Functional Roles of Three Cutin Biosynthetic Acyltransferases in Cytokinin Responses and Skotomorphogenesis

Cytokinins (CKs) regulate plant development and growth via a two-component signaling pathway. By forward genetic screening, we isolated an Arabidopsis mutant named grow fast on cytokinins 1 (gfc1), whose seedlings grew larger aerial parts on MS medium with CK. gfc1 is allelic to a previously reported cutin mutant defective in cuticular ridges (dcr). GFC1/DCR encodes a soluble BAHD acyltransferase (a name based on the first four enzymes characterized in this family: Benzylalcohol O-acetyltransferase, Anthocyanin O-hydroxycinnamoyltransferase, anthranilate N-hydroxycinnamoyl/benzoyltransferase and Deacetylvindoline 4-O-acetyltransferase) with diacylglycerol acyltransferase (DGAT) activity in vitro and is necessary for normal cuticle formation on epidermis in vivo. Here we show that gfc1 was a CK-insensitive mutant, as revealed by its low regeneration frequency in vitro and resistance to CK in adventitious root formation and dark-grown hypocotyl inhibition assays. In addition, gfc1 had de-etiolated phenotypes in darkness and was therefore defective in skotomorphogenesis. The background expression levels of most type-A Arabidopsis Response Regulator (ARR) genes were higher in the gfc1 mutant. The gfc1-associated phenotypes were also observed in the cutin-deficient glycerol-3-phosphate acyltransferase 4/8 (gpat4/8) double mutant [defective in glycerol-3-phosphate (G3P) acyltransferase enzymes GPAT4 and GPAT8, which redundantly catalyze the acylation of G3P by hydroxyl fatty acid (OH-FA)], but not in the cutin-deficient mutant cytochrome p450, family 86, subfamily A, polypeptide 2/aberrant induction of type three 1 (cyp86A2/att1), which affects the biosynthesis of some OH-FAs. Our results indicate that some acyltransferases associated with cutin formation are involved in CK responses and skotomorphogenesis in Arabidopsis.

Seeds were treated at 4°C for 2 days in water. Before inoculation on MS, seeds were sterilized with 0.1% w/v HgCl 2 . For growth under long day condition, germination and plant growth took place with a 16-h light (60-70 μmol m −2 s −1 ) / 8-h dark cycle. For growth in darkness, germination took place under white light (60-70 μmol m −2 s −1 ) at 22°C for 3 h. Seeds were plated on MS (pH 5.8) with agar at either 1.0% (w/v) (vertical plate) or 0.8% (w/v) (horizontal plate) and 1% (w/v) sucrose and grown at 22°C.
To introduce a reporter gene into the mutant, pCYCB1:GUS was crossed with gfc1-1 and double homozygotes were identified in the F 3 generation.
For DIC, 5 DAG seedlings were grown on vertical plate under long day condition. For adventitious root formation assays and callus induction assays, 11 DAG seedlings grown on horizontal plate under long day condition were used.
For Toluidine blue-O staining, 3-week-old plants grown in soil under long day condition were used. Relative humidity was approximately 60%.

Gene cloning and sequence analysis
Inverse-PCR (I-PCR) was adopted to clone the mutant gene. Genomic DNA was digested completely with HindIII and ligated with T4 DNA ligase. Two rounds of PCR were performed with two sets of nested primers, LBb 1 / Z4 and LBb 1 / Z3 (S1 Appendix). The PCR fragments were subcloned and sequenced. The downstream flanking sequence was amplified by PCR with LBb 1.3 and 3362-F primers (S2 and S3 Appendix). Flanking sequences (S1 Appendix) were used for designing gene-specific primers to determine hetero-/homozygosity and co-segregation (S3 Appendix).

Phenotype characterization
For DIC, roots were cleared by the chloral hydrate method as described by Inagaki et al. [41].
For hypocotyl inhibition assays, seedlings were germinated and grown on vertical MS plates in darkness with different concentrations of CKs, ACC (Sigma), IAA (Sigma) and GA3 (Sigma) at 5 DAG.
For adventitious root formation assays and callus induction assays, the methods were performed according to [42].
For Toluidine blue-O staining, plants were immersed for 5 min in 0.05% Toluidine blue-O (Sigma) and rinsed with water at room temperature.
Molecular complementation and pGFC1:GUS transgenic plants A 2.2-kb promoter sequence was amplified using PGFC1-F2 / PGFC1-R primers (S2 Appendix) and sub-cloned into a modified pCAMBIA1300 binary vector harboring a GUS gene to generate a promoter:GUS reporter gene construct. The 1.6-kb full-length cDNA fragment of GFC1 gene was amplified by RT-PCR using GFC1-F / GFC1-R primers (S2 Appendix) and ligated downstream of the 2.2-kb GFC1 promoter to construct pGFC1:GFC1 in a pCAMBIA1300 vector. All amplified DNA fragments were confirmed by sequencing, and the constructed binary vectors were introduced into either WT plants (for pGFC1:GUS) or gfc1-1 plants (for pGFC1:GFC1) by an Agrobacterium tumefaciens-mediated (strain GV3101) floral-dip transformation method [43]. Primary transformants were isolated on MS containing 25mg/L hygromycin (Sigma) and transferred to soil to grow to maturity.
For CK-induced type-A ARR expression, seeds were germinated on MS plates and grown for 7 days. Treatments were carried out by incubating seedlings in liquid 1/2 MS culture medium containing 1% sucrose and supplemented with 10 μM tZ for 30 min [45].
For CK-induced GFC1 expression, we treated seedlings with 1 μM tZ for 30 min.

Endogenous CK measurement
Endogenous levels of CKs were determined by LC-MS/MS methods according to [46] with modifications. Briefly, ice-cold modified Bieleski buffer (methanol/water/formic acid, 15/4/1, v/v/v; [47]) and two SPE columns (C18 column-500 mg/Applied Separations, and MCX column-30 mg/Waters; [48]) were used to extract 100 mg seedlings (7 DAG). To each extract the stable isotope-labeled CK internal standards (0.5 pmol of CK bases, ribosides, N-glucosides, 1 pmol of O-glucosides and nucleotides) were further added as a reference. Analytes were eluted by two-step elution using a 0.35 M NH 4 OH aqueous solution and 0.35 M NH 4 OH in 60% (v/v) MeOH solution. All samples were then evaporated under vacuum at 37°C to dryness. Purified samples were analyzed by the LC-MS/MS system consisting of an ACQUITY UPLC System (Waters) and Xevo TQ-S (Waters) triple quadrupole mass spectrometer. Quantification was obtained using a multiple reaction monitoring (MRM) mode of selected precursor ions and the appropriate product ion. The linear range was over at least five orders of magnitude with a correlation coefficient of 0.9989-0.9998. For each mutant, four independent biological replicates were performed.

Measurement of ethylene production
Ethylene level was measured by gas chromatography as described [49]. Arabidopsis seedlings (5 DAG) grown on MS or MS with 10 μM tZ in darkness were used. Seedlings (100 mg) were placed in a 2 ml vial (Agilent Technologies, http://www.agilent.com) and sealed. After 2 h, 0.5 ml samples of the air inside the vials were used for the determination of ethylene production (Varian 450-GC, http://www.varian.com). For each treatment, four independent biological replicates were performed.

Mutant isolation and molecular identification of GFC1 gene
To isolate genes potentially involved in CK responses, we used a forward genetics approach. Approximately 40,000 T-DNA insertion lines [50] were screened to isolate mutants with altered responses to CK in terms of root/shoot growth. The gfc1-1 mutant was identified by the significantly large aerial components of seedlings grown on MS containing 10 μM tZ (grow fast on cytokinins 1, gfc1, Fig. 1A). In darkness, gfc1-1 mutants had short hypocotyls, which were insensitive to exogenous tZ, an opened apical hook, and over-grown cotyledons-the so-called de-etiolated-like phenotypes (Fig. 1B, details see below). Furthermore, gfc1-1 mutant had higher percentage of non-germinating seeds than WT (S2 Fig.).
To clone the GFC1 gene, we first back crossed the gfc1-1 mutant into WT for two generations. F 1 lines showed a WT-like phenotype and F 2 lines segregated for WT and gfc1 phenotypes in a 3:1 ratio (588:189), indicating that gfc1 is a single recessive mutation. Inverse PCR amplification of T-DNA flanking sequences (S1 Appendix) identified a T-DNA insertion in the only intron (1289 bp downstream of ATG) of the gene At5g23940, which has been reported to encode a soluble BAHD acyltransferase essential for normal cuticle formation on epidermis [37]. Additional sequence analysis of regions surrounding the T-DNA insertion site indicated no sequence alterations in the adjacent regions (S1 Appendix). Co-segregation analysis of the F 2 backcross progeny showed a linkage between the gfc1 mutant phenotype and the insertion in the GFC1 gene.
In order to verify that the phenotype of the gfc1-1 mutant indeed originated from a lesion in the At5g23940 gene, we identified an additional mutant allele, salk_128228c, with the T-DNA insertion at the second exon (2904 bp downstream of ATG) (Fig. 1D). Homozygous salk_128228c seedlings displayed the same gfc1 phenotypes ( Fig. 1A-C) and this allele was named gfc1-2, (dcr-2 in [37]). We also observed the cutin-associated phenotypes in gfc1 mutants (see below). When the homozygous gfc1-1 plants were transformed with the GFC1 genomic sequence under the control of the GFC1 promoter (pGFC1:GFC1), its mutant phenotypes were fully rescued when observed in T 2 homozygous transgenic plants ( Fig. 1A-C).

gfc1 mutant exhibited increased root meristematic activity
When grown on vertical MS plates, gfc1 seedlings showed a variety of phenotypes. The primary root length of gfc1 seedlings was significantly longer than that of WT ( Fig. 2A and 2B). The fresh weight of shoots was also significantly increased in the two mutants ( Fig. 2C).
To determine whether the increased primary root growth in gfc1 mutants is a result of increased root meristematic activity or not, we compared the root structures of gfc1 mutants with WT. DIC images showed that the root meristem zone length (MZ, extending from the quiescent center (QC) to the first elongated cell) [41] was significantly increased in gfc1 when compared with that in WT ( Fig. 2D and 2E). We introduced pCYCB1:GUS, a cell cycle marker for G2/M transition [51,52] into the gfc1-1 mutant. Consistent with meristem phenotype, the expression of pCYCB1:GUS gene was greatly increased in gfc1-1 seedlings (Fig. 2F).

gfc1 seedlings have altered responses to CKs
We examined the sensitivity of gfc1 to CK in several assays, focusing on CK-mediated growth and development. External application of CK significantly inhibits the growth of WT seedlings. When treated with tZ, root elongation of the gfc1 seedlings was normally inhibited ( Fig. 3A and 3B). While tZ inhibited WT shoot growth in terms of fresh weight (Fig. 3C), it strikingly induced significant fresh weight increases in the gfc1 shoots ( Fig. 3A and 3C). Other CKs had similar effects on gfc1-1 seedlings (S3 Fig.). However, unlike CKs, other hormones or their biosynthetic precursors, including the auxin indole-3-acetic acid (IAA), gibberellic acid (GA 3 ), and the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC), failed to induce significantly differential responses between gfc1 and WT (S4 Fig.).

Callus culture and adventitious root formation
CKs are used to stimulate cell division and greening/shoot initiation of callus tissue [53]. On media containing 0.05 μM 2,4-D with different concentrations of tZ, the CK-induced cell division and greening of hypocotyl-derived calli were reduced in both ahk3-1 and gfc1-1 mutants when compared to WT, with ahk3-1 showing the least response (Fig. 4A).
CKs normally inhibit adventitious root formation near the cut end of hypocotyls [54]. The gfc1 mutants were less sensitive to CKs in the adventitious root formation assay, similar to the ahk3-1 mutant (Fig. 4B). Skotomorphogenesis and CK treatment of gfc1 seedlings in darkness CK can induce dark-grown seedlings to develop some photomorphogenic or de-etiolation characteristics, including reduced hypocotyl growth, apical hook opening, cotyledon expansion, and the induction of leaf development (Fig. 5A) [55][56][57]. The gfc1 mutant was observed to have those obvious de-etiolation-like phenotypes without CK (Fig. 5A). However, gfc1  mutant seedlings displayed completely normal photomorphogenic responses to red light (Rc), blue light (Bc) and far red light (FRc) (S5 Fig.), suggesting that gfc1 is defective only in skotomorphogenesis.
In darkness, the gfc1 mutant exhibited short hypocotyls compared with WT ( Fig. 5A and 5B). They showed overgrown cotyledons and formed the first true leaves 14 days after germination (Fig. 5A). When treated with CKs, the hypocotyl length in dark-grown WT seedlings decreased significantly, while there was no significant decrease in hypocotyl length in gfc1 mutant ( Fig. 5A-C and S3 Fig.). However, gfc1 showed WT-like responses to the ethylene biosynthesis precursor ACC (Fig. 5D) and to IAA (Fig. 5F), both of which suppress the hypocotyl elongation of both gfc1 and WT seedlings in darkness [58]. GA 3 slowly stimulates rather than inhibits hypocotyl elongation in darkness. The gfc1-1 mutant showed a WT-like response to application of GA 3 (Fig. 5E).

CK-response genes: Elevated basal expression and reduced CK induction
CKs induce the transcription of type-A ARRs in Arabidopsis [22]. To determine whether the induction levels of these primary CK response genes were compromised in gfc1-1, real-time quantitative PCR (qRT-PCR) analysis was performed on RNA prepared from WT and gfc1-1 mutants. Exogenous CK treatment of gfc1-1 did up-regulate the transcription of type-A ARRs, although the relative induction levels were lower than those in WT (Fig. 6), indicating that the mutant has a reduced CK response. Comparison of the basal expression levels of these ARRs in WT and gfc1-1 on MS without CK showed that most type-A ARRs displayed significantly higher levels in gfc1-1 ( Fig. 7A and  7C). Overexpression of type-A ARRs is reported to lead to decreased CK sensitivity [27,40]. In ARR15-OX plant, the CK induction levels of ARR4 and ARR7 were significantly decreased [27]. Basal up-regulation of some type-A ARRs and their relatively reduced responses to CK may be only part of the reason for gfc1 CK-insensitivity (Fig. 8). Unlike type-A ARRs, the expressions of type-B ARRs (ARR1, 10 and 12) were not significantly different between gfc1-1 and WT ( Fig. 7B and 7D).
Defective GPATs for cutin biosynthesis can phenocopy gfc1 GFC1/DCR has been shown to be essential for the assembly of cutin polyesters [37] and encodes a soluble BAHD acyltransferase that can catalyze the incorporation of dicarboxylic fatty acids (DFA) or OH-FAs into diacylglycerol to form TAG in vitro [38]. This raised the question of whether other enzymes in the cutin biosynthesis pathway, such as GPATs for sn-2 MAG synthesis and some P450 monooxygenases for FAs hydroxylation [34], have the same or similar roles in CK response.
It has been reported that GPAT4 and GPAT8 are functionally redundant and that a gpat4/8 double mutant is defective in cutin biosynthesis [59]. Strikingly, we found that the gpat4/8 double mutant had the same or similar dark-grown and CK-induced phenotypes as gfc1 (Fig. 9A-B and S6A-D Fig.). Significantly, the phenotypes were more severe in the gfc1-1/gpat4/8 triple mutants (Fig. 9A-B and S6A-D Fig.).
Several cytochrome P450 monooxygenases, such as CYP86As, can catalyze the ω-hydroxylation of fatty acyl chains, which are the major cutin monomers. Similar to gfc1 and gpat4/8, the cyp86A2/att1 mutant has epidermal cuticle defects with increased permeability as revealed by toluidine blue-O staining (Fig. 9C) [59,60]. In contrast to gfc1 and gpat4/8, however, when treated with CK or grown in darkness, the cyp86A2/att1 mutant failed to show gfc1-like phenotypes (Fig. 9A-B and S7A-D Fig.), suggesting that the gfc1 phenotypes were not necessarily related to general cuticle defects in epidermis.

Tissue distribution of GFC1 expression
The expression of GFC1 was followed by introducing a GFC1 promoter-GUS reporter fusion (pGFC1:GUS) into WT. Staining of the GUS activity in independent transgenic lines in T 4 progenies showed that the GFC1 promoter was highly active in cotyledon and the apical hook (Fig. 10A-B and 10K) in seedlings. In mature plants, strong GUS staining was observed in the upper stem, inflorescences, and siliques ( Fig. 10C-J), results which are similar to those reported for DCR in [37]. Notably, the expression level of GFC1 is not elevated by exogenous tZ (Fig. 10L). These results are consistent with the gene expression data retrieved from the GENE-VESTIGATOR using Meta-Analyzer. Three Acyltransferases and Cytokinin Responses Endogenous CKs in gfc1-1 and gpat4/8 mutants In Arabidopsis, CKs can induce de-etiolation in WT plants [55,57,61,62]. To know if the de-etiolation and altered CK-response phenotypes in gfc1 and gpat4/8 mutants were associated with any changes in CK metabolism, we compared the levels of various endogenous CKs between 7DAG seedlings of WT and the mutants by LC-MS/MS. The results showed no significant differences between the two mutants and the WT in either total CK content or each of the total tZ/cZ/DHZ/iP types (Table 1). Although both mutants had higher levels of active tZ-type cytokinins than WT, they showed reduced levels in free iP but increased levels in iPRMP. Some of these differences were not statistically significant. iPR was significantly decreased in gfc1-1, but increased in gpat4/8. Because of the strong and very similar CK-associated gfc1 phenotypes between these two mutants, the CK quantification results revealed no obvious link between CK homeostasis and the gfc1 phenotype.

Discussion
Cutin biosynthetic GFC1 and GPATs are essential for normal seedling development and CK responses Our present results reveal that GFC1/DCR and at least two GPATs, genes encoding acyltransferases that catalyze incorporation of OH-FAs into cutin monomers or polymer [63], significantly influence CK responses and plant development, including skotomorphogenesis in darkness. Previously, DCR was identified as a candidate gene whose expression is closely associated with cutin metabolism. Mutation of DCR resulted in many typical phenotypes associated with defective cuticle (Fig. 9C), such as altered epidermal cell differentiation and post genital organ fusion, as well as sensitivity to saline, osmotic, and water stress conditions [37], which we also observed in the T-DNA gfc1 mutants.
In our present study, we found that in darkness gfc1 had the de-etiolated phenotypes of short hypocotyls, early opening apical hooks, and overgrown cotyledons (Fig. 5A). Despite its defect in skotomorphogenesis, gfc1 responded normally to FRc, Rc and Bc light (S5 Fig.), indicating that GFC1/DCR is important for skotomorphogenesis in darkness but dispensable for photomorphogenesis under light. Under normal growth conditions, gfc1 mutant seedlings exhibited stronger staining of the cell cycle reporter pCYCB1:GUS in root and shoot apexes (Fig. 2F), longer primary root length ( Fig. 2A and 2B) and higher fresh weight of shoots (Fig. 2C), suggesting functional roles of GFC1/DCR in controlling cell division and differentiation. Pleiotropic mutants with alterations to not only epidermal cuticle integrity but also nonepidermal cell division and differentiation have also been observed in other cuticle mutants [64][65][66].
gfc1 was isolated as a CK response mutant in our present study by a forward genetic screen. While the primary roots of gfc1 responded normally to exogenous CK treatment (Fig. 3A and  3B), the shoots were completely insensitive, even becoming larger in size when treated with CKs ( Fig. 3A and 3C). The gfc1 mutant also showed full insensitivity to CK in darkness, as indicated by hypocotyl length (Fig. 5A-C). Notably, such strong differential responses between gfc1 Three Acyltransferases and Cytokinin Responses and WT were not observed after treatment with IAA, ACC or GA 3 (Fig. 5D-F), indicating that these gfc1 phenotypes are CK-specific. Similar to the CK signaling mutant ahk3, gfc1 mutant was less sensitive to CK in adventitious root formation than WT (Fig. 4B). It should be noted that the effect of gfc1 mutation on its responses to CK in adventitious root production (Fig. 4) was weaker than those on shoots (Fig. 3) or dark-grown hypocotyls (Fig. 5), implying that either GFC1/DCR plays limited roles in those processes or its functional loss can be compensated by other related processes.
Although the photomorphogenic phenotype of gfc1 in darkness (Fig. 5A) implies that the plant should have high endogenous CK levels [55,57,61,62], the mutant did lack other typical high CK phenotypes, such as a bushy appearance. Paradoxically, gfc1 had longer primary roots with an enlarged MZ, which has been associated with reduced endogenous CK level or signaling [67]. These conflicting phenotypes between gfc1 and mutants affecting CK levels indicate that the gfc1 phenotypes are not likely linked to changes in general CK homeostasis. This conclusion is consistent with our CK quantification results (Table 1), in which no causal link was found between levels of active CKs and the gfc1-like phenotypes in gfc1 and gpat4/8 mutants. Taken together, these results suggest that the gfc1 mutation leads to defects in CK responses. Asterisks indicate statistically significant differences between the mutant lines (gpat4/8 or gfc1-1) and the wild type In the CK signaling pathway, the type-B ARRs positively regulate CK responses by activating the transcription of their downstream targets, including the type-A ARR genes. Type-A ARRs are rapidly activated in response to exogenous CK and then down-regulate the CK responses through a negative-feedback loop [10]. Consistent with that, multiple loss-of-function mutants in type-A ARR genes are hypersensitive to CK in various assays, including inhibition of root elongation, lateral root initiation and callus formation, while over-expression of type-A ARR genes can lead to decreased CK sensitivity [20,40]. Our present results show that the basal expression levels of most type-A ARR genes were higher in gfc1, but that their relative induction levels by exogenous CK were lower than those in WT (Figs. 6, 7A and 7C), Like gfc1, ARR-OX lines show decreased CK-sensitivity (Fig. 8), which indicates that up-regulation of type-A ARRs in gfc1 and their reduced responses to CK is only part of the reason for the insensitivity of gfc1 to CK.

Interaction between CK and ethylene in hypocotyl elongation is disrupted in gfc1
There is strong cross talk between CK and ethylene in plant growth and development. The application of CK to dark-grown plants can result in the 'triple response' that is characteristic of ethylene, and a major part of the effect of CK on root and hypocotyl growth has been reported to be mediated by ethylene [68]. In this process, CK post-transcriptionally increases the activity of the ethylene biosynthesis gene ACS5, leading to an elevated level of ethylene biosynthesis [30,31]. We showed that gfc1 hypocotyls are fully insensitive to CK but positively respond to the ethylene precursor ACC (Fig. 5C and 5D). Consideration of the CK-ethylene cross-talk suggests that signaling from CK to ethylene biosynthesis is disrupted in the gfc1 mutant. Consistent with that, tZ failed to significantly increase the production of ethylene in dark-grown gfc1-1 seedlings (Fig. 11).

Concurrence of gfc1 phenotypes and blocked acyl-transfer in cutin biosynthesis
Biosynthesis of the epidermal cutin occurs through a complex process that consists of FA synthesis, activation into acyl-CoA, ωand/or in-chain oxygenation, sn-2 MAG synthesis, monomer/oligomer transport out of the cell to the surface and polymerization into cutin polyester [34]. While the effects of gfc1/dcr and gpat4/8 mutations on cuticle structure can easily be explained by the loss of their respective acyltransferase activities, the biochemical/molecular mechanisms of their effects on CK responses and skotomorphogenesis, which were not reported when the other cuticle mutants were isolated, remain elusive.
The gfc1/dcr, gpat4/8, and cyp86A2/att1 mutants presented similarly defective cuticles, in terms of decreased cutin monomer loads and increased permeability to solutes (Fig. 9C), but cyp86A2/att1 had normal CK responses and skotomorphogenesis ( Fig. 9A and 9B), indicating that the phenotypes of gfc1 and gpat4/8 are not simply due to the lack of an intact cuticle layer in the epidermis. Cuticle mutants often display pleiotropic phenotypes, some of which appear to be not directly linked to their primary effects on cuticle integrity, such as alteration to nonepidermal cell development [37,[64][65][66]. Another example is the sensitivity to pathogen infection. Xiao et al. (2004) reported that Arabidopsis CYP86A2/ATT1 is required for cuticle development and can represses Pseudomonas syringae type III genes and that cyp86A2/att1, but not wax2 mutation, could lead to enhanced avrPto-luc expression. Both mutants had similar cuticle defects, but cyp86A2/att1 had an additional gene-specific but not cuticle-specific effect. They suggested that certain cutin-related FAs synthesized by CYP86A2 may repress bacterial type III gene expression [69]. GPAT4/8 catalyzes the formation of sn2-OH-MAG, one of the major monomers of cutin, by sn-2-specific G3P: acyl-CoA acyltransferase as well as phosphatase activities [70]. Although the natural substrate and product of GFC1/DCR in vivo is still unknown [34], it has an in vitro diacylglycerol acyltransferase activity [38]. GFC1/DCR may function in acyltransfer of cutin monomers to form precursor intermediates or oligomeric structures [37]. With free-OH group, tZ is likely to be a substrate of these acyltransferases. However, such hypothesis conflicts with the similar effects of tZ and other CKs without free-OH group on gfc1 and gpat4/8 (S3 Fig.). As a P450 monooxgenase, CYP86A2/ATT1 has been shown to be associated with the ωhydroxylation of FA [69] and the production of α,ω-dicarboxylic acid (DCA) [71]. Therefore, gfc1 phenotypes are produced by blocking the GPATs or/and GFC1/DCR catalyzed acyl-transfer but not by preventing the CYP86A2-mediated FA ω-hydroxylation steps in the cutin biosynthesis pathway.
Hydroxylation of FAs by CYP86As, transfer of the OH-FA to acylate G3P by GPATs, and probably their further acylation into cutin polymer by GFC1/DCR are closely linked in the core reaction of cutin biosynthesis [34], making it likely that cyp86A2/att1 has decreased OH-FAs levels and gfc1/dcr and gpat4/8 have increased OH-FAs. The involvement of OH-FAs as constituents of membrane lipids, and as biosynthetic precursors for biologically active compounds such as jasmonates, and the existence of OH-FA-dependent signaling in plant cells, have been noted previously in the literature [64].
Cutin biosynthesis is complex and its organization and regulation remain largely uncertain. It is unknown how gfc1 and gpat4/8 link lipid metabolism or signaling pathways, cutin-associated acyltransferase, and CK response and skotomorphogenesis. One possibility is that the accumulation of OH-FAs may perturb cell membranes and affect CK responses. More studies at the biochemical, molecular, physiological and genetic levels are needed to uncover the mechanisms underlying the roles of acyltransferases in CK responses and skotomorphogenesis.