SEARCH

SEARCH BY CITATION

Keywords:

  • Rop GTPase;
  • plasma membrane;
  • localization signal;
  • GFP;
  • prenylation;
  • maize

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Signals in the carboxy-terminal hypervariable region (HVR) of Rho and Ras GTPases target these proteins to specific membrane compartments, where they function in signal transduction. ROP6 and ROP7 are closely related maize Rops (a plant-specific Rho subgroup) that share HVR sequences divergent from other Rho HVRs. Both ROPs terminate in CAA, instead of the consensus C-terminal CaaX motif required for membrane association of all characterized Ras and Rho GTPases. The ROP6/7 HVR contains two additional cysteines, potential sites for post-translational modification that leads to membrane association; one is in an internal CaaX motif, which would be at the C-terminus if the final intron in both genes were not removed. Transient expression of a GFP–ROP7 fusion revealed its near-total association with the plasma membrane (PM). Furthermore, the ROP7 HVR is sufficient to target GFP to the PM. Surprisingly, the cysteine in the terminal CAA is not required for PM targeting of GFP–ROP7. In contrast, an internal HVR cysteine is essential for proper targeting of the fusion, and the cysteine in the internal CaaX is required for complete membrane association. Interestingly, this CaaX motif can also direct PM association when placed at the fusion C-terminus by addition of an internal stop codon. Fractionation experiments confirm that maize ROPs associate with membranes in maize seedlings. Our analysis suggests that the ROP7 HVR directs PM localization by a mechanism independent of a C-terminal CaaX motif; this mechanism may have evolved through addition of 3′ intron/exon sequences to a rop progenitor.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Rho (Rho/Rac/Cdc42/Rop) small GTPases make up a family of regulatory proteins that share about 30% identity with the Ras oncogene and, as with Ras, their active forms are thought to be GTP-bound and membrane-associated ( Hall, 1994; Mackay and Hall, 1998; Olofsson, 1999; Seabra, 1998). GTP-bound Rho proteins are hypothesized to interact with and activate downstream effectors, thus regulating local changes at cellular membranes, e.g. formation of a tip-high Ca2+ gradient in growing pollen tubes ( Li et al., 1999 ) or nucleation of actin polymerization ( Rohatgi et al., 1999 ). Whereas no true ortholog of Ras, Rho, or Cdc42 has yet been identified in plants, a large plant-specific Rho subfamily, designated Rop, has been characterized ( Li et al., 1998 ; Lin et al., 1996 ; Winge et al., 1997 ; our unpublished results). Members of the Rop family have been implicated as regulators of pollen tube growth ( Kost et al., 1999 ; Li et al., 1999 ; Lin and Yang, 1997), cell death ( Kawasaki et al., 1999 ), and cell wall synthesis ( Delmer et al., 1995 ; Potikha et al., 1999 ). Due to the localized nature of Rho activity, the mechanisms that control subcellular localization of Rho proteins are important for influencing cellular responses. Although Rop membrane-targeting mechanisms are only beginning to be dissected, some are probably similar to the well characterized ones in mammalian and fungal cells.

In mammalian and yeast cells, both Ras and Rho family GTPases are targeted to specific membrane compartments by signals in a short stretch of carboxy-terminal amino acids (approximately 10–30 residues), termed the hypervariable region (HVR) ( Glomset and Farnsworth, 1994; Magee and Marshall, 1999; Zhang and Casey, 1996). Ras is almost completely membrane-associated in vivo, primarily with the plasma membrane (PM) and the ER/Golgi ( Choy et al., 1999 ; Hancock et al., 1989 ; Hancock et al., 1990 ; Hancock et al., 1991 ); in contrast, Rho GTPases have been found in both cytosolic and membrane-associated fractions. These data suggest that Rho proteins cycle on and off the membranes, as do members of the Rab and Arf small GTPase families ( Adamson et al., 1992b ; Olofsson, 1999). In accordance with this model, Rho GTPases form complexes with their corresponding Rho GDP-dissociation inhibitors (GDIs), which are primarily cytosolic ( Olofsson, 1999). Several Rho proteins have been localized to various sites at the PM (e.g. yeast Cdc42p at the bud tip; Ziman et al., 1993 ), but certain family members have been found associated with other membrane compartments (e.g. mammalian Rac1 with the mitochondria; Boivin and Beliveau, 1995).

Ras and Rho targeting to the PM is thought to require two components in the HVR: (i) an isoprenyl lipid group attached post-translationally to a cysteine residue near the C-terminus, and (ii) either five to seven basic residues, or one or two palmityl groups attached to internal cysteines ( Figure 1) ( Glomset and Farnsworth, 1994; Magee and Marshall, 1999; Resh, 1996). The best characterized HVR component is the CaaX box, a four amino-acid motif at the C-terminus of all previously described Ras and Rho proteins (C, cysteine; a, aliphatic amino acid; X, serine, methionine or leucine), which signals for post-translational modification ( Glomset and Farnsworth, 1994; Zhang and Casey, 1996). The C-terminal amino acid (X) directs association with distinct prenyltransferases, specifying addition of either a farnesyl (X is serine or methionine) or a geranylgeranyl (X is leucine) group ( Casey and Seabra, 1996). Almost all Rho family GTPases characterized to date are geranylgeranylated; substituting serine for the prenylated cysteine prevents prenylation and membrane targeting, and thus inactivates both Rho and Ras proteins ( Choy et al., 1999 ; Ridley et al., 1992 ; Ziman et al., 1991 ). Interestingly, mammalian RhoE, the only Rho GTPase known to have an alternative prenylation motif (CTVM), is farnesylated and localized to distinct regions of the PM ( Foster et al., 1996 ; Guasch et al., 1998 ). Such unique localization patterns (specified by HVR signals) may contribute to the functional specificity of each GTPase ( Hall, 1998).

image

Figure 1. HVR sequences of certain Rops are diverged from conventional targeting signal sequences.

Mutagenized cysteines in ROP7 are labeled. CaaX box sequences at the C-terminus of mammalian K-Ras, H-Ras and RhoB, cotton Rac13, and pea Rop1Ps are all prenylated ( Adamson et al., 1992a ; Hancock et al., 1989 ; Lin et al., 1996 ; Trainin et al., 1996 ), a requirement for membrane association. The mammalian proteins also require either a polybasic domain (K-Ras) or palmitoylation of cysteines (H-Ras, RhoB) for PM targeting ( Adamson et al., 1992a ; Choy et al., 1999 ). Four monocot Rops (maize ROP6 and ROP7; rice OsRac1 and OsRac2) lack the final amino acid of the CaaX motif; internal cysteines in five Rops are conserved. Both ROP6 and ROP7 contain a full-length CaaX motif (boxed) near the C-terminus, and their genes contain an intron with an in-frame stop (asterisk) just 5′ of the CSIM coding sequence. The Arabidopsis thaliana Arac7 gene has an identically placed intron and stop codon; the genomic sequence for the rice genes is unavailable.

Download figure to PowerPoint

The second targeting component within the HVR is located N-terminal to the CaaX motif, and is comprised of either one or two cysteine residues that undergo post-translational palmitoylation ( Hancock et al., 1989 ), or stretches of five to seven basic amino acids, referred to as the polybasic region ( Hancock et al., 1990 ). Although the CaaX box is necessary for membrane association, it is not sufficient for PM targeting of Ras in mammalian cells; prenylated Ras proteins lacking the second signal are associated with endomembranes ( Choy et al., 1999 ). Both the Ras and Rho families include palmitoylated and polybasic region proteins. The functional significance of this is not clear ( Seabra, 1998), however, distinct mechanisms may localize the two Ras types to the PM ( Thissen et al., 1997 ).

In plants, the importance of prenylation in cellular physiology (including GTPase targeting) is being addressed ( Nambara and McCourt, 1999; Qian et al., 1996 ; Randall et al., 1993 ; Rodriguez-Concepcion et al., 1999 ). For example, the PM localization of a plant calmodulin is CaaX-dependent ( Rodriguez-Concepcion et al., 1999 ). Consistent with the models established in mammalian and fungal cells, both pea Rop1 and cotton Rac13, which terminate in CaaL, are geranylgeranylated ( Lin et al., 1996 ; Trainin et al., 1996 ). The Arabidopsis thaliana proteins Rop1At ( Li et al., 1999 ) and At-Rac2 ( Kost et al., 1999 ) contain both a polybasic region and a CaaL motif, and localize to both the cytosol and PM at the pollen-tube tip. In addition, the prenylation motif of At-Rac2 is necessary for its localization to the PM ( Kost et al., 1999 ), and the Rop1At prenylation motif is necessary for its activity ( Li et al., 1999 ). However, whether the PM-targeting mechanisms for these Rops require a second signal (e.g. the polybasic region) has not been addressed, nor have any palmitoylated Rops been identified.

Several recently identified monocot Rops contain what appears to be a variant CaaX motif at their C-termini ( Figure 1). We identified and focused on ROP6 and ROP7, two maize proteins that contain this unconventional HVR sequence, and tested whether the HVR could direct membrane association. Our analysis suggests that ROP6 and ROP7 localize exclusively to the PM, in contrast to other characterized Rho proteins, but similarly to the mammalian Ras. Moreover, PM targeting of ROP7 does not require the CaaX motif, and thus may involve a unique prenylation-independent mechanism.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

ROP6 and ROP7 contain unique, conserved sequences in their C-terminal hypervariable region

ROP6 and ROP7 share 96% amino-acid identity, based on conceptual translation of cDNA and EST sequences (data not shown). They differ by only three of 32 amino acids in their HVRs ( Figure 1), which we will thus abbreviate as the ROP6/7 HVR. The C-terminal amino-acid sequence (CAA, containing Cys-210) of maize ROP6 and ROP7 is distinct from the CaaX motif of most other Rho and Rop proteins. No data yet demonstrate that CAA is a target for the known prenyltransferases. In addition, two internal cysteines (Cys-199 and Cys-206) are present in the ROP6/7 HVR; these cysteines could be palmitoylated, as are internal cysteines in the human H-Ras and RhoB HVRs ( Adamson et al., 1992a ; Hancock et al., 1989 ). Sequences similar to the ROP6/7 HVR are found in some Rops from rice and A. thaliana ( Figure 1).

Interestingly, Cys-199 is part of a sequence (CSIM) that would comprise a Ras-like farnesylation motif were it at the C-terminus of ROP6 or ROP7. Moreover, in both maize rop6 and rop7, an intron lies immediately 3′ to the CSIM coding sequence, and if it were not spliced out an adjacent, in-frame stop codon in the introns of both genes would truncate the protein. The internal CaaX, intron junction, and immediate in-frame stop codon are also present in the A. thaliana Arac7 gene (GenBank accession number CAB43909), suggesting that each gene could produce two transcripts encoding ROP isoforms with distinct HVRs. The ROP6/7 HVR also has a five amino-acid polybasic region.

GFP–ROP fusions localize to the plasma membrane

To test whether ROP6 and ROP7 could direct membrane targeting, we established a method for assaying protein localization using transient expression of green fluorescent protein (GFP) fusions in maize leaf epidermal cells, which is similar to a method used in onion ( Scott et al., 1999 ). GFP fusions with other Ras-like GTPases, including members of the Rop family, have accurately reported the in vivo localization of the endogenous protein ( Choy et al., 1999 ; Kost et al., 1999 ; LaRochelle et al., 1997 ; Li et al., 1999 ). Thus, we cloned full-length rop6 and rop7 cDNAs in-frame with the GFP C-terminus in a 355 CaMV promoter construct. For our localization assay, cells in immature leaf explants were transiently transformed by particle bombardment ( Bilang and Bogorad, 1996), enabling us to observe tens to hundreds of transformed cells per explant ( Figure 2). Transcripts from both rop6 and rop7 are normally expressed in leaves at this stage of development (T. Christensen and J. Fowler, unpublished results).

image

Figure 2. Transient expression of GFP in immature maize epidermal cells observed using conventional epifluorescence microscopy.

(a–c) The same field of cells, following co-bombardment with pBC17 (anthocyanin-inducing) and pMGR7 (GFP–ROP7) plasmids. Many of the transformed cells detectably express from both plasmids (e.g. asterisks); cells expressing anthocyanin only (arrowhead) demonstrate that fluorescence is not induced by biolistic transformation. Due to differences in fluorescence intensity and focal plane among transformed cells, (c) does not accurately portray the localization patterns in some transformed cells (e.g. the out-of-focus cell in the upper right corner). (d,e) A GFP–ROP7-expressing cell fluoresces around the cell periphery. (f,g) A GFP-expressing cell has fluorescent signal primarily in the cytoplasm; the vacuole does not fluoresce. (h) Cells immediately adjacent to a biolistically transformed cell often show weak, specific fluorescence (arrowheads). (a,e,g) Differential interference contrast; (c,d,f,h) epifluorescence; (b) merged image from (a) and (c). N = nucleus. Bars, 10 μm.

Download figure to PowerPoint

We found that the entire periphery of GFP–ROP6- and GFP–ROP7-transformed cells fluoresced distinctively, whereas the GFP-only control cells fluoresced primarily in the cytoplasm and nucleus ( Figure 2). Confocal laser scanning microscopy (CLSM) imaging confirmed that the GFP–ROP6 and –ROP7 fluorescence was almost exclusively at the cell periphery ( Figure 3a–d), and thus the fusion was associated with either the PM or the cell wall. All leaf cell types that we transformed, including abaxial and adaxial epidermal cells at early and late stages of differentiation, macrohairs and prickle hairs, stomatal guard cells and some mesodermal cells, as well as root epidermal cells, exhibited the same pattern of subcellular localization (data not shown). We occasionally observed what appeared to be a strongly fluorescent transformed cell with a tungsten DNA carrier particle in the nucleus, surrounded by weakly fluorescent cells that lacked visible tungsten ( Figure 2h). Fluorescence in the surrounding cells is probably due to intercellular transport of either GFP fusion protein or mRNA ( Jackson and Hake, 1997; Oparka et al., 1999 ). Bombardments with control vectors confirmed that the fluorescent signal was due to expression of GFP. The pBC17 plasmid encodes the B and C1 anthocyanin regulatory proteins and induces anthocyanin pigments in transformed cells ( Bilang and Bogorad, 1996). Neither cells transformed by pBC17 alone ( Figure 2a–c), nor those transformed by control vectors expressing ROP6 or ROP7 alone (data not shown), exhibited detectable above-background fluorescence.

image

Figure 3. Confocal images demonstrate that GFP–ROP7 is associated with the PM.

(a–d) Single optical section: GFP–ROP7 colocalizes with the cell periphery, a faint nuclear signal is present. (e–h) GFP–ROP7 is associated with the retracted PM (arrow) when cells are plasmolyzed. (a,e) Transmitted light image; (b,f) GFP fluorescence; (c,g) cell-wall autofluorescence; (d,h) merge of GFP and cell-wall fluorescence images. (i) Projection of cell in (a–d). GFP–ROP7 is present throughout the PM; punctate regions of intense fluorescence may be associated with biolistic particles on the cell surface (data not shown). (j) Projection of plasmolyzed cell in (e–h). PM tubules (‘Hechtian reticulation’, arrowheads) associated with the inner face of all cell walls are revealed upon PM retraction, for example the discontinuous fluorescence along the wall opposite the retracted PM in (h). This network of cell wall-associated PM is interconnected and continuous with the protoplast. (k) GFP–HVR localizes to the PM. (l) GFP alone is present in both the cytoplasm and nucleus. N = nucleus. Bar, 10 μm.

Download figure to PowerPoint

To distinguish whether the peripheral fluorescence corresponded to association with the PM or with the surrounding cell wall, transformed cells were incubated in hypertonic solution to induce plasmolysis. GFP–ROP7 ( Figure 3e–h,j) and –ROP6 (data not shown) fluorescence co-localized with the retracted PM. Thus ROP6 and ROP7 have localization signals that lead to their near-exclusive association with the PM; cytosolic fluorescence was notably weak. PM association of GFP–ROP7 allowed visualization of a complex network of PM tubules associated with the inner face of the cell walls from which the cell protoplast had retracted ( Figure 3j). This network may represent sites of stable attachment of the PM to the cell wall, and is probably analogous to similar networks (‘Hechtian reticulation’) observed in onion ( Oparka et al., 1994 ; Pont-Lezica et al., 1993 ).

To test whether the ROP7 HVR was sufficient for PM targeting, we fused the final 30 amino acids of ROP7 directly to GFP, and transiently expressed this fusion. GFP–HVR localized as did GFP–ROP7, with the vast majority of fluorescence at the cell periphery ( Figure 3k), which also retracted upon plasmolysis (data not shown). Although the GFP–HVR fusion may not localize quite as efficiently to the PM as the full-length fusion, our current data do not show a statistically significant difference (see Figure 5, below). A few GFP–HVR-expressing cells contained rapidly moving fluorescent particles in the cytoplasm, not observed in cells expressing GFP–ROP7 (data not shown). Nonetheless, a C-terminal ROP7 HVR is clearly sufficient for targeting proteins to the PM.

image

Figure 5. Statistical analysis indicates that the Cys[RIGHTWARDS ARROW]Ser mutant localization patterns are significantly different, and characteristic of each construct.

(a) Fluorescent signal intensity at the cell periphery (the PM) and in the nucleus of transformed cells for each construct (n = 6) were determined and ratios calculated. Ratios (○) are a relative indicator of signal distribution in each cell (>1 indicates more peripheral fluorescence, <1 more nuclear fluorescence). (b) Ratios were log-transformed to allow statistical comparison of the means (◆) by anova; 0 represents equal signal intensity at the nucleus and the periphery. Error bars show the standard deviation of the mean. GFP–HVR and GFP–ROP7 are not significantly different; however, by the F-test and LSD procedure, all single mutants are significantly different from the wild type and from each other; ROP7–C199S/C210S is significantly different from its constituent single mutants.

Download figure to PowerPoint

Cysteines in the ROP7 HVR participate in plasma membrane targeting

We reasoned that if the fusions were localized to the PM by the same mechanisms as the endogenous proteins, they should be sensitive to mutations in the putative localization signals in the ROP6/7 HVR. Therefore we constructed a series of site-directed mutations in which serine replaced the three ROP7 HVR cysteines, thus preventing possible post-translational prenylation or palmitoylation at the altered residues. We then assayed the localization of the GFP–ROP7 mutants ( Figure 4).

image

Figure 4. C-terminal cysteines are required for PM targeting of GFP–ROP7.

Single optical sections, positioned to include the nucleus, of representative cells expressing various GFP constructs. Mutation of ROP7 HVR cysteines to serines, or insertion of a premature stop codon in the ROP7–HVR, alters the localization pattern of the fusion protein. GFP fusions with (a) ROP7 (wild type); (b) ROP7–C199S; (c) ROP7–C206S; (d) ROP7–C210S; (e) ROP7–C199S/C210S; (f) ROP7–C199S/C206S; (g) ROP7–N203*; (h) ROP7–C199S/N203*. (i) GFP alone. N = nucleus. Bar, 10 μm.

Download figure to PowerPoint

Eliminating each of the three cysteines affected the distribution of fluorescence in transformed cells, but to different extents. Replacing the central cysteine (GFP–ROP7–C206S) eliminated PM localization of the fusion, indicating that this cysteine was required for correct targeting ( Figure 4c). GFP–ROP7–C199S exhibited a partial defect in PM targeting ( Figure 4b). Replacing the final cysteine (C210S) did not greatly affect PM localization ( Figure 4d). However, in the C199S/C210S double mutant PM localization was abolished, indicating that these two cysteines do play a role in protein targeting ( Figure 4e). In contrast, the localization of the C199S/C206S double mutant was indistinguishable from that of the C206S single mutant ( Figure 4f). Mutant fusions that were not PM-associated appeared to accumulate preferentially in the nucleus, but were also in the cytosol.

To substantiate our observations, randomly chosen transformed cells from each experiment were assigned values that were a relative measure of fluorescence localization pattern. This value was based on the ratio of fluorescence intensity at the cell periphery to that in the nucleus, two readily definable subcellular structures (see Experimental procedures). Thus values greater than one indicate that more fluorescence is associated with the cell periphery, and values less than one indicate more with the nucleus. Ratios were necessary because different cells exhibited different levels of fluorescence, probably due to differences in protein (GFP) levels. Because the periphery is defined based on the outermost extent of each cell's fluorescence, the value does not distinguish between signal associated with the PM and signal in the cytoplasm adjacent to the PM (e.g. in the C206S mutant; Figure 4c).

The distribution of peripheral/nuclear signal ratios from different transformed cells supports our observation that each construct led to a characteristic, reproducible localization pattern ( Figure 5a). Log-transformed values for each construct were compared by analysis of variance ( anova). All single mutants exhibited statistically significant differences from one another and from the wild type ( Figure 5b). Furthermore, the C199S/210S double mutant is significantly different from each of its constituent single mutants.

The internal CSIM in the ROP7 HVR can function as a conventional CaaX motif

The perfect internal CaaX motif (CSIM) 10 amino acids from the ROP7 C-terminus, paired with a stop-codon-containing intron ( Figure 1), raised the question of whether this motif could play a role in PM localization when placed at the ROP7 C-terminus. Therefore we inserted a stop codon immediately after the CSIM-coding region, creating the GFP–ROP7–N203* construct. This construct encodes a ROP7 variant that would presumably be expressed if the final rop7 intron were not spliced out of the ROP7-mRNA. The N203* variant was targeted at the PM and the cytosol, but was conspicuously absent from the nucleus ( Figures 4g and 6a,b). PM localization of this ROP7 variant was dependent on Cys-199, as the C199S/N203* double mutant was in the nucleus and cytosol, but not associated with the PM ( Figure 4h). Statistical analysis again indicated that these localization patterns were significantly different ( Figure 6c).

image

Figure 6. The internal CaaX motif in ROP7 directs a distinct localization pattern when at the ROP7 C-terminus.

(a,b) Single optical sections of plasmolyzed cells allow better visualization of cytoplasmic fluorescence. N = nucleus; arrows mark retracted PM; arrowheads mark cell wall with associated Hechtian reticulation. Insertion of a premature stop codon, N203* (a) leads to both cytoplasmic and PM-associated localization, but exclusion from the nucleus. This contrasts with the localization of GFP–ROP7–C199S (b) which is cytoplasmic and PM-associated, but is not excluded from the nucleus. Bar = 10 μm. (c) The N203* mutation leads to a localization pattern significantly different from the wild type; elimination of the CaaX box cysteine (ROP7–C199S/N203*) significantly shifts fluorescence to the nucleus.

Download figure to PowerPoint

Maize ROPs co-purify with the P100 particulate fraction

In an attempt to confirm membrane localization of endogenous maize ROP6 and ROP7, antiserum was raised against a peptide from a relatively diverged region of ROP6 ( Figure 7a), and affinity-purified using a bacterially expressed ROP6 fusion protein. These antibodies do not recognize other bacterial proteins, nor an overexpressed control MBP fusion. However, they do recognize bacterially expressed MBP–ROP4 ( Figure 7d) and MBP–ROP7 (data not shown), and are thus not specific for ROP6 and ROP7, recognizing at least one other member of the highly conserved ROP family.

image

Figure 7. Maize ROPs are associated with cellular membranes.

(a) Sequence analysis of maize ROPs in the region corresponding to the immunogenic ROP6 peptide; dots indicate identical amino acids. This region of ROP6 shows no significant homology to non-Rop proteins, as determined by blast search (data not shown). (b) An affinity-purified ROP antiserum recognizes a prominent band (arrow) in the membrane fraction (100 000 g pellet) that is barely detectable in the cytosolic fraction (100 000 g supernatant). This band runs at 22 kDa, the predicted molecular weight for ROPs. (c) Coomassie-stained gel of the 100 000 g pellet and supernatant fractions, as in (b), showing approximately equal loading. (d) Affinity-purified antiserum is specific for ROPs, although not specific for ROP6 and ROP7. When induced (i) in bacteria as a fusion with maltose binding protein (MBP), ROP4 and ROP6 are each recognized by the purified antiserum (arrowhead). The serum does not recognize proteins in uninduced bacteria (u), or in bacteria expressing an MBP fusion with the control D. discoideum ABP-120 protein.

Download figure to PowerPoint

Maize proteins extracted from immature leaves and shoot apices were fractionated into cytosolic (100 000 g supernatant) and membrane-associated (100 000 g pellet) portions and probed with the purified anti-ROP antibodies ( Figure 7b). A band at approximately 22 kDa (the predicted molecular weight of ROP proteins) was primarily associated with the membrane fraction and not the cytosol. Two unidentified proteins of 55–58 kDa, primarily cytosolic, were also detected by the antibodies. These data support our GFP–ROP fusion experiments, indicating that at least some endogenous maize ROPs are associated with cellular membranes.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Using GFP fusions and site-directed mutagenesis, we have characterized a unique PM-targeting signal in the C-terminus of maize ROP7, a member of the Rop subfamily of Rho GTPases (based on predicted amino-acid sequence, J. Fowler, T. Christensen and R.S. Quatrano, unpublished results). The Rop GTPases are a plant-specific subgroup of the Rho family, and have been implicated in such diverse processes as pollen-tube growth ( Kost et al., 1999 ; Li et al., 1999 ; Lin and Yang, 1997); signal transduction in the meristem ( Trotochaud et al., 1999 ); and the generation of reactive oxygen species in disease response ( Kawasaki et al., 1999 ). It is not known whether specific Rop proteins, which share a high degree of sequence identity (74–95%; Li et al., 1998 ; Winge et al., 1997 ), act in only a subset of signaling pathways; however such specificity might be dependent on distinct subcellular distribution of family members.

This idea is supported by the observation that the Rop C-terminal HVRs account for much of the divergence among family members. According to our current understanding, the HVR contains structural features that determine GTPase localization. Motifs in the HVR can direct post-translational addition of either prenyl or palmityl lipid groups, which aid in the association of the GTPase with cell membranes. Indeed, the CaaX box motifs in the HVRs of two of the better characterized Rop proteins are important for PM localization and protein function in pollen-tube growth ( Kost et al., 1999 ; Li et al., 1999 ). The unusual HVR sequences of ROP7, and the closely related ROP6, place them in a subgroup of Rop GTPases with certain rice and A. thaliana Rops ( Figure 1). These proteins do not contain a conventional C-terminal CaaX box motif (although the A. thaliana Arac7 is close). However, they do contain one or two conserved cysteines that are absent from other Rop HVRs; these internal cysteines could be palmitoylation sites. The amino acids adjacent to the internal cysteines also are somewhat conserved and resemble the loosely defined mammalian palmitoylation motif: aaCa (cysteine surrounded by aliphatic amino-acid residues), with a hydrophobic element (e.g. a post-translationally added lipid group) within 12 amino acids ( Hansen et al., 1999 ). Thus we asked whether ROP6/ROP7 localization was unusual, and whether the conserved cysteines functioned in protein targeting.

Our initial approach was to develop an antibody to ROP6 and ROP7 for immunolocalization. In immature leaves, the proteins recognized by the anti-ROP6 serum are primarily associated with cellular membranes ( Figure 7). However, our reagent was not specific for ROP6/ROP7 because it recognizes at least one other member of the maize Rop family, ROP4. Thus at least a subset of the ROP proteins in maize are primarily associated with membranes in vivo, as predicted by current models.

To circumvent our antiserum's lack of specificity, we transiently expressed fusions of the GFP to ROP6 and ROP7 in maize leaf epidermal cells. These fusions are targeted almost exclusively at and uniformly throughout the PM in leaf cells ( Figures 3 and 4). Although we have not identified the specific immature leaf cell types that express rop6 and rop7 in vivo, our data suggest that protein localization is independent of cell type, as we observed the same fluorescence pattern in all cell types transformed (including mesophyll cells and root epidermal cells). Thus the GFP fusion results were consistent with the antibody fractionation, and suggest that ROP6 and ROP7 are PM-localized. Furthermore, as predicted, the final 30 amino acids of the ROP7 HVR are sufficient for PM targeting when placed at the GFP C-terminus.

The lack of significant GFP–ROP6 and –ROP7 signal in the cytosol is in contrast to the location of other Rho GTPases, including Rop1Ps (by immunolocalization; Lin et al., 1996 ) and Rop1At and At-Rac2 (by GFP fusion; Kost et al., 1999 ; Li et al., 1999 ) in pollen tubes. These Rops are associated with the PM at the pollen tube tip but are also found in the cytosol. Thus the ROP6/7 HVR has unique localization signals that specify exclusive association with the PM; this pattern is similar to the distribution of Ras in mammalian cells ( Choy et al., 1999 ). Although overexpression leads to ectopic localization of GFP–Rop1At ( Li et al., 1999 ), we observed uniform PM fluorescence, even in GFP–ROP7 cells with barely detectable signal, suggesting that expression level does not dramatically affect distribution of the fusion at the PM. However, because we have no functional assay for ROP6/7 we cannot test whether fusion of GFP to ROP7 alters its function or localization in some manner. Thus, although we never observed non-uniform accumulation of GFP–ROP6/7 fluorescence at the PM, we cannot exclude the possibility that ROP6 and/or ROP7 localize to specific PM sites in vivo (as do Rop proteins in the pollen tube, for example).

If the GFP fusions were localized to the PM by the same mechanisms as were ROP6 and ROP7, they should also be sensitive to mutations in the HVR. Therefore we constructed a series of site-directed mutations in which serine replaced the ROP7 HVR cysteines, preventing post-translational modification (both prenylation and/or palmitoylation) of these residues. We found that Cys-206 was necessary for PM targeting, and that replacing Cys-199 partially eliminated PM targeting. In contrast to all previous mutational analyses of Rho- and Ras-like GTPases, replacing the final cysteine (Cys-210) did not strongly affect PM localization, except in combination with the Cys-199 mutation. By calculating a value for each transformed cell based on the ratio of fluorescence signal intensity at the cell periphery to that in the cell nucleus, we were able to evaluate statistically whether the fluorescence patterns from each of our different constructs were significantly different ( Figure 5). This analysis supports our observation that the various constructs resulted in characteristic, reproducible localization patterns. These data suggest that all three cysteines function in PM targeting; their functional importance can be ordered Cys-206 > Cys-199 > Cys-210. The mutagenesis results also support our inference that ROP7 localizes to the PM in vivo, via a localization mechanism relying on signals in the HVR.

Due to the presence of the internal CaaX motif and stop codon-containing intron in rop6 and rop7 ( Figure 1), we also constructed a ROP7 mutant (N203*) with a stop codon immediately following Met-202. This mutant terminates in CSIM, a sequence likely to be farnesylated. Interestingly, the localization pattern of this GFP fusion is similar to that of Rop1Ps, Rop1At and At-Rac2: it is present both at the PM and in the cytosol. Thus the exclusive PM association of the full-length GFP–ROP7 fusion is not an artifact produced by fusion to GFP, nor is it due to sequences in the ROP7 N-terminus. In contrast to GFP–ROP7, PM localization of GFP–ROP7-N203* is dependent on the cysteine in its CaaX motif (Cys-199), as predicted by current models for prenylation-dependent GTPase localization. Furthermore, the N203* mutant, unlike the three Cys[RIGHTWARDS ARROW]Ser mutants in the full-length ROP7, is excluded from the nucleus. We suggest that the N203* localization represents a pattern that is due to a fully functional, prenylated CaaX targeting sequence at the truncated ROP7 C-terminus, partitioning protein to both PM and cytosol.

Our results implicate the final 10 amino acids of the ROP6/7 HVR as the source of a unique, CaaX motif-independent signal for exclusive PM localization, with Cys-206 as a crucial component. Because replacing the final cysteine (Cys-210) has little effect on localization patterns, we speculate that ROP7 PM targeting occurs through a mechanism that is not dependent on prenylation, making it a novel variation for Rho and Ras membrane targeting. A negative regulator of Rho in mammalian and yeast cells, RhoGDI, sequesters inactive, GDP-bound Rho GTPases in the cytoplasm through interactions with the C-terminal prenyl group ( Hoffman et al., 2000 ; Olofsson, 1999); an unprenylated ROP7 might not be subject to RhoGDI cytoplasmic sequestration. Alternatively, the final CAA sequence could be recognized by a variant maize prenyltransferase. However, preliminary data argue against this alternative: the PM localization of GFP–ROP7–N203*, but not GFP–ROP7, is affected by melvinolin, an inhibitor of prenyl group biosynthesis ( Rodriguez-Concepcion et al., 1999 ; Z. Vejlupkova, K. Arthur and J. Fowler, unpublished results). Given their internal location, we suggest that Cys-206 and Cys-199 may well be palmitoylated. Only a single plant protein has been shown to be palmitoylated: a chloroplast protein which is palmitoylated coincident with its translocation to a distinct membrane compartment ( Mattoo and Edelman, 1987). Thus definitive tests of these hypotheses will require the establishment of palmitoylation and prenylation assays in maize cells.

Our data also suggest that the unique ROP6 and ROP7 HVR sequences were generated during evolution by addition of an intron and exon to the 3′ end of a rop progenitor gene that originally coded for CSIM at its C-terminus. Furthermore, these observations raise the possibility that ROP6 and ROP7 localization could be regulated by production of alternative transcripts coding for the two variant HVRs. The conservation of the intron junction and adjacent stop codon in A. thaliana Arac7 indicate that such a mechanism could be conserved. An analogous situation has been documented for the pumpkin hydroxypyruvate reductase (HPR) gene ( Mano et al., 1999 ). Light regulates the choice of splice junction at the final intron of the gene, and the alternatively spliced transcripts encode proteins with two different C-terminal sequences; these sequences appear to target the HPR isoforms to either the cytoplasm or the peroxisome. We are currently ascertaining whether such alternative rop6 and rop7 transcripts exist in vivo.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Cloning of maize rop genes

Plasmids with inserts corresponding to the maize rop4, rop6 and rop7 genes were purified from a Lambda ZAP maize shoot apical meristem cDNA library (provided by S. Hake and L. Smith, USDA-PGEC, Albany, CA, USA) using standard methods ( Sambrook et al., 1989 ). One rop6 and two independent rop7 cDNA clones were sequenced. Further details will be provided (J. Fowler, T. Christensen and R.S. Quatrano, unpublished results). Primers for rop6 and rop7 cDNAs were used to PCR amplify genomic DNA (B73 inbred). Amplified products were cloned using the PCR-Script kit (Stratagene, La Jolla, CA, USA) and sequenced using automated methods. Sequences were analyzed using the gcg package (Madison, WI, USA) and web-based blast programs ( Altschul et al., 1997 ) at the National Center for Biotechnology Information (NCBI) .

GFP expression plasmids and mutagenesis

The GFP–ROP fusion constructs were created as follows. Plasmid pBC17 ( Bilang and Bogorad, 1996) was a gift from S. Goff, Novartis (Research Triangle Park, NC, USA). pBCΔ, containing the CaMV 35S promoter, the adh1 intron and a multiple cloning site, is a derivative of pBC17, created by deletion of B and C1 sequences. The stop codon of a modified GFP (SGFP-TYG; Chiu et al., 1996 ), provided by J. Sheen, Massachusetts General Hospital, was replaced by an in-frame EcoRI site via PCR amplification. The GFP coding sequence was unaltered, except for two silent base pair changes. This fragment was cloned into the BamHI/EcoRI sites of pBCΔ to create the GFP-expressing plasmid pMGFP. A rop6/rop7-specific primer with an EcoRI site just upstream of the start codon was used to amplify the maize rop coding regions with Pfu Turbo (Stratagene). Both rop fragments were cloned into EcoRI/KpnI-cut pMGFP to create the expression plasmids pMGR6 and pMGR7. To create a GFP fusion to the C-terminal 30 amino acids of the ROP7-HVR, pMGR7 was cut with SacII and EcoRI to drop out the rop7 5′ coding region, blunted, and religated. The rop6 and rop7 coding sequences were ligated into pBCΔ to create the control plasmids pMR6 and pMR7.

To create the mutated plasmids pMGR7-C199S, pMGR7-C206S, and pMGR7-C210S, the cysteines at ROP7 positions 199, 206, and 210 were mutated to serine ( Kunkel et al., 1987 ) using primers 5′-GCTCCGGG AGCTCTATCAT-3′, 5′-ACCTCATGT CCGGCAGCACG-3′, and 5′-GCAGCACGT CTGC AGCTTAGGAGTC-3′, respectively. In pMGR7-N203*, a stop codon was introduced using the primer 5′-TCTATCATG TA GCTCATGTGTG-3′ (mutagenic nucleotides are underlined). The double mutant constructs pMGR7-C199S/C206S, pMGR7-C199S/C210S, and pMGR7-C199S/N203* were created by mutagenizing a plasmid carrying rop7-C199S sequence. The sequence of all amplified and mutagenized rop7 constructs was confirmed.

Plant material and transformation

The protocol for transient expression in maize leaves was modified from that of Bilang ( Bilang and Bogorad, 1996; Gallagher and Smith, 1999). Immature leaves were dissected from 1–2-month-old maize plants (inbred line W22). Sections of 5–15 mm were taken from the base of the leaf blade and placed, abaxial side up, on MS medium (Life Technologies, Rockville, MD, USA) containing 30 g l−1 sucrose and 15 g l−1 Phytagar (Life Technologies). Transformations were performed with a PDS-1000 helium biolistic system (Bio-Rad, Hercules, CA, USA) using the Bio-Rad protocol at helium pressure of 1100 psi, with a tungsten 17 microcarrier. To enhance transformation, some samples were pretreated in the dark with an osmotic solution ( Vain et al., 1993 ). Transient expression of GFP fluorescence was observed as early as 4 h after transformation, and observations were made over the next 3 days.

Imaging and image analysis

Bombarded leaf samples were placed flat on cover slips, in liquid MS medium, or in halocarbon oil (Sigma H-8773, St. Louis, MO, USA). In some cases samples were plasmolyzed by four 30-min incubations in a 0.25–1 m sucrose series. For conventional microscopy we used a Zeiss Axiovert microscope with differential interference contrast optics, or with epifluorescence using a GFP-optimized band-pass filter set (Chroma 41017). Digital images were acquired using a SPOT CCD camera and software (Diagnostic Instruments, Sterling Heights, MI, USA) . Confocal laser scanning microscopy (CLSM) was on a Leica DM IRBE with a 100×/NA1.4 objective, equipped with the tcs 4 d software and scanning package. GFP fluorescence was imaged using an Ar/Kr laser with the standard Leica FITC settings; cell-wall autofluorescence was imaged using a UV laser and DAPI settings. Digital images were imported into NIH Image 1.60 ( http://rsb.info.nih.gov/nih-image/) for image analysis. Further image processing was done in either Photoshop 4 (Adobe) or Canvas 6 (Deneba).

To confirm that the localization patterns for individual cells were characteristic for all cells expressing the same construct, and to better compare patterns between different constructs, we devised a method for assigning a single value to distinct localization patterns in each transformed cell. To allow comparison of cells with differing levels of GFP expression, we calculated a ratio of fluorescence intensity (measured by confocal imaging) in two distinct regions, the cell periphery (P) and nucleus (N). Using CLSM, a series of 1 μm optical sections was taken for randomly chosen, transformed cells using a standardized set of parameters in at least two separate experiments per construct. The section from each series that included the largest cross-sectional area of the nucleus was chosen for detailed analysis.

Due to the irregular cell shapes it was impossible to obtain an accurate mean value for signal intensity based on measurements of the entire periphery. Instead, we chose to rely on an arbitrary, objective method for assigning P and N values. A line was drawn from the periphery through the nucleus as specified below; the P and N values were determined by taking the highest pixel value, both at the cell periphery (within 1.3 μm of the cell edge) and in the nucleus, that lay on the line. For each image this line was drawn perpendicular to the cell edge that was opposite the nucleus, and was placed to bisect the nucleus; for almost all cells only one line could be drawn to satisfy these conditions. At the magnification used, the pixel and line width corresponded to 0.2 μm. The P/N ratio indicated the signal distribution within each cell. Ratios were log-transformed for statistical analysis using the statgraphics plus 3 software package (Manugistics, Inc., Rockville, MD, USA) .

Antibody preparation and purification

In an attempt to develop antisera specific for ROP6 (and/or ROP7), a rabbit was immunized with a peptide of sequence DFRDHRAYFADHPGASAV (ROP6 amino acids No. 121–138). This region of ROP6 (and ROP7) is the next most divergent from other Rops (after the HVR) ( Figure 7a). Coding regions of the rop4, rop6 and rop7 cDNAs were amplified and cloned into the EcoRI/SalI sites of the bacterial fusion protein expression vector pMAL-c2 (New England Biolabs, Beverly, MA, USA) using Pfu Turbo polymerase (Stratagene). Maltose-binding protein (MBP)–ROP fusions were produced in the bacterial strain UT5600 and, following sonication, purified by standard protocols using amylose resin (New England Biolabs). Fusions were eluted in 50 m m Tris pH 7.6, 50 m m NaCl, 5 m m MgCl2, 10 m m DTT. IPTG-induced and uninduced overnight cultures were mixed with 2× SDS loading buffer and boiled before separation by SDS–PAGE.

MBP–ROP6 and MBP–ROP4 fusion proteins were attached to Actigel resin (Sterogene Bioseparations, Inc., Carlsbad, CA, USA) for affinity-purification columns. Crude serum, which detected the MBP-ROP6 fusion in bacterial extracts, was subjected to a two-step purification over the affinity columns ( Williams et al., 1995 ). First, antibodies that bound to the MBP–ROP6 column were eluted using Actisep elution buffer (Sterogene). Second, eluted antibodies were immuno-adsorbed against the MBP–ROP4 column. Flowthrough containing purified anti-ROP was concentrated using an Ultrafree-4 (Millipore).

Maize protein fractionation and detection

Maize protein preparation and fractionation was done using established protocols ( Gallagher et al., 1988 ) and tissue from immature leaves and shoot apices of 3–4-leaf-stage maize seedlings (W22 inbred). Proteins were extracted in 20 m m MES pH 6.1, 4 m m MgCl2, 0.25 m sucrose, 0.1 m m EGTA, 1 m m DTT, 0.1 m m PMSF, 0.2 mg ml−1 BSA, supplemented with a plant protease inhibitor (Sigma P-9599), and separated into membrane-associated and cytosolic fractions by centrifugation ( Gallagher et al., 1988 ). Protein concentrations were determined by Bio-Rad Protein Assay, and equal amounts of protein from each fraction were separated by SDS–PAGE and transferred to Nitrobind nitrocellulose (MSI, Westboro, MA, USA) ( Sambrook et al., 1989 ). Primary antiserum was incubated with Western blots at a 1 : 250 dilution and detected using alkaline phosphatase-conjugated goat antirabbit IgG.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We would like to thank S. Goff for providing the pBC17 plasmid, J. Sheen for SGFP, and P. Nagpal (UNC-Chapel Hill) for the PCR-modified SGFP. We also thank J. Duvick and Y. Sharma (Pioneer Hi-Bred International, Inc.) for the ROP6 crude antiserum, A. Qu (OSU Statistics Department) for consultation on statistical analysis, and V. Peremyslov, K. Latham and K. Arthur for technical advice and assistance. The biolistic transformation facility was provided by T. Chen and S. Strauss. The manuscript was critiqued by M. Foss, S. Shaw, L. Smith and Z. Yang. The ONR funded some of the microscopy equipment; CLSM was done at the Confocal Facility of the OSU Center for Gene Research and Biotechnology. This work was begun in the lab of R.S.Q. while J.E.F. was supported by a postdoctoral fellowship from NSF (BIR-9404025), and pursued at OSU, supported by the USDA NRICGP (98-35304-6670).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Adamson, P., Marshall, C.J., Hall, A., Tilbrook, P.A. (1992a) Post-translational modifications of p21rho proteins. J. Biol. Chem. 267, 20033 20038.
  • Adamson, P., Paterson, H.F., Hall, A. (1992b) Intracellular localization of the p21rho proteins. J. Cell Biol. 119, 617 627.
  • Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J. (1997) Gapped blast and psi- blast: a new generation of protein database search programs . Nucl. Acids Res. 25, 3389 3402.
  • Bilang, R. & Bogorad, L. (1996) Light-dependent developmental control of rbcS gene expression in epidermal cells of maize leaves. Plant Mol. Biol. 31, 831 841.
  • Boivin, D. & Beliveau, R. (1995) Subcellular distribution and membrane association of Rho-related small GTP-binding proteins in kidney cortex. Am. J. Physiol. 269, 180 189.
  • Casey, P.J. & Seabra, M.C. (1996) Protein prenyltransferases. J. Biol. Chem. 271, 5289 5292.
  • Chiu, W., Niwa, Y., Zeng, W., Hirano, T., Kobayashi, H., Sheen, J. (1996) Engineered GFP as a vital reporter in plants. Curr. Biol. 6, 325 330.
  • Choy, E., Chiu, V., Silletti, J., Feoktistov, M., Morimoto, T., Michaelson, D., Ivanov, I., Philips, M. (1999) Endomembrane trafficking of Ras: the CAAX motif targets proteins to the ER and Golgi. Cell, 98, 69 80.
  • Delmer, D.P., Pear, J.R., Andrawis, A., Stalker, D.M. (1995) Genes encoding small GTP-binding proteins analogous to mammalian rac are preferentially expressed in developing cotton fibers. Mol. Gen. Genet. 248, 43 51.
  • Foster, R., Hu, K.Q., Lu, Y., Nolan, K.M., Thissen, J., Settleman, J. (1996) Identification of a novel human Rho protein with unusual properties: GTPase deficiency and in vivo farnesylation. Mol. Cell Biol. 16, 2689 2699.
  • Gallagher, S., Short, T.W., Ray, P.M., Pratt, L.H., Briggs, W.R. (1988) Light-mediated changes in two proteins found associated with plasma membrane fractions from pea stem sections. Proc. Natl Acad. Sci. USA , 85, 8003 8007.
  • Gallagher, K. & Smith, L.G. (1999) discordia mutations specifically misorient asymmetric cell divisions during development of the maize leaf epidermis. Development, 126, 4623 4633.
  • Glomset, J.A. & Farnsworth, C.C. (1994) Role of protein modification reactions in programming interactions between ras-related GTPases and cell membranes. Annu. Rev. Cell Biol. 10, 181 205.
  • Guasch, R.M., Scambler, P., Jones, G.E., Ridley, A.J. (1998) RhoE regulates actin cytoskeleton organization and cell migration. Mol. Cell Biol. 18, 4761 4771.
  • Hall, A. (1994) Small GTP-binding proteins and the regulation of the actin cytoskeleton. Annu. Rev. Cell Biol. 10, 31 54.
  • Hall, A. (1998) Rho GTPases and the actin cytoskeleton. Science, 279, 509 514.
  • Hancock, J.F., Magee, A.I., Childs, J.E., Marshall, C.J. (1989) All ras proteins are polyisoprenylated but only some are palmitoylated. Cell, 57, 1167 1177.
  • Hancock, J.F., Paterson, H., Marshall, C.J. (1990) A polybasic domain or palmitoylation is required in addition to the CAAX motif to localize p21ras to the plasma membrane. Cell, 63, 133 139.
  • Hancock, J.F., Cadwallader, K., Paterson, H., Marshall, C.J. (1991) A CAAX or a CAAL motif and a second signal are sufficient for plasma membrane targeting of ras proteins. EMBO J. 10, 4033 4039.
  • Hansen, S.G., Grosenbach, D.W., Hruby, D.E. (1999) Analysis of the site occupancy constraints of primary amino acid sequences in the motif directing palmitylation of the vaccinia virus 37-kDa envelope protein. Virology, 254, 124 137.DOI: 10.1006/viro.1998.9543
  • Hoffman, G.R., Nassar, N., Cerione, R.A. (2000) Structure of the Rho family GTP-binding protein Cdc42 in complex with the multifunctional regulator RhoGDI. Cell, 100, 345 356.
  • Jackson, D. & Hake, S. (1997) Morphogenesis on the move: cell-to-cell trafficking of plant regulatory proteins. Curr. Opin. Genet. Dev. 7, 495 500.
  • Kawasaki, T., Henmi, K., Ono, E., Hatakeyama, S., Iwano, M., Satoh, H., Shimamoto, K. (1999) The small GTP-binding protein rac is a regulator of cell death in plants. Proc. Natl Acad. Sci. USA , 96, 10922 10926.DOI: 10.1073/pnas.96.19.10922
  • Kost, B., Lemichez, E., Spielhofer, P., Hong, Y., Tolias, K., Carpenter, C., Chua, N.-H. (1999) Rac homologues and compartmentalized phosphatidylinositol 4,5-bisphosphate act in a common pathway to regulate polar pollen tube growth. J. Cell Biol. 145, 317 330.
  • Kunkel, T.A., Roberts, J.D., Zakour, R.A. (1987) Rapid and efficient site-specific mutagenesis without phenotypic selection. Meth. Enzymol. 154, 367 382.
  • LaRochelle, D.A., De Vithalani, K.K., Lozanne, A. (1997) Role of Dictyostelium racE in cytokinesis: mutational analysis and localization studies by use of green fluorescent protein. Mol. Biol. Cell , 8, 935 944.
  • Li, H., Wu, G., Ware, D., Davis, K.R., Yang, Z. (1998) Arabidopsis Rho-related GTPases: differential gene expression in pollen and polar localization in fission yeast. Plant Physiol. 118, 407 417.
  • Li, H., Lin, Y., Heath, R.M., Zhu, M.X., Yang, Z. (1999) Control of pollen tube tip growth by a Rop GTPase-dependent pathway that leads to tip-localized calcium influx. Plant Cell, 11, 1731 1742.
  • Lin, Y., Wang, Y., Zhu, J.-K., Yang, Z. (1996) Localization of a Rho GTPase implies a role in tip growth and movement of the generative cell in pollen tubes. Plant Cell, 8, 293 303.
  • Lin, Y. & Yang, Z. (1997) Inhibition of pollen tube elongation by microinjected anti-Rop1Ps antibodies suggests a crucial role for Rho-type GTPases in the control of tip growth. Plant Cell, 9, 1647 1659.
  • Mackay, D.J. & Hall, A. (1998) Rho GTPases. J. Biol. Chem. 273, 20685 20688.
  • Magee, T. & Marshall, C. (1999) New insights into the interaction of Ras with the plasma membrane. Cell, 98, 9 12.
  • Mano, S., Hayashi, M., Nishimura, M. (1999) Light regulates alternative splicing of hydroxypyruvate reductase in pumpkin. Plant J. 17, 309 320.DOI: 10.1046/j.1365-313x.1999.00378.x
  • Mattoo, A.K. & Edelman, M. (1987) Intramembrane translocation and posttranslational palmitoylation of the chloroplast 32-kDa herbicide-binding protein. Proc. Natl Acad. Sci. USA , 84, 1497 1501.
  • Nambara, E. & McCourt, P. (1999) Protein farnesylation in plants: a greasy tale. Curr. Opin. Plant Biol. 2, 388 392.DOI: 10.1016/s1369-5266(99)00010-2
  • Olofsson, B. (1999) Rho guanine dissociation inhibitors: pivotal molecules in cellular signalling. Cell Signal. 11, 545 554.DOI: 10.1016/s0898-6568(98)00063-1
  • Oparka, K., Prior, D., Crawford, J. (1994) Behavior of plasma membrane, cortical ER and plasmodesmata during plasmolysis of onion epidermal cells. Plant Cell Environ. 17, 163 171.
  • Oparka, K.J., Roberts, A.G., Boevink, P. et al. (1999) Simple, but not branched, plasmodesmata allow the nonspecific trafficking of proteins in developing tobacco leaves. Cell, 97, 743 754.
  • Pont-Lezica, R., McNally, J., Pickard, B. (1993) Wall-to-membrane linkers in onion epidermis: some hypotheses. Plant Cell Environ. 16, 111 123.
  • Potikha, T.S., Collins, C.C., Johnson, D.I., Delmer, D.P., Levine, A. (1999) The involvement of hydrogen peroxide in the differentiation of secondary walls in cotton fibers. Plant Physiol. 119, 849 858.
  • Qian, D., Zhou, D., Ju, R., Cramer, C.L., Yang, Z. (1996) Protein farnesyltransferase in plants: molecular characterization and involvement in cell cycle control. Plant Cell, 8, 2381 2394.
  • Randall, S.K., Marshall, M., Crowell, D. (1993) Protein isoprenylation in suspension-cultured tobacco cells. Plant Cell, 5, 433 442.
  • Resh, M.D. (1996) Regulation of cellular signalling by fatty acid acylation and prenylation of signal transduction proteins. Cell Signal, 8, 403 412.DOI: 10.1016/s0898-6568(96)00088-5
  • Ridley, A.J., Paterson, H.F., Johnston, C.L., Diekmann, D., Hall, A. (1992) The small GTP-binding protein rac regulates growth factor-induced membrane ruffling. Cell, 70, 401 410.
  • Rodriguez-Concepcion, M., Yalovsky, S., Zik, M., Fromm, H., Gruissem, W. (1999) The prenylation status of a novel plant calmodulin directs plasma membrane or nuclear localization of the protein. EMBO J. 18, 1996 2007.
  • Rohatgi, R., Ma, L., Miki, H., Lopez, M., Kirchhausen, T., Takenawa, T., Kirschner, M.W. (1999) The interaction between N-WASP and the Arp2/3 complex links Cdc42-dependent signals to actin assembly. Cell, 97, 221 231.
  • Sambrook, J., Fritsch, E.F., Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor: Cold Spring Harbor Laboratory Press.
  • Scott, A., Wyatt, S., Tsou, P.L., Robertson, D., Allen, N.S. (1999) Model system for plant cell biology: GFP imaging in living onion epidermal cells. Biotechniques, 26, 1128 1132.
  • Seabra, M.C. (1998) Membrane association and targeting of prenylated Ras-like GTPases. Cell Signal. 10, 167 172.DOI: 10.1016/s0898-6568(97)00120-4
  • Thissen, J.A., Gross, J.M., Subramanian, K., Meyer, T., Casey, P.J. (1997) Prenylation-dependent association of Ki-Ras with microtubules. Evidence for a role in subcellular trafficking. J. Biol. Chem. 272, 30362 30370.
  • Trainin, T., Shmuel, M., Delmer, D.P. (1996) In vitro prenylation of the small GTPase Rac13 of cotton. Plant Physiol. 112, 1491 1497.
  • Trotochaud, A.E., Hao, T., Wu, G., Yang, Z., Clark, S.E. (1999) The CLAVATA1 receptor-like kinase requires CLAVATA3 for its assembly into a signaling complex that includes KAPP and a Rho-related protein. Plant Cell, 11, 393 406.
  • Vain, P., McMullen, M.D., Finer, J.J. (1993) Osmotic treatment enhances particle bombardment-mediated transient and stable transformation of maize. Plant Cell Reports, 12, 84 88.
  • Williams, J.A., Langeland, J.A., Thalley, B.S., Skeath, J.B., Carroll, S.B. (1995) Production and purification of polyclonal antibodies against proteins expressed in E. coli. In DNA Cloning 2: Expression Systems (Glover, D.M. and Hames, B.D., eds). Oxford: IRL Press, pp. 15 58.
  • Winge, P., Brembu, T., Bones, A.M. (1997) Cloning and characterization of rac-like cDNAs from Arabidopsis thaliana. Plant Mol. Biol. 35 , 483 495.
  • Zhang, F.L. & Casey, P.J. (1996) Protein prenylation: molecular mechanisms and functional consequences. Annu. Rev. Biochem. 65, 241 269.
  • Ziman, M., O'brien, J.M., Ouellette, L.A., Church, W.R., Johnson, D.I. (1991) Mutational analysis of CDC42Sc, a Saccharomyces cerevisiae gene that encodes a putative GTP-binding protein involved in the control of cell polarity. Mol. Cell Biol. 11, 3537 3544.
  • Ziman, M., Preuss, D., Mulholland, J.O., Brien, J.M., Botstein, D., Johnson, D.I. (1993) Subcellular localization of Cdc42p, a Saccharomyces cerevisiae GTP-binding protein involved in the control of cell polarity. Mol. Biol. Cell , 4, 1307 1316.

EMBL sequence accession numbers AJ278666(ROP4), AJ278665(ROP6) and AJ278664(ROP7).