Composition and physiological function of the chloroplast NADH dehydrogenase-like complex in Marchantia polymorpha


  • Minoru Ueda,

    1. Department of Botany, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Tetsuki Kuniyoshi,

    1. Department of Botany, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Hiroshi Yamamoto,

    1. Department of Botany, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    2. CREST, Japan Science and Technology Agency, Chiyoda-ku, Tokyo 102-0076, Japan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Kazuhiko Sugimoto,

    1. Department of Botany, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Kimitsune Ishizaki,

    1. Division of Integrated Life Science, Graduate School of Biostudies, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Takayuki Kohchi,

    1. Division of Integrated Life Science, Graduate School of Biostudies, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Yoshiki Nishimura,

    1. Department of Botany, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Toshiharu Shikanai

    Corresponding author
    1. Department of Botany, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
    2. CREST, Japan Science and Technology Agency, Chiyoda-ku, Tokyo 102-0076, Japan
    Search for more papers by this author



The chloroplast NADH dehydrogenase-like (NDH) complex mediates cyclic electron transport and chloro-respiration and consists of five sub-omplexes, which in angiosperms further associate with photosystem I (PSI) to form a super-complex. In Marchantia polymorpha, 11 plastid-encoded subunits and all the nuclear-encoded subunits of the A, B, membrane and ferredoxin-binding sub-complexes are conserved. However, it is unlikely that the genome of this liverwort encodes Lhca5 and Lhca6, both of which mediate NDH–PSI super-complex formation. It is also unlikely that the subunits of the lumen sub-complex, PnsL1–L4, are encoded by the genome. Consistent with this in silico prediction, the results of blue-native gel electrophoresis showed that NDH subunits were detected in a protein complex with lower molecular mass in Marchantia than the NDH–PSI super-complex in Arabidopsis. Using the plastid transformation technique, we knocked out the ndhB gene in Marchantia. Although the wild-type genome copies were completely segregated out, the ΔndhB lines grew like the wild-type photoautotrophically. A post-illumination transient increase in chlorophyll fluorescence, which reflects NDH activity in vivo in angiosperms, was absent in the thalli of the ΔndhB lines. In ruptured chloroplasts, antimycin A-insensitive, and ferredoxin-dependent plastoquinone reduction was impaired, suggesting that chloroplast NDH mediates similar electron transport in Marchantia and Arabidopsis, despite its possible difference in structure. As in angiosperms, linear electron transport was not strongly affected in the ΔndhB lines. However, the plastoquinone pool was slightly more reduced at low light intensity, suggesting that chloroplast NDH functions in redox balancing of the inter system, especially under low light conditions.


Plants convert light energy into chemical energy in the forms of NADPH and ATP via electron transport in the thylakoid membrane. This light-dependent process of photosynthesis consists of linear electron transport driven by two photosystems that function in series, and cyclic electron transport mediated solely by photosystem I (PSI) (Shikanai, 2007). Whereas linear electron transport generates both ATP and NADPH, PSI cyclic electron transport preferentially produces ATP and balances the production ratio of ATP and NADPH. In Arabidopsis thaliana, PSI cyclic electron transport consists of two partly redundant pathways, a PROTON GRADIENT REGULATION 5 (PGR5)/PGR5-LIKE PHOTOSYNTHETIC PHENOTYPE 1 (PGRL1)-dependent antimycin A-sensitive pathway and an NADH dehydrogenase-like (NDH) complex-dependent pathway (Munekage et al., 2002, 2004; DalCorso et al., 2008). The former is essential for induction of the energization-dependent quenching (qE) component of non-photochemical quenching (NPQ) of chlorophyll fluorescence and is likely to contribute to ATP supply for CO2 fixation (Munekage et al., 2002). In contrast, NDH is required for protection of the photosynthetic machinery against oxidative stress (Endo et al., 1999); a severe phenotype was observed in double mutants lacking both pathways of PSI cyclic electron transport, even under growth chamber conditions (Munekage et al., 2004).

The presence of NDH in chloroplasts was first implied by complete sequencing of two plastid genomes in tobacco (Nicotiana tabacum) and liverwort (Marchantia polymorpha) (Ohyama et al., 1986; Shinozaki et al., 1986). It was not clear why chloroplast genomes encode proteins homologous to subunits of mitochondrial NADH dehydrogenase, which is involved in respiratory electron transport (Matsubayashi et al., 1987), but, on the basis of characterization of the ndhB mutant (M55) in cyanobacteria (Ogawa, 1991), it was later suggested that chloroplast NDH is involved in PSI cyclic electron transport. This hypothesis was tested by knockout of plastid-encoded ndh genes in tobacco (Burrows et al., 1998; Kofer et al., 1998; Shikanai et al., 1998). These studies were followed by identification and characterization of a series of Arabidopsis chlororespiratory reduction (crr) mutants that are specifically defective in NDH activity (Hashimoto et al., 2003) and also a rice mutant defective in the assembly of chloroplast NDH (Yamori et al., 2011). Because chloroplast NDH is structurally related to mitochondrial NADH dehydrogenase (Matsubayashi et al., 1987), and electron transport is linked to the plastid terminal oxidase (Okegawa et al., 2010), NDH-mediated electron transport is often called chlororespiration; this name is especially appropriate when this electron transport occurs in the dark (Peltier and Cournac, 2002).

In cyanobacteria, distinct types of NDH are present in a cell. They facilitate formation of complexes involved in multiple processes, including respiration, PSI cyclic electron transport and CO2 uptake, by modifying their subunit composition (Battchikova et al., 2011). The sequence similarity of chloroplast NdhD and NdhF with their cyanobacterial orthologs, together with the absence of subunits functioning in CO2 uptake, suggests that chloroplast NDH originated from cyanobacterial NDH involved in respiration and PSI cyclic electron transport, rather than from that functioning in CO2 uptake. The mutant phenotypes of Arabidopsis and tobacco are consistent with this possibility (Burrows et al., 1998; Shikanai et al., 1998; Hashimoto et al., 2003). Recently, we reported that chloroplast NDH consists of four sub-complexes [A, B, membrane and lumen (Peng et al., 2009)] and further includes the most fragile part of NDH involved in electron donor binding, i.e. NdhS–NdhU (Ifuku et al., 2011; Yamamoto et al., 2011), as summarized in Table 1. This NDH complex interacts with PSI to form a super-complex via the minor light harvesting complex I (LHCI) Lhca5 and Lhca6 (Peng et al., 2008, 2009). The subunits included in the B and lumen sub-complexes are not encoded by cyanobacterial genomes (Table 1), suggesting that chloroplast NDH has changed its structure during evolution of eukaryotic phototrophs. This process required co-evolution of many specific assembly factors in chloroplasts (Peng et al., 2010, 2011a, 2012).

Table 1. Summary of the genome survey of NDH subunit genes
  Synechocystis sp. PCC6803 Marchantia polymorpha Arabidopsis thaliana Referencesb
  1. ✓Indicates the presence of genes in the genome. aPlastid-encoded. bResults of the sequence analyses are deposited as Figures S1–S3.

Membrane subcomplex
 NdhA-NdhGaa Ohyama et al. (1986)
Subcomplex A
 NdhH-Kaa Ohyama et al. (1986)
 NdhL-NdhOFigure S1a–d
Fd-binding subcomplex
 NdhSFigure S1e
 NdhT, NdhU Figure S1f, g
Subcomplex B
 PnsB1-PnsB5  Figure 1h–l
Lumen subcomplex
 PnsL1-4  Figure S2a–c
 PnsL5 Figure S2d
 Lhca5, Lhca6  Figure S3

How did the huge chloroplast NDH complex of angiosperms evolve from ancestral cyanobacterial NDH? To answer this question, the liverwort Marchantia polymorpha may be a key organism. To study the early evolution of land plants, especially the water-to-land transition, Marchantia is emerging as a premier model system, in which gene manipulation is feasible (Ishizaki et al., 2008). Based on phylogenetic analysis, Marchantia is considered to be the earliest divergence from the land plant lineage (Qiu et al., 2006). Plastid transformation has been also established in Marchantia (Chiyoda et al., 2007), and it has become possible to disrupt the plastid ndh gene. In this paper, we report the unique structure of chloroplast NDH in Marchantia. Furthermore, we knocked out the plastid ndhB gene and characterized the mutant phenotype.


The Marchantia genome lacks some NDH subunit genes assigned in Arabidopsis

In Arabidopsis, chloroplast NDH includes extra sub-complexes that are absent in cyanobacteria, and further associates with PSI to form a super-complex (Peng et al., 2009; Table 1). To assess whether the Marchantia genome encodes all the NDH subunit genes assigned in Arabidopsis (Ifuku et al., 2011; Peng et al., 2011b), the EST database was surveyed using subunit sequences of Arabidopsis as queries. All the subunits in the membrane sub-complex (NdhA–NdhG) and four subunits of sub-complex A (NdhH–NdhK) are encoded in the plastid genome (Ohyama et al., 1986). ESTs that encode the putative functional orthologs of NdhL–NdhO were also discovered in the nuclear genome (Table 1), suggesting that sub-complex A is conserved among cyanobacteria, Marchantia and angiosperms. Recently, we described three subunits, NdhS/CRR31, NdhT/CRRJ and NdhU/CRRL, that are involved in high-affinity ferredoxin (Fd) binding of NDH (Yamamoto et al., 2011); these were also conserved in the Marchantia genome (Table 1). Sub-complex B, which is absent in cyanobacterial NDH, includes five subunits, PnsB1–B5, in Arabidopsis (Ifuku et al., 2011) (Table 1). ESTs that encoded the putative functional orthologs were discovered for all of the sub-complex B subunits, although the similarity was lower in PnsB2 than in the other subunits (Figure S1h–l).

In Arabidopsis, the lumen sub-complex consists of at least five subunits (Ifuku et al., 2011) (Table 1). PnsL1 is a PsbP-like protein, PPL2 (Ishihara et al., 2007), whereas PnsL2 and PnsL3 are two PsbQ-like proteins (Majeran et al., 2008; Suorsa et al., 2010; Yabuta et al., 2010). Because PsbP-like and PsbQ-like proteins form small families that are conserved in phototrophs (Ifuku et al., 2010), it is not easy to conclude whether the Marchantia genome encodes or does not encode functional orthologs of PnsL1–L3. In fact, some EST clones of Marchantia encode proteins weakly similar to PnsL1–L3. However, the Marchantia protein detected by a BLAST search using AtPnsL1 as a query was most similar to AtPPL1 involved in PSII repair in Arabidopsis (Ishihara et al., 2007), and is most likely to be MpPPL1 (Figure S2a). Similarly, the proteins detected in the search using PnsL2 and PnsL3 as queries were most similar to AtPsbQ1 or AtPsbQ2, which encode subunits of PSII (Figure S2b). An additional two lumen subunits are members of the peptidyl-prolyl cis/trans isomerase (PPIase) super-family, namely PnsL4/FKBP16-2 (Majeran et al., 2008; Peng et al., 2009) and PnsL5/AtCYP20-2 (Majeran et al., 2008; Sirpiöet al., 2009). Because these immunophilins are ubiquitous in eukaryotic cells, determining whether or not genes encoding functional orthologs of PnsL4 and PnsL5 are present in the Marchantia genome is a complicated issue. However, the Marchantia protein detected in the BLAST search using AtPnsL4/FKBP16-2 as a query was most similar to AtFKBP13 and is unlikely to encode the NDH subunit (Figure S2c). In contrast, an EST clone with high similarity to PnsL5/AtCYP20-2 was found in Marchantia, implying that its function is related to that of NDH (Table 1). The lack of genes encoding the functional orthologs of PnsL1–L4 was also supported by the results of a tBLAST search of the available genome information in May 2012 (see Experimental Procedures). Thus, it is unlikely that the entire lumen sub-complex reported in Arabidopsis is conserved in Marchantia.

Two minor LHCI proteins, Lhca5 and Lhca6, mediate super-complex formation between NDH and PSI in Arabidopsis (Peng et al., 2009), and this interaction is required to stabilize NDH (Peng and Shikanai, 2011). A BLAST search detected many ESTs and EST contigs encoding proteins similar to both LHCIs in Marchantia. Clones with E values greater than e−30 were more similar to the light harvesting complex II (LHCII) genes of Arabidopsis. EST sequences were summarized into four EST contigs likely to encode LHCI in Marchantia (Figure S3). These are likely to be functional orthologs of Arabidopsis Lhca1–4, which form the PSI–LHCI super-complex (Ben-Shem et al., 2003). However, ESTs corresponding to Arabidopsis Lhca5 and Lhca6 were not discovered (Table 1). We also performed a tBLAST search of the available genome information on Marchantia. The top three hits detected using AtLhca5 and AtLhca6 as queries corresponded to MpLhca2–4.

Structure of the chloroplast NDH complex in Marchantia

In Arabidopsis, chloroplast NDH forms a huge super-complex with PSI (Peng et al., 2009). This NDH–PSI super-complex may therefore be separated from other pigment–protein complexes of the thylakoid membrane in blue native (BN) gels (Peng et al., 2008). However, our in silico survey suggested that some of the NDH subunits assigned in Arabidopsis were absent in the Marchantia genome (Table 1). In particular, we did not discover genes corresponding to Lhca5 and Lhca6, which mediate super-complex formation in Arabidopsis. To assess whether chloroplast NDH in Marchantia forms a super-complex with PSI, protein complexes solubilized from the thylakoid membrane were separated using a BN gel (Figure 1). As a control, protein complexes solubilized from the Arabidopsis thylakoid membrane were also separated on a BN gel (Figure 1a,b). Consistent with previous reports (Peng et al., 2008, 2009), an NDH–PSI super-complex was detected as a high-molecular-mass green band (Figure 1a), which was more clearly detected in the gel stained with Coomassie Brilliant Blue (Figure 1b). However, the corresponding low-mobility band was missing in Marchantia (Figure 1a,b). To specify the localization of NDH subunits in the BN gel, the blot was probed using specific antibodies raised against recombinant NdhH and NdhM of Marchantia, as well as with PsaA and CP47 antibodies (Figure 1c). Consistent with the absence of band I, neither NDH subunit antibody recognized any protein complexes with a molecular mass >1000 kDa. The PsaA antibody recognized protein complexes that were broadly distributed from approximately 500–800 kDa, mainly co-localized with the major green band corresponding to the PSII dimer and PSI. The NdhM and NdhH signals also peaked at the same location as PsaA (Figure 1c).

Figure 1.

 BN gel analysis of thylakoid membrane protein complexes. Protein complexes were solubilized from thylakoid membranes isolated from Arabidopsis leaves and Marchantia thalli.
(a, b) Protein complexes were separated by BN-PAGE (a) and stained with Coomassie Brilliant Blue (b).The approximate size and identity of each band on the BN gel was estimated on the basis of a previous paper (Peng et al., 2008) and is indicated by vertical arrows.
(c) The BN gel was further subjected to 2D SDS–PAGE. Their protein blots were probed with specific antibodies against PsaA (PSI), NdhH, NdhM and CP47 (PSII). Horizontal red arrows indicate the positions of signals detected by the antibodies.

We could not completely eliminate the possibility that the interaction of NDH and PSI is fragile and was not maintained in BN gel electrophoresis. However, under the same experimental conditions used in this study, all of the NDH subunits of the NDH–PSI super-complex in Arabidopsis were detected (Peng et al., 2009), with the exception of NdhS/CRR31 that easily dissociates from the main complex in BN gel electrophoresis (Yamamoto et al., 2011). Taken together with the results of the in silico analysis (Table 1), we propose that chloroplast NDH does not form a super-complex with PSI, and some subunits, including PnsL1–L4, Lhca5 and Lhca6, are missing in the chloroplast NDH complex in Marchantia.

Knockout of the chloroplast ndhB gene

The results of both the genome survey (Table 1) and the BN gel analysis (Figure 1) suggested that the structure of the chloroplast NDH complex was divergent between Arabidopsis and Marchantia. This possible structural difference may reflect the difference between two organisms in terms of the physiological function of chloroplast NDH. To test this possibility, we took advantage of the established technology of plastid transformation in Marchantia (Chiyoda et al., 2007). We selected ndhB as a target because it is essential for activity in Synechocystis sp. PC6803, tobacco and Arabidopsis (Ogawa, 1991; Burrows et al., 1998; Kofer et al., 1998; Shikanai et al., 1998; Hashimoto et al., 2003). The pCSΔndhB vector includes the 4.9 kb region including rps12, rps7, ndhB, psbM, trnL and the 5′ exon of ycf66 cloned from the Marchantia plastid genome; the ndhB gene was disrupted by insertion of the chimeric aadA cassette (Shikanai et al., 1998) into the 3′ exon (Figure 2a). The intact ndhB sequence was replaced by the disrupted version by plastid transformation. Among 38 spectinomycin-resistant plants, 31 lines were shown by PCR to be transplastomic (Figure 2b). Primers 1 and 2 were designed for regions outside that cloned in pCSΔndhB, and were used with a combination of primers 3 and 4, designed for the aadA cassette (Figure 2a). Fragments of the expected sizes were amplified in the 31 lines (results for 10 representative lines are shown), indicating that the aadA cassette was inserted into the genome via homologous recombination (Figure 2b, Prs1-4 and Prs2-3). All lines contained both the wild-type and transgenic copies, indicating the heteroplasmic status of the genomes (Figure 2b, Prs1-2). Lines 13, 14 and 20 were used for asexual propagation through a gemma, which arises from a single initial cell in cupules (Barnes and Land, 1908). Although growth was maintained on non-selective medium without spectinomycin, the wild-type genome copies were totally segregated out in the fresh thalli, resulting in establishment of homoplasmic lines lacking the ndhB gene (Figure 2c). We selected two independent lines, ΔndhB14-5 and ΔndhB20-9 (male plants) for further analysis. Although both lines completely lacked ndhB, they grew like the wild-type on medium containing sucrose (Figure 3a). In the absence of sucrose in the medium, the ΔndhB lines had similar growth rates, indicating that ndhB was not involved in photoautotrophic growth (Figure 3b).

Figure 2.

 Knockout of the plastid ndhB gene using plastid transformation.
(a) Schematic representation of the WT and ΔndhB genomes. The region surrounded by two triangles was subcloned into the pCSΔndhB vector, and the 3′ exon of ndhB was disrupted by insertion of the aadA cassette at the SpeI site. Gray and white boxes indicate exons and introns, respectively. Positions of primers used in the genome analysis are indicated. The sizes of expected PCR products are also indicated.
(b) PCR analysis of the genomic DNA isolated from the first generation of transgenic thalli. Of 38 independent lines, representative results for 10 (lines 11-20) are shown.
(c) The genome was analyzed again using genomic DNA isolated from thalli originating from asexual reproduction through gemma. The culture was maintained on medium without spectinomycin. Lines 13-1, 14-1, 14-2, 14-5 and 20-9 were male plants, whereas lines 13-2, 13-4 and 20-8 were female plants.

Figure 3.

 Growth of ΔndhB lines. Both line ΔndhB14-5 and line ΔndhB20-9 grew like the wild-type (WT) on medium with sucrose (a) and medium without sucrose (b).

To study the impact of ndhB knockout on accumulation of thylakoid membrane proteins, blots of thylakoid proteins isolated from the wild-type and the two ΔndhB lines were probed with antibodies (Figure 4). The levels of PsaA (PSI), PsbA (PSII) and the cytochrome b6f complex were unaffected in the ΔndhB lines. In contrast, the levels of NdhH, NdhM and PnsB1 (NDH subunits) were below the detection limit (12.5%). This result strongly suggests that these proteins are included in the same protein complex. The presence of sub-complexes A and B was suggested by the results of the in silico analysis (Table 1); this likelihood was experimentally supported by the instability of NdhH, NdhM and PnsB1 in the ΔndhB lines. As in cyanobacteria (Zhang et al., 2004), tobacco (Peng et al., 2008) and Arabidopsis (Peng et al., 2009), disruption of ndhB probably destabilized the entire NDH complex, but this defect did not affect accumulation of other thylakoid membrane complexes.

Figure 4.

 Immunoblot analysis of thylakoid proteins in wild-type and ΔndhB lines. Membrane protein extracts corresponding to 2 μg chlorophyll (100%) were loaded onto each lane, as well as a dilution series of wild-type (WT) proteins. For PsbA, protein extracts corresponding to 0.4 μg chlorophyll (100%) were loaded. Antibodies used are indicated on the right. The asterisk indicates non-specific signals.

Chloroplast NDH mediates antimycin A-resistant, Fd-dependent plastoquinone reduction in Marchantia

In spinach and Arabidopsis, NDH activity is monitored as Fd-dependent plastoquinone (PQ) reduction in ruptured chloroplasts (Endo et al., 1997; Munekage et al., 2004). To test whether chloroplast NDH mediates similar electron transport in Marchantia, chloroplasts were isolated from wild-type and ΔndhB thalli and osmotically ruptured. Fd-dependent PQ reduction was monitored as an increase in chlorophyll fluorescence.

Addition of NADPH alone slightly increased the fluorescence level; this NADPH-dependent PQ reduction was not affected in the ΔndhB20-9 line, but was partially affected in the ΔndhB14-5 line (Figure 5). Addition of antimycin A, which inhibits PGR5-dependent cyclic electron transport, also partially inhibited NADPH-dependent PQ reduction in all lines, including the wild-type. Subsequent addition of Fd induced a further increase in chlorophyll fluorescence in the wild-type (Figure 5). This Fd-dependent PQ reduction was partially impaired in the ΔndhB lines; ΔndhB14-5 was more severely affected than ΔndhB20-9. Because ndhB was completely knocked out (Figures 2c and 4), this difference is unlikely to represent a difference in NDH activity between the lines. Consistently, the remaining Fd-dependent PQ reduction activity was completely arrested by addition of antimycin A. Antimycin A also partly impaired PQ reduction in the wild-type; the remaining activity probably corresponded to chloroplast NDH-dependent electron flow. As observed in Arabidopsis (Munekage et al., 2004), PGR5-dependent, antimycin A-sensitive activity and NDH-dependent, antimycin A-insensitive activity constitute all of the Fd-dependent PQ reduction in ruptured chloroplasts. Unlike in Arabidopsis, minor NADPH-dependent activity was detected, although this PQ reduction activity was almost negligible in the presence of Fd.

Figure 5.

 Fd-dependent PQ reduction in ruptured chloroplasts from wild-type and ΔndhB lines.
(a) Schematic model of Fd-dependent PQ reduction in ruptured chloroplasts. Antimycin A (AA) inhibits PGR5–PGRL1 complex-dependent PQ reduction. PQ reduction was monitored as an increase in chlorophyll fluorescence.
(b) Fd-dependent PQ reduction activity in ruptured chloroplasts isolated from various genotypes as indicated. The increase in chlorophyll fluorescence level (inline image) was monitored after consecutive addition of 0.25 mm NADPH and 5 μm Fd under weak illumination (1 μmol photons m−2 sec−1) in osmotically ruptured chloroplasts. +AA indicates addition of 5 μm antimycin A to the reaction before measurements. The increase in fluorescence reflects NDH-dependent PQ reduction in the presence of antimycin A.

In vivo evaluation of NDH function

In Arabidopsis and tobacco, absence of the chloroplast NDH complex scarcely affects photosynthetic electron transport in vivo. However, the transient increase in chlorophyll fluorescence that occurs after actinic light (AL) is turned off is suppressed in angiosperm mutants defective in chloroplast NDH (Burrows et al., 1998; Shikanai et al., 1998; Hashimoto et al., 2003). We are unsure of the exact mechanism by which chloroplast NDH induces this transient fluorescence. However, this fluorescence change is strongly correlated with the in vivo NDH activity estimated from the growth phenotype observed in the pgr5 mutant background, i.e. in double mutants (Munekage et al., 2004; Peng et al., 2009; Yamamoto et al., 2011). In wild-type Marchantia, a transient increase in chlorophyll fluorescence was observed in the dark after AL illumination, as in tobacco and Arabidopsis (Figure 6). However, it was absent in both the ΔndhB lines. Consistent with the results of the in vitro assay (Figure 5), NDH activity monitored during the transient chlorophyll change was impaired in the ΔndhB lines in Marchantia.

Figure 6.

 Monitoring of in vivo NDH activity by chlorophyll fluorescence analysis. A representative trace of chlorophyll fluorescence in the wild-type (WT) is shown. Thalli were exposed to AL (50 μmol photons m−2 sec−1) for 5 min, and the subsequent transient increase in chlorophyll fluorescence (boxed area) was monitored in the dark. The inset shows a magnified trace from the boxed area. inline image, maximum chlorophyll fluorescence; inline image, minimum chlorophyll fluorescence; ML, measuring light; SP, saturating light pulse of white light.

To evaluate the contribution of NDH-mediated electron transport to photosynthesis, the light intensity dependence of two chlorophyll fluorescence parameters, PSII yield (ΦPSII) and NPQ, were compared between the wild-type and the ΔndhB lines (Figure 7a,b). No major difference was observed in either parameter, suggesting that the contribution of chloroplast NDH to steady-state photosynthetic electron transport is subtle. Similar results have been obtained in tobacco and Arabidopsis mutants defective in chloroplast NDH (Shikanai et al., 1998; Hashimoto et al., 2003); these results contrast with the finding that an Arabidopsis pgr5 mutant defective in the main pathway of PSI cyclic electron transport exhibited severe mutant phenotypes with regard to both NPQ and ΦPSII at high light intensity (Munekage et al., 2002). However, ΦPSII was slightly affected at low light intensities of <100 μmol photons m−2 sec−1 in the ΔndhB lines (Figure 7a). To focus further on this weak phenotype in electron transport at low light intensities, the 1 – qL parameter, which reflects the level of reduction of the PQ pool, was compared between the wild-type and the ΔndhB lines (Figure 7c). The PQ pool was significantly more reduced in the ΔndhB lines than in the wild-type. In Marchantia, the chloroplast NDH complex may be required for the redox balance of the PQ pool, which is important at low light intensity.

Figure 7.

 Chlorophyll fluorescence parameters of wild-type and ΔndhB thalli.
(a) Quantum yield of PSII (ΦPSII), (b) NPQ and (c) proportion of closed centers of PSII (1−qL). Values are mean ± SD (= 4).


In the BN gel analysis of thylakoid membrane protein complexes in Marchantia, we did not detect any high-molecular-mass complex corresponding to the NDH–PSI super-complex of Arabidopsis (Figure 1). Consistently, our in silico survey did not identify Marchantia genes encoding Lhca5 or Lhca6 (Table 1). Lhca6 is believed to be specific to angiosperms, but Lhca5 was also detected in Physcomitrella patens (Alboresi et al., 2010; Neilson and Durnford, 2011). Although Lhca6 is specifically localized to the NDH–PSI super-complex in Arabidopsis, Lhca5 also interacts with PSI independently of NDH (Lucinski et al., 2006). Although Lhca6 is essential for stabilization of NDH in Arabidopsis, the contribution of Lhca5 is minor (Peng et al., 2011a,b). We consider that Lhca6 acquired a key role in formation of the NDH–PSI super-complex during the evolution of land plants.

Our EST or genome survey did not detect any strong candidates for functional orthologs of AtPnsL1–L4 (Table 1). An exception was PnsL5/CYP20-2, to which the Marchantia genome encodes a highly similar protein (Figure S2ds). PnsL5 is an unusual NDH subunit, and a defect in this subunit did not affect NDH stability or even activity, although PnsL5 is unstable in the absence of the other NDH subunits (Sirpiöet al., 2009) and has been detected in the NDH–PSI super-complex (Peng et al., 2009). PnsL5 may have an evolutional origin that differs from those of the other lumen subunits. As in the case of super-complex formation with PSI, our hypothesis is that the chloroplast NDH has been equipped with the entire lumen sub-complex, as observed in angiosperms, during the evolution of land plants.

In the Arabidopsis lhca5 lhca6 double mutant, chloroplast NDH does not interact with PSI and exists as a monomer (Peng et al., 2011a,b). The mobility of the protein complex including NdhH and NdhM in Marchantia appears to be faster than that of the NDH monomer in the Arabidopsis lhca5 lhca6 mutant (Figure 1). The molecular mass of this NDH monomer was estimated to be approximately 700 kDa in Arabidopsis (Peng et al., 2011a,b). The lack of genes encoding PnsL1–L4 may explain this difference in size. We hypothesize that, in Marchantia, the chloroplast NDH complex consists of the A, B, membrane and Fd-binding sub-complexes and does not interact with PSI. A possible exception is PnsL5/CYP20-2, which may be a component of NDH in Marchantia.

As reported in Arabidopsis (Munekage et al., 2004), Fd-dependent PQ reduction activity was partially impaired in ruptured chloroplasts of the ΔndhB lines (Figure 5). Because addition of antimycin A to the ΔndhB chloroplasts completely impaired PQ reduction activity in the presence of Fd, the remaining activity in the ΔndhB lines is likely to depend on PGR5-dependent, antimycin A-sensitive electron flow. This method does not provide quantitative information (Okegawa et al., 2008), but, as in Arabidopsis (Munekage et al., 2004), NDH-dependent and PGR5-dependent pathways explain almost all of the Fd-dependent PQ reduction activity in Marchantia (Figure 5). Unlike in Arabidopsis, addition of NADPH alone slightly reduced the PQ pool. This NADPH-dependent PQ reduction may depend on type II NDH, as in Chlamydomonas reinhardtii (Desplats et al., 2009). Puzzlingly, however, some of this activity was sensitive to antimycin A (Figure 5). The PQ pool may be reduced via the PGR5-dependent pathway with endogenous Fd and reverse reaction of Fd-NADP+ reductase (FNR). It is also possible that antimycin A non-specifically affected PQ reduction via type II NDH.

Even in the complete absence of ndhB, mutant thalli grew like the wild-type (Figure 3), as is the case in tobacco and Arabidopsis (Shikanai et al., 1998; Munekage et al., 2004). This is consistent with lack of a strong mutant phenotype with regard to chlorophyll fluorescence parameters (Figure 7). However, we detected a subtle but consistent phenotype with regard to the 1 – qL parameter: the PQ pool was reduced more at low light intensity (<100 μmol photons m−2 sec−1). Chloroplast NDH may be more important at low light intensity in Marchantia. This NDH function may not be specific to Marchantia, because a greater contribution of chloroplast NDH to PSI cyclic electron transport has been suggested to occur in rice under low light intensity (Yamori et al., 2011).

How does chloroplast NDH regulate the redox state of the PQ pool under low light intensity? Because NDH reduces the PQ pool (Okegawa et al., 2010) and the PQ pool was reduced more in the ΔndhB lines (Figure 7c), this mutant phenotype cannot be explained simply. In Arabidopsis, PGR5-dependent PSI cyclic electron transport is required for the balance of ATP/NADPH production; its defect causes a reduction in linear electron transport, especially at high light intensity (Munekage et al., 2002). However, this is unlikely to explain the phenotype observed with regard to 1 – qL, because the mutant thalli looked healthy under the light intensity used for growth conditions (50 μmol photons m−2 sec−1) (Figure 3) and also because the phenotype was not enhanced at higher light intensities (Figure 7). The exact mechanism by which chloroplast NDH balances the redox state of the PQ pool needs to be studied further.

In angiosperms, chloroplast NDH acts to alleviate oxidative stress under excessive light conditions (Shikanai, 2007). However, the phenotype of the ΔndhB lines of Marchantia was observed at low light intensity. Nevertheless, it is still probable that chloroplast NDH functions at high light intensity in Marchantia. Because the mutant phenotype of NDH-deficient mutants was observed in the pgr5 mutant background in Arabidopsis (Munekage et al., 2004), it would be informative to characterize the double mutant pgr5ΔndhB in Marchantia to assess the contribution of chloroplast NDH under the oxidative stress caused by the pgr5 defect. In Arabidopsis, NDH–PSI super-complex formation is required to stabilize NDH especially at high light intensity (Peng et al., 2011a,b). We propose that chloroplast NDH does not form a super-complex with PSI in Marchantia. A question remains as to how chloroplast NDH is stabilized under stress conditions in Marchantia.

Experimental Procedures

Plant materials and growth conditions

Male and female accessions of Marchantia polymorpha Takaragaike-1 (Tak-1) and Takaragaike-2 (Tak-2), respectively (Ishizaki et al., 2008), were grown on half-strength Gamborg’s B5 medium containing 1.2% agar in the presence or absence of 1% sucrose under continuous light of 50 μmol photons m−2 sec−1.

Database search

EST and genome sequences of M. polymorpha used in the database search include the dataset from previous analyses (Nagai et al., 1999; Nishiyama et al., 2000; Yamato et al., 2007), and the on-going Joint Genome Institute genome-sequencing project ( The data consist of over three million ESTs from thallus and sexual organs, and genome shotgun reads that cover approximately the nuclear genome 35-fold (280 Mb). Sequence alignments were generated using genetyx-mac version 12 (genetyx, Phylogenetic trees were constructed by the neighbor-joining method using the same software, with 1000 bootstrap trials.

Thylakoid membrane preparation, BN-PAGE and immunoblot analysis

Chloroplasts and thylakoids were isolated from Tak-1 thalli; the methods were essentially identical to those used in Arabidopsis (Munekage et al., 2002). BN-PAGE and subsequent 2D SDS–PAGE followed by immunoblot analysis were performed as previously described (Peng et al., 2008, 2009). For immunoblot analysis, thylakoid proteins were loaded on an equal chlorophyll basis. Signals were detected using ECL Prime Western blotting detection reagent (GE Healthcare, and visualized using an LAS3000 chemiluminescence analyzer (Fuji,

Antibody preparation

cDNAs encoding Marchantia NdhM (amino acids 73–226) and NdhH (amino acids 14–392) were amplified by RT-PCR and cloned into a pColdI expression vector (Takara Bio,, which adds a His tag at the N-terminus of expressed proteins. Expression of recombinant proteins was induced using 1 mm isopropyl thio-β-d-galactoside at 15°C for 24 h in the host Escherichia coli strain Rosetta (DE3) pLysS, and the recombinant proteins were purified by using HisTrap FF Crude (GE Healthcare) under denaturing conditions with 4 m urea. Purified NdhM and NdhH proteins were used to raise polyclonal antiserum in rabbits. Antibodies against PsaA and PsbA were purchased from Agrisera ( Antibodies against Arabidopsis PnsB1 and rice Cytf were kindly provided by T. Endo (Graduate School of Biostudies, Kyoto University, Japan) and A. Makino (Graduate School of Agricultural Science, Tohoku University, Japan), respectively.

Vector construction

A DNA fragment containing rps12, rps7, ndhB, psbM, trnL and the 5′ exon of ycf66 (positions 1–4904) was amplified from liverwort plastid DNA (Ohyama et al., 1986) using the primers 5′-TTGAACAGAAGCCGTATGAAATG-3′ and 5′-TAGGATCTAAGCGCCAACCCTG-3′, and then cloned into the pTAC-2 vector (Biodynamic Laboratory, Cloning reactions were performed using a DynaExpress TA cloning kit (BioDynamics Laboratory). The aadA cassette that confers spectinomycin resistance on chloroplasts (Shikanai et al., 1998) was inserted into the SpeI site that is present in the 3′ exon of ndhB so that the gene was inactivated. The resulting plasmid was digested using NotI to linearize it for plastid transformation.

Plastid transformation and genome analysis

Spores used for transformation were prepared as described by Chiyoda et al. (2008). Plastids were transformed using 7-day-old sporelings, as described previously (Shikanai et al., 1998; Chiyoda et al., 2008) with minor modifications. Sheets of thalli were bombarded using a biolistic delivery system (Bio-Rad, Bombarded cells were incubated overnight under continuous light (50–60 μmol photons m−2 sec−1) at 20°C, then divided and spread evenly onto four sucrose-free selective 0-M51C agar medium plates containing 300 mg l−1 spectinomycin. Homoplasmic lines were obtained by asexual reproduction through gemma on non-selective 0-M51C agar medium without spectinomycin.

Total DNA was isolated from the thalli using Isoplant II (Nippon Gene, The genome was analyzed by PCR using primers Pr1 (5′-CACACAGATAAGCTCACGCTAAC-3′), Pr2 (5′-CACTTGTTGTATGGGGACGAAGTGG-3′), Pr3 (5′-CCACTACGTGAAAGGCGAGATCAC-3′) and Pr4 (5′-CTTCGGCGATAACCGCTTCACGAG-3′).

Chlorophyll fluorescence analysis

Chlorophyll fluorescence from 10-day-old Tak-1 thalli (wild-type, ΔndhB16-5 and ΔndhB20-1) was measured using a MINI-PAM portable chlorophyll fluorometer (Walz, The transient increase in chlorophyll fluorescence after AL was turned off was monitored as previously described (Shikanai et al., 1998). To investigate the dependence of ΦPSII, NPQ and 1 – qL on light intensity, measuring light (650 nm, 0.1 μmol photons m−2 sec−1) was used to induce the minimum fluorescence at open PSII centers in the dark-adapted state (Fo). A saturating pulse of white light (0.8 sec, 8000 μmol photons m−2 sec−1) was applied to determine the maximum fluorescence at closed PSII centers in the dark (Fm) or during illumination (inline image). The steady-state fluorescence level (Fs) was recorded during AL illumination (15–1000 μmol photons m−2 sec−1). These photosynthetic parameters were recorded 2 min after the change in AL intensity. ΦPSII was calculated as inline image. NPQ was calculated as inline image. qL, the fraction of the open PSII center, was calculated as inline image (Miyake et al., 2009). Fd-dependent PQ reduction activity was measured in ruptured chloroplasts (20 μg chlorophyll ml−1) as previously described (Okegawa et al., 2008). As electron donors, 5 μm spinach Fd (Sigma-Aldrich, and 0.25 mm NADPH (Sigma-Aldrich) were used. Antimycin A (Sigma-Aldrich) at a concentration of 5 μm was added before measurement. Fluorescence levels (Fo’) were normalized against Fm levels.

Accession numbers

The EMBL accessions for Marchantia EST sequences are as follows: HE855888 (ndhL), HE855889 (ndhM), HE855890 (ndhN), HE855891 (ndhO), HE855892 (ndhS), HE855893 (ndhT), HE855894 (ndhU), HE855895 (pnsB1), HE855896 (pnsB2), HE855897 (pnsB3), HE855898 (pnsB4), HE855899 (pnsB5), HE855900 (ppl1), HE855901 (psbQ), HE855902 (flbp13), HE855903 (pnsL5), HE855904 (lhca1), HE855905 (lhca2), HE855906 (lhca3) and HE855907 (lhca4).


T.S. was supported by grant number 22247005 from the Ministry of Education, Culture, Sports, Science and Technology of Japan, the Strategic International Research Cooperative Program from Japan Science and Technology Agency, and by a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Genomics for Agricultural Innovation; GPN0008). M.U. was supported by a research fellowship from the Japan Society for the Promotion of Science for young scientists (21-764). Y.N. was supported by the Funding program for Next Generation World-Leading Researchers (NEXT Program).