*3.1. The RLCK VI\_A Kinases Are Required for Cell Elongation and Organ Growth in Addition to Their Role in Stress Responses*

The RLCK VI\_A3 kinase of Arabidopsis and its barley homologue, HvRBK1, have been implicated in ROP-GTPase-dependent pathogen resistance/susceptibility reactions [6,7,9], while the AtRLCK VI\_A4/AtRBK1 and AtRLCK VI\_A6/AtRBK2 kinases were shown to be expressed in response to pathogen infection [7], suggesting their primary role in plant defence. A considerable number of DEGs of the *rlck vi\_a2* mutant are also related to plant defence (Table 1), strengthening this view. In addition, the transcript analysis indicated that the kinase might function during abiotic stress responses as well (Table 1). However, the involvement of several RLCK VI\_A kinases in the regulation of plant growth and development has also been reported: the T-DNA mutant of the related *AtRLCK VI\_A3* kinase is stunted and has over-branched trichomes [9]; the Arabidopsis RBK1 protein kinase (RLCK VI\_A4) has been implicated in auxin-responsive cell expansion, due to the reduced auxin responsiveness of its T-DNA insertion mutant [14].

Our results show that the mutant seedlings producing no full length RLCK VI\_A2 kinase (Figure 1) have limited cell expansion as compared to the wild type (Figure 3) resulting in shorter hypocotyls and cotyledons (Figure 2). Independent experiments using RNA interference to reduce *RLCK VI\_A2* expression (Supplementary Figure S2), as well as the complementation of the T-DNA-caused mutation

in transgenic lines (Figure 2) confirmed that the observed phenotypes are indeed associated with the absence of the kinase. Greenhouse-grown mutant plants also exhibited smaller rosette/leaf size than the wild type similarly to the reduced plant size of the related AtRLCK VI\_A3 kinase [9].

Altogether, these observations strengthen the view that RLCK VI\_A members have a general role in cell expansion and plant growth. Interestingly, while mutation in *AtRBK1*/*AtRLCK VI\_A4* resulted in increased auxin sensitivity, in our experiments, the *atrlck vi\_a2* mutant showed no altered auxin response (Supplementary Figure S3), but the mutant phenotypes could be rescued by exogenous GA<sup>3</sup> (Figure 5 and Supplementary Figure S4). This indicates that the various RLCK VI\_A kinases might influence cell expansion via various pathways.

#### *3.2. How the RLCK VI\_A2 Kinase May A*ff*ect Cell Expansion?*

How RLCK VI\_A2 kinase regulate cell expansion is not known at present. It has to be mentioned that the barley HvRBK1/HvRLCK VI\_A3 kinase has been shown to be required for proper cortical microtubule organisation; the silencing of HvRBK1 was shown to result in a fragmented cortical microtubule network [6]. Since cortical microtubules are known to control directional cell elongation [34,35] in a ROP GTPase-dependent manner [36], the ROP-binding kinases might be involved in this process. Transcriptome analysis during the skotomorphogenesis of *rlck vi\_a2* mutant seedlings indicates that high portion of the DEGs modulated in the mutant code for proteins located at membranes, at the cell periphery or in the apoplast, and may have a role in transport or cell wall organisation (Table 1 and Supplementary Tables S2 and S6). Therefore, it is conceivable that the RLCK VI\_A2 kinase has a regulatory role in these processes in relation to cell elongation. The transcriptomic data, however, do not provide a clear view about the role of the kinase in cell elongation. The upregulation of several expansins and xyloglucan endotransglucosylase/hydrolase coding genes implicated in cell wall loosening [37,38] and the downregulation of those coding for extensin-like family proteins rather contributing to cell wall stiffening [39,40] are not consistent with the observed phenotype of the mutant having restricted cell elongation. Moreover, PIF4, a positive regulator of skotomorphogenesis including hypocotyl cell elongation is upregulated in the *rlck vi\_a2* mutant, despite its short-hypocotyl phenotype [30]. These contradictions might be resolved by keeping in mind that the kinase primarily modulates posttranslational and not transcriptional regulation. The observed transcriptional changes might be indirect responses to the missing kinase function: blocking cell elongation at the posttranslational level (e.g., phosphorylation-dependent degradation of PIF4 or other regulators) might give a feedback to increase the transcription of genes promoting cell expansion. For example, PIF4 has been shown to be phosphorylated by the brassinosteroid signalling kinase BRASSINOSTEROIDINSENSITIVE 2 (BIN2), marking it for proteasomal degradation [41]. Considering the number of genes affected by the *rlck vi\_a2* mutation, and being at the same time the direct targets of the PIF4/BZR1/ARF6 transcriptional factor circuit that centrally controls cell expansion and hypocotyl growth (Figure 7 and Supplementary Table S5), one can suppose that the kinase directly and/or indirectly modulates the downstream processes controlled by these factors.

#### *3.3. Gibberellin might Indirectly Complement for the Missing Kinase Function*

Although exogenous GA<sup>3</sup> treatment could rectify the absence of RLCK VI\_A2 function (Figure 5 and Supplementary Figure S4), the kinase mutant exhibited similar bioactive GA levels than the wild type (Figure 6a). Moreover, there was no decrease in the GA sensitivity of the mutant (Figure 6b) despite a ca. 1.5-fold increase in the expression of the GRAS-domain GAI protein gene (Table 2, Supplementary Figures S3 and S5a), a negative regulator of GA signalling. Transcriptomic analysis confirmed that genes implicated in gibberellin metabolism are hardly affected by *RLCK VI\_A2* expression. However, the same analysis indicated the misregulation of 15% of *RLCK VI\_A2*-dependent DEGs also in the GA synthesis defective mutant *ga1-3* [28,29] (Figure 7a and Supplementary Table S4). Furthermore, the RLCK VI\_A2 kinase might be involved in the modulation of GA transport, since the expression of four potential GA transporters [25,26,28] was found to be regulated in the mutant background (Table 2).

The above observations indicate that, although RLCK VI\_A2 functions interfere with GA action, this is likely not through the direct modulation of the synthesis of bioactive GAs. Transcriptomic analysis revealed that the absence of the RLCK VI\_A2 kinase affected the signalling of auxin and BR, the two other hormones also centrally involved in the light and developmental regulation of hypocotyl elongation [31]. Both the auxin and BR hormones are well known to crosstalk with GA modulating each other's metabolism and signalling [37–39]. Although all three hormones (GA, BR and auxin) act on distinct TFs governing hypocotyl elongation (PIF4, ARF6, and BZR1, respectively), the target genes of these transcription factors largely overlap [31] (Figure 7b). A considerable fraction of the DEGs of the *rlck vi\_a2* mutant (31.6%) are direct targets of at least one of the above TFs, while 10% of the DEGs is direct target of all three. The observed gene expression changes might be indirect consequences of blocked cell expansion/seedling growth. The data support the view that the RLCK VI\_A2 kinase might control basic cell elongation processes downstream of the PIF4/BZR1/ARF6 TFs, rather than the regulatory proteins themselves.

Why exogenous GA3, but not auxin or brassinosteroid, rescue the *rlckvi\_a2* mutant phenotypes is not known. Exogenous GA<sup>3</sup> might induce parallel pathways that can overcome the cell elongation defects caused by missing protein phosphorylations in the absence of the RLCK VI\_A2 kinase (e.g., via other kinases and/or microtubule/protein stability and/or cell wall organisation, etc.). One of such shared pathways is the regulation of auxin transport, since several auxin transport protein genes are regulated in the mutant background (Supplementary Table S3), and GA is known to affect auxin transport stabilizing these proteins [42].

#### **4. Materials and Methods**

#### *4.1. Plant Material and Growth Conditions*

Seeds of *Arabidopsis thaliana* (L.) Columbia-0 and mutant lines of the GABI-Kat Arabidopsis T-DNA insertional mutant collection [15], GABI\_435H03 and GABI\_676D12, were obtained from the Nottingham Arabidopsis Stock Centre (NASC).

The full-length cDNA of At2G18890 was obtained from the Arabidopsis Biological Resource Centre (ABRC, Columbus, OH, USA; stock number U67191). The cDNA was amplified (denaturation 94 ◦C 10 s, annealing temperature 62 ◦C for 30 s, elongation 1 min at 72 ◦C) by the proof-reading PHUSION™ II polymerase (Thermo Fisher Scientific, Waltham, MA, USA), using specific primers, having added EcoRI and XhoI sites (Supplementary Table S7). The gel-purified PCR fragment was digested by FastDigest™ EcoRI and XhoI enzymes (Thermo Fisher Scientific) and inserted into similarly cut and purified pENTR2B vector (Thermo Fisher Scientific). The cDNA was cloned into the plant expression vectors pN-TAPa (Gene bank accession: AY788908 [17]) and pMDC7 [18,43] via Gateway recombination, using standard LR Clonase™ II (Thermo Fisher Scientific, catalogue number: 11791019) reaction, as recommended by the supplier. The binary vectors were transformed into GV3101/pMP90 Agrobacterium strain with tri-parental mating [44], which were used for transgenic *Arabidopsis thaliana* (GABI\_435H03 and/or Col-0) production via floral dip agroinfiltration [45]. Seeds were selected using appropriate antibiotics (gentamycin and hygromycin, respectively). Plants (including the T-DNA insertion mutants) were characterized for *RLCK VI\_A2* (AT2G18890) expression, using the same primers as for cloning in reverse transcription polymerase chain rection (RT-PCR) (see later). Transgenic lines with appropriate expression were selfed and propagated. Stable, homozygous T3/T4 transgenic plants were used in the experiments.

Seeds of the wild type, mutant and transgenic lines were sterilized in 2% bleach, resuspended in sterile water and stratified (4 ◦C for 48 h). Germination was performed in vertically oriented square Petri dishes with half strength Murashige and Skoog (MS) medium containing 0.5% sucrose 0.8% agar, pH 5.7 (Duchefa Biochemie, Haarlem, The Netherlands). When estradiol inducible lines were used, the growth medium contained 5 µM β estradiol (E2758, Sigma, St. Louis, Mo, USA) for gene expression induction. The experiments were done in growth chambers (Aralab, Rio de Mouro, Portugal) under

short days (8 h 22 ◦C, 120 µE light intensity and 16 h 21 ◦C darkness); in complete darkness at 22 ◦C to investigate skotomorphogenesis; or in continuous low white light at 60 µE for the gibberellin sensitivity assay.

#### *4.2. Analysis of Hypocotyl Length and Rosette Size Measurement*

The hypocotyl and cotyledon lengths were measured from digital photographs by the ImageJ software (NIH, Bethesda, MD, USA). At least 60 wild-type and mutant seedlings were analysed in three biological replicates. For the determination of plant rosette size, 28 or 33-days-old Arabidopsis plants grown in pots were photographed, and rosette diameters were measured with the "straight line" function of the ImageJ software [46]. The distance between the tips of the two longest rosette leaves were measured and exported to MS Excel file for further analysis. Images were taken with a digital camera (Fuji FinePix S1000fd) using the same parameters (focus distance, resolution, ISO).

#### *4.3. Cell Size and Shape Analysis*

The analysis of hypocotyl cell length was done after Acridine orange (100 µg/mL) staining. Images on the three different part of the hypocotyls (basal, middle, top) were taken by confocal laser scanning microscopy (TCS SP5, Leica Microsystems, Heidelberg, Germany), and the cell length was determined using the ImageJ software [46].

Cotyledon epidermal cell size and shape were visualized uncoated in a JSM-7100F/LV scanning electron microscope (JEOL Ltd., Akishima, Tokyo, Japan) at low-vacuum, by detecting backscattered electrons according to [47]. The images were taken from the same zone of the cotyledons (at the centre, next to the main vein) and perimeter, area, and circularity of cells were analysed using the ImageJ software [46].

Altogether, 150–200 cells were measured per line, in three repetitions.
