Bioinformatic tools

TargetP 1.1 server (www.cbs.dtu.dk/services/TargetP) was used to predict the presence of the chloroplast transit peptides (cTP) in the protein sequences and the location of potential cTP cleavage sites. TargetP has a substantially greater ability to discriminate between signal peptides and uncleaved signal anchors [23].

Kyte–Doolittle hydropathy plots [24] were generated using an online tool available at the ExPASy molecular biology server (https://web.expasy.org/protscale/). For in-silico multiple sequence alignments, the MAFFT version 7-alignment software (https://mafft.cbrc.jp/alignment/server/) was used. [25]. PredSL (http://aias.biol.uoa.gr/PredSL/) was used to predict a protein’s localization [26]. The PSORT Prediction tool (http://psort1.hgc.jp/form.html) that uses the overall amino acid composition, the N-terminal targeting sequence information, and the motifs, was used as a hybrid approach. This latter software uses a set of knowledge-based “if-then” rules [27].

Isolation of thylakoid membranes and stroma fractions from intact chloroplasts

Pea seeds (Pisum sativum, cv Calvedon) were grown hydroponically and intact chloroplast from leaves was isolated as described previously [28]. Washed thylakoids and stromal extract was prepared from isolated chloroplasts as described previously [29]. Total chlorophyll (CHL) content was measured as described previously [30]. Prior to the experiments, reconstituted lysates were prepared by suspending the thylakoid membranes in the freshly prepared stroma.

Preparation and analysis of free and bound ribosomes

Free ribosomes were prepared as described [22] and thylakoid-bound ribosomes were detached as described previously [31].

Cell free transcription-translation insertion assays

The recombinant plasmid pET25b-SUIV containing the gene PetD was used as a template for the creation of plasmid pT7CFE1-SUIV (Additional file 1: Figure S1). The pT7CFE1-CHis based expression construct, pT7CFE1-SUIV, was used to express the PetD gene. PetD was cloned between the MscI and BamHI restriction sites of the PT7CFE1-CHis vector (Thermo Fisher Scientific). To eliminate the need for affinity tags (His tag), a double stop codon was added at the 3′ end of the PetD sequence. The pT7CFE1-SUIV expression vector utilizes the T7 viral promoter and an encephalomyocarditis virus (EMCV) internal ribosome entry site (IRES) that is critical for high levels of cap-independent protein expression. Moreover, this expression vector possesses a T7 RNA polymerase promoter that functions with the T7 RNA polymerase that is included in the transcription reaction mixture. The obtained pT7CFE1-SUIV construct was confirmed to be correct by sequence analysis. The plasmid pT7CFE1-b ₆ is described in [11].

Coupled transcription/translation reactions of the pea PetD were performed as described in [11] with the following modifications: the PetD gene was firstly transcribed using pT7CFE1-SUIV construct as a template and the 2.0 μg of DNA-purified mRNA was used in the translation process. PetD mRNA was then translated in vitro in the presence or absence of thylakoid membranes, with or without the freshly prepared stromal fraction for 60 min at 30 °C in the presence of 10 μCi [³⁵S]-methionine (Hartmann Analytic, Germany). Following the translation reaction, another 2.0 μg of the DNA-purified mRNA and 1 μL of manufacture “energy mix” were added for each reaction and incubated for another 30 min in order to increase the protein yield. The translation in vitro assays that were required for the post-translational insertion of PetD were performed accordingly to the method of Houben, et al. [32] with the following modifications: after 60 min of translation in the presence of 10 μCi [³⁵S]-methionine, puromycin and lincomycin were added to a final concentration of 5 mM and 15 mM, respectively, and the samples were incubated for another 30 min at 30 °C. Subsequently, membranes and protease inhibitors were added, and the samples were kept for 30 min at 30 °C. The membranes were separated from the translation mix and then purified as described previously [11]. Finally, in order to remove the endogenous RNA as well as the external domains of the endogenous proteins the translation reactions, the thylakoid membranes or stroma were pre-incubated as described in [11].

Assessment of integration of PetD into the thylakoid membranes

Assay mixtures containing 45 μL thylakoid suspension (250 μg CHL /mL) in HM buffer and 25 μL of translation mixture were incubated for 20 min at 25 °C. Following double washing of the thylakoids suspensions, the samples were analyzed. Where appropriate, incubations were pretreated with 1 unit of apyrase or sodium azide (10 mM) or nigericin (2 μM) plus KCl (10 mM).

To monitor the integration of the targeted proteins, membranes (200 μg CHL /mL) were resuspended in 2 M NaBr or 2 M NaSCN in 50 mM HEPES-KOH pH 7.7 buffer containing a protease inhibitor cocktail [33] and incubated on ice for 30 min. Next, the membranes were washed and collected (20,000×g for 10 min). Although the resulting membranes were solubilized in SDS sample buffer directly (pellet fraction), the supernatant fraction was precipitated in 80% acetone and then analysed by SDS–PAGE. The proteolytic assessment of protein import into the thylakoid membranes was performed as described in [34] with modifications made by Kroliczewski, et al. [11].

Integration of the single span subunit W of PS II (PsbW)

In order to validate PetD insertion assays, we followed the thylakoid membrane integration of PsbW using the methods described in [11]. PsbW protein inserts into the thylakoid membrane by the spontaneous pathway [11].

Crosslinking and co-immunoprecipitation of the ribosomal fractions

The ribosome-nascent chain complexes (RNCs) were cross-linked on ice using the homobifunctional N-hydroxysuccinade ester bis[2-(succinimidyloxycarbonyloxy) ethyl] sulfone (BSOCOES, Pierce), a membrane-permeable cleavable crosslinker or the heterobifunctional cleavable (by DTT) crosslinker N-succinimidyl-3-[2-pyridyldithio]propionate (SPDP, Pierce) by the method of Kroliczewski, et al. [11] followed by co-immunoprecipitation reactions using an antibody (Ab) against ALB3 or cpSRP54 proteins [11, 28]. To perform the immunoprecipitation reaction, we used 2 μg of antibody. The immunoprecipitated proteins were analysed by SDS-PAGE and Western blot or subjected to autoradiography using X-ray cassettes with a sheet of X-ray film (Kodak) [35]. The gels were dried prior to being placed on X-ray film. Furthermore, the detected bands were excised from a polyacrylamide gel and analysed by Q-TOF MS/MS spectrometry (MS) with peptide mass fingerprints (PMF), (see: Mass spectroscopy analysis section).

PetD and PetB co-expression in cell-free assays

The cell-free expression assays were performed for 60 min to ensure synchronised expression of PetD and PetB [11] in the presence of stroma and membrane fractions followed by crosslinking with 25 mM BSOCOES or 40 mM SPDP. BSOCOES was added to the protein samples to achieve a concentration equal to 10–50 times the molar concentration of the proteins in solution. The reaction mixtures were incubated on ice for 2 h and then in Quenching Solution (1 M Tris∙HCl, pH 7.5) to a final concentration of 20–50 mM. The samples were then incubated for 15 min to quench non-reacted crosslinker and byproducts. SPDP was added to the protein samples to achieve a 4-fold molar excess of SPDP and was incubated for 60 min at room temperature. Excess of non-reacted crosslinker and reaction byproducts were removed by extensive dialysis.

The isolated membranes were then solubilized in HMS buffer containing 1% of DDM. Immunoprecipitation was performed with an antibody against PetB as described in [28]. Afterwards, base-cleavage of BSOCOES was performed by increasing the pH of the solution containing the BSOCOES conjugate to 11.6 with NaOH, which was incubated for 2 h at 37 °C, but cross-links created using SPDP reagents can be cleaved with 25 mM DTT at pH 4.5 Obtained samples were dialyzed into HMS buffer for subsequent analysis. The proteins were then identified with MS.

Preparation of stroma fractions without the cpSRP54 protein

Protein A Sepharose CL-4B was prepared according to the standard protocol of GE Healthcare. An antibody against cpSRP54 (5 μg mL⁻¹) was incubated 30 min at room temperature with 2 ml of isolated stroma fraction. Subsequently, 20 μL of protein A-Sepharose CL-4B resin in 20 mM sodium phosphate buffer, pH 7.0 was added and the incubation was continued with very mild rotation at 4 °C for an additional 90 min. The resin was collected by centrifugation at 500×g, 5 min, 4 °C and the supernatant fraction containing stromal proteins without cpSRP54 was collected and immediately used in a PetD insertion experiment.

Mass spectroscopy analysis

For identification, the proteins were analyzed as described in [11, 28]. Tandem mass spectrometry experiments were carried out on a Q-Tof Premier mass spectrometer in conjunction with a nanoACQUITY UPLC system (Waters Corp) to obtain peptide information. Tandem mass spectrometry output lists of precursor and product ion information were used for NCBI database searching using the program Mascot Distiller (version 2.1, Matrix Science) [36].

In order to identify inserted proteins using mass spectroscopy following the insertion assays, the proteins were analyzed by SDS/PAGE and autoradiography. A gel band containing the radioactive polypeptide was excised. Moreover, at the same time, positive and negative control samples for mass spectroscopy analysis were prepared. As positive control, proteins from the molecular weight marker were used, while a piece of empty gel with no protein present was used as a negative control in order to determine any contamination that may have occurred during sample processing. Proteins were first extracted from polyacrylamide gels using electroelution (D-Tube, Novagene). Following electroelution, salts, SDS, and dye were removed by dialysis in D-Tube and the resulting proteins samples were then concentrated using an Amicon Ultra spin column.

All samples were analyzed by MALDI-TOF using the 4800 Plus MALDI-TOF/TOF™ Analyzer (Applied Biosystems) as well. Mass spectra for each gel slice were obtained in three replicates. Additionally, the proteins were digested by trypsin and analyzed by MS with PMF. The analysis procedure included washing, destaining, reduction and alkylation as described in [37].

Protein analyses

Protein concentrations were determined with the BCA™ protein assay kit (Pierce). Following the normalization of protein concentrations, proteins were separated via Tricine/Tris SDS-PAGE [11, 38] and transferred to nitrocellulose membranes using a semidry blotting system [11].

For protein detection, the following antibodies were used: polyclonal antibodies against PetB (antibodies raised against the 10 C-terminal amino acids, IRKQGIFGPL, of pea PetB) [28], cpSecY (antibodies raised against the 21 C-terminal residues of CRAEIISQKYKNIELYDFDKY of pea cpSecY) [39], ALB3 (antibodies raised against the 50 amino acid, PLTKQQVESTLAMQNPQPKIKAIQERYAGNQERIQLETSRLYTQAGVNPL, of the stromal protein sequence located between the first and second transmembrane (TM) domains) [40]. Antibodies were prepared in rabbits injected with the appropriate synthetic peptide (GenScript). The overexpressed and purified cpSRP54 [28] and a 21-amino acid synthetic peptide (YPIFAQQGYENPREATGRIVCANC), corresponding to the highly conserved N-terminus of the mature pea PetA [41], were used to prepare chicken (IgY) anti-cpSRP54 and anti-PetA antibodies, respectively. Antiserum was purified using the Pierce™ Chicken IgY Purification Kit (Thermo Scientific) followed by affinity-purification method against the antigen (Thermo Scientific). IgY antibodies was labelled with horseradish peroxidase using EZ-Link™ Plus Activated Peroxidase Kit (Thermo Scientific).

For antibody preparation, the antigen affinity purification (GenScript) or AminoLink™ Immobilization Kit (Thermo Scientific) were used. Affinity purification permits isolation of antibodies against the immunizing antigen and thereby eliminates cross-reactivity (‘unintentional’ binding) that would be present in the unpurified serum. The results of cross-reactivity studies are shown in Additional file 2: Figure S2. Antibodies were evaluated based on yield after affinity purification and the analysis of the specific activity of the purified antibodies. The titers of antigen-specific antibodies were determined by ELISA in 96-well Immulon 2HB flat bottom microtiter plates using the immunogen (Thermo Scientific).

The PetD protein presence in membrane and in immunoprecipitated complexes was confirmed by autoradiography followed by MS analysis with PMF since the antibodies directed against the PetD protein are unavailable [11]. Prior to autoradiography, gels were stained with Coomassie brilliant blue to confirm equal protein loading and then dried under vacuum. Importantly, we always used the same exposure time so that the image intensity was used as a consistent guide to the exposure required for subsequent autoradiography. Molecular masses were determined by comparison to molecular weight ladder (MW). For autoradiography, the MW bands were gently marked on the X-ray film (Kodak) with a pencil and the X-ray film was placed over the dry gel and exposed for 24 h. Quantitative densitometry of Western blot and autoradiography were performed using Image Lab software v. 4.1 (BioRad). The antibody bound proteins were detected by a chemiluminescence reaction using ECL (enhanced chemiluminescence, Amresco).

Bioinformatic analysis of PetD did not identify a membrane targeting signal sequence

Using TargetP tools, PetD was predicted to be a protein without a specific signal sequence or anchor sequence (Fig. 1d). Additionally, we used PredSL and PSORT software tools which combine multiple methods in order to predict a protein’s localization to the chloroplast and the thylakoids as well as to the mitochondrion or the secretory pathway [26]. Unfortunately, neither of the predictive methods used could identify the presence of a signal sequence potentially responsible for PetD transfer and its insertion into the thylakoid membrane (Fig. 1d).

The N-terminal regions of membrane proteins often serve as a membrane anchor, whereas the majority of known membrane proteins are translocated into membranes by a cleavable or non-cleavable signal sequence. In order to detect a potential non-cleavable signal within the PetD primary structure, the Kyte–Doolittle analysis was performed with a sequence window size of 5 amino acids that is appropriate for detecting a potential hydrophobic signal at N-terminus. The resulting data were aligned relative to the TM2-L18-TM3 amino acid sequence of the light-harvesting chlorophyll a/b binding proteins subunit B (LHCB1) (Fig. 1a). During the post-translational cpSRP-dependent insertion of LHCP, LHCP affiliates with cpSRP54 and cpSRP43 to form a stromal ‘transit complex’. Formation of this complex is governing by a sequence of specific recognition and interaction events. The interaction between LHCP and cpSRP43 involves a conserved 18 amino acid span, termed L18, that is located between two TMH, (TM2 and TM3) of LHCP. Hence, the L18 provides an internal targeting signal for a highly specific interaction between LHCPs and cpSRP43. After transit complex assembly, a third protein, cpFtsY, is assumed to target the transit complex to the thylakoid membrane [8].

The results of the Kyte–Doolittle analysis indicate that PetD does not contain a non-cleavable signal sequence. However, as shown in Fig. 1b, the average hydropathy plots of the compared PetD and LHCB1 sequences show similar profiles within the first 40 amino acids, but at a different hydrophobicity level. A consensus amino acid sequence, PLEIL (ankyrin repeats motive), is present in both proteins. Figure 1c shows the topography of the PLEIL sequence, marked in blue in the PetD structure [20]. However, the PetD sequence is missing DPLG motif that is responsible for cpSRP43 binding by LHCB1 [42]. Moreover, the average hydropathicity (GRAVY) values of these PLEIL sequences were similar (0.817 and 0.989 respectively).

For the most of the outer envelope proteins, their targeting plastid compartments (inner envelope membrane, stroma, and thylakoid) depend on the presence of a cleavable transit sequence in the precursor N-terminal region. Furthermore, the reports of Miras, et al. [43] suggest that neither the N-terminal nor the C-terminal sequences are essential for chloroplastic membrane localization of the ceQORH (Chloroplast Envelope Quinone Oxidoreductase Homologue) that has non-canonical sequence for chloroplast membrane integration. Therefore, we compared the PetD sequence with the non-canonical non-cleavable transit peptides of ceQORH. These analyses did not identify any sequence similarities between PetD and ceQORH within the range of 75 N-terminal amino acids. However, the Kyte–Doolittle hydropathy profiles of the first 75 amino acid of PetD and ceQORH protein (Additional file 3: Figure S4A) show some similarities between first 35 amino acid region of the PetD amino acid sequence is shifted by 8 amino acids toward the C-terminal end (Additional file 3: Figure S4B).

PetD is not inserted directly into the membrane in an unassisted posttranslational manner

To verify that the PetD protein integration into the pea thylakoid membrane can occur by the spontaneous pathway, the protein was expressed in a cell-free system as described in [11] in the presence of proteinase K-treated thylakoid membranes, but in the absence of the stromal fraction. During this translation protocol, 30% of the synthesized PetD protein was associated with the thylakoid membrane fraction (Fig. 2a). The molecular weight of inserted and integrated PetD protein was determined by mass spectrometry (Additional file 4: Figure S5A). MS analysis confirmed that the PetD construct is correctly expressed in the in vitro translation system. The molecular weight observed for PetD (17,487 Da) in MS analysis agrees well with the theoretical molecular weight of the monomeric species (17,476 Da).

The PetD protein integration with thylakoid membrane was further tested with use of strong chaotropic salts like NaBr and NaSCN that remove proteins peripherally bound to the membrane without membrane disassembly [32, 44, 45]. As shown in Fig. 2b, the membrane associated PetD was almost completely removed. Moreover, in order to facilitate spontaneous membrane integration, the synthesized PetD protein was solubilised in the presence of 0.2% DDM, and then mixed with purified thylakoid membrane. This approach, however, did not increase the resistance of the membrane-associated PetD to chaotropic removal (Fig. 2c). As a control for spontaneous protein insertion, we tested whether PsbW of the photosystem II or PetB can be spontaneously inserted into the thylakoid membrane [11, 46]. Both control experiments were performed under the same conditions as those for the petD protein. These experiments confirmed the previous observation that PsbW is indeed inserted into the thylakoid membranes by the spontaneous pathway. Furthermore, we also confirmed that PetB cannot be spontaneously inserted into thylakoid membranes (Additional file 5: Figure S6). All above results indicate that the native PetD cannot incorporate into thylakoid membranes spontaneously.

Post-translation import of PetD into the thylakoid membrane is SRP-dependent

The previous comparative analysis of ribosomal fractions bound to thylakoid membranes and from stroma indicate that PetD is translated on free ribosomes exclusively [22], suggesting a post-translational mechanism for protein membrane integration [12, 22]. Therefore, since the results of the thylakoid import assays excluded the spontaneous pathway for PetD membrane integration, we next verified the proposed post-translational translocation of PetD using a cell free in vitro system.

PetD expression was performed in a cell-free system by using the transcription–translation procedure where transcription and translation reaction were separated in space and time (see Materials and Methods, section: The cell free transcription-translation insertion assay). Moreover, during the post-translational insertion experiments, translations were performed in the absence of thylakoid membranes. As shown in Fig. 3a, without the stroma fraction the translation product (PetD) did not integrate into the membranes and was detected in the supernatant only. However, when the stroma fraction was present together with the membranes, the PetD was localized in the membrane fraction (Fig. 3b). Importantly, both inserted PetD and PetA (control) are resistant against all extraction procedures (Fig. 3c and f) [33]. Moreover, this result clearly shows that proper PetD membrane incorporation occurs via the post-translational mechanism and requires the stroma factors.

Hence using specific antibodies, we examined the mechanisms responsible for the membrane integration of PetD. Pretreatment of the thylakoids membrane with Ab against ALB3 prevented insertion of PetD into the membranes (Fig. 3d), suggesting that ALB3 is required for this process. However, although a significant level of PetD insertion was achieved in the presence of cpSecY Ab (Fig. 3e), it was still lower than in the control (Fig. 3e). Hence, the Ab against cpSecY only partially prevents cpSecA-dependent translocation of PetD into the membrane by the Sec pathway since this antibody blocks cpSecA binding to cpSecY without functional inhibition of cpSecY [39, 47, 48].

The current model of cytochrome b ₆ f complex assembly assumes the partial transcriptional activation of psbBNH-petBD operon within the PetB and PetD genes [49]. Following transcription, the PetB and PetD mRNA are translated into the polypeptides. As a result of this process, both proteins are incorporated into the membrane and form the polytopic monomeric core of the cytochrome b ₆ f complex. Hence, in order to validate the proposed mechanism for formation of PetD-PetB core complex, both genes were transcribed/translated in the cell-free system in the presence of the stroma and thylakoid membranes fractions (see PetD and PetB co-expression in cell-free assay in Materials and Methods section). After 1 h of translation, a membrane permeable and cleavable crosslinker (BSOCOES) were added. As shown in Fig. 4, the cotranslation of PetB and PetD resulted in detection of the PetD translation product with higher molecular weight than compared to PetD translated alone (Fig. 4a). Furthermore, the immunoprecipitation with an Ab against PetB of cross-linked isolated membrane complexes, followed by MALDI-TOF mass analysis confirmed that the radiolabeled complex contains a PetD-PetB complex (Fig. 4b and Additional file 4: Figure S5B).

As shown in Fig. 5, when the posttranslational insertion of PetD was performed in the presence of thylakoid membrane and stroma fractions devoid of the cpSRP54 protein, the PetD protein did not incorporate into membranes. Although a very weak signal is observed (Fig. 5a), it probably results from the incomplete removal of the cpSRP54 from stroma (as shown in Fig. 5b). Furthermore, two control experiments were set up to ensure the integrity of the method. In the first experiment, shown in Fig. 5c, we tested the co-translational insertion of PetB in the presence or absence of the cpSRP54 protein. As expected, we did not observe PetB insertion in the absence of the cpSRP54 protein. The goal of the second experiment (Fig. 5d) was to determine the impact of using a different concentration of antibody against cpSrp54 during PetD posttranslational insertion. We found that with increasing concentrations of antibody, there was corresponding decrease in the level of PetD embedded in the thylakoid membrane.

These MS results were further confirmed by Western blot analyses of the cross-linked complexes followed by immunochemistry. As presented in Fig. 6a, the Ab against cpSRP54 detected large amounts of cpSRP54 in the isolated membrane fractions. A smaller, but significant pool of cpSRP54 was detected in the immobilized ALB3 antibody unbound fraction as well. Importantly, the immunopurification of the cross-linked PetD complexes with antiserum directed against the cpSRP54 (Fig. 6b) allowed for the detection of ALB in the purified PetD complexes.

The proteins we identified by MS were previously recognized as crucial for SRP-mediated post-translational transport of LHC proteins into the thylakoid membrane [16]. Hence, our results of MS (Table 1) and autoradiograph analyses (Fig. 3) are strong indicators that PetD import into the membrane occurs through the post-translational SRP pathway.

MALDI-TOF mass analysis (Fig. 7) of intact proteins crosslinked by SPDP, immunoprecipitated and then cleaved with 25 mM DTT from the petD complexes shows two of CBB proteins (CCB1 and CCB3) that play important roles in guidance of apocytochromes and in haem groups for their covalent linkage by the cytochrome-c-hem lyase [6]. Importantly this analysis confirmed that after insertion PetD and PetB form a complex at the membrane with predicted size of 38.67 kDa (Fig. 4a, Table 2 and Additional file 4: Figure S5B). This result was confirmed by MS analysis of a tryptic digested PetD complex (Table 2).

No.	Proteina	Protein Scoreb	Peptide identified by MSc	Annotation in database
1	PetB	480	LEIQAIADDITSK VYLTGGFK IVTGVPDAIPVIGSSVVELLR	AIK21467
2	PetD	628	KPDLTDPVLRAK LLGVLLMVSVPAGLLTVPFLENVNK FQNPFR RPVATTVFLIGTVVALWLGIGATLPIEK	AAD41889
3	CCB1	225	IVNKTFVK LLDEVGNKAPNQVAGEVLSFFTR EDGTLSEIVVQGDDQQVEQMRK	Pisum_sativum_v1_Contig2650 [64]
4	CCB3	99	LMILADLDPATAK FPYVIAYAPTEPLLVPTRK	Pisum sativum_csfl_reftransV1_0082495 [64]
5	GUFP	43	VIASEALSAIR	Q9FNM5
6	GAPDH	117	TFAEEVNEAFR ELGIDLVIEGTGVFVDR	CAA33264

For the immunoprecipitation of complexes, antibody against cytochrome b ₆ was used. Bound proteins were directly analyzed by MS with PMF (data shown in Additional file 6: Table S2

^aChloroplast proteins which produced the highest scores are shown

^bOthers proteins with higher scores than 40 were not observed after cross-linking

^cPeptide identifications were accepted if they could be established at greater than 80.0% probability. Protein shown are from the double expression experiments. Analyses were repeated at least twice. MASCOT search and protein identification criteria were previously published in [11, 28, 36]