# Source paper archive - Source title: Redox regulation of 2-carboxy-D-arabinitol 1-phosphate phosphatase in Fabaceae is associated with high Rubisco dark inhibition - Author(s): Connor Nehls-Ramos, Elizabete Carmo-Silva, Douglas J Orr - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.05.764501v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.05.764501v1.full.pdf - PDF SHA-256: 1db844eff307bdb2c7a8c174648454674b46f7887f518dfe40b9a83ad81b8484 - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1 **Redox regulation of 2-carboxy-D-arabinitol 1-phosphate phosphatase in** 2 **Fabaceae is associated with high Rubisco dark inhibition** 3 4 **Authors:** Connor Nehls-Ramos, Elizabete Carmo-Silva, Douglas J. Orr 5 Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ, UK 6 7 **Emails:** 8 c.nehls@lancaster.ac.uk 9 e.carmosilva@lancaster.ac.uk 10 d.j.orr@lancaster.ac.uk 11 12 **ORCID ID:** 13 Connor Nehls-Ramos (0009-0000-3156-4819) 14 Elizabete Carmo-Silva (0000-0001-6059-9359) 15 Douglas J. Orr (0000-0003-1217-537X) 16 17 **Author for correspondence:** d.j.orr@lancaster.ac.uk 18 **Date of Submission:** **05/11/2026** 19 **Table Count:** 1 20 **Figure Count:** 5 21 22 **Word Count:** **4321** 23 **Supplementary Data:** In file or otherwise accessible by Github: 24 cwnehls/NehlsRamos_etal_JxB_2027: Supplementary data and code for the 25 publication Nehls-Ramos _et al._ 2027 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 26 **Short Running Title:** Cysteine pair motif regulates CA1Pase 27 28 **Highlight:** 29 A cysteine pair motif confers redox regulation to a chloroplast phosphatase 30 that degrades Rubisco inhibitors. _In silico_ and _in vitro_ analyses suggest this 31 mechanism correlates with Rubisco dark inhibition in Fabaceae. 32 33 **Abstract:** 34 Improving Rubisco regulation is an avenue for improving crop output. 35 Minimizing the buildup of Rubisco inhibitors can allow improved catalytic activity and 36 CO 2 assimilation. The enzyme 2-carboxy-D-arabinitol 1-phosphate phosphatase 37 (CA1Pase) degrades certain Rubisco inhibitors, yet the mechanisms for its 38 regulation, including the function of its C domain, are not well understood. In this 39 study we analysed _Vigna unguiculata_ (cowpea) CA1Pase _in silico_ and _in vitro_ for 40 structural regions and motifs related to regulation of phosphatase activity. We 41 identified a highly conserved cysteine pair motif tied to redox regulation and potential 42 homodimer formation. Variations in one or both cysteine residues in this motif across 43 species correlate with lower dark inhibitor accumulation levels as compared to 44 related species. The C domain is identified as having a regulatory function and 45 potential evolutionary origins are explored. These insights may inform future efforts 46 toward optimising Rubisco’s complex regulatory processes including inhibitor 47 degradation. 48 49 **Keywords:** 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 50 CA1P, CA1Pase, dark inhibition, enzyme activity, inhibitors, photosynthetic 51 metabolism, photosynthesis, Rubisco, sugar-phosphates, chloroplast, redox 52 53 **Abbreviations:** 54 CA, 2-carboxy-D-arabinitol 55 CA1P, 2-carboxy-D-arabinitol 1-phosphate 56 CA1Pase, 2-carboxy-D-arabinitol 1-phosphate phosphatase 57 CBB, Calvin-Benson-Basham 58 CRBP, 2-carboxy-D-ribitol-1,5-bisphosphate 59 cTP, chloroplast transit peptide 60 FBP, fructose 1,6-bisphosphate 61 GSH, reduced glutathione 62 GSSG, oxidised glutathione 63 PDBP, D-glycero-2,3-pentodiulose-1,5-bisphosphate 64 PGM, phosphoglycerate mutase 65 PSP1, phosphoserine phosphatase 1 66 Rca, Rubisco activase 67 ROS, Reactive oxygen species 68 RuBP, ribulose 1,5-bisphosphate 69 TSP, total soluble protein 70 VuCA1Pase, _Vigna unguiculata_ 2-carboxy-D-arabinitol 1-phosphate phosphatase 71 72 **Introduction** 73 One strategy to increase crop output is by improving photosynthesis through 74 fixing inefficiencies associated with Rubisco, the imperfect enzyme essential to 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 75 carbon fixation in photosynthesis (Parry _et al._, 2013; Croce _et al._, 2024). The activity 76 of Rubisco is regulated in response to environmental factors such as light availability, 77 which is impacted within the crop canopy by cloud cover, wind speed and sunflecks 78 frequency and duration, as well as sun movement across the day (Yamori, 2016; 79 Slattery _et al._, 2018). Despite advances in understanding the complex regulation of 80 Rubisco activity, there remain many knowledge gaps (Parry _et al._, 2008; Leister _et al._, 81 2023; Orr _et al._, 2023; Amaral _et al._, 2024; Pasch _et al._, 2025). While it has long 82 been known that certain sugar phosphate inhibitors bind tightly to Rubisco catalytic 83 sites and prevent catalysis throughout the day, it is less clear what factors regulate 84 their abundance and degradation (Parry _et al._, 2008; Orr _et al._, 2023; Pasch _et al._, 85 2025). 86 For many but not all crops, accumulation of nighttime Rubisco inhibitors, 87 thought to be primarily 2-carboxy-D-arabinitol 1-phosphate (CA1P), has been 88 reported when leaves experience an extended period of low light or darkness, as 89 seen in a recent compilation of dark inhibition across angiosperms (Nehls-Ramos _et_ 90 _al._, 2026). For cowpea specifically, this dark inhibitor has been found to accumulate 91 to significant amounts (Servaites _et al._, 1986). There are two known chloroplast 92 localised phosphatases which dephosphorylate Rubisco inhibitors, 2-carboxy-D 93 arabinitol 1-phosphate phosphatase (CA1Pase; EC: 3.1.3.63) and xylulose-1,5 94 bisphosphate phosphatase (Parry _et al._, 2008; Bracher _et al._, 2015). CA1Pase 95 dephosphorylates not only the diurnal inhibitor CA1P, but also Rubisco misfire 96 products including D-glycero-2,3-pentodiulose-1,5-bisphosphate (PDBP) after 97 release by Rubisco activase (Parry _et al._, 2008; Andralojc _et al._, 2012). 98 CA1Pase is a histidine phosphatase and part of the ancient phosphoglycerate 99 mutase (PGM) family. This enzyme is believed to reside solely in the chloroplast 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 100 stroma (Salvucci _et al._, 1988; Holbrook _et al._, 1989; Gutteridge & Julien, 1989; 101 Moore _et al._, 1995; Bayer _et al._, 2011). The mature enzyme in flowering plants 102 typically ranges between 440-460 amino acids in length, with a monomeric size near 103 50kDa. A previous study identified two distinct domains of similar length in wheat 104 and Arabidopsis (Andralojc _et al._, 2012). The N domain contains the histidine 105 phosphatase catalytic region, in which reside five main catalytic residues: two 106 arginines (R), two histidines (H), and a glutamic acid (E) (Rigden, 2007; Andralojc _et_ 107 _al._, 2012). The role of the C domain has not been defined, though it has been 108 previously hypothesised to be a phosphofructokinase-like domain (Andralojc _et al._, 109 2012). 110 Despite nearly 40 years of investigation, insights on the _in vivo_ role and 111 regulation of CA1P and CA1Pase are limited, especially regarding the intersection of 112 enzyme activity regulation and structural biology (Salvucci _et al._, 1988; Holbrook _et_ 113 _al._, 1989; Gutteridge & Julien, 1989; Salvucci _et al._, 1989; Holbrook _et al._, 1991; 114 Kingston Smith _et al._, 1992; Moore _et al._, 1995; Charlet _et al._, 1997; Heo & Holbrook 115 1999; Andralojc _et al._, 2012; Lobo _et al._, 2019). Regulation of CA1Pase has been 116 directly investigated in five species: tobacco ( _Nicotiana tabacum_ & _N. rustica_ ), 117 French bean ( _Phaseolus vulgaris_ ), spinach ( _Spinacia oleracea_ ), and wheat ( _Triticum_ 118 _aestivum_ ), with some other investigations into the kinetics in other species (Table 1; 119 Charlet _et al._, 1997). While there are large variations in dark inhibitor accumulation in 120 plants, CA1Pase appears to be ubiquitous across angiosperms, and may be present 121 in more ancient lineages including algae (Andralojc _et al._, 2012; Nehls-Ramos _et al._, 122 2026). An ancient CA1Pase would imply a role deeply entrenched in chloroplast and 123 Rubisco homeostasis, which may explain why modifications to its abundance have 124 caused unexpected shifts in Rubisco abundance in plants (Lobo _et al._, 2019). To 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 125 better understand the importance of CA1Pase in the chloroplast as well as retained 126 features in structure, its potential ancestral origins should be explored further. 127 While studies show conflicting results within and across species for factors 128 effecting _in vitro_ regulation of CA1Pase activity, some common trends are present 129 (Table 1). CA1Pase shows a pH optima around or below 7.5, suggesting peak 130 activity when chloroplasts are transitioning between darkness and high light (Trinh & 131 Masuda, 2022). Several Calvin Benson Basham (CBB) cycle related metabolites 132 have been shown to act as positive effectors of CA1Pase, consistent with a 133 stimulation of activity as photosynthetic metabolism increases in the light (Table 1). 134 One such effector is fructose-1,6-bisphosphate (FBP), a known precursor for the 135 formation of 2-carboxy-D-arabinitol (CA) and CA1P (Andralojc _et al._, 2002). Evidence 136 suggests CA1Pase partially deactivates in oxidising conditions and reactivates in 137 reducing conditions (Salvucci _et al._, 1988; Holbrook _et al._, 1989; Gutteridge & Julien, 138 1989; Salvucci _et al._, 1989; Holbrook _et al._, 1991; Kingston Smith _et al._, 1992; 139 Charlet _et al._, 1997; Heo & Holbrook, 1999). This has been hypothesised to occur 140 due to a key pair of cysteine residues, which remain unidentified despite availability 141 of CA1Pase sequences more than a decade later (Heo & Holbrook,1999; Andralojc 142 _et al._, 2012). 143 In the chloroplast, redox regulation is key to linking photosynthetic metabolism 144 with electron transport chain activity during photosynthesis (Yu _et al._, 2020; Yoshida 145 & Hisabori, 2023; Müller-Schüssele, 2024). Several CBB enzymes, including 146 Rubisco activase (Rca) and fructose-1,6-bisphophatase, have been shown to be 147 subject to this type of regulation (Portis _et al._, 2008; Gütle _et al._, 2016). During active 148 photosynthesis, the electron transport chain transfers photon energy into chemical 149 energy by generating ATP and a proton gradient via photosystem II, as well as 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 150 allowing for the generation of NADPH from NADP+ in photosystem I (Yoshida & 151 Hisabori, 2023; Trinh & Masuda, 2022). The NADPH generated allows for the 152 reduction of thioredoxin (Trx) and glutathione (GSH)—or homoglutathione (hGSH) in 153 Fabaceae—which in turn regulates redox sensitive enzymes through the breaking of 154 key disulphide bonds and the formation of mixed disulphides by S-glutathionylation 155 (Musaogullari & Chai, 2020). These stromal reducing agents also buffer for reactive 156 oxygen species (ROS), such as H 2 O 2, which are generated during photosynthesis 157 and via cell signalling pathways (Foyer & Hanke, 2022). During dark periods, the 158 reducing potential of the stroma greatly diminishes, resulting in the reversible 159 oxidation of stromal enzyme cysteine residues (Müller-Schüssele _et al._, 2020; 160 Cejudo _et al._, 2021). This cycling of oxidising (dark and low light) and reducing (light) 161 conditions allows for a cyclical regulation of photosynthetic enzymes. 162 In this study we elucidate the mechanism for redox regulation in CA1Pase 163 and clarify the function of the regulatory C domain, using CA1Pase from the staple 164 legume cowpea ( _Vigna unguiculata_ ). We use a combination of _in silico_ sequence, 165 predicted structure analyses, _in_ _vitro_ activity and oligomerisation measurements of 166 recombinant cowpea CA1Pase (VuCA1Pase). Key cysteines were identified _in silico_ 167 and mutated to investigate their effect on redox regulation of CA1Pase activity. We 168 propose that oxidation of this cysteine pair motif (CXXXXC) in CA1Pase is required 169 for the high levels of Rubisco dark inhibition observed in flowering plants. 170 171 **Materials and methods** 172 CA1Pase sequences & identification 173 The CA1Pase sequence was sourced from the cowpea genome (Phytozome 174 13 database; Goodstein _et al._, 2012) for _Vigna unguiculata_ 1.2 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 175 (ViGun11g130800.1.10) using known CA1Pase sequences from _Arabidopsis_ 176 _thaliana_ and _Triticum aestivum_ (Andralojc _et al._, 2012). To generate the mature 177 sequence, the chloroplast transit peptide (cTP) predicted by Andralojc _et al._ (2012) 178 was removed preceding the “KRVV…” motif. Key residues are numbered throughout 179 using the mature protein sequence of cowpea CA1Pase (VuCA1Pase). Homologous 180 CA1Pase sequences were further identified by performing a NCBI BLAST search 181 against ViGun11g130800.1.10 to gather the top 1000 sequences related to the 182 cowpea CA1Pase sequence, including searches for sequences identified in Uniprot 183 with “2-carboxy-D-arabinitol 1-phosphatase” as the protein name (Sayers _et al._, 184 2025; The UniProt Consortium, 2025). The resulting sequences were curated for 185 length between 400-600 amino acids, aligned using Geneious MUSCLE alignment 186 and manually trimmed to remove those lacking the RHGQSTWN catalytic motif or 187 the R/CXG and XXNXTXHL motifs unique to the Regulatory domain—highly 188 conserved across all complete sequences. Using a consensus tree alignment, the 189 sequences grouped into three main branches: cyanobacteria and chlorophytes, 190 bryophytes, and all other streptophytes. The latter contains a branch with all 191 angiosperm sequences (Supplementary Dataset S1). 192 193 Predicted structure modelling and visualisation 194 Structural predictions were performed using Deepmind AlphaFold 3, with 195 comparisons against D-I-Tasser modelling (Abramson _et al._, 2024; Zheng _et al._, 196 2026). Multimeric structures, including ligand interactions, were also modelled in 197 AlphaFold 3. Similar predicted structures for the monomer and dimer forms of 198 VuCA1Pase can be found with cTP included on the AlphaFold DB for _Vigna_ 199 _unguiculata_ (AF-A0A4D6LHI6-F1-v6; AF-0000000079567473-v1; Bertoni _et al._, 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 200 2026). Individual predicted structures depicting model confidence, including 201 predicted local distance difference test, can be found in Supplementary Fig. S1 and 202 Supplementary Datasets S2 and S3. Structures were visualised and annotated in 203 PyMol (version 2.5.7, Schrödinger, LLC., USA; www.pymol.org). Sequence 204 alignments and trees were generated using Geneious (Geneious 9.1.8; 205 www.geneious.com). 206 207 Cloning of CA1Pase variants and expression vector creation 208 The mature VuCA1Pase sequence was inserted into a 6xHis TEV vector 209 (pETHisTEVGG) (Gjindali _et al._, 2025). The domain boundary for designing the 210 regulatory domain only variant, VuC 207-445, was determined using D-I-Tasser 211 modelling (Supplementary Table S4). Modified cysteine variants are based on the 212 6xHis TEV VuCA1Pase sequence with one or both cysteine residues at C292 and 213 C297 changed to Alanine. 214 To create the expression construct for the wild type form of Cowpea CA1Pase 215 (Vigun11g130800.1), the oligonucleotide sequence, as well as the regulatory domain 216 variant, was ordered from Aruru Molecular (United Kingdom) and amplified using 217 primers ordered from Integrated DNA Technologies. The amplified sequence was 218 inserted into the pETHisTEVGG golden gate compatible plasmid and expressed in 219 BL21(DE3)pLysS _E. coli_ with expression inducible T7 promoter. The Cysteine 220 mutant variants were generated using the QuikChange II Site-Directed Mutagenesis 221 Kit (200523; Agilent Technologies, Inc., Santa Clara, California, US; 222 www.agilent.com), as per manufacturer specifications. Sequence inserted 223 expression vectors were verified by Plasmidsaurus Whole Plasmid Sequencing 224 (Oxford Nanopore, R10.4.1; Plasmidsaurus, UK; www.plasmidsaurus.com) and 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 225 Source Bioscience (Nanopore Whole Plasmid Sequencing; Cambridge, UK; 226 www.sourcebioscience.com). 227 228 CA1Pase expression, purification and quantification 229 Transformed BL21(DE3)pLysS cells were expressed and the recombinant 230 protein purified using the same method as per Gjindali _et al._ (2025). Purified protein 231 was stored in 50 mM HEPES, pH 8.0, and 5% glycerol and aliquoted prior to flash 232 freezing. Total soluble protein (TSP) was determined using the Bradford assay as 233 well as 280 nm absorption via a CLARIOstar Plus spectrophotometer with an LVis 234 plate (BMG Labtech, Ortenberg, Germany; www.bmglabtech.com) (Perdomo _et al._, 235 2018; Supplementary Dataset S5). For soluble protein calculations determined from 236 the absorbance at 280 nm absorption, absorption coefficients for the 6xHis TEV 237 tagged proteins were 63940 M [-1] cm [-1] for VuCA1Pase, 33460 M [-1] cm [-1] for VuC 1-206, 238 and 31970 M [-1] cm [-1] for VuC 207-445 as determined using the Expasy Prot Param 239 webtool (Gasteiger _et al._, 2005, https://web.expasy.org/protparam/). Protein purity 240 determination by SDS-PAGE gel and western blot were performed as per Perdomo 241 _et al._ (2018). Western blots used a mouse anti His antibody (Bio-Rad 242 Cat#MCA1396D800GA, RRID: AB_11152596; www.bio-rad-antibodies.com. All gels 243 were imaged using a Licor Odyssey Fc Imager (Li-Cor Bioscience, USA; 244 www.licorbio.com). Purity of protein was quantified by densitometry using the Licor 245 software ImageStudio comparing the peak density of the expressed protein in the 246 Coomassie stained SDS-Page gels as a percentage of the remaining 700 nm 247 absorption in the lane. Purity for all assayed proteins was >50% (Supplementary Fig. 248 S2; Supplementary Table S2). 249 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 250 CA1Pase activity assays 251 All reagents in this section and below were sourced from Sigma-Aldrich (UK; 252 www.sigmaaldrich.com) unless otherwise specified. Activity was quantified using a 253 modified version of a malachite phosphate detection assay (Martínez Gache _et al._, 254 2020). The colour detection reagent used (Supplementary Fig. S3) was modified to 255 use malachite chloride in place of malachite green carbinol chloride (the latter being 256 unavailable from commercial sources at the time). Final concentrations of reagents 257 were: 14.24 mM ammonium molybdate tetrahydrate (A7302), 501.5 µM malachite 258 chloride (38800), and 1.19 mM pluronic F68 in a solution of 16.8% (v/v) HClO 4 259 (1.00517.0250), 0.5% (v/v) absolute ethanol (Fisher-Scientific, USA, E/0650DF/17), 260 and the remainder milli-Q water. The solution was then filtered with ashless filter 261 paper (Whatman, UK; www.cytivalifesciences.com) and stored in a brown 262 polyethylene bottle. This solution is dark yellow-orange, clear, and can be used for 263 up to one year. 264 Thawed protein was re-checked for TSP by A280, diluted in buffer (50 mM 265 HEPES [Melford, H75030, UK], pH 8, and 200 mM NaCl [31434]), factoring for the 266 purity of the sample, to a concentration of 1.186 μ M of recombinant CA1Pase. A 267 reaction buffer was prepared from a 10x stock (500 mM HEPES, pH 8.0, 50% 268 glycerol w/v [G7757]) diluted in Mili-Q water. For assays involving glutathione, 269 200mM reduced (GSH) (G4251) and oxidised (GSSG) (G4376) glutathione stocks 270 prepared in Mili-Q water were combined in differing ratios and included in the final 271 buffer at a final concentration of 10mM. The sample buffer was added to a 96 well 272 plate at 25°C on a plate incubator at a volume of 57 μ L. In the same plate, a set of 273 standards that contained 0.1-5 nmoles of pH 7.0 NaPi (0.4167 mM NaH 2 PO 4 [S5011] 274 and 0.5833 mM Na 2 HPO 4 [S0876] for a 1 mM solution) were prepared and used to 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 275 quantify release of phosphate. The reaction substrate 2-carboxy-D-ribitol 1,5 276 bisphosphate (CRBP) was used in place of CA1P as a more easily accessible 277 reagent that is dephosphorylated by CA1Pase at similar rates and has been used as 278 a substitute in other studies (Andralojc _et al._, 2012; Lobo _et al._, 2019). CRBP was 279 added at 3 μ L and 20 mM to initiate the reaction with 500 rpm agitation. Reactions 280 were quenched using 40 μ L of the colour detection reagent after 10 minutes, unless 281 specified otherwise. One hour after the reaction was quenched, colour development 282 was quantified using a microplate reader for absorption at 640 nm (Supplementary 283 Fig. S3). 284 285 Size exclusion chromatography 286 An ENrich 650 column (BioRad, #7801650, USA) was equilibrated with 2 287 column volumes (CVs) of a 50 mM HEPES, pH 8, 200 mM NaCl buffer prior to each 288 run. For runs that included reduced glutathione, 10 mM GSH was included in the 289 buffer by adding 1 mL of 200mM GSH for every 19 mL of buffer. 100-200 μ L of 290 purified enzyme (~20-100 μ M) was injected into the column and ran at a flow rate of 291 0.5 mL min [-1] . Protein Standards (Protein Standard Mix 15 - 600 kDa, #69385, 292 Supelco, USA) were used to quantify protein peak sizes using the same flow rate 293 (Supplementary Table S3). For cases requiring further quantification, diluted 294 fractions were collected at 10 °C in 1.5 mL micro tubes (Sarstedt, Germany, 295 72.690.001) to quantify protein, TSP, and activity as described above. Some proteins 296 were run at 0.125 mL min [-1 ] in attempt to improve peak resolution, however this did 297 not appear to impact peak separation or elution volume. The program findpeaks() 298 from the R package pracma (Borchers 2023) was used to identify all peaks, with 299 visual confirmation (Supplementary Dataset S6). 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 300 301 Data analyses 302 Geneious 9.1.8 was used to generate protein alignments for Fig. 1b, Fig. 2a, 303 Fig. 2b, and Fig 5. Further data was processed in R 4.4.2 (R core team, 2024) and 304 RStudio 2025.05.0 (Posit Team, 2025). Data was processed and graphs were 305 prepared using ggplot 2 (Wickham, 2016), agricolae (de Mendiburu, 2023), dplyr 306 (Wickham _et al._, 2023a), readr (Wickham _et al._, 2024), viridis (Garnier _et al._, 2024), 307 pracma (Borchers, 2023), multcomp (Hothorn _et al._, 2008), multcompview (Graves _et_ 308 _al._, 2024), forcats (Wickham, 2023b), FSA (Ogle _et al._, 2025), and rcompanion 309 (Mangiafico, 2025) R packages. For scatter plots, dots represent technical replicates, 310 with mean lines and error bars indicating standard error. All code and data can be 311 found on github (https://github.com/cwnehls/NehlsRamos_etal_JxB_2027). 312 313 **Results** 314 **CA1Pase domains are structurally similar, with C domain region lacking key** 315 **phosphatase catalytic residues** 316 To better understand the role and function of the two putative domains of 317 CA1Pase, the structure and sequence of cowpea CA1Pase was compared with 318 related histidine phosphatase family proteins (Fig. 1). These enzymes were 319 screened through structural search by D-I-TASSER (Zheng _et al._, 2026; 320 Supplementary Dataset S4). The two domains previously identified in CA1Pase, a 321 catalytic N domain and C domain (Andralojc _et al._, 2012), resemble those in the 322 predicted cyanobacterial CA1Pase-like structure (Fig. 1a; Supplementary Table S4; 323 Supplementary Table S5). While the catalytic N domain in both proteins strongly 324 resembles the bacterial monomer of PhoE, a broad specificity bacterial phosphatase, 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 325 their complete monomeric structures parallel the homodimer form of the bacterial 326 phosphoserine phosphatase 1 (PSP1), (Supplementary Table S5). Previously the C 327 domain of CA1Pase was predicted to be a phosphofructokinase domain (Andralojc 328 _et al._, 2012). The NCBI Conserved Domain database, however, identifies the domain 329 as a general histidine phosphatase-like phosphoglycerate mutase (PGM) 330 (COG0406; Lu _et al._, 2020). 331 For all four proteins, the known catalytic residues - two arginines (R), two 332 histidines (H), and one glutamate (E) - are conserved in the catalytic domains (Fig. 333 1B). These are conserved in the C domain for the cyanobacterial CA1Pase-like 334 enzyme, but not in the C domain for VuCA1Pase, which retains only the last histidine, 335 H375. While similar to the structure of the catalytic domain of PSP1, the C domain 336 has several differences: an extended alpha helix on the side facing the catalytic 337 domain opening and a lack of conserved alpha helices in the region mirroring the 338 catalytic pocket of the cyanobacterial C domain (Fig. 1A). The structural similarity 339 that is retained may allow for allosteric binding of ligands in the predicted 340 pseudophosphatase pocket (Supplementary Fig. S4). Due to this and the expected 341 lack of catalytic capacity, we investigated if the C domain may function as a 342 regulatory domain. 343 **Conserved domain structures suggest an ancient gene duplication and fusion** 344 **event** 345 To identify key and unique regions of the regulatory domain, the two domains 346 were investigated via superimposition of predicted structures and AlphaFold 347 sequence alignments. To then establish key residues and potential ancestral motif 348 shifts within these domains, a selection of CA1Pase and CA1Pase-like sequences 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 349 were identified by BLAST from organisms ranging from cyanobacteria to plants for 350 comparison. 351 Initial alignment of the two domains by sequence and secondary structure 352 reinforced observed similarities of the regulatory domain and the C domain of the 353 cyanobacterial sequence (Fig. 2A). While the pairwise identity of residues between 354 the domains is low (~23%), greater identity is conserved in the secondary structure. 355 All five beta sheets in the catalytic domain are conserved in the regulatory domain 356 along with five of nine alpha helices. Superimposing the structures reveals high 357 conservation in the position of these secondary structures (Fig. 2B). Interestingly, 358 shortly after the extended alpha helix (A11) is a conserved cysteine pair motif 359 (CXXXXC) predicted to form a disulphide bond (Fig. 1A-B). This motif is not present 360 in the catalytic domain, suggesting a unique potential redox role for the regulatory 361 domain. 362 A phylogenetic tree comparing the CA1Pases identified by BLAST highlighted 363 high conservation of the catalytic domain active site residues from cyanobacteria 364 through to land plant lineages (Fig. 2C). This was not true for the equivalent residues 365 in the regulatory domain: the first arginine is only conserved in green algae; the first 366 catalytic histidine is retained only in certain lineages down to flowering plants; the 367 second arginine is lost with the onset of land plants; the key glutamate is present 368 solely in the cyanobacterial sequence. Only the previously mentioned second 369 histidine, H375, is conserved throughout all lineages. The CXXXXC motif is highly 370 conserved beginning with the charophytic, suggesting redox regulation for CA1Pase 371 may originate within Streptophytes. The lack of conserved residues across the 372 regulatory domain, combined with the less well predicted structure of the region that 373 mimics the catalytic substrate binding pocket, may suggest that substrate binding 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 374 specificity could be tied to individual lineages. This may, in turn, allosterically 375 regulate CA1Pase activity depending on species or lineage differences in stromal 376 metabolite content. 377 **Structural predictions and** _**in**_ _**vitro**_ **analyses suggest the regulatory domain** 378 **influences CA1Pase dimer formation** 379 Several enzymes in the chloroplast form multimeric structures that can 380 regulate catalytic activity (Podesta _et al._, 1990; Yu _et al_ ., 2020; Waheeda _et al_ ., 381 2023). To investigate potential regulatory oligomer formation, predicted multimeric 382 structures for mature VuCA1Pase were generated and stable forms identified 383 (Supplementary Fig. S1). Beyond the monomer, the homodimer was predicted to be 384 the most stable form. Sequences for CA1Pases from 252 species were aligned to 385 identify well conserved residues that may support dimer formation (Supplementary 386 Dataset S1). Residues conserved in over 95% of those sequences are highlighted in 387 the predicted dimer structure (Fig. 3A). Of the conserved residues in the regulatory 388 domain, the majority are located in the predicted dimer interface, suggesting a well 389 conserved ability to form dimers across flowering plants i _n vivo_ . The previously 390 identified CXXXXC motif was also found to be part of this dimer interface (Fig. 3B). 391 The predicted surface facing position of the disulphide bond suggests a redox role in 392 dimer formation and stabilisation. 393 When analysed by size exclusion chromatography, recombinant VuCA1Pase 394 repeatedly showed two peaks with sizes consistent with a monomer (near 40kDa) 395 and a larger peak that may be a dimer (near 70kDa; Fig. 3D). While the peaks are 396 estimated to be smaller than the near 50 and 100kDa expected, sizes determined by 397 size exclusion are not exact and can be influenced by a variety of factors including 398 the shape of the protein (Fekete _et al._, 2014; Burgess, 2018). To assess the ability of 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 399 each domain to independently form multimers, the domains were expressed 400 separately as truncated proteins. The catalytic domain variant VuC H8A,1-206, showed 401 consistent peaks near their monomer size (Fig. 3E), while the regulatory domain 402 variant, VuC 207-445 showed peaks that were consistent with multimeric sizes, typically 403 approximating trimers and dimers (Fig. 3F). Given the regulatory domain uniquely 404 shows affinity for oligomer formation, this further supports a role in dimer formation 405 for VuCA1Pase _in vivo_ . 406 **Two key regulatory domain cysteines influence redox sensitive CA1Pase** 407 **activity and dimer formation** 408 The combined roles of dimer formation and redox regulation of the regulatory 409 domain were investigated through the generation of recombinantly expressed 410 cysteine mutants, whereby one or both key cysteines (C292, C297) were replaced 411 with alanine (Supplementary Fig. S5). CA1Pase activity of these mutants was then 412 compared to the recombinant wild type VuCA1Pase across a gradient of reducing 413 conditions. From 100% oxidised glutathione (GSSG) to 100% reduced glutathione 414 (GSH), unmodified VuCA1Pase showed a distinct increase in activity in reducing 415 conditions (Fig. 4A). This supports previous findings in tobacco and French bean 416 which showed a stimulatory effect of GSH, and inhibitory effect of GSSG (Holbrook 417 _et al._, 1991; Heo & Holbrook, 1999). The reducing conditions also appeared to shift 418 the oligomeric state of VuCA1Pase, favouring the dimer form in the presence of 419 5mM DTT (Fig. 4A) and GSH (Supplementary Fig. S6). All cysteine mutants showed 420 a lower activity in fully reducing conditions compared to the wild type (Fig. 4B-D), 421 with those containing the C297A mutation showing an inhibitory effect when fully 422 reduced (Fig. 4C-D). The C292A mutation alone showed an unexpected increase in 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 423 activity from oxidised to reducing conditions despite the presumed loss of disulphide 424 formation potential (Fig. 4B). 425 For all the cysteine mutants (Fig. 4B-D), no monomer peak was seen in non 426 reducing conditions. With VuC C297A, reducing conditions resulted in a near total loss 427 of activity along with the presence of a monomer sized peak. This monomer peak 428 could relate to the loss of activity. Results from _in vitro_ activity of the individual peaks 429 of wild type VuCA1Pase supports the dimer as the more active variant 430 (Supplementary Fig. S7). Why the C297A variant would so drastically lose activity in 431 highly reducing conditions is unclear. 432 **Variation in two key cysteines is correlated with low Rubisco dark inhibition in** 433 **Fabaceae** 434 To identify the potential impact of the CA1Pase redox regulation, natural 435 variants of CA1Pase were identified lacking one or both key cysteines in species 436 where dark inhibition data is available (Supplementary Table S2; Nehls-Ramos _et al._, 437 2026). While cysteine mutants were found across angiosperms, a more distinct trend 438 was seen in Fabaceae, the most well sampled family (Fig. 5; Nehls-Ramos _et al._, 439 2026). For the bean family, fifteen species were identified that had both data for dark 440 inhibition and a known CA1Pase sequence. Of these, five had serine or threonine in 441 place of one or both of C292 and C297 (Fig. 5). All five had low accumulation of dark 442 inhibitors despite other related species having much higher accumulation. This trend 443 suggests CA1Pases lacking redox regulation capacity from the CXXXXC motif may 444 retain activity regardless of light or stress related redox state, mirroring our _in vitro_ 445 cysteine mutant work. It is therefore possible that these CA1Pase variants may 446 continue to degrade inhibitors such as CA1P during night and/or dark conditions. 447 18 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 448 **Discussion** 449 CA1Pase plays an important role in dephosphorylating certain sugar 450 phosphate derivatives that inhibit the activity of Rubisco. We used a combination of 451 _in silico_ and _in vitro_ protein characterisation to elucidate the function and 452 mechanisms of CA1Pase regulation in the legume crop cowpea. Our results suggest 453 an ancient monomeric origin for this enzyme, later developing capacity for 454 dimerisation and conversion of the C domain into a pseudophosphatase regulatory 455 domain in Viridiplantae. Appearance of a key CXXXXC motif suggests redox 456 regulation arose in the Charophytes. We found evidence of CA1Pase forming redox 457 regulated homodimers, which may confer heightened activity compared to the 458 monomer form. This redox regulation was eliminated _in vitro_ through mutating one or 459 both cysteines, and our results suggest an association between natural cysteine 460 variants of CA1Pase and previously identified low dark inhibitor accumulation in 461 some plant species. 462 While previously the evolutionary origins of CA1Pase have been difficult to 463 parse, comparison with its bacterial and cyanobacterial relatives enables a clearer 464 picture. In histidine phosphate phosphoglycerate mutase (PGM) family enzymes, a 465 single phosphatase domain is sufficient, as with PhoE enzyme (Fig. 1A). In some 466 cases, these enzymes form dimers or higher order homooligomers as with PSP1, 467 which can improve activity and stability of the enzyme complex (Fig. 1A; Marianaya 468 _et al._, 2004). It is therefore not so surprising that this type of phosphatase has the 469 potential to form a fused dimer via a gene fusion duplication event that could retain 470 this stable form (Yanai _et al._, 2002; Muthahari _et al._, 2025). Such events are not 471 uncommon and can provide diversification of domain roles, while also improving the 472 stability of a potentially higher activity structure (Rullán-Lind _et al._, 2018). It is 19 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 473 possible that the binding of ligands to one of the catalytically active domains may 474 also help to regulate the activity of the opposite domain allosterically, providing a 475 possible sensor function to further regulate activity of this phosphatase-like enzyme. 476 For example, such allosteric regulation occurs with phosphofructokinase (Compton & 477 Patrick, 2025). This fusion event in CA1Pase likely occurred in cyanobacteria prior to 478 the endosymbiotic origin of chloroplasts or was transferred to an early chloroplast via 479 horizontal gene transfer. This is supported by the ubiquity of CA1Pase-like 480 sequences across cyanobacterial classes (Supplementary Dataset S1). 481 The redox-sensitive dimerisation function of the regulatory domain in land 482 plants may have first occurred in charophytes, with the development of the extended 483 alpha helix 11, and the nearby CXXXXC motif (C292, C297) that is absent in 484 chlorophytic and cyanobacterial CA1Pase-like sequences (Fig. 2A; Fig. 2C; 485 Supplementary Dataset S1). Despite this, the predicted structures for the 486 chlorophyte algae _Tetradesmus obliquus_ and _Trebouxia_ suggest dimerisation 487 occurring in this division, and that this potentially arose prior to gaining the redox 488 sensitive trait (Supplementary Dataset S1; Supplementary Dataset S2). Regulation 489 of day-night dephosphorylation of specific metabolites in early land plants and their 490 charophytic ancestors may have driven an adaptation of CA1Pase. 491 Given the loss of catalytic residues in the regulatory domain appears to have 492 occurred in both chlorophytes and charophytes, and that the regulatory domain 493 pocket faces towards the proposed dimer interface, the regulatory domain may have 494 first functioned as a pseudophosphatase sensor for certain sugar-phosphate 495 metabolites in the chloroplast, potentially linked to dimer interactions (Fig. 3A; 496 Reiterer _et al._, 2020). This function is supported by the evidence of various sugar 497 phosphate CBB metabolites, such as FBP, enhancing CA1Pase activity _in vitro_ in 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 498 previous studies, along with _in silico_ predictions suggesting certain chloroplast 499 metabolites may interact with the regulatory pocket (Table 1; Supplementary Fig. S4). 500 These predicted structures however do not give the complete picture and are limited 501 to the related resolved structures they are based on. The resolved structure of land 502 plant CA1Pase with FBP would help elucidate this potential interaction, and any 503 impacts it may have on the proposed dimer structure. 504 In agreement with previous _in vitro_ studies, cowpea CA1Pase also appears to 505 be redox sensitive, with reducing conditions enhancing activity (Table 1; Fig. 4A). 506 Here we show that the disruption of one or both key cysteines in the regulatory 507 domain, C292 and C297, either decreases or reverses the positive effect on 508 CA1Pase activity. This is likely tied to the disruption of the formation of the more 509 active dimer form (Fig. 3B-D). The cysteine redox state impact on dimer formation 510 may be due to this covalent disulphide bond providing a more rigid loop structure, 511 hindering more stable conformations for the dimer interface (Fig. 3A). Currently no 512 experimental CA1Pase structures are available; future structures of the reduced and 513 oxidised forms would help to better elucidate this mechanism. 514 Between the two cysteines, C297 appears the more important residue for 515 disrupting dimer formation. Mutating this residue to alanine resulted in a distinct loss 516 of activity in reducing conditions that was not seen with the C292A variant (Fig. 4B 517 D). This may be due to C297 being more solvent available, allowing for further 518 interactions by this residue when forming disulphide bonds with the dimer interface 519 that is clearly seen in the predicted structure. Reduced glutathione can form mixed 520 disulphides with cysteine by S-glutathionylation, which may explain the partial 521 increase of activity with the C292A variant that retained C297 (Heo & Holbrook, 522 1999; Musaogullari & Chai, 2020). Formation of mixed disulphides for C297 may 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 523 result in a conformation that further stabilizes dimer formation, given it does not 524 directly interact with the dimer interface to cause a disruption by disulphide formation. 525 The opposite may be true for C292 which may disrupt the dimer interface when 526 converted to a mixed disulphide. An explanation for the loss of activity in the dual 527 mutant C292A,C297A in reducing conditions is less clear, potentially suggesting the 528 slight deleterious effect of mixed disulphides in other parts of the CA1Pase structure 529 for solvent available cysteines. A more active CA1Pase may be possible by retaining 530 the hydrophobic C292 while exchanging C297 for a hydrophilic residue. This could 531 result in a conformation of this region that further stabilises the dimer interface. 532 Across plant species, some CA1Pase cysteine variants of C297 possess 533 hydrophilic, polar uncharged residues at this position, such as serine or threonine 534 (Fig. 5). These amino acids are encoded by codons with a single nucleotide different 535 from cysteine, requiring only a single nucleotide mutation to occur. These species 536 are known to have low Rubisco dark inhibition, potentially indicating increased 537 inhibitor degradation in dark conditions due to lacking a low-activity oxidised state. 538 This trend of sudden phylogenetic loss of dark inhibition had been previously 539 identified in our work and can now be consistently linked with these cysteine variants 540 within the Fabaceae (Nehls-Ramos _et al._, 2026). Similar evidence points to this link 541 in the limited data available for Brassicaceae. It should be noted that there is limited 542 current information on the expression and abundance of CA1Pase in these species, 543 which would further elucidate this lower dark inhibition across species. 544 Given crops regularly face stress conditions which in some species has been 545 shown to increase production of Rubisco inhibitors (Parry _et al._, 2008; Carmo-Silva 546 _et al_ ., 2010), the downregulation of CA1Pase activity in periods of decreased 547 photosynthesis and elevated stromal ROS means a decreased potential for Rubisco 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 548 inhibitor degradation (Slattery _et al._, 2018). This could be relevant during heat and 549 drought stress conditions, where both ROS and Rubisco inhibitor generation is 550 expected to be higher (Parry _et al._, 2008; Carmo-Silva _et al_ ., 2010; Slattery _et al_ ., 551 2019). Such inhibitors include those produced as misfire products from increased 552 oxygenation rates (Pearce, 2006; Parry _et al._, 2008; Orr _et al._, 2023). A CA1Pase 553 that is resistant to the impacts of oxidising dark and stress conditions could mean 554 continued inhibitor degradation regardless of environment. Increased CA1Pase 555 activity may also help with the recycling of phosphate from other inhibitory 556 compounds such as PDBP, supporting sustained chloroplast phosphate 557 homeostasis. Regular degradation of inhibitors may be essential for sustaining 558 available phosphate for the light reactions of photosynthesis (Leister _et al._, 2023). 559 The direct benefits of the proposed redox insensitive CA1Pase cysteine variants in 560 some Fabaceae remains to be explored and may further elucidate the utility and role 561 of this trait (Fig. 5). 562 CA1Pase contributes to regulating the abundance of inhibitors and the activity 563 of Rubisco. Our investigation of the structure and function of cowpea CA1Pase, 564 identified evolutionary trends and highlighted a key CXXXXC motif that impacts 565 redox regulation. The physiological impact of altering the abundance of such 566 inhibitors through manipulating the abundance and activity of CA1Pase remain to be 567 fully elucidated _in planta_ . Future studies are warranted to interrogate and verify the 568 specific mechanism for sugar phosphate allosteric binding in the regulatory domain 569 and redox-dependant dimerisation via experimentally resolved structures. Metabolic, 570 proteomic, and physiological studies, combined with modified expression of 571 CA1Pase in species presenting low and high Rubisco dark inhibition may further 572 elucidate the combined role of CA1P and CA1Pase as regulators of carbon 23 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 573 assimilation by Rubisco as well as any ancillary role in the metabolic regulation of 574 the chloroplast. Through identifying new avenues for improved Rubisco inhibitor 575 degradation, we can gain further insight into optimisation for improved photosynthetic 576 activity towards maximising crop resilience and output. 577 578 **Supplementary data** 579 **Supplementary Fig. S1.** AlphaFold 3 predicted structures with confidence indicated 580 **Supplementary Fig. S2.** SDS-PAGE gels of purified VuCA1Pase and variants 581 **Supplementary Fig. S3.** Phosphate detection diagram 582 **Supplementary Fig. S4.** Alphafold 3 predicted structures for regulatory domain 583 ligand binding 584 **Supplementary Fig. S5.** Labelled AlphaFold 3 predicted structures for VuCA1Pase 585 cysteine mutants 586 **Supplementary Fig. S6.** Effect of glutathione on SEC elution peaks for VuCA1Pase 587 and VuC C292A,C297A 588 **Supplementary Fig. S7.** VuCA1Pase eluted fraction activity and SDS gel bands 589 **Supplementary Table S1.** Full list of species with known CA1Pase sequences and 590 dark inhibition levels 591 **Supplementary Table S2.** Recombinant protein purity and date of purification 592 **Supplementary Table S3.** Size exclusion column protein elution expected protein 593 sizes with treatments 594 **Supplementary Table S4.** Pairwise identity of aligned peptide sequences 595 **Supplementary Table S5.** Root mean square deviation of aligned predicted and 596 resolved structures 597 **Supplementary Dataset S1.** Geneious alignments and sequences 24 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 598 **Supplementary Dataset S2.** AlphaFold 3 predicted structures 599 **Supplementary Dataset S3.** Crystal structures for broad specifcity phosphatases 600 **Supplementary Dataset S4.** D-I-Tasser predicted protein structures 601 **Supplementary Dataset S5.** _in vitro_ data for biochemical assays 602 **Supplementary Dataset S6.** SEC runs for CA1Pases and variants 603 604 **Acknowledgements** 605 The authors thank colleagues in the Photosynthesis team at Lancaster for useful 606 discussions and data interpretation, especially Dr. Armida Gjindali, Dr. Joana Amaral, 607 Dr. Clayton Dilks, Dr. Felix Townsend, and Dr. Dawn Worrall. We also thank Dr. 608 Mitchell Altschuler (University of Illinois at Urbana-Champaign) for useful discussions. 609 610 **Author contributions** 611 CN, ECS, and DJO: conceptualisation; CN: data curation; CN: formal analysis; ECS 612 and DJO: funding acquisition; CN: investigation; CN, ECS and DJO: methodology; 613 ECS and DJO: resources; ECS and DJO: supervision; CN: visualisation; CN: writing 614 - original draft; CN, ECS, and DJO: writing - review & editing. 615 616 **Conflict of interest** 617 None declared. 618 619 **Funding** 620 This work was supported by the project Realizing Increased Photosynthetic 621 Efficiency (RIPE), that is funded by Gates Agricultural Innovations grant investment 622 57248, awarded to Lancaster University by the University of Illinois, USA. CNR is 25 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 623 also supported by a Lancaster Environment Centre (LEC) Postgraduate Research 624 Studentship from Lancaster University. 625 626 **Data availability** 627 The data collated and generated for this paper is included in the excel file found in 628 (https://github.com/cwnehls/NehlsRamos_etal_JxB_2027). Within the github file the 629 code used for analyses and figure generation are included. 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **References** **Abramson J, Adler J, Dunger J,** _**et al.**_ 2024. Accurate structure prediction of biomolecular interactions with AlphaFold 3. _Nature_ **630**, 493–500. **Amaral J, Lobo AKM, Carmo-Silva E** . 2024. Regulation of Rubisco activity in crops. _New Phytologist_ **241**, 35–51. **Andralojc PJ, Keys AJ, Kossmann J, Parry MAJ** . 2002. Elucidating the biosynthesis of 2-carboxyarabinitol 1-phosphate through reduced expression of chloroplastic fructose 1,6-bisphosphate phosphatase and radiotracer studies with 14 CO 2 . _Proceedings of the National Academy of Sciences, USA_ **99**, 4742–4747. **Andralojc PJ, Madgwick PJ, Tao Y,** _**et al.**_ 2012. 2-Carboxy-D-arabinitol 1 phosphate (CA1P) phosphatase: evidence for a wider role in plant Rubisco regulation. _Biochemical Journal_ **442**, 733–742. **Bayer RG, Stael S, Csaszar E, Teige M** . 2011. Mining the soluble chloroplast proteome by affinity chromatography. _PROTEOMICS_ **11**, 1287–1299. **Bertoni D, Tsenkov M, Magana P,** _**et al.**_ 2026. AlphaFold Protein Structure Database 2025: a redesigned interface and updated structural coverage. _Nucleic_ _Acids Research_ **54**, D358–D362. **Borchers HW** . 2023. _pracma: Practical Numerical Math Functions_ . **Bracher A, Sharma A, Starling-Windhof A, Hartl FU, Hayer-Hartl M** . 2015. Degradation of potent Rubisco inhibitor by selective sugar phosphatase. _Nature_ Plants **1**, 1–7. **Burgess RR** . 2018. A brief practical review of size exclusion chromatography: Rules of thumb, limitations, and troubleshooting. _Protein Expression and Purification_ **150**, 81–85. **Carmo-Silva AE, Keys AJ, Andralojc PJ, Powers SJ, Arrabaça MC, Parry MAJ.** 2010. Rubisco activities, properties, and regulation in three different C4 grasses under drought. Journal of Experimental Botany 61, 2355–2366. 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Cejudo FJ, González M-C, Pérez-Ruiz JM** . 2021. Redox regulation of chloroplast metabolism. _Plant Physiology_ **186**, 9–21. **Charlet T, Moore B d., Seemann JR** . 1997. Carboxyarabinitol 1-Phosphate Phosphatase from Leaves of _Phaseolus vulgaris_ and Other Species. _Plant and Cell_ _Physiology_ **38**, 511–517. **Compton JA, Patrick WM** . 2025. The more we learn, the more diverse it gets: structures, functions and evolution in the Phosphofructokinase Superfamily. _Biochemical Journal_ **482**, 467–483. **Croce R, Carmo-Silva E, Cho YB,** _**et al.**_ 2024. Perspectives on improving photosynthesis to increase crop yield. _The Plant Cell_ **36**, 3944–3973. **de Mendiburu F.** 2023. _agricolae: Statistical Procedures for Agricultural Research_ . R package version 1.3-7. **Fekete S, Beck A, Veuthey J-L, Guillarme D** . 2014. Theory and practice of size exclusion chromatography for the analysis of protein aggregates. _Journal of_ _Pharmaceutical and Biomedical Analysis_ **101**, 161–173. **Foyer CH, Hanke G** . 2022. ROS production and signalling in chloroplasts: cornerstones and evolving concepts. _The Plant Journal_ **111**, 642–661. **Garnier, Simon, Ross,** _**et al.**_ 2024. _viridis(Lite) - Colorblind-Friendly Color Maps for_ _R_ . **Gasteiger E, Hoogland C, Gattiker A, Duvaud S, Wilkins MR, Appel RD, Bairoch** **A** . 2005. Protein Identification and Analysis Tools on the ExPASy Server. In: Walker JM, ed. The Proteomics Protocols Handbook. Totowa, NJ: _Humana Press_, 571–607. **Gjindali A, Page R, Ashton CJ, Robertson I, Page MT, Bloemers D, Gould PD,** **Worrall D, Orr DJ, Carmo-Silva E** . 2025. Two cowpea Rubisco activase isoforms for crop thermotolerance. _New Phytologist_ **247**, 1199–1217. **Goodstein DM, Shu S, Howson R,** _**et al.**_ 2012. Phytozome: a comparative platform for green plant genomics. _Nucleic Acids Research_ **40**, D1178–D1186. 28 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Graves S, Piepho H-P, Dorai-Raj LS with help from S** . 2024. _multcompView:_ _Visualizations of Paired Comparisons_ . **Gütle DD, Roret T, Müller SJ,** _**et al.**_ 2016. Chloroplast FBPase and SBPase are thioredoxin-linked enzymes with similar architecture but different evolutionary histories. _Proceedings of the National Academy of Sciences of the United States of_ _America_ **113**, 6779–6784. **Gutteridge S, Julien B** . 1989. A phosphatase from chloroplast stroma of _Nicotiana_ _tabacum_ hydrolyses 2 ′ -carboxyarabinitol 1 -phosphate, the natural inhibitor of Rubisco to 2 ′ -carboxyarabinitol. _FEBS Letters_ **254**, 225–230. **Heo J, Holbrook GP** . 1999. Regulation of 2-carboxy-d-arabinitol 1-phosphate phosphatase: activation by glutathione and interaction with thiol reagents. _Biochemical Journal_ **338**, 409–416. **Holbrook GP, Bowes G, Salvucci ME** . 1989. Degradation of 2-Carboxyarabinitol 1 Phosphate by a Specific Chloroplast Phosphatase. _Plant Physiology_ **90**, 673–678. **Holbrook GP, Galasinski SC, Salvucci ME** . 1991. Regulation of 2 Carboxyarabinitol 1-Phosphatase 1. _Plant Physiology_ **97**, 894–899. **Hothorn T, Bretz F, Westfall P** . 2008. Simultaneous Inference in General Parametric Models. _Biometrical Journal_ **50**, 346–363. **Kingston-Smith AH, Major I, Parry MAJ, Keys AJ** . 1992. Purification and properties of a phosphatase in French bean ( _Phaseolus vulgaris_ L.) leaves that hydrolyses 2 ′ -carboxy-d-arabinitol 1-phosphate. _Biochemical Journal_ **287**, 821–825. **Leister D, Sharma A, Kerber N,** _**et al.**_ 2023. An ancient metabolite damage-repair system sustains photosynthesis in plants. _Nature Communications_ **14**, 3023. **Lobo AKM, Orr DJ, Gutierrez MO, Andralojc PJ, Sparks C, Parry MAJ, Carmo-** **Silva E** . 2019. Overexpression of ca1pase Decreases Rubisco Abundance and Grain Yield in Wheat. _Plant Physiology_ **181**, 471–479. **Lu S, Wang J, Chitsaz F,** _**et al.**_ 2020. CDD/SPARCLE: the conserved domain database in 2020. _Nucleic Acids Research_ **48**, D265–D268. 29 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Mangiafico SS** . 2025. _rcompanion: Functions to Support Extension Education_ _Program Evaluation_ . New Brunswick, New Jersey: _Rutgers Cooperative Extension_ . **Marianayagam NJ, Sunde M, Matthews JM** . 2004. The power of two: protein dimerization in biology. _Trends in Biochemical Sciences_ **29**, 618–625. **Martínez Gache SA, Recoulat Angelini AA, Sabeckis ML, González Flecha FL** . 2020. Improving the stability of the malachite green method for the determination of phosphate using Pluronic F68. _Analytical Biochemistry_ **597**, 113681. **Moore B d., Kiirats O, Charlet T, Seemann JR** . 1995. Levels of Carboxyarabinitol 1-Phosphate and Related Photosynthetic Components during Leaf Development in _Phaseolus vulgaris_ . _Plant and Cell Physiology_ **36**, 1097–1103. **Müller-Schüssele SJ** . 2024. Chloroplast thiol redox dynamics through the lens of genetically encoded biosensors. (G Noctor, Ed.). _Journal of Experimental Botany_ **75**, 5312–5324. **Müller-Schüssele SJ, Wang R, Gütle DD,** _**et al.**_ 2020. Chloroplasts require glutathione reductase to balance reactive oxygen species and maintain efficient photosynthesis. _The Plant Journal_ **103**, 1140–1154. **Musaogullari A, Chai Y-C** . 2020. Redox Regulation by Protein S-Glutathionylation: From Molecular Mechanisms to Implications in Health and Disease. _International_ _Journal of Molecular Sciences_ **21**, 8113. **Muthahari YA, Magnus L, Laurino P** . 2025. From duplication to fusion: Expanding Dayhoff’s model of protein evolution. _Protein Science_ **34**, e70054. **Nehls-Ramos C, Carmo-Silva E, Orr DJ** . 2026 Rubisco Dark Inhibition in Angiosperms Shows a Complex Distribution Pattern. _Journal of Experimental_ _Botany_ **77**, 2914-2924. **Ogle DH, Doll JC, Wheeler AP, Dinno A** . 2025. _FSA: Simple Fisheries Stock_ _Assessment Methods_ . 30 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Orr DJ, Robijns AKJ, Baker CR, Niyogi KK, Carmo-Silva E** . 2023. Dynamics of Rubisco regulation by sugar phosphate derivatives and their phosphatases. _Journal_ _of Experimental Botany_ **74**, 581–590. **Parry MAJ, Andralojc PJ, Scales JC, Salvucci ME, Carmo-Silva AE, Alonso H,** **Whitney SM** . 2013. Rubisco activity and regulation as targets for crop improvement. _Journal of Experimental Botany_ **64**, 717–730. **Parry MAJ, Keys AJ, Madgwick PJ, Carmo-Silva AE, Andralojc PJ** . 2008. Rubisco regulation: a role for inhibitors. _Journal of Experimental Botany_ **59**, 1569– 1580. **Pasch V, Leister D, Rühle T** . 2025. Synergistic role of Rubisco inhibitor release and degradation in photosynthesis. _New Phytologist_ **245**, 1496–1511. **Pearce FG** . 2006. Catalytic by-product formation and ligand binding by ribulose bisphosphate carboxylases from different phylogenies. The Biochemical journal 399 3, 525–34. **Perdomo JA, Sales CRG, Carmo-Silva E** . 2018. Quantification of Photosynthetic Enzymes in Leaf Extracts by Immunoblotting. In: Covshoff S, ed. Photosynthesis: Methods and Protocols. New York, NY: _Springer_, 215–227. **Podesta FE, Iglesias AA, Andreo CS** . 1990. Oligomeric enzymes in the C4 pathway of photosynthesis. _Photosynthesis Research_ **26**, 161–170. **Portis AR, Li C, Wang D, Salvucci ME** . 2008. Regulation of Rubisco activase and its interaction with Rubisco. _Journal of Experimental Botany_ **59**, 1597–1604. **Posit team** . 2025. _RStudio: Integrated Development Environment for R_ . Boston, MA: Posit Software, PBC. **R Core Team** . 2024. _R: A Language and Environment for Statistical Computing_ . Vienna, Austria: R Foundation for Statistical Computing. **Reiterer V, Pawłowski K, Desrochers G, Pause A, Sharpe HJ, Farhan H** . 2020. The dead phosphatases society: a review of the emerging roles of pseudophosphatases. _The FEBS Journal_ **287**, 4198–4220. 31 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Rigden DJ** . 2007. The histidine phosphatase superfamily: structure and function. _Biochemical Journal_ **409**, 333–348. **Rullán-Lind C, Pietri RB, Vázquez-Cintrón M, Baerga-Ortiz A** . 2018. Fused dimerization increases expression, solubility, and activity of bacterial dehydratase enzymes. _Protein Science_ **27**, 969–975. **Salvucci ME, Holbrook GP** . 1989. Purification and Properties of 2-Carboxy-d Arabinitol 1-Phosphatase 1. _Plant Physiology_ **90**, 679–685. **Salvucci ME, Holbrook GP, Anderson JC, Bowes G** . 1988. NADPH-dependent metabolism of the ribulose bisphosphate carboxylase/oxygenase inhibitor 2 carboxyarabinitol 1-phosphate by a chloroplast protein. _FEBS Letters_ **231**, 197–201. **Sayers EW, Beck J, Bolton EE,** _**et al.**_ 2025. Database resources of the National Center for Biotechnology Information in 2025. _Nucleic Acids Research_ **53**, D20–D29. **Servaites JC, Parry MAJ, Gutteridge S, Keys AJ** . 1986. Species Variation in the Predawn Inhibition of Ribulose-1,5-Bisphosphate Carboxylase/Oxygenase. _Plant_ _Physiology_ **82**, 1161–1163. **Slattery RA, Walker BJ, Weber APM, Ort DR** . 2018. The Impacts of Fluctuating Light on Crop Performance. _Plant Physiology_ **176**, 990–1003. **Slattery RA, Ort DR. 2019.** Carbon assimilation in crops at high temperatures. Plant, Cell & Environment 42, 2750–2758. **The UniProt Consortium. 2025.** UniProt: the Universal Protein Knowledgebase in 2025. Nucleic Acids Research 53, D609–D617. **Trinh MDL, Masuda S. 2022.** Chloroplast pH Homeostasis for the Regulation of Photosynthesis. Frontiers in Plant S _cience_ **13** . **Waheeda K, Kitchel H, Wang Q, Chiu P-L** . 2023. Molecular mechanism of Rubisco activase: Dynamic assembly and Rubisco remodeling. Frontiers in Molecular Biosciences Volume 10-2023. 32 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Wickham H** . 2016. _ggplot2: Elegant Graphics for Data Analysis_ . Springer-Verlag New York. **Wickham H, François R, Henry L, Müller K, Vaughan D** . 2023a. _dplyr: A Grammar_ _of Data Manipulation_ . **Wickham H** . 2023b. _forcats: Tools for Working with Categorical Variables (Factors)_ . **Wickham H, Hester J, Bryan J** . 2024. _readr: Read Rectangular Text Data_ . **Yamori W** . 2016. Photosynthetic response to fluctuating environments and photoprotective strategies under abiotic stress. _Journal of Plant Research_ **129**, 379– 395. **Yanai I, Wolf YI, Koonin EV** . 2002. Evolution of gene fusions: horizontal transfer versus independent events. _Genome Biology_ **3**, research0024.1. **Yoshida K, Hisabori T** . 2023. Current Insights into the Redox Regulation Network in Plant Chloroplasts. _Plant and Cell Physiology_ **64**, 704–715. **Yu A, Xie Y, Pan X, Zhang H, Cao P, Su X, Chang W, Li M** . 2020. Photosynthetic Phosphoribulokinase Structures: Enzymatic Mechanisms and the Redox Regulation of the Calvin-Benson-Bassham Cycle. The Plant Cell 32, 1556–1573. **Zheng W, Wuyun Q, Li Y, Liu Q, Zhou X, Peng C, Zhu Y, Freddolino L, Zhang Y** . 2026. Deep-learning-based single-domain and multidomain protein structure prediction with D-I-TASSER. _Nature Biotechnology_ **44**, 641–653. 33 **Tables** **Table 1.** Available data characterising CA1Pase: Michaelis Menten constant (Km) for CA1P degradation; molecules that increase, decrease or do not significantly affect CA1Pase activity; substrates in addition to CA1P; and pH optima. Bold indicates positive effectors identified in multiple studies or species. NA, not available. |Species|CA1P<br>Km<br>(mM)|CA1P<br>Activity<br>(µmolmin-1<br>mg-1)|PositiveEffectors|Negative<br>Effectors|NoEffect|OtherSubstrates|pH<br>Optima|Reference| |---|---|---|---|---|---|---|---|---| |_Nicotiana_<br>_rustica_|NA|0.0005|**NADPH**, **DTT**|NA|NADP+, NADH, ATP|NA|NA|Salvucci_et al._, <br>1988; Holbrook_et_<br>_al._, 1989| |_Spinacia_<br>_oleracea_|NA|NA|**NADPH**|NA|ATP|NA|NA|Salvucci_et al._, <br>1988| |_Nicotiana_<br>_tabacum_|0.15;<br>0.033;<br>0.0126|0.248|**DTT**, KCl, K2SO4, Na2SO4, (NH4)SO4, <br>NH4Cl,**NADPH**, **RuBP**, **FBP**, **3PGA**, **ADP;**<br>Fructose-2,6-BP, Glucose-1,6-BP, 2**-**<br>**phosphoglycolate**, Cysteine,**GSH**, 2-<br>mercaptoethanol|MgSO4, MnCl2, <br>ZnCl2, ZnSO4, <br>CuSO4, KH2PO4, <br>Ribulose-5-P, ATP;<br>Pi,**GSSG**, Sodium<br>tetrathionate|6-Phosphogluconate,<br>Fructose-6-P, Glucose-<br>6-P, Ribulose-5-P;<br>NADPH, Mg2+, ATP;<br>Glyceraldehyde-3-P,<br>Glucose-1-P|CA 1,5-bisphosphate,<br>CRBP, 2-carboxy-D-<br>ribitol-1-phosphate|Near<br>7.5;<br>near 8|Gutteridge &<br>Julien, 1989;<br>Salvucci_et al._, <br>1989; Holbrook_et_<br>_al._, 1991| |_Phaseolus_<br>_vulgaris_|0.433;<br>0.237|6.83; 6.1; 1.4|**NADPH, FBP**, (NH4)2SO4, KCl, NaCl,<br>MgCl2 (10mM);** DTT**, KCl, Ribulose-5-P,<br>**RuBP**. Pyrophosphate, Galacturonic acid<br>1-P,**3-phosphoglycerate,** **2-**<br>**phosphoglycerate**, 2,3-<br>diphosphoglycerate,**2-phosphoglycolate**, <br>carboxyethyl-phosphonic acid, ATP,**ADP;**<br>**Glutathione**|CaCl2 (<0.8mM);<br>Orthophosphate;<br>**GSSG**, FBP +<br>GSSG, Cysteine,<br>Thioredoxin|MgCl2 (<3.3mM);<br>Fructose-1-P, Fructose-<br>6-P, AMP, NADP + Pi, <br>KCl (oxidising<br>conditions)|carboxy pentitol,<br>carboxyhexitol<br>bisphosphate, CRBP,<br>2-carboxymannitol 1,6-<br>bisphosphate, CA 1,5-<br>bisphosphate|Near 7|Kingston-Smith_et_<br>_al._, 1992; Charlet<br>_et al._, 1997; Heo<br>& Holbrook, 1999| |_Triticum_<br>_aestivum_|0.01|4.2±0.1|**DTT**, ATP,**ADP**, Pyrophosphate,**FBP**, <br>**RuBP**, 3-phosphoglycerate.**2-**<br>**phosphoglycolate**|**GSSG**, 6-<br>phosphoglycerate|Pi, Glucose-6-P,<br>Fructose-6-P, Glucose-<br>3-P|CA 1,5-bisphosphate,<br>CRBP, 2,3-<br>diphosphoglycerate|near 7|Andralojc_et al._, <br>2012| 34 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Figure legends** **Fig. 1.** Structural predictions support two distinct domains within CA1Pase. (A) Left to-right: Predicted AlphaFold 3 structures for cowpea CA1Pase ( _Vigna unguiculata_ ; VuCA1Pase) and a cyanobacterial CA1Pase-like protein from _Myxacorys_ _almenidensis_, crystal structures of a bacterial broad specificity phosphatase (PhoE, 1H2F) from _Geobacillus stearothermophilus_, and a bacterial phosphoserine phosphatase (PSP1, 4IJ5) from _Hydrogenobacter thermophilus_ . The predicted transit peptide for VuCA1Pase is removed. Catalytically active residues are highlighted in orange. All structures are shown as monomers, excluding PSP1 which has been resolved as a homodimer. (B) Sequence alignment for the protein sequences of the four enzymes shown, excluding predicted chloroplast transit peptide sequences (Vigun11g130800.1.10, NDJ19632.1, AAK07665.1, BAI68570.1). Sequence alignment and numbering of residues start at the first Lysine (K) of VuCA1Pase. White, grey, and black indicate 100% homology, 50-75%, and not conserved, respectively. The green bar indicates the catalytic domain of VuCA1Pase, yellow bar highlights the regulatory domain. Orange annotations below residue positions denote catalytic residues for phosphatase activity conserved across all four sequences. In the regulatory domain, orange annotations indicate residues that mirror the conserved catalytic residues in the CA1Pase-like sequence. **Fig. 2.** A putative pseudophosphatase regulatory domain mirrors the catalytic domain. (A) Sequence alignment of VuCA1Pase catalytic and regulatory domains. White, grey, and black indicate 100%, 50-75%, and not conserved, respectively. The green bar indicates the catalytic domain of VuCA1Pase, yellow bar highlights the regulatory domain. Black bars above a black “C” indicate the position of the two 35 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) cysteine residues of the CXXXXC motif. Grey bars indicate secondary structures: arrows indicate beta sheets; cylinders indicate alpha helices. “*” indicate secondary structures conserved in both domains. (B) Comparison of the AlphaFold 3 predicted structures of the VuCA1Pase domains, showing similarity of secondary structures and the position of catalytic residues. Orange indicates catalytic residues and the mirrored residues found in the regulatory domain. (C) Tree alignment of 10 CA1Pase sequences and 1 CA1Pase-like sequence, representing ancestral branches from a wide range of organisms from cyanobacteria to several plant families. Any predicted chloroplast transit peptides were removed, starting alignment and numbering at the lysine of the conserved “KRVV” motif of VuCA1Pase. Green and yellow bars indicate the location of the catalytic and regulatory domains. The positions of catalytic (orange) and mirror (dark yellow) residues, and the putative regulatory cysteine residues (black) are shown in the boxes containing sequence alignment sections at the bottom. **Fig. 3.** The regulatory domain of CA1Pase is required for dimer formation. (A) Side and top view of the AlphaFold 3 predicted dimer structure of VuCA1Pase. Coloured regions represent residues conserved in >95% of 252 CA1Pase sequences, ranging from monocots to dicots. Green and yellow represent conserved catalytic and regulatory domain residues, respectively. The presence of the predicted disulfide bond formation is shown within the black box and indicated by black arrows. (B) A zoom into the redox motif region shown in the dotted line box in Figure 3A. Cysteines 292 and 297 are shown along with the presence of the predicted disulfide bond. This motif neighbours conserved residues that are part of the predicted dimer interface. (C-F) Size exclusion chromatography (SEC) traces for (C) protein standards ( γ 36 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) globulins, Ovalbumin, Ribonuclease A type I-A), (D) VuCA1Pase, (E) VuCA1Pase lacking the regulatory domain (VuC H8A,1-206 ), and (F) VuCA1Pase lacking the catalytic domain (VuC 207-445 ). Estimated size in kilodaltons (kDa) is indicated below each peak, shown by a dotted grey line. **Fig. 4.** Redox regulation of CA1Pase is linked to dimer formation. (A-D) Graphs at the top of each panel display CA1Pase specific activity in presence of increasing concentrations of reduced glutathione (GSH, 0-100%). Graphs below show the traces of size exclusion chromatography in the absence (left) or presence of 5 mM DTT reducing agent (right), with predicted peak sizes. (A) VuCA1Pase, (B) single cysteine mutant VuC C292A, (C) single cysteine mutant VuC C297A, (D) double cysteine mutant VuC C292A,C297A . Technical replicates for each GSH treatment (n) = 3-4. Identified peaks are denoted by a grey dotted line underneath. **Fig. 5.** C297 variants in Fabaceae CA1Pases correlate with low Rubisco dark inhibition. Alignment of mature CA1Pase from 15 Fabaceae species, ordered by tree alignment (left), highlighting conservation of the CXXXXC motif. Species represent those from Fabaceae with both sequence and dark inhibition data available (see Nehls-Ramos _et al._, 2026). Positions of the regulatory domain CXXXXC motif (C292 & C297) identified in _Vigna unguiculata_ are annotated. Residues are numbered as described in Materials and Methods. Black highlight indicates residue differences from consensus. Coloured boxes at right indicate known prevalence of dark inhibition, as classified by Nehls-Ramos _et al._ (2026). 37 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Figures and Legends** **Fig. 1.** Structural predictions support two distinct domains within CA1Pase. (A) Left to-right: Predicted AlphaFold 3 structures for cowpea CA1Pase (Vigna unguiculata; VuCA1Pase) and a cyanobacterial CA1Pase-like protein from Myxacorys almenidensis, crystal structures of a bacterial broad specificity phosphatase (PhoE, 38 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1H2F) from Geobacillus stearothermophilus, and a bacterial phosphoserine phosphatase (PSP1, 4IJ5) from Hydrogenobacter thermophilus. The predicted transit peptide for VuCA1Pase is removed. Catalytically active residues are highlighted in orange. All structures are shown as monomers, excluding PSP1 which has been resolved as a homodimer. (B) Sequence alignment for the protein sequences of the four enzymes shown, excluding predicted chloroplast transit peptide sequences (Vigun11g130800.1.10, NDJ19632.1, AAK07665.1, BAI68570.1). Sequence alignment and numbering of residues start at the first Lysine (K) of VuCA1Pase. White, grey, and black indicate 100% homology, 50-75%, and not conserved, respectively. The green bar indicates the catalytic domain of VuCA1Pase, yellow bar highlights the regulatory domain. Orange annotations below residue positions denote catalytic residues for phosphatase activity conserved across all four sequences. In the regulatory domain, orange annotations indicate residues that mirror the conserved catalytic residues in the CA1Pase-like sequence. 39 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 2.** A putative pseudophosphatase regulatory domain mirrors the catalytic domain. (A) Sequence alignment of VuCA1Pase catalytic and regulatory domains. White, grey, and black indicate 100%, 50-75%, and not conserved, respectively. The green bar indicates the catalytic domain of VuCA1Pase, yellow bar highlights the regulatory domain. Black bars above a black “C” indicate the position of the two cysteine residues of the CXXXXC motif. Grey bars indicate secondary structures: 40 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) arrows indicate beta sheets; cylinders indicate alpha helices. “*” indicate secondary structures conserved in both domains. (B) Comparison of the AlphaFold 3 predicted structures of the VuCA1Pase domains, showing similarity of secondary structures and the position of catalytic residues. Orange indicates catalytic residues and the mirrored residues found in the regulatory domain. (C) Tree alignment of 10 CA1Pase sequences and 1 CA1Pase-like sequence, representing ancestral branches from a wide range of organisms from cyanobacteria to several plant families. Any predicted chloroplast transit peptides were removed, starting alignment and numbering at the lysine of the conserved “KRVV” motif of VuCA1Pase. Green and yellow bars indicate the location of the catalytic and regulatory domains. The positions of catalytic (orange) and mirror (dark yellow) residues, and the putative regulatory cysteine residues (black) are shown in the boxes containing sequence alignment sections at the bottom. 41 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 3.** The regulatory domain of CA1Pase is required for dimer formation. (A) Side and top view of the AlphaFold 3 predicted dimer structure of VuCA1Pase. Coloured regions represent residues conserved in >95% of 252 CA1Pase sequences, ranging from monocots to dicots. Green and yellow represent conserved catalytic and 42 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) regulatory domain residues, respectively. The presence of the predicted disulfide bond formation is shown within the black box and indicated by black arrows. (B) A zoom into the redox motif region shown in the dotted line box in Figure 3A. Cysteines 292 and 297 are shown along with the presence of the predicted disulfide bond. This motif neighbours conserved residues that are part of the predicted dimer interface. (C-F) Size exclusion chromatography (SEC) traces for (C) protein standards ( γ globulins, Ovalbumin, Ribonuclease A type I-A), (D) VuCA1Pase, (E) VuCA1Pase lacking the regulatory domain (VuCH8A,1-206), and (F) VuCA1Pase lacking the catalytic domain (VuC207-445). Estimated size in kilodaltons (kDa) is indicated below each peak, shown by a dotted grey line. 43 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 4.** Redox regulation of CA1Pase is linked to dimer formation. (A-D) Graphs at the top of each panel display CA1Pase specific activity in presence of increasing concentrations of reduced glutathione (GSH, 0-100%). Graphs below show the traces of size exclusion chromatography in the absence (left) or presence of 5 mM DTT reducing agent (right), with predicted peak sizes. (A) VuCA1Pase, (B) single cysteine mutant VuC C292A, (C) single cysteine mutant VuC C297A, (D) double cysteine mutant VuC C292A,C297A . Technical replicates for each GSH treatment (n) = 3-4. Identified peaks are denoted by a grey dotted line underneath. 44 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764501; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764501) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 5.** C297 variants in Fabaceae CA1Pases correlate with low Rubisco dark inhibition. Alignment of mature CA1Pase from 15 Fabaceae species, ordered by tree alignment (left), highlighting conservation of the CXXXXC motif. Species represent those from Fabaceae with both sequence and dark inhibition data available (see Nehls-Ramos _et al._, 2026). Positions of the regulatory domain CXXXXC motif (C292 & C297) identified in _Vigna unguiculata_ are annotated. Residues are numbered as described in Materials and Methods. Black highlight indicates residue differences from consensus. Coloured boxes at right indicate known prevalence of dark inhibition, as classified by Nehls-Ramos _et al._ (2026). 45 bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license](http://creativecommons.org/licenses/by/4.0/) Bacterial . Bacterial Cowpea CA1Pase (monomer) Cyanobacterial available under a (monomer) Cowpea Cyanobacterial [available under a CC-BY 4.0 International license](http://creativecommons.org/licenses/by/4.0/) Bacterial . Bacterial ## **A** Bacterial CA1Pase-like PhoE (monomer) PSP1 (homodimer) Catalytic Domain Regulatory Domain ## **B** bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 1.** Structural predictions support two distinct domains within CA1Pase. (A) Left-toright: Predicted AlphaFold 3 structures for cowpea CA1Pase ( _Vigna unguiculata_ ; VuCA1Pase) and a cyanobacterial CA1Pase-like protein from _Myxacorys almenidensis_, crystal structures of a bacterial broad specificity phosphatase (PhoE, 1H2F) from _Geobacillus stearothermophilus_, and a bacterial phosphoserine phosphatase (PSP1, 4IJ5) from _Hydrogenobacter thermophilus_ . The predicted transit peptide for VuCA1Pase is removed. Catalytically active residues are highlighted in orange. All structures are shown as monomers, excluding PSP1 which has been resolved as a homodimer. (B) Sequence alignment for the protein sequences of the four enzymes shown, excluding predicted chloroplast transit peptide sequences (Vigun11g130800.1.10, NDJ19632.1, AAK07665.1, BAI68570.1). Sequence alignment and numbering of residues start at the first Lysine (K) of VuCA1Pase. White, grey, and black indicate 100% homology, 50-75%, and not conserved, respectively. The green bar indicates the catalytic domain of VuCA1Pase, yellow bar highlights the regulatory domain. Orange annotations below residue positions denote catalytic residues for phosphatase activity conserved across all four sequences. In the regulatory domain, orange annotations indicate residues that mirror the conserved catalytic residues in the CA1Pase-like sequence. ## **A** **B** bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m # * * * * available under a CC-BY 4.0 International license * * . * * * * # *** * * * * * * * * *** Catalytic Domain 180° |Col1|Col2|Col3|Col4| |---|---|---|---| |Cyanobac~~te~~ria)<br>~~G~~reen Alga~~e~~)<br>~~G~~reen Alga~~e~~)<br>~~C~~harophyt~~e)~~<br>~~M~~oss)<br>~~F~~ern)<br>~~Y~~ew)<br>~~W~~heat)<br>~~To~~bacco)<br>~~Pea~~)<br>C~~o~~wpea)|||| bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 2.** A putative pseudophosphatase regulatory domain mirrors the catalytic domain. (A) Sequence alignment of VuCA1Pase catalytic and regulatory domains. White, grey, and black indicate 100%, 50-75%, and not conserved, respectively. The green bar indicates the catalytic domain of VuCA1Pase, yellow bar highlights the regulatory domain. Black bars above a black “C” indicate the position of the two cysteine residues of the CXXXXC motif. Grey bars indicate secondary structures: arrows indicate beta sheets; cylinders indicate alpha helices. “*” indicate secondary structures conserved in both domains. (B) Comparison of the AlphaFold 3 predicted structures of the VuCA1Pase domains, showing similarity of secondary structures and the position of catalytic residues. Orange indicates catalytic residues and the mirrored residues found in the regulatory domain. (C) Tree alignment of 10 CA1Pase sequences and 1 CA1Pase-like sequence, representing ancestral branches from a wide range of organisms from cyanobacteria to several plant families. Any predicted chloroplast transit peptides were removed, starting alignment and numbering at the lysine of the conserved “KRVV” motif of VuCA1Pase. Green and yellow bars indicate the location of the catalytic and regulatory domains. The positions of catalytic (orange) and mirror (dark yellow) residues, and the putative regulatory cysteine residues (black) are shown in the boxes containing sequence alignment sections at the bottom. bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m 90° bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 3.** The regulatory domain of CA1Pase is required for dimer formation. (A) Side and top view of the AlphaFold 3 predicted dimer structure of VuCA1Pase. Coloured regions represent residues conserved in >95% of 252 CA1Pase sequences, ranging from monocots to dicots. Green and yellow represent conserved catalytic and regulatory domain residues, respectively. The presence of the predicted disulfide bond formation is shown within the black box and indicated by black arrows. (B) A zoom into the redox motif region shown in the dotted line box in Figure 3A. Cysteines 292 and 297 are shown along with the presence of the predicted disulfide bond. This motif neighbours conserved residues that are part of the predicted dimer interface. (C-F) Size exclusion chromatography (SEC) traces for (C) protein standards (γ-globulins, Ovalbumin, Ribonuclease A type I-A), (D) VuCA1Pase, (E) VuCA1Pase lacking the regulatory domain (VuC H8A,1-206 ), and (F) VuCA1Pase lacking the catalytic domain (VuC 207-445 ). Estimated size in kilodaltons (kDa) is indicated below each peak, shown by a dotted grey line. bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) ## A available under a B ## VuCA1Pase VuC **Control** **5mM DTT** **Control** **5mM DTT** ## C VuC C297A D VuC C292A,C297A **Control** **5mM DTT** **Control** **5mM DTT** bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 4.** Redox regulation of CA1Pase is linked to dimer formation. (A-D) Graphs at the top of each panel display CA1Pase specific activity in presence of increasing concentrations of reduced glutathione (GSH, 0-100%). Graphs below show the traces of size exclusion chromatography in the absence (left) or presence of 5 mM DTT reducing agent (right), with predicted peak sizes. (A) VuCA1Pase, (B) single cysteine mutant VuC C292A, (C) single cysteine mutant VuC C297A, (D) double cysteine mutant VuC C292A,C297A . Technical replicates for each GSH treatment (n) = 3-4. Identified peaks are denoted by a grey dotted line underneath. bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) # 292 297 bo pep t do [ttps//doog/06898/06](https://doi.org/10.64898/2026.11.05.764501) 05 650 ; tseso posted oe be9, 06 ecopygt ode otspep (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is m [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) **Fig. 5.** C297 variants in Fabaceae CA1Pases correlate with low Rubisco dark inhibition. Alignment of mature CA1Pase from 15 Fabaceae species, ordered by tree alignment (left), highlighting conservation of the CXXXXC motif. Species represent those from Fabaceae with both sequence and dark inhibition data available (see Nehls-Ramos _et al._, 2026). Positions of the regulatory domain CXXXXC motif (C292 & C297) identified in _Vigna_ _unguiculata_ are annotated. Residues are numbered as described in Materials and Methods. Black highlight indicates residue differences from consensus. Coloured boxes at right indicate known prevalence of dark inhibition, as classified by Nehls-Ramos _et al._ (2026).