Light signal transduction: an infinite spectrum of possibilities

SUMMARY

The past 30 years has seen a tremendous increase in our understanding of the light-signaling networks of higher plants. This short review emphasizes the role that Arabidopsis genetics has played in deciphering this complex network. Importantly, it outlines how genetic studies led to the identification of photoreceptors and signaling components that are not only relevant in plants, but play key roles in mammals.

INTRODUCTION

A few key events can shape a scientific life. For me, one of these incidents occurred in 1987, after I had just given my first job seminar. In it, I boldly claimed that I would use Arabidopsis genetics to determine a complete signal transduction pathway from phytochrome to the control of a light-regulated gene. In 2010, this may not seem like an ambitious goal, but in 1987, phytochrome was the only known receptor in plants (Hershey et al., 1984), there was only a single report of successful transformation of Arabidopsis (Lloyd et al., 1986), and the Arabidopsis genetic map consisted of only about 100 phenotypic markers (Koornneef et al., 1983).

It was still the days of slide projectors, so the lights in the small seminar room were dimmed. My job talk was over and it was time for probing questions, but the first speaker did not ask me for clarification; rather, out of the darkness in the back of the room came the proclamation ‘Why, you are a very brave and naive girl! You know, there’s a reason that we all fled from phytochrome research’. The speaker was Harold William Siegelman, an icon in the phytochrome field, who, in the 1960s, not only gave the intriguing photoperiodic pigment its name, but also partially purified ‘holo-phytochrome’ and visualized the blue–green pigment in a cuvette (Butler et al., 1959; Siegelman et al., 1966). Dr Siegelman sat me down with his old notebooks and showed me many puzzling results that scientists could not explain from the 1960’s – observations that could not be explained by the action of a single photoreceptor acting in a linear signaling pathway. A perfect problem for a geneticist! I left Brookhaven with words of encouragement from Dr Siegelman and our agreement that the tools were in place to unravel some of the paradoxes of light signaling. I was extremely excited that I could make a difference in a major area of plant physiology.

Fast forward to the bar at Cold Spring Harbor Laboratory (CSHL) during the summer of 1993. I was with Elliot Meyerowitz, a guest speaker at the first Arabidopsis lab course, and my fellow instructors, Joe Ecker and Sakis Theologis. We were having beers after a long day in the lab. Joe and Sakis were animatedly trying to convince Elliot that we needed to sequence the Arabidopsis genome. Elliot was resisting the idea, voting for individual investigator grants, rather than a multi-national funding initiative. I don’t remember exactly what happened, but at some point Elliot reversed his opinion, and became enthusiastic that sequencing Arabidopsis would be well worth the investment of tens of millions of dollars of public funds. Jim Watson, then head of CSHL, together with the leaders in Arabidopsis research, went on to lobby the US National Science Foundation to take the lead in an initiative that would fund three groups in the United States to sequence chromosomes 1, 2 and 4. The rest is history….

These stories provide a major lesson for young scientists: bringing new approaches to an old problem can lead to major breakthroughs in a field. Mutants clarified the overlapping, and unique, functions of the various phytochromes and led to the discovery of cryptochromes, phototropins and other photoreceptors in plants. The availability of the fully sequenced and annotated genome of Arabidopsis gave us a snapshot of what ‘makes a plant a plant’. It also provided the framework for reverse genetic studies, and greatly facilitated research in crop plants. I also learned that the leaders in a field can provide guidance and advocacy for funding that benefits an entire community.

In this review, I provide a brief history and summarize the current state of knowledge of light-signaling research in plants, focusing on how the sequence of Arabidopsis enabled discovery of the principle components and dissection of the molecular mechanisms. My purpose is not to review the many papers in the literature, but to recount the chronology of some of the major discoveries in an effort to uncover the role that Arabidopsis played in deconstruction of one of the most complex signaling networks in plants. Dissection of light-signaling networks has had profound implications, not only for all flowering plants, but for human biology as well.

A BRIEF HISTORY OF LIGHT SIGNAL TRANSDUCTION PATHWAYS: THE 1980s

The early 1980s marked the dawn of modern plant molecular biology. 1983 was an especially good year: the first reports of stable plant transformation were published (Barton et al., 1983; Caplan et al., 1983; Fraley et al., 1983; Herrera-Estrella et al., 1983; Murai et al., 1983), maize transposons were cloned (Burr and Burr, 1982; Geiser et al., 1982; Fedoroff et al., 1983), and an Arabidopsis genetic map with 100 phenotypic markers scattered across the five chromosomes became available (Koornneef et al., 1983). The icing on the cake was Barbara McClintock’s Nobel prize, awarded for the discovery of transposable elements. There was great excitement regarding the potential for paradigm-shifting discoveries in plant biology, so much that the US National Science Foundation started a post-doctoral program specifically in the area of plant molecular biology. This program was targeted to those new to the field, and would eventually fund over 200 researchers.

It was a very exciting time in the light-signaling field as well. After almost three decades of struggling, a full-length phytochrome A holoprotein was purified from etiolated oat seedlings by the Lagarias and Quail labs in 1983 (Litts et al., 1983; Vierstra and Quail, 1983) (see Figure 1a for timeline). This was followed a year later by cloning of the oat PHYA gene (Hershey et al., 1984). Phytochrome, a red/far-red light photoreceptor, thus became the first plant receptor whose identity was known at the molecular level. Work at the other end of the signal transduction pathway was also making great progress. Several genes encoding the small subunit of RuBisCO were cloned, as well as genes for the light-harvesting chlorophyll a/b binding proteins. Work in multiple laboratories led to the exciting discovery that light positively regulates transcription of these genes (Simpson et al., 1985; Timko et al., 1985; Gilmartin et al., 1990; Lam and Chua, 1990). The response was rapid (shorter than 15 min) and was universal among the angiosperms. Having cloned phytochrome from oat, the Quail lab then went on to show that light also negatively regulates transcription of this gene (Bruce et al., 1989). Elaine Tobin’s lab weighed in with the significant finding that light acts through phytochrome to regulate gene expression in plants (Silverthorne and Tobin, 1984).

Figure 1.

A 30-year timeline of light-signaling research.

(a) Major discoveries in light signaling since 1980 that were made independently of the Arabidopsis system.

(b) Discoveries made during the same period using Arabidopsis.

Note that, during the past decade, it is hard to distinguish Arabidopsis-dependent contributions from independent ones.

By now, Agrobacterium-mediated transformation of the solanaceous plants (tobacco, petunia, tomato, etc.) was routine. The availability of histochemical markers such as GUS, as well as facile transformation protocols, led to the identification of multiple light-regulatory enhancers acting in the promoters of these genes, such as the ‘L’, ‘I’ and ‘G’ boxes identified in Tony Cashmore’s lab, box II, and many more (Donald and Cashmore, 1990). One principle soon emerged: more than one light-regulatory enhancer was required for a gene to become light-regulated (Donald and Cashmore, 1990; Chattopadhyay et al., 1998). Yet, despite these exciting advances, the events between photon excitation of the photoreceptor and regulation of gene expression remained a mystery (Figure 2).

Figure 2.

Model for phytochrome signaling circa 1989.

Lurking in the shadows, a small, insignificant weed was about to burst onto the scene. However, this insignificant weed, Arabidopsis, had not yet made much of an impact on the light-signaling field. A few papers, such as Koornneef’s paper on long-hypocotyl mutants of Arabidopsis, are notable (Koornneef et al., 1980). In this paper, Koornneef described the first large genetic screen for mutants with reduced sensitivity to light. The mutants defined five complementation groups, and led to the identification of cryptochrome, phytochrome’s chromophore biosynthetic enzymes, a downstream transcriptional regulator, and phytochrome B (discussed in more detail below) (see Figure 1b). Later, Arabidopsis genomic libraries proved useful in identifying a phytochrome gene family (five members in Arabidopsis) (Sharrock and Quail, 1989). The decade ended with the description of a class of mutants [de-etiolated (det) mutants] that developed like light-grown plants even in the absence of light (Chory et al., 1989; Chory and Peto, 1990). These loss-of-function mutations were epistatic to the long-hypocotyl mutants, implying that negative regulators existed downstream from light perception that prevented photomorphogenesis from occurring when no light was available (Chory, 1992). Identification of this class of mutants added an important regulatory component to the pathway shown in Figure 2.

ARABIDOPSIS RULES THE 1990s: YOU CAN NEVER HAVE TOO MANY MUTANTS

Arabidopsis dominated light-signaling research during the 1990s (Figure 1b). Mutants in both receptors and positively acting signaling components were identified from genetic screens for light-hyposensitive mutants (seedlings that have long hypocotyls under various wavelengths and fluence rates of light). Mutations in individual phytochromes were characterized, solving some of the dilemmas in Dr Siegelman’s notebooks, while at the same time revealing even more complexity in the light-signaling network (reviewed by Neff et al., 2000; Briggs and Olney, 2001). Higher-order genetic analysis revealed a complex web of interactions within and between the various classes of photoreceptors, including redundancy, antagonism and effector/modulator relationships. Although the mechanisms of these interactions are not all clear, there is mounting evidence for both direct physical contacts between photoreceptors and common interacting partners (Figure 3).

Figure 3.

Genetic model for light signaling circa 2000.

Mutation in early signaling intermediates are expected to have a phenotype only under the specific light conditions activating their photoreceptor, and known molecules fitting this requirement include several transcription factors that directly interact with photoreceptors, as well as other molecules that are phosphorylated by phytochromes. The subcellular localization of phytochrome itself is light-regulated. Thus, in light signaling, there are direct interactions of photoreceptors, as well as cross-talk and integration of pathways both early and late in the signaling network.

Unlike phytochrome, which was purified based on an elegant photoreversibility assay, no receptor for blue light had been identified at the end of the 1980s. Moreover, researchers in this field were caught up in an abstract debate as to what the chromophore for a blue light receptor might be (a carotenoid or a flavin-type molecule), a problem exacerbated by the erroneous notion that there was a single blue light receptor. The first elusive blue/UV-A light photoreceptor, cryptochrome, was identified by cloning of the long hypocotyl 4 (hy4) locus, originally reported in the Koornneef et al., 1980 paper (Ahmad and Cashmore, 1993). HY4, which was subsequently renamed cryptochrome 1 (CRY1), uses an N-terminal domain that is shared with photolyases to bind two chromophores, a flavin adenine dinucleotide with catalytic functions and a deazaflavin or pterin for light harvesting (Cashmore et al., 1999). However, CRY1 has no photolyase activity. Cashmore and colleagues showed that Arabidopsis has two cry-like genes, and performed a brute force screen to identify cry2 mutants (Lin et al., 1998). Like CRY1, CRY2 appears to play a role in blue light perception in hypocotyls. In addition to this role, CRY2 also plays a significant role in flowering time. To date, the photochemistry of the cryptochromes has proven difficult to crack, although the prevalent idea is that cryptochromes act through some sort of redoxdriven reaction. Very recently, Liu et al. identified the first binding partner of CRY2, a nuclear protein called CIB1 that binds to G-boxes in the promoters of light-regulated genes (Liu et al., 2008).

The discovery of the blue light receptor family for phototropism was a decade-long labor of love, mostly by Winslow Briggs’ lab, that culminated in identification of phototropin 1 in 1998 by cloning of the Arabidopsis nph1 locus (non-phototropic hypocotyl 1) (Liscum and Briggs, 1995; Christie et al., 1998). There is irony in this story. I have it from good sources that Winslow strongly resisted working on Arabidopsis, preferring larger plants, such as maize and pea. While these larger plants ultimately contributed to the story, the discovery of phototropin would have been significantly delayed without Arabidopsis mutants. In 1988, Gallagher et al. showed that blue light could activate phosphorylation of a plasma membrane protein from growing regions of etiolated pea seedlings (Gallagher et al., 1988). Various biochemical experiments suggested that this protein could autophosphorylate in response to blue light (Short and Briggs, 1994). At the same time, genetic experiments performed in Briggs’ lab and later in Liscum’s lab, identified a number of phototropism-defective mutants, nph1, 2, 3 and 4 (Motchoulski and Liscum, 1999; Harper et al., 2000). The cloning of NPH1 suggested it to be the photoreceptor for phototropism, as nph1 mutants are impaired in light-activated phosphorylation and phototropism (Christie et al., 1998). Arabidopsis has two phototropins, PHOT1 and PHOT2. Genetic studies indicate that these two phototropins have partially redundant functions. In addition to controlling phototropism, characterization of phot1 phot2 double mutants implied a role for phototropins in blue light-dependent chloroplast relocation (Kagawa et al., 2001), stomatal opening (Kinoshita et al., 2001) and growth (Folta et al., 2003). Not only did the protein have a canonical serine/threonine kinase domain, but it also contained two repeated domains of 100 amino acids, with 40% amino acid sequence identity to each other (Zhulin et al., 1997). These domains also show sequence similarity to domains found in a number of signaling proteins in organisms from all kingdoms of life. Because all these proteins are regulated by light, oxygen or voltage, the domains were assigned the acronym LOV (Christie et al., 2002; Briggs, 2007). LOV domains are a subset of the PAS domain super-family, which is known to mediate both ligand binding and protein–protein interactions.

In addition to the work on photoreceptors, the 1990s was a decade of discovery of light-signaling components (Figure 3) (reviewed by Neff et al., 2000). Genetic and molecular screens, most of which focused on seedling responses, especially those that are part of de-etiolation, identified many dozens of genes acting downstream of photoreceptors. Because different spectral qualities trigger the same developmental responses using different photoreceptors, it was hypothesized that common late-acting signaling intermediates are used. The best-studied class of proteins that fit this description is the class of negative regulators, which, when mutated, cause seedlings to de-etiolate even in the absence of light (the det, cop and fus class). The screen was performed to saturation in Xing-Wang Deng’s lab, culminating in the identification of more than ten ‘cop’ loci (‘constitutively photomorphogenic’) (Kwok et al., 1996). Soon after, most of the COP genes had been cloned. However, it was not until the Deng lab purified the major ‘COP’ complex that it became apparent that these proteins play a role in protein turnover (Chamovitz et al., 1996). Eight of the proteins are members of an evolutionarily conserved complex called the CSN (‘COP9 signalosome’) (Wei et al., 1998). The CSN may have multiple activities, but the best characterized is its ability to cleave and remove the ubiquitin-like protein, Nedd8, from cullin ubiquitin ligase subunits (Wei and Deng, 2003). COP1 is an E3 ligase that targets both labile photoreceptors, e.g. PHYA, as well as downstream effectors, e.g. HY5 (Figure 3) (Yi and Deng, 2005). More recently, several researchers have shown that DET1 is also part of an E3 ligase that is in close association with chromatin, and complexed with COP10, DDB1 and CUL4 (Benvenuto et al., 2002; Schroeder et al., 2002; Yanagawa et al., 2004; Zhang et al., 2008). Also identified were many positively acting signaling components, including transcription factors, chaperones and scaffolds (reviewed by Neff et al., 2000).

The decade ended with two remarkable discoveries. The first is that phytochromes are light-regulated protein kinases. With the discovery of cyanobacterial phytochromes in the mid-1990s, Yeh et al. were able to show quite clearly that bacterial phytochromes are light-regulated kinases (Yeh et al., 1997). Higher-plant phytochromes also have a histidine-kinase related domain, but no histidine kinase activity could be found. Yeh and Lagarias showed that two plant phytochromes are light- and chromophore-regulated kinases, but, unlike their cyanobacterial counterparts, they autophosphorylate on serine/threonine (Yeh and Lagarias, 1998). PHYA is a phosphoprotein in vivo, and at least one serine is phosphorylated in a light-dependent manner. In vitro kinase assays using phytochrome A identified other substrates of PHYA, including CRY1 and CRY2, and some newly discovered signaling proteins (Fankhauser et al., 1999; Fankhauser, 2000).

Yeast two-hybrid screens identified the first potential signaling partners of phytochrome. The first three included a bHLH transcription factor, PIF3 (Ni et al., 1999), a cytoplasmic protein, PKS1 (Fankhauser et al., 1999), and a nucleoside diphosphate kinase, NDPK2, which is localized to both nuclear and cytoplasmic fractions (Choi et al., 1999). We know now that PIF3 is a member of a sub-family of bHLH proteins that bind preferentially to photo-activated phytochrome through a phytochrome interaction domain (Huq et al., 2004; Khanna et al., 2004, 2007; Al-Sady et al., 2006; Monte et al., 2007; Shen et al., 2007). PIF family members control light-regulated gene expression. Light induces the phosphorylation of some PIFs, which is necessary for their rapid degradation (Al-Sady et al., 2006; Shen et al., 2007); however, ten years after their discovery, it is still unclear whether PHY kinase activity is responsible for this degradation or whether phytochrome recruits another kinase to perform the degradation (Castillon et al., 2007; Bae and Choi, 2008).

LIGHT SIGNALING IN THE POST-SEQUENCE ERA: GENETIC REDUNDANCY, MICROARRAYS, CELL AND STRUCTURAL BIOLOGY

The completion of the Arabidopsis genome sequence in 2000 (Arabidopsis Genome Initiative, 2000) has facilitated proteomic studies, design of gene expression and full-genome tiling arrays for global gene expression and epigenetic studies, as well as the positional cloning of genes (Clark et al., 2007; Yazaki et al., 2007; Lister et al., 2009). This greatly increased the speed of new discoveries in terms of new signaling components; in addition, it helped to elucidate the degree to which these proteins may be redundantly encoded (Alonso and Ecker, 2006).

A major step forward for those of us studying photomorphogenesis was the design and creation of various types of platforms to interrogate transcription of the full genome of Arabidopsis. A number of studies were published reporting that hundreds, if not thousands, of genes were induced or repressed by changes in light quality, intensity or dark/light transitions (Tepperman et al., 2004, 2006). Judicious use of various mutants in the analysis helped explain why different photoreceptors have both overlapping and distinct functions (Ma et al., 2003; Tepperman et al., 2004, 2006; Jiao et al., 2007).

The Nagatani lab paper indicating that phytochromes move to the nucleus from the cytosol after excitation by light has completely changed researchers’ views of the signaling pathway (Sakamoto and Nagatani, 1996; Yamaguchi et al., 1999; Nagatani, 2004). Within a few years, it was noted that phytochromes are located in sub-nuclear particles, called phytochrome speckles or phytochrome nuclear bodies (reviewed by Chen, 2008). These nuclear bodies change in size and number with the fluence rate of red light, the total number of photons or the time of day (an indicator of how much active phytochrome there is) PHYB has been found in the same nuclear body as CRY2, and also in a NB with PIF3 (Mas et al., 2000). The current view is that these nuclear bodies are sites of protein turnover, and that turnover of transcription factors, such as the PIFs, as well as of some photoreceptors, is required to achieve a light response.

NATURAL VARIATION IN LIGHT SIGNALING: A UNIQUE SYSTEM FOR ELUCIDATING GROWTH REGULATORY NETWORKS, AS WELL AS THE EVOLUTION OF SUCH COMPLEX NETWORKS

The correct response to a specific light cue depends on the environmental context. Thus, plants native to different light environments have evolved different adaptive responses (Maloof et al., 2000). Although mutational studies have defined a number of genes involved in light perception and signaling, the genes and molecular changes responsible for adaptive changes in light response remain mostly unknown. In a collaboration that has lasted more than a decade, researchers in the labs of Detlef Weigel and myself have shown that a wide range of heritable differences in light response can be found among isolates of Arabidopsis thaliana (e.g. Maloof et al., 2001). Moreover, there is a significant inverse correlation between latitude and light sensitivity, suggesting adaptation to an environmental factor that varies over latitudinal clines. Thus, natural variation studies in Arabidopsis may be very informative, for both new gene discovery (Loudet et al., 2008) and determining which proteins in the signaling pathway may be under selection.

To date, these studies have identified that changes in individual photoreceptor family members, including phyA, phyB, phyC and cry2, are important determinants in the natural variation of light sensitivity. One accession, Lm-2, is insensitive to far-red light, similar to phytochrome A (phyA) mutants. We found that Lm-2 does not complement phyA due to a single amino acid change, and that this change causes production of a protein that is less sensitive to light (Maloof et al., 2001). In separate studies, we used QTL mapping to identify loci involved in light-response variation for seedling emergence between two other strains, Ler and Cvi, and identified an average of four loci per light condition examined (Borevitz et al., 2002). One QTL is the phytochrome B gene, which is known to be important for response to red light. Strikingly, association testing suggests that the PHYB region is an important determinant of the light response across many A. thaliana accessions (Filiault et al., 2008). Finally, in support of a potential adaptive role of PHYC, we found that the more active Col-0 PHYC haplotype group was more frequent at northern latitudes (Balasubramanian et al., 2006). Combined with the work of Koornneef and colleagues on CRY2 (El-Din El-Assal et al., 2001; El-Assal et al., 2004), and Weinig et al. on shade avoidance (Brock et al., 2007), these studies demonstrate that Arabidopsis is an excellent organism for studying the molecular basis of natural variation in light response, and suggest that some changes in the light response in Arabidopsis and its relatives are adaptive.

BEYOND PLANTS

Studying the mechanisms by which plants respond to light has had an impact beyond plants. In this section, I describe a few examples of how plant research on light signaling has had an impact on research in metazoans. Jones et al. (2008) provide a more thorough treatment of this subject.

Protein turnover

The study of light signaling in plants has not only provided insight into plant growth and development, but has also led to the discovery of conserved proteins that regulate transcription, tumorigenesis and lipid metabolism in metazoans. Many of these proteins function in the regulation of protein stability and are members of conserved signaling modules found in both plants and animals. Importantly, the COP9 signalosome (CSN), which shares structural similarity with the lid sub-complex of the 26S proteasome, was first recognized and purified from plants, but is now known to play an essential role in many aspects of both plant and animal development (Wei et al., 1998; Menon et al., 2007). Similarly, the human homologs of Arabidopsis COP1 and DET1 proteins were recently shown to be important negative regulators of the human tumor suppressor p53 (Yi and Deng, 2005). It was a great personal satisfaction to me to open Science and find a paper entitled: ‘Human De-etiolated-1 regulates c-Jun by assembling a CUL4A ubiquitin ligase’ (Wertz et al., 2004).

Cryptochrome and the circadian clock

Identification of Arabidopsis cryptochrome mutants enabled identification of the many functions for cryptochromes in plants, including its roles in photoperiodism and flowering, seedling emergence, and as an input to the circadian clock. Cloning the Arabidopsis gene allowed predictions of the protein sequence, and enabled the discovery of cryptochromes from both flies and humans (Ceriani et al., 1999). In flies, cryptochromes are both photoreceptors and components of the circadian clock, while, in humans, cryptochromes appear to be solely clock components. A number of diseases that depend on a functioning circadian clock (e.g. cancer) are suspected to be caused by defects in the cryptochrome pathway.

Phytochromes and phototropins as photoactivatable switches

Over the past few years, a number of X-ray and NMR structures for bacterial phytochromes and LOV domains have been obtained, and have assisted in clarifying how these photoreceptors work (Harper et al., 2004; Rockwell et al., 2006; Chen et al., 2007; Wagner et al., 2007; Cornilescu et al., 2008; Essen et al., 2008; Ogura et al., 2008; Tokutomi et al., 2008; Yang et al., 2008; Hitomi et al., 2009). These studies have enabled use of these two photoreceptors as photo-activatable switches in mammalian systems. Two recent papers explored actin dynamics using the chromophore-binding domain of phytochrome B (Leung et al., 2008) or the LOV domain of phototropin (Wu et al., 2009). A phytochrome switch is being used to alter cell shape and motility (Levskaya et al., 2009). LOV domain fusions are also being used to probe conformational changes in proteins (Strickland et al., 2008). Finally, fluorescent phytochromes have allowed the live imaging of internal organs in intact mice (Shu et al., 2009).

SUMMARY AND PROSPECTS FOR THE FUTURE

The past 30 years have seen a tremendous gain in our understanding of the light-signaling networks of higher plants. These networks allow plants to gauge their location, monitor daily and seasonal time, and to adjust their growth habit accordingly, thereby conferring a fitness advantage. Because of the contributions of scientists using Arabidopsis, we now have a mechanistic model for photomorphogenesis, in which photoreceptors, mostly acting in the nucleus, regulate gene expression by initiating turnover of key transcription factors. This triggers a transcriptional cascade, leading to the regulated expression of thousands of nuclear genes.

I think Bill Siegelman and his mentors would have been proud of where their field has gone. Yet they would also be shocked by the complexity, in terms of sheer numbers of photoreceptors and their complex relationships with each other. So, was sequencing the A. thaliana genome worth the tens of millions of dollars? The answer is ‘yes’. Did a genetic approach make a difference? Definitely ‘yes’. Arabidopsis has truly risen to the ranks where it is on equal footing with the other model genetic systems, e.g. yeast, Drosophila and C. elegans, especially in terms of natural variation and as a toolkit for molecular genetic studies.

Are the golden years of Arabidopsis over? My answer is a resounding ‘no’! The functions of about 50% of Arabidopsis genes are still unknown, and no other plant system comes close to Arabidopsis in terms of gene discovery. Moreover, we are just beginning to see the impact of 20 years of Arabidopsis research on agriculture. The ease and power of Arabidopsis transformation allows over- or under-expression of specific genes in various mutant backgrounds, thereby conferring unique phenotypes. Moreover, knowing the molecular defects of various Arabidopsis mutants has explained traits that breeders have enriched for, such as the high-pigment mutants of tomato (Lazarova et al., 1998; Mustilli et al., 1999; Liu et al., 2004), or the sorghum maturation genes (Finlayson et al., 1999, 2007). Reduction of losses in yield due to shade avoidance in maize has benefitted greatly from Arabidopsis research on shade avoidance mutants (Kebrom and Brutnell, 2007).

In addition, using tissue- and cell type-specific Arabidopsis promoters to express components of light-signaling pathways in a subset of cells is allowing us to understand which portions of a plant’s response to light are cell autonomous (Tanaka et al., 2002; Endo et al., 2005, 2007; Endo and Nagatani, 2008). Arabidopsis is the only plant system for which the tools are in place to successfully perform such experiments.

The extensive set of mutants for Arabidopsis light, disease and hormone responses will be put to good use over the next decade to understand and test hypotheses about fitness trade-offs during evolution of plants growing in specific environments. Coupled with natural variation studies (Weigel and Mott, 2009), and the Arabidopsis molecular genetics toolkit (Lister et al., 2009), the possibilities are limitless.