Knockout of Arabidopsis thaliana VEP1, Encoding a PRISE (Progesterone 5β-Reductase/Iridoid Synthase-Like Enzyme), Leads to Metabolic Changes in Response to Exogenous Methyl Vinyl Ketone (MVK)

Small or specialized natural products (SNAPs) produced by plants vary greatly in structure and function, leading to selective advantages during evolution. With a limited number of genes available, a high promiscuity of the enzymes involved allows the generation of a broad range of SNAPs in complex metabolic networks. Comparative metabolic studies may help to understand why—or why not—certain SNAPs are produced in plants. Here, we used the wound-induced, vein patterning regulating VEP1 (AtStR1, At4g24220) and its paralogue gene on locus At5g58750 (AtStR2) from Arabidopsis to study this issue. The enzymes encoded by VEP1-like genes were clustered under the term PRISEs (progesterone 5β-reductase/iridoid synthase-like enzymes) as it was previously demonstrated that they are involved in cardenolide and/or iridoid biosynthesis in other plants. In order to further understand the general role of PRISEs and to detect additional more “accidental” roles we herein characterized A. thaliana steroid reductase 1 (AtStR1) and compared it to A. thaliana steroid reductase 2 (AtStR2). We used A. thaliana Col-0 wildtype plants as well as VEP1 knockout mutants and VEP1 knockout mutants overexpressing either AtStR1 or AtStR2 to investigate the effects on vein patterning and on the stress response after treatment with methyl vinyl ketone (MVK). Our results added evidence to the assumption that AtStR1 and AtStR2, as well as PRISEs in general, play specific roles in stress and defense situations and may be responsible for sudden metabolic shifts.


Introduction
Specialized metabolism, sometimes still termed "secondary metabolism", has been studied extensively in plants and microorganisms. In the twentieth century, the vast chemical diversity of the natural products formed by plants attracted organic chemists even more than plant biologists. A lot of the complex natural products appearing in certain taxa during different developmental stages were labelled as "secondary" metabolites because a fundamental role in cell metabolism and growth was not seen. Today, several small (or specialized) natural plant products (SNAPs), such as jasmonates, brassinosteroids, and strigolactones have lost their stigma of being secondary metabolites. They are now regarded as important regulatory elements of plant development and growth. In some instances, it is difficult to determine which metabolite is "secondary" and which is "primary", particularly when they share a common biosynthetic route and the same structural scaffold. Examples are psilocybin (a toxic alkaloid) and indole-3-acetic acid (a phytohormone), both derived MVK is a powerful inducer of defense-related gene expression [21] and PRISEs may therefore modify stress responses through metabolizing MVK and related compounds. Consequently, a more general physiological role for PRISEs has been proposed in stress responses and detoxification processes [9,14].
In order to investigate the general role of PRISEs and to elucidate their "accidental" roles we characterized AtStR1 and compared it to AtStR2. We used A. thaliana Col-0 wildtype (wt) plants as well as AtStR1 knockout mutants (AtStR1-) and knockout mutants overexpressing either AtStR1 (35S:AtStR1/AtStR1-) or AtStR2 (35S:AtStR2/AtStR1-) to investigate the effects of MVK treatment.

Characterization of AtStR2
AtStR2 is encoded by locus At5g58750. An appropriate cDNA was deduced and cloned into a pDEST17 expression vector and then functionally expressed in E. coli. The recombinant enzyme was termed rAtStR2. It was purified ( Figure S1) and then used for enzyme characterization. We demonstrated that rAtStR2 converted progesterone enantioselectively into 5β-pregnane-3,20-dione ( Figure S2). Progesterone was converted equally well by rAtStR2 and rAtStR1, but AtStR1 converted MVK much better than rAtStR2 (Table 1). AtStR1 and AtStR2 were aligned together with other members of the PRISE family using ClustalW [22] and all PRISE motifs described by [23,24] could be identified ( Figure 2). MVK is a powerful inducer of defense-related gene expression [21] and PRISEs may therefore modify stress responses through metabolizing MVK and related compounds. Consequently, a more general physiological role for PRISEs has been proposed in stress responses and detoxification processes [9,14].
In order to investigate the general role of PRISEs and to elucidate their "accidental" roles we characterized AtStR1 and compared it to AtStR2. We used A. thaliana Col-0 wildtype (wt) plants as well as AtStR1 knockout mutants (AtStR1-) and knockout mutants overexpressing either AtStR1 (35S:AtStR1/AtStR1-) or AtStR2 (35S:AtStR2/AtStR1-) to investigate the effects of MVK treatment.

Characterization of AtStR2
AtStR2 is encoded by locus At5g58750. An appropriate cDNA was deduced and cloned into a pDEST17 expression vector and then functionally expressed in E. coli. The recombinant enzyme was termed rAtStR2. It was purified ( Figure S1) and then used for enzyme characterization. We demonstrated that rAtStR2 converted progesterone enantioselectively into 5β-pregnane-3,20-dione ( Figure S2). Progesterone was converted equally well by rAtStR2 and rAtStR1, but AtStR1 converted MVK much better than rAtStR2 (Table 1). AtStR1 and AtStR2 were aligned together with other members of the PRISE family using ClustalW [22] and all PRISE motifs described by [23,24] could be identified ( Figure 2).
Motifs I to III are typically present in all NADP + -dependent SDRs, the protein family to which the PRISEs belong. NADPH 2 but not NADH 2 was accepted as a cosubstrate by AtStR2 (Table 1). In all PRISEs investigated so far, NADPH 2 was the preferred cosubstrate. Cosubstrate specificity in AtSt5βR muteins was previously changed with limited success modifying the typical NADPH 2 -binding motif GVARR [25].
AtStR2 shares 92% amino acid identity with a PRISE from Erysimum crepidifolium (AGT29343.1). This high interspecific similarity and the much lower intraspecific similarity (less than 40% identity between AtStR1 and AtStR2, as well as EcP5βR1 and EcP5βR2, respectively) led to the assumption that at least two clusters (cluster I and cluster II) of PRISEs needed to be differentiated [7]. Only two PRISE genes were identified in the genome of A. thaliana, whereas up to seven VEP1-like genes were reported in other plant genomes [9]. Examples of cluster I and cluster II PRISEs are shown in Table 1. When the substrate spectra of AtStR1 and other cluster I PRISEs were studied, it was demonstrated that these enzymes converted small RES more efficiently [7,9,19]. Methyl vinyl ketone (MVK) belongs to this group of compounds. When calculating the progesterone/MVK quotient (Q P/M ) of the catalytic efficiencies shown in Table 1, it became obvious that PRISEs of cluster I usually prefer MVK over progesterone as a substrate, indicated by a smaller  [23] and [24] are shown in turquoise shading. The residues of the phenylalanine clamp defined by [10] are depicted in bold. Asterisk, fully conserved; colon, strongly conserved; period, weakly conserved.
This was achieved by the removal of the phenylalanines forming a clamp structure at the rim of the binding pocket [9]. With the disruption of this feature, small, rather hydrophilic 1,4-enones may not be "trapped" in the binding pocket long enough to allow for the conversion [9,10]. Since the phenylalanine clamp is highly conserved and found in most of the class I PRISEs described so far, it seems as if these enzymes were not optimized for larger molecules such as progesterone [9].
To analyze this in more detail, we here modeled AtStR2 on the crystal structure of AtStR1 (PDB entry 6el3.1.A) cocrystallized with NADP + [9], which is available in the RSCB protein data bank [28]. The model was created using Swissmodel [29] and progesterone was docked into the active site ( Figure 3) with AutoDock Vina [30]. The results were visualized in USCF Chimera [31] and demonstrated a different fit of progesterone in the active site of AtStR1 and AtStR2, respectively. visualized in USCF Chimera [31] and demonstrated a different fit of progesterone in the active site of AtStR1 and AtStR2, respectively.
In AtStR2, as well as in other cluster II PRISEs, the phenylalanine adjacent to motif 5 was substituted for other amino acids, hence a phenylalanine clamp could not be formed. As a consequence, the active site could accommodate larger molecules such as progesterone ( Figure 3). . The coulombic surface is displayed for important residues in the active site according to [32]. NADP + (red), progesterone (yellow for AtStR1, blue for AtStR2) and the phenylalanine clamp residues (green for AtStR1, grey for AtStR2) are shown as stick models.
The loss of activity regarding small molecules, which cannot be trapped in the active site has already been described [10]. Their interpretation of the important role of a phenylalanine clamp at the rim of the binding pocket is consistent with the observation that the cluster II PRISEs we analyzed here ( Figure 2) lack at least one of the two phenylalanines. This explains their lower catalytic efficiency for MVK. With this in mind, the rAtStR1 mutants rAtSt5βR1_F153A and rAtSt5βR1_F153A_F342A [10] could be regarded as cluster I PRISEs converted to cluster II PRISEs. The substrate preferences of rAtStR1, and its mutants have already been characterized quite well [10,18,19] whereas more data are still required for rAtStR2.
Paralogues obtained after gene duplication may have been formed in metaboliteenzyme coevolution [2] towards the preference of larger molecules and the discrimination of small RES released in stress reactions. Other substrate-promiscuous enzymes such as BAHD-type malonyltransferases may also have morphed from enzymes capable of detoxifying xenobiotics to enzymes capable of malonylating 21-hydroxypregnanes, yielding precursors of butenolide ring formation in the cardenolide biosynthesis [33]. In a similar way, PRISEs may have evolved helping to establish cardenolide formation in E. crepidifolium, a close relative of A. thaliana. Figure 3. Docking of progesterone into the active site of AtStR1 (a) and homology-modelled AtStR2 (b). The coulombic surface is displayed for important residues in the active site according to [32]. NADP + (red), progesterone (yellow for AtStR1, blue for AtStR2) and the phenylalanine clamp residues (green for AtStR1, grey for AtStR2) are shown as stick models.
In AtStR2, as well as in other cluster II PRISEs, the phenylalanine adjacent to motif 5 was substituted for other amino acids, hence a phenylalanine clamp could not be formed. As a consequence, the active site could accommodate larger molecules such as progesterone ( Figure 3).
The loss of activity regarding small molecules, which cannot be trapped in the active site has already been described [10]. Their interpretation of the important role of a phenylalanine clamp at the rim of the binding pocket is consistent with the observation that the cluster II PRISEs we analyzed here ( Figure 2) lack at least one of the two phenylalanines. This explains their lower catalytic efficiency for MVK. With this in mind, the rAtStR1 mutants rAtSt5βR1_F153A and rAtSt5βR1_F153A_F342A [10] could be regarded as cluster I PRISEs converted to cluster II PRISEs. The substrate preferences of rAtStR1, and its mutants have already been characterized quite well [10,18,19] whereas more data are still required for rAtStR2.
Paralogues obtained after gene duplication may have been formed in metaboliteenzyme coevolution [2] towards the preference of larger molecules and the discrimination of small RES released in stress reactions. Other substrate-promiscuous enzymes such as BAHD-type malonyltransferases may also have morphed from enzymes capable of detoxifying xenobiotics to enzymes capable of malonylating 21-hydroxypregnanes, yielding precursors of butenolide ring formation in the cardenolide biosynthesis [33]. In a similar way, PRISEs may have evolved helping to establish cardenolide formation in E. crepidifolium, a close relative of A. thaliana.
AtStR1 was expressed in Ath Col-0 wt but not in AtStR1mutant plants, verifying the stable knockout of AtStR1 in the homozygous plants used here ( Figure 4). AtStR2 was strongly expressed in seedlings ( Figure 4b) and cold-treated adult leaves ( Figure S3) but poorly in adult leaves of non-stressed A. thaliana plants (Arabidosis eFP browser 2.0; Figures 4a and S3). Since AtStR2 is not expressed in leaves (in normal conditions), AtStR2 activity did not interfere with the determination of AtStR1 activities in the enzyme assays of leaf extracts ( Figure S4). Other enzymes possibly converting progesterone to 5β-pregnane-3,20-dione seemed to be inactive ( Figure S4). stable knockout of AtStR1 in the homozygous plants used here ( Figure 4). AtStR2 was strongly expressed in seedlings ( Figure 4b) and cold-treated adult leaves ( Figure S3) but poorly in adult leaves of non-stressed A. thaliana plants (Arabidosis eFP browser 2.0; Figures 4a and S3). Since AtStR2 is not expressed in leaves (in normal conditions), AtStR2 activity did not interfere with the determination of AtStR1 activities in the enzyme assays of leaf extracts ( Figure S4). Other enzymes possibly converting progesterone to 5βpregnane-3,20-dione seemed to be inactive ( Figure S4).  Insertion lines in which AtStR2 on locus At5g58750 was inactivated were not available to us. Attempts to generate plants lacking the AtStR2-encoding gene using either RNAi-mediated gene knockdown or CRISPR/Cas9-mediated gene knockout have been unsuccessful. It is possible that AtStR2 might be essential for plant development and a lack of it stops seed germination or seedling development.
Progesterone 5β-reductase activity was recovered in the 35S:AtStR1/AtStR1as well as in 35S:AtStR2/AtStR1lines, indicating that the respective enzymes were produced in an active form ( Figure S4). Insertion lines in which AtStR2 on locus At5g58750 was inactivated were not available to us. Attempts to generate plants lacking the AtStR2-encoding gene using either RNAi-mediated gene knockdown or CRISPR/Cas9-mediated gene knockout have been unsuccessful. It is possible that AtStR2 might be essential for plant development and a We investigated the effects of MVK treatment on Ath Col-0 wt and mutant plants, because MVK is a powerful inducer of defense-related gene expression [21] and PRISEs could be involved in the stress response simply by metabolizing MVK itself. We analyzed surrogate parameters for the stress response, such as H 2 O 2 levels and glutathione reductase (GR1) expression. Glutathione reductase (GR1) is an enzyme influencing the redox state of a cell by releasing reduced glutathione (GSH) from oxidized glutathione (GSSG) [34]. Redox states can therefore be evaluated by measuring GR1 expression [14].
AtStR2 was not induced under MVK stress and rAtStR2 only poorly accepted MVK (Table 1). Moreover, the redox state as estimated by GR1 expression and H 2 O 2 production was not altered in AtStR1mutants transformed with 35S:AtStR2 ( Figure 6). The A. thaliana glutathione reductase gene GR1 is expressed twice as high in AtStR1mutants than in Ath Col-0 wt. H 2 O 2 levels were also higher in AtStR1mutants than in Ath Col-0 wt ( Figure 6). This indicates reduced GSH availability. The effect could be reduced to wild type levels after curing the AtStR1mutant with 35S:AtStR1 but not with 35S:AtStR2. A disturbed redox state, indicated by a changed expression of a cytosolic glutathione reductase and increased H 2 O 2 levels, was also seen in D. lanata RNAi shoot culture lines in which VEP1like genes were knocked down. In these mutants, four times higher GSH levels compared to wild type controls have been demonstrated [14].

Roles of AtStR1 and AtStR2 in Plant Reactive Electrophilic Species (RES) Stress
Cardenolide and/or iridoid pathways are not known to be present in A. thaliana. Iridoids are not known to occur in the Brassicaceae and only few genera, including Erysimum, accumulate cardenolides. AtStRs cannot be involved in pathways that are not existent in A. thaliana. Consequently, though capable of reducing the double bond of progesterone enantioselectively, here PRISEs simply cannot be connected to 5β-cardenolide formation. Enzymes as well as pregnane precursors may, however, be part of a "silent metabolism" in which putative intermediates and enzymes are present that have not yet been detected or are not associated with known metabolic pathways [1,35]. Studies in comparative metabolism (including metabolomics techniques) may help to better understand and evaluate these "silent" resources.

Roles of AtStR1 and AtStR2 in Plant Reactive Electrophilic Species (RES) Stress
Cardenolide and/or iridoid pathways are not known to be present in A. thaliana. Iridoids are not known to occur in the Brassicaceae and only few genera, including Erysimum, accumulate cardenolides. AtStRs cannot be involved in pathways that are not existent in A. thaliana. Consequently, though capable of reducing the double bond of progesterone enantioselectively, here PRISEs simply cannot be connected to 5βcardenolide formation. Enzymes as well as pregnane precursors may, however, be part of Small reactive carbonyl species, also known as reactive electrophilic species (RES) formed during stress reactions, have already been identified as putative substrates for PRISEs. Examples are small 1,4-enones such as methyl vinyl ketone (MVK) or 1-cyclohexen-2-one which are converted by AtStR1 very efficiently [7,9,10,19]. MVK is a product of trienoic fatty acid peroxidation [20]. MVK accumulates in various stress situations [10,[36][37][38] and is partly responsible for heat stress symptoms in cyclamen [20]. It is known that MVK causes tissue damage in A. thaliana [38] and Cyclamen persicum [20] leaves. The enzymatic reduction of MVK yields methyl ethyl ketone (MEK) which is harmless to plants [38]. There are other enzymes and mechanisms known to detoxify or inactivate RES. For example, RES can be detoxified by spontaneous reaction or enzymatic conjugation with reduced glutathione (GSH) [39]. RES, specifically 1,4-enones can also be reduced by aldehyde oxidoreductases (AOR) and 2-alkenal reductases (AER). We here considered these options as alternative routes for MVK detoxification and analyzed the expression pattern of the respective A. thaliana genes.
VEP1, AER and AOR are differentially expressed in A. thaliana. In the absence of stress, AER is expressed only weakly, whereas AOR is strongly expressed in young and VEP1 in mature leaves (Arabidopsis eFP browser 2.0). The respective enzymes differ in their substrate preferences, especially the size of potential substrates. For example, an AOR of Cucumis sativa reduced 4-hexen-3-one about 100 times more efficiently than MVK [27]. On the other hand, the AtAOR characterized in the same paper was equally active towards these two substrates. The authors discussed that AOR together with other enzymes might cooperatively detoxify reactive carbonyls. We here provide evidence that VEP1-encoded PRISEs may also be important players in this scenario.
AtAER (At5g16970) encodes a cytosolic alkenal reductase in A. thaliana. As far as the kinetic constants are concerned AtStRs and AtAER have similar K m values but differ in their catalytic efficiency, AtAER being about 200 times more efficient than AtStR1. Expression of AtAER and AtAOR was equally high in Ath Col-0 wt and in AtStR1mutants and could be amplified by MVK, in this way paralleling enhanced AtStR1 expression (Figure 7). Since AtAER is only weakly expressed in leaves and is enhanced moderately only under MVK stress, the strong expression of AtStR1 may lead to an amount of enzyme that can efficiently compete with AtAER for MVK detoxification. AtAOR (At1g23740), in contrast to AtAER, is strongly and constitutively expressed in the chloroplast and is assumed to be involved in the general elimination of reactive carbonyls [27]. All in all, detoxification of reactive carbonyls seems to be a more complex situation with correlating enzymes working in a group and providing a detoxifying network system. However, this issue requires further investigation. A 3 h exposure to MVK resulted in a fourfold increase of AtStR1 expression in Ath Col-0 wt (Figure 7). Expression stayed high even after 24 h of treatment (data not shown). On the other hand, AtStR2 was neither expressed in the control plants nor was its expression induced after MVK treatment, even after prolonged exposure up to 24 h.
Metabolites 2022, 12, 11 10 of 18 enzymes might cooperatively detoxify reactive carbonyls. We here provide evidence that VEP1-encoded PRISEs may also be important players in this scenario. AtAER (At5g16970) encodes a cytosolic alkenal reductase in A. thaliana. As far as the kinetic constants are concerned AtStRs and AtAER have similar Km values but differ in their catalytic efficiency, AtAER being about 200 times more efficient than AtStR1. Expression of AtAER and AtAOR was equally high in Ath Col-0 wt and in AtStR1-mutants and could be amplified by MVK, in this way paralleling enhanced AtStR1 expression (Figure 7). Since AtAER is only weakly expressed in leaves and is enhanced moderately only under MVK stress, the strong expression of AtStR1 may lead to an amount of enzyme that can efficiently compete with AtAER for MVK detoxification. AtAOR (At1g23740), in contrast to AtAER, is strongly and constitutively expressed in the chloroplast and is assumed to be involved in the general elimination of reactive carbonyls [27]. All in all, detoxification of reactive carbonyls seems to be a more complex situation with correlating enzymes working in a group and providing a detoxifying network system. However, this issue requires further investigation. A 3 h exposure to MVK resulted in a fourfold increase of AtStR1 expression in Ath Col-0 wt (Figure 7). Expression stayed high even after 24 h of treatment (data not shown). On the other hand, AtStR2 was neither expressed in the control plants nor was its expression induced after MVK treatment, even after prolonged exposure up to 24 h. The AtPR4 gene (also termed hel gene) has successfully been used as a surrogate to estimate MVK effects [21,40,41]. AtPR4 encodes a PR (pathogen-related) protein. PR proteins are induced in response to pathogen attacks and other stresses. When treated with MVK, AtPR4 was expressed around five times higher in AtStR1-mutants than in Ath Col-0 wt. This difference was taken as an indication for impaired MVK detoxification in AtStR1-mutants. Support for this conclusion was obtained from experiments with AtStR1-mutants overexpressing 35S:AtStR1, where AtPR4 expression levels were again comparable with those of Ath Col-0 wt (Figure 8). Overexpressing 35S:AtStR2 in AtStR1mutants did not result in the attenuation of AtPR4 expression. This showed that mainly The AtPR4 gene (also termed hel gene) has successfully been used as a surrogate to estimate MVK effects [21,40,41]. AtPR4 encodes a PR (pathogen-related) protein. PR proteins are induced in response to pathogen attacks and other stresses. When treated with MVK, AtPR4 was expressed around five times higher in AtStR1mutants than in Ath Col-0 wt. This difference was taken as an indication for impaired MVK detoxification in AtStR1mutants. Support for this conclusion was obtained from experiments with AtStR1mutants overexpressing 35S:AtStR1, where AtPR4 expression levels were again comparable with those of Ath Col-0 wt (Figure 8). Overexpressing 35S:AtStR2 in AtStR1mutants did not result in the attenuation of AtPR4 expression. This showed that mainly AtStR1 but not AtStR2 is involved in MVK detoxification. AtStR1-mutant plants show a phenotype with altered vein patterning [5]. This phenotype could also be verified in the plants used in this study (Figure 9). The expression of the auxin transporter PIN1 is dependent on the availability of GSH, i.e., influenced by the tissue's redox state. PIN1-deficient mutants also showed aberrant vein patterning [42]. We here demonstrated that PIN1 expression was reduced by MVK to less than 30% of the control values after 3 h of treatment. PIN1 expression was almost AtStR1mutant plants show a phenotype with altered vein patterning [5]. This phenotype could also be verified in the plants used in this study (Figure 9). AtStR1-mutant plants show a phenotype with altered vein patterning [5]. This phenotype could also be verified in the plants used in this study (Figure 9). The expression of the auxin transporter PIN1 is dependent on the availability of GSH, i.e., influenced by the tissue's redox state. PIN1-deficient mutants also showed aberrant vein patterning [42]. We here demonstrated that PIN1 expression was reduced by MVK to less than 30% of the control values after 3 h of treatment. PIN1 expression was almost The expression of the auxin transporter PIN1 is dependent on the availability of GSH, i.e., influenced by the tissue's redox state. PIN1-deficient mutants also showed aberrant vein patterning [42]. We here demonstrated that PIN1 expression was reduced by MVK to less than 30% of the control values after 3 h of treatment. PIN1 expression was almost switched off after prolonged MVK treatment. In AtStR1mutant plants, PIN1 expression levels were about 50% of the Ath Col-0 wt level. Curing the AtStR1mutant with 35S:AtStR1 yielded plants showing equally high PIN1 expression as Ath Col-0 wt, whereas recovery was not observed in plants overexpressing 35S:AtStR2 (Figure 10).  All Ath Col-0 wt plants germinated in the presence of MVK (n = 60) displayed aberrations in vein patterning, whereas 50% of the cotyledons of non-treated controls looked like the one shown in Figure 9a. This might be in direct correlation with the reduced PIN1 expression (about 70%) in MVK treated Ath Col-0 wt plants.
AtStR1-mutant overexpressing 35S:AtStR1 or 35S:AtStR2 could not cure aberrant vein patterning ( Figure S5). This may be due to the CaMV 35S promoter which constitutively expresses genes in all transformed tissues. The venation network in Arabidopsis leaves develops in a multistage process, which requires a sophisticated pattern of auxin transport through the leaf tissue. We demonstrated that the process is to some degree perturbed in the absence of functional VEP1 and it seems likely that proper venation network development requires a proper spatiotemporal regulation of VEP1 expression. This may explain why with restored global PIN1 expression levels, constitutive overexpression of VEP1 in our transgenic plants still does not sufficiently mimic the required pattern of VEP1 activity necessary to rescue the phenotype.
This implies that the role of VEP1 in vein patterning is related to the ability of the encoded PRISE (AtStR1) to reduce small reactive 1,4-enones, as demonstrated here for MVK released in stress reactions. Other PRISEs, like AtStR2 in this study, may play completely different roles considering their individual expression patterns, their inability to cure symptoms caused by AtStR1 deficiency and their different substrate specificities with a tendency to discriminate for larger molecules. RES are cytotoxic substances causing cell damage [20,38]. RES have also been demonstrated to activate defense-related gene expression and to be involved in signaling pathways ( [21,43,44] among others). RES may therefore be regarded as archaic plant signals acting in a phytohormone-like manner. We here demonstrated that AtStR1 gene knockout led to stronger responses of defense-related genes after exposure to MVK (Figure 8). We assume that PRISE-independent modulators of defense-related gene expression have evolved in plants. All Ath Col-0 wt plants germinated in the presence of MVK (n = 60) displayed aberrations in vein patterning, whereas 50% of the cotyledons of non-treated controls looked like the one shown in Figure 9a. This might be in direct correlation with the reduced PIN1 expression (about 70%) in MVK treated Ath Col-0 wt plants.
AtStR1mutant overexpressing 35S:AtStR1 or 35S:AtStR2 could not cure aberrant vein patterning ( Figure S5). This may be due to the CaMV 35S promoter which constitutively expresses genes in all transformed tissues. The venation network in Arabidopsis leaves develops in a multistage process, which requires a sophisticated pattern of auxin transport through the leaf tissue. We demonstrated that the process is to some degree perturbed in the absence of functional VEP1 and it seems likely that proper venation network development requires a proper spatiotemporal regulation of VEP1 expression. This may explain why with restored global PIN1 expression levels, constitutive overexpression of VEP1 in our transgenic plants still does not sufficiently mimic the required pattern of VEP1 activity necessary to rescue the phenotype.
This implies that the role of VEP1 in vein patterning is related to the ability of the encoded PRISE (AtStR1) to reduce small reactive 1,4-enones, as demonstrated here for MVK released in stress reactions. Other PRISEs, like AtStR2 in this study, may play completely different roles considering their individual expression patterns, their inability to cure symptoms caused by AtStR1 deficiency and their different substrate specificities with a tendency to discriminate for larger molecules. RES are cytotoxic substances causing cell damage [20,38]. RES have also been demonstrated to activate defense-related gene expression and to be involved in signaling pathways ( [21,43,44] among others). RES may therefore be regarded as archaic plant signals acting in a phytohormone-like manner. We here demonstrated that AtStR1 gene knockout led to stronger responses of defense-related genes after exposure to MVK (Figure 8). We assume that PRISE-independent modulators of defense-related gene expression have evolved in plants.
In correlation with comparative metabolism and the evolution of specialized natural products, PRISEs encoded by VEP1-like genes may facilitate pathways leading to SNAPs [1] such as 5β-cardenolides [14] and iridoids [7]. Similarly, BAHD acyltransferases and aromatic prenyltransferases, involved in the biosynthesis of paclitaxel and monoterpenes, respectively, showed an expansion and high rates of evolution favoring some plant-specialized metabolic processes [45]. This might also be the case for the PRISE family in individual plant species.
Enzymes engaged in basic metabolism can be recruited into specialized metabolism in different plant lineages, their substrate promiscuity allowing these enzymes to contribute catalytic steps to new pathways if acceptable substrates are present. Promiscuous enzymes are important elements in pathway evolution and may complement other mechanisms such as changes in transcriptional regulation, defense strategies, changes in protein-protein interactions or protein folding [46].

Plant Material
An Transformed GVs carrying the cloned constructs were used for the transformation of SALK_097798 by floral-dip transformation [47]. Seeds of dipped plants were stratified and germinated as described above. One week after transfer into a climate chamber for germination, the seedlings were treated with glufosinate-ammonium (1 g L −1 ). This was repeated 3 times within 10 days. Surviving plants were tested for the integration of T-DNA by PCR with primers against bar gene (BAR for: TGC ACC ATC GTC AAC CAC TAC ATC GAG; BAR rev: CAG GCT GAA GTC CAG CTG). Plants with integrated T-DNA were transferred into the greenhouse for propagation. Seeds of these plants were stratified, germinated and cultivated as described above. All plants were tested for the integration of T-DNA before their use in experiments.
Primer pairs were derived using rf-cloning.org [48]. The rf-cloning method [48] had successfully been used for integrating PRISEs into pDEST17 before by [14]. The restrictionfree cloning strategy followed the procedures described previously [14,49] using E. coli DH5α cells for the primary transformation and E. coli soluBL21 for the recombinant protein expression. Correctly integrated coding sequence (in frame with His-Tag of pDEST17 vector) was confirmed by sequencing (Eurofins, Brussels, Belgium) using standard primers against T7 promoter and T7 terminator. AtStR2 expression was induced in 1 L E. coli soluBL21 cultures (optical density of 0.5 to 0.7) by adding IPTG (final concentration of 0.1 mM). To avoid inclusion bodies [50] the culture was incubated at 4 • C for 96 h as described [13]. This was previously used for the expression of PRISEs [7]. Cells were harvested and protein was extracted as described before [14]. An ÄKTA purifier chromatography system (GE Healthcare, Upsala, Sweden) and HisTrap™ HP columns were used for the purification of recombinant AtStR2 by immobilized metal-chelate chromatography, as described by [14]. SDS-PAGE and semidry immunoblotting ( Figure S1) were carried out following QIAexpress Detection and Assay Handbook (QIAgen), which was used successfully for PRISEs by [7]. We used 12% Bis-Tris polyacrylamide gels for SDS-PAGEs and nitrocellulose membranes for electroblotting. Recombinant AtStR2 was detected using mouse anti-His (mixture of RGS-, Tetra-, and Penta-His antibodies; dilution 1:2000; QiAgen, Hilden, Germany) and anti-mouse IgG-peroxidase antibodies (Sigma, Munich, Germany). As a result, we detected recombinant AtStR2 as the chemiluminescence of 3-aminophtalate released from luminol.

Enzyme Kinetics
The reductase activity of recombinant AtStR2 was measured spectrophotometrically by following the conversion of NADPH/H + at 340 nm [7,14,32]. Enzyme kinetics were determined by using protein concentrations from 0.01 to 0.1 mg/mL of recombinant AtStR2 (depending on the substrate analyzed), 0.2 mM NADPH/H + and varying concentrations of the different substrates (0.05 to 1.6 mM) [14].

In Silico Analysis
The sequence alignment was generated using ClustalW [22]. The homology model of AtStR2 was built on the crystal structure of AtStR1 (Schmidt et al. 2018, PDB entry 6el3.1.A) with Swissmodel [29]. Docking simulations of progesterone were performed with AutoDock Vina [30]. The search area for appropriate docking was defined as a cube (side length = 30 Å), the coordinates (5/−5/−60) were optimized manually. UCSF Chimera [31] was used to visualize the docking calculations.

MVK Stress Treatment
A total of 16 µL of a MVK solution (Merck KGaA, Darmstadt, Germany) was diluted with 50 mL of water. Plants were transferred into a 1 L gas chamber. Diluted MVK was added to a total concentration of 2 µmol L −1 air volume and the chamber sealed during a 3 h incubation. Control plants were treated with tap water. Plant material was harvested after a 3 h treatment [14].  Table S2).

RNA Isolation and cDNA-Synthesis
For RNA isolation, the 4 youngest leaves and the shoot apical meristem or 7 days old seedlings were used. Isolation was done with the use of Monarch ® Total RNA Miniprep Kit (New England Biolabs, Ipswich, MA, USA). gDNA was digested with a DNase I (New England Biolabs, Ipswich, USA) treatment. A total of 1500 µg of the resulting RNA was transcribed into cDNA by using a RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific Inc., Waltham, MA, USA). cDNA was diluted 3:1 and used for quantitative real-time PCR.

Quantitative Real-Time PCR and Semiquantitative PCR
All quantitative real-time PCRs (qPCR) were realized in a Rotor-Gene Q thermocycler (Qiagen, Venlo, The Netherlands) using the following program (180 s at 95 • C followed by 40 cycles of 15 s at 95 • C; T A 40 s) and Brilliant III Ultra-Fast SYBR ® Green QPCR Master Mix (Agilent Technologies, Santa Clara, CA, USA). Quantification of gene expression was calculated using the 2 −∆∆ct -method [51]. Primers designed for quantitative real-time PCR are listed in a data supplement (Supplementary Table S1). Semiquantitative PCR was carried out in a peqSTAR 96 Universal Gradient thermocycler (Peqlab, Erlangen, Germany) using either primers shown in Supplementary Table S1 and a mRNA template isolated from adult leaves and seedlings, or primers shown in Supplementary Table S2 and an mRNA template extracted from adult leaves after cold treatment. The PCR program was designed as follows: 1 × 95 • C for 180 s, 40 cycles, 95 • C for 20 s, 50 • C for 30 s, 68 • C for 60 s, 1 × 68 • C for 600 s. Taq DNA Polymerase (New England Biolabs, Ipswich, MA, USA) was used for semiquantitative PCR. PCR products were separated on 1% agarose gels.

Quantification of Protein Activity in Plant Extracts
One gram of fresh leaves was snap frozen in liquid nitrogen. Frozen plant material was ground to a fine powder. To a total of 0.2 g of plant material the same amount of PVPP was added. Protein extraction from plant material followed the method of [52].
Total protein concentrations were determined according to [53]. Progesterone 5β-reductase activity was assayed in the presence of 1 mg mL −1 total leaf protein, 0.3 mM progesterone and a NADPH/H + -regenerating system (1.1 mM glucose 6-phosphate, 6.4 mM NADP+ and 4.2 nkat glucose-6-phosphate dehydrogenase) as described by [14,52]. To monitor the consumption of substrate we used a Shimadzu GC2010/QP-2010S in IE mode with helium as carrier gas (flow: 1.2 mL min −1 ) and DB-5ms columns (JW GC column) from Agilent (30 m × 0.25 mm × 0.25 µm). Progesterone and testosterone (internal standard) were from Steraloids Inc. (Newport, RI, USA). The program used is described by [14]. Additionally, products of the above-mentioned enzyme assays were blotted on TLC plates (silica gel 60; Merck) and run in GC-MS analysis ( Figure S4), as described by [53].

Analysis of the Vein Patterning in Cotyledons
Plant tissue was fixed by immersion in a mixture of ethanol and acetic acid (3:1) overnight to remove chlorophyll and subsequently washed with water. The vein patterning was analyzed with an Axioskop HBO 50 (Zeiss).

Statistical Analysis
Experiments were conducted as three or more independent biological replicates with three technical replicates each. One-way analyses of variance (ANOVA followed by Tukey's post hoc test) were used to compare the means of various groups. A post hoc Bonferroni correction was applied for multiple comparisons. p values ≤ 0.05 were considered as statistically significant and indicated by an asterisk. GraphPad Prism 8 Software (GraphPad) was used for data analysis and graph design.

Conclusions
Our results added evidence that AtStR1, AtStR2 and PRISEs of different clusters in general play different roles in stress and defense mechanisms. Based on our findings, we conclude that PRISEs are involved in the detoxification of small reactive 1,4-enones and are therefore part of the complex detoxification machinery of RES in plants. In the context of comparative metabolism and the evolution of specialized natural products (SNAPs), promiscuous enzymes, such as PRISEs, can provide important elements in pathway evolution. A comparison of closely related plant species, accumulating different sets of SNAPs but sharing homologues sets of genes/enzymes, adds a valuable new way of looking at comparative metabolism and "silent" or "underground" metabolism.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10.3390/ metabo12010011/s1, Figure S1: SDS-PAGE and Western blot of rAtStR2, Figure S2: Enantioselective conversion of progesterone by rAtStR2, Figure S3: Expression of AtStR1 and AtStR2 after cold shock treatment, Figure S4: Analysis of progesterone 5β-reductase activity assayed in plant leaf extracts of Ath Col-0 wt and mutant lines, Figure S5: Analysis of vein patterning in Ath Col-0 wt and mutant lines, Table S1: List of primer pairs for quantitative real-time PCR experiments, Table S2: List of primers used in cold treatment experiment.