Evidence for dual targeting of Arabidopsis plastidial glucose-6-phosphate transporter GPT1 to peroxisomes via the ER

Former studies on Arabidopsis glucose-6-phosphate/phosphate translocator isoforms GPT1 and GPT2 reported viability of gpt2 mutants, however an essential function for GPT1, manifesting as a variety of gpt1 defects in the heterozygous state during fertilization/seed set. Among other functions, GPT1 is important for pollen and embryo-sac development. Since previous work on enzymes of the oxidative pentose phosphate pathway (OPPP) revealed comparable effects, we investigated whether GPT1 might dually localize to plastids and peroxisomes. In reporter fusions, GPT2 was found at plastids, but GPT1 also at the endoplasmic reticulum (ER) and around peroxisomes. GPT1 contacted oxidoreductases and also peroxins that mediate import of peroxisomal membrane proteins from the ER, hinting at dual localization. Reconstitution in yeast proteoliposomes revealed that GPT1 preferentially exchanges glucose-6-phosphate for ribulose-5-phosphate. Complementation analyses of heterozygous gpt1 plants demonstrated that GPT2 is unable to compensate for GPT1 in plastids, whereas genomic GPT1 without transit peptide (enforcing ER/peroxisomal localization) increased gpt1 transmission significantly. Since OPPP activity in peroxisomes is essential during fertilization, and immuno-blot analyses hinted at unprocessed GPT1-specific bands, our findings suggest that GPT1 is indispensable at both plastids and peroxisomes. Together with the G6P-Ru5P exchange preference, dual targeting explains why GPT1 exerts functions distinct from GPT2 in Arabidopsis. One sentence summary In contrast to plastidial GPT2, GPT1 exhibits slightly different exchange preferences and alternatively targets the ER, from where the protein can be relocated to peroxisomes on demand.


peroxisomes. 478
In this respect, GPT1 release to peroxisomes may require interaction with Grx c1 (and Trx h7 ), 479 known to engage in monothiol-dithiol mechanisms, including glutathionylation (Riondet et al.,480 2012; . The latter is known to be triggered by oxidative transients that 481 accompany stress signaling and developmental change (2GSHGSSG). Sensible cysteine 482 residues (-S  at physiological pH) may become sulfenylated (-S-OH in the presence of H 2 O 2 ) or 483 glutathionylated (-S-SG), which protects from over-oxidation (reviewed in . 484 Reversion (de-glutathionylation) by GSH alone is slow, but fast together with Grx and Trx (as 485 recently shown for plastidial Amy3; Gurrieri et al., 2019). Perhaps this mechanism regulates 486 GPT1 interaction with Pex16 and/or Pex3, given that biochemically distinct ER vesicles were 487 shown to fuse and form new peroxisomes (Van Der Zand et al., 2012). In any case, GPT1 488 transport in monomeric form within the ER makes sense, since a potentially active translocator -489 still en route to its final destination -is likely not tolerated. This idea is supported by aberrant ER 490 structure in analyses with enforced GPT1-dimer formation ( Figure 3) protein partially resides at the ER. Grx c1 promoted ER targeting of GPT1, also without N-519 myristoylation motif (G2A) in grx c1 mutant protoplasts (not shown), indicating functional 520 redundancy with (an)other isoform/member(s) of the Grx/Trx superfamily. Interestingly, GPT1 is 521 listed as palmitoylation candidate by the plant membrane protein database Aramemnon 522 (http://aramemnon.uni-koeln.de) with high score. Protein S-acylation (via cysteine residues) is 523 still a poorly understood posttranslational process that is usually preceded by N-myristoylation, 524 to promote membrane association, targeting, and/or partitioning into membrane subdomains 525 (Aicart-Ramos et al., 2011; Hemsley, 2015). A potential role of Grx/Trx N-myristoylation for 526 putative S-palmitoylation of GPT1 will have to be analyzed by a complex experimental setup, a 527 difficult task considering partial redundancy among cytosolic Trx h2, h7, h8, h9 as well as Grx c1 528 and c2 isoforms Majeran et al., 2018). Clearly, GPT1 529 is inserted into the ER membrane in monomeric form, and may be modified at C65 ( Figure 9A, 530 question mark) for retention. Dimer formation beyond the perER would occur after de-protection, 531 likely triggered by cytosolic redox signaling that accompanies a/biotic stress responses 532 Foyer et al., 2009) or specific developmental stages, like pollen tube 533 elongation (Considine and Foyer, 2014) and navigation to ovules (Hölscher et al., 2016). 534

536
Our BiFC data suggested that GPT1 contacts at least two of the three early peroxins (Kim and 537 Mullen, 2013). Interaction with Pex3 and Pex16 was detected at the ER and PerMs, whereas 538 interaction with Pex19 was mostly distributed across the cytosol, reflecting its function as 539 cytosolic cargo receptor (Hadden et al., 2006). Since simple co-expression with Pex19-reporter 540 fusions did not show any change in GPT1 localization, dot-like structures labeled by GPT1-541 Pex19 BiFC analyses might be a false-positive result. This would be in line with Pex19 being 542 mainly involved in targeting of class I, but not class II PMPs. Focal localization of GPT1 at the 543 ER, previously described for Pex3 in yeast and for pxAPX in cottonseed/APX3 in Arabidopsis 544 (Lisenbee et al., 2003;Narendra et al., 2006), was mainly seen upon BiFC, indicating that 545 dimerization occurs beyond the perER. GPT1 dimers may therefore represent a forced 546 interaction at the ER, which does not (yet) occur under physiological conditions. As a side note, 547 Pex3 of plant cells had not been detected at the ER before (Hunt and Trelease, 2004). 548 Usually, GPT1 distributed evenly across the ER, unless co-expressed with Pex16 that coexists 549 at both the ER and PerMs (Lin et al., 2004;Karnik and Trelease, 2005). Interestingly, presence 550 of Pex16 influenced GPT1 localization at the ER, resulting in a similar but distinct patternalso 551 when driven by the own promoter (dark incubation in the presence of sugars activates GPT1 552 mRNA expression, Supplemental Figure 18). Considering that BiFC is not dynamic, and fluores-553 cent signals persist once the split YFP halves are reconstituted , 554 GPT1 was likely dragged to PerMs upon (otherwise transient) interaction with the peroxins. In 555 any case, this demonstrated that GPT1 can reach PerMs (although not detected there, unless 556 triggered), wherefore the transporter may first interact with Pex16 (for ER insertion/transport to 557 the perER; Hua et al., 2015), and then Pex3 (and possibly Pex19, during sorting to PerMs). By 558 contrast to APX3, GPT1 is only needed at peroxisomes when the OPPP is required (Meyer et

GPT1 transport preference differs from GPT2
564 After plastid import, TP sequences are cleaved off by the essential stromal processing peptidase 565 (SPP), which is usually important for maturation, stabilization, and activation of the proteins (van 566 Wijk, 2015). Here we show that also unprocessed GPT1 is an active transporter. Addition of a 567 small tag or large reporter did not influence transport activity. Furthermore, topology analyses of 568 roGFP fusions indicated that upon ER insertion, both N-and C-termini of GPT1 face the cytosol 569 (Supplemental Figure 12), similar to Arabidopsis PMP22 (Murphy et al., 2003) and the human 570 glucose transporter (Mueckler and Lodish, 1986). These findings support the theory of Shao and 571 Hegde (2011) that during post-translational ER import of membrane proteins, type-I topology (N-572 terminus facing the lumen) is strongly disfavored. This leads to obligate type-II topology (N-573 terminus facing the cytosol), and integration of the following MDs owing to the 'positive inside 574 rule' (von Heijne, 1986;Goder et al., 2004) for the cytosolic hinge regions. The latter is not 575 entirely true for the GPT proteins (marked red in Supplemental Figure 1 and the topology 576 models), which may facilitate posttranslational ER insertion. 577 The phosphate translocator family is known to form dimers that mediate strict counter-exchange 578 of various phosphorylated metabolites with inorganic phosphate (Pi). The ability to transport 579 other OPPP intermediates, although possible (e.g. triose-phosphates), is usually disfavored due 580 to the prevailing metabolite concentrations or competition with the preferred substrate (Flügge, 581 1999;Eicks et al., 2002). Here we show that GPT1 and GPT2 can exchange G6P for Ru5P, but 582 GPT1 has a stronger preference for Ru5P. Thus, import of the OPPP substrate and export of its 583 product is warranted across PerMs ( Figure 9B). Moreover, poor rates obtained with 6-phospho-584 gluconate (6PG) as counter-exchange substrate strongly suggest that sugar-derived NADPH  In principle, the discovered transport preference should also apply to metabolite exchange at 589 plastids. This may explain why Arabidopsis tpt xpt double mutants are viable (although strongly 590 growth-compromised; Hilgers et al., 2018) and why rpi2 mutants, lacking one of the two cytosolic 591 ribose-phosphate isomerase (RPI) isoforms form less starch in leaves . 592 Minute amounts of active GPT1 could drain G6P from chloroplasts due to preferred exchange 593 with Ru5P, likely more abundant in rpi2 mutants (Supplemental Figure 19). Besides G6P 594 exchange needed to stabilize the Calvin cycle (Sharkey and Weise, 2016), this argues for a role 595 of ubiquitously expressed GPT1, considering that GPT2 is absent from unstressed leaves 596 (Supplemental Figure 14F). On the other hand, lower transport capacity of GPT1 compared to 597 GPT2 is not surprising, since our data confirm a specialization of the two transporters. For 598 GPT1's function, flux rates are not necessarily a limiting parameter, but substrate specificity 599 obviously is. This is in line with our complementation analyses, demonstrating that GPT2 cannot 600 compensate for the absence of GPT1. Browser , in this stage mRNA expression of GPT2 is up to 3.5-fold higher 609 than of GPT1 (Supplemental Figure 17), which can explain the observed starch accumulation 610 upon GPT1 loss. 611 In accordance with these premises, we suspected that ectopic GPT2 expression may rescue 612 some plastidial functions, but not all phenotypes of the mutant gpt1 alleles, because swap 613 constructs headed by GPT2 were never detected at the ER. For heterozygous gpt1-2 614 transformed with GPT2 (driven by the GPT1 promoter), filled siliques with green, non-aborted 615 embryos, and fertilized, but later aborted brownish embryos were observed. Plants homozygous 616 for the gpt1-2 T-DNA were absent from the progeny of this line and also from ER/peroxisomal 617 compensated Pro35S:GFP-GPT1_C-mat. 618 Upon reciprocal crossing of these two lines, only one direction worked (Table 3), indicating that 619 besides partial rescue of the female gpt1 defects (showing as filled siliques), plastid-confined 620 GPT2 was unable to fully rescue GPT1's functions during pollen maturation/tube growth. Pollen 621 grains appeared normal, but no homozygous gpt1-2 plants were found among the progeny of 622 combined complementation constructs. This suggested that the remaining defects result mainly 623 from absence of GPT1 from plastids, due to a unique function GPT2 cannot fulfill. Furthermore, 624 GPT1 transfer from the ER to peroxisomes might be impeded by artificial construct composition. 625 Of note, Pro35S:GFP-GPT1_C-mat (transport-competent ER/PerM control) did not rescue ovule 626 abortion (Table 2), but led to a substantial increase in heterozygous offspring compared to the 627 parental line (Table 3). This may be even an underestimation, since the CaMV-35S promoter is 628 not well expressed in pollen, and generally fluctuates in floral tissues . By 629 contrast, the ProGPT1-driven GPT1_N-long mat construct (without TP) rescued seed set and 630 raised gpt1 transmission up to 43%, independent of additional GPT2 in plastids. Thus, together 631 with the pollination defect (mentioned above) and complementation by a genomic GPT1 632 construct (Niewiadomski et al., 2005), our results indicate that for full rescue GPT1 is additionally 633 needed in plastids, where the OPPP is mainly required for Ru5P provision to nucleotide bio-634 synthesis ( Figure 9B), as recently shown by Andriotis and Smith (2019). 635 The findings nicely support our previous analyses that loss of Ru5P formation in peroxisomes 636 of these aspects to dual targeting of GPT1 will require more detailed studies. 656

657
In summary, our data present compelling evidence for dual targeting of GPT1 to both plastids 658 and peroxisomes. Imported G6P is converted by the oxidative OPPP part to NADPH and Ru5P, 659 which is the preferred exchange substrate (likely at both locations), thus contributing to 660 gametophyte and embryo development as well as pollen-tube guidance to ovules. Since the 661 latter dominates the reproductive success, further analyses are required to determine the exact 662 physiological context of GPT1's presence at the ER/peroxisomes.

695
For in vitro-uptake studies, full-length or mature GPT1 and GPT2 versions were amplified with 696 the corresponding primers from cDNA and inserted into yeast vectors pYES2 or pYES-NTa via 697 Acc65I (KpnI)/BamHI sites (Thermo Scientific). For full-length GPT1, primer combinations were 698 GPT1_Acc65I_s with GPT1+S_BamHI_as; for mature GPT1, GPT1_C-mat_Acc65I_s with 699 GPT1+S_BamHI_as; and for mature GPT2, GPT2_C-mat_Acc65I_s with GPT2+S_BamHI_as 700 (Supplemental Table 2). For the GFP-GPT1_C-mat version, PCR fragments (primers: GPT1_C-701 mat_SpeI_s and GPT1+S_BamHI_as) were first inserted into pGFP2-SDM via SpeI/BamHI 702 sites, released with KpnI/BamHI, and cloned in pYES2. The resulting constructs were 703 transformed into strain INVSc1 (MATa, his3∆1, leu2, trp1-289, ura3-52/MATα,his3∆1, leu2, trp1-704 289, ura3-52) using the lithium acetate/PEG method (Gietz and Schiestl, 2007). Yeast cells were 705 selected on synthetic complete medium (SC-Ura; 0.67% (w/v) YNB supplemented with 706 appropriate amino acids and bases for uracil auxotrophy and 2% (w/v) glucose as carbon 707 source). Since protein expression is under control of the galactose-inducible promoter pGAL1, 708 yeast cells were grown aerobically in SC-Ura supplemented with 2% (w/v) galactose for 6 h at 709 30°C. Harvest and enrichment of total yeast membranes without and with recombinant GPT 710 proteins was performed according to Linka et al. (2008).  Table 2).    Tables (601 words) 884 885     (compare His-matGPT1 to GFP-matGPT1). In all graphs, the arithmetic mean of 3 technical 1014 replicates (±SD) was plotted against time (see Table 1    The schemes illustrate different orientation of the candidate proteins with respect to free N-and C-terminal ends. GPT1 interacts with both oxidoreductases (green signals) at the endoplasmic reticulum (ER) and its spherical sub-structures (arrowheads), except when the N-terminus of Grx c1 is masked (B, panels c and d). Note that these substructures differ from those labelled in Figure 3B. Note that transport rates of GPT1 are not influenced by the N-terminal tag (compare His-matGPT1 to GFP-matGPT1). In all graphs, the arithmetic mean of 3 technical replicates (±SD) was plotted against time (see Table 1  . GPT1 detection at the ER is increased by stress treatment and in reproductive Arabidopsis tissues. A, Arabidopsis protoplasts were co-transfected with the indicated GPT-GFP fusions and the peroxisome marker (Per, OFP-PGL3_C-short), samples were split in half, one was treated with 0.2 µM flagellin peptide (+flg22), and the other mockincubated for 24 h. Note that flg22 treatment did not change GPT localization to plastids, but enhanced the ER fraction of GPT1-GFP (arrowheads). All images show maximal projections of approximately 30 single sections (Merge; for single channel images, see Supplemental Figure S13). GFP fusions in green, peroxisome marker in magenta, and chlorophyll fluorescence in blue. Co-localization of magenta and green or very close signals (less than 200 nm) appear white in the Merge of all channels. Bars = 3 µm. B-C, Protein extracts (without detergent) of flower, leaf, and (green) silique tissue were prepared from wild-type plants (Col, Ws) and the indicated homozygous mutant lines. Supernatant fractions were separated on 10% SDS gels and blotted to nitrocellulose. After Ponceau-S staining, the blots were developed with GPT1-specific antibodies (-GPT1) raised against the N-terminus with His-tag (Supplemental Figure S14). Arrowheads mark double bands of full-length GPT1 (predicted size: 42.3 kDa) and mature GPT1 (ca. 37-39 kDa, depending on TP processing). Red arrowheads point to bands suspected to represent a largely 'off' situation and black arrowheads the corresponding 'on' situation at either location (as deduced from comparison of leaf to silique tissue), likely due to protein modification. C, Immunoblot of seedlings harvested from germination plates (1% sucrose) after 1-or 4-week (w) growth in short-day regime.   (Jones et al., 1992). The tree with highest log likelihood (-5414.98) is shown. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 34 amino acid sequences (Supplemental Table 3 . Whether this involves 65 C in the GPT1 N-terminus is unclear (question mark). ER insertion involves Sec61 and sorting to peroxisomal membranes specific peroxins (Pex). Brefeldin A (BFA) blocked ER import of GPT1. B, Scheme of sugar metabolism in a physiological sink state. Sucrose (suc) is cleaved by cytosolic invertase yielding two hexoses (hex) that are activated by hexokinase (HXK), consuming ATP provided by glycolysis and mitochondrial respiration (not shown). By contrast to GPT2, GPT1 imports G6P into both plastids (in exchange for Pi released by GPT2-driven starch synthesis) and peroxisomes (in exchange for Ru5P that may also enter plastids via GPT1, dashed red arrows), yielding 2 moles of NADPH in the oxidative part of the OPPP. NADP inside peroxisomes is formed by NAD kinase (NADK3) that relies on ATP and NAD imported into peroxisomes via PNC (At3g05290; At5g27520) and PXN (At2g39970). The cytosolic OPPP reactions are usually linked via RPE and XPT to the complete pathway in the plastid stroma. Abbreviations: G6PD, glucose-6-phosphate dehydrogenase; PGL, 6-phosphogluconolactonase; PGD, 6-phosphogluconate dehydrogenase; RPE, ribulosephosphate-3-epimerase; RPI, ribose-5-phosphate isomerase.