The proanthocyanidin (PA) branch of the flavonoid pathway affects rhizobacteria-mediated growth promotion in Arabidopsis
Considering our results and other reports on the negative impact of flavonoids on auxin transport, auxin distribution and turnover (Murphyet al. 2000; Brown et al. 2001; Besseau et al.2007; Santelia et al. 2008; Kuhn et al. 2011, 2016; Bueret al. 2013), we investigated if Pbg can establish effective partnership with Arabidopsis transparent testa (tt) mutants that harbor mutations in biosynthetic and transport genes along the main trunk route of the flavonoid pathway. Before assessing the response of the tt- mutants to Pbg, we evaluated their inherent variation in biomass accumulation. These twelve mutants showed significant variation in biomass accumulation. In general, mutants in the downstream steps of the flavonoid pathway showed reduced shoot biomass when compared to mutants from the upstream steps of the pathway (tt4 , tt5 and tt6 ) and the WT (Figures 4a, 4b and 4c ). The upstream mutants produce no detectable level of flavonols that are often implicated to interfere with auxin transport and turnover (Bowerman et al. 2012). However, that did not lead to growth promotion in these mutants. The tt7 mutant that accumulates kaempferol-derived flavonols (Routaboul et al. 2006) showed the lowest shoot biomass accumulation (Figure 4c ) and when treated with Pbg showed no significant difference in shoot biomass from the WT (Figures 4b, c and d ). Interestingly, both the ineffective partnerships between Arabidopsis-Pbg and Broccoli-Pf SS101 exhibited high accumulation of kaempferol-derived flavonols (Figure 2a2 and2b2 ). In Arabidopsis, kaempferol diglycosides are reported to act as an endogenous polar auxin transport inhibitor (Buer & Muday 2004; Peer et al. 2004; Yin et al. 2014; Kuhn et al. 2016, 2017). This may suggest that kaempferol-derived flavonols could negatively influence the inherent growth of Arabidopsis and also potentially compromise its responsiveness to Pbg .
From the early steps of the flavonoid pathway, mutants tt 4,tt 5 and tt 6 are defective in the biosynthesis of the three main branches of the flavonoid pathway such as flavonols, anthocyanins and PA (Lepiniec et al. 2006; Bowerman et al. 2012). Of the three mutants, tt 4 and tt 6 responded the least toPbg in terms of shoot biomass accumulation suggesting that some group of flavonoids might be required for Pbg -mediated growth promotion. Although the tt 5 mutant was expected to behave similarly to tt 4 and tt 6, it showed a significant increase in shoot biomass accumulation upon treatment with Pbg probably owing to its ‘leaky’ phenotype (Peer et al. 2001).
Mutants from the downstream branches of the flavonoid pathway defective in anthocyanin (tt 18, tds 4) and proanthocyanidin biosynthesis (tt 18, tds 4, tt 15, tt 12,aha 10 and tt 10) established effective partnership withPbg leading to significant increases in shoot biomass (Figure 4b ). The role of anthocyanin can be excluded in the establishment of effective partnership as the ban mutant that accumulates anthocyanin at the expense of PA (Devic et al. 1999; Xie et al. 2003) showed no significant influence on the growth phenotype. Similarly, the tt 7 that does not accumulate anthocyanin showed no significant growth promotion when treated withPbg. Hence, the influence of tt 18 and tds 4 on the establishment of effective partnership could be due to their influence on the PA pathway.
Among the PA mutants, the tt 12 gene encodes a transporter that is homologous to the Multidrug and Toxic compound Extrusion (MATE) secondary transporter (Marinova et al. 2007), the aha 10 gene encodes a H+-ATPse that is also involved in PA metabolism (Baxter et al. 2005) and tt 15 encodes the UDP-Glc:sterol glycosyltransferase UGT8B (DeBolt et al. 2009). The tt 10 gene is involved in the formation of polymeric pigments from epicatechin and may catalyze the oxidative browning of colorless PAs (Pourcel et al. 2005). Pbg had the highest growth-promoting effect on the tt 10 mutant that was reported to accumulate epicatechin (CE) and procyanidin polymers and soluble PA but lacks oxidized PA (Pourcel et al. 2005). Mutant tt 15 showed reduced PA and oxidized PA, cyanidin and quercetin (Fockset al. 1999; Routaboul et al. 2012), tt 12 revealed absence of CE, PA, oxidized PA and reduction of the major flavonol quercetin-3-O- rhamnoside (Marinova et al. 2007) andaha 10 accumulates CE and shows highly reduced PA and oxidized PA (Baxter et al. 2005). When evaluated together, the tt 15,tt 12 and aha 10 mutants lack or have a highly reduced PA and a concomitant reduction of oxidized PA. The tds 4 andtt 18 mutants that established effective partnership withPbg are reported to lack or have highly reduced anthocyanin levels, have reduced CE, highly reduced PA and lack oxidized PA. Hence, the common denominator of all Arabidopsis transparent testa (tt) mutants that established an effective partnership with Pbg is the absence of oxidized PA (tannins). In conclusion, our results indicate that the flavonoid pathway is a prime target of rhizobacteria and various branches of this pathway can have different impacts on the inherent growth and responsiveness of Arabidopsis to growth promoting rhizobacteria. Based on these results, we postulate that kaempferol-derived flavonols and the oxidized products of PA can negatively influence rhizobacteria-mediated plant growth promotion.