Report

Regulation of Vegetative Phase Change in Arabidopsis thaliana by Cyclophilin 40

See allHide authors and affiliations

Science  23 Mar 2001:
Vol. 291, Issue 5512, pp. 2405-2407
DOI: 10.1126/science.1057144

Abstract

During its development, a plant shoot progresses from a juvenile to an adult phase of vegetative growth and from a reproductively incompetent to a reproductively competent state. InArabidopsis, loss-of-function mutations in SQUINT(SQN) reduced the number of juvenile leaves and had subtle effects on inflorescence morphology but had no effect on flowering time or on reproductive competence. SQN encodes theArabidopsis homolog of cyclophilin 40 (CyP40), a protein found in association with the Hsp90 chaperone complex in yeast, mammals, and plants. Thus, in Arabidopsis, CyP40 is specifically required for the vegetative but not the reproductive maturation of the shoot.

All plants go through qualitatively different stages of vegetative development before they enter the reproductive phase (1, 2). In some species, the vegetative phases are marked by dramatic differences in a large number of traits, whereas in others the transition is more subtle (3, 4). This phenomenon is known as heteroblasty (5) or phase change (6). Mutations affecting phase-specific features of vegetative development have been identified in a number of plants, but only one gene involved in the regulation of this process has been identified on a molecular level (7). Thus, the molecular mechanism of phase change remains almost completely unknown.

In Arabidopsis, leaves produced during the development of the rosette can be distinguished from one another by their size and shape (8), number of hydathodes (9), and pattern of trichome (leaf hair) distribution (10, 11). The first two leaves are small and round, have only a few hydathodes, and do not produce trichomes on their abaxial (lower) surface. As the shoot develops, leaf size increases, the leaf blade becomes elliptical and more serrated, hydathode number increases, and leaves begin to produce abaxial trichomes. To identify genes that regulate this transition, we screened for mutations that cause these changes to occur precociously, focusing in particular on the production of abaxial trichomes, because this trait is relatively easy to score and quantify. Three mutant alleles of SQN were identified in this screen. Two of these alleles were induced in the Columbia (Col) ecotype, one by ethyl methane sulfonate (sqn-1) and the other by carbon ionization (sqn-3). A third allele (sqn-2) was induced by transferred DNA (T-DNA) transformation in a Wassilewskije background and was crossed six times to Col before being used for the studies described here. In a Col background, these alleles were phenotypically identical.

The effect of sqn mutations on leaf identity was particularly obvious in leaves one and two, because these leaves are normally quite different from all other rosette leaves. In mutant plants, leaves one and two were elliptical rather than round and were larger and more serrated than the first two leaves of wild type (WT) plants (Fig. 1, A and B); leaves at other positions on sqn shoots also resembled leaves at higher nodes on WT plants. Col plants normally begin producing abaxial trichomes on leaf 5 and do not produce trichomes under leaf one or two (10, 11). In contrast,sqn plants usually produced abaxial trichomes on leaf three and produced abaxial trichomes on leaf two when grown in high light (Fig. 1B and Table 1). sqnmutations also accelerated the normal increase in both adaxial (upper leaf surface) and abaxial trichome density on successive rosette leaves (Fig. 1E). These changes were accompanied by a decrease in total leaf number that was largely attributable to a reduction in the number of leaves lacking abaxial trichomes (that is, juvenile leaves) (Table 1); sqn did not affect, or only slightly reduced, the number of leaves with abaxial trichomes (adult leaves). Thus, sqn accelerated vegetative phase change rather than causing a homeotic transformation of juvenile leaves into adult leaves. In addition, sqn disrupted the phyllotaxis of the inflorescence (Fig. 1C) and increased carpel number in the first few siliques (Fig. 1D). These floral phenotypes were displayed by 88 to 100% of mutant plants, but their expressivity was weak and variable. Although the effect of sqn mutations on vegetative and reproductive morphology was recessive, their effect on abaxial trichome production was semidominant; on average, sqn/+ plants produced abaxial trichomes 0.5 leaves earlier than didsqn/sqn plants. Because these mutations probably produce a loss of function, this observation suggests thatSQN may be haplo-insufficient for abaxial trichome production.

Figure 1

Phenotype of sqn and Col plants. (A) Col (left) and sqn-1 (right) plants at 8 days after planting. (B) Rosette and inflorescence leaves from Col (top) and sqn-1 (bottom) plants. Light gray, no abaxial trichomes; dark gray, abaxial trichomes; black, inflorescence leaves. (C) Col (left) and sqn-1 (right) inflorescences. (D) Col (top) and sqn-1 (bottom) siliques. The arrowhead indicates an abnormal third carpel in the sqn-1silique. (E) Trichome density on the adaxial and abaxial surfaces of Col (diamonds) and sqn-1 (squares) plants. (F) The rate of leaf initiation (solid symbols) and leaf emergence (open symbols) in Col (squares) and sqn-1(circles) plants.

Table 1

The effect of sqn-1 and 35S::LFYon vegetative development and flowering time. Values are the mean ± SE.

View this table:

Mutant seedlings initially appeared “eyelike” (Fig. 1A) because immature leaves in the rosette were unusually small relative to leaves one and two. To determine the basis for this phenotype, we examined the effect of sqn-1 on the rate of leaf initiation and leaf emergence (12) (Fig. 1F). We found thatsqn-1 had no effect on the initiation or time of emergence of the first two leaves but transiently delayed leaf production after the initiation of leaf two; sqn-1 had no effect on the rate of leaf initiation or leaf emergence after this point. sqn-1also caused rosette leaf production to cease earlier than normal. Thus, the unusual morphology of sqn-1 seedlings is attributable to a transient delay in leaf initiation after the production of leaf two, whereas the reduced leaf number of mature sqn-1 plants is due both to this early delay and to a precocious cessation of leaf production.

To determine whether sqn-1 accelerated reproductive phase change in the same way in which it accelerated vegetative phase change, we examined its effect on flowering time and its interaction with a 35S::LFY transgene. In WT plants,35S::LFY induces flowering immediately after plants have made the transition to the adult vegetative phase (13, 14), which implies that vegetative phase change is accompanied by a change in the reproductive competence of the shoot. We found that sqn-1 had no effect on flowering time in either the absence or the presence of35S::LFY (Table 1). sqn-1; 35S::LFY plants flowered at exactly the same time as35S::LFY plants. These results suggest thatSQN is primarily involved in regulating vegetative phase change, and has no effect on the reproductive competence of the shoot.

SQN was mapped to chromosome 2 and cloned using a map-based approach (15). Hybridization of bacterial artificial chromosomes (BACs) to Southern blots of genomic DNA fromsqn-3 demonstrated that this allele is a ≈10-kb deletion encompassing four predicted genes, one of which was found to be mutant in sqn-1 and sqn-2. A 1313–base pair (bp) cDNA (GenBank accession number AY026065) corresponding to this gene rescued the mutant phenotype of sqn-1 and sqn-3 when introduced into plants under the regulation of the CaMV 35Spromoter, confirming that it corresponds to SQN. RNA blots probed with this cDNA (Fig. 2C) revealed a single 1.3-kb transcript that was present at low levels in WT plants and was significantly reduced in sqn-1. Different amounts ofSQN mRNA were detected in seedlings, adult leaves, and floral buds, but it is not clear whether this reflects a tissue-specific pattern of expression or the difference in the ages of the tissues (immature juvenile leaves and floral buds versus expanded adult leaves) used in this experiment.

Figure 2

Molecular analysis of SQN. (A) Exon-intron structure of SQN. The locations of the mutations in sqn-1, sqn-2, andsqn-3 are indicated. sqn-1 is a G → A transition, and sqn-2 is a single base (T) insertion that introduces stop codons at amino acid positions 207 (sqn-1) and 220 (sqn-2). (B) Alignment of SQN with other known CyP40 amino acid sequences (37). The underlined segments represent the position of the TPR repeats described in (17). Gray shading indicates similar residues; black shading indicates identical residues. (C) Northern analysis of SQN expression in 5-day-old seedlings (sdl), 21-day-old rosettes (ros), and flower buds (fl) of WT and sqn-1. SQN mRNA is significantly reduced in sqn-1 plants. (D) Northern analysis of the expression of HSP genes in WT and sqn-1seedlings. There is 1.5 times more HSP81.2,HSP81.3, HSP81.4, and HSC70-1 mRNA insqn-1 than in WT seedlings. This difference was computed after normalizing the HSP81 and HSC70-1 band intensities to the 18S ribosomal RNA signal.

SQN consists of nine exons (Fig. 2A) and is predicted to encode a 361–amino acid protein that is 48% identical and 64% similar to bovine CyP40 (Fig. 2B). SQN is the onlyCyP40-like gene in the Arabidopsis genome. Like other CyP40 proteins, SQN has an NH2-terminal peptidyl prolyl isomerase domain that is very similar to low-molecular-weight cyclophilins (16). In the COOH-terminal half of the predicted protein, the most highly conserved domains correspond to the three tetratricopeptide repeats (TPRs) (17) that are required for the interaction of CyP40 with Hsp90 (18, 19, 20), and are sufficient for the function of the CyP40 genes Cpr7(21) in Saccharomyces cerevisiae andwis2 (22) in Schizosaccharomyces pombe.

In yeast, the CyP40-related gene CPR7 negatively regulates the expression of the heat-shock protein (HSP) geneHSP104 (23). To determine whether Cyp40 has a similar function in Arabidopsis, we studied (24) the thermotolerance of sqn-1 and its effect on the expression of HSP genes that are transcribed under non-stress conditions (HSP81.2, HSP81.3,HSP81.4, and HSC70-1) (25,26) and one that is only transcribed in response to heat shock (HSP101) (27). In three independent experiments, Northern analysis revealed a small but reproducible increase (1.2 to 1.8 times) in HSP81 and HSC70-1expression in sqn-1 but no increase in the expression of HSP101 (Fig. 2D). This minor change in HSP gene expression did not produce a change in the thermotolerance of mutant seedlings. Thus, in Arabidopsis, CyP40 negatively regulates the transcription of some HSPs.

CyP40 and the immunophilin, FK506-binding protein 52 (FKBP52), were originally discovered in unactivated steroid receptor complexes in mammals (17, 28) and compete for a binding site on Hsp90 in both animals (29–31) and plants (19). It is believed that these immunophilins function as protein chaperones either on their own or in association with Hsp90, because they interact directly with a number of signaling molecules (32) and are present in varying amounts in steroid receptor complexes isolated from different tissues (33). Alternatively, they may be involved in the subcellular trafficking of Hsp90 complexes (34). In yeast, the loss of all 12 cyclophilin and FK506-binding proteins produces slow growth but is not lethal (35). Mutations in an Arabidopsis homolog of FKPB52, PASTICCINO1, have a highly pleiotropic phenotype (36), suggesting that this protein is critical for many aspects of plant growth and development. In contrast, CyP40 appears to have a more specific role inArabidopsis development. The sqn mutant phenotype suggests that CyP40 plays a major role in promoting the expression of the juvenile phase of vegetative development and is involved to a more limited extent in regulating the positioning of floral buds, floral morphogenesis, and the expression of HSPs. This phenotype is consistent with the hypothesis that CyP40 regulates the activity of specific signaling pathways. Future studies should reveal the components of these pathways and provide insight into the biochemical function of Cyp40 in Arabidopsis.

  • * To whom correspondence should be addressed. E-mail: spoethig{at}sas.upenn.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article