor 1-NOA (Fig. 4A), in fact, mimicked the DR5-GFPm expression patterns obtained by 1-NAA application (Fig. 2D). IAA applied with the efflux inhibitor NPA (Fig. 4B), on the other hand, induced patterns like those obtained after 2,4-D treatment (Fig. 2G). We therefore conclude that the DR5-GFPm expression patterns obtained after application of the auxin analogues result from a difference in accumulation of these auxins in the root tissues. Fig. 5. Effect of auxin transport inhibitor application on vertically grown and gravistimulated roots. Auxin biosensor signal in roots grown on 10 M NPA (A and B), 20 M BFA (C and D), and 50 M 1-NOA (E and F) in vertically grown (B, C, and E) and 24-h gravistimulated (B, D, and F) roots. Images are aligned with root tips to the vertical for better comparison. Note that the lower half of gravistimulated roots is at the right side (B, D, and F). (Bar 20 m.) 2,4-D showed no or even tenuous upward curvature. Reports have described the relative effectiveness of auxins IAA, 1-NAA, and 2,4-D on inhibition of root gravitropism in Arabidopsis plants (18). They identified 2,4-D as exerting the strongest disturbance on root gravitropism, whereas 1-NAA showed the weakest effects. This finding is in accordance with the data presented here. Moreover, our data correlate the degree of gravitropic root curvature with the extent of gravity-induced auxin asymmetry as reflected by the DR5-GFPm expression pattern. Remarkably, the asymmetric DR5-GFPm expression pattern is observed only in the presence of IAA and 1-NAA, which are, in contrast to 2,4-D, substrates for the auxin efflux carrier. Auxin-Induced DR5-GFPm Expression Reflects Auxin Accumulation. Treatment with the three different auxins, IAA, 1-NAA, and 2,4-D, resulted in qualitatively distinct DR5-GFPm expression patterns (see previous section). To address the argument that the observed differences may be due to DR5 sensitivity and not auxin transport and accumulation within the tissue, we compared the DR5-GFPm pattern induced by IAA in combination with either the auxin efflux inhibitor 1-naphthylphthalamic acid (NPA) or the auxin influx inhibitor 1-naphthoxyacetic acid (1-NOA). We reasoned that such combined treatments would copy the different transport characteristics of the auxin analogues. Changes in DR5-GFPm expression pattern would then point to differential auxin accumulation rather than DR5 sensitivity. We observed that simultaneous treatments of IAA together Auxin Influx and Efflux Carriers Contribute Differently to Gravity- Regulated Auxin Fluxes. To assess further the contribution of specific carriers to auxin transport during the graviresponse, we incubated roots with inhibitors of auxin efflux [NPA and brefeldin A (BFA)] and auxin influx (1-NOA), respectively. The inhibitors were applied in concentrations suitable to block root gravitropism. NPA (10 M) affected the regular auxin maximum within the root apex. GFPm signal intensity was reduced in S3 but increased in S1, CI, and QC cells. Additional signals appeared in the adjacent meristem (Fig. 5A), equivalent to NPAinduced signal expansion in DR5-GUS plants (10, 13). Gravistimulation did not induce signal asymmetry (Fig. 5B), and gravitropic curvature was impaired (Fig. 3A). We further analyzed the effect of 20 M BFA on DR5-GFPm expression. BFA was recently shown to interfere with auxin efflux by abolishing polar localization of efflux carrier AtPIN proteins (19 21). In vertically grown roots, the signal pattern on BFA (Fig. 5C) was comparable to controls (Fig. 1B). After gravistimulation, no signal asymmetry was detected (Fig. 5D), and root curvature was reduced as with NPA (Fig. 3A). Roots incubated with 50 M 1-NOA (Fig. 5E) displayed staining patterns as seen in controls (Fig. 1B). Gravistimulation induced asymmetric biosensor accumulation on the lower half of the LRC (Fig. 5F). The signal, however, was restricted to the dlrc, and could not be detected in plrc cells. Root curvature was entirely blocked (Fig. 3A). Thus, lateral transport of auxin from columella to dlrc was not visibly affected, whereas basipetal auxin transport from dlrc to plrc cells and the EZ was strongly impaired. That root gravitropism is inhibited, despite formation of an auxin gradient across the dlrc, indicates that auxin asymmetry needs to be established close to the EZ to induce gravitropic root curvature. Furthermore, these observations suggest that auxin accumulation in plrc cells and the EZ depends on auxin supplies derived from the dlrc. We conclude that lateral auxin transport in the 2990 www.pnas.org cgi doi 10.1073 pnas.0437936100 Ottenschläger et al.

root cap exclusively requires efflux carriers, whereas basipetal auxin transport depends on efflux and influx carriers. These findings are consistent with the localization of efflux carrier protein AtPIN3 (6) in the columella, and efflux carrier protein AtPIN2 (P.W., unpublished data) as well as influx carrier protein AUX1 (22) in the plrc. Increased Auxin Levels Are Detected in Root Tips of Atpin2 eir1-1 Mutants. Roots of agravitropic Atpin2 eir1-1 mutants are impaired in basipetal auxin transport (23). We investigated biosensor signals in roots of eir1-1 mutants. Vertically oriented eir1-1 roots exhibited strong signals in the dlrc cylinder with only weak extensions toward the plrc (Fig. 6B). We measured free IAA levels in the first millimeters of root tips by mass spectrometry. Compared with wild-type, mutant root tips contained 2.5 times more auxin (Fig. 6A), corresponding to enhanced DR5-GFPm expression in the mutant root cap. We conclude that a constant and even lateral auxin flow occurs from the columella to LRC cells, which in the case of Atpin2 eir1-1 mutants cannot be compensated by efficient basipetal transport. 1. Digby, J. & Firn, R. D. (1995) Plant Cell Environ. 18, 1434 1440. 2. Sack, F. D. (1997) Planta 203, S63 S68. 3. Firn, R. D., Wagstaff, C. & Digby, J. (2000) J. Exp. Bot. 51, 1323 1340. 4. Muday, G. (2001) J. Plant Growth Regul. 20, 226 243. 5. Ishikawa, H. & Evans, M. L. (1995) Plant Physiol. 109, 725 727. 6. Friml, J., Wisniewska, J., Benkova, E., Mendgen, K. & Palme, K. (2002) Nature 415, 806 809. 7. Ulmasov, T., Murfett, J., Hagen, G. & Guilfoyle, T. (1997) Plant Cell 9, 1963 1971. 8. Haseloff, J., Siemering, K. R., Prasher, D. C. & Hodge, S. (1997) Proc. Natl. Acad. Sci. USA 94, 2122 2127. 9. Müller, A., Guan, C., Gälweiler, L., Tänzler, P., Huijser, P., Marchant, A., Parry, G., Bennett, M., Wisman, E. & Palme, K. (1998) EMBO J. 17, 6903 6911. 10. Friml, J., Benkova, E., Bililou, I., Wisniewska, J., Hamann, T., Ljung, K., Woody, S., Sandberg, G., Scheres, B., Jürgens, G. & Palme, K. (2002) Cell 108, 661 673. 11. Mullen, J. L., Turk, E., Johnson, K., Wolverton, C., Ishikawa, H., Simmons, C., Soll, D. & Evans, M. L. (1998) Plant Physiol. 118, 1139 1145. 12. Verkhusha, V. V., Akovbian, N. A., Efremenko, E. N., Varfolomeyev, S. D. & Vrzheshch, P. V. (2001) Biochemistry (Moscow) 66, 1342 1351. 13. Sabatini, S., Beis, D., Wolkenfelt, H., Murfett, J., Guilfoyle, T., Malamy, J., Benfey, P., Leyser, O., Bechtold, N., Weisbeek, P. & Scheres, B. (1999) Cell 99, 463 472. 14. MacCleery, S. A. & Kiss, J. Z. (1999) Plant Physiol. 120, 183 192. Conclusion The results presented here, together with other recent articles (6, 22, 24), are on the verge of a conceptual revolution to settle the continuing controversy regarding the Cholodny Went theory of gravitropism (3). Here we show that gravitropic root curvature correlates with asymmetric auxin flux through the LRC. By using the fluorescent auxin biosensor we were able to (i) image auxin transport in root tips at cellular resolution, (ii) reveal exogenously applied auxins as being efficiently compensated by endogenous transport mechanisms, and (iii) differentiate between gravity-induced lateral and basipetal auxin transport. Block of lateral auxin transport by BFA together with gravitydependent AtPIN3 localization in columella cells (6) suggest lateral transport regulation at the level of posttranslational protein targeting. Further basipetal transport from the dlrc across the plrc to the EZ is mediated by AtPIN2 EIR1, and facilitated by the influx carrier. Moreover, we provide additional evidence that the auxin sink in root caps is controlled by efflux carrier activity supplying auxin from the root stele (10). The road map for auxin flux from the stele through the QC and columella into LRC cells raises questions as to how the positional information provided by statolith sedimentation affects regulation of auxin carrier location activity. Recent findings of rapid gravity-induced ph changes (25) might indicate the involvement of stretch-activated ion transport processes (26), resulting in activation of ph-sensitive protein kinase phosphatase signaling (27). It is tempting to speculate that such signaling cascades control endomembrane traffic and thereby relocation of auxin transport regulators. Future analyses of these processes are likely to provide deep insights into control and regulation of tropistic plant growth. We thank G. Jach for providing the GFP-LT and for helpful suggestions; T. Ulmasov for providing the DR5-promoter; the Automated DNA Isolation and Sequencing service group for DNA sequencing; U. Ringeisen for schematic representation of the Arabidopsis root apex; and B. Ruperti and M. Godde for helpful comments and critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (Schwerpunktprogramm Phytohormone), the Fonds der Chemischen Industrie, the European Communities Biotechnology Programs, the Fifth Framework Programme of the European Community: Quality of Life and Management of Living Resources (QLK5-CT- 2001-01871), the European Space Agency Microgravity Applications Program Biotechnology, the Deutsche Zentrum für Luft- und Raumfahrt e.v. (DLR) 50WB0222, and National Aeronautics and Space Administration Grant NAG2-1411. 15. Mullen, J. L., Wolverton, C., Ishikawa, H. & Evans, M. L. (2000) Plant Physiol. 123, 665 670. 16. Deichsel, H., Friedel, S., Detterbeck, A., Coyne, C., Hamker, U. & MacWilliams, H. K. (1999) Dev. Genes Evol. 209, 63 68. 17. Delbarre, A., Muller, P., Imhoff, V. & Guern, J. (1996) Planta 198, 532 541. 18. Yamamoto, M. & Yamamoto, K. T. (1998) Plant Cell Physiol. 39, 660 664. 19. Geldner, N., Friml, J., Stierhof, Y.-D., Jürgens, G. & Palme, K. (2001) Nature 413, 425 428. 20. Delbarre, A., Muller, P. & Guern, J. (1998) Plant Physiol. 116, 833 844. 21. Morris, D. A. & Robinson, J. S. (1998) Planta 205, 606 612. 22. Swarup, R., Friml, J., Marchant, A., Ljung, K., Sandberg, G., Palme, K. & Bennett, M. (2001) Genes Dev. 15, 2648 2653. 23. Rashotte, A. M., Brady, S. R., Reed, R. C., Ante, S. J. & Muday, G. K. (2000) Plant Physiol. 122, 481 490. 24. Rashotte, A. M., DeLong, A. & Muday, G. K. (2001) Plant Cell 13, 1683 1697. 25. Fasano, J. M., Swanson, S. J., Blancaflor, E. B., Dowd, P. E., Kao, T. & Gilroy, S. (2001) Plant Cell 13, 907 922. 26. Yoder, T., Zheng, H., Todd, P. & Staehelin, L. (2001) Plant Physiol. 125, 1045 1060. 27. DeLong, A., Mockaitis, K. & Christensen, S. (2002) Plant Mol. Biol. 49, 285 303. 28. Sack, F. D. & Kiss, J. Z. (1989) Am. J. Bot. 76, 454 464. 29. Blancaflor, E. B., Fasano, J. M. & Gilroy, S. (1998) Plant Physiol. 116, 213 222. Ottenschläger et al. PNAS March 4, 2003 vol. 100 no. 5 2991