RESULTS
Bacillusspp. released VCs promoted Arabidopsis LR development
Arabidopsis seedlings were grown in the presence of 7 differentBacillus strains in bi-compartmented Petri dishes. After 8 d of co-cultivation, more than half of the bacterial strains, includingB. amyloliquefaciens SQR9, FZB42, SCmB and B. subtilis168, significantly increased the PR elongation (Fig. 1A and B). A roughly 2~3 folds increase in LR number was observed in the seedlings treated with all of the strains (Fig. 1C). To exclude the possibility that increased LR number was only caused by longer PR length, we calculated the LR density and the result denied this possibility (Fig 1D). Moreover, except for B. subtilis (GZtD and168), all these strains increased total seedlings biomass production (Fig 1E).
Since B. amyloliquefaciens SQR9 is an excellent PGPR strain (L. Chen et al., 2017; Shao et al., 2014; Wu et al., 2018), we chose SQR9 as a typical model to carry out the following research. We found the regulation of SQR9 VCs on Arabidopsis growth is dose-dependent (Supplemental Fig. S1). All different doses of inoculum can promote PR elongation (Supplemental Fig. S1B). The promotion effect on LR formation and total biomass production per seedling was highest with three drops of 3 μl inoculum, while a low bacterial dose (1 drop) or spreading the inoculum on the half of the dishes caused fewer promoting effects (Supplemental Fig. S1, C and D).
To exclude the side-effect of bacterial produced CO2 or auxin for the LR formation induced by SQR9 VCs, we employed Ba(OH)2 to absorb CO2 (Ditengou et al., 2015) and tested an auxin biosynthesis deficient SQR9 mutant strain ΔysnE (Shao et al., 2014). These results showed that LR increase is not solely an effect of elevated CO2 levels produced by SQR9 or not a direct effect of auxin from SQR9 (Supplemental Fig. S2). Together, these results demonstrated that the VCs produced byBacillus spp. are practical factors to promote LR formation.
SQR9 released VCs promoted Arabidopsis LR initiation
To probe the LR development stages that are affected by SQR9 VCs, we used pCYCB1:GUS reporter lines to determine the LRP development stages. After 6 d co-cultivation, the total LRPs number increased by 33% under SQR9 VCs-treated condition compared to control treatment (Fig. 2B). Intriguingly, LRP stages I-Ⅱ were notably increased, but LRP stages V-Ⅶ were decreased in SQR9 VCs treated seedlings (Fig. 2B). These results indicated that SQR9 VCs increased LR branching might by inducing thede novo formation of LRPs and accelerating the LRP emergence. Besides, we observed an enhanced expression of CYCB1 in the meristem by SQR9 VCs treatment (Fig. 2D), indicating that SQR9 VCs promote cell division in the root apical meristem. This observation may explain why SQR9 VCs can promote Arabidopsis PR elongation.
Auxin signaling is required for LR initiation (De Rybel et al., 2010; Laskowski et al., 2008; W. Xuan et al., 2020). We detected the auxin response inArabidopsis root tip and LRPs by using pDR5:GUS marker lines. After 6 d of co-cultivation, we observed an enhanced expression of pDR5:GUS in the protoxylem at the meristem and also LRPs by SQR9 VCs treatment (Fig. 2, A and C).
SQR9 released VCs affected oscillation and prebranch site formation
DR5 oscillation in OZ is an inherent mechanism to pre-pattern LR initiation sites along the PR (Laskowski & Ten Tusscher, 2017; Moreno-Risueno et al., 2010; Van Norman et al., 2013; W. Xuan et al., 2020). Since SQR9 VCs treatment increased LRP number and LR initiation, we wondered whether SQR9 VCs could affect the DR5 oscillation and prebranch site formation. To address this question, we quantified the number of prebranch sites using pDR5:Luciferase reporter lines (Fig. 3A). Compared with control treatment, SQR9 VCs treatment significantly increased the number of prebranch sites by 29% after 6 d of co-cultivation (Fig. 3B). Notably, when co-cultivation for a longer time (8 d), SQR9 VCs treatment resulted in a reduction in prebranch site number (Supplemental Fig. S3). The density of the prebranch sites and LR along the whole PR was also increased by SQR9 VCs treatment (Fig. 3D). Moreover, we measured the periodic production of the prebranch sites in 48 h and found that SQR9 VCs treatment significantly increased the prebranch site number (Fig. 3C), indicating SQR9 VCs-induced prebranch site formed in a higher frequency. Furthermore, we observed enhanced expression of theDR5:Luciferase in the OZ of SQR9 VCs-treated roots (Fig. 3E), indicating an induction of oscillation amplitude by SQR9 VCs.
Auxin isindispensable for SQR9 VCs-induced LR formation
LR development is tightly controlled by auxin (Du & Scheres, 2018). We determined the auxin-related genes expression levels, which had significant changes in RNA-Seq (Supplemental Table S1). After SQR9 VCs treatment, most of the genes had a similar expression trend with that in RNA-seq, and IAA6 , IAA19 , IAA28 , PIN2 ,PIN3 , YUC9 , GH3.3 were up-regulated andGH3.10 was down-regulated (Fig. 4A). Furthermore, we examined the effects of SQR9 VCs-induced LR formation on a series of auxin related mutants, which are defective in auxin biosynthesis, transport, or signaling. The results showed that in all of the auxin signaling mutants, including tir1afb2 , in which oscillation amplitude and prebranch site are drastically compromised (W. Xuan et al., 2015),arf7arf19 and slr-1/iaa14, which completely abolished LR formation (Fukaki et al., 2002; Yoko Okushima et al., 2007), andCASP1pro:shy2-2 , in which a gain-of-function allele of SHOOT HYPOCOTYL2 (SHY2)/IAA3 is restrictedly expressed in the endodermis and LRP emerged is impaired (J. E. Vermeer et al., 2014), SQR9 VCs treatment did not influence LR formation as compared with the control treatment (Fig. 4, B and C). Moreover, even though SQR9 VCs could induce the LR formation in auxin transport mutants (pin2 ,pin3, and aux1-7 ), the effect of promoting LR formation on auxin efflux mutants pin2 and pin3 were weaker than that on Col-0 (Fig. 4, C and D). However, in the auxin influx mutantaux1-7 , which has fewer LRs compared with Col-0, SQR9 VCs treatment induced a comparable effect of LR formation to that in Col-0 (Fig. 4, C and D). In mutant gh3.3-1gh3.5-2gh3.6-1 , which is defective in auxin conjugation to amino acids (Nakazawa et al., 2001; Staswick et al., 2005), the effects of SQR9 VCs on promoting LR formation had no change compared with Col-0 (Fig. 2, C and D), while in mutant ech2ibr1ibr3ibr10 , which is defective in IBA-to-IAA conversion (Strader et al., 2011), SQR9 VCs has an enhanced effects on promoting LR formation compared with that on Col-0 (Fig. 4, C and D). Meanwhile, adding a well-known auxin efflux inhibitor NPA (Kim et al., 2010) to the medium with 2μM repressed the promoting effects of SQR9 VCs on LR formation. When the concentration of NPA reached 5μM, the effects of SQR9 VCs on LR formation was disappeared (Fig. 5, A and B). NPA also suppressed SQR9 VCs-promoted prebranch site formation (Fig. 5, C and D). These observations suggested a functional auxin efflux was required for SQR9 VCs-stimulated LR formation. To confirm this, we analyzed the expression of PIN1 , PIN2 ,PIN3 and PIN7 in primary roots of seedlings expressingpPIN1:PIN1:GFP , pPIN2:PIN2:GFP , pPIN3:PIN3:GFP andpPIN7:PIN7:GFP . The results showed that SQR9 VCs treatment increased the expression of pPIN2:PIN2:GFP in epidermal and cortex cells, and slightly induced the expression ofpPIN3:PIN3:GFP in the stele of primary roots (Fig. 5, E to G), whereas no changes in pPIN1:PIN1:GFP and pPIN7:PIN7:GFPexpression and localization at the root tip was observed when treated with SQR9 VCs (Supplemental Fig. S4).
IAA is mainly synthesized in a two-step pathway from tryptophan (Y. Zhao, 2012). As a rate-limiting step, indole-3-puruvate (IPA) is converted to indole-3-acetic acid (IAA) by the catalysis of flavin monooxygenases encoded by YUCCA (YUC ) genes (Mashiguchi et al., 2011; Y. D. Zhao et al., 2001). To test whether the increase in LR formation was caused by YUCs-mediated auxin biosynthesis under SQR9 VCs treatment, we analyzed the YUCs expression with SQR9 VCs treatment and employed yucasin to inhibit the function of YUCs specifically (Nishimura et al., 2014). In addition to YUC9 , the expression of YUC3 , YUC5 , YUC6 , YUC7 were also up-regulated under SQR9 VCs treatment (Fig. 6B). Moreover, the promoting effect of SQR9 VCs on LR formation was repressed with the increasing concentration of yucasin in the medium (Fig. 6, A and C). These results indicated that YUCs -mediated auxin biosynthesis is required for SQR9 VCs-promoted LR formation.
Identification of active chemicals regulating LR formation in SQR9 VCs
In order to find specific components of SQR9 VCs that promote LR formation, we analyzed the SQR9 VCs ingredients with a solid-phase microextraction (SPME) technique and GC-MS. The result showed that SQR9 produced various classes of compounds; most of them belong to ketones, hydrocarbons, and alcohols (Supplemental Table S3). Since 3-hydroxy-2-Butanone (also known as acetoin) was the most abundant compound within the chromatographic profile of the VCs (Supplemental Table S3), and it has been reported to promote plant growth (Perez-Flores et al., 2017; Ryu et al., 2003). We thus tested the effect of acetoin on LR information inArabidopsis by adding pure acetoin to the medium and using acetoin biosynthesis defective mutant ΔalsD (Wu et al., 2018) in this experiment. The results showed that acetoin slightly promoted LR formation at lower concentration (10 and 30 μM), and higher concentration (100, 300 and 1000 μM) has no effects on that (Fig. 7, A and B). The effect of SQR9 VCs on inducing LR formation is mildly repressed when inoculated ΔalsD (Fig. 7 C and D) and acetoin accounts for 19% of the effects of SQR9 VCs on promoting LR formation (Fig. 7E). Moreover, the effects of other components detected in SQR9 VCs on LR formation, whilst those pure compounds cannot induce LR formation (Supplemental Table S4). We further investigated whether the mechanism of acetoin promoting LR increase in Arabidopsis is the same as that of SQR9 VCs. Acetoin did not affect the number of prebranch sites formed within 48 hours in DR5:Lucferase seedlings (Fig. 8, C and D) and also not affect the CYCB1 expression pattern at the PR tips (Fig. 8B). Although acetoin slightly increased the total number of LRP, the stage distribution of LRPs was not significantly affected by treatment with it (Fig. 8A). These data indicated that acetoin in SQR9 VCs plays a certain role in promoting LR formation, but it is feeble compared with the effect of SQR9 VCs as a whole and thus not explain the role of SQR9 VCs in promoting LR formation.