Evidence for dual targeting control of Arabidopsis 6-phosphogluconate dehydrogenase isoforms by N-terminal phosphorylation

Abstract

The oxidative pentose-phosphate pathway (OPPP) retrieves NADPH from glucose-6-phosphate, which is important in chloroplasts at night and in plastids of heterotrophic tissues. We previously studied how OPPP enzymes may transiently locate to peroxisomes, but how this is achieved for the third enzyme remained unclear. By extending our genetic approach, we demonstrated that Arabidopsis isoform 6-phosphogluconate dehydrogenase 2 (PGD2) is indispensable in peroxisomes during fertilization, and investigated why all PGD-reporter fusions show a mostly cytosolic pattern. A previously published interaction of a plant PGD with thioredoxin m was confirmed using Trxm2 for yeast two-hybrid (Y2H) and bimolecular fluorescent complementation (BiFC) assays, and medial reporter fusions (with both ends accessible) proved to be beneficial for studying peroxisomal targeting of PGD2. Of special importance were phosphomimetic changes at Thr6, resulting in a clear targeting switch to peroxisomes, while a similar change at position Ser7 in PGD1 conferred plastid import. Apparently, efficient subcellular localization can be achieved by activating an unknown kinase, either early after or during translation. N-terminal phosphorylation of PGD2 interfered with dimerization in the cytosol, thus allowing accessibility of the C-terminal peroxisomal targeting signal (PTS1). Notably, we identified amino acid positions that are conserved among plant PGD homologues, with PTS1 motifs first appearing in ferns, suggesting a functional link to fertilization during the evolution of seed plants.

Keywords: Arabidopsis PGD2; NADPH provision; OPPP; dual targeting; monomeric import; peroxisomes; protein phosphorylation.

Fig. 1.

PGD2 activity is needed inside peroxisomes during fertilization. (A) Scheme of genomic complementation constructs and frequency of genotypes for the given transformed PGD2 pgd2-1 lines (SALK_036751) in the T₂ generation. No homozygous plants were obtained for the ∆SKI version (dash), even when the corresponding recombinant His-tag protein was expressed in E. coli (B), while weak PTS1 (-SEI) sufficed for normal Mendelian distribution. Pro, promoter (genomic PGD2 fragment); Asterisks (*) mark data from Hölscher et al. (2016). (B) Activity of the PGD2 cDNA variants cloned in pET16b and purified from G6PDH-deficient E. coli strain BL21^minus (Meyer et al., 2011). Aliquots (5 µl) were used for SDS–PAGE separation and blot transfer. The Ponceau S-stained blot (protein) was developed with PGD2 antiserum (α-PGD; Hölscher et al., 2016) using elution fractions of equivalent signal strength (see Supplementary Fig. S1) for comparison on the same blot; pkat, 6PGDH activity (pmol s^–1); molecular masses in kDa, PageRuler™ Prestained Protein Ladder (Fermentas).

Fig. 2.

All PGD isoforms interact with thioredoxin m2 in yeast and plant cells. (A) Yeast two-hybrid assay of the indicated binding and activation domain (BD, AD) combinations in strain SMY3. With the empty vector, Trx_m2 showed autoactivation (auto; Meyer et al., 2011) that was revoked by all PGD isoforms (+, top), but among the N-terminally truncated versions only by PGD2_Δ15 (bottom). Cat2, catalase 2 (main Arabidopsis isoform of green tissues) served as negative control. (B and C) Bimolecular fluorescent complementation (BiFC) analyses in Arabidopsis protoplasts isolated from leaves (48 h post-transfection). (B) Trx_m2 combined with the PGD isoforms; top, with C-terminal split YFP; bottom, with N-terminal split YFP (yellow); chlorophyll autofluorescence is in blue. (C) With N-terminal split YFP, Trx_m2 and PGD2_Δ15 also localized in peroxisomes. All images show maximal projections of ~35 single sections. BiFC is in green and the peroxisomal marker (OFP–PGL3_C-short) is in magenta. Co-localization of green and magenta, or very close signals (<200 nm), appear white. Scale bars 3 µm.

Fig. 3.

Exchange of the unique cysteine speeds up PGD2 localization in peroxisomes. (A) Bimolecular fluorescence complementation (BiFC) signals of PGD2_Δ15 in Arabidopsis protoplasts isolated from leaves (24–48 h post-transfection). Left panels without and right panels with the peroxisome marker (OFP–PGL3_C-short). WT, wild-type PGD2_Δ15; C420S, unique cysteine mutated to serine. All images show maximal projections of ~35 single sections (for single channel images, see Supplementary Fig. S3). BiFC signals are in yellow/green, peroxisomal marker in magenta, and chlorophyll autofluorescence in blue. Co-localization of green and magenta, or very close signals (<200 nm), appear white. Scale bars 3 µm. (B) Frequency of cells with a peroxisomal pattern (% per signal) indicates faster uptake of the C420S version. (C) 3D model of PGD2 (based on X-ray crystallography of the enzyme from sheep; Adams et al., 1991) with unique cysteines on opposite sides of the dimer interface (red frame).

Fig. 4.

A new medial GFP fusion of PGD2 is active but does not localize in peroxisomes. (A) PGD2 monomer (PDB file obtained from Alphafold, and modifed by Protean 3D) with a large dimer interface (dashed line), the C-terminally extended part with GFP insertion positions (position 476 of Eubel et al., 2008), and C420 indicated by arrows. (B) PGD2 constructs with the medial GFP reporter (numbers refer to amino acids of the N- and C-terminal parts) and their analysis in Arabidopsis protoplasts isolated from leaves (48 h post-transfection). Note that old medial and new medial display opposite localization patterns with an influence on catalytic activity, indicated on the left. All images show maximal projections of ~35 single sections. GFP fusions are in green, peroxisomal marker (OFP–PGL3_C-short) in magenta, and chlorophyll autofluorescence in blue. Merge, co-localization of green and magenta, or very close signals (<200 nm), appear white. Scale bars 3 µm.

Fig. 5.

Analysis of N- and C-terminal amino acid deletions in PGD2_new medial. (A) Alignment of the PGD2_new medial versions with gradual deletion of N-terminal amino acid residues. (B) Localization of N-terminally truncated PGD2_new medial versions in Arabidopsis protoplasts isolated from leaves (48 h post-transfection) showed that the border for a clear peroxisomal pattern lies between Δ4 and Δ5 amino acids, hinting at the importance of residue T6 in PGD2. All images show maximal projections of ~35 single sections (merge, for single channel images, see Supplementary Fig. S4). PGD2 fusions are in green, peroxisomal marker (OFP–PGL3_C-short) in magenta, and chlorophyll autofluorescence in blue. Co-localization of green and magenta, or very close signals (<200 nm), appear white. Scale bars 3 µm. (C) Logo plot for PGD sequences from different vascular plant clades (61 sequences in total), highlighting amino acid conservation within the N-terminus by letter size. Colors correspond to frames in (A).

Fig. 6.

N-terminal phosphomimetic changes promote PGD targeting to organelles. (A) Exchange of T6 to phosphomimetic E or D (but not V or M) in PGD2_new medial (arrow) led to a switch in localization (right) from largely cytosolic to a 100% peroxisomal pattern (left). Note that also for the wild tpye (WT) and the C420S version, cells with a partial peroxisomal pattern were scored (white arrows). (B) In GFP–PGD2_full, the T6E version showed a partial peroxisomal pattern, suggesting that binding of a cytosolic factor is affected by the N-terminally fused reporter. C420S (black arrow) had no effect. (C) In PGD1_full–GFP, exchange of S7 by phosphomimetic D (but not A) led to an exclusive plastidial pattern. All images show maximal projections of ~35 single sections (merge; for single channel images, see Supplementary Fig. S6). GFP fusions are in green, peroxisomal marker (OFP–PGL3_C-short) or plastidial marker (GPT2_5MD–OFP; Baune et al., 2020) in magenta, and chlorophyll autofluorescence in blue. Co-localization of green and magenta or very close signals (<200 nm) appear white. Scale bars 3 µm.

Fig. 7.

The phosphomimetic change T6E lowers PGD2 solubility and catalytic activity. (A) AlphaFold prediction of intramolecular bonds (within a radius of 5 Å) in the N-terminal part of PGD2. The magnification shows that Thr6 (red) is bound by surrounding Ile8, Ser32, Ser68, and Gln70 (green). (B) Relative protein amounts of PGD2 medial variants compared with endogenous PGD signals in cleared extracts of Arabidopsis protoplasts. Note that compared with PGD2_new medial, T6E and old medial versions are depleted from the supernatant (SN) but less from the pellet (Pe) fractions, indicative of lower solubility. (C) Relative activity of the indicated PGD2–GFP fusions compared with the corresponding wild type (WT, 100%) upon pull-down from supernatant fractions; n.d., not detected. The immunoblot was developed with anti-PGD2 antiserum (Hölscher et al., 2016). Note that with N-terminal GFP, PGD2-T6E showed catalytic activity. Compilation of different blots (marked by outlines). The lighter lanes stem from the same blot from which empty lanes and the marker were removed. (D) SDS–PAGE of His-PGD2 variants from cleared E. coli BL21^minus extracts without (–) or with (+) SDS and DTT in the sample buffer (plus boiling at 100 °C). The immunoblot was developed with anti-His–HRP conjugate to visualize only recombinant PGD proteins. kDa, PageRuler™ Prestained Protein Ladder (Fermentas).

Fig. 8.

Phylogenetic analyses support a role for PGD2 in peroxisomes of seed plants. The tree was built with amino acid sequences selected from different species. Left, maximum likelihood tree with highest log likelihood (–25 210.27); 70 sequences were compiled with bifunctional glyoxylate reductase isoforms (AtGLYR1/2) as outgroup. Additional features of the PGD sequences are depicted in table format, highlighting relevant positions based on AtPGD2. Note that a PTS1 motif is common among the spermatophyta (first appearing in ferns, yellow background), except for the seagrass Zostera marina (red letters) that lives submerged in shallow marine terrain. Asterisks (*) mark the oldest clades of the angiosperms. Subtrees were flipped to present the peroxisomal -TRI- group of A. thaliana on top.

Fig. 9.

Model of dual targeting regulation by N-terminal PGD phosphorylation. (A) Early homodimerization results in retention of most plastidial PGD1/3 and peroxisomal PGD2 precursors in the cytosol. Binding by chaperons (chap, grey/green) and thioredoxin (Trx, orange) aids in plastid import of PGD1/3 precursors and infrequent peroxisomal uptake of PGD2 monomers (possibly due to cysteine modification/oxidation), as suggested by localization analyses and proteomics data. (B) An unknown stimulus activates kinases (flashes) that phosphorylate the PGD N-termini (blue circles), enforcing uptake into plastids and peroxisomes, respectively (possibly involving a 14-3-3 protein, red). In the target organelles, phosphatase activity is needed to form active PGD dimers. Note that the N-termini of PGD1/3 must not be cleaved upon plastid import (renders them inactive), while the dimerization interface of folded but monomeric PGD2 has to be shielded from unspecific aggregation to enable Pex5 binding to the PTS1 motif (-SKI) for peroxisomal import.