Disruption of Fumarylacetoacetate Hydrolase Causes Spontaneous Cell Death under Short-Day Conditions in Arabidopsis

Fumarylacetoacetate hydrolase (FAH) hydrolyzes fumarylacetoacetate to fumarate and acetoacetate, the final step in the tyrosine (Tyr) degradation pathway that is essential to animals. Deficiency of FAH in animals results in an inborn lethal disorder. However, the role for the Tyr degradation pathway in plants remains to be elucidated. In this study, we isolated an Arabidopsis (Arabidopsis thaliana) short-day sensitive cell death1 (sscd1) mutant that displays a spontaneous cell death phenotype under short-day conditions. The SSCD1 gene was cloned via a map-based cloning approach and found to encode an Arabidopsis putative FAH. The spontaneous cell death phenotype of the sscd1 mutant was completely eliminated by further knockout of the gene encoding the putative homogentisate dioxygenase, which catalyzes homogentisate into maleylacetoacetate (the antepenultimate step) in the Tyr degradation pathway. Furthermore, treatment of Arabidopsis wild-type seedlings with succinylacetone, an abnormal metabolite caused by loss of FAH in the Tyr degradation pathway, mimicked the sscd1 cell death phenotype. These results demonstrate that disruption of FAH leads to cell death in Arabidopsis and suggest that the Tyr degradation pathway is essential for plant survival under short-day conditions.Programmed cell death (PCD) has been defined as a sequence of genetically regulated events that lead to the elimination of specific cells, tissues, or whole organs (Lockshin and Zakeri, 2004). In plants, PCD is essential for developmental processes and defense responses (Dangl et al., 1996; Greenberg, 1996; Durrant et al., 2007). One well-characterized example of plant PCD is the hypersensitive response occurring during incompatible plant-pathogen interactions (Lam, 2004), which results in cell death to form visible lesions at the site of infection by an avirulent pathogen and consequently limits the pathogen spread (Morel and Dangl, 1997).To date, a large number of mutants that display spontaneous cell death lesions have been identified in barley (Hordeum vulgare), maize (Zea mays), rice (Oryza sativa), and Arabidopsis (Arabidopsis thaliana; Marchetti et al., 1983; Wolter et al., 1993; Dietrich et al., 1994; Gray et al., 1997). Because lesions form in the absence of pathogen infection, these mutants have been collectively termed as lesion-mimic mutants. Many genes with regulatory roles in PCD and defense responses, including LESION SIMULATING DISEASE1, ACCELERATED CELL DEATH11, and VASCULAR ASSOCIATED DEATH1, have been cloned and characterized (Dietrich et al., 1997; Brodersen et al., 2002; Lorrain et al., 2004).The appearance of spontaneous cell death lesions in some lesion-mimic mutants is dependent on photoperiod. For example, the Arabidopsis mutant lesion simulating disease1 and myoinositol-1-phosphate synthase1 show lesions under long days (LD; Dietrich et al., 1994; Meng et al., 2009), whereas the lesion simulating disease2, lesion initiation1, enhancing RPW8-mediated HR-like cell death1, and lag one homolog1 display lesions under short days (SD; Dietrich et al., 1994; Ishikawa et al., 2003; Wang et al., 2008; Ternes et al., 2011).Blockage of some metabolic pathways in plants may cause cell death and result in lesion formation. For example, the lesion-mimic phenotypes in the Arabidopsis mutants lesion initiation2 and accelerated cell death2 and the maize mutant lesion mimic22 result from an impairment of porphyrin metabolism (Hu et al., 1998; Ishikawa et al., 2001; Mach et al., 2001). Deficiency in fatty acid, sphingolipid, and myoinositol metabolism also causes cell death in Arabidopsis (Mou et al., 2000; Liang et al., 2003; Wang et al., 2008; Meng et al., 2009; Donahue et al., 2010; Berkey et al., 2012).Tyr degradation is an essential five-step pathway in animals (Lindblad et al., 1977). First, Tyr aminotransferase catalyzes the conversion of Tyr into 4-hydroxyphenylpyruvate, which is further transformed into homogentisate by 4-hydroxyphenylpyruvate dioxygenase. Through the sequential action of homogentisate dioxygenase (HGO), maleylacetoacetate isomerase (MAAI), and fumarylacetoacetate hydrolase (FAH), homogentisate is catalyzed to generate fumarate and acetoacetate (Lindblad et al., 1977). Blockage of this pathway in animals results in metabolic disorder diseases (Lindblad et al., 1977; Ruppert et al., 1992; Grompe et al., 1993). For example, human FAH deficiency causes hereditary tyrosinemia type I (HT1), an inborn lethal disease (St-Louis and Tanguay, 1997). Although the homologous genes putatively encoding these enzymes exist in plants (Dixon et al., 2000; Lopukhina et al., 2001; Dixon and Edwards, 2006), it is unclear whether this pathway is essential for plant growth and development.In this study, we report the isolation and characterization of a recessive short-day sensitive cell death1 (sscd1) mutant in Arabidopsis. Map-based cloning of the corresponding gene revealed that SSCD1 encodes the Arabidopsis putative FAH. Further knockout of the gene encoding the Arabidopsis putative HGO completely eliminated the spontaneous cell death phenotype in the sscd1 mutant. Furthermore, we found that treatment of Arabidopsis wild-type seedlings with succinylacetone, an abnormal metabolite caused by loss of FAH in the Tyr degradation pathway (Lindblad et al., 1977), is able to mimic the sscd1 cell death phenotype. These results demonstrate that disruption of FAH leads to cell death in Arabidopsis and suggest that the Tyr degradation pathway is essential for plant survival under SD.     

Spore Density Determines Infection Strategy by the Plant Pathogenic Fungus Plectosphaerella cucumerina

Necrotrophic and biotrophic pathogens are resisted by different plant defenses. While necrotrophic pathogens are sensitive to jasmonic acid (JA)-dependent resistance, biotrophic pathogens are resisted by salicylic acid (SA)- and reactive oxygen species (ROS)-dependent resistance. Although many pathogens switch from biotrophy to necrotrophy during infection, little is known about the signals triggering this transition. This study is based on the observation that the early colonization pattern and symptom development by the ascomycete pathogen Plectosphaerella cucumerina (P. cucumerina) vary between inoculation methods. Using the Arabidopsis (Arabidopsis thaliana) defense response as a proxy for infection strategy, we examined whether P. cucumerina alternates between hemibiotrophic and necrotrophic lifestyles, depending on initial spore density and distribution on the leaf surface. Untargeted metabolome analysis revealed profound differences in metabolic defense signatures upon different inoculation methods. Quantification of JA and SA, marker gene expression, and cell death confirmed that infection from high spore densities activates JA-dependent defenses with excessive cell death, while infection from low spore densities induces SA-dependent defenses with lower levels of cell death. Phenotyping of Arabidopsis mutants in JA, SA, and ROS signaling confirmed that P. cucumerina is differentially resisted by JA- and SA/ROS-dependent defenses, depending on initial spore density and distribution on the leaf. Furthermore, in situ staining for early callose deposition at the infection sites revealed that necrotrophy by P. cucumerina is associated with elevated host defense. We conclude that P. cucumerina adapts to early-acting plant defenses by switching from a hemibiotrophic to a necrotrophic infection program, thereby gaining an advantage of immunity-related cell death in the host.Plant pathogens are often classified as necrotrophic or biotrophic, depending on their infection strategy (Glazebrook, 2005; Nishimura and Dangl, 2010). Necrotrophic pathogens kill living host cells and use the decayed plant tissue as a substrate to colonize the plant, whereas biotrophic pathogens parasitize living plant cells by employing effector molecules that suppress the host immune system (Pel and Pieterse, 2013). Despite this binary classification, the majority of pathogenic microbes employ a hemibiotrophic infection strategy, which is characterized by an initial biotrophic phase followed by a necrotrophic infection strategy at later stages of infection (Perfect and Green, 2001). The pathogenic fungi Magnaporthe grisea, Sclerotinia sclerotiorum, and Mycosphaerella graminicola, the oomycete Phytophthora infestans, and the bacterial pathogen Pseudomonas syringae are examples of hemibiotrophic plant pathogens (Perfect and Green, 2001; Koeck et al., 2011; van Kan et al., 2014; Kabbage et al., 2015).Despite considerable progress in our understanding of plant resistance to necrotrophic and biotrophic pathogens (Glazebrook, 2005; Mengiste, 2012; Lai and Mengiste, 2013), recent debate highlights the dynamic and complex interplay between plant-pathogenic microbes and their hosts, which is raising concerns about the use of infection strategies as a static tool to classify plant pathogens. For instance, the fungal genus Botrytis is often labeled as an archetypal necrotroph, even though there is evidence that it can behave as an endophytic fungus with a biotrophic lifestyle (van Kan et al., 2014). The rice blast fungus Magnaporthe oryzae, which is often classified as a hemibiotrophic leaf pathogen (Perfect and Green, 2001; Koeck et al., 2011), can adopt a purely biotrophic lifestyle when infecting root tissues (Marcel et al., 2010). It remains unclear which signals are responsible for the switch from biotrophy to necrotrophy and whether these signals rely solely on the physiological state of the pathogen, or whether host-derived signals play a role as well (Kabbage et al., 2015).The plant hormones salicylic acid (SA) and jasmonic acid (JA) play a central role in the activation of plant defenses (Glazebrook, 2005; Pieterse et al., 2009, 2012). The first evidence that biotrophic and necrotrophic pathogens are resisted by different immune responses came from Thomma et al. (1998), who demonstrated that Arabidopsis (Arabidopsis thaliana) genotypes impaired in SA signaling show enhanced susceptibility to the biotrophic pathogen Hyaloperonospora arabidopsidis (formerly known as Peronospora parastitica), while JA-insensitive genotypes were more susceptible to the necrotrophic fungus Alternaria brassicicola. In subsequent years, the differential effectiveness of SA- and JA-dependent defense mechanisms has been confirmed in different plant-pathogen interactions, while additional plant hormones, such as ethylene, abscisic acid (ABA), auxins, and cytokinins, have emerged as regulators of SA- and JA-dependent defenses (Bari and Jones, 2009; Cao et al., 2011; Pieterse et al., 2012). Moreover, SA- and JA-dependent defense pathways have been shown to act antagonistically on each other, which allows plants to prioritize an appropriate defense response to attack by biotrophic pathogens, necrotrophic pathogens, or herbivores (Koornneef and Pieterse, 2008; Pieterse et al., 2009; Verhage et al., 2010).In addition to plant hormones, reactive oxygen species (ROS) play an important regulatory role in plant defenses (Torres et al., 2006; Lehmann et al., 2015). Within minutes after the perception of pathogen-associated molecular patterns, NADPH oxidases and apoplastic peroxidases generate early ROS bursts (Torres et al., 2002; Daudi et al., 2012; O’Brien et al., 2012), which activate downstream defense signaling cascades (Apel and Hirt, 2004; Torres et al., 2006; Miller et al., 2009; Mittler et al., 2011; Lehmann et al., 2015). ROS play an important regulatory role in the deposition of callose (Luna et al., 2011; Pastor et al., 2013) and can also stimulate SA-dependent defenses (Chaouch et al., 2010; Yun and Chen, 2011; Wang et al., 2014; Mammarella et al., 2015). However, the spread of SA-induced apoptosis during hyperstimulation of the plant immune system is contained by the ROS-generating NADPH oxidase RBOHD (Torres et al., 2005), presumably to allow for the sufficient generation of SA-dependent defense signals from living cells that are adjacent to apoptotic cells. Nitric oxide (NO) plays an additional role in the regulation of SA/ROS-dependent defense (Trapet et al., 2015). This gaseous molecule can stimulate ROS production and cell death in the absence of SA while preventing excessive ROS production at high cellular SA levels via S-nitrosylation of RBOHD (Yun et al., 2011). Recently, it was shown that pathogen-induced accumulation of NO and ROS promotes the production of azelaic acid, a lipid derivative that primes distal plants for SA-dependent defenses (Wang et al., 2014). Hence, NO, ROS, and SA are intertwined in a complex regulatory network to mount local and systemic resistance against biotrophic pathogens. Interestingly, pathogens with a necrotrophic lifestyle can benefit from ROS/SA-dependent defenses and associated cell death (Govrin and Levine, 2000). For instance, Kabbage et al. (2013) demonstrated that S. sclerotiorum utilizes oxalic acid to repress oxidative defense signaling during initial biotrophic colonization, but it stimulates apoptosis at later stages to advance necrotrophic colonization. Moreover, SA-induced repression of JA-dependent resistance not only benefits necrotrophic pathogens but also hemibiotrophic pathogens after having switched from biotrophy to necrotrophy (Glazebrook, 2005; Pieterse et al., 2009, 2012).Plectosphaerella cucumerina ((P. cucumerina, anamorph Plectosporum tabacinum) anamorph Plectosporum tabacinum) is a filamentous ascomycete fungus that can survive saprophytically in soil by decomposing plant material (Palm et al., 1995). The fungus can cause sudden death and blight disease in a variety of crops (Chen et al., 1999; Harrington et al., 2000). Because P. cucumerina can infect Arabidopsis leaves, the P. cucumerina-Arabidopsis interaction has emerged as a popular model system in which to study plant defense reactions to necrotrophic fungi (Berrocal-Lobo et al., 2002; Ton and Mauch-Mani, 2004; Carlucci et al., 2012; Ramos et al., 2013). Various studies have shown that Arabidopsis deploys a wide range of inducible defense strategies against P. cucumerina, including JA-, SA-, ABA-, and auxin-dependent defenses, glucosinolates (Tierens et al., 2001; Sánchez-Vallet et al., 2010; Gamir et al., 2014; Pastor et al., 2014), callose deposition (García-Andrade et al., 2011; Gamir et al., 2012, 2014; Sánchez-Vallet et al., 2012), and ROS (Tierens et al., 2002; Sánchez-Vallet et al., 2010; Barna et al., 2012; Gamir et al., 2012, 2014; Pastor et al., 2014). Recent metabolomics studies have revealed large-scale metabolic changes in P. cucumerina-infected Arabidopsis, presumably to mobilize chemical defenses (Sánchez-Vallet et al., 2010; Gamir et al., 2014; Pastor et al., 2014). Furthermore, various chemical agents have been reported to induce resistance against P. cucumerina. These chemicals include β-amino-butyric acid, which primes callose deposition and SA-dependent defenses, benzothiadiazole (BTH or Bion; Görlach et al., 1996; Ton and Mauch-Mani, 2004), which activates SA-related defenses (Lawton et al., 1996; Ton and Mauch-Mani, 2004; Gamir et al., 2014; Luna et al., 2014), JA (Ton and Mauch-Mani, 2004), and ABA, which primes ROS and callose deposition (Ton and Mauch-Mani, 2004; Pastor et al., 2013). However, among all these studies, there is increasing controversy about the exact signaling pathways and defense responses contributing to plant resistance against P. cucumerina. While it is clear that JA and ethylene contribute to basal resistance against the fungus, the exact roles of SA, ABA, and ROS in P. cucumerina resistance vary between studies (Thomma et al., 1998; Ton and Mauch-Mani, 2004; Sánchez-Vallet et al., 2012; Gamir et al., 2014).This study is based on the observation that the disease phenotype during P. cucumerina infection differs according to the inoculation method used. We provide evidence that the fungus follows a hemibiotrophic infection strategy when infecting from relatively low spore densities on the leaf surface. By contrast, when challenged by localized host defense to relatively high spore densities, the fungus switches to a necrotrophic infection program. Our study has uncovered a novel strategy by which plant-pathogenic fungi can take advantage of the early immune response in the host plant.     

PLASTID MOVEMENT IMPAIRED1 and PLASTID MOVEMENT IMPAIRED1-RELATED1 Mediate Photorelocation Movements of Both Chloroplasts and Nuclei

Organelle movement and positioning play important roles in fundamental cellular activities and adaptive responses to environmental stress in plants. To optimize photosynthetic light utilization, chloroplasts move toward weak blue light (the accumulation response) and escape from strong blue light (the avoidance response). Nuclei also move in response to strong blue light by utilizing the light-induced movement of attached plastids in leaf cells. Blue light receptor phototropins and several factors for chloroplast photorelocation movement have been identified through molecular genetic analysis of Arabidopsis (Arabidopsis thaliana). PLASTID MOVEMENT IMPAIRED1 (PMI1) is a plant-specific C2-domain protein that is required for efficient chloroplast photorelocation movement. There are two PLASTID MOVEMENT IMPAIRED1-RELATED (PMIR) genes, PMIR1 and PMIR2, in the Arabidopsis genome. However, the mechanism in which PMI1 regulates chloroplast and nuclear photorelocation movements and the involvement of PMIR1 and PMIR2 in these organelle movements remained unknown. Here, we analyzed chloroplast and nuclear photorelocation movements in mutant lines of PMI1, PMIR1, and PMIR2. In mesophyll cells, the pmi1 single mutant showed severe defects in both chloroplast and nuclear photorelocation movements resulting from the impaired regulation of chloroplast-actin filaments. In pavement cells, pmi1 mutant plants were partially defective in both plastid and nuclear photorelocation movements, but pmi1pmir1 and pmi1pmir1pmir2 mutant lines lacked the blue light-induced movement responses of plastids and nuclei completely. These results indicated that PMI1 is essential for chloroplast and nuclear photorelocation movements in mesophyll cells and that both PMI1 and PMIR1 are indispensable for photorelocation movements of plastids and thus, nuclei in pavement cells.In plants, organelles move within the cell and become appropriately positioned to accomplish their functions and adapt to the environment (for review, see Wada and Suetsugu, 2004). Light-induced chloroplast movement (chloroplast photorelocation movement) is one of the best characterized organelle movements in plants (Suetsugu and Wada, 2012). Under weak light conditions, chloroplasts move toward light to capture light efficiently (the accumulation response; Zurzycki, 1955). Under strong light conditions, chloroplasts escape from light to avoid photodamage (the avoidance response; Kasahara et al., 2002; Sztatelman et al., 2010; Davis and Hangarter, 2012; Cazzaniga et al., 2013). In most green plant species, these responses are induced primarily by the blue light receptor phototropin (phot) in response to a range of wavelengths from UVA to blue light (approximately 320–500 nm; for review, see Suetsugu and Wada, 2012; Wada and Suetsugu, 2013; Kong and Wada, 2014). Phot-mediated chloroplast movement has been shown in land plants, such as Arabidopsis (Arabidopsis thaliana; Jarillo et al., 2001; Kagawa et al., 2001; Sakai et al., 2001), the fern Adiantum capillus-veneris (Kagawa et al., 2004), the moss Physcomitrella patens (Kasahara et al., 2004), and the liverwort Marchantia polymorpha (Komatsu et al., 2014). Two phots in Arabidopsis, phot1 and phot2, redundantly mediate the accumulation response (Sakai et al., 2001), whereas phot2 primarily regulates the avoidance response (Jarillo et al., 2001; Kagawa et al., 2001; Luesse et al., 2010). M. polymorpha has only one phot that mediates both the accumulation and avoidance responses (Komatsu et al., 2014), although two or more phots mediate chloroplast photorelocation movement in A. capillus-veneris (Kagawa et al., 2004) and P. patens (Kasahara et al., 2004). Thus, duplication and functional diversification of PHOT genes have occurred during land plant evolution, and plants have gained a sophisticated light sensing system for chloroplast photorelocation movement.In general, movements of plant organelles, including chloroplasts, are dependent on actin filaments (for review, see Wada and Suetsugu, 2004). Most organelles common in eukaryotes, such as mitochondria, peroxisomes, and Golgi bodies, use the myosin motor for their movements, but there is no clear evidence that chloroplast movement is myosin dependent (for review, see Suetsugu et al., 2010a). Land plants have innovated a novel actin-based motility system that is specialized for chloroplast movement as well as a photoreceptor system (for review, see Suetsugu et al., 2010a; Wada and Suetsugu, 2013; Kong and Wada, 2014). Chloroplast-actin (cp-actin) filaments, which were first found in Arabidopsis, are short actin filaments specifically localized around the chloroplast periphery at the interface between the chloroplast and the plasma membrane (Kadota et al., 2009). Strong blue light induces the rapid disappearance of cp-actin filaments and then, their subsequent reappearance preferentially at the front region of the moving chloroplasts. This asymmetric distribution of cp-actin filaments is essential for directional chloroplast movement (Kadota et al., 2009; Kong et al., 2013a). The greater the difference in the amount of cp-actin filaments between the front and rear regions of chloroplasts becomes, the faster the chloroplasts move, in which the magnitude of the difference is determined by fluence rate (Kagawa and Wada, 2004; Kadota et al., 2009; Kong et al., 2013a). Strong blue light-induced disappearance of cp-actin filaments is regulated in a phot2-dependent manner before the intensive polymerization of cp-actin filaments at the front region occurs (Kadota et al., 2009; Ichikawa et al., 2011; Kong et al., 2013a). This phot2-dependent response contributes to the greater difference in the amount of cp-actin filaments between the front and rear regions of chloroplasts. Similar behavior of cp-actin filaments has also been observed in A. capillus-veneris (Tsuboi and Wada, 2012) and P. patens (Yamashita et al., 2011).Like chloroplasts, nuclei also show light-mediated movement and positioning (nuclear photorelocation movement) in land plants (for review, see Higa et al., 2014b). In gametophytic cells of A. capillus-veneris, weak light induced the accumulation responses of both chloroplasts and nuclei, whereas strong light induced avoidance responses (Kagawa and Wada, 1993, 1995; Tsuboi et al., 2007). However, in mesophyll cells of Arabidopsis, strong blue light induced both chloroplast and nuclear avoidance responses, but weak blue light induced only the chloroplast accumulation response (Iwabuchi et al., 2007, 2010; Higa et al., 2014a). In Arabidopsis pavement cells, small numbers of tiny plastids were found and showed autofluorescence under the confocal laser-scanning microscopy (Iwabuchi et al., 2010; Higa et al., 2014a). Hereafter, the plastid in the pavement cells is called the pavement cell plastid. Strong blue light-induced avoidance responses of pavement cell plastids and nuclei were induced in a phot2-dependent manner, but the accumulation response was not detected for either organelle (Iwabuchi et al., 2007, 2010; Higa et al., 2014a). In both Arabidopsis and A. capillus-veneris, phots mediate nuclear photorelocation movement, and phot2 mediates the nuclear avoidance response (Iwabuchi et al., 2007, 2010; Tsuboi et al., 2007). The nuclear avoidance response is dependent on actin filaments in both mesophyll and pavement cells of Arabidopsis (Iwabuchi et al., 2010). Recently, it was shown that the nuclear avoidance response relies on cp-actin-dependent movement of pavement cell plastids, where nuclei are associated with pavement cell plastids of Arabidopsis (Higa et al., 2014a). In mesophyll cells, nuclear avoidance response is likely dependent on cp-actin filament-mediated chloroplast movement, because the mutants deficient in chloroplast movement were also defective in nuclear avoidance response (Higa et al., 2014a). Thus, phots mediate both chloroplast (and pavement cell plastid) and nuclear photorelocation movement by regulating cp-actin filaments.Molecular genetic analyses of Arabidopsis mutants deficient in chloroplast photorelocation movement have identified many molecular factors involved in signal transduction and/or motility systems as well as those involved in the photoreceptor system for chloroplast photorelocation movement (and thus, nuclear photorelocation movement; for review, see Suetsugu and Wada, 2012; Wada and Suetsugu, 2013; Kong and Wada, 2014). CHLOROPLAST UNUSUAL POSITIONING1 (CHUP1; Oikawa et al., 2003) and KINESIN-LIKE PROTEIN FOR ACTIN-BASED CHLOROPLAST MOVEMENT (KAC; Suetsugu et al., 2010b) are key factors for generating and/or maintaining cp-actin filaments. Both proteins are highly conserved in land plants and essential for the movement and attachment of chloroplasts to the plasma membrane in Arabidopsis (Oikawa et al., 2003, 2008; Suetsugu et al., 2010b), A. capillus-veneris (Suetsugu et al., 2012), and P. patens (Suetsugu et al., 2012; Usami et al., 2012). CHUP1 is localized on the chloroplast outer membrane and binds to globular and filamentous actins and profilin in vitro (Oikawa et al., 2003, 2008; Schmidt von Braun and Schleiff, 2008). Although KAC is a kinesin-like protein, it lacks microtubule-dependent motor activity but has filamentous actin binding activity (Suetsugu et al., 2010b). An actin-bundling protein THRUMIN1 (THRUM1) is required for efficient chloroplast photorelocation movement (Whippo et al., 2011) and interacts with cp-actin filaments (Kong et al., 2013a). chup1 and kac mutant plants were shown to lack detectable cp-actin filaments (Kadota et al., 2009; Suetsugu et al., 2010b; Ichikawa et al., 2011; Kong et al., 2013a). Similarly, cp-actin filaments were rarely detected in thrum1 mutant plants (Kong et al., 2013a), indicating that THRUM1 also plays an important role in maintaining cp-actin filaments.Other proteins J-DOMAIN PROTEIN REQUIRED FOR CHLOROPLAST ACCUMULATION RESPONSE1 (JAC1; Suetsugu et al., 2005), WEAK CHLOROPLAST MOVEMENT UNDER BLUE LIGHT1 (WEB1; Kodama et al., 2010), and PLASTID MOVEMENT IMPAIRED2 (PMI2; Luesse et al., 2006; Kodama et al., 2010) are involved in the light regulation of cp-actin filaments and chloroplast photorelocation movement. JAC1 is an auxilin-like J-domain protein that mediates the chloroplast accumulation response through its J-domain function (Suetsugu et al., 2005; Takano et al., 2010). WEB1 and PMI2 are coiled-coil proteins that interact with each other (Kodama et al., 2010). Although web1 and pmi2 were partially defective in the avoidance response, the jac1 mutation completely suppressed the phenotype of web1 and pmi2, suggesting that the WEB1/PMI2 complex suppresses JAC1 function (i.e. the accumulation response) under strong light conditions (Kodama et al., 2010). Both web1 and pmi2 showed impaired disappearance of cp-actin filaments in response to strong blue light (Kodama et al., 2010). However, the exact molecular functions of these proteins are unknown.In this study, we characterized mutant plants deficient in the PMI1 gene and two homologous genes PLASTID MOVEMENT IMPAIRED1-RELATED1 (PMIR1) and PMIR2. PMI1 was identified through molecular genetic analyses of pmi1 mutants that showed severe defects in chloroplast accumulation and avoidance responses (DeBlasio et al., 2005). PMI1 is a plant-specific C2-domain protein (DeBlasio et al., 2005; Zhang and Aravind, 2010), but its roles and those of PMIRs in cp-actin-mediated chloroplast and nuclear photorelocation movements remained unclear. Thus, we analyzed chloroplast and nuclear photorelocation movements in the single, double, and triple mutants of pmi1, pmir1, and pmir2.     

A Suppressor Screen of the Chimeric AtCNGC11/12 Reveals Residues Important for Intersubunit Interactions of Cyclic Nucleotide-Gated Ion Channels

To investigate the structure-function relationship of plant cyclic nucleotide-gated ion channels (CNGCs), we identified a total of 29 mutant alleles of the chimeric AtCNGC11/12 gene that induces multiple defense responses in the Arabidopsis (Arabidopsis thaliana) mutant, constitutive expresser of PR genes22 (cpr22). Based on computational modeling, two new alleles, S100 (AtCNGC11/12:G459R) and S137 (AtCNGC11/12:R381H), were identified as counterparts of human CNGA3 (a human CNGC) mutants. Both mutants lost all cpr22-mediated phenotypes. Transient expression in Nicotiana benthamiana as well as functional complementation in yeast (Saccharomyces cerevisiae) showed that both AtCNGC11/12:G459R and AtCNGC11/12:R381H have alterations in their channel function. Site-directed mutagenesis coupled with fast-protein liquid chromatography using recombinantly expressed C-terminal peptides indicated that both mutations significantly influence subunit stoichiometry to form multimeric channels. This observation was confirmed by bimolecular fluorescence complementation in planta. Taken together, we have identified two residues that are likely important for subunit interaction for plant CNGCs and likely for animal CNGCs as well.Cyclic nucleotide-gated ion channels (CNGCs) were first discovered in retinal photoreceptors and olfactory sensory neurons (Zagotta and Siegelbaum, 1996; Kaupp and Seifert, 2002). CNGCs play crucial roles for the signal transduction in these neurons that are excited by photons and odorants, respectively. In mammalian genomes, six CNGC genes have been found and named CNGA1 to CNGA4, CNGB1, and CNGB3 (Kaupp and Seifert, 2002). It has been reported that in mammalian cells, CNGCs function as heterotetramers that are composed of A and B subunits with cell-specific stoichiometry (Kaupp and Seifert, 2002; Cukkemane et al., 2011). For example, CNGCs in rod photoreceptors are composed of three A1 subunits and one B1a subunit, whereas in cone photoreceptors, they are believed to be composed of two A3 and two B3 subunits (Zhong et al., 2002; Peng et al., 2004). The structure of each subunit is similar to that of the voltage-gated K⁺-selective ion channel (Shaker) proteins, including a cytoplasmic N terminus, six membrane-spanning regions (S1–S6), a pore domain located between S5 and S6, and a cytoplasmic C terminus (Zagotta and Siegelbaum, 1996). However, CNGCs are only weakly voltage dependent and are opened by the direct binding of cyclic nucleotides (cAMP and cGMP), which are universally important secondary messengers that control diverse cellular responses (Fesenko et al., 1985). The cytoplasmic C terminus contains a cyclic nucleotide-binding domain (CNBD) and a C-linker region that connects the CNBD to the S6 domain. CNGC activity is also regulated by feedback inhibitory mechanisms involving the Ca²⁺ sensor protein, calmodulin (CaM). CaM-binding sites in animal CNGCs have been found in various regions of both the C- and N-terminal domains (Ungerer et al., 2011). It has been reported that the subunit composition has significant influence on the mode of CaM-mediated regulation (Kramer and Siegelbaum, 1992; Bradley et al., 2004; Song et al., 2008).On the other hand, plant CNGCs have only been investigated much more recently. The first plant CNGC, HvCBT1, was identified as a CaM-binding transporter protein in barley (Hordeum vulgare; Schuurink et al., 1998). Subsequently, several CNGCs were identified from Arabidopsis (Arabidopsis thaliana) and tobacco (Nicotiana tabacum; Arazi et al., 1999; Köhler and Neuhaus, 1998; Köhler et al., 1999). Interestingly, the Arabidopsis genome sequencing project identified a large family comprising 20 members (AtCNGC1–AtCNGC20), indicating a significant expansion of Arabidopsis CNGCs that suggests a higher level of diversity and functional importance in plants (Mäser et al., 2001). To date, possible biological functions of Arabidopsis CNGCs in development, ion homeostasis, thermal sensing, as well as pathogen resistance have been reported (Kaplan et al., 2007; Chin et al., 2009; Dietrich et al., 2010; Moeder et al., 2011; Finka et al., 2012). With respect to structure, plant CNGCs are believed to have a similar architecture to their animal counterparts (Chin et al., 2009). However, only a handful of studies on the structure-function analysis of plant CNGCs have been published so far, and this field is still very much in its infancy (Hua et al., 2003; Bridges et al., 2005; Kaplan et al., 2007; Baxter et al., 2008; Chin et al., 2010).Previously, we have reported two functionally important residues in plant CNGCs (Baxter et al., 2008; Chin et al., 2010). These residues were discovered using a suppressor screen of the rare gain-of-function Arabidopsis mutant constitutive expresser of PR genes22 (cpr22; Yoshioka et al., 2006). The cpr22 mutant, which has a deletion between AtCNGC11 and AtCNGC12 resulting in a novel but functional chimeric CNGC (AtCNGC11/12), exhibits multiple resistance responses without pathogen infection in the hemizygous state and conditional lethality in the homozygous state (Yoshioka et al., 2001, 2006; Moeder et al., 2011). It has been reported that the cpr22 phenotype is attributable to the expression of AtCNGC11/12 and its channel activity (Yoshioka et al., 2006; Baxter et al., 2008), thereby making the suppressor screen an invaluable tool for identifying intragenic mutants to further elucidate the structure-function relationship of plant CNGCs (Baxter et al., 2008; Chin et al., 2010).In this study, we describe a total of 29 mutant alleles of AtCNGC11/12, including the three previously published alleles (Baxter et al., 2008; Chin at al., 2010), and compare their predicted three-dimensional structural positions with equivalent mutations of a human CNGC, CNGA3. In this analysis, two AtCNGC11/12 mutations emerged as counterparts of human mutations (Wissinger et al., 2001). Both the AtCNGC11/12 as well as the human CNGA3 mutations were computationally predicted to affect intersubunit interactions. This prediction was experimentally validated by size-exclusion chromatography (FPLC) as well as bimolecular fluorescence complementation (BiFC) in combination with site-direct mutagenesis using recombinant C-terminal peptides.     

Multiscale Metabolic Modeling: Dynamic Flux Balance Analysis on a Whole-Plant Scale

Plant metabolism is characterized by a unique complexity on the cellular, tissue, and organ levels. On a whole-plant scale, changing source and sink relations accompanying plant development add another level of complexity to metabolism. With the aim of achieving a spatiotemporal resolution of source-sink interactions in crop plant metabolism, a multiscale metabolic modeling (MMM) approach was applied that integrates static organ-specific models with a whole-plant dynamic model. Allowing for a dynamic flux balance analysis on a whole-plant scale, the MMM approach was used to decipher the metabolic behavior of source and sink organs during the generative phase of the barley (Hordeum vulgare) plant. It reveals a sink-to-source shift of the barley stem caused by the senescence-related decrease in leaf source capacity, which is not sufficient to meet the nutrient requirements of sink organs such as the growing seed. The MMM platform represents a novel approach for the in silico analysis of metabolism on a whole-plant level, allowing for a systemic, spatiotemporally resolved understanding of metabolic processes involved in carbon partitioning, thus providing a novel tool for studying yield stability and crop improvement.Plants are of vital significance as a source of food (Grusak and DellaPenna, 1999; Rogalski and Carrer, 2011), feed (Lu et al., 2011), energy (Tilman et al., 2006; Parmar et al., 2011), and feedstocks for the chemical industry (Metzger and Bornscheuer, 2006; Kinghorn et al., 2011). Given the close connection between plant metabolism and the usability of plant products, there is a growing interest in understanding and predicting the behavior and regulation of plant metabolic processes. In order to increase crop quality and yield, there is a need for methods guiding the rational redesign of the plant metabolic network (Schwender, 2009).Mathematical modeling of plant metabolism offers new approaches to understand, predict, and modify complex plant metabolic processes. In plant research, the issue of metabolic modeling is constantly gaining attention, and different modeling approaches applied to plant metabolism exist, ranging from highly detailed quantitative to less complex qualitative approaches (for review, see Giersch, 2000; Morgan and Rhodes, 2002; Poolman et al., 2004; Rios-Estepa and Lange, 2007).A widely used modeling approach is flux balance analysis (FBA), which allows the prediction of metabolic capabilities and steady-state fluxes under different environmental and genetic backgrounds using (non)linear optimization (Orth et al., 2010). Assuming steady-state conditions, FBA has the advantage of not requiring the knowledge of kinetic parameters and, therefore, can be applied to model detailed, large-scale systems. In recent years, the FBA approach has been applied to several different plant species, such as maize (Zea mays; Dal’Molin et al., 2010; Saha et al., 2011), barley (Hordeum vulgare; Grafahrend-Belau et al., 2009b; Melkus et al., 2011; Rolletschek et al., 2011), rice (Oryza sativa; Lakshmanan et al., 2013), Arabidopsis (Arabidopsis thaliana; Poolman et al., 2009; de Oliveira Dal’Molin et al., 2010; Radrich et al., 2010; Williams et al., 2010; Mintz-Oron et al., 2012; Cheung et al., 2013), and rapeseed (Brassica napus; Hay and Schwender, 2011a, 2011b; Pilalis et al., 2011), as well as algae (Boyle and Morgan, 2009; Cogne et al., 2011; Dal’Molin et al., 2011) and photoautotrophic bacteria (Knoop et al., 2010; Montagud et al., 2010; Boyle and Morgan, 2011). These models have been used to study different aspects of metabolism, including the prediction of optimal metabolic yields and energy efficiencies (Dal’Molin et al., 2010; Boyle and Morgan, 2011), changes in flux under different environmental and genetic backgrounds (Grafahrend-Belau et al., 2009b; Dal’Molin et al., 2010; Melkus et al., 2011), and nonintuitive metabolic pathways that merit subsequent experimental investigations (Poolman et al., 2009; Knoop et al., 2010; Rolletschek et al., 2011). Although FBA of plant metabolic models was shown to be capable of reproducing experimentally determined flux distributions (Williams et al., 2010; Hay and Schwender, 2011b) and generating new insights into metabolic behavior, capacities, and efficiencies (Sweetlove and Ratcliffe, 2011), challenges remain to advance the utility and predictive power of the models.Given that many plant metabolic functions are based on interactions between different subcellular compartments, cell types, tissues, and organs, the reconstruction of organ-specific models and the integration of these models into interacting multiorgan and/or whole-plant models is a prerequisite to get insight into complex plant metabolic processes organized on a whole-plant scale (e.g. source-sink interactions). Almost all FBA models of plant metabolism are restricted to one cell type (Boyle and Morgan, 2009; Knoop et al., 2010; Montagud et al., 2010; Cogne et al., 2011; Dal’Molin et al., 2011), one tissue or one organ (Grafahrend-Belau et al., 2009b; Hay and Schwender, 2011a, 2011b; Pilalis et al., 2011; Mintz-Oron et al., 2012), and only one model exists taking into account the interaction between two cell types by specifying the interaction between mesophyll and bundle sheath cells in C4 photosynthesis (Dal’Molin et al., 2010). So far, no model representing metabolism at the whole-plant scale exists.Considering whole-plant metabolism raises the problem of taking into account temporal and environmental changes in metabolism during plant development and growth. Although classical static FBA is unable to predict the dynamics of metabolic processes, as the network analysis is based on steady-state solutions, time-dependent processes can be taken into account by extending the classical static FBA to a dynamic flux balance analysis (dFBA), as proposed by Mahadevan et al. (2002). The static (SOA) and dynamic optimization approaches introduced in this work provide a framework for analyzing the transience of metabolism by integrating kinetic expressions to dynamically constrain exchange fluxes. Due to the requirement of knowing or estimating a large number of kinetic parameters, so far dFBA has only been applied to a plant metabolic model once, to study the photosynthetic metabolism in the chloroplasts of C3 plants by a simplified model of five biochemical reactions (Luo et al., 2009). Integrating a dynamic model into a static FBA model is an alternative approach to perform dFBA.In this study, a multiscale metabolic modeling (MMM) approach was applied with the aim of achieving a spatiotemporal resolution of cereal crop plant metabolism. To provide a framework for the in silico analysis of the metabolic dynamics of barley on a whole-plant scale, the MMM approach integrates a static multiorgan FBA model and a dynamic whole-plant multiscale functional plant model (FPM) to perform dFBA. The performance of the novel whole-plant MMM approach was tested by studying source-sink interactions during the seed developmental phase of barley plants.     

Polyamine Oxidase5 Regulates Arabidopsis Growth through Thermospermine Oxidase Activity

The major plant polyamines (PAs) are the tetraamines spermine (Spm) and thermospermine (T-Spm), the triamine spermidine, and the diamine putrescine. PA homeostasis is governed by the balance between biosynthesis and catabolism; the latter is catalyzed by polyamine oxidase (PAO). Arabidopsis (Arabidopsis thaliana) has five PAO genes, AtPAO1 to AtPAO5, and all encoded proteins have been biochemically characterized. All AtPAO enzymes function in the back-conversion of tetraamine to triamine and/or triamine to diamine, albeit with different PA specificities. Here, we demonstrate that AtPAO5 loss-of-function mutants (pao5) contain 2-fold higher T-Spm levels and exhibit delayed transition from vegetative to reproductive growth compared with that of wild-type plants. Although the wild type and pao5 are indistinguishable at the early seedling stage, externally supplied low-dose T-Spm, but not other PAs, inhibits aerial growth of pao5 mutants in a dose-dependent manner. Introduction of wild-type AtPAO5 into pao5 mutants rescues growth and reduces the T-Spm content, demonstrating that AtPAO5 is a T-Spm oxidase. Recombinant AtPAO5 catalyzes the conversion of T-Spm and Spm to triamine spermidine in vitro. AtPAO5 specificity for T-Spm in planta may be explained by coexpression with T-Spm synthase but not with Spm synthase. The pao5 mutant lacking T-Spm oxidation and the acl5 mutant lacking T-Spm synthesis both exhibit growth defects. This study indicates a crucial role for T-Spm in plant growth and development.Polyamines (PAs) are low-molecular mass aliphatic amines that are present in almost all living organisms. Cellular PA concentrations are governed primarily by the balance between biosynthesis and catabolism. In plants, the major PAs are the diamine putrescine (Put), the triamine spermidine (Spd), and the tetraamines spermine (Spm) and thermospermine (T-Spm; Kusano et al., 2008; Alcázar et al., 2010; Mattoo et al., 2010; Takahashi and Kakehi, 2010; Tiburcio et al., 2014). Put is synthesized from Orn by Orn decarboxylase and/or from Arg by three sequential reactions catalyzed by Arg decarboxylase (ADC), agmatine iminohydrolase, and N-carbamoylputrescine amidohydrolase. Arabidopsis (Arabidopsis thaliana) does not contain an ORNITHINE DECARBOXYLASE gene (Hanfrey et al., 2001) and synthesizes Put from Arg via the ADC pathway. Put is further converted to Spd via an aminopropyltransferase reaction catalyzed by spermidine synthase (SPDS). In this reaction, an aminopropyl residue is transferred to Put from decarboxylated S-adenosyl-Met, which is synthesized by S-adenosyl-Met decarboxylase (SAMDC; Kusano et al., 2008). Spd is then converted to Spm or T-Spm, reactions catalyzed in Arabidopsis by spermine synthase (SPMS; encoded by SPMS) or thermospermine synthase (encoded by Acaulis5 [ACL5]), respectively (Hanzawa et al., 2000; Knott et al., 2007; Kakehi et al., 2008; Naka et al., 2010). A recent review reports that T-Spm is ubiquitously present in the plant kingdom (Takano et al., 2012).The PA catabolic pathway has been extensively studied in mammals. Spm and Spd acetylation by Spd/Spm-N¹-acetyltransferase (Enzyme Commission no. 2.3.1.57) precedes the catabolism of PAs and is a rate-limiting step in the catabolic pathway (Wallace et al., 2003). A mammalian polyamine oxidase (PAO), which requires FAD as a cofactor, oxidizes N¹-acetyl Spm and N¹-acetyl Spd at the carbon on the exo-side of the N⁴-nitrogen to produce Spd and Put, respectively (Wang et al., 2001; Vujcic et al., 2003; Wu et al., 2003; Cona et al., 2006). Mammalian spermine oxidases (SMOs) perform oxidation of the carbon on the exo-side of the N⁴-nitrogen to produce Spd, 3-aminopropanal, and hydrogen peroxide (Vujcic et al., 2002; Cervelli et al., 2003; Wang et al., 2003). Thus, mammalian PAOs and SMOs are classified as back-conversion (BC)-type PAOs.In plants, Spm, T-Spm, and Spd are catabolized by PAO. Plant PAOs derived from maize (Zea mays) and barley (Hordeum vulgare) catalyze terminal catabolism (TC)-type reactions (Tavladoraki et al., 1998). TC-type PAOs oxidize the carbon at the endo-side of the N⁴-nitrogen of Spm and Spd to produce N-(3-aminopropyl)-4-aminobutanal and 4-aminobutanal, respectively, plus 1,3-diaminopropane and hydrogen peroxide (Cona et al., 2006; Angelini et al., 2008, 2010). The Arabidopsis genome contains five PAO genes, designated as AtPAO1 to AtPAO5. Four recombinant AtPAOs, AtPAO1 to AtPAO4, have been homogenously purified and characterized (Tavladoraki et al., 2006; Kamada-Nobusada et al., 2008; Moschou et al., 2008; Takahashi et al., 2010; Fincato et al., 2011, 2012). AtPAO1 to AtPAO4 possess activities that convert Spm (or T-Spm) to Spd, called partial BC, or they convert Spm (or T-Spm) first to Spd and subsequently to Put, called full BC. Ahou et al. (2014) report that recombinant AtPAO5 also catalyzes a BC-type reaction. Therefore, all Arabidopsis PAOs are BC-type enzymes (Kamada-Nobusada et al., 2008; Moschou et al., 2008; Takahashi et al., 2010; Fincato et al., 2011, 2012; Ahou et al., 2014). Four of the seven PAOs in rice (Oryza sativa; OsPAO1, OsPAO3, OsPAO4, and OsPAO5) catalyze BC-type reactions (Ono et al., 2012; Liu et al., 2014a), whereas OsPAO7 catalyzes a TC-type reaction (Liu et al., 2014b). OsPAO2 and OsPAO6 remain to be characterized, but may catalyze TC-type reactions based on their structural similarity with OsPAO7. Therefore, plants possess both TC-type and BC-type PAOs.PAs are involved in plant growth and development. Recent molecular genetic analyses in Arabidopsis indicate that metabolic blocks at the ADC, SPDS, or SAMDC steps lead to embryo lethality (Imai et al., 2004; Urano et al., 2005; Ge et al., 2006). Potato (Solanum tuberosum) plants with suppressed SAMDC expression display abnormal phenotypes (Kumar et al., 1996). It was also reported that hydrogen peroxide derived from PA catabolism affects root development and xylem differentiation (Tisi et al., 2011). These studies indicate that flux through metabolic and catabolic PA pathways is required for growth and development. The Arabidopsis acl5 mutant, which lacks T-Spm synthase activity, displays excessive differentiation of xylem tissues and a dwarf phenotype, especially in stems (Hanzawa et al., 2000; Kakehi et al., 2008, 2010). An allelic ACL5 mutant (thickvein [tkv]) exhibits a similar phenotype as that of acl5 (Clay and Nelson, 2005). These results indicate that T-Spm plays an important role in Arabidopsis xylem differentiation (Vera-Sirera et al., 2010; Takano et al., 2012).Here, we demonstrate that Arabidopsis pao5 mutants contain 2-fold higher T-Spm levels and exhibit aerial tissue growth retardation approximately 50 d after sowing compared with that of wild-type plants. Growth inhibition of pao5 stems and leaves at an early stage of development is induced by growth on media containing low T-Spm concentrations. Complementation of pao5 with AtPAO5 rescues T-Spm-induced growth inhibition. We confirm that recombinant AtPAO5 catalyzes BC of T-Spm (or Spm) to Spd. Our data strongly suggest that endogenous T-Spm levels in Arabidopsis are fine tuned, and that AtPAO5 regulates T-Spm homeostasis through a T-Spm oxidation pathway.     

MUCILAGE-RELATED10 Produces Galactoglucomannan That Maintains Pectin and Cellulose Architecture in Arabidopsis Seed Mucilage

Plants invest a lot of their resources into the production of an extracellular matrix built of polysaccharides. While the composition of the cell wall is relatively well characterized, the functions of the individual polymers and the enzymes that catalyze their biosynthesis remain poorly understood. We exploited the Arabidopsis (Arabidopsis thaliana) seed coat epidermis (SCE) to study cell wall synthesis. SCE cells produce mucilage, a specialized secondary wall that is rich in pectin, at a precise stage of development. A coexpression search for MUCILAGE-RELATED (MUCI) genes identified MUCI10 as a key determinant of mucilage properties. MUCI10 is closely related to a fenugreek (Trigonella foenumgraecum) enzyme that has in vitro galactomannan α-1,6-galactosyltransferase activity. Our detailed analysis of the muci10 mutants demonstrates that mucilage contains highly branched galactoglucomannan (GGM) rather than unbranched glucomannan. MUCI10 likely decorates glucomannan, synthesized by CELLULOSE SYNTHASE-LIKE A2, with galactose residues in vivo. The degree of galactosylation is essential for the synthesis of the GGM backbone, the structure of cellulose, mucilage density, as well as the adherence of pectin. We propose that GGM scaffolds control mucilage architecture along with cellulosic rays and show that Arabidopsis SCE cells represent an excellent model in which to study the synthesis and function of GGM. Arabidopsis natural varieties with defects similar to muci10 mutants may reveal additional genes involved in GGM synthesis. Since GGM is the most abundant hemicellulose in the secondary walls of gymnosperms, understanding its biosynthesis may facilitate improvements in the production of valuable commodities from softwoods.The plant cell wall is the key determinant of plant growth (Cosgrove, 2005) and represents the most abundant source of biopolymers on the planet (Pauly and Keegstra, 2010). Consequently, plants invest a lot of their resources into the production of this extracellular structure. Thus, it is not surprising that approximately 15% of Arabidopsis (Arabidopsis thaliana) genes are likely dedicated to the biosynthesis and modification of cell wall polymers (Carpita et al., 2001). Plant walls consist mainly of polysaccharides (cellulose, hemicellulose, and pectin) but also contain lignin and glycoproteins. While the biochemical structure of each wall component has been relatively well characterized, the molecular players involved in their biogenesis remain poorly understood (Keegstra, 2010). The functions of the individual polymers, and how they are assembled into a three-dimensional matrix, are also largely unknown (Burton et al., 2010; Burton and Fincher, 2012).Significant breakthroughs in cell wall research have been achieved through the examination of specialized plant tissues that contain elevated levels of a single polysaccharide (Pauly and Keegstra, 2010). Some species, particularly legumes, accumulate large amounts of the hemicellulose galactomannan during secondary wall thickening of the seed (Srivastava and Kapoor, 2005). Analysis of the developing fenugreek (Trigonella foenumgraecum) endosperm led to the purification of a GALACTOMANNAN GALACTOSYLTRANSFERASE (TfGMGT), the first glycosyltransferase (GT) whose activity in plant cell wall synthesis was demonstrated in vitro (Scheller and Ulvskov, 2010). TfGMGT catalyzes the decoration of mannan chains with single α-1,6-galactosyl residues (Edwards et al., 1999). A similar approach in guar (Cyamopsis tetragonoloba) seeds revealed that the β-1,4-linked mannan backbone is synthesized by a member of the CELLULOSE SYNTHASE-LIKE A (CSLA) protein family (Dhugga et al., 2004).Galactomannan functions as a storage polymer in the endosperm of the aforementioned seeds, analogous to starch in cereal grains (Dhugga et al., 2004), but it also has important rheological properties in the cell wall that have been exploited to produce valuable stabilizers and gelling agents for human consumption (Srivastava and Kapoor, 2005). The Man-to-Gal ratio is essential for the application of galactomannan gums in the food industry (Edwards et al., 1992). This is because unsubstituted mannan chains can interact via hydrogen bonds to produce crystalline microfibrils similar to cellulose (Millane and Hendrixson, 1994). Indeed, some algae that lack cellulose employ mannan fibrils as a structural material (Preston, 1968). The addition of Gal branches to the smooth, ribbon-like mannan chains creates hairy regions that limit self-association and promote gelation (Dea et al., 1977). All mannans are likely synthesized as highly substituted polymers that are trimmed in the cell wall (Scheller and Ulvskov, 2010).Generally, polysaccharides containing backbones of β-1,4-linked Man units can be classified as heteromannan (HM). Galactoglucomannan (GGM) is the main hemicellulose in gymnosperm secondary walls and, in contrast to galactomannan, has a backbone that contains both Glc and Man units (Pauly et al., 2013). HM is detected in most Arabidopsis cell types (Handford et al., 2003) and facilitates embryogenesis (Goubet et al., 2009), germination (Rodríguez-Gacio et al., 2012), tip growth (Bernal et al., 2008), and vascular development (Benová-Kákosová et al., 2006; Yin et al., 2011). In the last 10 years, in vitro mannan synthase activity has been demonstrated for recombinant CSLA proteins from many land plants (Liepman et al., 2005, 2007; Suzuki et al., 2006; Gille et al., 2011; Wang et al., 2012). HM synthesis may also involve CELLULOSE SYNTHASE-LIKE D (CSLD) enzymes and MANNAN SYNTHESIS-RELATED (MSR) accessory proteins (Yin et al., 2011; Wang et al., 2013), but their precise roles in relation to the CSLAs have not been established. Arabidopsis CSLA2, like most other isoforms, can use both GDP-Man and GDP-Glc as substrates in vitro (Liepman et al., 2005, 2007) and is responsible for stem glucomannan synthesis in vivo along with CSLA3 and CSLA7 (Goubet et al., 2009). CSLA2 also participates in the synthesis of glucomannan present in mucilage produced by seed coat epidermal (SCE) cells (Yu et al., 2014).Arabidopsis SCE cells represent an excellent genetic model in which to study the synthesis, polar secretion, and modification of polysaccharides, since these processes dominate a precise stage of seed coat development but are not essential for seed viability in laboratory conditions (Haughn and Western, 2012; North et al., 2014; Voiniciuc et al., 2015). Hydration of mature seeds in water releases a large gelatinous capsule, rich in the pectic polymer rhamnogalacturonan I, which can be easily stained or extracted (Macquet et al., 2007). Biochemical and cytological experiments indicate that Arabidopsis seed mucilage is more than just pectin and, in addition to cellulose, is likely to contain glycoproteins and at least two hemicellulosic polymers (Voiniciuc et al., 2015). There is mounting evidence that, despite their low abundance, these components play critical functions in seed mucilage architecture. The structure of homogalacturonan (HG), the major pectin in primary cell walls but a minor mucilage component, appears to be a key determinant of gelling properties and mucilage extrusion (Rautengarten et al., 2008; Saez-Aguayo et al., 2013; Voiniciuc et al., 2013). Mucilage attachment to seeds is maintained by the SALT OVERLY SENSITIVE5 glycoprotein and cellulose synthesized by multiple CELLULOSE SYNTHASE (CESA) isoforms (Harpaz-Saad et al., 2011; Mendu et al., 2011; Sullivan et al., 2011; Griffiths et al., 2014, 2015). From more than 35 genes that are reported to affect Arabidopsis seed mucilage properties (Voiniciuc et al., 2015), only CSLA2, CESA3, CESA5, GALACTURONOSYLTRANSFERASE11 (GAUT11; Caffall et al., 2009), and GAUT-LIKE5 (GATL5; Kong et al., 2013) are predicted to encode GTs. This highlights that, despite many detailed studies about mucilage production in SCE cells, the synthesis of its components remains poorly understood.To address this issue, we conducted a reverse genetic search for MUCILAGE-RELATED (MUCI) genes that may be required for polysaccharide biosynthesis. One of these, MUCI10, encodes a member of the Carbohydrate Active Enzymes family, GT34 (Lombard et al., 2014), which includes at least two enzymatic activities and seven Arabidopsis proteins (Keegstra and Cavalier, 2010). Five of them function as XYLOGLUCAN XYLOSYLTRANSFERASES (XXT1–XXT5) in vivo and/or in vitro (Faik et al., 2002; Cavalier et al., 2008; Vuttipongchaikij et al., 2012). MUCI10/GT7 (At2g22900) and its paralog GT6 (At4g37690) do not function as XXTs (Vuttipongchaikij et al., 2012) and are more closely related to the TfGMGT enzyme (Faik et al., 2002; Keegstra and Cavalier, 2010). MUCI10, also called GALACTOSYLTRANSFERASE-LIKE6 (GTL6), served as a Golgi marker in multiple proteomic studies of Arabidopsis callus cultures (Dunkley et al., 2004, 2006; Nikolovski et al., 2012, 2014). Nevertheless, the role of TfGMGT orthologs in Arabidopsis remained unknown. We show that MUCI10 is responsible for the extensive galactosylation of glucomannan in mucilage and influences glucomannan backbone synthesis, cellulose structure, and the distribution of pectin.