Chloroplast FBPase and SBPase are thioredoxin-linked enzymes with similar architecture but different evolutionary histories

Abstract

The Calvin-Benson cycle of carbon dioxide fixation in chloroplasts is controlled by light-dependent redox reactions that target specific enzymes. Of the regulatory members of the cycle, our knowledge of sedoheptulose-1,7-bisphosphatase (SBPase) is particularly scanty, despite growing evidence for its importance and link to plant productivity. To help fill this gap, we have purified, crystallized, and characterized the recombinant form of the enzyme together with the better studied fructose-1,6-bisphosphatase (FBPase), in both cases from the moss Physcomitrella patens (Pp). Overall, the moss enzymes resembled their counterparts from seed plants, including oligomeric organization-PpSBPase is a dimer, and PpFBPase is a tetramer. The two phosphatases showed striking structural homology to each other, differing primarily in their solvent-exposed surface areas in a manner accounting for their specificity for seven-carbon (sedoheptulose) and six-carbon (fructose) sugar bisphosphate substrates. The two enzymes had a similar redox potential for their regulatory redox-active disulfides (-310 mV for PpSBPase vs. -290 mV for PpFBPase), requirement for Mg(2+) and thioredoxin (TRX) specificity (TRX f > TRX m). Previously known to differ in the position and sequence of their regulatory cysteines, the enzymes unexpectedly showed unique evolutionary histories. The FBPase gene originated in bacteria in conjunction with the endosymbiotic event giving rise to mitochondria, whereas SBPase arose from an archaeal gene resident in the eukaryotic host. These findings raise the question of how enzymes with such different evolutionary origins achieved structural similarity and adapted to control by the same light-dependent photosynthetic mechanism-namely ferredoxin, ferredoxin-thioredoxin reductase, and thioredoxin.

Keywords: Calvin–Benson cycle; fructose-1,6-bisphosphatase; redox regulation; sedoheptulose-1,7-bisphosphatase; thiol–disulfide exchange.

Fig. 1.

Structural overview of PpFBPase (PDB ID code 5IZ1) and PpSBPase (PDB ID code 5IZ3). Regulatory cysteines are highlighted. The active sites are represented as surface areas for each monomer. (A) PpFBPase. (B) PpSBPase. (C) Superposition of PpFBPase (green) and PpSBPase (orange) monomers.

Fig. S1.

Alignment of selected FBPase and SBPase amino acid sequences with Clustal Omega (www.ebi.ac.uk/Tools/msa/clustalo/). At, Arabidopsis thaliana; Cr, Chlamydomonas reinhardtii; Gt, Guillardia theta; Os, Oryza sativa; Pp, Physcomitrella patens; Ps, Pisum sativum; Pt, Populus trichocarpa; Tb, Trypanosoma brucei; Tg, Toxoplasma gondii; Tt, Tetrahymena thermophilia. Chloro, chloroplastic; Cyto, cytosolic. Red/bold, redox-sensitive cysteines; bold/underlined, conserved residues in FBPase and SBPase in the potential sugar bisphosphate binding site; brown/underlined, start of the recombinant proteins.

Fig. S1.

Alignment of selected FBPase and SBPase amino acid sequences with Clustal Omega (www.ebi.ac.uk/Tools/msa/clustalo/). At, Arabidopsis thaliana; Cr, Chlamydomonas reinhardtii; Gt, Guillardia theta; Os, Oryza sativa; Pp, Physcomitrella patens; Ps, Pisum sativum; Pt, Populus trichocarpa; Tb, Trypanosoma brucei; Tg, Toxoplasma gondii; Tt, Tetrahymena thermophilia. Chloro, chloroplastic; Cyto, cytosolic. Red/bold, redox-sensitive cysteines; bold/underlined, conserved residues in FBPase and SBPase in the potential sugar bisphosphate binding site; brown/underlined, start of the recombinant proteins.

Fig. S2.

ConSurf analysis for PpFBPase (A) and PpSBPase (B). Multiple sequence alignment was built with the alignment method MAFFT using the CS-BLAST search algorithm for homologs (ranging from 10% to 90% sequence identity) from the UniRef90 database. Residue conservation is plotted onto the surface of both enzymes and is colored according to conservation scores ranging from 1 (cyan, low conservation) to 9 (purple, high identity). The figure reveals that the active sites of both enzymes are highly conserved and that the C1–C2 and C3–C4 interfaces are well-conserved, unlike the C1–C4 and C2–C3 interfaces of PpFBPase. The information was obtained and is shown using the PyMOL output file from the ConSurf server (24).

Fig. S3.

Electrostatic potential of the molecular surface of PpFBPase and PpSBPase monomers. Negatively charged residues are colored in red; positively charged areas are colored in blue; and neutral regions are indicated in white.

Fig. S4.

Topology diagram of P. patens FBPase and comparison with SBPase. Structural elements shared between PpFBPase and PpSBPase are colored in black, additional elements present in PpFBPase are colored in blue, and additional secondary structures present in PpSBPase are colored in red. The extended loop present in PpFBPase between the β1- and β2-strands is indicated.

Fig. 2.

Regulatory aspects of PpFBPase and PpSBPase. (A) Dependency of phosphatases on thioredoxin. Dark gray bars show the activity of the enzymes reduced with TRX f, and light gray bars show the activity with TRX m. Activities are depicted in mol substrate transformed per s/mol enzyme. Both FBPase and SBPase activities were evaluated using the coupled spectrophotometric assay and the “alternate” FBP substrate for SBPase. (B) Redox potential. Midpoint redox potentials estimated by SDS/PAGE following methoxy-PEG (mPEG)-maleimide labeling. Both proteins were treated with various ratios of oxidized and reduced DTT and then labeled with mPEG-maleimide. The oxidation–reduction potential was read at the point indicating that the protein was half-oxidized and half-reduced. (C) Time course of reduction. mPEG-maleimide labeling by reduction with 10 mM DTT and 3 µM TRX f at pH 7.0. (D) Time course of activation. Experimental conditions were as in C. Red, reduced; ox, oxidized. Error bars in A and D represent standard deviation.

Fig. S5.

Requirement of PpFBPase (A) and PpSBPase (B) for Mg²⁺. Results are shown for both the oxidized (ox) and reduced (red) enzymes with FBP as substrate. Error bars represent standard deviation.

Fig. S6.

Close-up view of both FBP and SBP surface binding sites based on homology modeling from 3D known structures. (A) Putative FBP binding site of PpFBPase with monomers A and B colored green and cyan, respectively. (B) Putative SBP binding site of PpSBPase with monomers A and B colored blue and orange, respectively. In each case, secondary structures are shown as cartoons, residues potentially involved in substrate binding are shown as sticks, and substrate molecules (FBP or SBP) are colored white and shown in a ball-and-stick representation. Both FBP and SBP molecules are stabilized through hydrogen bonds formed by ∼15 residues, including Arg336 for PpFBPase (Arg291 for PpSBPase) located within the other monomer.

Fig. 3.

Evolutionary origin of eukaryotic FBPase and SBPase. (A) Simplified version of the phylogenetic analysis performed (a detailed version is in Fig. S7). (B) Scheme illustrating the most parsimonious scenario for the acquisition and loss of FBPase and SBPase enzymes during evolution. A, SBPase; B, FBPase; C, cyanobacterial bifunctional enzyme; cp, chloroplastic; ct, cytosolic.

Fig. S7.

Phylogenetic tree of FBPases and SBPases. The 48 amino acid sequences used include 4 archaeal FBPases, 3 ε-proteobacterial FBPases, 10 eukaryotic SBPases, 3 α-proteobacterial FBPases, 4 eubacterial FBPases, 3 cyanobacterial FBPases class I, 11 eukaryotic cytosolic FBPases, and 10 eukaryotic chloroplast FBPases (for details, see Materials and Methods and Dataset S1). All α-proteobacterial FBPases consistently clustered with eukaryotic FBPases, and all archaeal FBPases with plant SBPases. Considering that horizontal gene transfer is continuously happening between prokaryotes (resulting in pangenomes) and that most prokaryote-derived genes were acquired during endosymbiotic events in eukaryotic evolution (38), the phylogenetic distribution we observed indicates that plant and animal FBPases originated from an α-proteobacterial ancestor, whereas plant SBPases originated from an archaeal ancestor. The arrow depicts the loss of one regulatory cysteine in a putatively secondary heterotrophic lineage, the alveolate T. thermophila.