- Top of page
- Materials and Methods
- Supporting Information
The phytohormone auxin, represented predominantly by indole-3-acetic acid (IAA), plays a crucial role in plant growth and development. Auxin needs to be transported from the sites of synthesis, mainly in the apices and young leaves, to the distal part of the plant to exert its function (Berleth et al., 2007). Auxin transporters, including ABCB proteins, AUX1/LAX family members and PIN proteins (Noh et al., 2001; Friml, 2003; Petrasek et al., 2006; Yang et al., 2006; Zažímalová et al., 2010; Peer et al., 2011), are responsible for the auxin fluxes and patterning in plants (Friml et al., 2003). Flavonoids, phenylpropanoic secondary metabolites, have been implicated in the blocking of auxin transport (Peer & Murphy, 2007). The supply of flavonols to detached zucchini hypocotyls resulted in decreased polar auxin transport (PAT) (Jacobs & Rubery, 1988). In Arabidopsis, PAT was increased in transparent testa4 (tt4), a flavonoid-deficient mutant defective in the first step of flavonoid production (Shirley et al., 1995; Buer & Muday, 2004; Peer et al., 2004) (Fig. 1). Accordingly, tt4 roots exhibited delayed gravitropism, which was reversed by chemical complementation by naringenin, an intermediate of flavonoid biosynthesis (Buer & Muday, 2004). By contrast, PAT was reduced in the flavonol over-production mutant tt3 defective in dihydroflavonol reductase (Fig. 1), consistent with an inhibitory role of flavonols in PAT (Peer et al., 2004). Despite these and other substantial pieces of evidence supporting a role of flavonols in the modulation of auxin transport (Kuhn et al., 2011; Lewis et al., 2011; Grunewald et al., 2012), neither specific flavonol aglycones nor their conjugates active in this process in vivo have been identified so far.
Figure 1. Flavonoid biosynthesis pathway in Arabidopsis thaliana. (a) Scheme of flavonoid biosynthesis. CHS (TT4), chalcone synthase; F3′H (TT7), flavonoid 3′-hydroxylase; DFR (TT3), dihydroflavonol 4-reductase; FLS, flavonol synthase; ANS (TT18), anthocyanidin synthase; UGT78D1, flavonol 3-O-rhamnosyltransferase; UGT78D2, flavonoid 3-O-glucosyltransferase; UGT89C1, flavonol 7-O-rhamnosyltransferase. As shown by Yin et al. (2012), the combined loss of UGT78D1 and UGT78D2 does not imply an accumulation of flavonol aglycones because of a feedback inhibition of flavonol biosynthesis. (b) Glycosylation reactions catalyzed by UGT78D1, UGT78D2 and UGT89C1 (Jones et al., 2003; Tohge et al., 2005; Yonekura-Sakakibara et al., 2007). Abbreviations: Kaempferol (k); rhamnoside (rha); glucoside (glu). k1, k-3-O-rha-7-O-rha; k2, k-3-O-glu-7-O-rha; k3, k-3-O-[rha (1->2 glu)]-7-O-rha; q1, q2 and q3 are quercetins structurally equivalent to k1, k2 and k3, respectively.
Download figure to PowerPoint
The difficulty in relating specific flavonols to auxin transport modulation is, in part, a result of the complex flavonol modification in planta. Flavonol aglycones are intensively modified by UDP-dependent glycosyltransferases (UGTs), which include UGT78D1, UGT78D2, UGT78D3, UGT73C6 and UGT89C1 in the model plant Arabidopsis thaliana (Fig. 1) (Jones et al., 2003; Tohge et al., 2005; Yonekura-Sakakibara et al., 2007, 2008). The glycosides are distributed in an organ-specific manner. In contrast with the complex flavonol profile in flowers, it is rather simple in inflorescence stems (Yonekura-Sakakibara et al., 2008; Stracke et al., 2010b). Thus, the inflorescence stem, which is implicated in basipetal auxin movement, is the optimal organ for searching for the flavonol derivative(s) active in auxin transport modulation.
Here, we show that the loss of the flavonoid 3-O-glucosyltransferase UGT78D2 resulted in an altered flavonol glycoside pattern and reduced PAT in shoots, which was accompanied by a reduced plant height and increased branching. Blocking of flavonoid biosynthesis and/or glycosylation at specific steps clearly related the enhanced accumulation of kaempferol 3-O-rhamnoside-7-O-rhamnoside (k1) to the growth defects of ugt78d2. Through analyses of auxin transport in several genotypes, which contained different levels of k1, an inverse correlation between basipetal auxin transport and k1 level was identified. Therefore, we propose that k1 acts as an endogenous auxin transport inhibitor in Arabidopsis shoots.
- Top of page
- Materials and Methods
- Supporting Information
Over the past decade, strong evidence, both in vitro and in vivo, has accumulated implicating flavonols in plant growth regulation and auxin transport modulation (Jacobs & Rubery, 1988; Buer & Muday, 2004; Peer et al., 2004; Ringli et al., 2008; Santelia et al., 2008; Kuhn et al., 2011; Lewis et al., 2011; Grunewald et al., 2012; Buer et al., 2013). However, the physiologically active flavonol derivative(s) could not be pinpointed unequivocally. In vivo feeding studies were hampered by the possible modification and metabolism of the exogenously added compounds. Similarly, genetic studies affecting either flavonol biosynthesis or flavonol conjugation did not generate effects that could be traced back to a single flavonol moiety (Buer & Muday, 2004; Peer et al., 2004; Ringli et al., 2008; Buer et al., 2013). Nevertheless, the genetic identification of the kaempferol derivative k1 as an endogenous inhibitor of PAT in this work is in agreement with previous studies. These have indicated activity associated with exogenously applied flavonols that could be metabolized to k1 (Jacobs & Rubery, 1988; Mathesius et al., 1998). Furthermore, a release from PAT suppression in flavonol-free tt4 was observed, whereas PAT is further repressed in tt3, which exhibits an enhanced flavonol content at the expense of anthocyanins, and, importantly, in tt7, which shows specifically increased kaempferol glycoside levels at the expense of quercetins (Buer & Muday, 2004; Peer et al., 2004). Furthermore, the formation of k1 is dependent on the 3-O-rhamnosyltransferase UGT78D1, which is relatively abundant at the shoot apex in comparison with other tissues and almost absent from roots in agreement with a low k1 level in roots (Jones et al., 2003) (Figs S1, S8). Thus, the expression pattern of UGT78D1 correlates with an in situ activity of k1 in modulating auxin transport and affecting the shoot phenotype.
Nevertheless, the identification of k1 as an active compound does not exclude an impact of other flavonol derivatives on auxin fluxes. Flavonol biosynthesis is under developmental and organ-specific regulation by MYB and WRKY transcription factors (Stracke et al., 2007, 2010b; Grunewald et al., 2012). Accordingly, the active compounds could also be dependent on the developmental stage and/or on the organ. WRKY23 has been revealed to be part of a feedback loop of auxin to repress its own transport in roots, as it was induced by auxin and enhanced the biosynthesis of flavonols, in particular of quercetins. Thus, quercetins have been suggested to be active agents in roots (Grunewald et al., 2012; Buer et al., 2013). This finding was supported by the analysis of the quercetin-less tt7, which indicated that no kaempferol derivatives, but rather quercetins, were involved in the suppression of basipetal auxin flux in roots (Lewis et al., 2011; Grunewald et al., 2012). However, Buer et al. (2013) also found a repressed root PAT in both quercetin-accumulating tt3 and quercetin-deficient tt7. Nevertheless, these data and the very low k1 level in roots, as well as the lack of obvious ugt78d2-related phenotypes in hypocotyls and roots, including a similar auxin accumulation in mutant and wild-type root tips monitored by a DR5::β-glucuronidase reporter line, strongly support the notion that different flavonol compounds may affect PAT in root and shoot (Figs S1, S9; Table S1).
Suppression of rhamnose biosynthesis in rol1-2 strongly altered the flavonol profile and induced, for example, hyponastic growth of cotyledons and aberrant leaf cell development. This was related to enhanced auxin accumulation and a repressed auxin (non-IAA) efflux from mesophyll protoplasts (Ringli et al., 2008; Kuhn et al., 2011). As rol1-2 growth defects were rescued in tt4 rol1, but retained in tt7 rol1, they were attributed to kaempferols. However, k1 was not the causal metabolite for this particular phenotype, as rol1 ugt78d1 lacking k1 retained the rol1 cotyledon phenotype (Ringli et al., 2008).
Several possible mechanisms have been reported on how flavonols might affect auxin transport. Flavonol biosynthetic mutants led to an altered expression, subcellular localization and perhaps modified dynamics in the plasma membrane of some PIN proteins, which could have been caused by a direct or indirect impact of flavonols on these proteins (Peer et al., 2004; Santelia et al., 2008). At the transcriptional level, there was no difference in mRNA abundance of several PIN as well as ABCB1 and ABCB19 genes in ugt78d2 and wild-type plants (Fig. S6). ABCB1 and ABCB19 are ABC transporters which primarily function in the long-distance auxin transport streams and the movement of auxin out of apical tissues (Bandyopadhyay et al., 2007). Therefore, these ABCB transporters are potential targets of k1 in planta. Indeed, ABCB1/19 proteins are able to bind the quercetin aglycone (Murphy et al., 2002), and inhibition of ABCB1 auxin transport activity by the quercetin aglycone has been demonstrated in Arabidopsis protoplasts (Geisler et al., 2005). These observations are in agreement with previous in vitro experiments showing that flavonol aglycones can compete with the synthetic auxin transport inhibitor NPA for binding to isolated microsomal vesicles, although these results may not necessarily reflect an in vivo relevance (Jacobs & Rubery, 1988; Murphy et al., 2002). Nevertheless, these studies indicate that flavonol derivatives, including k1, could interact directly with and/or inhibit auxin transport proteins.
Despite these in vitro studies, a possible involvement of flavonol aglycones in auxin transport inhibition in vivo and, in particular, in the ugt78d2 growth phenotype appears to be unlikely. This notion is based on two lines of evidence: first, to the best of our knowledge, so far only two of numerous studies have detected flavonol aglycones in Arabidopsis by either HPLC (Peer et al., 2001) or mass spectrometry (Buer et al., 2013); however, the latter authors also discussed inconsistencies with respect to the occurrence and absence of aglycones in different tt mutants. Although flavonol aglycones exist at least as biosynthetic intermediates, only minute amounts of the hydrophobic molecules might be tolerated in living cells, and therefore free flavonols are not consistently detected; however, this would not preclude a regulatory role (Yin et al., 2012). The important second line of evidence stems from loss-of-function mutants affecting the two major UGTs in Arabidopsis, UGT78D1 and UGT78D2, which perform the initial 3-O-flavonol glycosylation (Fig. 1). The lack of either UGT is compensated by the remaining one, and results in a shift of the flavonol glycoside pattern with no aglycones being detected (this work; Jones et al., 2003; Yonekura-Sakakibara et al., 2008; Yin et al., 2012). The ugt78d1 ugt78d2 double mutant strongly enhances the chance of flavonol aglycone accumulation, yet no free flavonols were detected (Yin et al., 2012) and, more importantly, the ugt78d2-dependent growth retardation was reversed in the double mutant (Fig. 2).
Flavonols have also been shown to exert their function by modulating the interaction between TWD1, a regulatory protein, and ABCB1 transporters, thus suggesting another possible scenario (Bailly et al., 2008). Furthermore, it has been reported that PINOID modulates ABCB1-mediated auxin transport through its kinase activity, and that this effect is reversed by direct quercetin binding, that is, the flavonol acts as an endogenous kinase inhibitor to modulate PAT (Henrichs et al., 2012). In summary, these mechanistic studies on auxin transport inhibition allow for both different active flavonol moieties and multiple target proteins. The discovery of k1 as an active molecule in planta may serve as a valuable tool to obtain further insight into the detailed mechanism of flavonol-dependent auxin transport inhibition.
Apart from the genetic designation of k1, the dose-dependent PAT repression by k1 further supports its identification as an endogenous PAT inhibitor in Arabidopsis shoots (Fig. 4). Interestingly, there is also almost the same quantitative relationship in ugt78d2 relative to the wild-type between the concentrations of k1 (c. two-fold higher in the mutant), endogenous auxin (c. two-fold lower in the lower inflorescence stem segments of mutant plants) and basipetal auxin transport (c. two-fold lower in the basal part of the inflorescence stem of the mutant) (Figs 3b, 4a; Table 4). A further increase in k1 in tt7 ugt78d2 resulted in more severe growth defects. However, the complete lack of k1 (in tt4, ugt78d1, ugt89c1) did not result in a higher stature of the A. thaliana plants, suggesting a threshold for k1-dependent growth repression or an independent limitation of growth of these mutants.
In plants, a minor increase in k1 levels would only lead to moderate PAT inhibition, which might not lead to obvious phenotypic changes (as seen in the shoots of abcb1 and abcb19 single mutants; Noh et al., 2001). However, under harsh natural growth conditions, including high light, UV-B irradiation, extreme temperatures and drought, k1 as well as other flavonol glycosides could be strongly increased (Hannah et al., 2006; Rausher, 2006; Caldwell et al., 2007; Korn et al., 2008; Götz et al., 2010; Stracke et al., 2010a). Thus, regulation of flavonol glycoside levels may constitute a means to adjust plant growth and stature by environmental factors and in ecological contexts.