Cloning of the Arabidopsis WIGGUM gene identifies a role for farnesylation in meristem development

Abstract

Control of cellular proliferation in plant meristems is important for maintaining the correct number and position of developing organs. One of the genes identified in the control of floral and apical meristem size and floral organ number in Arabidopsis thaliana is WIGGUM. In wiggum mutants, one of the most striking phenotypes is an increase in floral organ number, particularly in the sepals and petals, correlating with an increase in the width of young floral meristems. Additional phenotypes include reduced and delayed germination, delayed flowering, maturation, and senescence, decreased internode elongation, shortened roots, aberrant phyllotaxy of flowers, aberrant sepal development, floral buds that open precociously, and occasional apical meristem fasciation. As a first step in determining a molecular function for WIGGUM, we used positional cloning to identify the gene. DNA sequencing revealed that WIGGUM is identical to ERA1 (enhanced response to abscisic acid), a previously identified farnesyltransferase β-subunit gene of Arabidopsis. This finding provides a link between protein modification by farnesylation and the control of meristem size. Using in situ hybridization, we examined the expression of ERA1 throughout development and found it to be nearly ubiquitous. This extensive expression domain is consistent with the pleiotropic nature of wiggum mutants and highlights a broad utility for farnesylation in plant growth and development.

The development of multicellular organisms relies on signal transduction for basic cellular functions and intercellular communication. The function of proteins involved in signal transduction pathways often depends on covalent attachment of chemical side groups. Protein prenylation, in which lipophilic isoprenyl groups like the 15-carbon farnesyl and 20-carbon geranylgeranyl moieties are covalently attached to target proteins, has been shown to be one such posttranslational modification (1, 2). The largest group of proteins known to be prenylated is the Ras superfamily of small GTP-binding proteins, with other prenylated proteins falling into the classes of lipopeptide pheromones, nuclear lamins, and trimeric G proteins (1, 2). The functional diversity of known prenylated proteins indicates a broad utility for isoprenylation in the control of cell growth and differentiation, cytokinesis, and membrane trafficking.

Prenylation of proteins is catalyzed by farnesyltransferase (FTase) or one of two types of geranylgeranyltransferases (GGTase-I and GGTase-II). The closely related FTase and GGTase-I enzymes function as heterodimers, sharing a common α-subunit with a distinct β-subunit responsible for providing substrate specificity (2, 3). After isoprenylation by FTase or GGTase-I, the substrate protein then is methylated. Prenylation and subsequent methylation strongly increase the hydrophobicity of proteins, facilitating membrane localization. Despite the observation that most prenylated proteins have been found to be associated with membranes (2), the isoprenoid group per se is not the determinant of localization. Prenylated proteins contain a variety of membrane targeting signals allowing for localization to distinct intracellular membranes. Other roles for prenylation are protein–protein interactions and assembly into multisubunit protein complexes (2).

Prenylation, a modification that has profound implications for signal transduction and intracellular trafficking pathways, has been studied extensively over the past two decades in both fungal and mammalian systems. cDNAs for α- and β-subunits of FTase and GGTase-I have been identified from yeast, rat, bovine, and human libraries (2), and numerous substrate proteins also have been identified (4). More recently, the enzymes responsible and pathways affected have begun to be studied in plants as well; FTase α and β genes have been cloned from pea, tomato, tobacco, and Arabidopsis thaliana (5).

A connection between prenylation and hormone signaling in plants was made when ERA1 (enhanced response to abscisic acid) was cloned from Arabidopsis (6). Abscisic acid is a plant hormone involved in modulating development, most notably with regard to seed dormancy and drought-induced water retention (7). era1 mutants were identified based on their increased seed dormancy, even in the absence of abscisic acid (6). era1 mutants also displayed increased tolerance to drought stress, through a reduction in stomatal apertures (8). The ERA1 gene encodes the β-subunit for protein FTase, and mutants were found to be devoid of FTase activity in an in vitro assay (6). Based on biochemical assays investigating expression, activity, and inhibition, farnesylation has been proposed to play other roles in plant growth and development, including nutrient allocation (9) and cell division (9–11).

The control of cell division and differentiation is critical for the proper development of multicellular organisms. Groups of undifferentiated stem cells at the growing tips of plants comprise meristems, in which cell division control is vital. The size of meristems generally remains constant throughout the lifespan of the plant, indicative of the highly controlled pattern of cell division within meristems, both in number and position (12). Cellular differentiation occurs at the periphery of meristems to produce organs flanking the meristem. In Arabidopsis, mutants that are unable to initiate and/or maintain viable shoot apical meristems [e.g., shoot meristemless (13) and wuschel (14)] and mutants that have enlarged shoot apical and floral meristems [e.g., clavata1 (clv1), clv2, and clv3 (15–17)] and wiggum (wig; ref. 18) have been identified. The SHOOTMERISTEMLESS (19) and WUSCHEL (20) genes have been cloned and shown to encode homeodomain proteins that are likely to act as transcription factors in shoot and floral meristem cells. The cloning of all three CLV genes also has been reported: CLV1 encodes a receptor kinase (21), CLV3 encodes the presumed extracellular protein ligand for CLV1 (22), and CLV2 encodes a receptor-like protein (23). These two classes of Arabidopsis meristem genes implicate specific transcription factors and signal transduction components likely to function in cell–cell communication in the control of meristematic growth.

wig mutants, like the clv mutants, have enlarged shoot apical and floral meristems, resulting in flowers with extra floral organs in all whorls. Arabidopsis flowers are arranged in concentric whorls in which the first and outermost whorl is comprised of sepals, the second of petals, the third of stamens, and the fourth of carpels. clv mutants have taller shoot and floral meristems than wild type and show the greatest number of extra floral organs in the inner whorls. wig plants, in contrast, have wider meristems and wig flowers have more sepals and petals than wild type. wig mutants exhibit additional phenotypes including reduced and delayed germination, delayed flowering, longer life-span, reduced internode elongation, aberrant sepal development, floral buds that open precociously, occasional fasciation, and modest phyllotaxy defects (18).

To expand our understanding of meristematic development, we have cloned the WIG gene. Sequencing and complementation experiments indicate that WIG is identical to ERA1 (encoding FTase β). We examined the ERA1 mRNA expression pattern in wild-type plant tissues by using in situ hybridization and found that ERA1 is broadly expressed throughout development, from embryogenesis through vegetative and reproductive growth. The wig alleles therefore have provided genetic and molecular evidence of a role for farnesylation in meristematic growth and development. Furthermore, the pleiotropic wig phenotypes indicate that farnesylation also is involved in numerous other aspects of development.

Materials and Methods

Plant Growth.

Seeds were sown on a 1:1:1 mixture of soil/vermiculite/perlite, imbibed for 5 days at 4°C, and grown under continuous cool-white fluorescent light at ≈22°C. wig-1 (renamed era1–4, see Results), gl3, and tt2 are in the Landsberg erecta (Ler) background. wig-2 and wig-3 (renamed era1–5 and era1–6, respectively) originally were identified in the Nossen ecotype and subsequently outcrossed two times to Ler (18).

Map-Based Cloning and Sequencing of WIGGUM.

Based on a preliminary map position for WIG in the middle of chromosome 5 (18), wig-1 (era1–4) was crossed to closely linked visible markers to obtain wig-1 gl3 and wig-1 tt2 double mutants. These lines in the Ler ecotype were crossed to wild-type plants of the Columbia ecotype to generate a mapping population. F₂ progeny were screened for loss of one of the two phenotypes, indicating that a recombination event between Ler and Columbia chromosomal DNA in the vicinity of WIG had occurred. DNA was prepared from recombinants as described (24). Using chromosome 5 sequence data from the Arabidopsis Genome Initiative supplied by the Kazusa group (http://www.kazusa.or.jp/arabi/), polymorphisms were identified to design PCR markers. Cleaved amplified polymorphic sequences (CAPS; ref. 25) and derived CAPS (26) markers were used to map the meiotic recombination breakpoints by PCR and restriction digestion. A list of CAPS markers designed for this study is provided in Table 1.

Table 1.

CAPS markers designed for this study

Marker	Size, bp	Polymorphism*	Primers
MYH19L	175	SspI: Col=175	5′-CTATATATTTTATACAAGTTTAAAATAAATA-3′
		Ler=145, 30	5′-CTACCATGTGATGACATG-3′
MYH19VR	880	XhoI: Col=610, 270	5′-CGCGTGGGTACTCATCTAGTG-3′
		Ler=880	5′-CGGATTAGCAGCTAGGACAG-3′
MSN9L	190	AflIII: Col=190	5′-GACAAGGCTTATTGTTTACGT-3′
		Ler=170, 20	5′-GAATGTATTATGCGTAGTGAC-3′
MSN9R	200	SpeI: Col=200	5′-TGGCAGTAATTTAATGATAAACTA-3′
		Ler=180, 20	5′-CGTATTGAGCCTACATAACAA-3′
MPO12VL	180	Fnu4HI: Col=160, 20	5′-AATTAATTTGAGTTTGTTTGGCAG-3′
		Ler=180	5′-TATGTTTCAAACAATAGTTTAGG-3′
MPO12.2	220	HinfI: Col=220	5′-GTCTTGTAGACTCTCCAAGACT-3′
		Ler=200, 20	5′-CTCTTTCTATTATGTTCACCTCT-3′
MPO12M	200	HinfI: Col=200	5′-GCTCGAGATAAATATTTGAGACT-3′
		Ler=180, 20	5′-AGGGATTCTTACTCTTTGTAGG-3′
MPO12R	305	Tth111I: Col=285, 20	5′-GTTTCTGATCTGTGGTGACGGA-3′
		Ler=305	5′-GGCTAATTCTCTCCGGCAC-3′

^*Indicated is the restriction enzyme used and the sizes of the digestion products for Col (Columbia) and Ler. 

DNA sequencing was performed by using the ABI PRISM Dye Terminator method (Perkin–Elmer). Two independent PCR amplifications of genomic DNA from the wig alleles and the corresponding wild-type ecotype were sequenced. macvector (Kodak IBI) was used to align our sequencing data with the Columbia sequence from the Arabidopsis Genome Initiative database.

A cDNA corresponding to the 3′ two-thirds of WIG (renamed ERA1, see Results) was isolated from a flower cDNA library (27) by using a genomic ERA1 probe. Rapid amplification of cDNA ends (RACE)–PCR (5′ RACE-PCR; GIBCO/BRL) was performed on RNA isolated from Ler inflorescences to identify the 5′ end of ERA1. PCR generation of a full-length ERA1 cDNA used high-fidelity Pwo DNA Polymerase (Boehringer Mannheim) and specific primers designed with BamHI ends to amplify from the starting ATG to the poly(A) addition site in the 3′ untranslated region. The PCR product then was cloned into pZero-Blunt (Invitrogen) to create p4B. The entire 1.6-kb insert was sequenced to confirm that no mutations had occurred. An overexpression construct was made by digesting p4B with BamHI and cloning the ERA1 cDNA into pCGN18 (28), placing ERA1 under control of the cauliflower mosaic virus 35S promoter. This construct 35S∷ERA1, was introduced into plants by Agrobacterium-mediated transformation using vacuum infiltration (29).

A genomic ERA1 clone was constructed by digesting MSN9 with HindIII and ligating a 6.8-kb DNA fragment into pPZP211 (30). This construct, ERA1–6.8, was introduced into plants by Agrobacterium-mediated transformation using vacuum infiltration (29).

In Situ Hybridization.

Tissues were fixed (31) and radioactive in situ hybridization was performed as described (32). Exposure time was ≈6 wk. The riboprobe template was constructed by digesting p4B with BamHI and cloning the ERA1 cDNA into pGEM7zf(t) (Promega). Antisense and sense probes were transcribed from HindIII- and SacI-linearized pERA1.7 by using SP6 and T7 RNA polymerase, respectively. This clone was demonstrated to be specific as it hybridized to a single band on a genomic Southern blot under low stringency conditions [washed with 2× standard saline phosphate/EDTA (0.18 M NaCl/10 mM phosphate, pH 7.4/1 mM EDTA) at 50°C, data not shown].

Results

Cloning and Sequence of WIGGUM.

To isolate the WIG gene, we used PCR-based CAPS markers (Table 1) to map recombination breakpoints between wig-1 and either tt2 or gl3. Using 38 wig-1 tt2 recombinants and 21 wig-1 gl3 recombinants, we were able to localize WIG to a region of less than 100 kb about 15 Mb from the top of chromosome 5, between markers MYH19VR and MPO12VL, respectively (Table 1). Genetic recombination frequencies showed that wig-1 is ≈1–2 centimorgans from gl3, and ≈8–10 centimorgans from tt2, indicating that the 21 wig-1 gl3 recombinants were at a higher density than the 38 wig-1 tt2 recombinants. The wig-1 gl3 recombinants mapped at a frequency of ≈1 per 15 kb, suggesting that WIG was most likely to lie in the right half of MSN9, a fully sequenced 61-kb bacterial artificial chromosome clone.

Annotation of MSN9 identified 15 putative genes (http://www.kazusa.or.jp/arabi/), providing candidate genes to sequence from the three known wig alleles to identify mutations. Nucleotide substitutions for all of these alleles were found in the coding region of MSN9.15; MSN9.15 corresponds to the ERA1 β-subunit FTase gene (Fig. 1). wig-1 and wig-2 contain premature stop codons and wig-3 contains a splice site mutation (Fig. 1). All of these alleles are likely to result in a complete loss of function. The premature stop codon in wig-1 occurs before the first glycine-rich amino acid repeat conserved in FTase βsubunits (33), the premature stop codon in wig-2 disrupts the fourth repeat, and wig-3 is inappropriately spliced (determined by reverse transcriptase–PCR, data not shown) after the first repeat (Fig. 1).

Figure 1.

The ERA1 gene and protein sequences are shown. Triangles above the cDNA sequence indicate the position of introns. The stop codon at position 1449 is indicated by an *. The C to T and G to A nucleotide substitutions resulting in premature stop codons in wig-1 (era1–4) and wig-2 (era1–5), respectively, are indicated. # denotes the position of the G to A substitution at the 3′ end of the fifth intron in wig-3 (era1–6). Conserved repeats with glycine-rich amino acid cores that are common to FTase β-subunits (33) are underlined. The additional N-terminal amino acid stretch found in Arabidopsis but not in pea and tomato is underlined with a dashed line.

Transgenic Complementation of wig.

To confirm by complementation that we had identified the correct gene, we transformed a genomic clone into the wig-1 mutant. This construct, ERA1–6.8, contains the 3,389-bp FTase β coding region plus ≈2.4 kb of upstream and ≈1.1 kb of downstream DNA sequences. wig-1 ERA1–6.8 T₁ plants were examined for complementation of the wig-1 phenotypes. Eight transgenic lines were examined closely, and all of these lines showed complete rescue of most wig-1 phenotypes including germination, flowering time, lifespan, internode elongation, phyllotaxy, sepal development, and splayed floral bud defects. The extra floral organ defect was only partially rescued (Fig. 2 A–C), suggesting that flower development may be more sensitive to the exact amount of functional FTase (which may vary because of possible position effects in the transgenic lines). Alternatively, not all of the promoter or regulatory elements may be present on the 6.8-kb genomic fragment. The nearly complete complementation of wig-1 by the genomic ERA1–6.8 clone provides genetic evidence that WIG truly is the FTase β gene and is thus synonymous with ERA1. Consequently, wig-1 has been renamed era1–4, wig-2 renamed era1–5, and wig-3 renamed era1–6.

Figure 2.

Transgenic complementation of era1 mutants. Shown are inflorescences of Ler (A), era1–4 (B), and T₁ plants of the following genotypes: era1–4 ERA1–6.8 (C), era1–4 35S∷ERA1 (D), era1–5 35S∷ERA1 (E), era1–6 35S∷ERA1 (F), and Ler 35S∷ERA1 (G). Wild-type Ler flowers consist of four sepals, four petals, six stamens, and two carpels (A). era1–4 mutants produce floral buds that open precociously and can produce flowers with many more sepals and petals than wild type (B). ERA1–6.8 partially complements the era1–4 floral phenotypes, evident by the reduced number of extra floral organs and the more enclosed floral buds. 35S∷ERA1 fully complements all mutant phenotypes in era1–4, era1–5, and era1–6, and results in no overexpression phenotype in Ler.

Identification of the 5′ End of ERA1 and Construction of a Full-Length ERA1 cDNA.

A discrepancy at the 5′ end of the FTase β gene exists between the MSN9.15 annotation and the reported cDNA for ERA1. MSN9.15 is predicted to have an additional 5′ exon encoding 78 aa. Others raised the suspicion that the 1.45-kb cDNA sequence reported for ERA1 (6) was incomplete because both pea and tomato FTase β genes encode an additional N-terminal sequence of approximately 40 aa (34). To reconcile this discrepancy, we performed 5′ RACE-PCR on RNA isolated from wild-type Ler inflorescences. The 5′ RACE product included the 14th predicted exon with ≈50 nts of upstream untranslated sequence. A stop codon in-frame and immediately before the start ATG indicated that the 5′ RACE product contains the bona fide 5′ end of the FTase β gene.

We also attempted to isolate a full-length ERA1 cDNA by screening a cDNA library (27) with an ERA1-specific genomic probe. The longest clone identified was ≈1 kb in length and corresponded to more than half of the gene at the 3′ end. To construct the full-length cDNA, PCR was used to combine the 5′ RACE product with the 1-kb partial 3′ cDNA. The resulting cDNA corresponds to an mRNA with 1.45-kb coding sequence and a 3′ untranslated region of ≈180 bp (Fig. 1). A protein alignment of the translated ERA1 cDNA with other FTase β subunits revealed that Arabidopsis FTase β contains an N-terminal stretch of 30–40 aa that is not present in tomato and pea but is present in rat and yeast (data not shown). The function of these N-terminal amino acids is unclear, as their identities are not conserved.

Overexpression of ERA1.

Because era1–4 through era1–6 loss-of-function alleles display a variety of developmental defects (18), we were interested in determining the effect of ERA1 overexpression. An overexpression cDNA clone was made by placing the full-length ERA1 cDNA under control of the cauliflower mosaic virus 35S promoter. The 35S∷ERA1 construct was transformed into wild-type Ler and era1–4, era1–5, and era1–6 mutant plants. Thirty-three of 35 transgenic Ler lines showed no obvious overexpression phenotypes (Fig. 2G), whereas two lines appeared to be cosuppressed as they exhibited weak era1–4-like phenotypes (data not shown). A total of 64 T₁ lines in era1–4, era1–5, and era1–6 showed complete rescue of all mutant phenotypes, including floral organ number (Fig. 2 D–F), providing yet another line of evidence that WIG and ERA1 are identical.

ERA1 Expression Pattern.

In the initial study of ERA1, mRNA was detected only in flower buds by Northern blot analysis (6). A subsequent study of transgenic plants expressing an ERA1 promoter-GUS fusion revealed expression in vegetative tissues, most notably in guard cells (8). Given the enlarged meristem phenotypes of era1–4, era1–5, and era1–6 (18), we predicted that ERA1 also would be expressed in the shoot apical meristem. Therefore, we performed a more thorough investigation of the ERA1 expression pattern by using in situ hybridization on various wild-type plant tissues.

We first examined ERA1 expression in inflorescence tissue. A sense control provided a measure of background (Fig. 3 A and B). An ERA1 antisense probe hybridized to wild-type Ler inflorescences with signal throughout the stem, shoot apical meristem, and developing flowers (Fig. 3 C and D). ERA1 was expressed throughout the inflorescence in all stages of development, with levels highest in the shoot apical meristem and in the center of developing flowers. ERA1 mRNA also was detected throughout the vegetative shoot apical meristem and developing leaves of 7-day-old Ler seedlings, but not in the cotyledons (Fig. 3 E and F). In addition, ERA1 expression was detected throughout embryogenesis in the embryo and the endosperm (Fig. 3 G–J), beginning in embryos just undergoing the first division to produce a two-cell embryo and continuing through the fully mature stage, although expression was reduced in the later stages. ERA1 expression also was detected throughout era1–4, era1–5, and era1–6 mutant inflorescences, seedlings, and embryos (data not shown), indicating that none of these alleles is an RNA null. Reverse transcription–PCR of RNA extracted from developing siliques, inflorescences, leaves, roots, and stems of wild-type plants showed comparable levels of ERA1 in all tissues (data not shown). Taken together, these results led us to conclude that ERA1 is expressed essentially ubiquitously in all actively growing tissues.

Figure 3.

ERA1 mRNA expression pattern as determined by in situ hybridization. (Left) Bright-field images photographed by using Nomarski optics. (Right) Bright-field/dark-field double exposures in which both the toluidine blue-stained tissue and the silver grains developed from the photographic emulsion are visible. Longitudinal sections of Ler inflorescences were hybridized with either a sense (A and B) or antisense (C and D) ERA1 probe. The sense control (B) showed no signal above background (regions of the slide where no tissue is present). The antisense probe revealed ERA1-specific signal throughout inflorescence tissue (D), including the inflorescence meristem, developing flowers, and the stem. The antisense probe also was hybridized to longitudinal sections of 7-day-old Ler seedlings (E and F) and various stage Ler embryos including heart (G and H) and bent cotyledon (I and J) stages. ERA1 mRNA is expressed throughout the vegetative shoot apical meristem and developing leaves in seedlings and throughout the embryo in all embryonic stages examined. (Scale bar, 100 μm.)

Discussion

WIG Encodes FTase β.

We have cloned the WIGGUM gene in Arabidopsis and discovered that it encodes a protein FTase β-subunit. As this gene was previously isolated as ERA1, we have renamed the wig alleles (wig-1, wig-2, and wig-3; ref. 18) with era1 designations (era1–4, era1–5, and era1–6, respectively). Low stringency genomic DNA gel blot analysis and Arabidopsis sequence database searches suggest that only one FTase β gene exists in Arabidopsis (data not shown).

Farnesylation, and isoprenylation in general, provides a mechanism to increase the hydrophobicity of substrate proteins posttranslationally. Such a lipophilic covalent modification facilitates membrane localization and protein–protein interactions, both important for signal transduction. In yeast, prenyltransferase activity by FTase or GGTase-I has been shown to be essential: RAM2, RAM1, and CDC43 encode the common α-subunit and the β-subunits of FTase and GGTase-I, respectively (1, 2). Although cells carrying a RAM1 deletion are viable, ram1 cdc43 double mutants are inviable, and deletion of RAM2 is also lethal (1).

It seems likely that loss-of-function mutants of FTase and GGTase-I α- and β-subunit genes in Arabidopsis would yield similar results to those obtained in yeast. Loss-of-function era1 mutants are viable, indicating that FTase β is not essential in Arabidopsis. The FTA gene, encoding the FTase and GGTase-I α-subunit, has been identified in Arabidopsis (35), but no mutational studies have been reported. Although the functional specificity of FTase and GGTase-I is generally quite strong, cross-specificity between these closely related enzymes has been detected in yeast (2). By analogy, it seems reasonable to speculate that the common α-subunit may be essential in Arabidopsis, as it is in yeast, and that GGTase-I likely compensates to some extent for the loss of FTase β in era1 mutants, although additional mutational studies will be required to test this hypothesis.

ERA1 Is Expressed Throughout Development.

In situ RNA hybridization revealed nearly ubiquitous expression of ERA1 throughout development in Arabidopsis. ERA1 expression was detected in the earliest stages of embryogenesis, in embryos undergoing the first division to produce two cells, and continues to be detectable throughout embryogenesis, with all tissues exhibiting expression. ERA1 continues to be expressed in actively growing and dividing tissues throughout vegetative and reproductive growth, with the highest levels detected in shoot and floral meristems. This broad expression domain suggests that ERA1 is not markedly regulated at the transcriptional level throughout development. The precise regulation of substrate proteins for FTase or availability of farnesyldiphosphate or the α-subunit are likely to be responsible for FTase activity on target proteins. Consistent with this idea is the observation that constitutive overexpression of ERA1 in otherwise wild-type plants results in no detectable phenotype.

Our ERA1 RNA expression results are consistent with those obtained for FTA and FTB RNA, protein, and FTase activity in both pea and tomato (10, 12, 36). Furthermore, the broadly distributed ERA1 expression pattern correlates well with the observation that era1–4, era1–5, and era1–6 mutants are defective in multiple aspects of growth and development (18). A role for FTase activity in abscisic acid-mediated responses of seed dormancy and drought resistance has been shown from the initial studies of era1–1, era1–2, and era1–3 mutants (6, 8). The pleiotropy exhibited by era1–4, era1–5, and era1–6 mutants reveals an additional requirement for farnesylation in controlling multiple, specific developmentally important signaling pathways including flowering time, senescence, internode elongation, phyllotaxy, sepal initiation, and meristem size maintenance (18).

Farnesylation in Plants.

Very little is known about the targets of prenylation in plants, although several lines of investigation are providing clues. To date only one plant protein, ANJ1 (the Atriplex nummularia DnaJ homolog), has been shown to be farnesylated in vivo (37). Additional farnesylated proteins have been identified in vitro by expression library screening (38). Several geranylgeranylated proteins also have been isolated in plants, including a subfamily of Rho-related GTPases from plants, or Rop proteins (39, 40). A Rop protein was identified in a CLV1/CLV3 signaling complex (41), implicating geranylgeranylation in meristem control. Furthermore, computer-assisted database searches for potential isoprenylation substrates in plants has led to the identification of numerous candidate proteins involved in diverse biological processes. These include proteins involved in DNA binding and transcriptional regulation, Ras-related GTP-binding and calcium signaling, cell cycle regulation, cell wall modification, and hormone, defense, and stress responses (5, 42). As the Arabidopsis genome sequencing project nears completion, additional candidates for protein prenylation will likely be identified for further study.

A large task for the future lies in the characterization of farnesylated proteins in the signal transduction pathways involved in meristem control and in the various other developmental pathways implicated by the era1 mutants. The pleiotropy of era1–4, era1–5, and era1–6 suggests that multiple, specific developmental processes rely on farnesylation. The expression library screening technique for the identification of farnesylated proteins by Biermann et al. (38) presents itself as a method of great potential, particularly because expression libraries could be made from specific Arabidopsis tissues at specific times during development. For example, using this technique for a vegetative or reproductive meristem-specific library would help to identify possible signaling proteins that require farnesylation to act in the control of meristem size maintenance.

Abbreviations

WIG

WIGGUM

ERA1

enhanced response to abscisic acid

FTase

farnesyltransferase

GGTase

geranylgeranyltransferase

Ler

Landsberg erecta

CAPS

cleaved amplified polymorphic sequences

RACE

rapid amplification of cDNA ends

CLV

CLAVATA