Introduction

Proteomics is, at present, an irreplaceable tool to study the response of living organisms to their environment. However, less attention has been applied to the plant proteome, although some reviews on this topic have been published in recent years (Rossignol 2001; Bagginsky and Gruissem 2004; Cánovas et al. 2004; Newton et al. 2004; Rose et al. 2004; van Wijk 2004; Rossignol et al. 2006; Jorrín et al. 2007, 2009). In order to improve the understanding of plant proteome expression, a number of studies have focused on organelle subproteomes, such as those of chloroplast and mitochondria responsible for the plant energy metabolism. Two-dimensional gel electrophoresis (2-DE), consisting of isoelectrofocusing (IEF) followed by SDS-PAGE electrophoresis (IEF/SDS-PAGE), was one of the first techniques applied to study the chloroplast proteome. This approach has proven successful for the analysis of both luminal and peripheral thylakoid proteins from Pisum sativum (Peltier et al. 2000) and Arabidopsis (Kieselbach et al. 2000; Peltier et al. 2002; Schubert et al. 2002). Blue-native polyacrylamide gel electrophoresis followed by SDS-PAGE (2-DE BN/SDS-PAGE), commonly used to characterize chloroplast protein complexes in plants and algae (Schägger et al. 1994), has also provided an excellent two-dimensional alternative for the IEF/SDS-PAGE in the case of the most hydrophobic thylakoid proteins. This technique has provided insight into the chloroplast proteome of Hordeum vulgaris (Ciambella et al. 2005) and Cucurbita pepo (Aro et al. 2005). Several strategies (fractionation followed by multi-dimensional protein-separation steps) have also been used to identify the hydrophobic thylakoid integral membrane proteome (Friso et al. 2004; Peltier et al. 2004), as well as to decipher the chloroplast envelope proteome (Ferro et al. 2002) of Arabidopsis. In addition, novel non-gel proteomic techniques have revealed the pathway abundance and novel protein functions in the photosynthetic membranes (Huber et al. 2004; Kleffman et al. 2004).

Fewer efforts have been made to identify the changes on the proteomic profile of chloroplast from plants undergoing different biotic and abiotic-stress factors. Phee et al. (2004) and Giacomelli et al. (2006) analyzed the response of the chloroplast proteome to high light in Arabidopsis thaliana; Andaluz et al. (2006) and Laganowsky et al. (2009) studied the proteomic changes induced by iron deficiency in the thylakoid membranes from Beta vulgaris and A. thaliana, respectively; Cui et al. (2005) investigated the chloroplast proteome as the main target of cold stress in Oryza sativa. Recently, Taylor et al. (2009) reviewed the changes induced by abiotic stress in the proteome of some organelles, including the chloroplast. Using 2-DE, Jones et al. (2006) detected significant changes in the chloroplast proteome of Arabidopsis plants infected with Pseudomonas syringae pv. tomato; photosystem II (PSII) and the Benson–Calvin cycle were revealed as the most affected processes. Zhou et al. (2006) carried out a proteomic analysis of the compatible interaction Fusarium graminearum and Triticum aestivum and concluded that the chloroplast is the organelle mostly affected by the infection. A leaf proteomic approach was also carried out to study pea responses to powdery mildew, and showed changes in proteins involved in photosynthesis and carbon metabolism in the host plant (Curto et al. 2006).

By using 2-DE, we have previously investigated the changes in the protein pattern of the oxygen-evolving complex (OEC) of PSII during the infection of Nicotiana benthamiana plants with the Spanish strain of the Pepper mild mottle tobamovirus (PMMoV-S) (Pérez-Bueno et al. 2004). We demonstrated that the OEC is one of the main targets of the viral infection in the chloroplast (Rahoutei et al. 2000). The proteomic analysis of the OEC from the infected plants showed that the PsbO and PsbP proteins correspond to protein families in N. benthamiana, and the content of the PsbP polypeptides decreased dramatically with respect to PsbO during the infection progress. Notably, there was a differential decrease of the various PsbP proteins, indicative of a different regulation of their expression during pathogenesis (Pérez-Bueno et al. 2004).

In the present study, we delve further into the analysis of the chloroplast proteome of N. benthamiana by using 2-DE and mass spectrometry (MS) followed by database searching. Nicotiana benthamiana is one of the most common experimental hosts and it is used in many studies on compatible plant–virus interactions, because of its susceptibility to infection by many different viral families. Progress in the knowledge of the N. benthamiana chloroplast proteome would be of high interest in the field of the impact of biotic stress on photosynthesis. In addition, PMMoV-S-induced changes in the chloroplast proteome after 14 days post-inoculation (dpi) were investigated. Our previous studies showed that this time span was one of the most representative of the PMMoV-S infections (Rahoutei et al. 2000; Pérez-Bueno 2003; Chaerle et al. 2006; Sajnani et al. 2007). The present proteomic study revealed that viral infection induced the down-regulation of several photosynthetic proteins involved in both the photosynthetic electron-transport chain and the Benson–Calvin cycle.

Materials and methods

Plant and virus material

Nicotiana benthamiana Gray plants were cultivated in a growth chamber at 100 μmol m−2 s−1 photosynthetically active radiation, provided by cool white fluorescent lamps, with a 16/8 h light/dark photoperiod, a temperature regime of 23/18°C (day/night), and a relative humidity of 65%. Plants at the stage of 6–7 fully expanded leaves were inoculated in the three lower leaves with 25 μl of inocula (50 μg/ml of purified PMMoV-S solution in 20 mM sodium phosphate buffer pH 7.0) per leaf. Mock-inoculated plants were treated only with buffer. The Spanish strain of PMMoV (genus Tobamovirus) was isolated in Almería, Spain (Alonso et al. 1989).

Visual symptoms and sample isolation

Plants infected with PMMoV-S developed new curly leaves at 5–6 dpi. Growth was inhibited in infected plants compared with the controls at 14 dpi. At 22–24 dpi, infected plants were totally senescent. For a more detailed description of symptoms, see Pineda et al. (2008b).

Chloroplast-enriched preparations from virus-infected plants and their corresponding controls were isolated according to Reche et al. (1997) at 14 dpi. For this purpose, we selected 8 g of plant material (curly and non-curly leaves from infected plants, as well as their homologous controls). Then, leaves were homogenized 4 s at 9.3 g (Polytron, Kinematica; Lucerne, Switzerland) in 35 ml of semi-frozen homogenization buffer [0,33 M sorbitol, 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)-KOH pH 7.5, 0.1% serum bovine albumin, 2 mM ethylenediaminetetraacetic acid (EDTA), and 1 mM Cl2Mg]. Next, the homogenate was filtered through four layers of miracloth and placed in Sorvall tubes. A layer of Percoll (40% in homogenization buffer) was deposited in the bottom of the tube (14 ml Percoll solution per 17 ml homogenate) and then centrifuged at 7,000g for 1 min at maximum speed and at 4°C (Sorvall RC5C centrifuge, SS-34 rotor; Thermo Fisher Scientific Inc., Waltham, USA). The supernant was removed and 400 μl of lysis buffer (5 mM Cl2Mg) were added to the chloroplast pellet before centrifuging at 13,000g for 10 min at 4°C. The pellet, containing the chloroplast-enriched preparations, was resupended in 200 μl of sample buffer (0,33 M sorbitol, 50 mM Hepes-KOH pH 7.5). Aliquots were stored at −70°C.

Sample solubilization for protein electrophoresis

Sample solubilization was carried out according to Schuster and Davies (1983) with minor modifications. First, 200 μl from a chloroplast-enriched preparation containing 1 μg Chl/μl was incubated with 200 μl extraction buffer [0.7 M sucrose, 0.5 M Tris, 30 mM HCl, 50 mM EDTA, 0.1 M KCl, 2% (v/v) 2-mercaptoethanol, and 2 mM phenylmethylsulphonyl fluoride (PMSF)] for 20 min at 4°C, vortexing several times. Then, 400 μl of 0.1 M Tris–HCl-saturated phenol (Amresco, Ohio, USA) were added. After 15 min of shaking at 4°C, the phases were separated by 10 min centrifugation at 16,000g and 4°C. The phenolic phase was recovered and re-extracted three times with an equal volume of extraction buffer, as mentioned above. Proteins were precipitated from the phenol phase by adding 6 volumes of 0.1 M ammonium acetate in methanol (HPLC grade) and incubated at −20°C overnight. The pellet, after a centrifugation for 10 min at 16,000g and 4°C, was washed three times with the ammonium acetate–methanol solution and then 8 volumes of 100% HPLC acetone was added. Proteins were kept at −20°C at least for 4 h and then precipitated under the same centrifugation conditions. The pellet was dried and solubilized in ReadyPrep Sequential Extraction Reagent 3 (BioRad; Hercules, CA, USA), containing 5 M urea, 2 M thiourea, 2% (w/v) CHAPS, 2% (w/v) SB 3–10, 40 mM Tris, 0.2% Bio-Lyte 3/10, and 2 mM TBP. The sample was incubated for 2 h at room temperature under gentle shaking. Insoluble material was removed by centrifugation for 10 min at 16,000g and 25°C. The protein content of the sample was measured using the 2-D Quant Kit (GE Health Care; St. Giles Chalfont, United Kingdom). After quantification, pI (isoelectrical point) standards (2-D SDS-PAGE Standards, BioRad) were added to samples used for alkaline (6–11) pH range gels.

First-dimension electrophoresis: isoelectrofocusing (IEF)

IEF was carried out on 11 cm ReadyStrip IPG Strip gels (BioRad) as well as on Immobiline DryStrip gels (GE Health Care) for the 4–7 and 6–11 pH ranges, respectively. The strips were allowed to rehydrate in a PROTEAN IEF Cell (BioRad) for 14–16 h at 50 V and 20°C with 200 μl of chloroplast preparation, containing 75 μg of proteins and a trace of bromophenol blue. Proteins were separated in the pH range 4–7 by IEF using a three-step procedure: 15 min at 250 V, followed by 2 h at 8,000 V and a final step of 8,000 Vh to complete 35,000 Vh. In the case of the 6–11 pH range, some intermediate steps were required according to Görg et al. (2000): 1 h at 150 V, 1 h at 300 V, 1 h at 1,000 V, 1 h at 3,000 V, 1 h at 6,000 V, and 6,000 Vh to complete 25,000 Vh. After focusing, the strips were immediately run or frozen at −80°C.

SDS-PAGE

Focused IEF strips were incubated at room temperature for 20 min with two changes of equilibration buffer, 50 mM Tris–HCl, pH 8.8, containing 6 M urea, 2% (w/v) SDS, 0.02% (w/v) bromophenol blue, 30% glycerol, and 0.5% (w/v) DTT (dithiothreitol). Strips were then layered on the top of a SDS 12% polyacrylamide gel and sealed with melted 0.8% agarose in running buffer. Then two gels were run simultaneously, one per treatment, in order to reduce variability. Running conditions were 17 mA for 30 min and 48 mA for 5–6 h. After electrophoresis, the gels were silver-stained (pH 6–11 map; Yan et al. 2000) or Sypro Ruby-stained (pH 4–7 map; BioRad, Steinberg et al. 1996). In order to ensure the reproducibility of the results, triplicate 2-DE gels were made from three independent chloroplast preparations, each from a different batch of plants. Gels were scanned and analysed using the ImageJ software (http://rsb.info.nih.gov/ij). The intensity of every spot of interest, as well as its corresponding area, were measured and then compared between gels from control and infected plants, according to the following equation:

$$ {\text{Difference}} = [ 1 - \left( {{\text{area}}_{{{\text{PMMoV}} - {\text{S}}}} \cdot {\text{intensity}}_{{{\text{PMMoV}} - {\text{S}}}} /{\text{area}}_{\text{control}} \cdot {\text{intensity }}_{\text{control}} } \right)] \cdot 100 $$

Only differences higher than 30% were considered as significant. Figures show the most representative gels.

Mass spectrometry and protein identification

Spots from silver- or Sypro Ruby-stained gels were manually excised and stored in milli-Q water at 4°C until MALDI-TOF/TOF (matrix-assisted laser desorption/ionization coupled to time-of-flight tandem mass spectrometry) analysis. Protein spots were digested automatically using a Proteineer DP protein digestion station (Bruker-Daltonics; Bremen, Germany). The digestion protocol used was that of Schevchenko et al. (1996) with minor variations. For peptide mass fingerprinting and TOF/TOF spectra acquisition an aliquot of a-cyano-4-hydroxycinnamic acid in 33% aqueous acetonitrile and 0.1% trifluoroacetic acid was mixed with an aliquot of the above digestion solution and the mixture was deposited onto an AnchorChip MALDI probe (Bruker-Daltonics). Peptide mass fingerprint spectra were measured on a Bruker Ultraflex MALDI-TOF/TOF mass spectrometer (Bruker-Daltonics) (Suckau et al. 2003) in positive ion reflector mode. When possible, mass measurements were performed automatically through fuzzy-logic-based software. Each spectrum was internally calibrated with mass signals of trypsin autolysis ions to reach a typical mass-measurement accuracy of ±25 ppm. The measured tryptic peptide masses were transferred through MS BioTools program (Bruker-Daltonics) as inputs to search the NCBInr database using Mascot software (Matrix Science; London, UK). MS/MS data from LIFT-TOF/TOF spectra were combined with MS peptide mass fingerprint data for database searching. Positively identified proteins are listed in Tables 1 and 2.

Table 1 Proteins identified by MALDI/TOF-TOF from the 2-DE showed in Figs. 1 and 2 and marked by the same number
Full size table
Table 2 pI standards identified by MALDI/TOF-TOF from gels displayed in Fig. 2
Full size table

Results

Chloroplastidic proteins from both control and PMMoV-S infected N. benthamiana plants were separated by 2-DE and the resulting gels were analyzed in order to detect the changes induced by the viral infection. In order to improve the gel resolution, separate 2-DE maps were made for the low (4–7) and high (6–11) pH ranges.

Chloroplast proteome of control Nicotiana benthamiana plants in the pH range 4–7

Most of the chloroplastidic proteins were found in this pH range. At least 150 spots could be detected in the 2-DE gel (Fig. 1a, b). Based on the reproducibility of the spot pattern and relative amount of the proteins, 78 spots were selected for the MALDI-TOF/TOF analysis. A number of 55 proteins could be identified (Table 1, spots # 1–55). Most of the proteins identified (31) were members of the different photosynthetic complexes of the thylakoid membrane: 16 polypeptides from PSII, three proteins from photosystem I (PSI), two cytochrome f (Cyt. f) polypeptides from the cytochrome complex and 10 spots corresponding to the ATP synthase. PSII proteins belonging to the OEC were one of the most abundant proteins in the gels. Six spots corresponding to the PsbO multiprotein family were found (# 1–6), showing the same M r (relative molecular mass) and different pI (isoelectrical point). The PsbP multiprotein family contributed with 10 spots (# 7–16), six of them (# 7–11, 16) with similar M r and differing in pI as well as four spots (# 12–15) of lower M r. From PSI, we detected the ferredoxin-NADP+-reductase (# 19), PsaC (# 31), and a cab protein from the light-harvesting complex of PSI (Lhca; # 32). In addition, we identified 2 Cyt. f polypeptides (# 18 and 20) from the cytochrome complex. The ATP synthase macrocomplex was represented by 10 proteins (# 21–30), corresponding to the five subunits of the CF1 catalytic portion: α (# 25–27), β (# 22–24), γ (# 21), δ (# 28), and ε (# 29 and 30). Spots corresponding to the same subunit showed the same M r but different pI.

Fig. 1
figure 1

a Sypro-stained 2-DE maps of chloroplast preparations from control Nicotiana benthamiana plants. Proteins were separated in the first dimension in linear (pH 4–7) IPG gel strips and in the second dimension in 12% polyacrylamide vertical gels. Identified spots by means of MALDI/TOF-TOF are listed in Table 1. LHCII proteins were detected by Western Blot analysis. The region enclosed by a dotted box is shown in detail in B. b Detail of the map in the pH 4–7 range, where proteins from # 42 to # 54 have better resolution

Full size image

Concerning the proteins involved in the Benson–Calvin cycle, the RuBisCO (ribulose 1,5 bisphosphate carboxylase oxygenase) was the most prominent protein in the 2-DE gel. RbcL (RuBisCO large subunit, # 36) was present in the high M r range of the gels with a pI of 6.9, and RbcS (RuBisCO small subunit) in the low M r range with two spots, showing the same M r but different pIs (# 34 and 35). We also detected other proteins from the same metabolic pathway: sedoheptulose-1,7-biphosphatase (SBPase, # 45), phosphoribulokinase (PRK, # 46), RuBisCO activase (Rca, # 47–50), phosphoglycerate kinase (PGK, # 49–51), and chloroplastidic sedoheptulose/fructose-bisphosphate aldolase (SFBA, # 52–54). Spots 49 and 50 contained a combination of two proteins: Rca and PGK. Other proteins identified corresponded to enzymes of the nitrogen metabolism (glutamine synthetase, GS, # 42–44).

Some proteins linked with stress processes were found in the gels: a RuBisCO-binding chaperonin (# 38), heat shock 70 kDa protein (Hsp70, # 40–41) and the FtsH-like protein (# 39), related to the D1 turnover and to the hypersensitive response in incompatible host–pathogen interactions. In addition, three proteins with unknown functions were present in the gels: hypothetic protein (# 37), thylakoid lumenal 16.5 kDa protein (# 33), and a plastid-lipid-associated protein (# 17).

Chloroplast proteome of control Nicotiana benthamiana plants in the pH range 6–11

In the 2-DE map of chloroplastidic proteins separated in the basic range, 53 spots appeared (Fig 2). All of these were analyzed by MALDI-TOF/TOF. Only 28 spots could be identified: seven spots corresponded to the pI standards (Table 2) and 21 spots were chloroplastidic proteins (Table 1, spots # 56–71). Due to the overlap between acidic (pH 4–7) and basic (pH 6–11) pH ranges, some proteins present in the acid 2-DE map also appeared in the basic one: PsbP from the OEC (# 16), the Cyt. f (# 18) the γ subunit of CF1 (# 21), Lhca (# 32) and RbcL (# 36). In the overlapping range (6–7) also appears a chlorophyll a/b binding protein from PSII (Lhcb; # 71) not identified in the acidic-range gels. New polypeptides corresponding to proteins already identified in the acidic range appeared in the 6–7 pH range: Cyt. f (# 56) and CF1 α subunit (# 70).

Fig. 2
figure 2

Silver-stained 2-DE maps of chloroplast preparations from control Nicotiana benthamiana plants. Proteins were separated in the first dimension in linear (pH 6–11) IPG gel strips and in the second dimension in 12% polyacrylamide vertical gels. Identified spots by means of MALDI/TOF-TOF are listed in Table 1. Dotted arrows represent pI standards, indicated by letters and listed in Table 2

Full size image

The other proteins identified in the pH range 6–11 were: two spots for the E subunit of PSI (PsaE) with the same M r and different pI (# 67 and 68); eight spots corresponding to D subunit of PSI (PsaD), four of these with similar M r and differing in pI (# 57–60); and 4 PsaD proteins with lower M r (# 61–64). The PsbQ protein from the OEC was detected in the low M r region of the gel, matching two spots (# 65 and 66). The ribosomal protein L14 was the lowest M r identified protein (# 69).

Comparative analysis of chloroplast proteomes from control and PMMoV-S infected plants in both pH ranges

Comparative analysis of 2-DE gels from control and PMMoV-S infected N. benthamiana plants at 14 dpi (Fig. 3; Table 3), enabled us to identify some proteins that were down-regulated by the viral infection (difference more than 30%, according to “Materials and methods” section) in the pH range 4–7: Cyt. f (# 18 and 20), FNR (# 19), δ and ε subunits of CF1 (# 28–30), RbcL (# 36), RbcS (# 34 and 35), hypothetical protein (# 37), GS (# 42–44), SBPase (# 45), PRK (# 46), Rca (# 47–50), PGK (# 49–51), Lhca (# 32), and SFBA (# 52-54). These are involved in the photosynthetic electron-transport chain, the Benson–Calvin cycle, and nitrogen metabolism. Only one new protein appeared in the chloroplast from infected plants: the viral-coat protein (CP; # 55) with a M r and pI of 17.6 kDa and 5.2, respectively. Other down-regulated proteins during pathogenesis were detected in the basic range 2D gels: PsaD forms and the ribosomal protein L14 (# 57–64, and 69).

Fig. 3
figure 3

Changes detected by 2-DE in the expression levels of different chloroplastidic proteins. Left panels correspond to control plants and right panels to PMMoV-S infected plants. All the proteins (excepting # 55, the coat protein of PMMoV-S) present lower expression levels in infected plants. Numbers correspond to those of Table 1. Spots from 1 to 55 are located in the pH range 4–7 gels (Sypro Ruby-stained), while proteins from 56 to 71 are placed in pH range 6–11 gels (silver-stained)

Full size image
Table 3 Content of chloroplastidic proteins from PMMoV-S infected plants relative to the controls (averages ± standard deviation)
Full size table

Discussion

Two-dimensional gel electrophoresis has been successfully applied to the analysis of the chloroplast proteome, and especially to the peripheral and luminal one in plants such as Arabidopsis and P. sativum (Kieselbach et al. 2000; Peltier et al. 2000; van Wijk 2001; Whitelegge 2003). Only a few authors have investigated changes in the chloroplast proteome under abiotic or biotic stress: Andaluz et al. (2006) and Timperio et al. (2007) analyzed the effect of iron deficiency in B. vulgaris and Spinacea oleracea, respectively; Cui et al. (2005) investigated the impact of cold stress in O. sativa; Jones et al. (2006) evaluated changes in the chloroplast proteome of Arabidopsis plants infected with P. syringae pv. tomato; Díaz-Vivancos et al. (2008) showed that changes in the expression levels of proteins from the electron-transport chain correlated with the appearance of reactive oxygen species during the infection of P. sativum with Plum pox virus.

Also, we performed a proteomic analysis of the OEC from N. benthamiana infected with PMMoV (Pérez-Bueno et al. 2004), finding that the OEC proteins were the most abundant ones in the chloroplast proteome. Previously, we had described them as a target of infection with this tobamovirus (Rahoutei et al. 1998). In the present study, we go further into the analysis of the chloroplast proteome of N. benthamiana at 14 dpi using 2-DE and MS followed by database searching. At 14 dpi, the effects of PMMoV-S infection in N. benthamiana plants are highly significant, as evaluated by visual symptoms (Pineda et al. 2008b), lower mRNA levels of chloroplastidic proteins (Pérez-Bueno 2003), imbalance of ATP + NADPH assimilatory potential (Sajnani et al. 2007), alterations in fluorescence emissions (Pineda et al. 2008a, b), and stomatal closure (Chaerle et al. 2006).

More than 200 spots were visualized in the 2-DE gel from chloroplast preparations isolated from control plants: 150 spots in the pH range 4–7, mostly in the 20–43 kDa M r region (Fig. 1a, b), and 53 spots in the pH range 6–11 (Fig. 2). From the MALDI-TOF/TOF analysis, 71 proteins were identified; in addition, several LHCII proteins were identified by Western Blot (Fig. 1a, data not shown). The 2-DE gels showed that proteins involved in carbon fixation were still present in our preparations after removing the stromal fraction. However, some authors have already found evidence that Benson–Calvin cycle enzymes can assemble into stable multi-enzyme complexes bound to the stromal face of thylakoid membranes in different plant species (Rodríguez Andrés et al. 1987; Gontero et al. 1988; Süss et al. 1993, 1995; Anderson et al. 1996; Andaluz et al. 2006), including N. tabacum (Jebanathirajah and Coleman 1998). This kind of association has been proposed to grant direct access to required cofactors such as ATP and NADPH, and to prevent interferences by other metabolic pathways (Süss et al. 1993; Anderson et al. 1996).

Particularly remarkable was the high number of PsbO and PsbP spots, 6 and 10, respectively (Fig. 1a; Table 1). We have previously identified four isoforms of each protein family (Rahoutei et al. 1998; Pérez-Bueno et al. 2004), both encoded by a multigene family with at least four isogenes (Pérez-Bueno 2003). PsbO protein families have been described in P. sativum (Wales et al. 1989), Lycopersicon esculentum (Görlach et al. 1993) and A. thaliana (Kieselbach et al. 2000; Peltier et al. 2002; Schubert et al. 2002). PsbP protein families were found in Nicotiana tabacum (Hua et al. 1991a, b; Takahashi et al. 1991; Hua et al. 1993), A. thaliana (Peltier et al. 2002) and Sinapis alba (Merkle et al. 1990). It has been demonstrated in N. tabacum that the four PsbP isoforms are functionally equivalent in vivo, but the expression of the four isogenes is required for optimal PSII activity (Ishihara et al. 2005). We have also demonstrated that the expression level of the various PsbP isoforms is differentially affected by the PMMoV infection in N. benthamiana (Pérez-Bueno et al. 2004). Analysis of the two PsbO isogenes from A. thaliana revealed that they were not equivalent; the down-regulation of the expression from every isogene induced that of other PSII proteins and a reduced activity of this photosynthetic complex (Murakami et al. 2005).

The higher number of PsbP and PsbO spots found in this study compared with our previous findings (Pérez-Bueno et al. 2004), could be because, in our earlier experiments, we used self-made disk gels for the IEF, whereas, in the present study, we carried out the IEF using ready-made IPG Strip gels. This technique allows higher resolution, more reproducible patterns and better focusing of the basic proteins (Corbett et al. 1994; Görg et al. 2000). From the six and ten spots detected for the PsbO and PsbP proteins, respectively, some may have been the result of post-transcriptional alterations, given that we have found only four isogenes codifying each protein family (Pérez-Bueno 2003). Further analysis of these spots carried out by electrospray ionization-MS will clarify this point.

The reported fall in the levels of PsbP isoforms due to the viral infection (Pérez-Bueno et al. 2004) was not evident in the present acidic 2-DE maps. With the OEC proteins being the most prominent ones in this pH range, the PsbO and PsbP signals were saturated (Fig. 1a), because we loaded a high amount of protein in these gels in order to identify minor proteins in the N. benthamiana chloroplast proteome.

In the alkaline 2-DE map (pH 6–11), we detected two spots for the PsbQ protein (pI 9.4 and 9.5; Fig. 2 and Table 1), in accordance with the two psbQ isogenes that we have reported previously for N. benthamiana (Pérez-Bueno 2003). Two PsbQ isoforms were identified by 2-DE on A. thaliana chloroplast preparations (Kieselbach et al. 2000; Schubert et al. 2002), while Peltier et al. (2000) detected only one spot in 2-DE gels from P. sativum chloroplasts.

The different subunits of the CF1, together with the OEC proteins, are the most abundant polypeptides in the acidic pH range (Fig. 1a, b). These data agree with the 2-DE maps from the thylakoid proteome of A. thaliana investigated by Friso et al. (2004). In our gels, all subunits of the CF1 fragment from the ATP synthase were found. Due to the analogy with Friso et al. (2004) gels, we propose that proteins with the same M r close to CF1 α and CF1 β (Fig. 1a) would be protein trains of these subunits.

We also identified several high M r chaperonins (Fig. 1; Table 1): the 60-kDa chaperonin, participating in RuBisCO assembly (Buchanan et al. 2000) and Hsp70, involved in PSII stress protection (Schroda et al. 2001). The FtsH-like related to PSII repair and turnover (Nixon et al. 2005; Cheregi et al. 2005; Zaltsman et al. 2005a, b) was also present in the 2-DE gels. Two of the spots identified in our 2-DE maps (hypothetical protein and the thylakoid lumenal 16.5 kDa proteins) have unknown functions; the hypothetical protein amount decreased in the plants infected with PMMoV-S (Fig. 3; Table 3).

Comparative analysis of the chloroplastidic proteome from control and PMMoV-S infected plants at 14 dpi revealed that the expression levels of some proteins implicated in Benson–Calvin cycle, electron-transport chain and nitrogen metabolism decreased during the PMMoV-S infection (Fig. 3; Table 3).

The proteins involved in the Benson–Calvin cycle affected by the PMMoV infection are: RbcL, RbcS, Rca, SFBA, SBPase, PGK, and PRK (Fig. 3; Table 3). The decrease in the amounts of these polypeptides is consistent with the lower transcript levels of these proteins during viral and bacterial pathogenesis (Pérez-Bueno 2003; Ishiga et al. 2009, respectively). The degradation of the RbcL subunit under different stress conditions (UV radiation, ozone, and osmotic, etc.,) is for some authors a consequence of oxidative alterations in this enzyme (Houtz et al. 2003 and references therein). Oxidative stress is usually induced in infected plants (Foyer and Harbinson 1994; Vranová et al. 2002). The Rca has a dual function: activating the RuBisCO complex; and, as a chaperonin, protecting protein synthesis under stress conditions (Sánchez de Jiménez et al. 1995; Rokka et al. 2001). SFBA is codified by two genes in Nicotiana with differential expression levels under salt stress (Yamada et al. 2000). A similar situation seems to occur in our PMMoV-S infected plants. We have identified three protein spots corresponding to SFBA, with number 52 being the most affected by the PMMoV-S infection. Probably, the additional spot is due to post-translational modifications. In order to evaluate the impact of the lower levels of the SBPase (this being one of the key regulatory enzymes of the Benson–Calvin cycle) and PRK in infected plants, we take as model a transgenic plant showing different expression levels of both enzymes. Decreases on SBPase amounts limit CO2 fixation and leaf carbohydrate content (Harrison et al. 1998), while the overexpression of this enzyme enhances photosynthesis against high-temperature stress in transgenic rice plants (Feng et al. 2007). In general, it is well documented that the fall in the levels of Benson–Calvin enzymes induced adverse effects on CO2 fixation, photosynthesis rate, carbohydrates accumulation, and plant growth (Haake et al. 1998; Banks et al. 1999; Raines 2003). We have also documented a decrease of the net photosynthesis rate of the host plant under infection with PMMoV (Pineda et al. 2008b); a similar situation was found by Zhou et al. (2006) and Ishiga et al. (2009) during bacterial challenge.

The plastid GS, a key protein in the nitrogen metabolism, is also diminished in plants infected by PMMoV-S (Fig. 3; Table 3). Senescent plants show also a fall in both protein and mRNA levels of this enzyme, whereas, the cytosolic one is induced (Masclaux-Daubresse et al. 2005; Teixeira et al. 2005). PMMoV-S infection on N. benthamiana shows similarities with senescence processes. In clover (Trifolium repens L.), senescence lowers levels of PsbO, both RuBisCO subunits and Rca, as well as GS (Wilson et al. 2002). This down-regulation pattern is similar to that found in symptomless leaves of PMMoV-infected N. benthamiana plants for the transcripts of these proteins (Pérez-Bueno 2003), confirming that viral infection accelerates senescence processes.

With respect to electron-transport chain proteins, in the 2-DE maps of the alkaline pH range from N. benthamiana chloroplasts, some PSI proteins appeared, such as the PsaD and PsaE (Fig. 2; Table 1). These highly hydrophilic proteins, together with PsaC, which could be distinguished in gels within a low pH range (Fig. 1a; Table 1), are located on the PSI stromal side, being involved in the ferredoxin-docking (Sétif et al. 2002). PsaD and PsaE are codified by nuclear genes, two isogenes for both existing in A. thaliana (Naver et al. 1999; Ihnatowicz et al. 2004) and Nicotiana spp. (Obokata et al. 1993, 1994; Yamamoto et al. 1993). In the case of Nicotiana sylvestris, the two psaE isogenes codify for four different polypeptides; however, we could detect only two PsaE spots in N. benthamiana. In N. sylvestris, PsaD is also a protein family with two isoforms (PsaD1 and PsaD2), functionally redundant, encoded by two isogenes. Nevertheless, we detected up to eight PsaD spots, the expression levels of which diminished due to the viral infection (Figs. 2, 3; Table 3). Because of the high homology between PsaD1 and PsaD2 proteins (96/95% similarity/identity, Ihnatowicz et al. 2004), they could not be distinguished by the MALDI-TOF/TOF analysis. The psaD genes are differentially expressed during leaf development in N. sylvestris (Yamamoto et al. 1993). Arabidopsis PsaD double mutants showed that this protein is essential for photosynthesis (Ihnatowicz et al. 2004). Its down-regulation induces both a high non-photochemical quenching and steady-state chlorophyll fluorescence (Haldrup et al. 2003) as well as a decrease in linear PSI electron-transport activity (Montesano et al. 2004); a similar phenomenon was observed in our PMMoV-infected plants (Chaerle et al. 2006; Pérez-Bueno et al. 2006; Sajnani et al. 2007; Pineda et al. 2008a, b). Other authors have also shown that the PsaD content decreases in Fe-deficient Arabidopsis plants (Laganowsky et al. 2009).

Other proteins from the electron-transport chain with expression levels diminished due to the PMMoV-S infection are Cyt. f, FRN, and some polypeptides belonging to the light-harvesting complexes from both photosystems (Fig. 3; Table 3), according to the recent results for Fe-deficient Arabidopsis plants (Laganowsky et al. 2009). While we detected only one spot corresponding to FNR, several isoforms of this oxidoreductase have been found in other plants. In Arabidopsis (two isoforms) and maize (three isoforms), FNR isoproteins are believed to play different roles in response to the nitrogen regime (Hanke et al. 2005; Okutani et al. 2005, respectively), whereas in wheat the two present isoforms may substitute for each other under certain physiological conditions (Grzyb et al. 2008). It is unknown how many genes code for FNR in tobacco (Lintala et al. 2007). Transgenic tobacco plants overexpressing FNR exhibited enhanced tolerance to photo-oxidative damage (Rodríguez et al. 2007), while lower activity of this protein in the chloroplast leads to impaired plant growth and restricts photosynthesis capacity (Hajirezaei et al. 2002). The impact of the PMMoV infection on the photosynthetic electron transport described in our previous studies (Rahoutei et al. 2000; Pérez-Bueno et al. 2004), could be explained by the down-regulation of proteins involved in this photosynthetic step as well as in light harvesting.

Another chloroplastidic protein present in the alkaline pH range and affected by the PMMoV-S infection is the ribosomal protein L14 (Fig. 3; Table 3), a protein from the chloroplastidic large ribosomal subunit (Tanaka et al. 1986). The significance of its decrease in PMMoV-S-infected plants remains unclear.

The presence of multigene families for different chloroplastidic proteins—PsbO, PsbP, PsbQ, PsaD, PsaE, and SFBA- in N. benthamiana appear to enable the host plant to react in a more flexible way to environmental stimuli and different stress factors (Koiwa et al. 2002).

The understanding of the chloroplast proteome of N. benthamiana, a promiscuous host for different viral families, is of high interest for the study of the virus–plant interactions. Despite its limitations, the 2-DE technique was powerful enough to generate maps with about 200 polypeptides. Taking into account that the N. benthamiana genome is not completely sequenced, we identified a high number of proteins, similar to the number reported in other publications cited above. The future applications of non-gel-proteomic techniques for the study of the N. benthamiana chloroplastidic proteome, together with the progress in sequencing the genome of this host plant, will help to enlarge an emerging investigation field lacking in these types of studies: the influence of biotic-stress factors in the proteome of the photosynthetic apparatus.