Impairment in O-acetylserine-(thiol) lyase A and B, but not C, confers higher selenate sensitivity and uncovers role for A, B and C as L-Cys and L-SeCys desulfhydrases in Arabidopsis

The role of the cytosolic O-acetylserine-(thiol) lyase A (OASTLA), chloroplastic OASTLB and mitochondrion OASTLC in plant resistance/sensitivity to selenate was studied in Arabidopsis plants. Impairment in OASTLA and B resulted in reduced biomass, chlorophyll and soluble protein levels compared with impaired OASTL C and Wild-Type treated with selenate. The lower organic-Se and protein-Se levels followed by decreased organic-S, S in proteins and total glutathione in oastlA and oastlB compared to Wild-Type and oastlC are indicative that Se accumulation is not the main cause for the stress symptoms, but rather the interference of Se with the S-reduction pathway. The increase in sulfite oxidase, adenosine 5′-phosphosulfate reductase, sulfite reductase and OASTL activity levels, followed by enhanced sulfite and sulfide, indicate a futile anabolic S-starvation response to selenate-induced organic-S catabolism in oastlA and oastlB compared to Wild-Type and oastlC. Additionally, the catabolic pathway of L-cysteine degradation was enhanced by selenate, and similar to L-cysteine producing activity, oastlA and B exhibited a significant decrease in L-cysteine desulfhydrase (DES) activity, compared with WT, indicating a major role of OASTLs in L-cysteine degradation. This notion was further evidenced by sulfide dependent DES in-gel activity, immunoblotting, immunoprecipitation with specific antibodies and identification of unique peptides in activity bands generated by OASTLA, B and C. Similar responses of the OASTLs in Seleno-Cysteine degradation was demonstrated in selenate stressed plants. Notably, no L-cysteine and L-Seleno-Cysteine DES activity bands but those related to OASTLs were evident. These results indicate the significance of OASTLs in degrading L-cysteine and L-SelenoCysteine in Arabidopsis. Summary The cytosolic OASTLA and chloroplastic OASTLB have significantly higher desulfhydrase activity rates than the cytosolic DES1 and are able to degrade L-Cys and L-SeCys to sulfide and selenide, respectively in Arabidopsis.


ABSTRACT INTRODUCTION
shown in oastlA and oastlB mutants was significantly higher than in WT and oastlC ( Fig. 1A-D), 148 in spite of the lower Se level in the protein fraction of the oastlA and oastlB (Fig. 1E). 149 Since at high levels selenate and Se can act as pro-oxidants and cause oxidative stress 150 (Grant et al., 2011;Mroczek-Zdyrska and Wójcik, 2012), Se and selenate levels were determined 151 in the leaves. Yet, no higher total Se and selenate were detected in oastlA and oastlB as 152 compared to WT and oastlC (Supplemental Fig. S2B and C) which could be attributed to the 153 decreased biomass accumulation and higher chlorophyll degradation rates in oastlA and oastlB 154 mutants. These results indicate that the levels of Se in the protein fraction, total Se and selenate 155 in leaf tissue are not the only cause for the higher sensitivity to Selenate noticed in oastlA and 156 oastlB compared to WT and oastlC mutant.
WT leaves, and the enhancement of OASTLB in oastlA, OASTLA in oastlB and OASTLB in 178 AtoastlC as compared to their control unstressed plants (Supplemental Fig. S2A). 179 In-spite of the enhanced protein expression of OASTLB and C in the absence of selenate 180 and of OASTLB in the presence of selenate, the impairment in OASTLA resulted in a strongly 181 decreased cysteine production activity of OASTL in AtoastlA mutant, exhibiting a reduction in 182 activity by 87% and 72% as compared to WT under control and selenate treated conditions, 183 respectively. The AtoastlB mutant showed a significant decline of 28% in both conditions, while 184 in oastlC OASTL activity was similar to the activity in WT plants. Interestingly, the activity in 185 all the examined genotypes was significantly increased with selenate treatment, likely as the 186 result of the enhanced expression of at least one of the active OASTL proteins ( Fig. 2A). The 187 comparable activity rates to WT found in oastlC mutant might be the result of the enhanced 188 expression of the two major proteins, the OASTLA and B as compared to WT in control as well 189 as under stressed conditions (Supplemental Fig. S2A). Taken together, the enhanced OASTL 190 activity in leaves of selenate stressed WT and the OASTL mutants may indicate S-limitation type 191 response shown before in Arabidopsis leaves in response to limited S supply (Barraso et al., 192 1995;Hesse et al., 1999). 193 To examine this assumption the Cys, total glutathione and total organic S levels, as well 194 as the S level in protein fraction were detected. No differences were noticed in cysteine level 195 between WT and mutants in control and selenate treated plants, whereas in selenate treated 196 plants a higher cysteine level was evident with WT and oastlC compared to these plants grown 197 under control conditions (Fig. 2B). Determination of total glutathione revealed a decrease only in 198 oastlC leaves whereas the other mutants exhibited comparable levels to WT leaves in plants 199 grown in control conditions. In contrast, oastlA and oastB showed decreased total glutathione 200 level compared to WT and oastlC leaves in plants exposed to selenate. Significantly, all 201 genotypes exposed to selenate exhibited significantly lower glutathione levels than control plants were similar in the mutants compared to WT either in control or in selenate treated plants (Fig. 234 2B). Considering the lowest cysteine generating activity level of OASTL was evident in oastlA 235 and B (Fig. 2B), this result indicates a possible catabolic protein degrading activity which can act 236 as an additional source of free L-cysteine in oastlA and B. In support of this notion is the 237 significantly lower soluble protein content evident in oastlA and B compared to WT and oastlC 238 mutant grown with selenate in the growth medium (Fig. 3A).

239
Recently, a lower biomass accumulation rate, the result of reduced organic S content was 240 shown to be the consequence of higher L-cysteine degradation by desulfhydrase (DES) activity 241 of OASTL in the perennial halophyte Sarcocornia, but not in the annual halophyte Salicornia, 242 both fed with high sulfate (Kurmanbayeva et al., 2017). To examine whether OASTLs play a 243 role in cysteine degradation also in non-halophyte plants, we studied first the expression of the 244 cytosolic DES1 claimed previously to play an important role in cytosolic L-Cys degradation 245 (Álvarez et al., 2010). Unlike the OASTLs whose transcript levels were enhanced in WT plants in

250
Unlike the DES1 transcript abundance in WT and OASTLs mutants, but similar to the L-

251
Cys generation activities of the OASTLs ( Fig. 2A), the total L-Cys DES kinetic activity was 252 enhanced in all genotypes grown with selenate compared to unstressed control plants (Fig. 3C).

253
DES activity rate was significantly lower in oastlA and oastlB mutants compared to WT grown 254 under unstressed conditions, while when exposed to selenate oastlA exhibited a significant lower 255 activity rate than WT and the other genotypes (Fig.3C)   Notably, in spite of the lower or similar DES activity rate, oastlA and oastlB mutants 275 exhibited higher sensitivity to selenate stress, accumulating lower biomass. This can be attributed 276 to the lowest capacity of L-Cys biosynthesis by OASTL in oastlA and oastlB mutants grown with selenate that resulted in low levels of organic sulfur, total glutathione as well as low S in the 278 protein fraction ( Fig. 2A,C,D,E). This could result in S-starvation type responses such as the 279 higher sulfide level evident in these mutants compared to WT and oastlC when stressed by 280 selenate, exhibiting, excluding oastlC, higher sulfide than the control unstressed plants, (Fig. 281 3D). The higher sulfite level detected in these mutants compared to WT and oastlC grown with 282 selenate (Fig. 3E), shown before when detected on fresh weight basis to be toxic in Arabidopsis 283 and tomato plants (Brychkova et al., 2012;Brychkova et al., 2013;Yarmolinsky et al., 2014), is 284 likely another S-starvation type response and an additional cause for the mutants' sensitivity to 285 selenate. Interestingly, while under control unstressed conditions SO activity was similar in 286 mutants and in WT treated with selenate, the mutants exhibited a drastically higher SO activity 287 rate than WT (Fig. 3F). This enhanced SO activity in the mutants compared to WT is a typical 288 response of SO to sulfite increase as was shown before (Brychkova et al., 2012;2013;289 Yarmolinsky et al., 2013;2014). The selenate induced sulfite increase in oastlA and oastlB 290 mutants is most likely the result of the increased APR activity rate in these mutants compared to 291 WT and oastlC grown with selenate (Fig. 3G). The increase in APR activity rate in oastlA and     Table S1A-C). Together with the OAS-TL A peptides, a lower amount of 14 and 381 9 peptides of OASTLB were identified in the lowest sliced L-Cys and SeCys desulfhydrase 382 activity bands, respectively (Supplemental Table S1A PAGE. Yet, based on the identification by the use of the three mutants exposed to in-gel activity, 394 immunoblot and immunoprecipitation by the specific antibodies employed (Fig. 4), as well as the 395 majority of the overlapping unique peptides (Supplemental Table S1), we can conclude that 396 OASTLA, B and C are localized to the highest, second-highest and the lowest activity bands, 397 respectively of L-Cys and SeCys desulfhydrase activity.  Table S1). Considering that the number of theoretic tryptic peptides in DES1 is similar to that of 401 OASTLs (estimated by the use of Expasy: https://web.expasy.org/peptide mass/), these results 402 indicate that DES1 is a less abundant protein or it undergoes ionization. Yet, since the semi 403 quantitation was made by calculating the peak area of each peptide, whereas to overcome 404 differential ionization, the area of the protein was calculated as the average of the three most 405 intense peptides from each protein; the data indicate that OASTL A, B and C are more abundant 406 than DES1 protein (Supplemental Table S1D).

407
The less abundant DES1 protein in the activity bands does not necessarily indicate 408 decreased L-Cys and L-SeCys DES activity by the WT DES1 compared to the OASTLs. Yet, 409 firstly, the kinetic activity revealed a major significant decrease of 80 and 65% in the Cys-  Impairment in OASTLA and OASTLB leads to higher sensitivity to selenate 423 Non-specific incorporation of seleno-amino acids SeCys and SeMet into proteins instead 424 of cysteine and methionine is thought to be a main cause for Se induced toxicity in many plants 425 (Sors et al., 2005;Terry et al., 2000). Yet, lower Se in protein fraction was evident in apr2 KO 426 mutants exhibiting stronger sensitivity than WT to the Se treatment (Grant et al., 2011). 427 Similarly, determination of the total Se and Se content in protein fraction revealed a significantly ). Among the three Arabidopsis OASTL mutants grown for 2 weeks with 40µM selenate, a 438 decrease in biomass accumulation and chlorophyll content compared to WT was noticed in 439 oastlA and oastlB but not in oastlC mutant (Fig.1). This is likely since the cytosolic and 440 chloroplastic OASTL A and B respectively, exhibit significantly higher activity rates compared 441 with the mitochondrion localized OASTL C, the latter being responsible for only 5-12% of 442 OASTL activity in WT plants (Heeg et al., 2008;Kuske et al., 1996). The results demonstrate 443 that dissimilar to WT and impairment in OASTLC, in the absence of active OASTLA and 444 OASTLB, the selenate interference in S assimilation pathway leads to a stronger decrease in 445 organic S, S in protein fraction and total glutathione, and has a more negative effect on the 446 remaining chlorophyll level and the resulting biomass accumulation than the expected positive 447 effect of lower selenium content in the organic fraction (Figs 1 and 2).

448
The free cysteine levels were similar in WT and the mutants exposed to selenate 449 (Fig.2B), whereas the Cys generation activity rate differed as shown here in response to selenate 450 ( Fig.2A). This was also shown previously in an O-acetylserine (thiol) lyase reduced isoform by Riemenschneider et al., 2005), indicating the feasibility of an additional Cys source affecting free 453 Cys level. The source can be protein degradation, indicated by the lower soluble protein level in 454 oastlA and oastlB KO mutants compared to WT and oastlC treated with selenate (Fig. 3A). The 455 protein degradation is the result of selenate induced stress, being stronger in oastlA and B as 456 indicated by the lower remaining chlorophyll (Fig. 1D), organic-S, and S in protein fraction (Fig. 457 2D and E).

458
Sulfite is a strong nucleophile that should be tightly regulated to avoid its deleterious   (Fig. 4C), as well as by immunoprecipitation assay (Fig. 4D). Finally it was supported by the 508 identification of the tripsinized unique peptides (Supplemental Table S1) in the three sliced WT 509 activity bands (Fig.4A and B). Importantly, DES1 was shown to degrade L-Cys, estimated by the use of the des1 null mutants 511 to be responsible for up to 14% of the total L-Cys desulfhydrase activity compared to WT plants 512 grown in plated on MS medium (Álvarez et al., 2010). This suggests the existence of additional 513 source/s that can contribute to the majority (ca 86%) of H 2 S production by degrading L-Cys.
Notably, the impairment of oastlA and oastlB mutants led to a significantly decreased sulfide 515 production activity rate by L-Cys desulfhydrase [ca 76% and 18% reduction respectively 516 (Fig3C)] compared to WT, indicating the important role of the OASTLs in degrading L-Cys.

517
Intriguingly, only very few unique peptides of DES1 were identified in the major activity bands, 518 the highest and second-highest mobility bands, of L-Cys desulfhydrase activity and none in the 519 SeCys desulfhydrase activity bands, indicating that the OASTLs are more abundant proteins than 520 DES1 (Supplemental Table S1). Taking together i) the lower abundancy of DES1 compared to Additionally it was shown that the absence of active OASTLA or OASTLB confers higher 535 sensitivity to selenate compared to WT or the absence of active OASTLC. Further, it was shown 536 here that the cytosolic OASTLA, chloroplat localized OASTLB and the mitochondrion localized 537 OASTLC has an important role in desulfhydrase activity degrading not only L-Cys but also 538 SeCys to sulfide and selenide, respectively.

557
To quantify the transcripts using quantitative reverse transcriptase-polymerase chain reaction 558 (qRT-PCR), total RNA was prepared using the Aurum TM total RNA Mini Kit (Bio-Rad,  For detection of remaining content of chlorophyll (%) the leaf discs (7 mm diameter) 572 were immersed in 90% ethanol and incubated at 4°C for 2 d in the dark. Absorbance of the 573 extracted chlorophyll was measured at 652 nm, and total chlorophyll was estimated (Ritchie, For detecting H 2 O 2 , frozen leaves were extracted in 50 mM P buffer (pH 7.5) at a ratio of 576 1:8 (w/v) and centrifuged (Eppendorf 5417R) twice at 18,000g for 20 min. The reaction mixture 577 for detecting H 2 O 2 consisted of 0.85 mm 4-aminoantipyrine, 3.4 mm 3,5-dichloro-2-578 hydroxobenzene sulfonate, 4.5 U mL −1 HRP in 2 mL of 50 mM P buffer (pH 7.5) in the presence 579 or absence of 2 mM tungstic acid and 100 μM DPI as described in Yesbergenova et al. 2005. 580 Absorbance was measured after 5 min at 515 nm as described above.

581
The anthocyanin content was determined based on a modification of protocols described  To determine the total selenium content, 100 mg of the dried and powdered leaves were 647 placed in glass tubes, digested with 70% HN0 3 , heated at 220 O C for 4 hours and quantified by 648 inductively coupled plasma emission spectrometry (ICP-AES) (Kalra, 1997).

651
The homogenate was cleared by centrifugation (7,500g for 10 min). A small sample was taken 652 for protein determination, and the volume of the extract was measured. The proteins in the 653 extract were precipitated by adding TCA to a final concentration of 15% (w/v). The mixture was 654 incubated on ice for 30 min and then centrifuged for 20 min at 7,000g at 4°C. The pellet was 655 washed with ice-cold acetone, dried, and dissolved in 1 mL of concentrated nitric acid. After acid 656 digestion, the Se was determined by ICP (Pilon-Smits et al., 1999). To identify the proteins participating in the cysteine desulfhydrase activity, the activity bands 677 from the in-gel activity of DES were sliced from the native gel, and fractionated with 12.5% 678 SDS-PAGE (Fig. 4). Thereafter the proteins were stained by Coomassie Brilliant Blue, and the 679 stained bands were excised from the gel, trypsinized and the resulting peptides were separated by  Supplemental Table S1. Identified and overlapped unique peptides from L-Cys and L-SeCys 709 desulfhydrase activity. Supplemental Table S2. List of gene primers used for quantitative real-time PCR. 710

ACKNOWLEDGMENT: 711
We thank Prof. Dr. S. Kopriva (University of Cologne, Köln) and Prof. R.Hell, (Universität   The relative expression in the Atdes1 with the 2 coupled primers was analyzed using the normalized WT 787 control as reference. In-gel L-Cysteine desulfhydrase activities were employed after protein fractionation 788 in SDS gel performed in    Tables  806   Supplemental Table S1. Identified and overlapped unique peptides from L-Cys and L-SeCys 807 desulfhydrase activity. A. List of the number coverage percentage B. Identified and overlapped unique 808 peptides from L-Cys desulfhydrase activity. C. Identified and overlapped unique peptides from L-SeCys 809 desulfhydrase activity. D. Abundance of OASTL A, B and C as compared to DES1 protein at the lower, 810 middle and upper activity bands. The semi quantitation was made by calculating the peak area of 811 each peptide, whereas to overcome differential ionization, the area of the protein was calculated

Upper middle lower
Immunoprecipitation Native-SDS with L-Cys Native-SDS with SeCys