Biosynthesis of Cutin and Suberin

Biosynthesis of the Monomers
The aliphatic monomers of cutin and suberin derive from the general fatty acid biosynthetic pathway, that is, from palmitic (16:0), stearic (18:0), and oleic (18:1) acids synthesized in the plastids of the epidermal cell.

The biosynthetic pathway leading to the characteristic cutin monomers had been largely discovered by the group of Kolattukudy in the early 70s (Kolattukudy, 1981). The major cutin monomers are synthesized by multiple hydroxylation and epoxidation reactions. These reactions are catalyzed by oxygen and NADP-dependent enzyme systems that are inhibited by CO, a typical characteristics of cytochrome P450-dependent enzymes. The research on plant cytochrome P450 has advanced much during the recent years (Kahn and Durst, 2000). Different cytochrome P450-dependent enzymes have been characterized that catalyze the internal as well as the o-hydroxylation of fatty acids (Beneviste et al., 1998; Cabello-Hurtado et al., 1998; Pinot et al., 1992, 1998; Tijet et al., 1998). Several of these cytochrome P450-dependent monooxygenases have been cloned, including CYP86A1, CYP94A1, CYP81B1, CYP86A8 (LCR), and CYP86A2 (ATT1) (Beneviste et al., 1998; Cabello-Hurtado et al., 1998; Tijet et al., 1998; Wellesen et al., 2001; Xiao et al., 2004). A function in cutin biosynthsis has been confirmed for LCR and ATT1 (Yephremov, unpublished results) (Xiao et al., 2004). Mutations in ATT1 of Arabidsopsis lead to a 30% loss in cutin and a much looser cuticular ultrastructure (Xiao et al., 2004). Alteration in the monomer composition of residual-bound lipids has been found in lcr plants (Yephremov, unpublished results).

A lipoxygenase/peroxygenase/epoxide hydrolase pathway has also been demonstrated for the synthesis of cutin monomers (Ble´e and Schuber, 1993). A peroxygenase may catalyze a hydroxyperoxide-dependent epoxidation of unsaturated fatty acids after the action of a lipoxygenase (Blée and Schuber, 1993, 1990; Hamberg and Hamberg, 1990). The cis-epoxy group formed by the peroxygenase may then be hydrated in the trans position by an epoxide hydrolase, resulting in a threo-diol in mid-chain position of the cutin monomers (Blée and Schuber, 1992, 1995; Pinot et al., 1997; Morisseau et al., 2000).

For formation of the cutin monomers, fatty acids leave the plastid after release from the fatty acid synthetase since cytochrome P450-dependent enzymes are located at the endoplasmic reticulum (ER) membrane. Precursors for the unsaturated cutin monomers of Arabidopsis are provided by phospholipids of the ER (Bonaventure et al., 2004). Further details on the mechanism of the hydroxylation reactions have not yet been elucidated, that is, the order of hydroxylations and the substrates for the different enzymes in vivo or other enzymes and cofactors involved. The acyl-CoA synthetase LACS2 has been found to be involved in the synthesis of the cuticular membrane, indicating that changes in the activation status of the precursors of cutin monomers are necessary during cutin biosynthesis (Schnurr et al., 2004). Recombinant LACS2 has a higher activity with 16-hydroxypalmitate than with palmitate (Schnurr et al., 2004).

In Arabidopsis, the characterization of HOTHEAD/ADHESION OF CALYX EDGES (HTH/ACE) identified a gene encoding a long-chain fatty acid ω-alcohol dehydrogenase belonging to glucose-methanol-choline oxidoreductase domaincontaining proteins (Krolikowski et al., 2003; Kurdyukov et al., 2006a). HTH/ACE catalyzes the formation of oxo-acids that are the precursors for a significant proportion of α,ω-dicarboxylic acids in Arabidopsis stem cutin (Kurdyukov et al., 2006a).

The very-long-chain fatty acid derivatives of suberin are synthesized by fatty acid elongases that catalyze the elongation of the carbon chain of stearate to different lengths, as found in wax biosynthesis (Domergue et al., 1998). Rootspecific fatty acid elongases have been characterized from maize (Schreiber et al., 2000). The necessary hydroxylation steps may be introduced by cytochrome P450-dependent enzymes. The formation of α,ω-dicarboxylic acids from ω-hydroxyacids is catalyzed by a ω-hydroxy fatty acid dehydrogenase (Agrawal and Kolattukudy, 1978a,b). A cytochrome P450 that oxidizes fatty acids to the corresponding ω-alcohols and subsequently to the α,ω-dicarboxylic acids was described (Le Bouquin et al., 2001). Another possibility may be that HTH/ACE, or one of the closely related proteins, are involved in the formation of α,ω-dicarboxylic acids present in suberin (Kurdyukov et al., 2006a). While the major types of enzymes responsible for the synthesis of aliphatic suberin monomers have been identified, none of them have been shown to be directly involved in suberin biosynthesis.

Our knowledge of cutin biosynthesis is likely to increase rapidly in the future, since genome-wide transcriptional profiling has been combined with polyester analysis in the epidermis of Arabidopsis stem sections, while such approaches still need to be attempted for suberized cells (Suh et al., 2005).

Formation of the Polyesters
How the cutin and suberin monomers are transported to the place of polymerization is still to be elucidated. On the other hand, an ATP-transporter involved in the transport of wax molecules across the plasmalemma has been identified by cloning the CER5 gene of Arabidopsis (Kunst and Samuels, 2003; Pighin et al., 2004). The very-long-chain fatty acid derivatives of suberin may be transported by a similar mechanism. However, the transport mechanism of cutin monomers also remains to be discovered; cutin monomers have a shorter chain length and much lower hydrophobicity than wax molecules. For cutin formation, an additional transport through the cell wall is required, and this transport step may involve lipoproteins. This model was proposed after proteins with the activity to transport lipids in vitro (lipid-transfer proteins) were localized to the cell wall (Kader, 1996). Although lots of circumstantial evidence for this function of lipidtransfer proteins have been collected, the direct involvement of lipid-transfer proteins in cutin biosynthesis has not yet been substantiated (Hollenbach et al., 1997; Pyee and Kolattukudy, 1995). Instead, recent work has indicated that lipidtransfer proteins act in plant defense against pathogens (Garcia-Olemedo et al., 1995; Maldonado et al., 2002; Molina and Garcia-Olmedo, 1997).

In order to form the three-dimensional structure of cutin and suberin, the respective monomers have to be linked together, in part by ester bonds. Classical chemical studies showed that cutin is mostly held together by primary alcohol– ester linkages between the cutin monomers with about half of the secondary hydroxyl groups involved in ester cross-links resulting in a polymeric network. The recent finding that glycerol is a substantial monomer of cutin and suberin makes it likely that the polyesters have a more complex structure whose formation remains largely to be discovered (Graça et al., 2002).

Some early studies showed that the cutin monomers bound to CoA as cofactors are transferred to free hydroxyl groups present in the cutin polymer (Croteau and Kolattukudy, 1973, 1975; Kolattukudy, 1981). An hydroxyl-CoA:cutin transacylase activity has been detected in a crude extract that needs ATP for the reaction as well as cutin polymer as a primer. However, the transacylase has not been purified and no gene encoding the enzyme has been identified. A putative acyl-CoA:cutin transferase has been claimed to be purified from Agave epidermis (Reina and Heredia, 2001). After partial protein sequencing, a gene was isolated that encodes a novel small valine-rich protein with a putative HxxxE domain present in other acyltransferases (Reina and Heredia, 2001).No confirmation exists to date, however, that this protein has the proposed function.

BODYGUARD, an enzyme of the α,β hydrolase family, has been found to be critically involved in the formation of the cuticular membrane of Arabidopsis (Kurdyukov et al., 2006b). In the bdg mutant, the cuticular membrane is disrupted and the outer extracellular matrix is disorganized, with polysaccharides coming to the surface and polyester also deposited within the cell wall. These structural changes were accompanied by a higher amount of residual-bound lipids of totally extracted leaves. BDG is extracellularly localized and thus functions directly in the formation/organization of the cuticular membrane (Kurdyukov et al., 2006b). Since some members in the α,β hydrolase superfamily have synthase activity, it is hypothesized that BDG may also be capable of synthesizing reactions in the cuticular layer of the cell wall. Potentially, BDG may even be capable of catalyzing hydrolysis as well as synthesis, depending on the conditions in which the reaction takes place (Kurdyukov et al., 2006b).

Mutants Affected in Cutin Deposition
The best means for linking enzyme activities and the corresponding proteins and genes to their respective functions is by mutation. The Sorghum bicolor bloomless (bm) mutant was the first mutant identified having a thinner cuticuler membrane as well as a reduced wax deposition. The bm mutant exhibits a higher conductance to water vapor and an increased susceptibility to the fungal pathogen Exserohilum turcicum ( Jenks et al., 1994). Since several aspects of the cuticle were altered in bm, it could not be determined which cuticular component contributes which feature of the cuticle. Furthermore, Sorghum is not a species well suited for map-based cloning of genes.

Isolation of Arabidopsis mutants only affected in the deposition of either cutin or suberin by a chemical or ultrastructural screening method would be extremely work-intensive and not feasible. Thus, secondary phenotypes of plants having an altered cutin or suberin structure have to be identified in order to use the large resources available in Arabidopsis for this research area.

A phenotype that was at first unexpected but found to be related to cuticular changes was organ fusion (Lolle and Cheung, 1993; Lolle and Pruitt, 1999; Lolle et al., 1997, 1998). Support for the idea that a disrupted cuticular membrane structure and/or less cutin lead to organ fusions was originally obtained by an indirect approach using transgenic Arabidopsis plants expressing and secreting a fungal cutinase and therefore degrading their own cutin (Sieber et al., 2000). These transgenic plants show an altered ultrastructure and a higher permeability of the cuticle. When organs having a disrupted cuticular membrane are in close contact early during development, fusions form most likely by cross polymerization. These organ fusions are very strong so that organs do not separate during further growth, leading to distortions of the growth habit of the plant (Sieber et al., 2000). A number of organ fusion mutants were shown to be altered in the cuticular polyester (Kurdyukov et al., 2006a,b). Organ fusions are still used as a selection criterion for mutants having changes in cuticular structure or composition (Yephremov and Schreiber, 2005).

A very simple and much more direct way to identify mutants with an increased permeability of the cuticle is by staining of plant tissues with a dye, for example, with toluidine blue (Tanaka et al., 2004). In addition, mutants with alterations in the cuticular membrane have been identified by various other phenotypes that were often not obviously associated with cuticular function, such as either altered resistance to pathogens or a number of changes in cell morphology and differentiation (Yephremov and Schreiber, 2005).

The increasing number of well-characterized Arabidopsis plants having alterations in the cuticular membrane enables some phenotype comparisons to be made. The organ fusion mutant bdg shares most of the phenotypes with cutinaseexpressing plants, such as an increased permeability of the cuticle, higher wax accumulation, ectopic pollen germination, stunted growth, altered trichome formation, and increased resistance to Botrytis cinerea (Kurdyukov et al., 2006b; Sieber et al., 2000). The characteristic difference in the structure of the cuticular membranes between cutinase-expressing plants and bdg mutants, namely, that bdg accumulates, in addition, large amounts of osmophilic material deeper within the cell wall, lead to the hypothesis that BDG acts directly in the formation of the extracellular matrix, as discussed above (Kurdyukov et al., 2006b).

Surprising are the differences in the phenotypes of att1 and lcr, two Arabidopsis mutants affected in a cytochrome P450 of the same subfamily and having a higher permeability as well as ultrastructural changes in the cuticular membrane of leaves (C. Nawrath, unpublished results) (Wellesen et al., 2001; Xiao et al., 2004). While lcr shows frequently organ fusions and alterations in trichome formation, demonstrating a link between cuticle structure and the development of epidermal cells, att1 does not show either organ fusions or any developmental disorders (Wellesen et al., 2001; Xiao et al., 2004). A direct link between plant disease resistance and the formation of the cuticular membrane was found in att1 mutants. Pseudomonas syringae pv. phasaelicula expresses high levels of the type III genes when colonizing the att1 mutant, demonstrating that ATT1 is important for the repression of bacterial virulence genes in wild-type plants (Xiao et al., 2004). Therefore, att1 mutants are more susceptible to P. syringae pv. tomato DC3000. It was speculated that the alteration of the cuticle in the substomatal chamber in which the bacteria reside is of relevance for the mechanism (Xiao et al., 2004).

The organ fusion mutant fdh lead to the identification of a fatty acid biosynthetic enzyme with homology to condensing enzymes of which the exact substrate is unknown (Pruitt et al., 2000; Yephremov et al., 1999). The fdh mutant shows, similarly to lcr, bdg, and cutinase-expressing plants, alterations in trichome formation and ectopic pollen germination, in addition to organ fusions and a higher permeability of the cuticle (Pruitt et al., 2000; Yephremov et al., 1999). However, the ultrastructure of the fusion itself in fdh plants has a very different structure in comparison to all other characterized organ fusion mutants since the cuticular membranes are not disrupted but fuse directly to each other (C. Nawrath, unpublished results).

The analysis of the molecular basis underlaying the obvious differences in composition, structure, and function will surely be of great interest during the next years (Nawrath, 2006).

WAX2/YORE-YORE is an Arabidopsis protein with six-membrane spanning domains having homology to the sterol desaturase family at the N-terminus and the short-chain dehydrogenase/reductase family at the C-terminus as well as having an overall homology to CER1, a protein required for wax deposition of unknown function in Arabidopsis (Chen et al., 2003; Kurata et al., 2003). In contrast to the other mutants having a looser cuticular membrane structure and loss of cuticular membrane material, wax2/yore-yore has, in addition, a reduced wax deposition (Chen et al., 2003; Kurata et al., 2003). Other phenotypes of the wax2/ yore-yore mutants are typical for cutin mutants, such as an increased permeability of the cuticularmembrane, disorders in the development of epidermal cell types, and organ fusions (Chen et al., 2003). Thus,WAX2 plays a critical role in the synthesis of both cuticular components, cutin and wax.

In addition to enzymes that are directly involved in either cutin monomer biosynthesis or polyester formation, a number of genes have been identified by mutations that are regulators of epidermal development and therefore lead to an abnormal cuticle (Aharoni et al., 2004; Becraft et al., 1996; Broun et al., 2004; Jin et al., 2000; Tanaka et al., 2001, 2002; Watanabe et al., 2004).

ABNORMAL LEAF SHAPE (ALE1) is a subtilisin-like protease that is involved in the regulation of the formation of the cuticle in embryos and juvenile plants in Arabidopsis (Tanaka et al., 2001). ale1 mutants have a disrupted cuticular membrane in embryos, cotyledons, and juvenile leaves. The leaves of ale1 are crinkled, often have organ fusions, and are very susceptible to low humidity, resulting in conditional lethality. Interestingly, ALE1 is expressed in certain endosperm cells adjacent to the embryo, as well as in the young embryo, and may be essential for the separation of the two entities (Tanaka et al., 2001).

CRINKLY4 (CR4) is a receptor kinase with homology to tumor-necrosis factor receptors that is involved in proper epidermal formation that has first been identified in maize (Becraft et al., 1996). CR4 mutants have organ fusions as well as abnormal epidermal cell wall and cuticle deposition (Jin et al., 2000). ACR4, the CR4 homologue in Arabidopsis that is expressed in the outer cell layers of embryos and mature plants, has similar function in epidermal differentiation and cuticle development as CR4 (Tanaka et al., 2002; Watanabe et al., 2004). ALE1 and ACR4 affect synergistically the differentiation and function of the epidermis, since ale1/acr4 double mutants have a stronger phenotype than do both single mutants (Watanabe et al., 2004).

The overexpression of SHINE/WAX INDUCER1, an AP2 domain transcription factor, leads to an increased wax deposition (Aharoni et al., 2004; Broun et al., 2004). In addition, the ultrastructure of cuticular membrane as well as permeability of the cuticle is changed. Furthermore, diverse aspects of epidermal differentiation are altered, such as epidermal cell structure, trichome shape and number, and stomatal index (Aharoni et al., 2004). These diverse phenotypes make the interpretation of the physiological analyses difficult. However, the expression pattern of the different genes of the shine clade, whose overexpression all result in similar phenotypes, suggest diverse functions in lipid and/or cell wall metabolism, including cutin and suberin deposition (Aharoni et al., 2004).