1 An atypical aspartic protease modulates lateral root development in Arabidopsis 2 thaliana 3 4 André Soares1,2,3,4#, Stefan Niedermaier5, Rosário Faro3, Andreas Loos6§, Bruno

5 Manadas3, Carlos Faro3, Pitter F. Huesgen5*, Alice Y. Cheung4*, Isaura Simões2,3*, 6 1 7 PhD Programme in Experimental Biology and Biomedicine, Center for Neuroscience 8 and Cell Biology, University of Coimbra, Portugal; 2 9 Institute for Interdisciplinary Research, University of Coimbra, Portugal; 3 10 CNC-Center for Neuroscience and Cell Biology, University of Coimbra; 4 11 Department of Biochemistry and Molecular Biology, University of Massachusetts, 12 Amherst, MA 01003, USA; 5 13 Central Institute for Engineering, Electronics and Analytics, ZEA-3, 14 Forschungszentrum Jülich, Jülich, Germany; 15 6Department of Applied Genetics and Cell Biology, University of Natural Resources 16 and Life Sciences, Vienna, Austria; 17 18 #Current address: FairJourney Biologics, Porto, Portugal 19 §Current address: Aridis Pharmaceuticals, 5941 Optical Court, CA 95138, US.

20 *Corresponding authors 21 22 E-mail addresses: AS ([email protected]); SN ([email protected]); 23 RF ([email protected]); AL ([email protected]); BM 24 ([email protected]); CF ([email protected]); PFH ([email protected]); 25 AYC ([email protected]); IS ([email protected]) Phone: +351 231 26 249196 27 28 This is a pre-copyedited, author-produced version of the article accepted for publication 29 in the Journal of Experimental Botany following peer review. The version of record is 30 available online at: https://doi.org/10.1093/jxb/erz059 31 32

1 33 Running title: 34 An atypical AP in lateral root development 35 36 Highlight 37 Aspartic protease (AP) ASPR1 displays atypical enzymatic properties, including fungal- 38 like specificity preferences. A role in lateral root development is anticipated. This work 39 further strengthens a sophisticated specialization among atypical APs. 40 41 Abstract 42 Few atypical aspartic proteases (APs) present in plants have been functionally studied to 43 date despite having been implicated in developmental processes and stress responses, 44 and diverse biochemical properties were reported for these enzymes. Here we 45 characterize a novel atypical AP that we name ATYPICAL ASPARTIC PROTEASE IN 46 ROOTS 1 (ASPR1), denoting its expression in Arabidopsis roots. Recombinant ASPR1 47 produced by transient expression in N. benthamiana was active and displayed atypical 48 properties, combining optimum acidic pH, partial sensitivity to pepstatin and 49 pronounced sensitivity to redox agents, and unique specificity preferences resembling 50 those of fungal APs. ASPR1 overexpression suppressed primary root growth and lateral 51 root development, implicating a previously unknown biological role for an AP. 52 Quantitative comparison of WT and aspr1 root proteomes revealed deregulation of 53 proteins associated with both ROS and auxin homeostasis in the mutant. Together, our 54 findings on ASPR1 reinforce the varied pattern of enzymatic properties and biological 55 roles of atypical APs and raise exciting questions on how these distinctive features 56 impact functional specialization among these proteases. 57 58 Keywords 59 Arabidopsis thaliana; ASPR1; Atypical aspartic protease; auxin; lateral root; 60 magnICON expression system; root development; ROS 61 62 Abbreviations: AP, aspartic protease; LR, lateral root

2

63 Introduction 64 Aspartic proteases (APs) represent the second-largest class of plant proteases 65 (van der Hoorn, 2008). Members of the pepsin-like family (A1) are widely distributed 66 in a variety of plants (Simões and Faro, 2004), with a large representation of genes 67 encoding putative A1 members found in the genomes of Arabidopsis thaliana, Oryza 68 sativa and Vitis vinifera L. (Chen et al., 2009; Faro and Gal, 2005; Guo et al., 2013). 69 Depending on their domain organizations and active site sequence motifs, plant A1 70 family members are generally grouped into three classes – typical, atypical and 71 nucellin-like (Faro and Gal, 2005). Typical APs have been extensively studied and are 72 characterized by the presence of a signal peptide, a prosegment and an extra protein 73 domain known as the plant-specific insert, which is absent in the mature form of the 74 enzymes. These proteases are active at acidic pH, are specifically inhibited by pepstatin 75 A and are mostly localized to the vacuole (Simões and Faro, 2004). Strikingly, the vast 76 majority of putative plant APs are grouped as atypical and nucellin-like, exhibiting 77 distinct features when compared with typical APs (Faro and Gal, 2005). These features 78 include the absence of the plant-specific insert; an unusually high number of cysteine 79 residues; the nature of the amino acids preceding the first catalytic aspartate; and 80 unexpected localizations such as the chloroplast, mitochondria or endoplasmic 81 reticulum/extracellular (Gao et al., 2017; Paparelli et al., 2012; Phan et al., 2011). 82 Although studies on atypical APs are still nascent, unusual enzymatic properties have 83 already been documented for several of these enzymes, including more diverse optimal 84 pHs (less-acidic), very broad sensitivity to pepstatin A, proteolytic activity without 85 prosegment removal, as well as distinct specificities (Olivares et al., 2011; Paparelli et 86 al., 2012; Prasad et al., 2010; Simões et al., 2007; Yao et al., 2012). Therefore, the large 87 number of atypical and nucellin-like APs in plants is suggestive of greater 88 diversification of protein functions and a more regulatory role for these APs, as 89 compared to the housekeeping function generally attributed to typical plant APs 90 (Simões and Faro, 2004). Indeed, some functions are starting to be uncovered for these 91 atypical APs, with proposed roles in important and highly regulated processes. For 92 instance, several APs have been demonstrated to be important for responses to biotic 93 (Alam et al., 2014; Breitenbach et al., 2014; Li et al., 2016; Prasad et al., 2009; Xia et 94 al., 2004) and abiotic stresses (Yao et al., 2012). APs are also involved in a range of 95 other functions, including chloroplast metabolism (Nakano et al., 1997; Nakano et al., 96 2003; Paparelli et al., 2012), hybrid sterility (Chen et al., 2008; Ji et al., 2012), and

3

97 reproductive development (Chen and Foolad, 1997; Gao et al., 2017; Ge et al., 2005; 98 Huang et al., 2013; Niu et al., 2013; Phan et al., 2011). These findings implicate 99 functional specialization of plant APs and suggest tight activity regulation. 100 Root developmental processes depend on the complex interplay between various 101 transcriptional regulators, multiple hormones as well as other signaling molecules such 102 as reactive oxygen species (ROS). Auxin plays a central role in almost every aspect, 103 including the integration of developmental cues with changes in environmental 104 conditions (Du and Scheres, 2018; Kazan, 2013; Lavenus et al., 2013; Malekpoor 105 Mansoorkhani et al., 2014; Manzano et al., 2014; Peret et al., 2009; Petricka et al., 106 2012; Tognetti et al., 2012; Tsukagoshi, 2016). Proteolysis is an important post- 107 translational modification process known to modulate different plant developmental and 108 environmental cues (Schwechheimer and Schwager, 2004; van der Hoorn, 2008). For 109 example, during root development, the ubiquitin-26S proteasome system (UPS) 110 regulates auxin signaling through the degradation of Aux/IAA transcriptional repressors 111 that inhibit the activity of the transcription factors AUXIN RESPONSIVE FACTORS 112 (ARFs) (Rogg and Bartel, 2001; Schwechheimer and Schwager, 2004; Wang and 113 Estelle, 2014). Plant proteases are also critical regulators of different biological 114 processes. Although their role in root development remains largely unclear (van der 115 Hoorn, 2008), there are several examples of proteases that have been implicated in this 116 process, such as the putative zinc carboxypeptidase SOL1 (Casamitjana-Martinez et al., 117 2003), subtilase SBT6.1 (Ghorbani et al., 2016), the caspase family member separase 118 (Wu et al., 2010), and the membrane-associated aminopeptidase M1 (APM1) (Peer et 119 al., 2009) from Arabidopsis, and the rice sumo protease OTS1 (Srivastava et al., 2016). 120 Other proteases described in roots are linked to nutrient acquisition in response to 121 abiotic stress conditions or symbiotic or pathogenic relationships (Kohli et al., 2012). 122 From the atypical APs implicated in reproduction, the Arabidopsis PCS1 (Ge et 123 al., 2005), UNDEAD (Phan et al., 2011), A36 and A39 (Gao et al., 2017), and rice 124 OsAP25 and OsAP37 (Niu et al., 2013), and OsAP65 (Huang et al., 2013) have been 125 shown to modulate programmed cell death (PCD) events that determine the 126 development of male gametophyte. When evaluating the function of the transcription 127 factor MALE STERILITY 1, whose absence resulted in a severe male sterility phenotype, 128 Ito and co-workers identified a gene encoding another atypical AP significantly 129 repressed in ms1 plants (At2g03200), suggesting its possible function during male 130 reproductive development (Ito et al., 2007).

4

131 In this work, we explore the biochemical properties and functional role of this 132 novel putative atypical AP (herein named ASPR1, for Atypical Aspartic Protease in 133 Roots 1).We demonstrate that ASPR1 is indeed an active AP displaying a set of atypical 134 enzymatic properties. Moreover, overexpression mutants of ASPR1 display shorter 135 primary roots and a pronounced reduction in the number of lateral roots (LRs). ASPR1 136 emerges as the first AP with an anticipated role in the regulation of root development in 137 Arabidopsis.

5

138 Materials and methods 139 Generation of A. thaliana transgenic lines 140 Different transgenic lines were generated: oxASPR1 (At2g03200) 141 overexpression lines (35S::ASPR1-GFP) and ASPR1 promoter::GUS transgenic line 142 (pASPR1::GUS). Construct design is detailed in Supplementary Materials and Methods. 143 Plasmid DNA constructs 35S::ASPR1-GFP and pASPR1::GUS were independently 144 transformed into A. thaliana using the floral dip method (Clough and Bent, 1998) 145 (Supplementary Materials and Methods). 146 147 Heterologous expression of ASPR1 in N. benthamiana 148 The sequence encoding proASPR1 without the first 27 amino acids was 149 amplified (see Supplementary Fig. S2A) by PCR using the primers 7 and 8 150 (Supplementary Table S1), designed to introduce a BsmBI restriction site at both ends, 151 and a linker sequence and the HisTag coding sequence (at the 3’ end). The amplified 152 product was then inserted into the magnICON® vector under the control of TMVα 153 promoter, using the restriction enzyme BsmBI (construct TMVα::ASPR1-His). The 154 construct was confirmed by DNA sequencing. The magnICON® expression vector 155 carrying TMVα::ASPR1-His was transformed into A. tumefaciens strain GV3101 by a 156 combined heat and cold treatment. Positive clones were selected by colony PCR using 157 the primers 9 and 10 (Supplementary Table S1). Cells were grown overnight at 30°C, 158 harvested by gentle centrifugation at 3200g for 15 min at RT and resuspended in buffer

159 10 mM MES pH 5.6, 10 mM MgSO4, supplemented with acetosyringone to a final

160 OD600 of 0.2. Cell suspension was then used to infiltrate leaves of 5-weeks-old WT N. 161 benthamiana grown at 22°C under a 16 h/8 h light/dark cycle. Routinely, 2-3 leaves 162 were infiltrated per plant. Infiltrated leaves were collected after 3 days, immediately 163 frozen in liquid nitrogen and stored at -80°C. 164 165 Extraction and purification of recombinant ASPR1 (rASPR1) from N. benthamiana 166 leaves 167 Total soluble protein (TSP) was extracted from 15-20 g of frozen leaf material. 168 Plant material was ground under liquid nitrogen and resuspended in extraction buffer 169 (45 mM Tris.HCl buffer pH 7.4, 1.5 M NaCl, 1 mM EDTA, 40 mM ascorbic acid). The 170 extract was incubated on ice for 15 min, centrifuged at 104000g during 20 min at 4°C, 171 and TSP collected. TSP extract was filtered and loaded on a HisTrap HP 1 mL column

6

172 (GE Healthcare Life Sciences, Pittsburgh, PA, USA) pre-equilibrated with HisTrap 173 binding buffer (20 mM Tris.HCl buffer pH 7.5, 500 mM NaCl and 10 mM imidazole). 174 After extensive washing with binding buffer, protein elution was performed with an 175 imidazole gradient (50 mM, 100 mM and 500 mM) in the same buffer. Eluted fractions 176 were pooled together and buffer exchanged into 20 mM Tris.HCl buffer pH 7.5 by 177 dialysis at 4°C. Samples were then applied onto a Mono Q 5/50 GL column (GE 178 Healthcare Life Sciences, Pittsburgh, PA, USA), previously equilibrated with the same 179 buffer, and protein elution was carried out with a linear gradient of NaCl (0-1 M) in 20 180 mM Tris.HCl buffer pH 7.5. All chromatographic steps were performed in the FPLC 181 system BioLogic DuoFlow (BioRad, Hercules, CA, USA). 182 183 SDS-PAGE and Western blotting 184 Protein samples were separated by SDS-PAGE (12.5%) and either stained or 185 immunodetected as detailed in Supplementary Materials. 186 187 Enzymatic activity assays 188 The proteolytic activity of rASPR1 was determined by fluorescence assays using 189 the fluorogenic substrate [7-methoxycoumarin-4-yl)acetyl(MCA)-Lys]Leu-His-Pro- 190 Glu-Val-Leu-Phe-Val-Leu-Glu[Lys-2,4-Dinitrophenyl (DNP)] (GenScript, Piscataway, 191 NJ, USA) at a final concentration of 8.18 µM. The fluorescence was measured using 192 FluoroMax-3® spectrofluorimeter (Horiba, Kyoto, Japan) (excitation and emission 193 wavelengths of 328 nm and 393 nm, respectively). To assess activity at different pH 194 values, buffers ranging between pH 4 and pH 6 (50 mM sodium acetate pH 4, 4.5, 5, 5.5 195 and 6, supplemented with 100 mM NaCl) were used at 25°C. To determine activity in 196 the presence of different inhibitors (concentrations in Supplementary Table S2), 197 rASPR1 was pre-incubated with each compound for 2 min at room temperature, in 50 198 mM sodium acetate buffer pH 4, supplemented with 100 mM NaCl, before measuring 199 proteolytic activity at 25°C. The pull-down assay was performed using His-Mag 200 Sepharose™ Ni2+ magnetic beads (Ge Healthcare Life Sciences, Pittsburgh, PA, USA) 201 according to the manufacturer’s instructions. Briefly, purified rASPR1 was incubated 202 with the beads for 2 h, at 4°C. The non-bound fraction was collected for subsequent 203 analysis and after a washing step (20 mM sodium phosphate buffer pH 7.4, 500 mM 204 NaCl, 10 mM imidazole), protein was eluted with 20 mM sodium phosphate buffer pH

7

205 7.4, 500 mM NaCl, 500 mM imidazole. Proteolytic activity was checked in both initial 206 and non-bound fractions. All fractions were analyzed by Western blot. 207 208 Endoglycosidase digestion 209 Assays with EndoH and PNGase F (New England Biolabs, Ipswich MA, USA) 210 were performed according to the manufacturer’s instructions, with minor changes 211 (detailed in Supplementary material). 212 213 Proteomics identification of protease cleavage sites (PICS) 214 The specificity profiling of rASPR1 was determined using the Proteomics 215 Identification of Cleavage Sites (PICS) methodology (Schilling and Overall, 2008), 216 with some minor modifications. Tryptic and GluC peptide libraries were generated from 217 THP1 cells. rASPR1 was incubated with each peptide library at a ratio of 1:20 (w/w), 218 for 5 h at 25°C, in 50 mM sodium acetate buffer pH 4 supplemented with 100 mM 219 NaCl. The purification of the carboxy-terminal peptide cleavage products and the 220 analysis of samples were performed as described in Almeida et al. (Almeida et al., 221 2017) (detailed in Supplementary material). 222 223 A. thaliana material and growth conditions 224 A. thaliana ecotype Col-0 seeds were used in all experiments. The T-DNA 225 insertion lines Salk_038980.40.95.x (aspr1), Salk_072864.55.25.x and 226 WscDsLoxHs079_01H.0) were obtained from the Arabidopsis Biological Resource 227 Center (ABRC) (The Ohio State University, Columbus, OH, USA). Seeds were surface 228 sterilized, sown either in Gamborg media (B5) (Phyto Technology Laboratories, 229 Lenexa, KS, USA) (for seedlings transferred to soil), or half-strength Murashige and 230 Skoog media (1/2 MS) (Phyto Technology Laboratories, Lenexa, KS, USA) and 231 modified 1/2 MS media without ammonium nitrate (Phyto Technology Laboratories, 232 Lenexa, KS, USA) (for seedlings used for root analysis), both supplemented with 1% 233 sucrose and 0.7% agar, and incubated in the dark at 4°C for 48 h. Plates were then 234 transferred to a growth chamber and seeds were allowed to germinate at 22°C under 16 235 h/8 h light/dark cycles. 10-d-old tissue culture-grown seedlings were transferred to soil 236 and maintained in a growth chamber at 22°C under a 16 h/8 h light/dark cycle. 237 238 Histochemical GUS assays

8

239 Arabidopsis lines expressing GUS under the ASPR1 promoter (pASPR1::GUS) 240 were generated as described in Supplementary Materials and Methods. Whole 241 inflorescences and 13-d-old seedlings were collected and incubated in GUS staining

242 solution (0.1 M NaPO4 buffer pH 7, 0.5 mM K3[Fe(CN)6], 0.5 mM K4[Fe(CN)6].3H2O,

243 10 mM Na2EDTA) supplemented with 0.2 mg/mL of X-Gluc. The GUS reaction was 244 allowed to develop for 12-16 h, at 37°C. After staining, plant material was cleared in 245 70% ethanol and imaged (Olympus SZ61). 246 247 Protoplast transfection 248 Protoplasts were isolated from 3-week-old tissue Arabidopsis seedlings and 249 transfected as described elsewhere (Tao et al., 2002). Typically 0.1 mL of cells were

250 transfected with 10 µg of each reporter gene-fusion construct (ASPR1(1-39)-GFP; 251 ASPR1-GFP). Construct design is detailed in Supplementary Materials and Methods. 252 Protoplasts were incubated for 16 h in the dark, at room temperature, and protein 253 localization was observed by epifluorescence microscopy (Nikon Eclipse E800). 254 255 Root growth assays 256 Seedlings grown as described above were monitored for 10 days. Root length 257 was marked at days 3, 5, 7 and 10, and the number of LRs was determined at the end of 258 the experiment. Plates were imaged and root length measured using ImageJ 259 (http://imagej.nih.gov/ij/). In total, 100-150 seeds per WT and aspr1 lines, and 30 seeds 260 per OxASPR1 line were used in each experiment. Two independent replicate 261 experiments were performed. Statistical analysis was performed by the ANOVA test 262 using the Prism Software (GraphPad Software Inc). The number of LRs was determined 263 at the end of the experiment. For these growth-related analyses, experiments were 264 repeated twice with a total of 40 to 100 seeds from different seed batches with 265 comparable results. 266 267 Quantitative root proteome comparison 268 Proteins were extracted from frozen root tissue by mechanical grinding, reduced, 269 alkylated and digested with trypsin using standard procedures. Tryptic peptides were 270 differential stable isotope labeled by reductive dimethylation (Boersema et al., 2009). 271 500 ng desalted peptides were analyzed by nano-LC-ESI-MS/MS (as detailed in 272 Supplementary Materials and Methods).

9

273 274 Data availability 275 The mass spectrometry proteomics data have been deposited to the 276 ProteomeXchange Consortium via the PRIDE (Vizcaino et al., 2016) partner 277 repository, with the identifier PXD009580 for the root shotgun proteome dataset 278 (reviewer account username: [email protected]; Password: uccsHeVf) and 279 PXD010355 for the PICS dataset (reviewer account username: 280 [email protected]; Password: LJqSx9Pa).

10

281 Results 282 rASPR1 displays AP activity with atypical biochemical properties 283 To evaluate whether At2g03200/ASPR1 encodes an active AP, we first 284 attempted heterologous production in E. coli. Due to its high number of cysteine 285 residues, several strategies were tried to produce the protein in its soluble form. 286 However, in all approaches rASPR1 accumulated mostly in the insoluble fraction 287 (Supplementary Fig. S1). Since ASPR1 is also predicted to be glycosylated, we then 288 decided to express the protease in a plant expression host. We chose Nicotiana 289 benthamiana and the magnICON® expression system, which consists of a highly 290 engineered viral-based vector that promotes the transient expression of a protein of 291 interest in leaves (Marillonnet et al., 2005). The coding sequence of ASPR1 without the 292 putative signal peptide was fused in frame with a C-terminal His-Tag and subcloned 293 into the magnICON® expression vector under the control of TMVα promoter 294 (Supplementary Fig. S2A). Three days after Agrobacterium-mediated infiltration of the 295 viral-based vector, N. benthamiana leaves were harvested for protein purification. A 296 purification protocol was optimized, consisting in the extraction of the total soluble 297 protein fraction, its application to a His-Trap HP column, followed by further 298 purification by anion exchange chromatography (Fig. 1A, Supplementary Fig. S2B, C). 299 SDS-PAGE and Western blot analysis of Mono Q eluates confirmed the purification of 300 rASPR1, with one of the fractions accumulating mainly a His-Tag positive band above 301 50 kDa and a second fraction where a slightly lower molecular weight band was also 302 detected (Fig. 1B). N-terminal Edman sequencing was performed to both 303 immunoreactive bands. The upper band matches the N-terminal sequence Ile32-Asp- 304 Arg-Thr (rASPR1 full-length numbering) (corresponding to rASPR1 without signal 305 peptide), whereas the N terminus of the lower molecular weight band corresponds to a 306 processed form of rASPR1 without most of the putative prosegment region, matching 307 the sequence Ala85-Val-Ala-Ser (Supplementary Fig. S2A). Both fractions were 308 screened for proteolytic activity against several synthetic fluorogenic substrates. 309 However, activity was observed only towards [MCA-Lys]Leu-His-Pro-Glu-Val-Leu- 310 Phe-Val-Leu-Glu[Lys-DNP], the substrate specifically designed in our laboratory for 311 the biochemical characterization of the atypical AP CDR1 (Simões et al., 2007). The 312 proteolytic activity of rASPR1 was further confirmed with a pull-down assay with Ni2+ 313 magnetic beads, taking advantage of its C-terminal His-Tag. The protease was pulled

11

314 down, and activity towards the fluorogenic substrate was detected in the input but not in 315 the non-bound sample (Fig. 1C,D). 316 We then proceeded to rASPR1 biochemical characterization using the fraction 317 enriched in the high molecular weight band (rASPR1 without signal peptide; from now 318 on, rASPR1 will refer to this form). Treatment with EndoH and PNGaseF confirmed 319 that rASPR1 is indeed glycosylated (Fig. 2A). When tested for enzymatic activity at 320 different pH values (Fig. 2B), rASPR1 displayed its maximum activity at pH 4 towards 321 the same fluorogenic substrate and no activity was detected at pH 6. At pH values 322 below 4, the activity was very unstable. We then evaluated the effect of a broad range of 323 compounds on proteolytic activity (Fig. 2C and Supplementary Fig. S3). rASPR1 was 324 significantly but not completely inhibited by pepstatin A, the canonical inhibitor of APs. 325 Moreover, EDTA and pefabloc (a serine protease inhibitor) also partially inhibited 326 rASPR1 activity. Interestingly, protease activity was also strongly inhibited by redox 327 agents such as NADP and NADPH, with the latter completely inhibiting protease 328 activity. Other redox reagents, like NAD and oxidized glutathione (GSSG), also 329 resulted in about 50% reduction in activity, whereas nucleotides and ions did not 330 significantly affect activity (Supplementary Fig. S3). 331 In summary, our results demonstrate that rASPR1 shares an optimum pH with 332 typical APs but most of its enzymatic features are similar to those observed in other 333 atypical members of this family of proteases. 334 335 rASPR1 displays atypical amino acid preferences 336 To further explore the specificity preferences of rASPR1 we determined P4-P4’ 337 specificity profiles using the Proteomics Identification of Cleavage Sites (PICS) 338 approach (Schilling and Overall, 2008; Schilling et al., 2011). Two peptide libraries 339 derived from the digestion of human THP-1 cell protein extracts with either Trypsin or 340 GluC were used for profiling rASPR1 amino acid preferences. Both peptide libraries 341 were independently incubated with purified rASPR1, and it was possible to identify 342 1435 C-terminal cleavage products from the trypsin library and 729 C-terminal cleavage 343 products from the GluC library. These results were then analyzed using the WebPICS 344 tool (Schilling et al., 2011) where the corresponding non-prime side sequences were 345 obtained. The complete cleavage specificity profiles showed a good agreement between 346 the results obtained from these two peptide libraries (Fig. 3). At the P1 position, directly 347 preceding the cleaved peptide bond (Schechter and Berger, 2012), rASPR1 strongly

12

348 preferred leucine as well as aromatic amino acids like phenylalanine and tyrosine. 349 Interestingly, the neutral amino acid asparagine and positively charged lysine were also 350 accommodated at this position. Branched aliphatic amino acids like isoleucine, valine 351 and alanine, neutral amino acids (serine, threonine, glutamine), as well as tryptophan, 352 glycine and the positively charged histidine were all underrepresented, whereas proline 353 was never found at this position. On the C-terminal side of the scissile bond in P1’, a 354 strong preference for phenylalanine and tyrosine was observed although the positively 355 charged histidine was also shown to be overrepresented. Proline was again 356 underrepresented at this position and leucine was slightly disfavored. 357 Interestingly, and despite the preference for hydrophobic residues like 358 isoleucine, valine and carboxyamidomethylated cysteine (modified during library 359 preparation) at P2, the charged amino acids lysine and glutamate were also strikingly 360 overrepresented at this position. Proline was underrepresented at P2, together with 361 tryptophan, tyrosine, leucine, and alanine. At P2’, the preference for accommodating 362 hydrophobic amino acids (valine, alanine, carboxyamidomethylated cysteine) was 363 maintained, although other aliphatic amino acids like methionine and leucine, as well as 364 tryptophan, aspartate and lysine were disfavored at this position. 365 At positions more distant from the cleavage site (P3, P4, P3’ and P4’), rASPR1 366 appears to be less stringent. At P3, rASPR1 showed a slight preference for the 367 hydrophobic amino acids phenylalanine, tyrosine, and modified cysteine but, as 368 previously shown for P2, there were also charged amino acids overrepresented in this 369 position (in this case, histidine and glutamate). At P3’, there was a slight preference for 370 alanine, glycine, lysine, and asparagine and there were no major underrepresented 371 amino acids. Interestingly, at P4´ position preference for proline and asparagine was 372 also observed. 373 374 ASPR1 is expressed in root tissue as well as in mature pollen grains 375 To study the expression pattern of ASPR1, A. thaliana transgenic lines 376 pASPR1::GUS were generated. Whole inflorescences and 13-d-old seedlings of five 377 different transgenic lines were analyzed for GUS staining. The expression of the gene 378 was found to change in developing flowers, being absent in immature stages and mostly 379 observed in pollen grains of mature flowers (Fig. 4A-C). The gene was also expressed 380 in vegetative tissues, with apparently higher expression levels in roots, more specifically 381 at primary roots, LR primordia as well as LR tips (Fig. 4D-I). RT-PCR using specific

13

382 primers for ASPR1 in root and whole inflorescence RNA samples extracted from WT 383 Col-0 plants also confirmed these GUS expression observations (Fig. 4J). Furthermore, 384 the subcellular localization of ASPR1 was assessed in transfected protoplasts using two

385 constructs: ASPR1(1-39)-GFP comprising the first 39 amino acids of ASPR1 (putative 386 signal peptide plus 12 additional amino acids of ASPR1 N terminus), and ASPR1-GFP 387 corresponding to the full-length protein. Both fusion proteins showed a similar 388 distribution that suggested a localization in the ER, although we cannot exclude that it 389 may also reflect protein in transit for secretion (Supplementary Fig. S4). 390 391 ASPR1 affects primary root growth and lateral root number under normal growth 392 conditions 393 To gain insights into the physiological function of this atypical AP, we focused 394 further phenotypical analysis to the root system, given the apparent expression of 395 ASPR1 in this organ. We generated overexpression lines with the ASPR1 gene under 396 the control of the 35S constitutive promoter (35S::ASPR1-GFP) (Supplementary Fig. 397 S5), referred to as oxASPR1 from hereon. The root lengths of two independent lines of 398 oxASPR1 (oxASPR1-2 and oxASPR1-4) were compared with those of WT plants (Fig. 399 5). When grown on vertical plates in ½ MS media (Fig. 5A), the primary roots of 10- 400 day-old oxASPR1-2 and oxASPR1-4 seedlings were shorter than those of the WT. 401 Monitoring growth of these seedlings over time revealed significantly shorter primary 402 roots in both oxASPR1 lines compared to WT from day 5 with increasing difference 403 over time (Fig. 5B). At 10 days, the relative root length of oxASPR1-2, and oxASPR1-4 404 seedlings was 86.3% (+/- 15.8%) and 74.3% (+/-15.9%) of WT, respectively (Fig. 5C, 405 D). Given the observed expression of ASPR1 in LR tips and LR initiation sites (Fig. 4), 406 we counted the number of LRs per cm of primary root at day 10. Overexpression of 407 ASPR1 resulted in a significant reduction in the number of LRs compared to WT (Fig. 408 5E, F). oxASPR1-2 and oxASPR1-4 seedlings displayed only 32.9% (+/- 13.6%) and 409 37.6% (+/- 10.5%), respectively, of WT LR number. These results suggest that ASPR1 410 affects primary root growth and, in a more pronounced way, LR production or 411 development. 412 LRs are in general more sensitive to variations in environmental conditions than 413 primary roots, with nitrate signaling being an important regulator of LR development 414 (Sun et al., 2017). We tested the growth of oxASPR1-2, oxASPR1-4, and WT seedlings 415 on vertical plates in modified ½ MS media without ammonium nitrate (Fig. 6A). As

14

416 expected for mild nitrogen deprivation (Sun et al., 2017), elongated primary roots and a 417 slight increase in LR number per cm was observed in 10-days-old WT seedlings. The 418 same positive trend was observed in primary root length for oxASPR1 lines, and the 419 mutant lines were moderately but significantly shorter than the WT plants (Fig. 6B-D). 420 On the other hand, both overexpression lines did not display significant differences in 421 LR numbers from WT plants (Fig. 6E, F). In contrast with what was observed under 422 normal growth conditions, these results indicate that altered levels of ASPR1 do not 423 affect LR responses under low nitrate deficiency. 424 We next sought to perform a similar phenotypic analysis in loss-of-function 425 mutants. Two out of the three T-DNA insertion lines evaluated by RT-PCR analysis 426 revealed residual levels of expression of ASPR1 (Supplementary Fig. S6) and, therefore, 427 only the null mutant line Salk_038980.40.95.x (line hereon named aspr1) was further 428 evaluated in this study. We examined the root length and lateral root number of aspr1 429 seedlings and our results showed the same trend as observed for the overexpression 430 lines (Supplementary Fig. S7). Root length at 10 days in aspr1 seedlings was shorter 431 than WT and displayed a significant reduction in the number of LRs when compared 432 with WT seedlings (Supplementary Fig. S7A-F). Under low nitrate deficiency the aspr1 433 line displayed similar LR numbers as WT plants, as observed with the overexpression 434 lines (Supplementary Fig. S7G-L). Although further analysis of additional loss-of- 435 function alleles of ASPR1 lines will be required to validate our observations, the results 436 obtained with oxASPR1-2, oxASPR1-4, and aspr1 lines suggest that altered levels of 437 ASPR1 impact root growth and lateral root number. 438 To gain first insights into molecular changes associated with the observed root 439 phenotype, we compared protein abundance using chemical stable isotope labeling- 440 based quantitative proteomics approach between aspr1 and WT roots isolated from 10 441 day-old seedlings grown on ½ MS. We identified 2466 proteins (Supplementary Fig. 442 8A, B), of which 1676 proteins could be quantified in at least two of the three 443 independent replicate experiments (Supplementary Fig. 8B; Supplementary Table S3). 444 In agreement with the mild phenotype, the log2-transformed intensity ratios (aspr1/WT) 445 showed narrow and symmetric distributions (Supplementary Fig. 8B). 67 proteins were 446 significantly more abundant in aspr1 roots, while 66 proteins were significantly reduced 447 in abundance (LIMMA-moderated p-value <0.05; 20% change in abundance) 448 (Supplementary Fig. 8C, Supplementary Table S3). GO term analysis indicated that 449 accumulating proteins were involved in flavonoid biosynthesis, response to karrikin

15

450 (Nelson et al., 2012), and cadmium and salt stress responses, while proteins involved in 451 responses to biotic stimuli and cold stress were reduced in abundance (Supplementary 452 Fig. 8D, Supplementary Table S4). The broad GO categories “defense response”, 453 “response to oxidative stress” and “oxidation-reduction process” were represented both 454 among proteins with increased and reduced abundance.

16

455 Discussion 456 ASPR1 is an active aspartic protease displaying atypical enzymatic properties 457 We have demonstrated that ASPR1 encodes an active AP through heterologous 458 production in N. benthamiana using the magnICON® system. This expression system 459 has been extensively used to successfully express a broad range of proteins, mostly with 460 potential for pharmaceutical applications (Klimyuk et al., 2014; Larrimore et al., 2017; 461 Zahin et al., 2016). Our results suggest that it is an effective alternative also for the 462 production of plant atypical APs, which are generally very rich in cysteine residues and 463 challenging in many protein expression systems. Recombinant ASPR1 was shown to be 464 glycosylated and accumulated in two active forms, one containing the putative 465 prosegment domain and the other further processed lacking this region (Fig.1). 466 Although prosegment removal is often crucial for activation of typical APs, the 467 observed activity of pro-rASPR1 is not unexpected. For example, recombinant CDR1 468 was also active without propeptide removal (Simões et al., 2007). In typical APs, 469 removal of the prosegment is usually an autocatalytic event (Simões and Faro, 2004). 470 However, further studies are necessary to understand the nature and functional 471 relevance of this processing in rASPR1 as we could not rule out that this conversion 472 occurred during the purification process. Nevertheless, our results suggest that 473 proteolytic activity without irreversible prosegment removal might be a feature that is 474 more common among atypical APs than initially anticipated. 475 rASPR1 was most active at acidic pH [pH 4] (Fig. 2). This is a typical optimum 476 pH value for an AP (Simões and Faro, 2004), but very different from the pH profile of 477 CDR1, which shows maximal activity towards the same substrate at pH 6 (Simões et 478 al., 2007). These disparate optimal pH values are apparently common among atypical 479 APs, as ASPG1 and NANA also showed an optimal pH around 6, whereas nodulin 41 480 was most active at pH 4.5, similar to rASPR1 (Olivares et al., 2011; Paparelli et al., 481 2012; Yao et al., 2012). Moreover, rASPR1 was not completely inhibited by pepstatin 482 A. This is emerging as another distinguishing feature that separates typical APs from 483 the atypical members identified so far, as the latter were generally found not to be fully 484 inhibited by this canonical AP inhibitor (ASPG1, CND41 and nodulin 41) (Nakano et 485 al., 1997; Olivares et al., 2011; Simões and Faro, 2004; Simões et al., 2007; Yao et al., 486 2012). Various redox reagents inhibited rASPR1, suggesting that proteolytic activity 487 may be regulated by a mechanism involving redox-active Cys residues as previously 488 proposed for CDR1 (Simões et al., 2007). rASPR1 and CDR1 share the same number of

17

489 cysteines (12 in the proform of the enzymes) in conserved positions, reinforcing the 490 hypothesis that the presence of a thiol-disulfide redox switch regulates protease activity 491 in these atypical APs (Simões et al., 2007). Moreover, these results suggest that the in 492 vivo function of ASPR1 may be linked to a redox-sensing mechanism that merits further 493 exploration, especially in light of our proteomics results that point towards deregulation 494 of proteins involved in response to ROS in aspr1 roots (Supplementary Fig. 8, 495 Supplementary Table S3). 496 497 ASPR1 has unique specificity preferences resembling those of fungal aspartic 498 proteases 499 Our data showed that rASPR1 also displays unexpected specificity preferences 500 (Fig. 3). ASPR1 preferred hydrophobic amino acids in the S1 subsite, although with a 501 marked preference for leucine. This has been described for other APs (Almeida et al., 502 2017; Arnold et al., 1997; Beyer et al., 2005; Castanheira et al., 2005; Dunn and Hung, 503 2000; Koelsch et al., 2000; Leal et al., 2016; Turner et al., 2001), together with the 504 more stringent character of S1 in comparison with all other subsite binding pockets, 505 suggesting that the preferences at the P1 position are likely key specificity determinants. 506 Also, in line with what has been described for most of these proteases (Ahn et al., 2013; 507 Arnold et al., 1997; Beyer et al., 2005; Castanheira et al., 2005; Koelsch et al., 2000; 508 Leal et al., 2016; Turner et al., 2001), branched chain hydrophobic amino acids like 509 isoleucine and valine, as well as proline and the neutral amino acids serine, threonine 510 and glutamine were mostly disfavored at this position (Fig. 3). However, rASPR1 also 511 displayed a clear preference for accommodating asparagine and lysine in S1. This is 512 quite striking since similar preferences have been previously reported for 513 aspergillopepsin (Ahn et al., 2013). In fact, fungal APs like aspergillopepsin, 514 rhizopuspepsin, endothiapepsin, penicillopepsin as well as SAP2 and SAP3 from 515 Candida albicans differ from other eukaryotic APs because they not only favor 516 hydrophobic amino acids in positions P1 and P1’, but also cleave substrates with polar 517 residues like lysine (or histidine) in P1 (Shintani et al., 1996; Shintani et al., 1997). 518 Extensive characterization of this S1 subsite specificity in fungal APs has resulted in the 519 identification of Asp77 and Ser79 (pepsin numbering), two residues in the active site 520 flap, as critical for the accommodation of a basic amino acid at this position (Shintani et 521 al., 1996; Shintani et al., 1997). Strikingly, by comparing the amino acid sequences of 522 rASPR1 and fungal APs in the flap region, it was possible to confirm the presence of

18

523 the same residues (Asp and Ser) in ASPR1 (Table 1), suggesting these residues could 524 potentially account for the observed preference for lysine in P1. 525 Regarding P1’ specificity for ASPR1, there is a preference for accommodating 526 hydrophobic amino acids and for excluding proline, and also being much less stringent 527 when compared with P1 position, as described for other APs (Beyer et al., 2005; 528 Castanheira et al., 2005; Koelsch et al., 2000; Leal et al., 2016; Turner et al., 2001). 529 The S1’ subsite of ASPR1 appears more stringent than the corresponding subsite of 530 cardosin B, a typical plant AP, and accommodating bulkier hydrophobic amino acids 531 such as phenylalanine and tyrosine (Almeida et al., 2017). For positions distal to the 532 scissile bond, rASPR1 shares the same preferences with cardosin B, shewasins, pepsin, 533 plasmepsins,and BACE for P2’ position (Almeida et al., 2017; Beyer et al., 2005; Dunn 534 and Hung, 2000; Leal et al., 2016; Turner et al., 2001). However, P2 and P3 positions 535 for ASPR1 appear to be much more specific when compared with results available for 536 most eukaryotic APs. 537 Both primary and secondary specificity preferences of rASPR1 thus revealed 538 unique specificity requirements amongst plant APs but more akin to fungal APs. 539 540 Phenotype analysis and proteome profiling suggest a role for ASPR1 in LR formation 541 Analysis of ASPR1-promoter activity revealed strong expression in mature 542 flowers as well as root tips/lateral roots (Fig. 4). Phenotypic analysis found shorter 543 primary roots and fewer LRs in seedlings with altered expression levels of ASPR1 544 compared to WT (Fig. 5). This is not unexpected, since both absence and elevated 545 activity can negatively impact protein abundance and molecular pathways regulated by 546 proteolysis. This is in agreement with other reports on plant atypical APs, where 547 reported phenotypes resulted from their over-expression (Breitenbach et al., 2014; Ge et 548 al., 2005; Prasad et al., 2009; Xia et al., 2004; Yao et al., 2012). The formation of LRs 549 responds to environmental cues such as nitrogen availability that is integrated by auxin 550 and other signaling pathways (Fukaki and Tasaka, 2009; Kazan, 2013; Lavenus et al., 551 2013; Sun et al., 2017; Tsukagoshi, 2016). Surprisingly, mild nitrogen deprivation had a 552 strikingly similar positive impact on LR formation and LR density in oxASPR1 553 seedlings as in WT seedlings. Taken together, altered expression of ASPR1 indicated a 554 role of this AP in root elongation and as a modulator of LR development. Further work 555 will be needed to resolve the more precise stage of arrest of LR development in ASPR1 556 overexpression and loss-of-function mutants.

19

557 A quantitative comparison of aspr1 and WT root proteomes identified 133 558 differentially accumulating proteins, albeit mostly with modest changes in abundance 559 (Supplementary Table S3) that likely reflect the localized ASPR1 expression and the 560 mild quantitative phenotype. Nevertheless, the observed changes provided clues on 561 molecular mechanisms that might impair LR formation in aspr1 which can be pursued 562 in future functional analysis of additional mutant alleles. Several proteins known to be 563 involved in auxin biosynthesis, modification and transport showed differential 564 abundance in aspr1 roots, including accumulation of nitrilase 1 (AT3G44310) 565 (Lehmann et al., 2017) and rhamnosyl transferase UDP-glycosyltransferase 89C1 566 (AT1G06000) (Kuhn et al., 2016). In contrast, the lipid-binding protein ROSY1 567 (AT2G16005), which affects basipetal auxin transport (Dalal et al., 2016), was among 568 the proteins with the strongest reduction in aspr1. Moreover, several enzymes involved 569 in flavonoid biosynthesis over-accumulated in aspr1 roots, including chalcone synthase 570 (At5g13930; Transparent Testa 4 (TT4)), a known negative regulator of secondary root 571 development (Buer and Djordjevic, 2009). Finally, we observed that proteins involved 572 in response to ROS were deregulated in aspr1, including several peroxidases and 573 glutathione S-transferases (Supplementary Table S3). 574 Taken together, this work describes a new atypical AP with unique specificity 575 preferences that may be susceptible to redox-dependent regulation. Mutant lines with 576 altered ASPR1 abundance showed changes in root length and LR formation. These 577 findings open the exciting possibility that proteolytic regulation by ASPR1 could 578 directly or indirectly contribute to the integration of ROS, auxin and other 579 phytohormone signals during root growth and LR formation.

20

580 Supplementary data 581 Fig. S1. Expression of ASPR1 in E. coli resulted in the accumulation of recombinant 582 protein in the insoluble fraction. 583 Fig. S2. Amino acid sequence of rASPR1 construct expressed in N. benthamiana and 584 HisTrap purification. 585 Fig. S3. rASPR1 proteolytic activity in the presence of nucleotides and ions. 586 Fig. S4. Subcellular localization of ASPR1 in protoplasts of A. thaliana. 587 Fig. S5. Overexpression of ASPR1 in transgenic plants. 588 Fig. S6. ASPR1 T-DNA insertion lines. 589 Fig. S7. Root and lateral root phenotypes of aspr1 compared to WT seedlings. 590 Fig. S8. Quantitative comparison of protein abundance in aspr1 and WT roots. 591 Table S1. Primer sequences used during the experimental work. 592 Table S2. Compounds tested on rASPR1 enzymatic activity. 593 Table S3. Proteins included in the quantitative root proteome comparison. 594 Table S4. GO term analysis. 595 596 Acknowledgments 597 The authors kindly acknowledge the collaboration with Dr. Herta Steinkellner, 598 Department of Applied Genetics and Cell Biology, University of Natural Resources and 599 Life Sciences, Vienna, Austria that made possible the use of magnICON system. We 600 would like to thank Dr. Sandra Anjo for her advice on PICS MS data acquisition and the 601 Analytical Services Unit, ITQB, Universidade Nova de Lisboa, for Edman sequencing. 602 This work was supported by a PhD grant by the Foundation for Science and Technology 603 (FCT) to A.S. (SFRH/BD/51676/2011); an FCT Scientific and Technological Bilateral 604 Agreement 2015/2016 (Project 6818 - Programa de cooperação bilateral entre a FCT e o 605 DAAD da Alemanha, da competência do Dep. de Relações Internacionais da FCT to 606 I.S.); The European Regional Development Fund (ERDF) through the COMPETE 2020 607 - Operational Programme for Competitiveness and Internationalisation and Portuguese 608 national funds via FCT – Fundação para a Ciência e a Tecnologia, I.P., under projects: 609 POCI-01-0145-FEDER-007440 (ref.; UID/NEU/04539/2013), POCI-01-0145-FEDER- 610 016428 (ref.: SAICTPAC/0010/2015), and POCI-01-0145-FEDER-016795 (ref.: 611 PTDC/NEU-SCC/7051/2014); and by The National Mass Spectrometry Network 612 (RNEM) under the contract LISBOA-01-0145-FEDER-402-022125 (ref.: 613 ROTEIRO/0028/2013); the German Academic Exchange Service (DAAD) with

21

614 Funding from the Federal Ministry of Education and Research (PPP project 57128819 615 to P.F.H). Work in AYC lab is supported by US National Science Foundation (IOS- 616 1147165).

22

References

Ahn J, Cao M-J, Yu YQ, Engen JR. 2013. Accessing the Reproducibility and Specificity of Pepsin and other Aspartic Proteases. Biochimica et Biophysica Acta 1834, 1222-1229. Alam MM, Nakamura H, Ichikawa H, Miyao A, Hirochika H, Kobayashi K, Yamaoka N, Nishiguchi M. 2014. Response of an aspartic protease gene OsAP77 to fungal, bacterial and viral infections in rice. Rice 7, 9. Almeida CM, Manso JA, Figueiredo AC, et al. 2017. Functional and structural characterization of synthetic cardosin B-derived rennet. Applied Microbiology and Biotechnology 101, 6951-6968. Arnold D, Keilholz W, Schild H, Dumrese T, Stevanović S, Rammensee H-G. 1997. Substrate Specificity of Cathepsins D and E Determined by N-Terminal and C-terminal Sequencing of Peptide Pools. European Journal of Biochemistry 249, 171-179. Barkholt V. 1987. Amino acid sequence of endothiapepsin. European Journal of Biochemistry 167, 327-331. Beyer BB, Johnson JV, Chung AY, Li T, Madabushi A, Agbandje-McKenna M, McKenna R, Dame JB, Dunn BM. 2005. Active-Site Specificity of Digestive Aspartic Peptidases from the Four Species of Plasmodium that Infect Humans Using Chromogenic Combinatorial Peptide Libraries. Biochemistry 44, 1768-1779. Boersema PJ, Raijmakers R, Lemeer S, Mohammed S, Heck AJ. 2009. Multiplex peptide stable isotope dimethyl labeling for quantitative proteomics. Nature Protocols 4, 484-494. Breitenbach HH, Wenig M, Wittek F, et al. 2014. Contrasting roles of apoplastic aspartyl protease AED1 and legume lectin-like protein LLP1 in Arabidopsis systemic acquired resistance. Plant Physiology. Buer CS, Djordjevic MA. 2009. Architectural phenotypes in the transparent testa mutants of Arabidopsis thaliana. Journal of Experimental Botany 60, 751-763. Casamitjana-Martinez E, Hofhuis HF, Xu J, Liu CM, Heidstra R, Scheres B. 2003. Root-specific CLE19 overexpression and the sol1/2 suppressors implicate a CLV-like pathway in the control of Arabidopsis root meristem maintenance. Current biology : CB 13, 1435-1441.

23

Castanheira P, Samyn B, Sergeant K, Clemente JC, Dunn BM, Pires E, Van Beeumen J, Faro C. 2005. Activation, Proteolytic Processing, and Peptide Specificity of Recombinant Cardosin A. Journal of Biological Chemistry 280, 13047-13054. Chen F, Foolad MR. 1997. Molecular organization of a gene in barley which encodes a protein similar to aspartic protease and its specific expression in nucellar cells during degeneration. Plant Molecular Biology 35, 821-831. Chen J, Ouyang Y, Wang L, Xie W, Zhang Q. 2009. Aspartic proteases gene family in rice: Gene structure and expression, predicted protein features and phylogenetic relation. Gene 442, 108-118. Chen JJ, Ding JH, Ouyang YD, et al. 2008. A triallelic system of S5 is a major regulator of the reproductive barrier and compatibility of indica-japonica hybrids in rice. Proceedings of the National Academy of Sciences of the United States of America 105, 11436-11441. Clough SJ, Bent AF. 1998. Floral dip: a simplified method for Agrobacterium- mediated transformation of Arabidopsis thaliana. The Plant Journal 16, 735-743. Dalal J, Lewis DR, Tietz O, Brown EM, Brown CS, Palme K, Muday GK, Sederoff HW. 2016. ROSY1, a novel regulator of gravitropic response is a stigmasterol binding protein. Journal of Plant Physiology 196-197, 28-40. Du YJ, Scheres B. 2018. Lateral root formation and the multiple roles of auxin. Journal of Experimental Botany 69, 155-167. Dunn BM, Hung S-H. 2000. The two sides of enzyme–substrate specificity: lessons from the aspartic proteinases. Biochimica et Biophysica Acta 1477, 231-240. Faro C, Gal S. 2005. Aspartic proteinase content of the Arabidopsis genome. Current Protein and Peptide Science. 6, 493-500. Faro C, Ramalho-Santos M, Vieira M, et al. 1999. Cloning and Characterization of cDNA Encoding Cardosin A, an RGD-containing Plant Aspartic Proteinase. Journal of Biological Chemistry 274, 28724-28729. Fukaki H, Tasaka M. 2009. Hormone interactions during lateral root formation. Plant Molecular Biology 69, 437-449. Gao H, Zhang Y, Wang W, Zhao K, Liu C, Bai L, Li R, Guo Y. 2017. Two Membrane-Anchored Aspartic Proteases Contribute to Pollen and Ovule Development. Plant Physiology 173, 219-239.

24

Ge X, Dietrich C, Matsuno M, Li G, Berg H, Xia Y. 2005. An Arabidopsis aspartic protease functions as an anti-cell-death component in reproduction and embryogenesis. EMBO Reports 6, 282-288. Ghorbani S, Hoogewijs K, Pecenkova T, et al. 2016. The SBT6.1 subtilase processes the GOLVEN1 peptide controlling cell elongation. Journal of Experimental Botany 67, 4877-4887. Guo R, Xu X, Carole B, Li X, Gao M, Zheng Y, Wang X. 2013. Genome-wide identification, evolutionary and expression analysis of the aspartic protease gene superfamily in grape. BMC Genomics 14, 554. Hsu IN, Delbaere LTJ, James MNG, Hofmann T. 1977. Penicillopepsin from Penicillium janthinellum crystal structure at 2.8 [angst] and sequence homology with porcine pepsin. Nature 266, 140-145. Huang J, Zhao X, Cheng K, Jiang Y, Ouyang Y, Xu C, Li X, Xiao J, Zhang Q. 2013. OsAP65, a rice aspartic protease, is essential for male fertility and plays a role in pollen germination and pollen tube growth. Journal of Experimental Botany 64, 3351- 3360. Hube B, Turver CJ, Odds FC, Eiffert H, Boulnois GJ, Köchel H, Rüchel R. 1991. Sequence of the Candida albicans gene encoding the secretory aspartate proteinase. Journal of Medical and Veterinary Mycology 29, 129-132. Ito T, Nagata N, Yoshiba Y, Ohme-Takagi M, Ma H, Shinozaki K. 2007. Arabidopsis MALE STERILITY1 encodes a PHD-type transcription factor and regulates pollen and tapetum development. Plant Cell 19, 3549-3562. Ji Q, Zhang MJ, Lu JF, et al. 2012. Molecular Basis Underlying the S5-Dependent Reproductive Isolation and Compatibility of indica/japonica Rice Hybrids. Plant Physiology 158, 1319-1328. Kazan K. 2013. Auxin and the integration of environmental signals into plant root development. Annals of Botany 112, 1655-1665. Kervinen J, Tobin GJ, Costa J, Waugh DS, Wlodawer A, Zdanov A. 1999. Crystal structure of plant aspartic proteinase prophytepsin: inactivation and vacuolar targeting. EMBO Journal 18, 3947-3955. Klimyuk V, Pogue G, Herz S, Butler J, Haydon H. 2014. Production of Recombinant Antigens and Antibodies in Nicotiana benthamiana Using ‘Magnifection’ Technology: GMP-Compliant Facilities for Small- and Large-Scale Manufacturing. In: Palmer K,

25

Gleba Y, eds. Plant Viral Vectors. Berlin, Heidelberg: Springer Berlin Heidelberg, 127- 154. Koelsch G, Tang J, Loy JA, Monod M, Jackson K, Foundling SI, Lin X. 2000. Enzymic characteristics of secreted aspartic proteases of Candida albicans. Biochimica et Biophysica Acta 1480, 117-131. Kohli A, Narciso JO, Miro B, Raorane M. 2012. Root proteases: reinforced links between nitrogen uptake and mobilization and drought tolerance. Physiologia Plantarum 145, 165-179. Kuhn BM, Errafi S, Bucher R, Dobrev P, Geisler M, Bigler L, Zazimalova E, Ringli C. 2016. 7-Rhamnosylated Flavonols Modulate Homeostasis of the Plant Hormone Auxin and Affect Plant Development. Journal of Biological Chemistry 291, 5385-5395. Larrimore KE, Kazan IC, Kannan L, et al. 2017. Plant-expressed cocaine hydrolase variants of butyrylcholinesterase exhibit altered allosteric effects of cholinesterase activity and increased inhibitor sensitivity. Scientific Reports 7. Lavenus J, Goh T, Roberts I, et al. 2013. Lateral root development in Arabidopsis: fifty shades of auxin. Trends in Plant Science 18, 450-458. Leal AR, Cruz R, Bur D, Huesgen PF, Faro R, Manadas B, Wlodawer A, Faro C, Simoes I. 2016. Enzymatic properties, evidence for in vivo expression, and intracellular localization of shewasin D, the pepsin homolog from Shewanella denitrificans. Scientific Reports 6. Lehmann T, Janowitz T, Sanchez-Parra B, Alonso MP, Trompetter I, Piotrowski M, Pollmann S. 2017. Arabidopsis NITRILASE 1 Contributes to the Regulation of Root Growth and Development through Modulation of Auxin Biosynthesis in Seedlings. Frontiers in Plant Science 8, 36. Li YR, Kabbage M, Liu WD, Dickman MB. 2016. Aspartyl Protease-Mediated Cleavage of BAG6 Is Necessary for Autophagy and Fungal Resistance in Plants. Plant Cell 28, 233-247. Malekpoor Mansoorkhani F, Seymour GB, Swarup R, Moeiniyan Bagheri H, Ramsey RJ, Thompson AJ. 2014. Environmental, developmental, and genetic factors controlling root system architecture. Biotechnology & Genetic Engineering Reviews 30, 95-112.

26

Manzano C, Pallero-Baena M, Casimiro I, et al. 2014. The Emerging Role of Reactive Oxygen Species Signaling during Lateral Root Development. Plant Physiology 165, 1105-1119. Marillonnet S, Thoeringer C, Kandzia R, Klimyuk V, Gleba Y. 2005. Systemic Agrobacterium tumefaciens-mediated transfection of viral replicons for efficient transient expression in plants. Nature Biotechnology 23, 718-723. Nakano T, Murakami S, Shoji T, Yoshida S, Yamada Y, Sato F. 1997. A novel protein with DNA binding activity from tobacco chloroplast nucleoids. Plant Cell 9, 1673-1682. Nakano T, Nagata N, Kimura T, et al. 2003. CND41, a chloroplast nucleoid protein that regulates plastid development, causes reduced gibberellin content and dwarfism in tobacco. Physiologia Plantarum 117, 130-136. Nelson DC, Flematti GR, Ghisalberti EL, Dixon KW, Smith SM. 2012. Regulation of seed germination and seedling growth by chemical signals from burning vegetation. Annual Review of Plant Biology 63, 107-130. Niu N, Liang W, Yang X, Jin W, Wilson ZA, Hu J, Zhang D. 2013. EAT1 promotes tapetal cell death by regulating aspartic proteases during male reproductive development in rice. Nature Communications 4, 1445. Olivares JE, Díaz-Camino C, Estrada-Navarrete G, et al. 2011. Nodulin 41, a novel late nodulin of common bean with peptidase activity. BMC Plant Biology 11, 134. Paparelli E, Gonzali S, Parlanti S, et al. 2012. Misexpression of a Chloroplast Aspartyl Protease Leads to Severe Growth Defects and Alters Carbohydrate Metabolism in Arabidopsis. Plant Physiology 160, 1237-1250. Peer WA, Hosein FN, Bandyopadhyay A, et al. 2009. Mutation of the membrane- associated M1 protease APM1 results in distinct embryonic and seedling developmental defects in Arabidopsis. Plant Cell 21, 1693-1721. Peret B, De Rybel B, Casimiro I, Benkova E, Swarup R, Laplaze L, Beeckman T, Bennett MJ. 2009. Arabidopsis lateral root development: an emerging story. Trends in Plant Science 14, 399-408. Petricka JJ, Winter CM, Benfey PN. 2012. Control of Arabidopsis root development. Annual Review of Plant Biology 63, 563-590. Phan HA, Iacuone S, Li SF, Parish RW. 2011. The MYB80 Transcription Factor Is Required for Pollen Development and the Regulation of Tapetal Programmed Cell Death in Arabidopsis thaliana. The Plant Cell Online 23, 2209-2224.

27

Prasad BD, Creissen G, Lamb C, Chattoo BB. 2009. Overexpression of rice (Oryza sativa L.) OsCDR1 leads to constitutive activation of defense responses in rice and Arabidopsis. Molecular Plant-Microbe Interactions : MPMI 22, 1635-1644. Prasad BD, Creissen G, Lamb C, Chattoo BB. 2010. Heterologous expression and characterization of recombinant OsCDR1, a rice aspartic proteinase involved in disease resistance. Protein Expression and Purification 72, 169-174. Rogg LE, Bartel B. 2001. Auxin signaling: derepression through regulated proteolysis. Developmental Cell 1, 595-604. Schechter I, Berger A. 2012. ON THE SIZE OF THE ACTIVE SITE IN PROTEASES. I. PAPAIN (Reprinted from BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS, vol 27, pg 157, 1967). Biochemical and Biophysical Research Communications 425, 497-502. Schilling, Overall CM. 2008. Proteome-derived, database-searchable peptide libraries for identifying protease cleavage sites. Nature Biotechnology 26. Schilling O, auf dem Keller U, Overall CM. 2011. Factor Xa subsite mapping by proteome-derived peptide libraries improved using WebPICS, a resource for proteomic identification of cleavage sites. Biological Chemistry 392, 1031-1037. Schwechheimer C, Schwager K. 2004. Regulated proteolysis and plant development. Plant Cell Reports 23, 353-364. Sepulveda P, Marciniszyn J, Liu D, Tang J. 1975. Primary structure of porcine pepsin. III. Amino acid sequence of a cyanogen bromide fragment, CB2A, and the complete structure of porcine pepsin. Journal of Biological Chemistry 250, 5082-5088. Shintani T, Ichishima E. 1994. Primary structure of aspergillopepsin I deduced from nucleotide sequence of the gene and aspartic acid-76 is an essential active site of the enzyme for trypsinogen activation. Biochimica et Biophysica Acta 1204, 257-264. Shintani T, Kobayashi M, Ichishima E. 1996. Characterization of the S1 Subsite Specificity of Aspergillopepsin I by Site-Directed Mutagenesis. Journal of Biochemistry 120, 974-981. Shintani T, Nomura K, Ichishima E. 1997. Engineering of Porcine Pepsin: alteration of S1 substrate specificity of Pepsin to those of fungal Aspartic Proteinases by site- directed mutagenesis. . Journal of Biological Chemistry 272, 18855-18861. Simões I, Faro C. 2004. Structure and function of plant aspartic proteinases. European Journal of Biochemistry 271, 2067-2075.

28

Simões I, Faro R, Bur D, Faro C. 2007. Characterization of Recombinant CDR1, an Arabidopsis Aspartic Proteinase Involved in Disease Resistance. Journal of Biological Chemistry 282, 31358-31365. Srivastava AK, Zhang C, Sadanandom A. 2016. Rice OVERLY TOLERANT TO SALT 1 (OTS1) SUMO protease is a positive regulator of seed germination and root development. Plant Signaling & Behavior 11, e1173301. Sun CH, Yu JQ, Hu DG. 2017. Nitrate: A Crucial Signal during Lateral Roots Development. Frontiers in Plant Science 8, 485. Takahashi K. 1987. The amino acid sequence of rhizopuspepsin, an aspartic proteinase from Rhizopus chinensis. Journal of Biological Chemistry 262, 1468-1478. Tao LZ, Cheung AY, Wu HM. 2002. Plant Rac-like GTPases are activated by auxin and mediate auxin-responsive gene expression. Plant Cell 14, 2745-2760. Tognetti VB, Muhlenbock P, Van Breusegem F. 2012. Stress homeostasis - the redox and auxin perspective. Plant, Cell & Environment 35, 321-333. Tsukagoshi H. 2016. Control of root growth and development by reactive oxygen species. Current Opinion in Plant Biology 29, 57-63. Turner RT, Koelsch G, Hong L, Castenheira P, Ghosh A, Tang J. 2001. Subsite Specificity of Memapsin 2 (β-Secretase): Implications for Inhibitor Design. Biochemistry 40, 10001-10006. van der Hoorn RA. 2008. Plant proteases: from phenotypes to molecular mechanisms. Annual Review of Plant Biology 59, 191-223. Vizcaino JA, Csordas A, del-Toro N, et al. 2016. 2016 update of the PRIDE database and its related tools. Nucleic Acids Research 44, D447-456. Wang R, Estelle M. 2014. Diversity and specificity: auxin perception and signaling through the TIR1/AFB pathway. Current Opinion in Plant Biology 21, 51-58. Wu S, Scheible WR, Schindelasch D, Van Den Daele H, De Veylder L, Baskin TI. 2010. A conditional mutation in Arabidopsis thaliana separase induces chromosome non-disjunction, aberrant morphogenesis and cyclin B1;1 stability. Development 137, 953-961. Xia Y, Suzuki H, Borevitz J, Blount J, Guo Z, Patel K, Dixon RA, Lamb C. 2004. An extracellular aspartic protease functions in Arabidopsis disease resistance signaling. EMBO Journal 23, 980-988.

29

Yao X, Xiong W, Ye T, Wu Y. 2012. Overexpression of the aspartic protease ASPG1 gene confers drought avoidance in Arabidopsis. Journal of Experimental Botany 63, 2579-2593. Zahin M, Joh J, Khanal S, et al. 2016. Scalable Production of HPV16 L1 Protein and VLPs from Tobacco Leaves. PloS one 11.

30

Table 1. Amino acid sequences of the active-site flap region of several aspartic proteases

Protein Active-site flap Reference (Shintani and Aspergillopepsin I WDISYGDGSSASGD Ichishima, 1994)

(Hsu et al., Penicillopepsin WSISYGDGSSASGN 1977)

(Takahashi, Rhizopuspepsin WSISYGDGSSASGI 1987)

(Barkholt, Endothiapepsin WSISYGDGSSSSGD 1987)

(Hube et al., Candidapepsin FYIGYGDGSSSQGT 1991)

ASPR1 YLYTYGDYSSTRGL This work

(Faro et al., Cardosin A GAIIYGTGSI-TGF 1999)

(Kervinen et Phytepsin AAIQYGTGSI-AGY al., 1999)

(Sepulveda et Porcine Pepsin LSITYGTGSM-TGI al., 1975)

31

Figure Legends

Fig. 1. Purification of rASPR1 produced from N. benthamiana leaves using the magnICON expression system. (A) Anionic exchange chromatography. HisTrap eluates (Fig. S2B) were pooled, dialyzed and applied to a MonoQ column. (B) Fractions corresponding to the shaded area in (A) were analyzed by SDS-PAGE stained with Coomassie blue (left panel) and Western blot (anti-His antibody) (right panel). Asterisks indicate protein bands subjected to N-terminal Edman sequencing. (C) Purified rASPR1 pull-down using Ni2+ magnetic beads (C-terminal His-tag). Samples were monitored by Western blot (anti-His). I: input; NB: non-bound (protein fraction non-bound to the Ni2+ magnetic beads); E: eluted protein; Wash: washing steps. Molecular weight markers in kilodaltons (kDa) are shown. (D) Proteolytic activity of the input sample (I) and the non-bound (NB) sample tested against the fluorogenic substrate [MCA-Lys]Leu-His-Pro-Glu-Val-Leu-Phe-Val-Leu-Glu[Lys-DNP]. This figure is available in colour at JXB online.

Fig. 2. Glycosylation of rASPR1 and effect of pH, different protease inhibitors and redox agents on proteolytic activity. (A) Protein deglycosylation assays. Purified rASPR1 was treated with Endo H or PNGase F and digestion products were analyzed by SDS-PAGE stained with Coomassie blue. (B) The effect of pH on the proteolytic activity of purified rASPR1 was studied by incubating the enzyme at 25 °C with buffers between pH 4.0 and pH 6.5. (C) Effect of different compounds on the activity of rASPR1 was evaluated by pre-incubating the purified protease with each compound for 2 min at room temperature and enzymatic activity tested at pH 4.0, 25 °C. Protease inhibitors: pepstatin (aspartic proteases); bestatin (aminopeptidases); Pefabloc (serine proteases); E-64 (cysteine proteases). GSH: reduced glutathione; GSSG: oxidized glutathione. All activity assays were performed using the fluorogenic substrate [MCA- K]Leu-His-Pro-Glu-Val-Leu-Phe-Val-Leu-Glu[Lys-DNP]. The error bars represent standard deviation from the mean. This figure is available in colour at JXB online.

Fig. 3. rASPR1 specificity preferences profiled by PICS. Purified rASPR1 was incubated with either a tryptic or GluC peptide library derived from human THP1 cells, at a ratio of 1:20 (enzyme:library), at pH 4.0. The average amino acid occurrences in P4-P4’ were calculated from one experiment for each library. (A) Specificity profiles

32 represented in the form of two-dimensional heatmaps of log2 transformed values of fold-enrichment over natural abundance of amino acids. (B) Specificity profiles represented in the form of % difference IceLogos. Horizontal axis in IceLogos represents the amino acid position and vertical axis shows the over- and under- representation of amino acid occurrence compared with the Swiss-Prot Homo sapiens protein database. Cysteines are carboxyamidomethylated and lysines are dimethylated. Vertical dashed line indicate cleaved peptide bond between P1-P1’ positions. This figure is available in colour at JXB online.

Fig. 4. Expression pattern of ASPR1 gene. Analysis of GUS activity in A. thaliana pASPR1::GUS transgenic lines in whole inflorescence (A-C) and in 13-d-old transgenic seedlings (D-I) showing expression in LR primordia and LR tips. (J) Semi-quantitative RT-PCR analysis of mRNA levels of ASPR1 in whole inflorescence (I) and root (R) tissue in WT plants. ASPR1 and Actin (loading control) specific primers were used for RT-PCR analysis. This figure is available in colour at JXB online.

Fig. 5. Root and lateral root phenotypes of oxASPR1 compared to WT seedlings. (A) WT, oxASPR1-2, and oxASPR1-4 seedlings were grown vertically for 10 days in ½ MS media. Hypocotyl lengths were highly variable and thus not analyzed in detail in here. (B) Primary root length was measured at day 3, 5, 7 and 10. Each experiment was repeated twice and in total 100-150 roots were measured per WT and 30 roots per OxASPR1 line. (C) Total root length at day 10. (D) Relative total root length at day 10 (to each WT). (E) and (F) The number of lateral roots was counted under the dissection microscope. Absolute and relative number of lateral roots per cm of primary root, respectively. The error bars represent standard deviation from the mean (**** P <0.0001). This figure is available in colour at JXB online.

Fig. 6. Root growth and lateral root phenotypes of oxASPR1 seedlings under low nitrate conditions. (A) WT, oxASPR1-2, and oxASPR1-4 seedlings were grown vertically for 10 days in modified ½ MS media without ammonium nitrate . (B) Primary root length was measured at day 3, 5, 7 and 10. Each experiment was repeated twice and in total 100-150 roots were measured per WT and 30 roots per OxASPR1 line. (C) Total root length at day 10. (D) Relative total root length at day 10 (to each WT). (E) and (F) Absolute and relative number of lateral roots per cm of primary root, respectively. The

33 error bars represent standard deviation from the mean (ns, nonsignificant; **** P <0.0001). This figure is available in colour at JXB online.

34

Supplementary Material

An atypical aspartic protease modulates lateral root development in Arabidopsis thaliana

André Soares1,2,3,4#, Stefan Niedermaier5, Rosário Faro3, Andreas Loos6§, Bruno Manadas3, Carlos Faro3, Pitter F. Huesgen5*, Alice Y. Cheung4*, Isaura Simões2,3*,

1PhD Programme in Experimental Biology and Biomedicine, Center for Neuroscience and Cell Biology, University of Coimbra, Portugal; 2Institute for Interdisciplinary Research, University of Coimbra, Portugal; 3CNC-Center for Neuroscience and Cell Biology, University of Coimbra; 4Department of Biochemistry and Molecular Biology, University of Massachusetts, Amherst, MA 01003, USA; 5Central Institute for Engineering, Electronics and Analytics, ZEA-3, Forschungszentrum Jülich, Jülich, Germany; 6Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria;

#Current address: FairJourney Biologics, Porto, Portugal §Current address: Aridis Pharmaceuticals, 5941 Optical Court, CA 95138, US. *Corresponding authors

*Corresponding authors: Pitter F. Huesgen, Central Institute for Engineering, Electronics and Analytics, ZEA-3, Forschungszentrum Jülich, Jülich, Germany; E-mail: [email protected] Alice Y. Cheung, Department of Biochemistry and Molecular Biology, University of Massachusetts, Amherst, MA 01003, USA; E-mail: [email protected] Isaura Simões, Edifício UC-Biotech, Biocant-Parque Tecnológico de Cantanhede, Núcleo 04, Lote 8, 3060-197 Cantanhede, Portugal; E-mail: [email protected];

1

Methods and Materials

Generation of A. thaliana transgenic lines For the 35S::ASPR1-GFP construct, the cDNA encoding the full-length sequence of ASPR1 was amplified using the primers 1 and 2 (Supplementary Table S1) to include BamHI and SalI restriction sites (underlined sequences) at 5’- and 3’-ends, respectively. The sequence was then inserted into the plasmid pAC1352 (Rodermel et al., 1988) in frame with the C-terminal GFP coding sequence, and under the control of the 35S constitutive promoter. A silent mutation was introduced to eliminate ASPR1 BamHI recognition site prior amplification and cloning into pAC1352. Mutagenesis was performed using the Quick Change Site-Directed Mutagenesis Kit (Stratagene) using primers 3 and 4 (Supplementary Table S1). To generate the construct pASPR11::GUS, the ASPR1 promoter sequence (considered to be the 2000 bp sequence upstream the coding region) was amplified using the primers 5 and 6 (Supplementary Table S1) which included the restrictions sites for XbaI and BanHI at 5’ and 3’ ends, and inserted into the vector pBI121 (Clontech) upstream the GUS sequence, using XbaI and BamHI restriction enzymes. All constructs were confirmed by DNA sequencing. Plasmid DNA constructs 35S::ASPR1-GFP and pASPR1::GUS were independently transformed into Agrobacterium tumefaciens strain GV3101 by a combined heat and cold treatment and positive clones grown in 200 mL of LB supplemented with rifampicin (25 µg/mL) and ampicillin (100 µg/mL), 12-16 h at 30 °C. Cultures were pelleted at 5000g for 15 min and resuspended in dipping buffer (1 mM MgSO4, 5% sucrose) supplemented with 0.02% of silwet L- 77. Cell suspensions were used for A. thaliana transformation using the floral dip method (Clough and Bent, 1998). T0 seeds were surface sterilized and sown in B5 (Phyto Technology Laboratories) medium plates supplemented with kanamycin (50 µg/mL). Kanamycin-resistant seedlings were selected and transferred to soil. Plants were confirmed to harbor the intended transgene by assessing GUS activity or by Western blot analysis.

Plasmid DNA constructs for transient expression in Arabidopsis protoplasts

For the ASPR1(1-39)-GFP construct, the cDNA encoding the first 39 amino acids of ASPR1 was amplified using the primers 11 and 12 (Supplementary Table S1) to include PstI and SalI restriction sites (underlined sequences) at 5’- and 3’-ends, respectively. For the ASPR1-GFP construct, the cDNA encoding the full-length sequence of ASPR1 was amplified using the primers

2

11 and 13 (Supplementary Table S1), also with PstI and SalI restriction sites (underlined sequences) at 5’- and 3’-ends, respectively. The sequences were then inserted into a modified Bluescript pSK vector (Stratagene), in frame with the C-terminal GFP coding sequence, and under the control of the 35S constitutive promoter. Both constructs were confirmed by DNA sequencing.

Endoglycosidase digestion Assays with EndoH and PNGase F (New England Biolabs) were performed according to the manufacturer’s instructions, with minor changes. For each treatment, 25 µL of purified rASPR1 were denatured according to manufacturer’s instructions, the reaction volume was brought to 50 µL with ddH2O and the supplemented buffers, and incubated at 37 °C for 1 h with 0.5 µL of each enzyme, separately. Treated samples were analyzed by Coomassie stained SDS- PAGE.

SDS-PAGE and Western blotting Protein samples were separated by SDS-PAGE (12.5%) with a MiniProtean3 apparatus (Bio-Rad), in 100 mM Tris, 100 mM bicine buffer supplemented with 0.1% SDS, and . were either stained with Coomassie Brilliant Blue R-250 or electrotransferred to a polyvinylidene difluoride membrane for Western blot analysis, using a Trans-Blot® Electrophoretic Transfer Cell (Bio-Rad) (40 V, 12 hours, at 4 ⁰C), in the buffer 25 mM Tris, 192 mM glicine, 20% methanol. Membranes were blocked for 1 h in TBST buffer (20 mM Tris.HCl pH 7.6, 192 mM NaCl, 0.1% v/v Tween 20) supplemented with 5% skim milk and incubated (1 h) with primary antibody (mouse anti-His antibody from GenScript; 1:10000 dilution). After washing with TBST with 0.5% skim milk and incubated with the secondary antibody for 1 h [alkaline phosphatase-conjugated goat anti-(mouse IgG + IgM) from GE Healthcare Life Sciences; 1:10000 dilution], membranes were washed with TBST buffer, developed with ECFTM substrate (GE Healthcare Life Sciences) and scanned on a Molecular Imager FX (Bio-Rad).

Proteomic identification of protease cleavage sites (PICS) Samples were analyzed on a Triple TOFTM 5600 System (ABSciex®) in information-dependent acquisition (IDA) mode. Peptides were resolved by liquid chromatography (nanoLC Ultra 2D, Eksigent®) on a MicroLC column Halo Fused-Core C18 (300 μm ID × 15 cm length, 2.7 μm

3

particles, 90 Å pore size, Eksigent®) at 5μL/min with a 110 min linear gradient from 2% to 30% acetonitrile in 0.1% formic acid. Peptides were ionized into the mass spectrometer using an electrospray ionization source (DuoSprayTM Source, ABSciex®) with a 50 μm internal diameter (ID) stainless steel emitter (NewObjective). IDA experiment was performed (as described in Anjo et al (2015)), the mass spectrometer was set for IDA scanning full spectra (350-1250 m/z) for 250 ms, followed by up to 30 MS/MS scans (100-1500 m/z for 100 ms each). Candidate ions with a charge state between +2 and +5 and counts above a minimum threshold of 70 counts per second were isolated for fragmentation, and 1 MS/MS spectra was collected before adding those ions to the exclusion list for 15 sec (mass spectrometer operated by Analyst® TF 1.6, ABSciex®). Rolling collision energy was used with a collision energy spread of 5. Analysis of the peptide sequences was performed using MS Converter, followed by peptide sequence identification with X!Tandem in combination with IProphet at a confidence level > 99% (Craig and Beavis, 2004; Schilling and Overall, 2008; Shteynberg et al., 2011). Analysis parameters were established such as mass tolerances at 0.05 Da for parent ions and 0.1 Da for fragment ions, as well as for identification of the following static modifications: carboxyamidomethylation of cysteine residues (+ 57.02 Da), dimethylation of lysines (+ 28.03 Da) and thioacylation of peptide amino termini (+ 88.00 Da). Semi-style cleavage searches were applied with no constraints for the orientation of the specific terminus. The Web-based PICS service (Schilling et al., 2011) was used to derive nonprime sequences and to reconstruct the cleavage sites. Sequence logos were generated with IceLogo (Colaert et al., 2009), showing significant differences in amino acid occurrence at each position as compared to the expected amino acid frequency in the human proteome-derived peptide libraries (p-value <0.05).

Quantitative root proteome comparison 150-200 mg of frozen roots were mechanically ground (Polytron PT 2500E homogenizer (Kinematica) with a 12 mm diameter rotor, 24.000 rpm for 60 s) in ice cold protein extraction buffer (6M Guanidine-HCl, 5 mM EDTA, 150 mM HEPES pH 7.5, 1x Complete protease inhibitors (Roche)). Ground tissue was filtered through Miracloth membrane (Sigma Aldrich), centrifuged at 500g for 10 min at 4°C to remove insoluble cell debris and protein concentrations of in the supernatant estimated by BCA assay (Thermo). 200 µg of each sample were reduced with

4

12.5 mM dithiothreitol (DTT) at 55°C for 30min, Cys residues alkylated with 37.5 mM iodoacetic acid (IAA) for 30min at 20°C in the dark, and the reaction quenched with 25 mM DTT for 20 min at 37°C. Proteins were purified by chloroform/methanol precipitation (Wessel and Flugge, 1984), air dried and resuspended in 100 mM HEPES (pH 7.5) before digestion with proteomics-grade trypsin (Thermo) at a protease:proteome ratio of 1:125 (w/w) with added 5 mM CaCl2. Digestion was allowed to proceed at 37°C for 16 h before differential stable isotope labeling of primary amines at peptide N termini and at Lys side chains by reductive dimethylation (Boersema et al., 12 2009). In all cases, wild type samples were labeled with 40 mM light formaldehyde ( CH2O and 12 20 mM sodium cyanoborohydride (NaCNBH3), while aspr1 samples were labeled with CH2O 13 (replicates 1 and 2) or CD2O (replicate 3). In the first experiment, we included the ASPR1 12 overexpression line and labeled those peptides with CD2O and 20 mM sodium cyanoborodeuteride (NaCNBD3), but had excluded this line from later experiments due to an insufficient supply of seedlings because of a low germination seed lot. All labeling reactions were quenched by adding 100 mM TRIS pH 7.5 for 30min at 37°C before combination of the differentially labeled samples of each replicate and desalted with C18 reverse phase cartridges (Sep-

Pak, Waters). 500 ng desalted peptides were loaded onto a C18 reverse phase capillary trap column

(Acclaim PepMap C18, 75 µm x 2 cm, 3 µm particle size, Thermo) and separated on a C18 reverse phase analytical column (Acclaim PepMap C18, 75 µm x 50 cm, 2 µm particle size, 100 Å, Thermo) using a nano-HPLC system (UltiMate3000 nano RSLC, Thermo). Peptides were eluted at a flow rate of 300 nl/min with a 90 min gradient from 2% to acetonitrile (ACN) to 35% (v/v) acetonitrile in H2O. The nano-LC system was on-line coupled to a high-resolution quadrupole- time of flight tandem mass spectrometer (Impact II, Bruker) using a CaptiveSpray nano- electrospray source (Bruker) with acetonitrile as dopant in the nitrogen gas stream. Peptides were identified and quantified using the Max-Quant software package, version 1.5.5.1 (Tyanova et al., 2016a). Generic settings for Bruker Q-TOF instruments were used to match spectra to proteins sequence in the A. thaliana Uniprot proteome database (release 2016/11, 33463 entries). Dimethylation at peptide N-termini and lysine side chains was set as label for quantification (channel 1 +28.0313, channel 2 +32.0564 for replicates 1 and 2 or +34.0631 for replicate 3, and channel 3 +36.0756 for replicate 1) and the “requantify” feature was enabled. Trypsin was set as digestion protease allowing for up to two missed cleavages. N-terminal protein acetylation (+42.0106) and methionine oxidation (+15.9949) were considered as variable

5

modification and carbamidomethylation of cysteines was set as fixed modification. MaxQuant protein quantifications were further evaluated using the Perseus statistical software (v 1.6.0.7, (Tyanova et al., 2016b)). Contaminant, reverse and “only identified by site” hits were filtered out and decimal ratios were transformed into log2 values. The resulting log2 (aspr1/wt) distributions were further processed by linear model for microarray (LIMMA) fitting (Ritchie et al., 2015) as previously described (Gomez-Auli et al., 2016) using the Bioconductor R package (Huber et al., 2015). Proteins were considered altered in abundance if they passed the LIMMA-moderated t-test (p-val <0.05) and showed a >20% change in abundance (log2 fold change <-0.263 or >0.263). GO and KEGG annotation terms were added from the Perseus Arabidopsis thaliana annotation file (download 2016/11) and GO enrichment analysis was performed using the DAVID webserver (Huang et al., 2009). Generic DAVID annotation chart settings were used.

6

Supplementary Figures

Fig. S1. Expression of ASPR1 in E. coli resulted in the accumulation of recombinant protein in the insoluble fraction. (A) Schematic representation of constructs used for small-scale expression screenings. ASPR1: sequence without the putative signal peptide (first 27 amino acids); ΔPro_ASPR1: sequence without the putative signal peptide and prosegment (first 92 amino acids); SP (signal peptide), Pro (prosegment). (B) Western blot analysis of insoluble (I) and soluble (S) fractions of ASPR1 small-scale expression screening. Conditions tested: E. coli strains BL21 Star (DE3) and C41; Expression media: LB and TB; Expression temperatures: 37 oC and 20 oC; Expression times: 3 h and 16 h (on); IPTG concentrations: 0.1 mM and 0.5 mM. Upper panel: ASPR1 expression without pre-expression of Erv1p and DsbC (Nguyen et al., 2011); Botton panel: ASPR1 expression upon pre-expression of Erv1p and DsbC. (C) Western blot analysis of insoluble

7

(I) and soluble (S) fractions of ASPR1 and ΔPro_ASPR1 small-scale expression screenings. Conditions tested: E. coli strain C41 cells with pre-expression of Ervd1 and DsbC; Expression media: LB and TB; Expression temperature: 20 oC, Expression time: 3 h and 16 h (on); IPTG concentrations: 0.1 mM and 0.05 mM. For each condition we analyzed the insoluble (I) and the soluble (S) fraction. Western blot with anti-His antibody (α-His).

8

Fig. S2. Amino acid sequence of rASPR1 construct expressed in N. benthamiana and HisTrap purification. (A) Full-length proASPR1 (without its putative signal peptide) fused to the proprietary magnICON® signal peptide and with a 6xHis-tag at the C-terminal. Green: putative prosegment; yellow boxes: catalytic aspartates; blue: cysteine residues; red: putative glycosylation sites. Amino acids underlined in light blue and orange correspond to the N-terminal sequences determined for rASPR1 purified fraction (Fig. 1A). Light blue: N-terminal sequence of the higher molecular weight band in the SDS-PAGE analysis (Fig. 1A). Orange: N-terminal sequence of the lower molecular weight band (Fig. 1A). (B) Total soluble protein (TSP) extract from 15-20 g of infiltrated leaves 3 dpi was applied onto a HisTrap HP column. Proteins were eluted using stepwise increases in concentration of imidazole (50, 100 and 500 mM), corresponding to the shaded areas of the chromatogram. (C) SDS-PAGE analysis of the eluted fractions in (B). Protein samples were dialyzed and purified by anion-exchange chromatography (Fig. 1A).

9

Fig. S3. rASPR1 proteolytic activity in the presence of nucleotides and ions. Effect of different compounds on the activity of rASPR1 was evaluated by pre-incubating the purified protease with each compound for 2 min at room temperature. Enzymatic activity was tested towards the fluorogenic substrate [MCA-Lys]Leu-His-Pro-Glu-Val-Leu-Phe-Val-Leu-Glu[Lys-DNP] at pH 4.0, 25 °C. The error bars represent standard deviation from the mean.

10

Fig. S4. Subcellular localization of ASPR1 in protoplasts of A. thaliana. Fluorescence signal from the expression of (A) ASPR1(1-39)-GFP (first 39 amino acids of ASPR1) and of (B) ASPR1- GFP (full-length ASPR1).Chlorophyll autofluorescence is shown in red.

11

Fig. S5. Overexpression of ASPR1 in transgenic plants. (A) Schematic representation of the DNA construct used to generate ASPR1 overexpression lines (oxASPR1). WT Arabidopsis plants were transformed with plasmid DNA that consisted on the GFP gene (600bp) fused to the 3’ end of ASPR1 full length gene (1400bp) under the control of the Cauliflower mosaic virus 35S promoter. BamHI and SalI restriction sites were used for cloning ASPR1 CDS into pAC1352 vector. (B) Western blot analysis of total protein extracts of 14-day-old oxASPR1 lines 2 and 4, as well as WT seedlings using an anti-GFP antibody.

12

Fig. S6. ASPR1 T-DNA insertion lines. (A) Schematic representation of ASPR1 gene structure showing the T-DNA insertion position in the different lines evaluated in this study. 1: Salk_038980.40.95.x; 2: Salk_072864.55.25.x; 3: WscDsLoxHs079_01H.0. (B) Semi- quantitative RT-PCR analysis of mRNA levels of ASPR1 in root tissue in T-DNA insertion lines 1, 2, 3, and WT plants. ASPR1 and Actin (loading control) specific primers were used for RT-PCR analysis.

13

Fig. S7. Root and lateral root phenotypes of aspr1 compared to WT seedlings. WT and aspr1 seedlings were grown vertically for 10 days in ½ MS media (A-F) or in modified ½ MS media without ammonium nitrate (G-L). Hypocotyl lengths were highly variable and thus not analyzed in detail in here. (B, G) Primary root length was measured at day 3, 5, 7 and 10. Each experiment

14

was repeated twice and in total 100-150 roots were measured per WT and aspr1 lines. (C, I) Total root length at day 10. (D, J) Relative total root length at day 10 (to each WT). (E,F and K, L) The number of lateral roots was counted under the dissection microscope. Absolute and relative number of lateral roots per cm of primary root, respectively. The error bars represent standard deviation from the mean (**** P <0.0001).

15

Fig. S8. Quantitative comparison of protein abundance in aspr1 and WT roots. (A) Overlap of proteins identified in three independent biological experiments. (B) Distribution of protein abundance ratios in aspr1 roots compared to WT, visualized for independent biological experiment. Whiskers extend to 5th and 95th percentile. (C) Volcano plot highlighting proteins significantly enriched (red, n = 67) or reduced (blue, n = 66) in abundance. A LIMMA-moderated t-test with p-value < 0.05 and abundance change > 20% were chosen as cut-offs. (D) GO-term classification of proteins with significantly altered abundance. Only GO-biochemical pathway (BP) categories with at least four associated proteins are shown.

16

Supplementary Tables

Table S1. Primer sequences used during the experimental work

Primer number Sequence (5’ - 3’) 1* GGA TCC ATG GCT TCT TCT TCT TCT TCT (forward) 2* GTC GAC CAA TTT TCC ACA TTC AGT GGG (reverse) CCA GTA GAT GAC TCT GGT TCC ACG GGA CTT GAC (forward) 3 (mutation in bold) GTC AAG TCC CGT GGA ACC AGA GTC ATC TAC TGG (reverse) 4 (mutation in bold) 5* TCT AGA TTC ATA ACA ATG GAA C (forward) 6* GGA TCC GGT TTG TGT GTG TGT GAG TG (reverse) 7* GCA CGTCTCAAGGT AGA AGA TCA TTA ATC GAC CG (forward) GCA CGTCTCAAAGC CTA ATG GTG ATG GTG ATG GTG CAA 8* TTT TCC ACA (reverse) 9 ACA GGT GCA AGT TTA GAT GG (forward) 10 GAA AAT AGA CAT CCC ATT AGA (reverse) 11 CTG CAG ATG GCT TCT TCT TCT TCT TCT 12 GTC GAC AAG GTT TTT CGG AAG AGT ACG 13 GTC GAC CAA TTT TCC ACA TTC AGT GGG *Underlined: Restriction site sequences; Italic: Linker sequence; Bold: HisTag coding sequence

17

Table S2. Compounds tested on rASPR1 enzymatic activity.

Compound Concentration (mM) Pepstatin A 0.001 Bestatin 0.01 Protease Pefabloc 1 Inhibitors E-64 0.01 EDTA 10 DTT 1 GSH 2 GSSG 2 Redox Agents NAD 3 NADP 2 NADPH 2 dATP 1 dGTP 1 dTTP 1 dCTP 1 Nucleotides ATP 1 ADP 1 GTP 1 CTP 1 NaCl 20 CaCl2 2 Ions MnCl2 1 MgCl2 2

18

References

Boersema PJ, Raijmakers R, Lemeer S, Mohammed S, Heck AJ. 2009. Multiplex peptide stable isotope dimethyl labeling for quantitative proteomics. Nature Protocols 4, 484-494. Clough SJ, Bent AF. 1998. Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant Journal 16, 735-743. Colaert N, Helsens K, Martens L, Vandekerckhove J, Gevaert K. 2009. Improved visualization of protein consensus sequences by iceLogo. Nature Methods 6, 786-787. Craig R, Beavis RC. 2004. TANDEM: matching proteins with tandem mass spectra. Bioinformatics 20, 1466-1467. Gomez-Auli A, Hillebrand LE, Biniossek ML, Peters C, Reinheckel T, Schilling O. 2016. Impact of cathepsin B on the interstitial fluid proteome of murine breast cancers. Biochimie 122, 88-98. Huang DW, Sherman BT, Lempicki RA. 2009. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols 4, 44-57. Huber W, Carey VJ, Gentleman R, et al. 2015. Orchestrating high-throughput genomic analysis with Bioconductor. Nature Methods 12, 115-121. Nguyen VD, Hatahet F, Salo KEH, Enlund E, Zhang C, Ruddock LW. 2011. Pre-expression of a sulfhydryl oxidase significantly increases the yields of eukaryotic disulfide bond containing proteins expressed in the cytoplasm of E.coli. Microbial Cell Factories 10, 1-1. Ritchie ME, Phipson B, Wu D, Hu YF, Law CW, Shi W, Smyth GK. 2015. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Research 43. Rodermel SR, Abbott MS, Bogorad L. 1988. Nuclear-organelle interactions: Nuclear antisense gene inhibits ribulose bisphosphate carboxylase enzyme levels in transformed tobacco plants. Cell 55, 673-681. Schilling, Overall CM. 2008. Proteome-derived, database-searchable peptide libraries for identifying protease cleavage sites. Nature Biotechnology 26. Schilling O, auf dem Keller U, Overall CM. 2011. Factor Xa subsite mapping by proteome- derived peptide libraries improved using WebPICS, a resource for proteomic identification of cleavage sites. Biological Chemistry 392, 1031-1037.

19

Shteynberg D, Deutsch EW, Lam H, et al. 2011. iProphet: multi-level integrative analysis of shotgun proteomic data improves peptide and protein identification rates and error estimates. Molecular and Cell Proteomics 10, M111 007690. Tyanova S, Temu T, Cox J. 2016a. The MaxQuant computational platform for mass spectrometry-based shotgun proteomics. Nature Protocols 11, 2301-2319. Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, Mann M, Cox J. 2016b. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nature Methods 13, 731-740. Wessel D, Flugge UI. 1984. A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Analytical Biochemistry 138, 141-143.

20