The function of the P450 CYP98A gene subfamily has been extensively discussed in the literature for over 10 years. More than 60 different genes have been sequenced and referenced in the international P450 database (http://drnelson.uthsc.edu/cytochromep450.html
) as belonging to this subfamily. Various functions have been attributed to these proteins. Their activity is classically described as a C3’H catalyzing hydroxylation in the 3' position of p-
coumaroyl shikimate and p-c
oumaroyl quinate [15
]. However, other substrates have also been described, such as 4-p
-coumaroyl-3' 4'-dihydroxyphenyllactate [19
coumaroyl tyramine [18
] or spermidine esters [21
]. Among the enzymes that metabolize p-
coumaroyl shikimate, some were demonstrated to hydroxylate p-
coumaroyl quinate at a lower rate and efficiency [15
], while others were simply unable to metabolize both substrates [28
]. In some cases, these enzymes have been described as having a larger effect on metabolism. For instance, Schoch and collaborators [17
] indicated that the presence of C3’H is necessary for the biosynthesis of many divergent compounds such as, lignins, UV-absorbing pigments, antioxidants, flavors, fragrances and coumarins. Together, all these data show that this enzyme plays a central (Figure) and pivotal role in the phenylpropanoid pathway and that it has an indirect impact on the synthesis of several molecules. In this work, we hypothesized that an orthologous gene isolated from a furanocoumarin-producing plant might have impact the synthesis of some of these molecules and play an important role in the answer to various stresses.
Little is known concerning the genome of R. graveolens
, a furanocoumarin-producing plant. Only a few accessions (less than 80) are available in the nucleic acid or protein databases, and no information concerning a putative gene corresponding to a cyp98a
orthologous gene was found. However, each P450 subfamily shares several conserved consensus domains which makes it possible to use sequence-based PCR approaches for cloning and identifying plant cytochrome P450 genes. Accordingly, we used a PCR approach with degenerated primers [24
] to isolate a gene from the R. graveolens
genome that potentially encodes a C3’H. The translational product of the resulting open reading frame (ORF), cyp98a22
(Genbank JF799117), was compared with 25 peptidic sequences available in the databases. The results showed that the protein is clustering with enzymes described as classical C3’H involved in the synthesis of lignin precursor, among which the first C3’H CYP98A3 ever functionally described and not with divergent CYP98 as the Arabidopsis thaliana
CYP98A8 and CYP98A9 which are involved in the synthesis of N10
-dihydroxyferuloyl spermidine, an important constituent of pollen [21
]. This first set of data seemed to indicate that this enzyme might be a conventional C3’H that potentially metabolized p-
coumaroyl quinate and p-
The biochemical characterization of plant cytochrome P450s using a yeast expression system [29
] is often difficult due to the low abundance and instability of these membrane-bound proteins. Thus, several different heterologous systems for the expression of the P450s, including yeast, E. coli
and baculovirus, have been described in the literature. To perform the enzymatic characterization of CYP98A22, we selected the yeast expression system described by Pompon and collaborators [29
] which is extensively used for the heterologous expression and characterization of enzymatic activities of proteins belonging to the P450 family. This system is efficient for the production of certain enzymes [30
] but was limited for the production of other plant enzymes. Various improvements have been proposed in the literature, including changes in the media and culture conditions [31
], the modification or exchange of nucleotide sequences (partial or complete recoding of genes) [32
] or the replacement of the membrane anchor [32
]. A carbon monoxide differential spectrum is generally performed to assess the expression level of P450 in such a system. To achieve the expression of cyp98a22
, several of these strategies were employed. Although no peak at 450 nm could be observed with a CO spectrum for any strategy tested, we used the prepared microsomes to conduct metabolisation tests using the following substrates: p-
coumaroyl quinate, p-
coumaroyl shikimate, p-
coumaroyl tyramine, and p-
tricoumaroyl spermidine. No transformation of any substrate could be detected, whereas the metabolisation of p-
coumaroyl quinate and shikimate was observed in the presence of CYP98A3 yeast microsomes. This lack of expression could not been explained so far but led us to move to another protein heterologous expression system.
To determine the activity of CYP98A22, we used N. benthamiana
as an alternative heterologous expression system. The transient transformation of the plant petals with cytochrome P450s using particle bombardment as a technique for transferring a T-DNA has been described in the literature [34
]. The use of this system has also been reported for the functional characterization of a P450 belonging to the CYP71A subfamily involved in the biosynthetic pathway of alkaloid phytoalexins [36
To test the efficiency of this plant heterologous expression system to allow the expression of CYP98A22, we first constructed a fusion protein comprising CYP98A22 and the yellow fluorescent protein (YFP). This recombinant ORF was placed under the control of the CaMV 35S promoter and agroinfiltrated in the epidermal cells of N. benthamiana
. Because no fluorescence could be observed, a second set of infiltrations was performed in the presence of a plasmid allowing the expression of the Tomato Bushy Stunt Virus silencing suppressor protein P19 as described by Voinnet et al.
]. In this case, the analyses of the infiltrated leaves revealed a strong fluorescence signal localized in the endoplasmic reticulum. This result is consistent with the work of Ro and collaborators who demonstrated that the cinnamate 4-hydroxylase (C4H), a cytochrome P450 enzyme of the phenylpropanoid pathway, is bound in the ER membrane [37
]. Our results show that this plant system is efficient for the functional expression of CYP98A22 in the presence of the gene silencing suppressor protein.
To realize the functional characterization of the enzyme, a new set of plasmids were constructed containing the ORF of cyp98aA22
under the control of the 35S promoter in the pBIN vector prior to agroinfiltration to N. benthamiana
in the presence of TBSV P19. As a first attempt, we prepared microsomes from the leaves at 4 days post-infiltration, and the enzymatic tests showed that both p-
coumaroyl quinate and p-
coumaroyl shikimate were hydroxylated to their caffeic counterparts by CYP98A22 and CYP98A3. However, CYP98A22 showed a much higher efficiency with p-
coumaroyl quinate as a substrate than with p-
coumaroyl shikimate, whereas CYP98A3 seems to preferentially metabolize p-
coumaroyl shikimate. This first experiment revealed that the N. benthamiana
plant system is a more efficient tool for the expression and the in vitro
characterization of CYP98A22 than the yeast system. The use of this plant heterologous system for the expression of CYP98A22 also provides a rapid method to test the in vivo
activity of the concerned protein. In addition to a pool of p-
coumaroyl esters of quinate and shikimate, the dominant phenylpropanoids of tobacco is chlorogenic acid [38
]. The presence of these molecules in plants provides an interesting tool to determine the activity of the overexpressed CYP98A enzymes and to highlight the natural function of these enzymes. The analysis of the metabolic profile and the quantification of chlorogenic and caffeoyl shikimic acid were performed on leaves infiltrated with recombinant agrobacteria containing pBIN-CYP98A22 and on non-infiltrated leaves. The results showed a statistically significant increase of the chlorogenic acid content.
To complete this study, we examined the tissue-specific expression pattern using real-time PCR. Indeed, as discussed above, most of the members of the CYP98A P450 subfamily catalyze efficient 3’-hydroxylation of p-
coumaroyl shikimate and much slower synthesis of caffeoyl quinate [15
]. Because the hydroxylation of p
-coumaroyl-shikimate is a key step in the formation of the monolignol lignin precursor, this observed activity, along with the relatively high level of expression of most of the described C3’H in stem and vascular bundles, indicates that these genes play an important role in lignification [15
]. The results we obtained in our investigation concerning the expression pattern of cyp98a22
were somewhat different. Unlike other C3’H, cyp98a22
displays a broad expression pattern. Strong expression was observed in the petals and roots, whereas a relatively weak expression was observed in the petioles, stems and pistils.
Together, these results show that even if CYP98A22 is phylogenetically related to CYP98A3, it might play a more complex or, at least, a different role in planta.
Indeed, caffeoyl shikimate plays an important role in lignification, while the caffeoyl quinate derivatives, such as chlorogenic acid, are described as growth regulators, disease resistance factors, antioxidants and compounds affecting the organoleptic quality of fruits [40
]. The ability of CYP98A22 to be more dedicated to the synthesis of chlorogenic acid constitutes an evidence of its involvement in the stress response.
To explore this hypothesis, plantlets were exposed to UV-B light for 24 hours as described by Vialart et al.
]. Altough the analysis of the furanocoumarin content showed no increase in comparison with the non-treated plants, the UV-B elicitation stimulated the synthesis of umbelliferone, which is a precursor of this pathway. Moreover, these analyses demonstrated that the expression level of cyp98a22
gene was strongly increased (3-fold) in UV-treated leaves as compared with non-treated leaves. Recently, the elicitation experiments were utilized in Cynara cardunculus
to induce the expression of cyp98a49
using UV-C light [23
]. Although this does not provide irrefutable evidence, these elements are consistent with the hypothesis that CYP98A22 is involved in responses against several stresses in R. graveolens
. These preliminary results must now be explored in depth.
Scopoletin is an important defense compound ubiquitously found in higher plants for which two biosynthetic routes have been described (Figure). A first route, operating through feruloyl-CoA was demonstrated by Kai and collaborators [42
] in Arabidopsis, which implies the conversion of p
-coumaroyl-shikimic acid into caffeoyl-shikimic acid by CYP98A3, upstream to feruloyl-CoA [43
]. Scopoletin content was dramatically decreased in knock-out cyp98a3
mutants, confirming the role of the Arabidopsis C3’H in the synthesis of this compound. An alternate route to scopoletin was proposed in plants such as Daphne mezereum
] and Agathosma puberula
] where umbelliferone and esculetin are the direct precursors of scopoletin. It is unestablished in R. graveolens
if the two pathways are coexisting or if only one of the two is prevalently operating. To assess the involvement of CYP98A enzymes in the synthesis of scopoletin in R. graveolens
, transgenic Ruta plants overexpressing either cyp98a3
where generated. The analysis done on the in vitro
plants highlighted a statistically significant increase of the scopoletin content in 35S::CYP98A3 plants whereas plants overexpressing cyp98a22
did not show any modification in scopoletin content when compared to control plants. The results obtained on CYP98A3 are consistent with the findings of Kay et al.
] who described the key-role of C3’H in the synthesis of scopoletin in Arabidopsis. In parrallel, the results obtained on CYP98A22 fit well with our in vitro
investigations obtained with N. benthamiana
, which demonstrate the preferential role of this enzyme in the synthesis of chlorogenic acid vs
caffeoyl-shikimate, the precursor of scopoletin in the first biosynthetic route described above. The analyses were extended to other coumarin which led us to show that the overexpression of both CYP98A impact significantly the concentration of furanocoumarins in transgenic plants. This constitute an unexpected result because C3’H is not directly involved into the synthesis of furanocoumarins (Figure). The exact reason for this increase in furanocoumarin content remains unknowned and will require further metabolomic investigations to be elucidated.