Our initial characterization defined rsr1-1 as a mutant with impaired HXK-independent sugar induction of the patatin class I promoter from potato and altered regulation of proline metabolism and sensitivity (Martin et al., 1997; Hellmann et al., 2000). In addition to these phenotypes, rsr1-1 mutants displayed a variety of differences in appearance to the parental line Pat(B33)-Gus when grown on soil: First, the chlorophyll content of rosette leaves was reduced by approximately 25%, which caused a yellowish appearance of the leaves (Figure 1A). Second, the mutant also displayed spontaneous lesions on all leaves under long-day conditions in the greenhouse (Figure 1B). Staining of leaves with 3,3′-diaminobenzidine revealed increased H₂O₂ production in these lesions (Figure 1C). Furthermore, rsr1-1 mutants developed the first inflorescence about 10 days earlier than control plants (Figures 2A,C), which correlated with a reduced leaf number at the onset of bolting (Figure 2B), classifying rsr1-1 as a novel early flowering mutant. Additionally, changes in root morphology were observed when rsr1-1 mutants were cultured on MS-medium supplemented with 2% Suc: under such conditions roots of rsr1-1 seedlings were almost twice as long as observed for Pat(B33)-Gus plants (Figure 2C). In contrast, root hairs of rsr1-1 seedlings were slightly deformed and shorter in comparison to control plants (Figures 2D,E).

To determine if altered sugar sensitivity of rsr1-1 was manifested already at the embryonic stage, germination of Pat(B33)-Gus and rsr1-1 seeds was assayed on MS-plates containing either 30 mM Glc, 330 mM Glc, or 30 mM Glc plus 300 mM 3Omeg. As an osmotic control, a combination of 30 mM Glc and 300 mM sorbitol was applied. Germination was defined as the time point when the radicle breaks through the seed coat. At 30 mM Glc, more than 95% of the seeds germinated within 3 days after plating, and no difference was detected between rsr1-1 and Pat(B33)-Gus seeds (Figures 3A,B). All other treatments delayed germination of both lines markedly. Interestingly, high concentrations of 3Omeg or Glc slowed germination to a greater extent than sorbitol in Pat(B33)-Gus seeds (Figure 3A). A sugar-specific effect that is mimicked by 3Omeg indicates that the underlying signaling cascade is independent of Glc phosphorylation and further metabolism. In rsr1-1 seeds, the effects of Glc and 3Omeg did not differ from the osmotic control, indicating that the sugar-specific signal was not correctly transmitted (Figure 3B). The effect of 3Omeg was specific for germination, as seedlings of both lines, Pat(B33)-Gus and rsr1-1, did not display any developmental arrest on medium containing 30 mM Glc plus 300 mM 3Omeg (data not shown).

At later developmental stages, it is known that seedling establishment is inhibited by high sugar concentrations, and a variety of mutants have been described that are insensitive to this inhibitory effect (Pego et al., 2000; Rolland et al., 2002). To investigate whether rsr1-1 is also a Glc-insensitive mutant, Pat(B33)-Gus and rsr1-1 seedlings were cultured on growth medium supplemented with increasing d-Glc concentrations. While Pat(B33)-Gus plants were arrested in development at the cotyledon stage, and chlorophyll synthesis was blocked at concentrations higher than 300 mM (Figures 3C,D), rsr1-1 seedlings still developed green cotyledons followed by true leaves on medium containing up to 330 mM Glc. However, further increase in Glc (360–380 mM) also resulted in arrest of rsr1-1 development. After 6–8 weeks on 360 mM Glc, rsr1-1 plants developed callus-like tissue at the apical meristem, which was not observed in Pat(B33)-Gus (Figures 3E,F). These experiments demonstrate that mutation of RSR1 shifted the upper limit of Glc-tolerance, but did not result in complete Glc insensitivity.

Earlier studies had demonstrated that the inhibitory effect of high Glc concentrations on seedling establishment involves ABA signaling (Dekkers et al., 2008; Wingler and Roitsch, 2008), while low Glc concentrations were shown to reduce the inhibitory effect of ABA on germination and development (Finkelstein and Lynch, 2000). Addition of 3 μM ABA to the growth medium strongly delayed germination seeds from both, rsr1-1 mutants and the parental line Pat(B33)-Gus, but still allowed germination of nearly all seeds. Seedling establishment in both lines was completely blocked by 3 μM ABA on sugar-free MS-medium (Figure 4A). Addition of 30 mM Glc attenuated the delay in germination in both lines. While rsr1-1 plants accumulated chlorophyll and later on developed true leaves under these conditions, Pat(B33)-Gus seedlings remained white and cotyledons did not expand (Figure 4B). However, combinations of 5 μM ABA with 30 mM Glc or 3 μM ABA with more than 90 mM Glc inhibited development of rsr1-1 seedlings equal to Pat(B33)-Gus (data not shown).

The mutation in rsr1-1 plants had previously been mapped to a region between nga111 and ADH on the long arm of chromosome I (Martin et al., 1997), and was therefore not allelic to any gin or abi mutants that have been characterized at the molecular level (Zhou et al., 1998; Arenas-Huertero et al., 2000; Huijser et al., 2000; Laby et al., 2000; Rook et al., 2001). To identify the RSR1 locus, 18 markers polymorphic between C24 and Col-0 were analyzed in a population of 1315 F2 plants from a cross between rsr1-1 and Col-0. A total of 20 informative recombination events between markers nga111 and f1m20 were detected that placed RSR1 on a 130-kbp fragment between markers f25p22.3 and f2p9.46 (Figure 5A; Table A1 in Appendix). This region contains 40 predicted protein-coding genes, one micro-RNA and one pre-tRNA (TAIR10 genome annotation)¹. Double recombination events within this region indicated a higher probability for a position of RSR1 on the distal half of BAC clone F25P22.

Sub-clones of F25P22 containing 16 out of the 20 predicted genes in this region were tested for complementation of the sugar signaling deficiency of rsr1-1 mutant plants. Complementation was exclusively observed with a 5.6-kb HindIII fragment spanning nucleotides 88982–94626 of the genomic insert of BAC clone F25P22 (Figure 5; Table A2 in Appendix). This fragment contained a single complete gene, At1g73840, encoding for a predicted proline-rich protein of 388 amino acids. Complementation of the rsr phenotype was observed in two independently transformed lines as well as in rsr1-1 plants expressing the cDNA of At1g73840 under the control of a CaMV 35S promoter. In the presence of 90 mM sucrose, strong GUS-staining demonstrated induction of the patatin class I promoter in Pat(B33)-Gus plants, while no GUS-staining was observed in rsr1-1 plants (Figure 5B; Martin et al., 1997). Seedlings from all three complementation lines with either the genomic fragment or with CaMV 35S-driven expression of the cDNA of At1g73840 showed equally strong GUS-staining as the original line Pat(B33)-Gus. Moreover, the genomic fragment and the cDNA also complemented the gin, early flowering and spontaneous lesion formation phenotypes of rsr1-1 (Figures 5C,D and data not shown).

The ORF of At1g73840 contains a single base exchange at position 442 (C to T transition, typical for EMS mutagenesis) in rsr1-1, resulting in a premature stop codon in the fourth out of seven exons (Figure 6A). The Baulcombe lab had previously identified a different mutation in At1g73840 in a screen for genes involved in restricting RNAi-mediated gene silencing of a phytoene desaturase gene, and named this mutant enhanced silencing phenotype 1-1 (esp1-1; Herr et al., 2006). The esp1-1 mutant carries a point mutation 25 codons upstream of the base exchange in rsr1-1, which also leads to a premature stop codon.

To further confirm that the phenotypic changes in rsr1-1 plants are caused by the mutation in At1g73840, several T-DNA insertion lines from the Salk collection (Alonso et al., 2003) were characterized (Figure 6A). The Salk_084710 line carried a T-DNA insertion 54 bp upstream of the ATG of At1g73840 and did not show any alterations in phenotype or RSR1/ESP1 expression in comparison to wildtype plants (data not shown). The Salk_046443 line carried a larger deletion downstream of RSR1/ESP1, which also affects an unknown gene that is essential for embryo development (data not shown). Of the analyzed lines, only the Salk_078793 line, with an insertion in the fifth intron, specifically affected the expression of RSR1/ESP1. Homozygous plants of this line contained a much larger RSR1/ESP1-containing transcript and had a slightly increased organ initiation rate, while they were otherwise phenotypically very similar to wildtype plants under greenhouse conditions (Figure A1 in Appendix). In sterile culture, the Salk_078793 line displayed a decreased Glc sensitivity compared to the parental line Col-8 when grown on MS-medium containing 300 mM Glc, and was therefore named rsr1-2 (Figure 6B). At a Glc concentration of 330 mM, neither rsr1-2 nor Col-8 were able to de-etiolate and establish autotrophic growth (Figure 6C). rsr1-2 plants did not flower earlier than Col-8, which generally flowers earlier than C24 and other late flowering accessions due to allelic variations at the Frigida (Fri) locus (Gazzani et al., 2003; Shindo et al., 2005). Spontaneous lesion formation was not observed in rsr1-2, but this T-DNA insertion mutant was hypersensitive to Pro, similar to rsr1-1 (Figure 6D). In summary, our observations of early flowering, glucose-insensitivity and spontaneous lesion formation in rsr1-1 strongly support that this mutant is affected in At1g73840 and is therefore allelic to esp1-1.

A survey of publicly available microarray data showed that RSR1/ESP1 is expressed at similar levels in all plant tissues and expression does not respond strongly to developmental or environmental stimuli (Genevestigator V.3, Zimmermann et al., 2005). However, a post-translational modification of the RSR1/ESP1 protein was detected by high-resolution proteomics analysis, namely phosphorylation of the threonine residue at position 33 of the predicted protein sequence (PhosPhAt database, release 3.0; Durek et al., 2010). Computational analysis of the RSR1/ESP1 protein sequence did not reveal any well-defined targeting signals. However, SubLoc v1.0 and WolFPSort predicted a slight preference for a nuclear localization of RSR1/ESP1, which is in agreement with its proposed function in mRNA processing (Hua and Sun, 2001; Horton et al., 2007). To verify the computational predictions, the cDNA of RSR1/ESP1 was translationally fused with GFP at its C-terminus and expressed under the control of the CaMV 35S promoter. In both transiently transformed Nicotiana benthamiana leaves and in stable Arabidopsis transformants, GFP fluorescence was most prominent in the nucleus but also detectable in the cytosol (Figures 7A,B).

It had previously been demonstrated that sugar and ABA signaling is mediated by protein phosphorylation events. For example, the Glc-insensitive mutant gin4 affects the protein kinase CTR1 (Zhou et al., 1998), and SNF1-related protein kinases are critical factors in mediating ABA responses (Fujii et al., 2011). To investigate whether protein phosphorylation plays a role in the sugar- and RSR1/ESP1-dependent regulation of the patatin promoter, we applied protein kinase and phosphatase inhibitors alone or in combination with Suc.

After incubation with 200 mM Suc for 30 h, rsr1-1 mutants showed a 1.5-fold increase in GUS-activity, compared to a threefold induction in Pat(B33)-Gus (Figure 8A). Staurosporine (ST), a broad range inhibitor of protein kinases, induced GUS-activity in roots of Pat(B33)-Gus plants in the absence of Suc, whereas okadaic acid (OA), an inhibitor of protein phosphatases, reduced the sugar-dependent response. A combination of ST and Suc led to an additional increase in GUS-activity. In contrast, these responses to altered protein phosphorylation were severely reduced in rsr1-1 demonstrating that phosphorylation of either RSR1/ESP1 or upstream components in the same signaling cascade mediate sugar-dependent regulation of the Pat(B33)-Gus construct (Figure 8B).

Arabidopsis (A. thaliana (L.) Heynh.) lines Pat(B33)-Gus and rsr1-1 (descendent from ecotype C24; Martin et al., 1997) were grown in sterile culture under short-day conditions on MS-media (Murashige and Skoog, 1965) supplemented with different concentrations of Suc, Glc, 3Omeg, sorbitol, Pro, or ABA. Respective concentrations are given in the text or figure legends. Seeds were stratified for 24 h at 4°C in 0.1% agarose. Transformation by floral dip was performed according to (Clough and Bent, 1998).

Three lines (Salk lines 084710, 046443, and 078793) predicted to carry a T-DNA insertion in the RSR1 gene (At1g73840) were obtained from the Nottingham Arabidopsis Stock Center. The insertion sites for all three lines were confirmed by PCR with LB primer (gccctttgacgttggaggccac) and RSR1-specific primers (for 084710: ccacacggattgcagatttag, for 078793: atgttcgagctctctgattgg, and for 046443: cccagcaacctccttttcatc). The insertion sites were further confirmed by sequencing of the PCR products in the case of lines 084710 and 078793 (rsr1-2). All insertion lines were backcrossed three times to the parental line (ABRC accession Col-8) to eliminate secondary mutations.

Glc-tolerance and ABA-insensitivity were tested by germinating seedlings on MS-media supplemented with various concentrations of Glc, 3Omeg, sorbitol, or ABA. Seeds were scored as germinated when the radicle had emerged through the seed coat. Seedling establishment was scored after 6–28 days. For analysis of root growth and root hairs, the plates were placed vertically in a growth cabinet. To test for hypersensitivity to Pro, plants were cultivated on MS-medium with 30 mM Glc and 40 mM Pro (Hellmann et al., 2000). For bolting time analysis, transformation and pigment quantification plants were grown in the greenhouse (≥16 h light; 20°C). Pigments were extracted twice with 80% acetone from fully expanded rosette leaves ground in liquid nitrogen. Chlorophyll A and B were quantified photometrically under dimmed light according to (Lichtenthaler, 1987). For ROS detection, detached leaves were infiltrated by transpiration with 1 mM 3,3′-diaminobenzidine at pH 5.5 in dim light. After 4 h, the leaves were destained in 80% ethanol.

For mapping of the RSR1 locus, F2 plants from a cross between rsr1-1 (C24 ecotype) and Col-0 were assayed for sugar induction of the Pat(B33)-Gus construct using in vivo GUS-staining (see below). Genomic DNA was isolated from 1315 kanamycin-resistant plants and used for PCR-based genotyping in the proximity of marker nga111. Eighteen previously described or novel SSLP or CAPS markers were found to be polymorphic between C24 and Col-0 in the region of interest (see Table A1 in Appendix).

For complementation of the rsr1-1 mutant with genomic fragments, the BAC clone F25P22 (Medline AC012679) was digested with different restriction enzymes and fragments containing single or multiple genes were sub-cloned into pPZP221 (Hajdukiewicz et al., 1994; Table A2 in Appendix). For complementation assays using the RSR1 cDNA, the full length cDNA (RIKEN, clone RAFL 09-93-F21) was cloned between the CaMV 35S promoter and terminator in the EcoRI/BamHI sites of pRT101 (Töpfer et al., 1987). The cassette was excised with HindIII and transferred into a pCAMBIA-derived binary vector conferring phosphinothricin resistance².

To analyze the subcellular localization of RSR1/ESP1, the RSR1 ORF was amplified by PCR from RALF cDNA clone 09-93-F21 using the following primers: ataggtaccatggcgggtaagcagatcgg and cttggatccgctgtcggagaatctgttgaag. The resulting PCR fragment was cloned into the KpnI/BamHI restriction sites of vector pCF203, a derivative of the pPZP212 vector, creating a CaMV 35S-driven C-terminal GFP fusion construct. Localization of GFP expression was analyzed in N. benthamiana transiently transformed by Agrobacterium-mediated transformation, as well as in stable Arabidopsis Col-0 transformants. Cells expressing the RSR1–GFP fusion were imaged using a Nipkow spinning disk confocal microscope as described (Deuschle et al., 2006).

Plants were cultured on solid MS-medium containing 30 mM Glc. Twenty-day-old plants were transferred to glass jars, in which only the roots had contact to liquid medium. Plants were pre-treated for 2 h with 0.4 mM okadaic acid (Calbiochem) or 0.4 mM staurosporine (from a 40 mM stock in DMSO, Sigma) in MS-medium before Suc was added. After 30 h, roots were dissected and frozen in liquid nitrogen. Proteins were extracted from roots in GUS extraction buffer (Jefferson et al., 1987) in the presence of protease inhibitors (Complete, Roche). Fluorimetric GUS-assays were performed with 10 μg total protein by measuring the fluorescence of 4-methyl-umbelliferone as described by (Martin et al., 1997) using a Fluoroskan II fluorimeter (Lab-systems). In vivo GUS-staining and staining of whole seedlings was performed for 18 h at 37°C according to (Martin et al., 1997).