Endoplasmic reticulum stress controls PIN-LIKES abundance and thereby growth adaptation

To shed light on the control of PILS5 protein levels, we have used constitutive expression of PILS5 fused to a green fluorescent protein (GFP) (p35S::PILS5-GFP/PILS5-GFP^OX) and its growth repressive effects in Arabidopsis thaliana (6). Here, we have used the dark-induced elongation of hypocotyls (6), which is also a physiologically important response because it lifts photosynthetic organs through the soil. Using this growth model, we have conducted a nonsaturated, PILS5 enhancer screen, isolating mutants that presumably impact PILS protein abundance in a posttranslational manner [Fig. 1 A and B and (7)]. The here-identified imperial pils2 (imp2) mutant in the p35S::PILS5-GFP background (imp2; PILS5-GFP^OX) displayed an increase in PILS5-GFP abundance, correlating with reduced dark-grown hypocotyl elongation (Fig. 1 A–C). Rough mapping and next-generation sequencing of imp2 identified a mutation in the gene coding for the CHOLINE TRANSPORTER-LIKE1 (CTL1/CHER1) (Fig. 1 D and E and SI Appendix, Fig. S1A). Outcrossed imp2 mutants in the Col-0 wild-type background were not distinguishable from the cher1-4 loss-of-function mutants (SI Appendix, Fig. S1B). Moreover, the overexpression of PILS5-GFP in the cher1-4 allele induced growth retardation in dark-grown hypocotyls, being reminiscent of imp2; PILS5-GFP^OX mutants (Fig. 1D). In addition, the expression of pCHER1::CHER1-YFP in imp2; PILS5-GFP^OX rescued the growth retardation phenotype back to the level of PILS5 overexpressors (Fig. 1E). This set of data suggests that the mutation in CHER1 causes the defects observed in imp2. Accordingly, hereafter, imp2 refers to the here-identified point mutation in CHER1, which at least partially disrupts the function of CHER1. Notably, cher1-4 and imp2 mutants as well as PILS5 overexpressors show also shorter main root growth (SI Appendix, Fig. S1C). In contrast to hypocotyls, the repression of growth was not additive in imp2; PILS5-GFP^OX mutant roots (SI Appendix, Fig. S1C), suggesting a distinct mode of action in roots and dark-grown hypocotyls.

CHER1 contributes to multiple aspects, including vascular patterning (9), ion homeostasis (10), and differential growth control during apical hook development (11). Most of the pleiotropic phenotypes of cher1 mutants relate to altered levels of phosphatidylcholines in cellular membranes (reviewed in ref. 12). In agreement, imp2 and cher1-4 mutants displayed the expected alterations in phospholipid content, and these defects were not modified by ectopic PILS5 expression (SI Appendix, Fig. S1D).

Accordingly, the imbalance in membrane lipids may impact PILS protein abundance. This assumption was further supported by the usage of the ceramide inhibitor fumonisin B1 (FB1), which disrupts sphingolipid biosynthesis at the ER (13–16). FB1 applications similarly increased PILS5-GFP abundance (Fig. 2A) and also caused the enhancement of PILS5-induced growth repression in dark-grown hypocotyls (Fig. 2B). Notably, FB1 treatment of PILS5-GFP^OX seedlings phenocopied the imp2; PILS5-GFP^OX mutant (Fig. 2 A and B), but FB1 application did neither enhance the PILS5-GFP abundance nor the hypocotyl growth phenotype in imp2; PILS5-GFP^OX mutants (Fig. 2 A and B). We accordingly conclude that an imbalance in membrane lipids defines PILS protein abundance and growth.

cher1 mutants show severe growth repression in roots and shoots, but in contrast accelerated growth during apical hook opening, correlating with reduced auxin signaling rates at the concave (inner) side of the apical hook (11). In agreement, FB1 application reduced auxin signaling at the inner side of apical hooks (Fig. 2C), largely phenocopying cher1 mutants (11). This finding suggests that an alteration in membrane lipid composition affects auxin signaling in apical hooks. PILS2 and PILS5 redundantly contribute to apical hook opening kinetics, by reducing auxin signaling at the inner side of apical hooks (1). Correlating with its effect on PILS5 protein abundance, FB1-induced repression of nuclear auxin signaling was reduced in pils2 pils5 double mutants (Fig. 2C). This finding suggests that an imbalance in membrane lipids affects auxin signaling in a PILS-dependent manner.

The PILS-induced repression of auxin signaling initiates apical hook opening (1). In agreement with an increase in PILS levels and reduced auxin output signaling, cher1 mutants [Fig. 2D and SI Appendix, Fig. S2A (11)], as well as FB1 application, showed strongly accelerated apical hook opening in the dark (Fig. 2E and SI Appendix, Fig. S2B), which is reminiscent to PILS5 overexpression [(1, 11) SI Appendix, Fig. S2E]. Notably, FB1 treatments as well as the cher1-induced defects in apical hook opening were partly alleviated in pils2 pils5 mutants (Fig. 2 D and E).

We thus conclude that the interference with lipid homeostasis affects PILS protein abundance at the ER, thereby contributing to auxin-dependent growth regulation.

The imbalance in membrane lipid composition did not only affect PILS protein abundance but caused the ectopic accumulation of PILS protein-containing ER structures, which likely signifies a cellular stress response at the ER (Fig. 1C and SI Appendix, Fig. S2C). In accordance, defects in lipid metabolism, including fatty acid desaturation and phosphatidylcholine metabolism, is a cellular disturbance that causes ER stress in fungal, animal, and plant cells (17–23). In line with the published findings, imp2 mutants showed transcriptional activation of ER stress reporters (SI Appendix, Fig. S2D), also designated as Unfolded Protein Response (UPR) genes. We, hence, tested whether, in fact, ER stress affects the PILS5 protein abundance, using commonly used elicitors of ER stress, such as salt and tunicamycin (TM) treatments. Salt stress eventually limits biochemical processes, which lead, among others, to broad stress responses at the ER (24). TM is a specific inhibitor of N-linked glycosylation, thereby interfering more specifically with protein folding and consequently inducing ER stress (25). Salt as well as TM applications strongly up-regulated PILS5 proteins (Fig. 3 A and B and SI Appendix, Fig. S3 A and B). This set of data indicates that ER stress–inducing conditions, including imbalance in membrane lipids, salt stress, and unfolded proteins, lead to the upregulation of PILS5 proteins.

Subsequently, we tested whether this posttranslational effect is specific to PILS5. We observed that seedlings constitutively expressing GFP-PILS3 or PILS6-GFP showed a similar ER stress–induced upregulation (Fig. 3 A and B and SI Appendix, Fig. S3 A and B). Next, we addressed whether ER stress has a general impact on ER-localized proteins. ER stress did not increase, but reduced the abundance of the ER luminal GFP-HDEL and transmembrane ER marker DERLIN1 (DER1)-mScarlet (Fig. 3 C and D and SI Appendix, Fig. S3 A and B).

This set of data suggests that ER stress–inducing conditions exert a specific effect on PILS proteins in dark-grown hypocotyls.

To assess whether this response is possibly indirect, we addressed the response kinetics of ER stress–induced PILS protein abundance in dark-grown hypocotyls. Salt, as well as TM, increased p35S::GFP-PILS3 abundance within 1 h (SI Appendix, Fig. S4 A and B), suggesting a rather direct effect of ER stress on the posttranslational control of PILS protein abundance. Notably, we also observed a similar response for functional pPILS3::PILS3-GFP (1) in the pils3-1 mutant background (Fig. 4 A and B and SI Appendix, Fig. S4 C and D), suggesting that ER stress also affects physiologically relevant protein levels of PILS3.

In agreement with the stabilization of PILS proteins, we observed that chronic ER stress, such as germinating seedlings on TM-containing plates, also strongly enhanced the PILS5-induced growth repression in dark-grown hypocotyls (Fig. 4C). To provoke milder ER stresses, we transferred 3-d-old dark-grown seedlings for another 2 d to TM containing medium. During these 2 d, the growth of wild-type seedlings was only slightly affected, but PILS5 overexpressing seedlings still showed quantitatively enhanced growth repression (Fig. 4D), suggesting that the stabilization of PILS proteins contributes to TM-induced repression of growth rates. Similarly, we also observed hypersensitivity of PILS5 overexpressors when transferred to high salt-containing plates (Fig. 4E). The constitutive expression lines of PILS3 also showed hypersensitivity to ER stress–inducing conditions (Fig. 4 D and E), again pointing that the response is not specific to PILS5. This finding is also in agreement with a highly redundant function of PILS genes and at least PILS2 and PILS5 redundantly control hypocotyl growth in the dark (6). Conversely to the overexpression phenotypes, pils2 pils5 mutants were less sensitive when transferred to salt or TM when compared to the wild type (Fig. 4 D and E). This finding illustrates that ER stress signals repress growth in dark-grown hypocotyls at least partially in a PILS-dependent manner.

We noted that the imp2; PILS5-GFP^OX mutant enhanced PILS5-induced growth repression in dark-grown hypocotyls but not in roots of light-grown seedlings (SI Appendix, Fig. S1D). This finding points at a tissue-dependent effect and we hence tested whether the impact of ER stress on PILS5 proteins is similarly tissue specific. Notably, the induction of ER stress did not increase but lowered the PILS5 abundance in roots (SI Appendix, Fig. S5 A and B), correlating with PILS5-dependent root growth control (SI Appendix, Fig. S5C). Auxin defines plant growth in a concentration- and tissue-dependent manner, leading to a preferential stimulation and repression of growth in aerial and root tissues, respectively. We accordingly conclude that ER stress differentially affects PILS proteins in shoots and roots and thereby contributing to an overall retardation of growth.

In conclusion, we illustrate that ER stress perception defines the protein abundance of PILS proteins, which has consequences for auxin signaling rates. We accordingly conclude that the ER stress response machinery utilizes PILS proteins to provoke growth retardation.

The ER stress response machinery provides a fundamental mechanism to sense and react to environmental stresses. A variety of environmental conditions lead to the accumulation of misfolded proteins or altered composition of membrane lipids in the ER. The imbalance in these biochemical processes is sensed and activates the UPR (26, 27). Defects in the UPR sensor IRE1 affect auxin signaling output, which may relate to the transcriptional regulation of auxin receptors as well as auxin transport components (28). Here, we show that ER stress–inducing conditions define the turnover of PILS proteins and thereby link the fundamental ER stress machinery to auxin-dependent growth control. The here-uncovered posttranslational effect on PILS proteins could therefore in part mechanistically explain the interrelation of UPR and auxin signaling.

We uncover that ER stress specifically stabilizes PILS proteins in dark-grown hypocotyl, which consequently represses the nuclear auxin signaling output, leading to growth retardation. The ER stress–dependent control of PILS turnover is tissue specific, showing reduced and increased PILS turnover in shoot and root tissues. The underlying tissue-specific cues remain to be investigated, but they seem to guide the biphasic auxin responses in shoots and roots, where auxin acts as a promoter and repressor of growth, respectively. It remains however until now completely unknown how PILS turnover is molecularly defined, and hence, it is difficult to anticipate its tissue-specific regulation.

Increasing evidence already suggested that PILS proteins are important players to incorporate environmental signals into developmental growth programs (1, 2, 6–8). Here, we show that the posttranslational control of PILS protein levels also allows to integrate ER stress–inducing conditions, including imbalanced lipid homeostasis, salt stress, and unfolded proteins, with auxin signaling output. We accordingly propose that PILS proteins provide flexibility to adaptive plant development.

In conclusion, our work mechanistically links ER stress responses to PILS-dependent control of auxin-reliant growth. Accordingly, plant growth retardation under stressful environments is at least in part independent of biochemical limitations and depends on alterations in PILS-dependent auxin signaling output.

We are grateful to Ykä Helariutta for sharing published material; our team members for helpful discussions; to the Bordeaux Metabolome Facility MetaboHUB (ANR-11-INBS-0010); and the LIC Imaging Center Freiburg for expertise and support. This work was supported by the Vienna Research Group program of the Vienna Science and Technology Fund (WWTF to J.K.-V.), the Austrian Science Fund (FWF) (P29754 to J.K.-V., P33497 to S.W. and Hertha Firnberg T728-B16 and Elise Richter V690-B25 to E.F.), the European Research Council (639478-AuxinER to J.K.-V.), German Science Fund (DFG; 470007283 and via the Centre for Integrative Biological Signalling Studies- CIBSS – EXC-2189 – Project ID 390939984 to J.K.-V.), Austrian Academy of Sciences (25479 to J.F.D.S.S.), and the French National Research Agency (ANR-18-CE13-0025 to Y.B.).