• Chloroplasts contain a plastoquinone–NADH–oxidoreductase (Ndh) complex involved in protection against stress and the maintenance of cyclic electron flow. Inactivation of the Ndh complex delays the development of leaf senescence symptoms.
• Chlorophyll a fluorescence measurements, blue native gel electrophoresis, immunodetection and other techniques were employed to study tobacco (Nicotiana tabacum) Ndh-defective mutants (ΔndhF).
• The ΔndhF mutants compared with wild-type plants presented: (i) higher photosystem II : photosystem I (PSII : PSI) ratios; (ii) similar or higher levels of ascorbate, carotenoids, thylakoid peroxidase and superoxide dismutase, yield (ΦPSII) and maximal photochemical efficiency of PSII levels (Fv/Fm) than wild-type plant leaves of the same age; (iii) lower values of nonphotochemical quenching yield (ΦNPQ), but not at very high light intensities or during induced leaf senescence; (iv) a similar decrease of antioxidants during senescence; (v) no significant differences in the total foliar area and apical growth rate; and (vi) a production of viable seeds significantly higher than wild-type plants.
• These results suggest that the Ndh complex is involved in one of the redundant mechanisms that play a safety role in photosynthesis under stress, which has been conserved during evolution, but that its deletion increases fitness when plants are grown under favourable controlled conditions.
In the present work, we perform a more detailed study of the consequences of Ndh complex suppression on the life cycle, accumulation of antioxidant activities and substances, growth, photosynthetic capacity and productivity of tobacco plants. Our results show that the Ndh-defective plants show subtle phenotypic changes and increased fitness compared with wild-type plants.
Materials and Methods
Tobacco seeds (Nicotiana tabacum L. cv. Petit Havana) from wild-type and ΔndhF plants ( Martín et al., 2004) were germinated in a culture chamber (23°C, 60 µmol m−2 s−1, RH 98%) and 1 month after germination were grown in individual pots under controlled glasshouse conditions (20–25°C, mean light intensity at noon 300 µmol m−2 s−1, RH higher to 80%), avoiding blocked distribution of wild-type or ΔndhF plants on the glasshouse racks. For each experiment described here, unless otherwise stated, we randomly selected 10–30 plants per group from a pool of 40 plants for each genotype.
Measurement of in vivo Chla fluorescence
Chlorophyll a fluorescence was measured with a portable pulse-modulated fluorometer (PAM-2000, Walz, Effeltrich, Germany) at room temperature. Attached leaves were kept in the dark for 30 min before measurements were taken. The minimum (dark) fluorescence yield (F0) was obtained after excitation of the leaves with a weak measuring beam from a light-emitting diode. The maximum fluorescence yield (Fm) was determined with a 600-ms saturating pulse of white light (8000 µmol m−2 s−1). Variable fluorescence (Fv) was calculated as Fm − F0. Following a 2-min dark readaptation, actinic white light (260 µmol m−2 s−1 unless otherwise stated) was switched on, and saturating pulses were applied at 1-min intervals for 11 min to determine the maximum fluorescence yield during actinic illumination (), the level of modulated fluorescence during a brief interruption (3 s) of actinic illumination in the presence of 6 µmol m−2 s−1 far red (730 nm) light (), and the Chl fluorescence yield during actinic illumination (Fs). Quenching caused by the nonphotochemical dissipation of absorbed light energy (NPQ) was calculated at each saturating pulse, according to the equation NPQ = (Fm − )/. The fraction of open PSII centres (qL) was estimated according to Kramer et al. (2004) as qL =qP(/Fs), where qP represents the coefficient for photochemical quenching calculated as ( − Fs)/( − ) (Schreiber et al., 1989). The quantum efficiency of PSII photochemistry, φPSII, which is closely associated with the quantum yield of noncyclic electron transport, was estimated from ( − Fs)/ (Genty et al., 1989; Kramer et al., 2004). After the quenching parameters had reached a steady state, the actinic light was switched off and the variation in F0 level was recorded. The fraction of absorbed light dissipated in the PSII antennae (ΦNPQ = 1 − ΦPSII − ΦNO) and the fraction that is not used in photochemistry or dissipated in the PSII antenna (ΦNO = 1/(NPQ + 1 + qL(Fm/F0 − 1))) were estimated according to Kramer et al. (2004).
Isolation of chloroplasts and thylakoids
Chloroplasts were purified from tobacco leaves using Percoll gradients according to the protocol of Guéra et al. (2000) and thylakoids were purified by submitting the purified plastids to an osmotic shock in hypotonic buffer (50 mm HEPES–NaOH pH 7.6, 4 mm MgCl2, 2 mm Na2– ethylenediaminetetraacetic acid (EDTA), 0.5 mm phenylmethylsulphonylfluoride (PMSF)). The lysed plastids were centrifuged at 10 000 g for 10 min and the pellet was resuspended in a buffer containing 50 mm HEPES–NaOH pH 7.6, 4 mm MgCl2, 0.3 m NaCl, 2 mm Na–EDTA. After 30 min incubation on ice, thylakoid membranes were recovered by centrifugation at 10 000 g for 10 min and washed twice in buffer (50 mm HEPES–NaOH pH 7.6, 0.3 m sorbitol, 2 mm Na–EDTA).
Gel electrophoresis, zymograms and immunoassays
SDS–PAGE was carried out as described by O’Farrel (1975). Blue native PAGE was carried out in a linear gradient of 6–12% of polyacrylamide, as described by Schägger et al. (1994), after solubilization of samples with 1.5%n-dodecyl-β-d-maltoside. Superoxide dismutase activity from isolated chloroplasts was detected after native gel electrophoresis as described by Rao et al. (1995). Electroblotting and immunodetection were performed as described by Guéra et al. (2000). The anti-NDH-F antibody was the same as that described by Cataláet al. (1997). Polyclonal antibodies against the PsaA and PsbA proteins were purchased from AgriSera AB (Vännäs, Sweden).
Protein was determined after 2% sodium dodecyl sulfate (SDS) solubilization of samples, with a Bio-Rad (Hercules CA, USA) detergent-compatible protein assay kit based on the method of Lowry et al., 1951), using BSA as standard. Peroxidase activity of thylakoid solubilized fractions was determined from the rate of increase of absorbance at 250 nm caused by the oxidation of p-hydroquinone with H2O2, as described previously (Zapata et al., 1998).
Chlorophyll and carotenoids were determined according to Lichtenthaler (1987). The ascorbate content of leaves was measured as described by Luwe et al. (1993). Total nitrogen in seeds was analysed by Kjeldahl using 0.1 n sulphuric acid, and titration with 0.1 n NaOH and 5% methyl red as indicator. anova was performed on experimental data, and statistical significance (P < 0.05) was judged by the least significant differences (LSD) test. Semiquantitative analysis of PSI, PSII, light-harvesting complex II (LHCII), NDH-F, psaA and psbA was performed by densitometry using a UVP Easy digital image analyser (Ultra-Violet Products Ltd, Cambridge, UK). Statistical analyses were performed by using the spss program (SPSS Inc., Chicago, IL, USA).
Disruption of the ndhF gene affects the PSII : PSI ratio and parameters of Chla fluorescence
Plastid DNA insertional mutants (ΔndhF) defective in ndhF gene expression (Martín et al., 2004) were selected on spectinomycin during five generations. These mutants showed a delay in the development of leaf senescence symptoms, as measured by the loss of chlorophylls and an increase in lipid peroxidation (Zapata et al., 2005). In order to understand better the role of the Ndh complex during the plant life cycle, we first determined possible changes in the composition of the photosynthetic machinery related to the loss of this complex. Blue native gel electrophoresis showed that, for the same amount of total thylakoid protein, the ΔndhF mutant leaves contained higher levels of PSII and LHCII than the wild-type leaves (Fig. 1a). These results were confirmed by immunodetection after SDS–PAGE of the psbA (PSII) and psaA (PSI) proteins (Fig. 1b), which indicated that the total amount of core proteins in the PSII reaction centre, as determined by densitometric analysis, was about fivefold higher in the ΔndhF genotype than in the wild type. This relation was inverted for the PSI reaction centre protein. It could be expected that these changes in the PSII : PSI ratio would have an effect on photosynthetic electron transport and, indeed, the quantum efficiency of PSII electron transport (ΦPSII) was higher in ΔndhF leaves than in wild-type leaves (Fig. 2a) for increases of light intensity up to 1200 µmol m−2 s−1. The fluorescence parameter qL is proportional to the reduction state of QA (Kramer et al., 2004). As expected from an active Ndh complex, the qL values were lower (a more reduced quinone pool and therefore a higher reduction state of QA) in the wild type than in the ΔndhF leaves (Fig. 2b). In previous studies carried out with Ndh-deficient plants, the nonphotochemical dissipation of light energy (NPQ) was seen to be lower than in plants with a functional Ndh complex (Li et al., 2004; Guéra et al., 2005). In the case of the ΔndhF mutants (Fig. 2c), the values of ΦNPQ were lower than in the wild type for low light intensities, but were similar for the highest values. The proportion of light intensity not employed in the antenna for photochemical reactions, and not dissipated nonphotochemically (ΦNO), was higher in the wild type than in the mutants (Fig. 2d).
Effects of ndhF gene disruption on antioxidants in tobacco plants
The levels of peroxidase activity associated with thylakoid membranes were similar during development for wild-type and ΔndhF leaves. The maximum values were registered for the fourth leaf (wild type 4.16 ± 0.40 nkat mg−1 FW; ΔndhF 3.60 ± 0.20 nkat mg−1 FW) and the minimum values for the eighth leaf (wild type 1.20 ± 0.05 nkat mg−1 FW; ΔndhF 1.00 ± 0.02 nkat mg−1 FW) of 70-d-old plants. No differences were found in levels of the plastid-associated Cu/Zn SOD in leaves of the same age from mutant or wild-type plants (Fig. 3), but a parallel decrease in the level of plastid SOD was registered for the older leaves of both genotypes. These experimental data suggest that the onset of leaf senescence in tobacco is related to a decrease in superoxide dismutase and an increase in Ndh complex activities, as proposed by Zapata et al. (2005). No significant differences were found in β-carotene or lutein levels between ΔndhF (68.5 ± 0.7 mmol mol−1 Chla + b for β-carotene; 112 ± 10 mmol mol−1 Chla +b for lutein) and wild type (67.0 ± 1.0 mmol mol−1 Chla + b for β-carotene; 118 ± 7 mmol mol−1 Chla + b for lutein) tobacco leaves. However, the ascorbate/dehydroascorbate ratio was significantly higher in young ΔndhF (eighth leaf of 70-d-old plants; 2.70 ± 0. 0.15) than in wild-type leaves (1.50 ± 0.30), although the levels were similar in older leaves of both kinds of plant (fourth leaf of 70-d-old plants; 1.00 ± 0.60 and 0.96 ± 0.50 for ΔndhF and wild type, respectively).
Effects of ndhF gene disruption on growth and seed production of tobacco plants
When the development of ΔndhF plants was followed under glasshouse conditions, no significant differences were observed in the height of the main shoot (Fig. 4a). The increase in foliar area with time was also similar for both genotypes (Fig. 4b). The number of flowers or fruits did not change, but a significant difference was observed in the number of seeds per fruit (Table 1). The proportion of N per seed was not increased in the ΔndhF mutants, but the higher number of seeds (which had the same mean dry weight for wild type and ΔndhF) suggested an increase in the total amount of N translocated from vegetative to reproductive organs. Finally, the germination rate was similar for mutant and wild-type seeds. Therefore inactivation of the Ndh complex when plants were grown under favourable controlled conditions involved an increase in the fitness of tobacco.
Table 1. Quantitative comparison of characters related to fitness in tobacco (Nicotiana tabacum L. cv. Petit Havana) wild-type and ΔndhF plants
Values are means ± SE (n for germination = 50–100 seeds per plant from 15 to 20 different plants for each genotype; n for capsules per plant = 10 for each genotype; n for total seeds per capsule = 10 capsules per plant from five different plants for each genotype; n for N content per seed = 5 for each genotype). Different superscript letters indicate statistically significant differences between means (P < 0.001).
88.2 ± 1.0A
82.1 ± 0.9A
Capsules per plant
63 ± 5B
62 ± 6B
Total seeds per capsule
1006 ± 87C
774 ± 91D
Nitrogen content per seed (% of dry weight)
4.30 ± 0.23E
4.25 ± 0.39E
Effects of ageing, light-intensity fluctuation and disruption of the ndhF gene on photosynthetic electron transport
The first visual symptoms of senescence in the basal leaves of wild-type tobacco appeared during the ninth to tenth weeks after germination, while the appearance of the same symptoms was delayed approx. 30 ± 7 d in the ΔndhF mutants, as reported previously (Zapata et al., 2005). The sprouting time of wild-type and ΔndhF leaves has been compared, and no statistically significant difference (P > 0.05, n = 40 per genotype) was found between both genotypes. Thus, for plants of the same age (80 d after germination and 7–14 d after anthesis), it was observed that the older most basal leaves (second to fourth) of wild-type tobacco (mean Chl content 0.412 ± 0.120 µg g−1 FW) showed lower maximum activity of PSII electron transport (Fv/Fm) than the equivalent leaves (mean Chl content 0.635 ± 0.80 µg g−1 FW) of ΔndhF plants. For younger leaves (eighth to tenth), Fv/Fm values were similar for wild-type (mean Chl content 1.303 ± 0.150 µg g−1 FW) and ΔndhF (mean Chl content 1.408 ± 0.125 µg g−1 FW) leaves (Fig. 5a). Equivalent results were obtained (faster decay of Fv/Fm in wild type) following the development of individual leaves over time (Fig. 5b).
The results obtained for increasingly older leaves from the same plant clearly matched those for experiments with leaf senescence induction. For the latter experiments, 1-cm-diameter disks from the eighth leaf of 70-d-old plants were incubated at 23°C, floating on distilled water in total darkness, for the periods indicated (Fig. 6). Under these experimental conditions, the ΔndhF leaf disks retained higher values of Fv/Fm and ΦPSII (Fig. 6a,b) for a longer period, indicating that the delay in leaf senescence observed in the ΔndhF mutants is independent of light effects on this process. Furthermore, the degree of plastoquinone reduction, as indicated by qL (Fig. 6c), was more pronounced in the wild-type leaf disks until the third day of incubation in the dark, as might be expected in the presence of an activated Ndh complex. Both genotypes showed an increase in ΦNPQ during the first days of incubation in the dark, reaching a maximum on the fourth day, but in contrast with data obtained in intact leaves, the ΔndhF mutants presented higher values than wild-type leaf disks from the second day of incubation (Fig. 6d). Otherwise, the levels of ΦNO were constant for the ΔndhF until the fourth day of incubation in the dark, while they steadily increased in the wild type (Fig. 6e).
The function of the Ndh complex has also been related to adaptation against sudden changes in light intensity (Endo et al., 1999; Casano et al., 2000). To test this possibility we recorded Chla fluorescence in younger (eighth) and older (fourth) leaves of 80-d-old plants from both genotypes under natural conditions, during the hours of light on 1 day (28 June 2005), including cloudy and sunny periods (Fig. 7). At the beginning of the day, under a low light intensity of approx. 100 µmol m−2 s−1, the greatest proportion of light absorbed by the PSII antenna was employed for photochemistry (ΦPSII) in the young leaves (eighth leaf) of both genotypes, but the proportion of ΦPSII was lower in the wild type. For both genotypes the proportion of ΦPSII diminished in the older leaves, although this decrease was clearly more pronounced for wild-type than for ΔndhF leaves. On the other hand, ΦNPQ was always lower in ΔndhF than in wild-type leaves. ΦNPQ increased with leaf age, particularly in wild-type plants, as described previously for senescence of Arabidopsis thaliana leaves (Wingler et al., 2004). The values for ΦNO were almost constant throughout the day, independent of light intensity, genotype or leaf age. Otherwise the responses to sudden variations in sunlight were similar in leaves of similar age in both genotypes: a decrease in ΦPSII after a sudden burst of sunlight and rapid recovery after clouding over.
Many articles on ndh-disrupted mutants report irrelevant phenotypic differences when the mutants are compared with wild-type plants (Burrows et al., 1998; Shikanai et al., 1998; Horváth et al., 2000; Takabayashi et al., 2002), but detailed data on the developmental cycle of these mutants are scarce. Recently, Zapata et al. (2005) described ndhF-disrupted tobacco plants showing a delay in leaf senescence. The results presented by Zapata et al. (2005) focused mainly on the last stages of leaf development. Now we have included observations at different stages of leaf and whole-plant development of Ndh-defective tobacco mutants. The ΔndhF plants had a higher PSII : PSI ratio than wild-type plants (Fig. 1). These results are complementary to those described by Baena-González et al., 2003) for psbA-deficient mutants in which the lack of the corresponding core PSII protein was accompanied by a 10-fold relative increase in the Ndh complex and the plastid terminal oxidase. These data indicate the high capacity for adaptation of the photosynthetic machinery of plants, able to compensate (at least partially) for the loss or dysfunction of some components. In this way, the lack of function of the Ndh complex, which has been associated with resistance to oxidative stress conditions, could also be compensated for in young mutant leaves by the higher ascorbate/dehydroascorbate ratio observed.
The ΔndhF mutation had no effect on foliar area, apical growth, sprouting of new leaves, blooming or germination (Fig. 4; Table 1). Otherwise, significant differences between wild-type and ΔndhF plants were found for ΦPSII, ΦNPQ and the onset of leaf senescence (Figs 2, 5, 6). Lower levels of NPQ are characteristic of ΔndhF mutants ( Martín et al., 2004; Guéra et al., 2005; Zapata et al., 2005). NPQ is a general photoprotective mechanism in plants with a main component (qe) dependent on thylakoid lumen acidification (Krause & Weis, 1991; Horton et al., 2005; Szabo et al., 2005). It is shown that the proportion of photons absorbed in the PSII antenna, the energy of which is dissipated by NPQ (ΦNPQ), was higher in wild-type than in ΔndhF plants (Figs 2, 7) under moderate light intensity. This fact permits higher levels of ΦPSII in the ΔndhF mutants, as long as ΦNO is not increased. Although no alterations in NPQ were found in several studies with Ndh-deficient plants (Shikanai et al., 1998; Joët et al., 2001; Munekage et al., 2004), Burrows et al. (1998) and Li et al. (2004) described ΔndhCKJ or ΔndhB tobacco mutants, respectively, as having a reduced ability to quench fluorescence nonphotochemically. Also, lower levels of NPQ have been described for sunflower nonmutant plants under conditions of low nonphotochemical reduction of the plastoquinone pool, apparently associated with low activity of the chlororespiratory chain (Feild et al., 1998). Similar observations have been made on the youngest leaves of wild-type barley, which present no or very low levels of the Ndh complex (Guéra et al., 2005). A plausible explanation for the lower levels of NPQ observed in ΔndhF mutants than in other Ndh-defective plants could be the possible involvement of the NdhF subunit in H+ pumping (Casano et al., 2004). Whatever the case, the NPQ levels maintained by the ΔndhF mutants appear to be sufficient to allow the complete development of plants under controlled conditions. One of the main functions of NPQ is to protect against sudden fluctuations of light intensity (Demmig-Adams & Adams, 1993), and the same has been proposed for the Ndh complex in experiments performed under laboratory conditions (Endo et al., 1999; Casano et al., 2000; Guéra et al., 2005). Here it is shown that abrupt light variations did not produce a differential pattern in the responses of the ΔndhF mutants compared with wild-type plants (Fig. 7). Therefore, although the Ndh complex can contribute to generating NPQ, this is probably only a complementary mechanism. Indeed, as illustrated in Fig. 2, at high light intensities the levels of NPQ became similar for ΔndhF and wild-type plants, indicating that the Ndh complex modulates, but is not an indispensable factor for the generation of, NPQ.
The basal leaves of wild-type plants developed senescence symptoms (lower levels of photosynthetic electron transport, increased membrane lipid peroxidation, chlorophyll loss and necrosis) before the leaves of ΔndhF plants (Zapata et al., 2005). We show that the levels of several antioxidants (ascorbate, β-carotene, peroxidase, the chloroplastic Cu/Zn SOD) diminished in both genotypes in an ageing-dependent way. Thus the delay in leaf senescence in ΔndhF mutants cannot, in principle, be attributed to the promotion of higher levels of antioxidants. In fact, Zapata et al. (2005) hypothesized that a possible decay of chloroplastic SOD during leaf ageing, together with high Ndh complex activity (which should lead to more reduced electron transporters and so higher levels of the Mehler reaction, thus increasing superoxide production) would contribute to the rise of steady-state levels of ROS, which could trigger programmed cell death processes. A role for ROS in leaf programmed cell death is usually accepted (Navabpour et al., 2003; Yoshida, 2003; Woo et al., 2004; Zimmermann & Zentgraf, 2005), both for their toxicity and for their function as signalling molecules at low concentrations. However, leaf senescence is a process that can be reverted (Zavaleta-Mancera et al., 1999a, 1999b) and therefore can be distinguished from programmed cell death (Thomas et al., 2003). We followed the evolution of induced senescence, a traditional way to study leaf senescence, by incubating leaf disks from both genotypes on distilled water in the dark. As for the observations made in whole plants, results with leaf disks revealed that the decrease in photosynthetic capacity is delayed in the ΔndhF mutants (Fig. 6). Therefore inactivation of the Ndh complex can delay the processes of both natural and induced dark-leaf senescence. The latter fact is important in looking for a function for the Ndh complex in senescence because, under dark conditions, the light-induced photosynthetic mechanisms for generating ROS evidently cannot be accomplished and the levels of the Mehler reaction and generation of superoxide radical in the dark must be very low.
It must be borne in mind that the main function attributed to the Ndh complex is to act as an intermediate in chlororespiration, reducing the plastoquinone pool, which in a further step will be oxidized by the quinol-oxidase or peroxidase activity described for thylakoid membranes (Zapata et al., 1998; Carol & Kuntz, 2001; Joët et al., 2002; Peltier & Cournac, 2002). During the onset of leaf senescence, enhancement of this route could lead to the generation of H2O2, which could, in turn, function as a signalling molecule triggering leaf senescence. This mechanism could be largely homologous to the plasma membrane NADPH-dependent oxidase that generates H2O2 in response to pathogen attack, which develops similar symptoms to those of leaf senescence (Lamb & Dixon, 1997; Apel & Hirt, 2004). It might also be of significance that, under induced senescence, the NPQ steady-state values were higher in ΔndhF than in the wild type (Fig. 7d), indicating a possible change in the functional role of the Ndh complex as an element implicated in one or more chains of electron transfer, and probably of energy dissipation.
Finally, the delay in leaf senescence might explain why ΔndhF plants present a very significant increase in seed production (Table 1), as has been described for other senescence-delayed mutants under nonlimiting nutrient supply conditions (Gan & Amasino, 1995; Spano et al., 2003).
This work was supported by the Spanish MCYT grants REN2003-04465/GLO and BMC2003-01261.