Molecular characterization of a mutation affecting abscisic acid biosynthesis and consequently stomatal responses to humidity in an agriculturally important species

Mutants deficient in the phytohormone abscisic acid (ABA) have been instrumental physiological models for understanding both the biosynthesis and action of this critical hormone. The wilty mutant of the agriculturally important species Pisum sativum is a quintessential ABA deficient mutant which has remained molecularly uncharacterised in the 40 years since its discovery. We show that the wilty mutation affects the xanthoxin dehydrogenase step in the ABA biosynthetic pathway and that this step in ABA biosynthesis is critical for normal stomatal responses to changes in humidity.


Introduction
The phytohormone abscisic acid (ABA) is critical for land plant survival and is implicated in plant responses to water deficit. One of the earliest identified roles for ABA is closing the stomata of angiosperms (Mittelheuser and Van Steveninck 1969;Jones and Mansfield 1970). The signalling pathway for ABA-induced stomatal closure has been the subject of detailed investigation in recent years, through the use of single gene mutants, and is reasonably well understood (Geiger et al. 2011). Likewise, mutants deficient in ABA levels have been instrumental in revealing the pathway by which ABA is synthesized in plants (Nambara and Marion-Poll 2005;Taylor et al. 2005). The molecular characterization of mutants from diverse angiosperm species has provided strong evidence for the widely accepted, linear biosynthetic pathway for this hormone (Fig. 1).
Oxidative cleavage of 9-cis-neoxanthin and/or 9-cisviolaxanthin to xanthoxin is catalyzed by a 9-cisepoxycarotenoid dioxygenase (NCED), in the chloroplasts (Tan et al. 2001(Tan et al. , 2003. This is the rate-limiting step in ABA biosynthesis, and the first non-reversible step (Qin and Zeevaart 1999;Thompson et al. 2000). Of note, NCED mutants have been identified in Zea mays (vp14) (Schwartz et al. 1997), A. thaliana (nced3) (Iuchi et al. 2001;Ruggiero et al. 2004;Huang et al. 2008) and S. lycopersicon (notabilis) (Burbidge et al. 1999) and all have severe deficits in ABA, particularly under water stress. Once formed, xanthoxin moves across the membrane of the chloroplast to the cytosol and is converted to abscisic aldehyde by a short-chain alcohol dehydrogenase/reductase (SDR) (Fig. 1). Thus far, A. thaliana aba2 is the only mutant identified that affects this step (Cheng et al. 2002;Gonzá lez-Guzmá n et al. 2002). However, an O. sativa ortholog of ABA2 has been found to restore the ABA-deficient phenotype in A. thaliana aba2 mutants, suggesting that the function of this gene may be conserved across angiosperms (Endo et al. 2014). The final step in the ABA-biosynthetic pathway is the conversion of abscisic aldehyde to ABA by an abscisic aldehyde oxidase (AAO) gene (Fig. 1). Mutants of this final step are well represented in diverse species, including S. lycopersicon (sitiens) (Harrison et al. 2011), S. tuberosum (droopy) [based on synteny between the genomes of the two Solanum species (Duckham et al. 1989)] and A. thaliana (aao3) (Seo et al. 2000). The activity of AAO requires a molybdenum cofactor (MoCo) and mutants defective in the synthesis or maturation of this MoCo suffer from ABA deficiency in S. lycopersicon ( flacca) (Sagi et al. 2002), Hordeum vulgare (nar2a) (Walker-Simmons et al. 1989), N. plumbaginifolia (aba1 and cnxA) (Leydecker et al. 1995;Mendel and Schwarz 1999), A. thaliana (aba3 and other alleles) (Bittner et al. 2001;Xiong et al. 2001) and Z. mays (vp10 and vp15) Suzuki et al. 2006).
This collection of classical ABA-deficient mutants has also been critical for informing our understanding of the roles ABA plays in angiosperm physiology and survival, including driving stomatal closure, inducing seed dormancy and aiding desiccation and salinity tolerance. A quintessential ABA-deficient mutant used in physiological studies on Pisum sativum is wilty. wilty was identified from a spontaneous mutation in an unknown breeding line nearly 40 years ago (Marx 1976), but has not been previously characterized at a molecular level. The wilty mutant has provided insights into the physiological role of ABA in seed development (de Bruijn and Vreugdenhil 1992;de Bruijn et al. 1993b;Batge et al. 1999), plant growth (de Bruijn et al. 1993a) and stomatal behaviour (Donkin et al. 1983). wilty plants display a loss of turgor under conditions of moderate evaporative demand, and have particularly low ABA levels in seeds (10-fold less than wild-type plants) (Batge et al. 1999) and leaves, especially during drought-stress (Wang et al. 1984). Rapid reductions in the water status of the leaves of wilty compared with wild-type plants, during desiccation, have been well documented (Donkin et al. 1983;de Bruijn et al. 1993a). A single study has attempted to identify the biosynthetic step of ABA that is impaired in wilty plants, finding that wilty plants were able to rapidly synthesize ABA after the application of radiolabelled abscisic aldehyde (Duckham et al. 1989). This has led to the suggestion that the gene impaired in the wilty mutant is upstream of this step in the biosynthetic pathway and possibly due to a lesion in the NCED gene responsible for the conversion of 9-cis-neoxanthin or 9-cis-violaxanthin to xanthoxin . However, this has never been further examined.
In this study, we provide evidence regarding the molecular basis of the wilty mutation of P. sativum. The evidence strongly suggests that a mutation in the P. sativum ortholog of ABA2 underlies the P. sativum wilty mutant phenotype. We use this mutant to investigate the role of ABA biosynthesis in the rapid stomatal response to changes in humidity or vapour pressure deficit (VPD). There are mixed reports in the literature relating to the stomatal responses to VPD in ABA-biosynthetic mutants, although all work to date has focussed on A. thaliana and none have measured ABA levels (Assmann et al. 2000;Xie et al. 2006;Bauer et al. 2013;Merilo et al. 2015). Conclusions have ranged from ABAbiosynthetic mutants in A. thaliana having normal stomatal response to changes in VPD (Assmann et al. 2000) to ABA biosynthesis being critical for stomatal responses to such changes (Bauer et al. 2013). Recently, it has been shown that ABA levels in angiosperms can increase over a very short period (20 min) in response to a reduction in relative humidity, and that this increase in foliar ABA level is responsible for the closing of stomata, particularly in angiosperm herbs (McAdam and Brodribb 2015). Here we investigate whether plants that have an inability to synthesize ABA have functional stomatal response to changes in humidity.

Plant material
The original wilty line resulted from a spontaneous mutation that emerged in a breeding line belonging to L.G. Cruger from Del Monte Corporation, San Lenadro, CA, and was later sent to Prof. G.A. Marx in the early 1970s (Marx 1976). This original line, which has no known wild type, has been used in all physiological studies using wilty to date and is referred to in this study as L233. All physiological experiments here use line wil which was derived by introgression of L233 with the cultivar Torsdag over five backcrosses to create a near isogenic line. Eight lines of P. sativum were used to represent wild-type genotypes for sequencing of ABA2 in this study. These included JI281, WL1771, the argenteum mutant [also a spontaneous mutation that emerged in a breeding line belonging to L.G. Cruger, with an unknown background (Marx 1982)] and the cultivars Cameor, Champagne, Kaliski, Torsdag (Hobart Line 107) and Virtus. In addition to these diverse P. sativum lines, a single representative from the closely related genus Lathyrus odoratus [Hobart Line LO5, derived from the cultivar Grandiflora XD by Ross and Murfet (1985)] was also used.

Stomatal responses to changes in VPD
Plants were grown in a growth cabinet (PGC-105, Percival Scientific Inc.) in 2.6 L 14 × 17 cm slim-line pots in a 1 : 1 mix of vermiculite and dolerite gravel chips topped with 3 cm of an 8 : 1 mix of composted pine bark and course river sand. All plants were watered daily and received weekly applications of liquid fertilizer (Aquasol, Hortico Ltd). Conditions in the growth cabinet were regulated at 25 8C/16 8C day/night temperature and a 16 h photoperiod provided by mixed incandescent and fluorescent lights ensuring a minimum 300 mmol quanta m 22 s 21 at the pot surface. Before the experiment, a daytime VPD of 1.2 kPa (62 % relative humidity) was maintained while temperature and relative humidity were monitored every 5 min during this period by a data logger (HOBO Pro Series, Onset). Plants were initially grown at a VPD of 1.2 kPa to allow leaves to expand under relatively low humidity so as to avoid the potential stomatal dysfunction apparent in leaves that have expanded under very high humidity (Aliniaeifard et al. 2014).
When the plants were 3 weeks old, with approximately seven fully expanded leaves, the VPD in the cabinet was lowered to 0.7 kPa (+0.05 kPa) by the presence of containers of water and a 1 m 2 surface of wet hessian. After a 5-day acclimation period to this low VPD the simultaneous monitoring of leaf gas exchange and ABA levels were undertaken on leaves from seven individuals of wil and Torsdag (as described below). After this initial simultaneous measurement of leaf gas exchange and ABA level, VPD was increased to 1.5 kPa (42 % relative humidity) using a condensing dehumidifier (SeccoUltra 00563, Olimpia-Splendid) in the growth cabinet. Temperature and relative humidity were monitored every 30 s during the experimental period by a humidity probe (HMP45AC, Vaisala) and thermocouple connected to a data logger (CR10X, Campbell Scientific). A VPD of 1.5 kPa was maintained for 20 min after which leaf gas exchange and ABA level were again simultaneously measured in leaves from three individuals. Vapour pressure deficit was then returned to 0.7 kPa and measurements were again conducted after a further 20 min. The relatively small and contained volume of air in the growth cabinet (3 m 3 ) resulted in a relatively fast half time for the VPD transition of 150 s following transitions between 0.7 and 1.5 kPa (with no hysteresis when VPD was returned to 0.7 kPa).

Leaf gas exchange measurements
During transitions in VPD, leaf gas exchange was measured in fully irradiated (300 mmol quanta m 22 s 21 ) leaves using an infrared gas analyser (LI-6400, LI-COR AoB PLANTS www.aobplants.oxfordjournals.org Biosciences). Conditions in the leaf cuvette were maintained as close as possible to the conditions in the growth cabinet, with VPD regulated by a portable dew point generator (LI-610, LI-COR Biosciences). Leaves were enclosed in the cuvette and instantaneous gas exchange was logged following stability in cuvette conditions (after 30 s). Following gas exchange measurements, the same leaf was then excised and immediately sampled for ABA quantification (see below).

ABA purification and quantification
Samples harvested for ABA quantification were immediately weighed (+0.0001 g, MS204S, Mettler-Toledo) into 50 mL tubes, covered in 15 mL of cold (220 8C) 80 % methanol in water (v v 21 ) with 250 mg L 21 (m v 21 ) of added butylated hydroxytoluene and transferred to 220 8C. Foliar ABA levels in these samples were extracted, purified and quantified by physicochemical methods using an added internal standard and UPLC-MS according to the methods of McAdam and Brodribb (2014).
To investigate the evolution of ABA2 and the phylogenetic placement of PsABA2, amino acid sequences of SDRs from the SDR110C clade [sequence details in Moummou et al. (2012)] were assembled from 11 species spanning the diversity of the land plant phylogeny including A. thaliana, Populus trichocarpa, Glycine max, Vitis vinifera, O. sativa, Z. mays, Sorghum bicolor, Selaginella moellendorffii and Physcomitrella patens [sequences as per Moummou et al. (2012)], in addition to predicted protein sequences for SDR110C genes identified by reciprocal BLAST searches in Picea abies (http://congenie.org) and Amborella trichopoda (http://phytozome.jgi.doe.gov) for all known representatives). Amino acid sequences were aligned using ClustalX and distance and parsimonybased methods were used for phylogenetic analyses in PAUP version 4.0b10 (http://paup.csit.fsu.edu/).

Statistical analysis
The effect of changing VPD on g s and foliar ABA level of both wild-type and wilty mutant plants was tested by twoway analysis of variance (ANOVA). Pair-wise comparisons between the means of g s and foliar ABA level for wild-type and mutant plants over the VPD transition were carried out using one-way ANOVAs followed by a Tukey's test. Analyses were performed using R Statistical Software.

Results
The wilty mutant carries a lesion in PsABA2 The phenotype of the wilty mutant implies that a gene involved in the ABA biosynthesis pathway may be compromised in this mutant. In previous studies, close synteny between the genomes of P. sativum and Medicago (Kaló et al. 2004) has enabled the use of publicly available M. truncatula sequence to select specific candidate genes for P. sativum mutant loci using a comparative mapping strategy (Hecht et al. 2005;Zhu et al. 2005). Strong linkage between wilty and the morphological markers st and b was detected in the earliest studies of the wilty mutant, indicating that the WILTY locus is located on P. sativum linkage group III (Marx 1976). A recent consensus map for P. sativum, which included tentative locations for known mutations, suggested that st may be located close to the marker gene Pip1 (Bordat et al. 2011). We examined the corresponding region of the M. truncatula genome, searching for any genes known to be associated with ABA biosynthesis close to the M. truncatula Pip1 orthologue (MtPip1; Medtr3g070210; v4.0, http://phytozome.jgi. doe.gov/) on M. truncatula chromosome 3. We identified a homologue of ABA2 (MtABA2; Medtr3g020670) and further investigated the corresponding gene PsABA2 as a candidate for WILTY.
Sequencing of PsABA2 in eight diverse wild-type P. sativum lines, including a line containing the original wilty allele [L233 (Marx 1976)] and a near isogenic line derived by introgression of L233 with the cultivar Torsdag (wil), as well as the closely related legume L. odoratus, revealed a nucleotide change (CGT to GA) that causes a frameshift in codon 163 in wilty mutant plants relative to the wild-type P. sativum lines and L. odoratus (Fig. 2). This change, observed in both gDNA and cDNA, causes a premature stop after 11 missense amino acids, resulting in the loss of the highly conserved SDR-catalytic domain defined by the Tyr-XX-Ser-Lys motif ( Fig. 2; Supporting Information- Fig. S1). In A. thaliana, mutation to this key domain results in the loss of ABA2 function in the aba2-13 mutant (González-Guzmán et al. 2002). We predict that the mutation seen in the P. sativum wilty mutant results in loss of ABA2 function in a similar manner. The mutation in the PsABA2 gene was carried by both the original L233 and the wil line, the result of introgression over five generations with the unrelated wild-type line Torsdag.
Phylogenetic analysis showed that the PsABA2 gene is embedded in a uniquely angiosperm clade of SDRs that includes the ABA2 genes responsible for ABA biosynthesis in A. thaliana and O. sativa (Fig. 3). Intriguingly, this clade, dedicated to ABA biosynthesis, is only found in angiosperms. This clade is not represented in the conifer P. abies, lycophyte S. moellendorfii or moss P. patens, and the most basal genes in the ABA2 clade are from the most basal angiosperm, A. trichopoda (Fig. 3). The SDR genes in non-flowering plants were no more closely related to the ABA2 clade than to any of the other diverse clades of genes in the SDR110C family (Fig. 3).

Dysfunctional stomatal response to changes in humidity in the wilty mutant
The stomata of wilty mutant plants do not have the characteristic rapid response to changes in VPD observed in wild-type P. sativum ( Fig. 4; Table 1). In wild-type plants 20 min after a doubling in VPD (from 0.7 to 1.5 kPa), stomata closed significantly from an initial g s of 0.587 -0.2 mol m 22 s 21 (F 1, 4 ¼ 31.03, P ¼ 0.0051; Fig. 4). This reduction in g s in wild-type plants was accompanied by a significant 10-fold increase in foliar ABA levels (F 1, 16 ¼ 15.46, P ¼ 0.0012; Fig. 4). Significantly higher foliar ABA levels in wild-type plants did not noticeably fall after 20 min on returning to a VPD of 0.7 kPa, resulting in pronounced hysteresis in the recovery of g s (Fig. 4). In contrast to wild-type, the stomata of wilty plants did not close in response to a doubling in VPD and neither did foliar ABA levels significantly increase, remaining ,5 ng g 21 fresh weight over the transition in VPD ( Fig. 4; Table 2). Both g s and foliar ABA levels in wild-type and wilty plants at the start of the experiment were not significantly affected by genotype (g s F 1, 4 ¼ 0.27, P ¼ 0.62; foliar ABA F 1, 16 ¼ 2.43, P ¼ 0.14; Fig. 4).

Discussion
Here we present evidence that the classical wilty mutant of P. sativum is due to a null mutation in the key ABAbiosynthetic gene, ABA2, which encodes a xanthoxin dehydrogenase (an SDR), responsible for the conversion of xanthoxin to abscisic aldehyde (Fig. 2). To summarize (i) both wilty lines in our study show severe deficiency in ABA biosynthesis, having very low levels in most tissues (Fig. 4) (de Bruijn et al. 1993b;Batge et al. 1999); (ii) the ABA2 gene of P. sativum maps to the same chromosomal region as the wilty phenotype, at the base of linkage group III (Marx 1976); (iii) this ABA2 gene of P. sativum is the only ABA2 homologue in this species (Fig. 3); (iv) both of the wilty lines tested have a mutation in the key SDR-catalytic domain of the ABA2 gene (Fig. 2), similar to the aba2-13 mutant of A. thaliana and (v) the association between the wilty phenotype and this mutation in the ABA2 gene has persisted through five generations of introgression with the wild-type line Torsdag. To date, the key SDR-mediated step in ABA biosynthesis has been represented only by mutants from a single species, A. thaliana (Fig. 1). Our finding that wilty P. sativum lines  Fig. S1). Black bars represent amino acids that are conserved between wild-type P. sativum and A. thaliana ABA2 proteins; dark grey bars represent differences between these species; light grey bars represent amino acids unique to the wilty mutant. Key functional domains are shown in boxes including the cofactor binding domain (green) and the SDR-catalytic domain (blue) according to Joernvall et al. (1995).
AoB PLANTS www.aobplants.oxfordjournals.org & The Authors 2015 carry a severe mutation in ABA2 strongly suggests that disruption to ABA2 function underlies the wilty mutant phenotype, resolving the unknown identity of this mutant. Taylor et al. (2000), suggested on the basis of Duckham et al. (1989), that the wilty mutant phenotype could be the result of a mutation in the NCED gene, responsible for the carotenoid cleavage step in ABA biosynthesis and the formation of xanthoxin (Fig. 1). However, the experiment of Duckham et al. (1989), which involved the feeding of abscisic aldehyde, the penultimate precursor in ABA biosynthesis, to wilty plants, could not rule out the possibility of wilty being a mutant for ABA2, as the formation of ABA following the feeding of abscisic aldehyde would be possible in mutants of both ABA2 and NCED (Fig. 1). To rule out the unlikely possibility that another gene closely linked to PsABA2 may be contributing to the wilty mutant phenotype, a future study could confirm restoration of phenotype in this mutant using wild-type PsABA2.
The most obvious feature of the wilty mutant phenotype is a rapid loss of turgor on exposure to conditions of increased evaporation (Wang et al. 1984). We show that this is because ABA levels increase in wild-type plants in response to low humidity, closing stomata, but Bootstrap values from 1000 trees are shown next to each branch. Naming of angiosperm-specific clades follows that of Moummou et al. (2012). Full reference to the sequences compressed in each of the clusters is given in Supporting Information-Fig. S2 and Table S1. not in the wilty mutant (Fig. 4). Mutants of ABA2 in A. thaliana have a similar lack of stomatal closure under increased evaporative demand, with mutants of this gene identified as readily wilting (Léon-Kloosterziel et al. 1996). Mutants of the ABA1, ABA2 and ABA3 genes of A. thaliana display a varying degree of stomatal response to increased VPD (Assmann et al. 2000;Xie et al. 2006;Merilo et al. 2015), although these reports are not always consistent (compare the results of Bauer et al. 2013 with Merilo et al. 2015). However, as none of these studies measured ABA levels we cannot rule out the possibility that the A. thaliana biosynthetic mutants, all of which are known to have a degree of redundancy, had slight increases in foliar ABA levels at high VPD.
In addition to an obvious wilting phenotype, aba2 mutants in A. thaliana also show reduced growth, which is presumably due to factors other than increased water stress, being observed even when plants are grown under very humid conditions (Cheng et al. 2002). In contrast, wilty mutant plants in P. sativum do not appear to have any obvious deficiency in shoot growth, but do exhibit reduced root growth (de Bruijn et al. 1993a). Whether this reduction in root growth in wilty plants is because ABA normally promotes root growth in wild-type plants, or is the result of increased water stress and lower water potentials in the mutant, is yet to be investigated.
Recently we have shown that stomatal responses to changes in humidity or VPD in angiosperms are driven by functionally significant changes in foliar ABA level; this is particularly evident in angiosperm herbs (McAdam and Brodribb 2015). Previously, some have argued that ABA responsible for the rapid (,1 h) responses of stomata to changes in water status does not have an origin in de novo synthesis but rather release from internal stores in the chloroplasts (Georgopoulou and Milborrow 2012) or via a single step that converts the catabolite ABA -glucose ester to ABA (Lee et al. 2006). Two findings in our study support an alternative view to these two hypotheses. The first is that our solvent-based extraction method for ABA does not discriminate between ABA that is found in the cytosol or bound in the chloroplasts, so any increase observed in the level of ABA in the leaf cannot be due to the redistribution of ABA between compartments of the leaf. The second is that the lack of a stomatal response to the step change in VPD in the wilty mutant plants was due to a lack of increase in the foliar ABA level. These observations suggest that the rapid (20 min), 10-fold increase in foliar ABA level in response to a doubling of VPD in P. sativum is likely the result of de novo synthesis. However, further work is required to investigate Figure 4. The mean response of stomatal conductance (n ¼ 9 individuals, +SE) and foliar ABA level (n ¼ 9 individuals, +SE) to a reversible sequence of VPD transitions from 0.7 to 1.5 kPa and returning to 0.7 kPa, with each transition lasting 20 min in wild-type (black circles) and wilty mutant (white circles) plants. Different letters denote significant difference between means (P , 0.05, Oneway ANOVA followed by Tukey's test).  the importance of de novo biosynthesis in the rapid accumulation of ABA in response to a change in VPD. An interesting observation emerging from the phylogenetic analysis of the SDR110C clade, which includes ABA2 as well as related genes, is that a dedicated ABAbiosynthetic SDR (namely ABA2) evolved in the earliest angiosperms, including A. trichopoda (Fig. 3). This is surprising given that all clades of land plants are known to synthesize ABA (Johri 2008). Indeed, other studies investigating the evolution of ABA-biosynthetic genes have found that the genomes of the moss P. patens and the lycophyte S. moellendorfii also lack a specific AAO gene responsible for the final step in the biosynthetic pathway, the conversion of abscisic aldehyde to ABA (Hanada et al. 2011). There is, however, evidence that the conversion of xanthoxin to abscisic aldehyde can be catalyzed, to a very limited degree, by other SDRs in angiosperms. The null mutant of Arabidopsis aba2-11, for example, is still able to synthesize a small quantity of ABA (Gonzá lez-Guzmá n et al. 2002). Indeed, unstressed wilty mutants grown under high humidity have similar foliar ABA levels compared with wild-type plants (Fig. 4). Also, in wilty mutants ABA levels begin to increase after 5 days of drought-stress, albeit to levels only one-third of those observed in wildtype plants (Wang et al. 1984). If the currently accepted biosynthetic pathway for ABA is conserved across land plants then the conversion of xanthoxin to abscisic aldehyde in basal clades of land plants is likely catalyzed by non-specific SDRs. This, as well as the lack of a specific AAO gene (Hanada et al. 2011), suggests that the final two steps in the ABA-biosynthetic pathway may be the ratelimiting steps for ABA biosynthesis in the basal lineages of land plants. This would stand in contrast to the hypothesized rate-limiting process for ABA biosynthesis in angiosperms, the carotenoid cleavage step leading to the formation of xanthoxin (Qin and Zeevaart 1999). Several observations indicating slow or delayed increases in foliar ABA level in these basal lineages support this possibility, including: (i) foliar ABA levels in ferns and lycophytes do not increase until plants approach a lethal water stress (McAdam and Brodribb 2013); (ii) foliar ABA levels in seedless vascular plants and conifers do not increase on exposure to increased VPD (McAdam and Brodribb 2015) and (iii) foliar ABA levels do not increase in the conifer Metasequoia glyptostroboides unless maintained for at least 6 h beyond turgor loss point (McAdam and Brodribb 2014), which is unlike ABA biosynthesis rates in angiosperms (Pierce and Raschke 1981). It could be argued that a single rate-limiting step in ABA biosynthesis, as occurs in angiosperms, would enable rapid and dynamic regulation of ABA biosynthesis. This would make it much easier for angiosperms to control a stomatal response to VPD through changes in the foliar ABA level. To date, however, there has been inadequate monitoring of ABA-biosynthetic gene expression during a VPD transition to confirm the single rate-limiting step hypothesis. It could be possible, in fact, that angiosperms dynamically alter the expression of multiple genes in the ABA-biosynthetic pathway in response to VPD (Thompson et al. 2007) but further work is required to investigate this scenario.

Conclusions
Here we show that the classical ABA-biosynthetic mutant wilty in P. sativum contains a mutation in the ABAbiosynthetic gene ABA2. wilty mutant plants show no increase in foliar ABA levels during a whole-plant transition in VPD, and as a result do not close stomata. This lack of a stomatal response to changes in VPD strongly implicates ABA biosynthesis in the stomatal responses to changes in VPD in angiosperms. Phylogenetic analysis of the ABA2 gene across land plants shows that the evolution of an ABA-biosynthetic-specific short-chain dehydrogenase (ABA2) evolved in the earliest angiosperms. This lack of specificity in the later-stage ABA-biosynthetic genes may explain why lycophytes, ferns and conifers, unlike angiosperms, do not show a rapid increase in foliar ABA levels on exposure to high VPD (McAdam and Brodribb 2015).

Sources of Funding
Our work was funded by the Australian Research Council (DE140100946).

Supporting Information
The following additional information is available in the online version of this article - Figure S1. Amino acid sequence alignment of the ABA2 protein from A. thaliana, L. odoratus and eight diverse accessions and cultivars of P. sativum and the two lines carrying the wilty mutation. Figure S2. Phylogenetic tree of the SDR110C family of land plant SDRs. Table S1. Accession details for SDR110C family used to construct Figure 3 and Supporting Information- Fig. S2.