Three thioredoxin targets in the inner envelope membrane of chloroplasts function in protein import and metabolism

Sandra Bartsch*, Julie Monnet†, Kristina Selbach†, Franc¸oise Quigley†, John Gray‡, Diter von Wettstein§¶, Steffen Reinbothe†, and Christiane Reinbothe*†¶

*Lehrstuhl fu¨r Pflanzenphysiologie, Universita¨t Bayreuth, Universita¨tsstrasse 30, D-95447 Bayreuth, Germany; †Unite´Mixte de Recherche 5575, Centre d’Etudes et de Recherches sur les Macromole´cules Organiques, Universite´Joseph Fourier et Centre National de la Recherche Scientifique, BP53, F-38041 Grenoble Cedex 9, France; ‡Department of Biological Sciences, University of Toledo, 2801 West Bancroft Street, Toledo, OH 43606; and §Department of Crop and Soil Sciences and School of Molecular Biosciences, Washington State University, Pullman WA 99164-6420

Contributed by Diter von Wettstein, January 16, 2008 (sent for review December 11, 2007) Thioredoxins (Trxs) are ubiquitous small proteins with a redox- of Trxs in dark/light regulation of protein translocation and active disulfide bridge. In their reduced form, they constitute very chlorophyll (Chl) metabolism and provoke a scenario of deter- efficient protein disulfide oxidoreductases. In chloroplasts, two rence to pigment-sensitized oxidative stress in chloroplasts. types of Trxs (f and m) coexist and play central roles in the regulation of the Calvin cycle and other processes. Here, we Results identified a class of Trx targets in the inner plastid envelope Affinity Purification of Plastid Envelope Proteins Interacting with Trx membrane of chloroplasts that share a CxxC motif Ϸ73 aa from f and Trx m. Chloroplast Trxs contain two neighboring Cys their carboxyl-terminal end. Members of this group belong to a residues in a conserved motif, Trp-Cys-Gly-Pro-Cys (in some superfamily of Rieske iron–sulfur proteins involved in protein cases Trp-Cys-Pro-Pro-Cys), that is referred to as the Trx translocation and chlorophyll metabolism. These proteins include signature (summarized in refs. 1–3). We began by designing the protein translocon protein TIC55, the precursor NADPH:proto- mutant Trx f and Trx m molecules in which the buried Cys was oxidoreductase translocon protein PTC52, which exchanged by a Ser or Ala residue. These were bound on operates as a-oxygenase, and the lethal leaf Sepharose 4B and tested for their capability to bind known Trx spot protein LLS1, which is identical with pheophorbide a oxygen- target proteins present in stromal extracts of barley chloroplasts. ase. The role of these proteins in dark/light regulation and oxida- Fig. 1A (lane 1) shows proteins that could be eluted with DTT. tive control by the Trx system is discussed. Protein mass spectrometry verified the presence of sedoheptu- lose-1,7-bisphosphate bisphosphatase (75 kDa), phosphoribu- protein translocation ͉ biosynthesis ͉ Tigrina d12 ͉ lokinase (80 kDa), and NADP-dependent malate dehydrogenase FLU mutant ͉ photooxidative stress (85 kDa), which are known to be Trx targets in spinach chloro- plasts (1, 3). hioredoxins (Trxs) are small multifunctional redox-active In a next step, outer and inner plastid envelope membranes Tproteins that are widely, if not universally, distributed among were isolated from barley chloroplasts and solubilized in buffer living organisms (1–3). In chloroplasts and other plastid types, containing 3% SDS. Each protein extract in turn was passed over two main types of Trxs (f and m) operate in disulfide/dithiol the Trx affinity column, and protein was bound to Trx with interchange and thereby act as powerful regulators of enzyme eluted with DTT. The outer plastid envelope fraction gave rise activity. Trxs catalyze two successive one-electron transfer re- to three main and a few minor bands (Fig. 1A, lane 2). Mass actions. Themselves, they are reduced via ferredoxin with elec- spectrometry identified the 55-kDa band as a protein related to trons provided by photosynthetic electron transport. Trx targets Erv1 and Erv2, enzymes implicated in disulfide bridge formation involve enzymes of the Calvin cycle and oxidative pentose in mitochondria and bacteria (16–19). The role of this protein needs to be established. For the inner plastid envelope protein phosphate cycle that often contain redox-sensitive C(x)nC pairs (n ϭ 2–8, or longer spacing between the two Cys residues; see fraction, five major bands were obtained, of which the 52-kDa ref. 1 for review). The mechanism requires the formation of a band was by far more abundant than the other bands (Fig. 1A, transient heterodisulfide bond between the reduced Trx and the lane 3). Several additional minor bands appeared that were not oxidized enzyme before complete reduction of the target disul- characterized further. Upon 2D separation of proteins in the 50- fide (4–7). Affinity purification procedures take advantage of to 52-kDa range, six spots of different isoelectric points and molecular masses were found (Fig. 1B). Microsequence analysis the fact that mutation in one of the redox-active cysteines in the identified three of the spots to be caused by TIC55, PTC52, and Trx active site (the one buried in the molecule) stabilizes the PAO (Fig. 1B, spots d-f). To confirm their role as potential Trx normally transient heterodisulfide, thereby covalently linking targets, the eluates were reduced by Trx, and the change in redox the target protein to Trx via a bond that can be cleaved by DTT state was measured by fluorescence coupled with gel ele- (e.g., refs. 4, 6, 8, and 9). ctrophoresis. Free SH groups were first blocked with Using this approach, we identified three inner plastid envelope N-ethylmaleimide, then the eluates were treated with the NADP/ membrane proteins from barley chloroplasts as novel Trx tar- gets: the protein translocon protein TIC55 (10), the precursor

NADPH:protochlorophyllide (Pchlide) oxidoreductase A Author contributions: C.R. designed research; S.B., J.M., K.S., F.Q., S.R., and C.R. performed (pPORA) translocon protein PTC52 (11), and pheophorbide a research; J.G., D.v.W., and S.R. analyzed data; and C.R. wrote the paper. oxygenase [PAO; originally named lethal leaf spot protein The authors declare no conflict of interest. (LLS1) (12, 13)]. All three proteins contain redox-active CxxC ¶To whom correspondence may be addressed. E-mail: [email protected] or christiane. motifs and interacted with Trxs f and m. By contrast, the related [email protected]. Rieske nonheme oxygenase family member chlorophyllide a- This article contains supporting information online at www.pnas.org/cgi/content/full/

oxygenase (CAO) (14), which lacks the CxxC motiv (15), was 0800378105/DC1. PLANT BIOLOGY found not to be a Trx target. Together, our findings reveal a role © 2008 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800378105 PNAS ͉ March 25, 2008 ͉ vol. 105 ͉ no. 12 ͉ 4933–4938 Downloaded by guest on October 2, 2021 Fig. 2. Expression of PTC52 and PAO. (A) Identification of PTC52 and PAO among the messenger products of 5-day-old dark-grown barley plants that had been exposed to light for 24 h, by hybrid-selected translation in the presence of [35S]methionine and immunoprecipitation using corresponding antisera and preimmune sera (PIS). (B) Expression of PTC52 and PAO in 5-day-old etiolated barley plants (lane a), 5-day-old dark-grown plants that had been exposed to light for 24 h (lane b), and 5-day-old light-grown plants (lane c), as revealed by RT- PCR, Northern hybridization, and Western blotting. As control, expression of actin was assayed.

PTC52, and PAO share CxxC motifs at Ϸ73 aa from their COOH-terminal ends (Fig. 1E).

Effect of Trx on PTC52 and PAO Activity. Given that no full-length cDNAs were available for PTC52 and PAO of barley, we used corresponding EST clones for barley PTC52 and a partial cDNA encoding lls1 of Zea mays for hybrid-selected translation and subsequent activity measurements. Fig. 2A shows that PTC52 and PAO could be detected in the mRNA of light-grown barley plants. The size of the detected precursors was slightly different and presumably reflected differences in the length of their transit peptides. Immunoprecipitations using monospecific (PTC52) (11) or monoclonal (PAO) (15) antibodies confirmed the iden- tity of the different bands. Northern blotting revealed differ- ences in the abundance of ptc52 and pao transcripts in dark- grown, light-exposed, and light-grown plants. Whereas ptc52 Fig. 1. Identification of outer and inner plastid envelope membrane pro- mRNA was most abundant in dark-grown seedlings, the pao teins interacting with Trx. (A) Identification of Trx targets in stromal extracts mRNA was present at similar levels in all samples (Fig. 2B). (lane 1) and solubilized outer (lane 2) and inner (lane 3) plastid envelope membranes of barley chloroplasts. The bands marked a, b, and c represent PTC52 is normally active in the pPORA translocon where it NADP-dependent malate dehydrogenase (85 kDa), phosphoribulokinase (80 operates as Pchlide a oxygenase (11). In contrast to this anabolic kDa), and sedoheptulose-1,7-bisphosphate bisphosphatase (75 kDa), respec- role of PTC52 in the C5 pathway, PAO acts catabolically in Chl tively. (B) Re-electrophoresis of the inner plastid envelope membrane proteins breakdown (13). The reaction catalyzed by PAO is coupled to of Ϸ52 kDa obtained as in A. Protein spots were subjected to mass spectrom- that of red Chl catabolite (RCC) reductase, leading to the loss etry and identified by sequence comparisons to be caused by TIC55 (d), PTC52 of pigment color during leaf senescence. In vitro, the interme- (e), and PAO (f). (C) Monobromobimane labeling of TIC55 (d), PTC52 (e), and diate of the coupled reaction, RCC, does not accumulate to PAO (f) before reduction with bacterial Trx. (D) Monobromobimane labeling substantial amounts such that pFCC1 is the only detectable of TIC55 (d), PTC52 (e), and PAO (f) after reduction with bacterial Trx. (E) product (13, 21). Presence of a conserved CxxC motif in TIC55, PTC52, and PAO/LLS1 and related proteins from Hordeum vulgare (H.v.), Zea mays (Z.m.), Oryza sativa (O.s.), To test whether the activity of PTC52 and PAO may be Triticum aestivum (T.a.), Arabidopsis thaliana (A.t.), and cyanobacteria (ORFs sensitive to regulation by Trx, activity measurements were slr1747, alr4354, alr5007). Biocomputational methods and sequence data performed with established procedures (13, 14, 22). PTC52 and sources can be found in SI Text. PAO of barley produced by hybrid-selected translation were incubated with Pchlide a and pheophorbide a. As controls, in vitro-translated and purified PTC52-(His)6 and PAO-(His)6 of Trx system to reduce disulfide bonds and newly formed SH Arabidopsis thaliana were used. After incubation, the assays were groups labeled with monobromobinane. Using nonequilibrium subjected to a step of gel filtration to remove unbound pigments. pH gradient electrophoresis and SDS/PAGE (20), we found that Protein–pigment complexes eluted with the flow-through were all three proteins (TIC55, PTC52, and PAO) showed, relative to supplemented with O2, Fd, and a Fd-reducing system comprising the untreated control, an increase in fluorescence and thus glucose-6-phosphate, NADPH, glucose-6-phosphate dehydroge- qualified as Trx targets (Fig. 1 D versus C). Aside from the nase, Fd:NADPH oxidoreductase, and wheat germ extract. In presence of conserved Rieske and 2Fe-2S clusters, TIC55, case of PAO, in vitro-expressed RCC reductase was added. Fig.

4934 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800378105 Bartsch et al. Downloaded by guest on October 2, 2021 Fig. 3. In vitro activity of PTC52 and PAO. (A) PTC52 was expressed as described in Fig. 2A, and activity measurements were performed with Pchlide a,O2, and an energy-regenerating system (see Materials and Methods). Pigments formed at time point zero (solid line) and after 15 min (dashed line) were extracted with acetone and subjected to RP-HPLC. Pigment detection was made by absorbance measurements at 455 nm, and the identity of the different peaks was confirmed by absorption spectrometry (Inset) and mass spectrometry (data not shown). (Inset) Shown is the spectrum of the peak that eluted at 12.5 min, which identified this compound as Pchlide b. As a control, Fig. 4. Verification of TIC55, PTC52, and PAO as Trx targets. TIC55- (lanes a), in vitro-expressed PAO was used that failed to bind Pchlide a (stippled line). (B) PTC52- (lanes b), and PAO-(His)6 (lanes c) mutant proteins containing engi- The same as A, but showing activity measurements for PAO (F455)inthe neered CxxC motifs with Cys-Ala exchanges replacing the first Cys residue presence of pheophorbide a,O2, and the same energy-regenerating system as were produced by coupled in vitro transcription/translation in wheat germ described before, at time point zero (solid line) and after 15 min (dashed line). lysates, purified and used for Trx affinity chromatography. (A) Western blot of Controls with PTC52 showed a lack of pheophorbide a binding (A455) and thus Trx-related proteins bound to immobilized TIC55-, PTC52-, and PAO-(His)6 in proved the specificity of the enzyme assay (stippled line). wheat germ extracts. (B) Western blot of Trx-related proteins bound to immobilized TIC55-, PTC52-, and PAO-(His)6 in stromal extract of barley chlo- roplasts. After treatment with DTT and extensive washing of the same affinity 3A shows representative HPLC tracings of pigments formed in matrices used in A, stromal extract of isolated barley chloroplasts was passed the presence of PTC52 at time 0 (solid line) and after 15 min over the columns. Binding reactions were carried out either in the absence of competitor (Ϫcomp.) or in the presence of competitors mimicking the thiol- (dashed line). Pchlide a has a retention time of 15 min at the C18 reactive sites of Trx m and Trx f. RP material, whereas Pchlide b elutes at 12.5 min. Because the absorption coefficients of Pchlide a and Pchlide b are 5-fold different at 455 nm, the measured decrease in Pchlide a levels tion/reduction reactions involving redox-active protein SH after incubation correlated with an apparent large, 5-fold in- groups that in turn affected their activities, the reduced form crease in the amount of Pchlide b (Fig. 3A). Pchlide a to Pchlide being more active than the oxidized form. b conversion required O2 and a Fd-reducing system. In the TIC55, PTC52, and PAO are localized in the inner envelope absence of these factors, no Pchlide b was produced (data not membrane of chloroplasts (10, 11, 23, 24), and all contain the shown). When wheat germ lysate was omitted, low PTC52 CxxC motif that could have interacted with Trx. To confirm this activity was measurable, suggesting the presence of a factor that hypothesis, hexa-histidine-tagged versions of Arabidopsis TIC55, promoted the folding and assembly of the Rieske and 2Fe-2S PTC52, and PAO were produced containing an Ala residue in clusters. In the absence of PTC52 protein, no Pchlide b was place of the first or second Cys residue in the CxxC motif. When formed (data not shown). in vitro-translated and Ni-NTA-purified proteins were used as No binding of Pchlide a was detectable with the PAO protein, bait, small amounts of a low molecular mass, Ϸ12-kDa wheat and consequently no Pchlide b was produed (Fig. 3A, stippled germ protein were recovered for all three proteins (Fig. 4A). line). Conversely, no binding of pheophorbide a was achieved for DTT addition eluted this band. Protein mass spectrometry and PTC52 under conditions that led to the production of pFCC-1 Western blotting identified this band to be caused by wheat germ (Fig. 3B). pFCC-1 formation by PAO required RCCR, Fd, a Trx h (data not shown). Upon testing stromal extract of barley chloroplast for binding, two protein bands were found for TIC55, Fd-reducing system, and the substrates pheophorbide a and O2. We next tested the effect of stromal Trx f and Trx m on PTC52 PTC52, and PAO representing Trx f (Fig. 4B, top band, band I) and PAO activity. Either Trx was expressed in Escherichia coli, and Trx m (Fig. 4B, bottom band, band II). Consistent with this purified, and added to the in vitro-expressed proteins. Then the view, band II was sensitive to excess peptide mimicking the ϩ activity measurements were repeated as before. In a parallel presence of the redox-active sites of Trx m (Fig. 4B, Trx m), whereas band I was blocked by competition by Trx f-specific sample, the effect of DTT was studied. Table 1 shows that ϩ addition of Trx m or addition of DTT led to 12- and 15-fold peptide probe (Fig. 4B, Trx f). These results showed that increases in PTC52 and PAO activity, respectively. These data TIC55, PTC52, and PAO had no preference with regard to either Trx m or Trx f in the in vitro assay. Similar conclusions have been suggested that PTC52 and PAO may undergo reversible oxida- drawn from previous studies (8).

Table 1. The effects of Trx m and DTT on PTC52 and PAO Role of Trxs in the Regulation of Protein Import. It has been activities proposed that Trx could play a role in oxidative regulation (3). Taking into account the findings reported thus far, we assumed ⅐ Ϫ1 Enzyme activity, nkat mg protein that Trx could regulate the activity of TIC55, PTC52, and PAO Protein ϪTrx m ϩTrx m ϪDTT ϩDTT in response to oxidative stress. A way to test this hypothesis was provided by previous work on mutants that are de-regulated in PTC52 0.68 8.20 0.70 8.40 biosynthesis such as the fluorescent (FLU) mutant of PAO 1.04 15.60 1.08 16.20 Arabidopsis (25, 26) and its ortholog in barley, tigrina d12 (27), PTC52 and PAO were tested for their Pchlide a-oxygenase and pheophor- and other species. FLU and tigrina d12 seedlings overproduce

bide a-oxygenase activity, respectively, as described in Fig. 3. In two parallel free, nonprotein-bound protochlorophyllide (Pchlide) in dark- PLANT BIOLOGY sets of samples, the effects of isolated Trx m and DTT were examined. ness. Upon illumination, Pchlide operates as photosensitizer and

Bartsch et al. PNAS ͉ March 25, 2008 ͉ vol. 105 ͉ no. 12 ͉ 4935 Downloaded by guest on October 2, 2021 Fig. 5. In vitro protein import to monitor dark/light and oxidative regulation of TIC55, PTC52, and CAO activity. (A) 35S-labeled precursors were synthesized by coupled in vitro transcription/translation of cDNA clones and imported into chloroplasts of 5-day-old, light-grown tigrina d12 mutant plants of barley in a standard import reaction containing 2.5 mM Mg-ATP. (B and C) In two additional sets of incubation, chloroplasts were used from plants that had been shifted to a 14-h dark period before plastid isolation (B) or to a nonpermissive 10-h dark to 4-h light shift (C). To study substrate-dependent import of pPORA, aliquots of the plastid suspension were pretreated with 5-aminolevulinic acid (ϩ 5-ALA) or phosphate buffer (Ϫ5-ALA) and Mg-ATP, giving rise to intraplastidic Pchlide synthesis. Parallel assays were conducted that were treated with or without thermolysin after import. Levels of precursor and mature proteins were determined by SDS/PAGE and autoradiography. Percentages refer to the sum of precursor (light-gray columns) and mature proteins (dark-gray columns) in the assays.

gives rise to singlet oxygen (28). In seedlings kept under con- processed, mature form after import into chloroplasts from tinuous white light, Pchlide is constantly converted to chloro- darkened plants (Fig. 5 and data not shown). In the case of phyllide by the NADPH:Pchlide oxidoreductase such that no pPORA, higher import rates were measured for chloroplasts singlet oxygen is produced. Mature plants cultivated under isolated from darkened as compared with light-grown plants. alternative dark/light cycles or subjected to nonpermissive light- Chloroplasts from light-grown plants imported the precursor to-dark-to-light shifts suffer from the same, pigment-sensitized only in the presence of the Pchlide precursor 5-aminolevulinic photooxidative response as etiolated plants after illumination acid, a result that is consistent with our previous findings using (25), allowing oxidative regulation of TIC55 and PTC52 activity in vitro-translated precursor (31) but is at variance with the to be tested by in planta and in vitro experiments. As a measure results of Kim and Apel (32) using in planta assays. This issue for this regulation, protein import was explored given that needs to be resolved in subsequent work. With chloroplasts from TIC55, PTC52, and the related Rieske protein family member dark-shifted plants, the endogenously produced Pchlide substi- CAO (and perhaps also PAO) have established roles in protein tuted for the exogenous 5-ALA-derived Pchlide. Upon the translocation across the inner plastid envelope membranes (10, nonpermissive dark-to-light shift, no import of pPORA oc- 11, 24, 29). curred. For pLHCII, the CAO import substrate, a different Tigrina d12 seeds of barley were germinated in continuous result was obtained. Whereas chloroplasts from light-grown white light for 5 days. Chloroplasts were isolated from primary plants imported more pLHCII than chloroplasts from darkened leaves of such plants and tested for their in vitro-import capability plants, no deleterious effect on import was apparent after the for the small subunit precursor of ribulose-1,5-bisphosphate nonpermissive dark-to-light shift (Fig. 5). carboxylase/oxygenase (pSSU; a TIC55 import substrate; refs. 10 and 24), pPORA (a PTC52 import substrate; ref. 11), and the Discussion precursor to the major light-harvesting Chl a/b-binding protein In the present study, a family of inner plastid envelope Trx targets of photosystem II (PSII) (pLHCII, a CAO import substrate; ref. was discovered. This family comprises three proteins sharing 29). As controls, the precursors to the photosynthetic ferredoxin conserved Rieske and mononuclear iron binding domains and (pFdI) and nonphotosynthetic ferredoxin (pFdIII) were used. As CxxC motifs (TIC55, PTC52, and PAO). Similar Rieske and shown previously, pFdIII is mis-sorted to the intermembrane mononuclear iron binding motifs are present in choline mono- space of chloroplasts in white light but it is imported into the oxygenase CMO and CAO, but the latter lack the redox-active stroma and processed in darkness (30). In contrast, pFdI is CxxC motif present in TIC55, PTC52, and PAO and were not faithfully taken up under either condition (30). To see whether purified by the established affinity purification procedure. light-dark regulation and/or oxidative control would happen to At least two of the inner plastid envelope Trx targets, , namely the TIC55-, PTC52-, and CAO-containing import machineries, TIC55 and PTC52, have roles in protein import. TIC55 is part of one pot of the originally light-grown tigrina d12 seedlings was the general protein import machinery that is responsible for the transferred to darkness for 14 h, whereas another pot was shifted transport of pSSU and perhaps other photosynthetic precursors to a nonpermissive 10 h dark/4 h light shift. into the chloroplast (10, 24). PTC52 establishes a distinctive trans- Fig. 5A summarizes the results of three independent import locon. This translocon is most abundantly expressed in etiolated experiments. They showed that pSSU import, as measured as the plants and is involved in the Pchlide-dependent import of pPORA amount of mature protein inside thermolysin-treated, repurified, (11). Despite a light-induced decline in the amount of OEP16 (33) intact chloroplasts, was higher for light-grown than for darkened and PTC52 (this study) the PTC complex remains active in chlo- plants. In plants subjected to the nonpermissive dark-to-light roplasts. The observed dark/light regulation could ensure that shift, no import was detectable. Similar results were obtained for maximum amounts of pPORA are imported via the PTC complex pFdI, the photosynthetic ferredoxin. Consistent with previous at the end of the night where Pchlide reaccumulates (34). The reports, its nonphotosynthetic counterpart was undetectable as expression of TIC55 may be constitutive, but no data have been processed enzyme in import assays containing chloroplasts from published regarding the role of TIC55 at different stages of light-grown plants but accumulated in a thermolysin-resistant, development to our knowledge. Both TIC55 and PTC52 were

4936 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800378105 Bartsch et al. Downloaded by guest on October 2, 2021 sensitive to oxidative control and were inactivated in reilluminated bridges, either established intramolecularly or intermolecularly. tigrina d12 plants, presumably by a porphyrin (Pchlide)-sensitized Especially the risk of singlet oxygen formation by virtue of mechanism. The same holds true for PAO, the third identified Trx photoreactive , such as Chl, Pchlide and pheophor- target, which is senescence-induced with low basal protein and bide a, that accumulate at different stages of plant development, mRNA levels in presenescent plants (13, 35, 36). The observed large may require such mechanism. In fact, plastids need to adapt their Ϸ60-fold increase in PAO activity, which is at variance with the only metabolism to different requirements at the beginning of green- Ϸ6-fold raise in pao mRNA level in senescent plants (13) suggests ing, in light-adapted plants, and during leaf senescence. Accord- a posttranscriptional mode of control that could involve redox ing to recent work (42), even amyloplasts, the yellow, starch- control of enzyme activity. It is tempting to hypothesize that PAO, accumulating type of plastids, contain a full set of Trxs and in addition to its role in Chl breakdown, may be involved in the Trx:ferredoxin reductases, highlighting the ubiquitous role of the regulated import of senescence-induced gene products into the Trx system for plastid metabolism and plant development. chloroplast, although hard biochemical evidence for the existence of a PAO-containing import apparatus is still lacking. Materials and Methods CAO, which lacks the CxxC motif and was not a Trx target, Identification of Proteins Interacting with Trx f and Trx m. Chloroplasts were establishes a third TIC subcomplex (29). This complex was isolated from 5-day-old light-grown barley (Hordeum vulgare L. Carina) and involved in the regulated import of the Chl b-containing light- fractionated into stroma and membranes (11). The membranes were further harvesting proteins LHCII and CP29 (29). The in vitro import separated into outer and inner plastid envelope membranes (11), and protein experiments shown here revealed that pLHCII import was of each of these fractions was solubilized with 3% SDS. The resulting protein subject to dark/light regulation but did not respond to oxidative extracts in turn were subjected to Trx m affinity column by using wild-type Trx f and Trx m and their mutant f C49S and m C40A derivatives immobilized on control elicited by free, photoexcited Pchlide molecules accu- Sepharose 4B (Amersham Pharmacia Biosciences) (8). Bound outer and inner mulating in tigrina d12 plants. This result may be explained by the envelope membrane proteins were subsequently eluted with 10 mM DTT, lack of the CxxC motif that could be oxidized by singlet oxygen. dialyzed, concentrated, and subjected to SDS/PAGE. The amount of protein The rate of LHCII import may be controlled in a different way. collected on addition of DTT to the mutated Trx f and Trx m columns was Yamasato et al. (37) identified a pigment sensor domain in the similar (Ϸ120–150 ␮g, depending in the actual experiment). Similar binding intermembrane part of CAO that could operate as relay in and elution steps were performed with activated thiol Sepharose columns providing Chlide b to incoming pLHCII and pCP29 apoproteins. that, however, yielded an extremely low amount of protein (Ϸ5–10 ␮g). Only when Chlide b is produced would pLHCII and pCP29 Peptides mimicking the thiol-active sites of Trx f and Trx m (in bold) used for 26 competitive binding assays were as follows: Trx f,NH2- WPIVKAAGDKPVV- import occur. 57 18 The Chl b-containing light-harvesting proteins have important LDMFTQWCGPCKAMAPKYE -COOH and Trx m, NH2- KEFVLESEVPVMVDF- WAPWCGPCKLIAPVIDE49-COOH. functions not only in light trapping and energy transduction but also in photoprotection (38). It is attractive to hypothesize that Confirmation of Trx Targets. After elution from the Trx affinity columns with oxidative stress may signal a requirement for more LHC proteins DTT, excess DTT was removed by extensive dialysis against 50 mM Tris⅐HCl, pH and, therefore, no block but even an elevation of pLHCII and 7.5. Subsequently, the samples were treated with 1 mM N-ethylmaleimide to pCP29 import could be observed in reilluminated tigrina d12 block free SH groups. Excess N-ethylmaleimide was removed by addition of 1 plants. Once imported and processed, the mature LHCII nor- mM 2-mercaptopethanol. Half of the final samples was reduced with bacterial mally associates with PSII but, during state transition, its location (N-ethylmaleimide) or stromal (barley) Trx, whereas the remaining halves changes toward PSI (39). There is some evidence that the were incubated with water. The reduced and the control samples in turn were relocation of LHCII during state transition may be regulated by treated with monobromobimane to label newly formed SH groups (43, 44). a PSI thiol-based mechanism that could involve Trx, in addition to the one depending on plastoquinone and protein phosphor- 2D Gel Separation. Isoelectric focusing and SDS/PAGE were performed according to Scharf and Nover (45) using a pH range from 3 to 10 and 10–20% polyacryl- ylation (40). It is attractive to hypothesize that perhaps a need to amide gradients containing SDS. To separate inner envelope proteins, nonequi- tie Trx action to state transition may have led to a loss of Trx librium pH gradient electrophoresis was used (45). Proteins were stained with responsiveness of the CAO-containing TIC import machinery. Coomassie brilliant blue R-250 or subjected to Western blotting.

Conclusions Protein Sequencing. Individual protein spots were excised and in-gel digested The results of the present study strongly suggest two major with trypsin, and the resulting peptide mixtures were subjected to protein mechanisms of Trx-dependent control of protein import. The mass spectrometry. The obtained partial amino acid sequences were aligned first mechanism is dark/light regulation that could help adjust the with other proteins using the SEQUEST algorithm and the nonredundant amounts of freshly imported proteins to the photosynthetic protein database from the National Center for Biotechnology Information (8) needs. Examples for this type of control include pSSU and pFdI and other standard computer programs [ref. 15; see supporting information that were imported into chloroplasts with higher rates in the light (SI) Text]. than in the dark. By contrast, pPORA was less efficiently Activity Measurements of PTC52 and PAO. Activity measurements were carried imported into chloroplasts from light-grown versus darkened out according to Oster et al. (14) and Pruzinska et al. (13), respectively. Final plants. This result may be explained by the fact that pPORA- 50 ␮l-assays contained the following supplements: 2 mM Pchlide a or pheo- bound Pchlide b serves as light trap (41) and could impede phorbide a,10␮g of Fd (Sigma) and a Fd-reducing system [2 mM glucose-6- normal light harvesting during photosynthesis. For pFdIII, a phosphate; 1 mM NADPH; 50 milliunits of glucose-6-phosphate dehydroge- different type of regulation was observed. pFdIII was mis-sorted nase; 10 milliunits of Fd-NADPH-oxidoreductase (Sigma)]. Assays for testing to the inter envelope membrane space of chloroplasts of light- PAO activity likewise contained RCC reductase, allowing conversion of pheo- grown plants but was faithfully imported into the stroma and phorbide a to be measured by the appearance of pFCC-1 (13). As a source of processed in chloroplasts of darkened plants. Although much RCC reductase, the Arabidopsis enzyme was used that had been expressed less is known about the potential physiological background of from a respective cDNA in E. coli as described (13). Separation and identifica- this control, Hirohashi et al. (30) proposed that the limitation of tion of pigments by RP-HPLC has been described (13, 22). The production of pFCC-1 from pheophorbide a was linear for Ͼ1 h of incubation at 22°C and pFdIII import may be a means to prevent potential interferences followed Michaelis-Menten kinetics with an apparent Km value of 6 ␮M, with normal electron transport involving FdI. consistent with previous results (13). The second type of control in which Trx may be involved is oxidative control. Chloroplasts are a source of reactive oxygen Protein Import. Import assays were carried out by using established procedures 35 species that very easily and irreversibly inactivate proteins and and S-labeled, wheat germ-translated precursors (11). Plastids were isolated PLANT BIOLOGY enzyme activities by modifying the reduction state of disulfide from tigrina d12 plants that had been grown in continuous white light for 5 days

Bartsch et al. PNAS ͉ March 25, 2008 ͉ vol. 105 ͉ no. 12 ͉ 4937 Downloaded by guest on October 2, 2021 or subjected to prior 14 h dark or 10-h dark/4-h white light shifts. Protein ACKNOWLEDGMENTS. This is scientific paper no. 0201-08 from the College of detection and quantification was made by SDS/PAGE and autoradiography or Agricultural, Human, and Natural Resource Sciences of Washington State Uni- with a PhosphorImager. versity. We thank D. J. Schnell (University of Massachusetts, Amherst), F. Kessler (Universite´de Neuchaˆtel, Neuchaˆtel, Switzerland), and J. Soll (University of Mu- nich, Munich, Germany) for gifts of cDNA clones and antisera and B. Buchanan Miscellaneous. RNA preparation, Northern hybridization, and hybrid-selected (University of California, Berkeley) for critical reading of the manuscript, support, translation using filter-bound pao- and ptc52-specific DNA probes were car- and advice with the experiments. This study was supported by a Chaire ried out as described (46). d’Excellence program of the French Ministry of Research (C.R.).

1. Schu¨rmann P, Jacquot J-P (2000) Plant thioredoxin systems revisited. Annu Rev Plant 25. Meskauskiene R, et al. (2001) FLU: A negative regulator of chlorophyll biosynthesis in Physiol Plant Mol Biol 51:371–400. Arabidopsis thaliana. Proc Natl Acad Sci USA 98:12826–12831. 2. Buchanan BB, Schu¨rmann P, Wolosiuk RA, Jacquot J-P (2002) The thioredoxin/ 26. Meskauskiene R, Apel K (2002) TIGRINA d12, required for regulating the biosynthesis ferredoxin system: From discovery to molecular structures and beyond. Photosynth Res of tetrapyrroles in barley, is an ortholog of the FLU gene of Arabidopsis thaliana. FEBS 73:215–222. Lett 553:119–124. 3. Buchanan BB, Balmer Y (2005) Redox regulation: A broadening horizon. Annu Rev 27. von Wettstein D, Kahn A, Nielsen OF, Gough S (1974) Genetic regulation of chlorophyll Plant Biol 56:187–220. synthesis analyzed with mutants in barley. Science 184:800–802. 4. Brandes HK, Larimer FW, Hartman FC (1996) The molecular pathway for the regulation 28. op den Camp R, et al. (2004) Rapid induction of distinct stress responses after the of phosphoribulokinase by thioredoxin f. J Biol Chem 271:3333–3335. release of singlet oxygen in Arabidopsis. Plant Cell 15:2320–2332. 5. Verdoucq L, Vignols F, Jacquot, J-P, Chartier Y, Meyer Y (1999) In vivo characterization 29. Reinbothe C, et al. (2006) A role for chlorophyllide a-oxygenase in the regulated import of a thioredoxin h target protein defines a new peroxiredoxin family. J Biol Chem and stabilization of light-harvesting chlorophyll a/b proteins in chloroplasts. Proc Natl 274:19714–19722. Acad Sci USA 103:4777–4782. 6. Balmer Y, Schu¨rmann P (2001) Heterodimer formation between thioredoxin f and 30. Hirohashi T, Hase T, Nakai M (2001) Maize nonphotosynthetic ferredoxin precursor is fructode-1,6-bisphosphatase from spinach chloroplasts. FEBS Lett 492:58–61. missorted to the intermembrane space of chloroplasts in the presence of light. Plant 7. Goyer A, et al. (2002) The internal Cys207 of Sorghum leaf NADP-malate dehydroge- Physiol 125:2154–2163. nase can form mixed disulfides with thioredoxin. FEBS Lett 444:165–169. 31. Reinbothe S, Mache R, Reinbothe C (2000) A second, substrate-dependent site of 8. Balmer Y, et al. (2003) Proteomics gives insights into the regulatory function of protein import into chloroplasts. Proc Natl Acad Sci USA 97:9795–9800. chloroplast thioredoxins. Proc Natl Acad Sci USA 100:370–375. 32. Kim C, Apel K (2004) Substrate-dependent and organ-specific chloroplast protein 9. Hisabori T, et al. (2003) Thioredoxin affinity chromatography: A useful method for import in planta. Plant Cell 16:88–98. further understanding the thioredoxin network. J Exp Bot 56:1463–1468. 33. Reinbothe S, Quigley F, Springer A, Schemenewitz A, Reinbothe C (2004) The outer 10. Caliebe A, et al. (1997) The chloroplastic protein import machinery contains a Rieske- plastid envelope protein Oep16: Role as precursor translocase in import of protochlo- type iron-sulfur cluster and mononuclear iron-binding protein. EMBO J 16:7342–7350. rophyllide oxidoreductase A. Proc Natl Acad Sci USA 101:2203–2208. 11. Reinbothe S, Quigley F, Gray J, Schemenewitz A, Reinbothe C (2004) Identification of 34. Griffiths WT, Kay SA, Oliver RP (1985) The presence of photoregulation of protochlo- plastid envelope proteins required for import of protochlorophyllide oxidoreductase rophyllide reductase in green tissues. Plant Mol Biol 4:13–22. A into the chloroplast of barley. Proc Natl Acad Sci USA 101:2197–2202. 35. Ginsburg S, Schellenberg M, Matile P (1994) Cleavage of chlorophyll-porphyrin 12. Gray J, Close PS, Briggs SP, Johal GS (1997) A novel suppressor of cell death in plants (requirement for reduced ferredoxin and oxygen). Plant Physiol 105:545–554. encoded by the LLs1 gene of maize. Cell 89:25–31. 36. Ho¨rtensteiner S, Vicentini F, Matile P (1995) Chlorophyll breakdown in senescent 13. Pruzinska A, Tanner G, Anders I, Boca M, Ho¨rtensteiner S (2003) Chlorophyll break- cotyledons of rape, Brassica napus L.: Enzymatic cleavage of phaeophorbide a in vitro. down: Pheophorbide a oxygenase is a Rieske-type iron–sulfur protein, encoded by the New Phytol 129:237–246. accelerated cell death 1 gene. Proc Natl Acad Sci USA 100:15259–15264. 37. Yamasato A, Nagata N, Tanaka R, Tanaka A (2005) The N-terminal domain of chloro- 14. Oster U, Tanaka R, Tanaka A, Ru¨diger W (2000) Cloning and functional expression of phyllide a oxygenase confers protein instability in response to chlorophyll b accumu- the gene encoding the key enzyme for chlorophyll b biosynthesis (CAO) from Arabi- lation in Arabidopsis. Plant Cell 17:1585–1597. dopsis thaliana. Plant J 21:305–310. 38. Ku¨hlbrandt W, Wang DN, Fujiyoshi Y (1994) Atomic model of plant light-harvesting 15. Gray J, et al. (2004) A small family of LLS1-related nonheme oxygenases in plants with an origin amongst oxygenic photosynthesizers. Plant Mol Biol 54:39–54. complex by electron crystallography. Nature 367:614–621. 16. Lee J, Hofhaus G, Lisowsky T (2000) Erv1p from Saccharomyces cerevisiae is a FAD- 39. Allen JF (1992) Protein phosphorylation in regulation of photosynthesis. Biochim linked sulfhydryl oxidase. FEBS Lett 477:62–66. Biophys Acta 1098:275–335. 17. Gerber J, Muhlkenhoff U, Hofhaus G, Lill R, Lisowsky T (2001) Yeast ERV2p is the first 40. Aro EM, Ohad I (2003) Redox regulation of thylakoid protein phosphorylation. Anti- microsomal FAD-linked sulhydryl oxidase of the Erv1p/Alrp protein family. J Biol Chem oxid Redox Signal 5:55–67. 276:23486–23491. 41. Reinbothe C, Lebedev N, Reinbothe S (1999) A protochlorophyllide light-harvesting 18. Mesecke N, et al. (2005) A disulfide relay system in the intermembrane space of complex involved in de-etiolation of higher plants. Nature 397:80–84. mitochondria that mediates protein import. Cell 121:1059–1069. 42. Balmer Y, et al. (2006) A complete ferredoxin/thioredoxin system regulates fundamen- 19. Tokatlidis K (2005) A disulfide relay system in mitochondria. Cell 121:965–967. tal processes in amyloplasts. Proc Natl Acad Sci USA 103:298–2993. 20. Yano H, Wong JH, Lee YM, Cho M-J, Buchanan BB (2001) A strategy for the identifi- 43. Besse I, Wong JH, Kobrehel K, Buchanan BB (1996) Thiocalsin: A thioredoxin-linked, cation of proteins targeted by thioredoxin. Proc Natl Acad Sci USA 98:4794–4799. substrate-specific protease dependent on calcium. Proc Natl Acad Sci USA 93:3169– 21. Rodoni S, et al. (1997) Chlorophyll breakdown in senescent chloroplasts: Cleavage of 3175. pheophorbide a in two enzymic steps. Plant Physiol 115:669–676. 44. Balmer Y, et al. (2004) Thioredoxin links redox to the regulation of fundamental 22. Reinbothe S, Pollmann S, Reinbothe C (2003) In situ conversion of protochlorophyllide processes of plant mitochondria. Proc Natl Acad Sci USA 101:2642–2647. b to protochlorophyllide a in barley.Evidence for a novel role of 7-formyl reductase in 45. Scharf K-D, Nover L (1982) Heat shock-induced alterations of ribosomal protein phos- the prolamellar body of etioplasts. J Biol Chem 278:800–806. phorylation in plant cell cultures. Cell 30:427–437. 23. Matile P, Schellenberg M (1994) The cleavage of phaeophorbide a is located in the 46. Reinbothe C, Ortel B, Parthier B, Reinbothe S (1994) Cytosolic and plastid forms of envelope of barley gerontoplasts. Plant Physiol Biochem 34:55–59. 5-enolpyruvylshikimate-3-phosphate synthase in Euglena gracilis are differentially 24. Ku¨chler M, Decker S, Ho¨rmann F, Soll J, Heins L (2003) Protein import into chloroplasts expressed during light-induced chloroplast development. Mol Gen Genet 245:616– involves redox-regulated proteins. EMBO J 21:6136–6145. 622.

4938 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800378105 Bartsch et al. Downloaded by guest on October 2, 2021