Plant Development: From Cell Fate to Organ Formation

HORMONE SIGNALING

Hormones control not only many aspects of development, but also responses of plants to the environment. One of the emerging themes is the convergence and interaction of several hormone pathways. Jérôme Giraudat (Gif-sur-Yvette) discussed several genetic screens to identify new Arabidopsis loci mediating the responses to the hormone abscisic acid (ABA). Screens for enhancers and suppressors of ABA insensitive mutants led to the isolation of ethylene response mutants. Incorporating recent work from several laboratories, Giraudat presented a model in which ABA acts downstream of hexose sensing to affect seed germination, and the effects of ABA are modulated both by gibberellins and ethylene.

The effects of gibberellins on Arabidopsis development are mediated by a subgroup of GRAS transcription factors, including GAI and RGA, which are distinguished from other family members by an N-terminal extension (Nick Harberd, Norwich). Partial deletion of this N-terminal extension in GAI causes constitutive repression of gibberellin responses, resulting in dwarf plants. Similar deletions are found in the GAI ortholog of wheat, appropriately called Reduced height (Rht). The dwarfism caused by the dominant Rht alleles was the basis for the green revolution. While the agronomic utility of GAI and its orthologs has been aptly demonstrated, it is still unclear how GAI affects gibberellin responses at the molecular level. GAI has some similarity to the mammalian family of STAT transcription factors, many members of which reside in their inactive form in the cytoplasm, and are translocated into the nucleus when activated. Experiments to determine if gibberellin treatment similarly affects the subcellular localization of GAI are in progress. It has already been shown that gibberellins are required for nuclear localization of PHOR1, a potato protein with armadillo repeats (Jaime Martínez-García, Barcelona). Interestingly, gibberellins regulate PHOR1 at several levels, since its RNA is upregulated upon gibberellin treatment, which in turn is dependent on photoperiod. Gibberellins inhibit tuberization in potato, and PHOR1 is apparently like GAI and RGA a negative regulator of gibberellin response, since PHOR1 antisense plants are semidwarves that tuberize early. Another gene expressed differentially during tuberization was described by Christian Bachem (Wageningen). Upregulation of this gene, CB12, which encodes a protein with similarity to short chain alcohol dehydrogenases, closely correlates with tuber formation, and overexpression experiments confirmed that CBI2 has a regulatory role in tuber formation. The overexpressing plants are also spindly, mimicking the effects of exogenous gibberellin application, which provides a further possible link between gibberellin signaling and tuberization.

The gibberellin antagonist ABA acts also antagonistically with brassinosteroids, as discussed by Joanne Chory (La Jolla), whose laboratory has identified two Arabidopsis transcription factor genes that integrate brassinosteroid and ABA signals. These factors act downstream of the BRI1 leucine-rich receptor kinase, a candidate receptor for brassinosteroids. Chory described a series of elegant experiments proving that BRI is indeed the brassinosteroid receptor. First, bri1 mutants have reduced brassinosteroid binding, while overexpression of a BRI1:GFP fusion protein increases binding. Second, radioactively labeled brassinosteroid can be co-immunoprecipitated with the BRI1: GFP protein from plants. Third, brassino-steroid treatment induces BRI1 auto-phosphorylation. How the changes in BRI1 phosphorylation ultimately regulate downstream targets is not known, but candidates for signaling components have been identified through mutants that are resistant to the effects of the brassinosteroid biosynthesis inhibitor brassinazole.

In contrast to brassinosteroids, receptors for most other major plant hormones have not been identified. A candidate for a hormone receptor is GCR1, an Arabidopsis protein similar to G-protein coupled receptors from fungi and mammals (Richard Hooley, Long Ashton). Initial results suggested that this receptor has a role in cytokinin signaling, but this seems less likely now, as a knockout mutant has normal cytokinin sensitivity. However, GCR1 can couple to chimeric G protein alpha subunits in yeast, confirming that it plays a role in transducing a—yet to be identified—signal.

Most plant hormones act systemically, but are not necessarily synthesized throughout the plant. This is especially true for auxins, in which transport involves efflux carriers of the PIN family. Two of these have been shown before to be required for auxin-dependent processes in the shoot and root, and Klaus Palme (Cologne) presented work showing that knockouts of two other family members, PIN3 and PIN4, also cause auxin-related phenotypes. He showed beautiful images of PIN protein localization, including predominantly lateral localization of PIN3 in endodermal cells. The practical value of being able to manipulate auxins was discussed by Angelo Spena (Verona), who has been able to improve fruit quality and productivity in several vegetable and fruit species by inducing parthenocarpy, using ovule-specific expression of the bacterial iaaM auxin biosynthetic gene in transgenic plants. Parthenocarpy has several applications, including increasing the size of tasty fruits from wild plants; inducing fruit development during seasons in which plants do not easily self-fertilize; and producing seed-less fruits. Contrary to conventional wisdom, these results were achieved in the absence of mutations that cause male or female sterility.

LIGHT SIGNALING

One the most important external cues for plants is light. While the phytochrome family of red-light photoreceptors has been known for a long time, it has only been very recently that the molecular mechanisms of phytochrome signaling have been elucidated. Some of these signaling pathways have turned out to be surprisingly short (Peter Quail, Berkeley). Phytochrome B can interact with a basic helix-loop-helix protein, PIF3, both on and off DNA. Importantly, phytochrome B binds only to PIF3 in its biologically active Pfr form, but not in the inactive Pr form. In contrast to PIF3, the related HFR1 protein does not bind to either phytochrome A or B, even though HFR1 also acts downstream of phytochromes in mediating light response. Phytochromes are likely to modulate HFR1 activity through PIF3, which can heterodimerize with HFR1. Given the large number of basic helix-loop-helix proteins encoded in the Arabidopsis genome, there is an enormous potential for interactions of hetero- and homodimers of these proteins with the various phytochromes, providing a possible mechanism through which phytochromes can relatively directly affect a large number of target genes. Support for a role of phytochromes in transcriptional regulation comes from recent reports on dynamic, light-regulated nuclear localization of phytochromes (Eberhard Schäfer, Freiburg). However, phytochromes interact not only with DNA binding proteins. Schäfer reported that ARR4, a member of the response regulator-like family of proteins, interacts with phytochrome B both in vitro and in vivo, and that ARR4 stabilizes phytochrome B in its active, Pfr form. Another nuclear effector of phytochrome A was identified through mutations in the EID1 gene, which has just been cloned.

An important readout of light signaling is elongation growth, and Monica Carabelli (Rome) described the role of homeodomain-leucine zipper proteins in this process. The gene encoding one of these proteins, ATHB-2, is rapidly induced under light with a low red to farred ratio, which mimics shading by neighboring plants. As part of the phytochrome-dependent shade avoidance response, plants elongate in the shade, and overexpression of ATHB-2 was shown to be sufficient to induce elongation growth. As with hypocotyl elongation that is induced as part of the shade avoidance response, hypocotyl elongation induced by ATHB-2 requires polar auxin transport. Based on transgenic phenotypes and expression analyses, Carabelli proposed that other members of this family regulate similar processes as ATHB-2, but in response to other signals.

Another group of genes acting downstream of photoreceptors comprises the COP/DET/FUS genes, which suppress light-dependent development (Xing-Wang Deng, New Haven). Many of the COP/FUS proteins form a multimeric protein complex, known as the COP9 signalosome (originally called the COP9 complex). The similarity of the COP9 complex with the lid of the 26S proteasome suggests a role in ubiquitin-mediated degradation, and another protein, COP1, may act as a ubiquitin ligase with a regulatory role in this process. One substrate of COP1-targeted degradation is the HY5 transcription factor, which is phosphorylated by a light-regulated kinase. In the light, HY5 is predominantly unphosphorylated, which increases both its affinity for target promoters and for the interacting COP1 protein. It is also the unphosphorylated form of HY5 that is preferentially degraded. However, because COP1 is predominantly found in the cytoplasm in light, nuclearly localized HY5 accumulates in the light. COP1 can interact with several distinct substrates through its WD-40 repeats, and Deng discussed specific point mutations in the WD-40 repeats that either enhance or reduce interaction with subsets of COP1 partners. The relevance of these differential interactions was confirmed in vivo, where overexpression of the point mutants induced different, specific gain-of-function or loss-of-function phenotypes.

Another member of the COP/DET/FUS class is encoded by the tomato ortholog of the Arabidopsis DET1 gene (Chris Bowler, Naples). Although DET1 homologs exist in humans and flies, their biochemical function has remained enigmatic. A possible role in chromatin modification has been revealed through the ability of tDET1 to bind the N-terminal domain of one of the histones. Given that endogenous DET1 levels are very low, it is possible that the interaction with histones points to an enzymatic role of DET1 in chromatin modification. Bowler also described his very interesting research on marine diatoms, which provide about one quarter of all fixed carbon on the planet. In contrast to land plants, diatoms live in an environment that has only very little red light. Interestingly, although red light is rapidly extinguished as sun light travels through the water column of the ocean, red light intensities at depths of up to 100 meters are similar to those inducing low and very low-fluence responses in land plants. This red light is likely due to fluorescence of photosynthetic pigment, and might be exploited by diatoms in a neighbor-sensing quorum mechanism. It will be exciting to find out whether diatoms use phytochrome-like molecules for such responses.

STRESS SIGNALING

Because plants cannot move, they have to respond to environmental stress by modifying cellular and developmental parameters. One class of transcription factors that plays an important role in adjusting metabolic pathways in response to environmental stress is the Myb family (Chiara Tonelli, Milan). Because few Myb factors have been identified in mutant screens and because only a minority of knockout mutants has obvious phenotypes, a European consortium aims to characterize all Arabidopsis Myb genes at several different levels. Most Myb genes are under exquisite transcriptional control themselves, and by comparing their RNA expression profiles under different conditions with that of known response genes, it has been possible to assign several Myb genes to specific stress response pathways. A large-scale genomic approach to stress response is also taken by a US consortium, whose efforts were discussed by Ray Bressan (West Lafayette). In one experimental approach, several hundreds of thousands of transgenic plants have been generated, to isolate tagged loss- and gain-of-function mutants with altered stress responses. Perhaps the most impressive slide of the meeting showed one of Bressan’s coworkers holding a flask containing more than 10 liters of seeds from Agrobacterium-treated Arabidopsis plants. This batch, which comprised hundreds of millions of seeds, contained more than a million transformed seeds, aptly demonstrating the power of Arabidopsis transformation.

A surprising route to salt stress tolerance was discussed by Dirk Inzé (Ghent). Plants contain homologs of many cell cycle regulators that were first described in yeast and mammals, including several families of cyclins and cyclin dependent kinases (CDKs). To test the role of CDK Inhibitors (CKIs) in arresting cell cycle progression in response to abiotic stress, a CDK A variant that can no longer be phosphorylated by the CKI WEE1 was introduced into plants. These plants showed increased resistance of salt stress compared with wild-type plants. Another class of CKIs comprises the p27^KIP-related proteins (KRPs), one of which shows a particularly interesting expression pattern in the epidermis of young leaf primordia, but not in the shoot apical meristem. Overexpression of this KRP isoform caused a remarkable leaf phenotype, with larger cells and reduced ploidy levels.

A ubiquitous response to heat stress is the transcriptional activation of heat shock protein (HSP) genes encoding chaperones. However, these proteins play also important roles under other conditions. For example, HSP90 buffers against morphogenetic variation in Drosophila, and mutating HSP90 reveals phenotypic variation caused by small genetic differences that on their own do not cause phenotypic differences in wild-type animals. Christine Queitsch (Chicago) discussed that a similar phenomenon exists in plants, as treatment with the HSP90 inhibitor geldanamycin can induce many developmental abnormalities in Arabidopsis. Different effects were observed in different accessions and in different recombinant inbred lines, revealing a genetic component in Arabidopsis as well.

PLANT–MICROBE INTERACTIONS

Plants respond not only to abiotic, but also to biotic components of their environment. For example, many plants can form root symbioses either with fungi or bacteria, which requires complex signaling between microbes and plants. Rhizobium bacteria produce Nod factors, which induce a number of sequential responses in the host plant that are required for productive infection by Rhizobium, beginning with root hair curling tip growth and culminating in nodule formation (Ton Bisseling, Wageningen). Use of Nod factors with a fluorescent tag showed that Nod factors became predominantly associated with the cell wall, whereas very little Nod factor was associated with the plasma membrane of host cells. By applying a new single-molecule detection technique, fluorescence correlation microscopy, it could furthermore be demonstrated that Nod factors become up to fifty-fold concentrated in the cell wall of root hairs and are markedly immobilized. In this way, local secretion of Nod factors might provide positional information to direct root hair curling. Biochemical approaches to identifying Nod-factor receptors have so far not been successful, but genetic approaches to dissect the downstream events of Nod factor binding to the plant cell wall are very promising. Natural variants of pea that respond differently to different Nod factors have been identified and the responsible loci placed on the genetic map. Positional cloning of such alleles is possible by exploiting synteny with the model legume Medicago truncatula, for which many genomic resources are available.

Genes involved in the early events of nodulation have also been isolated from the semi-aquatic tropical legume Sesbania rostrata, in which stem nodulation occurs from dormant lateral root meristems, which are present at predetermined positions (Marcelle Holsters, Ghent). Among the genes identified because of their differential expression upon nodulation is an acidic chitinase that can degrade specific Nod factors in vitro. Another gene, which encodes a chitinase homolog that has lost its hydrolytic activity, is activated during early stages of the infection. Experiments to test whether this protein is involved in Nod factor binding are under way.

The formation of nodules with a longer life time can increase plant productivity, and one promising avenue is by increasing the delivery of auxin from endosymbiontic bacteria (Roberto Defez, Naples). This can be achieved by expressing auxin biosynthetic genes from Agrobacterium rhizogenes in Rhizobium under control of the rolA prokaryotic promoter. Infecting both temperate and tropical legumes with Rhizobium carrying the appropriate constructs led to an increase in nitrogen fixation, plant dry weight and seed production.

Apart from symbionts, plants have to respond to a large number of pathogenic bacteria. Nitric oxide is an important signaling molecule in this process (Francesca Fenzi, Freiburg). Using pharmacological experiments and plant cells expressing the calcium indicator aequorin, it was shown that calcium fluxes from both extra- and intracellular pools function downstream of nitric oxide.

EMBRYO AND SEED DEVELOPMENT

In contrast to similar genetic screens in animals, genetic screens for patterning mutants in Arabidopsis have turned up only a limited number of genes that are directly involved in early patterning events in the embryo. One possible reason for this is that many patterning mutants never develop to the seedling stage, at which mutants in previous screens were scored. As an alternative, the group of Chris Somerville (Stanford) has conducted an extensive screen for embryonic lethal mutants in Arabidopsis, collecting more than 7000 different lines, each of which was examined for early patterning defects by differential interference contrast microscopy. Some 300 lines showed detectable defects at the heart stage, and were examined further. For most loci, many alleles were recovered, indicating that the screen has been largely saturated. Somerville described several classes of mutants. One of the mutants with the clearest patterning defect was monopole, with early basal defects. Surprisingly, some mutant individuals can recover and form a relatively normal seedling from the upper tier of the octant-stage embryo, underscoring the plasticity of embryo development in plants. The gene has been cloned and found to encode a transcription factor. Relatively normal adult plants can stochastically produced by two other mutants, yoda and grave2, both of which show abnormal patterns of cell division during very early stages. Finally, a large class of mutants produced radially swollen embryos, resulting from direct or indirect effects on cellulose synthesis and thus an aberrant cell wall.

A different set of cell wall mutants was discussed by Herman Höfte (Versailles). A central enzyme in cell wall formation is cellulose synthase, and different isoforms are encoded by a moderately large family of genes in Arabidopsis. Interestingly, the different isoforms have distinct functions, and mutations at several cellulose synthase loci have been isolated in forward genetic screens. Some loci are required for primary cell wall formation, while others are required for secondary wall formation in the vasculature. Mutations in one of these loci, PROCUSTE, were originally identified because they cause a dark-specific effect in hypocotyl cell elongation. PROCUSTE RNA is expressed in both the dark and the light, indicating that bypass of the procuste phenotype in the light is mediated by a light-dependent activity. Experiments with light of different wavelengths and with double mutants showed that this light-dependent activity requires PHYA function.

Sacco de Vries (Wageningen) reported on a leucine-rich repeat transmembrane kinase, SERK, which can homodimerize in the plant plasma membrane, as shown by fluorescence-resonance energy transfer (FRET) with fusions of SERK to different fluorescent proteins. SERK was originally identified in carrot, where its expression marks embryogenic competence of cells in culture. The Arabidopsis SERK gene is expressed from before fertilization in the ovule primordium and in the female gametophyte, and expression persists throughout the embryo and suspensor until the heart stage. Ectopic overexpression of SERK increases embryogenic potential, demonstrating that SERK not only is a marker of embryogenic competence, but that it also has an instructive role in this process. Another gene expressed in ovules was identified by molecular means in tobacco and found to encode a carpel-specific ACC oxidase isoform (Domenico De Martinis, Rome/Nijmegen). Reducing expression of this gene causes an arrest of ovule development. The specificity of this phenotype was confirmed by rescuing the phenotype using an exogenous ethylene source, indicating that ethylene is important for ovule development at least in some species.

In mammals, certain genes are imprinted, such that only the paternally or maternally inherited allele is expressed in the embryo. Imprinting exists also in Arabidopsis, and one of the imprinted genes is MEDEA (MEA), which encodes a SET-domain protein of the Polycomb group (Ueli Grossniklaus, Zürich). During early stages of seed development, MEDEA is expressed only from the maternal allele in both endosperm and embryo, where it represses growth. Because imprinting-based silencing of the paternal allele may involve DNA methylation, the effect of several mutants differentially affecting methylation and transcriptional gene silencing phenomena were examined. For instance, experiments with DDM1, a chromatin remodeling factor gene required for DNA methylation, showed that embryos inheriting a maternally inherited mutant mea allele can survive when they are zygotically homozygous for ddm1. The interpretation is that, in the absence of DDM1, the paternally inherited non-expressed wild-type MEA allele becomes reactivated. ddm1, however, does not interfere with subsequent inactivation, as the paternal allele is initially silent again in the next generation. It is known that demethylated cytosines do not seem to be remethylated in ddm1 mutants, and experiments with methylation-sensitive restriction enzymes did not reveal changes in the methylation pattern across the MEA locus. Thus, it is likely that the observed reactivation and inactivation of the paternal MEA allele is independent of methylation and related to chromatin remodeling.

Maternal contribution to seed development was also discussed by Paolo Costantino (Rome). His group identified a zinc finger transcription factor of the Dof family, DAG1, which acts in maternal tissue to repress dormancy. dag1 mutant seeds are still sensitive to the inhibitory effects of ABA on germination, and require gibberellin biosynthesis for germination, confirming that DAG1 acts relatively far upstream in controlling germination. Interestingly, DAG1 is expressed in the vascular tissue (of the mother plant), suggesting that DAG1 controls the transport of a substance required to confer dormancy. Moreover, that dag1 mutant seeds do not require light for germination suggests that the activity of this substance is light regulated. Another pair of transcription factors that control seed maturation, ABI3 and FUS3, was discussed by Vered Raz (Wageningen). Raz showed that growth arrest and cessation of cell division in the embryo and endosperm are regulated by FUS3 and precede the induction of dormancy, which is regulated by ABI3 and ABA. Embryo growth arrest can be uncoupled from seed dormancy genetically, and aberration in these two sequential pathways causes vivipary.

POSTEMBRYONIC DEVELOPMENT

A developmental mechanism that has been intensively studied in animals is lateral inhibition, which allows specialized cells to emerge from a uniform field of undifferentiated cells at more or less even spatial intervals. The epidermis of dicot leaves contains two types of specialized structures, trichomes and stomata, whose distribution is likely to be at least partially regulated by lateral inhibition. A candidate for a protein involved in lateral inhibition of stomata was identified through Arabidopsis mutants with increased stomatal density (Thomas Altmann, Golm). This gene, SDD1, is expressed at low levels in the mesophyll of very young leaf primordia, and more strongly in guard mother cells and young, developing guard cells, consistent with SDD1 being involved in generating a negative signal emanating from these cells. The similarity of the SDD1 protein to subtilisin-like serine proteases suggests that SDD1 acts by activating a peptide ligand.

The question of whether plants make cells or cell make plants, was addressed by Andrew Fleming (Zürich). His laboratory has combined a tetracycline-inducible system for overexpression of cyclin A with local induction using tetracycline-loaded beads. Surprisingly, spatially restricted induction of cyclin A in the shoot meristem of tobacco had little effect, but placement of the tetracycline-loaded beads on young leaf primordia caused ectopic tissue to form, indicating that the role of cell division in morphogenesis is context dependent.

One of the most studied events in postembryonic development is the switch from the vegetative to the reproductive phase. The two most important environmental cues controlling flowering are vernalization and daylength. A requirement for vernalization, or exposure to a prolonged period of cold temperature, ensures that plants growing in Northern latitudes overwinter vegetatively and flower only in spring (Caroline Dean, Norwich). A locus determining vernalization requirement in wild accessions of Arabidopsis the FRIGIDA (FRI) gene. Accessions with a functional FRI allele are more common at Northern latitudes, suggesting that loss of FRI function was a major event in the loss of vernalization requirement in Southern accessions. Comparison of sequence polymorphisms at the FRI locus confirmed that the ancestral form of Arabidopsis likely had a functional FRI allele, and that early-flowering forms arose independently at least twice. It is not yet known how plants sense the cold temperature signal, but two loci required for the vernalization response, VRN1 and VRN2, have recently been cloned and found to encode apparent transcription factors.

Remarkable progress has also been made in understanding how Arabidopsis senses long days, which promote flowering. A central factor in this process is the CONSTANS gene (George Coupland, Norwich). CONSTANS encodes a protein with zinc fingers of the B-box type, which are known to mediate protein-protein interactions among mammalian transcription factors. CONSTANS RNA accumulation is circadian clock-regulated, with peak levels during the night. The CONSTANS peak is broader under long-day conditions than under short-day conditions, such that high CONSTANS RNA levels are still present at dawn in long-day grown plants, but not in short-day grown plants. Excitingly, CONSTANS protein function seems to be light regulated, as CONSTANS overexpression is lethal, when plants are grown in blue light, which is known to have a strong promoting effect on flowering. A simple model derived from these observations is that long-day dependent flowering occurs only when CONSTANS protein is present during the light phase of the day. Which photoreceptors regulate the circadian expression of CONSTANS RNA is not known, but their effects are mediated by several genes such as GI, LHY, CCA1 and ELF3. These genes are part of the general clock machinery, as they affect circadian expression not only of CONSTANS, but also of other genes not involved in flowering.

In Arabidopsis, the action of flowering-time genes including FRI and CONSTANS ultimately results in the upregulation of floral meristem-identity genes such as LEAFY. Detlef Weigel (La Jolla) discussed the use of activated versions that make a transcription factor independent of other coregulators, which presents a powerful method to probe the interaction of a transcription factor with its targets in vivo. This approach was used to demonstrate that homeotic genes such as APETALA1 and AGAMOUS are direct targets of LEAFY, which was subsequently confirmed by several other methods. Interestingly, the LEAFY transcription factor can move between cells, which raises the question of how LEAFY activity is limited. One possible mechanism involves other, non-mobile transcription factors that repress LEAFY targets in the shoot apical meristem. Orthologs of LEAFY as well as its coregulator UFO have been isolated in several other species including petunia (Erik Souer, Amsterdam). However, in contrast to Arabidopsis, the loss-of-function phenotype of the UFO ortholog DOT is as severe as that of the LEAFY ortholog ALF. DOT is expressed only in a subset of cells that express ALF, pointing to non-autonomous effects of DOT on floral identity. A feature that makes petunia particularly interesting is its cymose inflorescence, in which the meristem bifurcates at each node, generating a floral meristem and a new inflorescence meristem. This bifurcation requires the activity of the EXP gene, which encodes a MADS domain protein.

In summary, the workshop presented an excellent overview of many important directions in contemporary plant developmental biology. Perhaps the most impressive aspect of the workshop was that the vast majority of presentations dealt with molecular mechanisms that went beyond the mere cloning of genes. This together with the beautiful location and the intimate format, which mixed both senior and junior scientists, made the meeting a particularly successful and memorable one.