DNA binding and partial nucleoid localization of the chloroplast stromal enzyme ferredoxin:sulfite reductase


  • Note Nucleotide sequence data for PsSiR are available in the DDBJ/EMBL/GenBank databases under accession number AB168112

N. Sato, Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo 153–8902, Japan
Fax: +81 3 5454699
Tel: +81 3 54546631
E-mail: naokisat@bio.c.u-tokyo.ac.jp


Sulfite reductase (SiR) is an important enzyme catalyzing the reduction of sulfite to sulfide during sulfur assimilation in plants. This enzyme is localized in plastids, including chloroplasts, and uses ferredoxin as an electron donor. Ferredoxin-dependent SiR has been found in isolated chloroplast nucleoids, but its localization in vivo or in intact plastids has not been examined. Here, we report the DNA-binding properties of SiRs from pea (PsSiR) and maize (ZmSiR) using an enzymatically active holoenzyme with prosthetic groups. PsSiR binds to both double-stranded and single-stranded DNA without significant sequence specificity. DNA binding did not affect the enzymatic activity of PsSiR, suggesting that ferredoxin and sulfite are accessible to SiR molecules within the nucleoids. Comparison of PsSiR and ZmSiR suggests that ZmSiR does indeed have DNA-binding activity, as was reported previously, but the DNA affinity and DNA-compacting ability are higher in PsSiR than in ZmSiR. The tight compaction of nucleoids by PsSiR led to severe repression of transcription activity in pea nucleoids. Indirect immunofluorescence microscopy showed that the majority of SiR molecules colocalized with nucleoids in pea chloroplasts, whereas no particular localization to nucleoids was detected in maize chloroplasts. These results suggest that SiR plays an essential role in compacting nucleoids in plastids, but that the extent of association of SiR with nucleoids varies among plant species.


pea (Pisum sativum) sulfite reductase


sulfite reductase


maize (Zea mays) sulfite reductase

The assimilation of sulfur is an important process for the synthesis of various sulfur compounds such as amino acids, sulfolipids, and coenzymes. Sulfite reductase (SiR) is a central enzyme within the sulfur assimilation pathway. Sulfate ions taken up by the sulfate transporter are first activated with ATP by ATP sulfurylase, forming adenosine-5′-phosphosulfate. Adenosine-5′-phosphosulfate is further phosphorylated by adenosine-5′-phosphosulfate kinase, forming 3′-phosphoadenosine-5′-phosphosulphate. 3′-Phosphoadenosine-5′-phosphosulfate is reduced to sulfite by 3′-phosphoadenosine-5′-phosphosulfate reductase, and sulfite is further reduced to sulfide by SiR. The resultant sulfide is fixed into cysteine by cysteine synthase using O-acetylserine as an acceptor. SiR is localized to chloroplasts in green leaves and to nongreen plastids in nonphotosynthetic tissues. SiR has been identified as one of the main constituents of plastid nucleoids in pea [1] and soybean [2]. Chloroplast DNA was previously thought to occur dissolved in the stroma, but recent studies have revealed that the functional form of chloroplast DNA is a DNA–protein complex called a nucleoid [3]. Plant SiR contains a siroheme and a [4Fe-4S] cluster and catalyzes the six-electron reduction of sulfite to sulfide, depending on ferredoxin as an electron donor [4]. Plant SiR was considered to be a stromal protein [5–7], but localization to plastid nucleoids provides a new aspect of this enzyme and chloroplast molecular biology.

Plastid nucleoids are comprised of various proteins [3,8–13], and the protein composition changes during plastid development. Several plastid nucleoproteins have been studied. CND41 [14–17] is a bifunctional protease, considered to be a negative regulator of plastid gene expression. The PEND protein [18–21] is thought to anchor nucleoids to the envelope membrane in developing chloroplasts. MFP1 [22,23] binds to the thylakoid membrane in mature chloroplasts and is thought to function in a manner similar to PEND in developing chloroplasts. In a proteomic analysis of DNA–protein complexes, substantially similar to what we call nucleoids here, polypeptides derived from various enzymes were reported in addition to those involved in transcription, DNA replication, DNA topology, and DNA binding. These were iron superoxide dismutase, putative thioredoxin, pfkB-type carbohydrate kinase family protein, and Mur ligase family protein in Arabidopsis and mustard [24], and proteins involved in pyruvate dehydrogenase, acetyl-CoA carboxylase, ATP synthase, ribulose bisphosphate carboxylase, and Calvin cycle proteins in pea [25]. Different researchers use different criteria to judge if unexpected proteins are the result of contamination. Therefore, it is important to confirm either the direct or indirect DNA-binding properties of these putative components of nucleoids in vitro.

In previous studies of the role of SiR in chloroplast nucleoids, SiR was suggested to repress DNA synthesis [26] and transcription [27] within nucleoids. Relaxing the DNA compaction of nucleoids by release of SiR activates transcription, whereas increasing DNA compaction by the addition of exogenous SiR represses transcription. These observations led us to propose that transcription regulation occurs through DNA compaction by SiR in the plastid nucleoids [27]. The localization of SiR to nucleoids was confirmed by immunofluorescence microscopy of isolated pea and soybean chloroplasts, showing that SiR localizes to nucleoids within chloroplasts [2].

In our previous studies of the role of SiR in DNA compaction in nucleoids, we used maize SiR (ZmSiR), the sole recombinant holoenzyme successfully prepared at that time. Studies of SiR in other plants were forced to use recombinant, but enzymatically inactive, proteins lacking prosthetic groups. However, we have been using pea to isolate chloroplast nucleoids because a large amount of chloroplast nucleoid is obtained efficiently. Here, we isolated the cDNA of pea SiR (PsSiR), deduced the structure of PsSiR and prepared a recombinant holoenzyme of PsSiR. Using the active recombinant SiR, we examined basic DNA-binding properties of SiR and the relationship between SiR activity and DNA-binding activity. We also examined differences in the properties of PsSiR and ZmSiR, such as DNA binding and localization within chloroplasts.


Isolation and molecular characterization of a cDNA clone encoding pea SiR

The λ gt10 pea cDNA library was screened. One positive plaque was obtained from recombinant phages, and the insert was cloned to pCR2.1-TOPO. The insert, designated Seq1, had a length of 2216 bases and contained a reading frame encoding 663 amino acids, although the initiation codon was missing. The amino-terminal sequence of SiR from the pea plastid nucleoid, ‘VSTPAKS’[1], was found within the deduced amino acid sequence (Fig. 1). To determine the missing region, the λ gt11 pea cDNA library was screened. Three clones were obtained. The longest nucleotide sequence, designated Seq2, contained the other three inserts. The overlapping regions of Seq1 and Seq2, which had a length of 275 bases, matched completely. A putative initiation codon was found in the extended region. The sequence assembled from Seq1 and Seq2, designated PsSiR, therefore encodes the precursor of pea SiR consisting of 685 amino acids. Two candidates for a poly(A) signal, ATAAA and ATAAT, were found 12 bases and 55 bases upstream of the poly(A) start site, respectively (data not shown).

Figure 1.

 Comparison of the amino acid sequences of different SiRs. The precursor sequence of PsSiR is compared with those predicted from the cDNA of N. tabacum (NtSiR; GenBank accession number D83583), A. thaliana (GenBank accession number BT000593), Z. mays (ZmSiR; GenBank accession number D50679), O. sativa (OsSiR; GenBank accession number AK103289), C. merolae SiRA (CmSiRA), SiRB (CmSiRB), Anabaena sp. PCC7120 (AnSiR; CyanoBase accession number alr1348), Synechocystis sp. PCC6803 (CyanoBase accession number slr0963), and CysI, the hemoprotein of E. coli NADPH-SiR (EcCysI; GenBank accession number M23007). Residues similar to PsSiR are indicated by white letters on a black background. The N-terminal amino acid sequence ‘VSTPAKS’ is marked by an arrow. The amino acid residues that bind to the substrate are indicated by R and K. The cysteine residues predicted to be ligands for the [4Fe-4S] cluster and the siroheme are indicated by C.

The amino acid sequence deduced from PsSiR exhibited significant homology with SiRs from various other organisms. The amino acid sequences of SiRs from angiosperms, a red alga, and two cyanobacteria were aligned with the hemoprotein subunit (CysI) of the Escherichia coli SiR complex, which is another type of SiR using NADPH as an electron donor (Fig. 1). The similarity of the putative mature PsSiR was 91% to Nicotiana tabacum SiR, 85% to Arabidopsis thaliana SiR, 79% to Zea mays SiR, 79% to Oryza sativa SiR, 59% to Cyanidioschyzon merolae SiRA, 57% to SiRB, 67% to Anabaena sp. PCC7120 SiR, 66% to Synechocystis sp. PCC6803 SiR, and 48% to E. coli CysI. The three-dimensional structure of E. coli CysI determined using X-ray crystallography [28] revealed that the siroheme and the [4Fe-4S] cluster are retained within the active site of the enzyme through four cysteine ligands, Cys434, Cys440, Cys479, and Cys483, and that four basic residues, Arg83, Arg153, Lys215, and Lys217, are involved in the substrate coordination to siroheme. These residues are completely conserved in all SiRs.

To estimate the molecular phylogeny of the SiRs, we constructed phylogenetic trees using the maximum likelihood method, using treefinder. Essentially similar trees were obtained by the neighbor-joining and maximum parsimony methods. All SiRs of flowering plants were monophyletic and the SiRs of plants originated from cyanobacterial SiR (Fig. 2). The reliability of the tree was confirmed by bootstrapping. The important nodes for the above statement were supported with high confidence levels.

Figure 2.

 Phylogenetic tree of SiRs based on amino acids sequences. The tree was constructed using the maximum likelihood method. The three numbers on each branch show confidence levels for the maximum likelihood/neighbor-joining/maximum parsimony analyses, estimated by bootstrap analysis with 1000 replicates.

Basic characterization of the PsSiR gene

cDNA gel blot analysis was carried out with pea genomic DNA restricted with EcoRI or HindIII. A single band was detected in both of the digests (Fig. 3D), suggesting a single PsSiR gene per genome. To examine the expression level of pea SiR, RNA gel blot (Fig. 3A,B) and immunoblot (Fig. 3C) analyses were carried out using leaves, stems, and roots. Comparable amounts of SiR transcripts and proteins were detected in the three organs (Fig. 3A,C). The transcript expression level was not significantly different in light- and dark-grown leaves (Fig. 3B). These results suggest that the expression of SiR is constitutive in pea, in agreement with previous studies in tobacco [29], maize [30], and A. thaliana[31].

Figure 3.

 Basic characterization of PsSiR. RNA blot analyses of the expression of PsSiR in (A) various organs of pea plants and (B) pea leaves grown in light or dark. Total RNA (10 µg) prepared from green leaves, stems, or roots, and polyA + RNAs (3 µg) prepared from leaves grown in light or dark were electrophoresed on 1.2% agarose gel. The blots were probed with a DIG-labeled DNA fragment corresponding to the second exon of PsSiR. Lower panels indicate the staining of (A) 28S rRNA and (B) blotted mRNA of pea β-tubulin 1 as controls. (C) Immunoblot analysis of the distribution of PsSiR in various organs. Total extracts of 5 mg fresh weight of green leaves, stems, or roots in NaCl/Pi were separated on 10% gel by SDS/PAGE. The blots were probed with antibodies raised against PsSiR. As a loading control, the staining pattern of a 95-kDa major band (putative heat shock protein) is shown below. (D) DNA blot analysis of the pea genome. Genomic DNA of pea was digested with EcoRI and HindIII and separated on 0.8% agarose gel. The blots were probed with the same probe used for RNA blotting.

The presence of SiR in chloroplast nucleoids was detected previously using immunological and enzymological methods. To confirm that the cloned PsSiR was actually present within the chloroplast nucleoid, the band corresponding to PsSiR on the SDS/PAGE of chloroplast nucleoids from pea leaves was analyzed by MALDI-TOF/MS after in-gel digestion with lysyl endopeptidase or endoproteinase Asp-N (AspN) (Table S1). In lysyl endopeptidase digestion, 13 peptides corresponding to the predicted digestion products of PsSiR were detected. The sequence coverage was 30.3% of the mature protein. In AspN digestion, six peptides were detected with 13.9% sequence coverage. Importantly, the AspN digestion gave peptides corresponding to the first 18 residues of the N terminus and the last 16 residues of the C terminus of the predicted mature protein. No peptide corresponding to the predicted transit peptide sequence was detected. Because AspN cleaves peptide bonds at the N-terminal side of aspartic and glutamic acids under our conditions, the 18-mer peptide beginning with the N-terminal valine should not have resulted from enzymatic digestion with AspN. This result clearly indicates that PsSiR exists in chloroplast nucleoids and that it consists of the full-length mature polypeptide beginning with the valine residue.

Preparation of recombinant SiR and its DNA-binding activities

ZmSiR was overproduced in E. coli cells under coexpression of siroheme synthase and purified by a combination of three successive chromatographies on ion-exchange, hydrophobic, and ferredoxin-affinity resin, as described previously [32]. PsSiR was also produced in a similar way and could be purified by ferredoxin-affinity chromatography only. The final preparation of PsSiR showed a UV-visible absorption spectrum characteristic of siroheme-containing proteins and gave a single major band of around 70 kDa in SDS/PAGE (Fig. 4A,B).

Figure 4.

 (A) UV-visible absorption spectrum and (B) SDS/PAGE analysis of purified recombinant PsSiR. (A) Absorption maxima at 389 and 580 nm (arrows) indicate the presence of a siroheme-containing prosthetic group. (B) A single band (arrowhead) was observed by staining with Coomassie brilliant blue.

This preparation was used to examine the DNA-binding activity of SiR. A gel-mobility shift assay was carried out using radiolabeled 40-mer or 20-mer synthetic dsDNA as a probe (Fig. 5A,B). Various amounts of recombinant PsSiR and 40-mer or 20-mer dsDNA were mixed, and then the mixtures were electrophoresed. The intensity of the shifted bands (complexes of DNA and PsSiR) increased with the amount of PsSiR in both experiments. The most retarded band was very close to the origin, indicating that a large complex was formed (arrowheads in Fig. 5). The band could represent insoluble materials stuck on the upper edge of the gel. But the materials did enter agarose gels (Fig. 6B,C) and are not insoluble materials. The apparent dissociation constants (Kd) of 40-mer and 20-mer dsDNA were about 55 nm and 142 nm, respectively, which indicates a high affinity of PsSiR for long DNA. The recombinant PsSiR also shifted ssDNA (Fig. 5C).

Figure 5.

 DNA binding activity of SiRs. The 32P-labeled (A) 40-mer dsDNA, (B) 20-mer dsDNA, and (C) 40-mer ssDNA were incubated without (lanes 1 and 10) or with 25, 50, 100, and 200 nm PsSiR (lanes 2–5) or ZmSiR (lanes 6–9) before electrophoresis in 6% polyacrylamide gel. FD indicates free DNA. The bands very close to wells, which suggest large complexes, are indicated by arrowheads. The slightly shifted bands detected in ssDNA with PsSiR are indicated by an asterisk. The apparent dissociation constants (Kd) are shown below the lane numbers.

Figure 6.

 Sequence specificity for DNA-binding by SiRs. (A) Competition for DNA-SiR complex formation with poly(dI-dC)·poly(dI-dC). The 32P-labeled 20-mer dsDNA and none, two-, five-, or 10-fold mass excess of nonlabeled poly(dI-dC)poly(dI-dC) (lanes 2–5, respectively) were mixed with 100 nm PsSiR prior to electrophoresis in 6% polyacrylamide gel. The bands shifted by PsSiR and faded by the competitor are indicated by arrowheads. Lane 1 is DNA alone. (B) Binding of SiR to StyI-digested λ phase DNA (lanes 1–7) and XbaI-digested chloroplast DNA (lanes 8–14). Each digested DNA was incubated without (lanes 1 and 8) or with 200, 400, and 800 nm PsSiR (lanes 2–4 and 9–11) or ZmSiR (lanes 5–7 and 12–14) prior to electrophoresis in 1% agarose gel.

The sequence specificity of SiR during DNA binding was examined. Non-labeled poly(dI-dC)·poly(dI-dC) was added to the mixture of labeled 20-mer dsDNA and PsSiR as a competitor, and then the mixture was electrophoresed (Fig. 6A). The densities of shifted band signals at two-, five- and 10-fold excess of competitor were 34, 17 and 12%, respectively, of that without a competitor, indicating comparable affinities of SiR for the 20-mer dsDNA and the poly(dI-dC). The λ DNA digested with StyI and pea chloroplast DNA digested with XbaI were mixed with PsSiR and electrophoresed (Fig. 6B). All fragments were shifted and concentrated in a few slowly migrating bands in the vicinity of the wells. These results suggest that SiR binds to DNA with low sequence specificity.

Effects of DNA binding on sulfite reductase activity

We measured the sulfite reductase activity of the recombinant PsSiR by a cysteine synthase-coupled system using recombinant maize ferredoxin I as an electron donor for SiR and dithionite as a reductant for ferredoxin I (Fig. 7). The Michaelis constant (Km) of PsSiR for maize ferredoxin I was about 18 µm, higher than that of ZmSiR [33], which is about 4 µm. Unlike PsSiR, the activity of ZmSiR was assayed by measuring the increase in NADP+ oxidized by reduction of ferredoxin III donating electrons to SiR as described by Yonekura-Sakakibara et al. [30]. The difference in Km may be due to the differences in the methods used to measure activity.

Figure 7.

 Measurement of enzymatic activity of DNA-bound PsSiR. PsSiR was incubated without (open circles with solid thin line) or with 10 µg·mL−1 (filled squares with bold solid line) or 20 µg·mL−1 (filled triangles with dashed line) HindIII digested λ DNA before measurement. Various concentrations of recombinant maize ferredoxin I were added as an electron donor to SiR. The amount of cysteine produced per minute per mole of SiR was used as a measure of enzymatic activity.

To examine the effects of DNA binding on sulfite reductase activity, recombinant PsSiR was mixed with DNA to form a DNA–SiR complex, and then enzymatic activity was measured (Fig. 7). There was no significant difference in activity of DNA-bound and DNA-free PsSiR. This indicates that sulfite reductase remains functional when it is bound to DNA. Binding of PsSiR to DNA in the reaction mixture during the measurement of activity was confirmed by an experiment of coprecipitation of SiR with DNA-cellulose in the reaction medium (data not shown).

Characteristics of pea SiR with respect to maize SiR

We previously used recombinant ZmSiR in studies of the DNA binding of sulfite reductase [1,27]. Here, we compared the DNA-binding activity of ZmSiR and PsSiR (Fig. 5, right). An increase in the intensity of the shifted bands was found as the concentration of ZmSiR increased with 40-mer ds-, 20-mer ds- or 40-mer ssDNA. However, the apparent Kd values of ZmSiR were higher than those of PsSiR, suggesting that the binding of ZmSiR to DNA is weaker than that of PsSiR. It should be noted that very slowly migrating bands, as detected with PsSiR, were scarcely detected with ZmSiR. This result suggests that ZmSiR has a lower ability to compact DNA than PsSiR. An additional rapidly migrating band was detected with ssDNA and PsSiR (asterisk in Fig. 5C).

Differences in the ability to compact DNA were also examined for PsSiR and ZmSiR. Changes over time were examined by fluorescence microscopy after mixing DNA with PsSiR or ZmSiR. Immediately after mixing, a number of blurred particles were formed upon addition of PsSiR, whereas either ZmSiR or buffer alone did not engender such particles (Fig. 8). After incubation for >2 h, bright and structurally well-defined particles were formed with PsSiR. Some particles increased in brightness over time, indicating that the quantity of DNA per particle increased and/or DNA became more tightly compacted. After incubation for more than 6 h, blurred particles similar to those at 0 h in PsSiR were formed in ZmSiR. These results suggest that the DNA-compacting activity of ZmSiR is weaker than that of PsSiR.

Figure 8.

 Compaction of chloroplast DNA by SiRs. Isolated pea chloroplast DNA was incubated with recombinant PsSiR (A, B, C, D, E) or ZmSiR (F, G, H, I, J) on ice. The mixture was stained with 4′,6-diamidino-2-phenylindole and examined by fluorescence microscopy immediately (A, F) or at 1 h (B, G), 2 h (C, H), 6 h (D, I), or 20 h (E, J) after mixing. Chloroplast DNA incubated in buffer alone was also examined immediately (K) or at 20 h (L) after mixing as a control.

We previously performed an in vitro transcription assay using isolated chloroplast nucleoids and showed that UTP incorporation into RNA was repressed by the addition of recombinant ZmSiR [27]. We expected that PsSiR is more active in repressing the transcriptional activity of chloroplast nucleoids. Isolated pea chloroplast nucleoids were mixed with various concentrations of recombinant PsSiR or ZmSiR and incubated for 30 min on ice before the addition of radiolabeled UTP to initiate run-on transcription (Fig. 9, closed and open circles, respectively). The incorporation of radioactive UTP into the high molecular weight fraction was used as a measure of transcriptional activity. In Fig. 9, the activity is expressed in percentage of the control activity in the presence of 100 µg·mL−1 heparin. In a previous paper [27], we showed that addition of heparin releases the endogenous SiR from nucleoids, and the full transcriptional activity is obtained. When ZmSiR was used, transcriptional activity gradually decreased as ZmSiR concentration increased. The activity decreased by about one-half at a ZmSiR concentration of 1.6 µm (not shown). This result is consistent with the results of our previous experiment [27]. When PsSiR was added at a concentration of 0.2 µm, the transcriptional activity of the nucleoids was reduced dramatically to about 20% of that in the absence of exogenous SiR. Therefore, PsSiR more strongly repressed transcription than did ZmSiR. This is consistent with the higher affinity of PsSiR for DNA.

Figure 9.

 Effects of SiRs on the transcriptional activity of nucleoids from mature and developing chloroplasts. (A) The isolated pea chloroplast nucleoids were incubated with various concentrations of PsSiR (solid line) or ZmSiR (dashed line) on ice for 30 min and then added to the reaction mixture. After preincubation at 25 °C for 30 min, 3H-labeled UTP was added to start the reaction and the mixture was incubated at 25 °C for 30 min. The count rate (cpm) of radioactive UTP incorporated into the high-molecular weight fraction was measured. The transcriptional activity is expressed as a percentage of the activity in the presence of 100 µg·mL−1 heparin, which releases all endogenous SiR and makes the nucleoids fully active in transcription. We used such measure because the amount of cpDNA, as estimated by Southern blotting with rbcL as a probe, may not be very accurate for comparing different samples. The actual UTP incorporation in the nucleoids from mature and developing chloroplasts was 601 ± 83 and 824 ± 113 cpm·ng−1 cpDNA, respectively. ○,•, nucleoids from mature pea chloroplasts (14-day-old leaves); □,▪, nucleoids from developing pea chloroplasts (6-day-old leaf buds).○,□, addition of maize SiR; •,▪, addition of pea SiR. (B) A schematic view on the compaction status of nucleoids in developing and mature chloroplasts. In developing chloroplasts, a large amount of SiR is bound to the nucleoids and represses the transcription severely. In mature chloroplasts, the amount of SiR is reduced and the nucleoids are more active in transcription.

Figure 9 also shows comparison of nucleoids from mature (14-day-old) and developing (6-day-old) leaves. As previously reported [10,18], the leaves of 6-day-old seedlings are small buds that are pale green and are not yet open. The developing chloroplasts in such leaves contain a higher amount of chloroplast DNA, but are not active in photosynthesis. The results in Fig. 9 show that the nucleoids from developing leaves (closed square at zero SiR concentration) are less active in transcription than those from mature leaves (closed circle) due to stronger repression by endogenous SiR. This is consistent with the previous results of immunoblot [10]. The effects of exogenous SiR, either from pea or maize, were similar in both mature and developing nucleoids.

Intrachloroplast localization of SiR

To examine the intrachloroplast location of SiR, indirect immunofluorescence microscopy of isolated chloroplasts was performed (Fig. 10). Isolated pea or maize chloroplasts fixed in paraformaldehyde were incubated with an antibody raised against PsSiR and then with an AlexaFluor-tagged secondary antibody. The chloroplasts were also counterstained with 4′,6-diamidino-2-phenylindole to visualize chloroplast DNA. In pea chloroplasts, the AlexaFluor signal was detected nonuniformly within the chloroplast. The spots that were densely stained with AlexaFluor coincided with the areas of 4′,6-diamidino-2-phenylindole staining. These results indicate that PsSiR exists within nucleoids, as well as in the stroma. In maize chloroplasts, the AlexaFluor signal was located uniformly throughout the whole chloroplast and no dense AlexaFluor signal was colocalized with a 4′,6-diamidino-2-phenylindole signal. This indicates that ZmSiR is not particularly concentrated in nucleoids.

Figure 10.

 Intrachloroplast localization of SiR. Isolated pea (A, B, C, D) or maize (E, F, G, H) chloroplasts were fixed with paraformaldehyde, probed with antibodies raised against PsSiR (A, B, E, F) or preimmunized serum (C, D, G, H) and AlexaFluor488-tagged secondary antibodies, and stained with 4′,6-diamidino-2-phenylindole. Left-hand panels (A, C, E, G) display 4′,6-diamidino-2-phenylindole and chlorophyll fluorescence. Right-hand panels (B, D, F, H) display AlexaFluor488 fluorescence. The bar indicates 10 µm.


Comparative aspects of SiR

The sequence alignment (Fig. 1) indicates that PsSiR has the archetypal molecular structural characteristics of plant ferredoxin-dependent SiRs, i.e. four cysteines as ligands for the prosthetic groups and two lysines and two arginines involved in the substrate coordination to siroheme are completely conserved. Region A in Fig. 1 indicates the common insertion in ferredoxin-dependent SiRs and ferredoxin-dependent nitrite reductases, with respect to E. coli CysI, and is reported as a candidate for an interaction site with ferredoxin [29]. PsSiR retains this insertion from Ala234 to Phe261 with high homology, as expected. However, region A sequences in C. merolae SiRA and SiRB are poorly conserved, with an additional insertion: YWK(R/K)(D/E)(I/L). It will be interesting to determine whether SiRA and SiRB have affinity for ferredoxin.

The phylogenetic tree (Fig. 2) indicates that the SiRs of cyanobacteria and plants (using ferredoxin as an electron donor) originate from bacterial SiRs (using NADPH as an electron donor via the flavin subunit). The plant SiRs originate from cyanobacterial SiRs, with Gloeobacter as the root. However, cyanobacterial SiRs are divided into two clades, one consisting of Synechococcus and Prochlorococcus, and the other consisting of Anabaena–Nostoc, Synechocystis, and Thermosynechococcus. These two clades correspond to the two major lineages of cyanobacteria [34]. Phylogenetic analysis with plastid-encoded protein genes suggested that plastids originate from the AnabaenaSynechocystis clade. However, plant SiRs are associated with the SynechococcusProchlorococcus clade, although the confidence level of the branches near Synechococcus sp. PCC 6301 is low (Fig. 2). The angiosperm SiRs form a monophyletic cluster distinct from the SiRs of cyanobacteria or C. merolae. The C. merolae SiRs and Thalassiosira (diatom) SiR are monophyletic, which suggests that the SiR gene was transferred from a red algal endosymbiont during secondary endosymbiosis. In C. merolae, an ORF (CMR 440 C) homologous to the α-component (flavoprotein) of NADPH-dependent SiR is also found [35]. One of the SiRs in C. merolae (possibly SiRB) could function with this flavoprotein, rather than ferredoxin.

The genomic DNA blot analysis (Fig. 3D) suggests that there is a single SiR gene in the haploid pea genome. A single SiR gene is also found in A. thaliana[36], rice [37], and N. tabacum[29]. Another copy must be present in tobacco because it is an amphidiploid. Except in tobacco, the SiR gene occurs as a single-copy gene in all known flowering plants. SiR was known as a stromal enzyme before it was found localized to plastid nucleoids [1,2]. The nucleoid localization of SiR could have been explained by an isozyme encoded by a different gene, but the copy number analysis indicates that this is not likely.

Reductase activity and DNA binding

Here, the large-scale production of enzymatically active recombinant SiR containing prosthetic groups enabled detailed experiments on the relationship between enzymatic activity and DNA-binding activity.

The DNA gel-mobility shift assay using recombinant SiR revealed that SiR is an authentic DNA-binding protein, with high DNA-binding affinity. Our data demonstrate that SiR directly binds to dsDNA, as well as to ssDNA (Fig. 5). This suggests that SiR binds to DNA during replication, which may cause the repression of DNA synthesis by SiR in isolated nucleoids [26]. Radiolabeled DNA and poly(dI-dC)·poly(dI-dC) competed comparably for binding to SiR, and all the restriction fragments of both λ and chloroplast DNA were shifted by SiR binding (Fig. 6). This shows that SiR binds to DNA without notable sequence specificity and supports our argument that SiR is a global regulator of nucleoid functions such as transcription [27] and replication [26].

Intrachloroplast localization

Indirect immunofluorescence microscopy of isolated chloroplasts demonstrated the presence of PsSiR in the nucleoids of pea chloroplasts. Our data are basically consistent with previous results [2]. In the previous study, however, the fluorescence signal of SiR clearly coincides with DNA in pea chloroplast nucleoids, and essentially no fluorescence was detected in the stroma [2]. We found an SiR signal throughout the whole chloroplast and in dense patches that coincided with the nucleoids. Chi-Ham et al. [2] fixed chloroplasts in a buffer containing formaldehyde and then dehydrated them with ethanol and acetone on slides, whereas we used formaldehyde fixation without dehydration and performed immunoreaction and 4′,6-diamidino-2-phenylindole staining in a test tube. We suspect that the stromal components were washed away during the washing and dehydration process in the experiments of Chi-Ham et al. [2]. The mild processing in our experiments made the localization of SiR slightly obscure, but this indicates that SiR is not confined to the nucleoids and is also present in the stroma.

Comparison of PsSiR and ZmSiR

In the gel-mobility shift assay of PsSiR, shifted bands that remained very close to the origin were detected. These bands represent potentially large DNA–SiR complexes formed by the intermolecular aggregation of DNA fragments. The large complexes are only found with PsSiR, indicating high DNA-compacting ability. In contrast, ZmSiR formed no such complex, indicating that the DNA-compacting ability of ZmSiR is inferior to that of PsSiR.

The difference in DNA-compacting ability between PsSiR and ZmSiR was clearly demonstrated by the in vitro compaction assay (Fig. 8). Previously, we showed that nucleoid-like particles were formed only several hours after the mixing of recombinant ZmSiR and pea chloroplast DNA [1]. Our current results reproduced the previous findings, but the compaction by ZmSiR was not as tight as that found in the previous study because we did not add spermidine in the present study. Nevertheless, PsSiR did form blurred particles immediately after mixing, and well-defined particles were eventually formed, demonstrating the even higher DNA-compaction ability of PsSiR than of ZmSiR.

The two SiRs had different effects on the transcriptional activity of nucleoids. PsSiR repressed the transcriptional activity of chloroplast nucleoids more strongly than did ZmSiR (Fig. 9). We previously reported that ZmSiR reversibly repressed transcriptional activity by enhancing DNA compaction. However, PsSiR had far stronger transcriptional repression than ZmSiR. Based on the results of the DNA-binding study, the tighter compaction of DNA induced by PsSiR was a result of its stronger binding to DNA.

The results of immunofluorescence microscopy need additional explanation. In the isolated maize chloroplasts, SiR was detected throughout the whole chloroplast and the distribution was almost uniform. This result was different from that in pea chloroplasts. In developing soybean chloroplasts [2], similarly diffuse SiR-indicative fluorescence was observed. This was explained by the differentiation status of cells because fluorescence coincided with less-defined nucleoids. However, our result cannot be explained by the same reasoning because the maize chloroplasts contained well-defined nucleoids. Because ZmSiR clearly has DNA-binding activity, SiR may partly associate with nucleoids in maize chloroplasts. This association may not be detected clearly by immunofluorescence because of the presence of SiR within the stroma. In proteomic mass spectroscopy, a number of peptides, including SiR, were identified from the Triton-insoluble fraction prepared from pea chloroplasts [25], whereas only a few peptides that do not completely satisfy identification criteria were identified from transcriptionally active chromosomes purified from Arabidopsis and mustard chloroplasts [24]. The nucleoid fraction, which was prepared from Arabidopsis chloroplasts in the same way, did not contain a detectable amount of SiR (unpublished results). This could be due to treatment with a high concentration of detergent during nucleoid preparation. These different results in different plants can be explained by differences in the affinity of SiRs to DNA.

Relationship between DNA binding and enzymatic activity of SiR

What is the role of SiR within the nucleoids? Here, the enzymatic activity assay using recombinant ferredoxin demonstrated that the reductase activity of PsSiR is not affected by its DNA binding. Although there is currently no decisive evidence for a physiological advantage of the association of SiR with nucleoids, one probable explanation is that the nucleoid SiR, containing a siroheme and a [4Fe-4S] cluster as catalytically active redox centers, could act as a sensor of the redox state of the chloroplast because the expression of some chloroplast genes is regulated by the chloroplast redox state [38,39]. It is intriguing that FrxB, a subunit of NAD(P)H dehydrogenase containing a [4Fe-4S] cluster, was isolated as a plastid DNA-binding protein [40]. This provides a reasonable hypothesis for the redox control of the functional and/or morphological state of chloroplast nucleoids. Various metabolic enzymes, which do not have obvious activity related to nucleoid functions such as DNA replication, DNA maintenance, or transcription, constitute mitochondrial nucleoids in baker's yeast [41–44], Xenopus laevis[41,45], and human HeLa cells [46]. Interestingly, aconitase, a citric acid cycle enzyme, is essential for the stability of mitochondrial DNA in yeast [42,47]. Aconitase also contains a [4Fe-4S] cluster that is essential for its catalytic activity [48]. Shadel [49] showed that the disassembly or oxidation of the aconitase iron–sulfur cluster induced by reactive oxygen species generated by oxidative phosphorylation results in its reallocation from the matrix to mitochondrial nucleoids, stabilizing mitochondrial DNA under oxidative stress. In contrast, Chen and Butow [41] suggested that, in respiratory conditions, an increased level of aconitase might substitute for Abf2, which tightly packages mitochondrial DNA and changes the structure of nucleoids into a metabolically favorable conformation and protects DNA in the remodeled conformation. In chloroplasts, various reactive oxygen species are produced during photosynthesis [50,51]. We hypothesize that the DNA binding of SiR might be dependent on the metabolic state, e.g. photosynthesis, to optimize nucleoid conformation for metabolic activity by regulating gene expression and/or for maintenance of chloroplast DNA by protecting it from harmful agents.

Experimental procedures

Plant materials and growth conditions

Seeds of pea (Pisum sativum L. cv. ‘Alaska’) and maize (Zea mays L. cv. ‘Golden Cross Bantam’) were soaked in tap water overnight at room temperature and then sown on moist vermiculite and allowed to germinate at about 25 °C for pea and 32 °C for maize. The seedlings were grown under white fluorescent lamps at a fluence rate of about 50 µmol·m−2·s−1.

Preparation of chloroplasts and nucleoids

Pea chloroplasts were prepared from the leaf buds of epicotyls of 6- or 7-day-old seedlings or the mature leaves of 14-day-old plants as described previously [10]. Maize chloroplasts were isolated from the cotyledons of 6-day-old plants in a similar way. Nucleoids were prepared as described previously [10] using TAN buffer (20 mm Tris/HCl, pH 7.5, 0.5 mm EDTA, 0.5 m sucrose, 7 mm 2-mercaptoethanol, 0.4 mm phenylmethylsulfonyl fluoride, 1.2 mm spermidine) and then stored in the presence of 33% glycerol at −80 °C until use.

Screening cDNA Libraries

PCR was carried out with genomic DNA (laboratory stock) using the degenerate primers 5′-ATCAAGTTYCAYGGWAGCTA-3′ and 5′-AYACGACCAYTRTCWACRTG-3′, which were designed based on the highly conserved regions of SiR. The PCR products were cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, CA, USA). DNA sequencing and a subsequent homology search in GenBank were used to identify the PCR products as a portion of an SiR gene. This fragment was used as a probe to screen two types of cDNA libraries [19], a cDNA library in λ gt10, which was primed with oligo-dT, and a cDNA library in λ gt11, which was primed with random hexamers. The details of the cloning are described in the Results section. The inserts in the isolated λ clones were amplified by PCR using vector primers, and the products were then cloned into pCR2.1-TOPO. DNA sequencing was performed by the dideoxy chain-termination method using a DNA sequencing kit, Big-Dye Terminator 3 (Applied Biosystems, Foster City, CA, USA), with a DNA sequencer (Model 377; Applied Biosystems). The DNA sequence was assembled with AutoAssembler (Applied Biosystems) and manipulated with genetyx (Software Development, Tokyo, Japan).

Phylogenetic analysis

Various SiR sequences were retrieved from the GenomeNet website (http://www.genome.ad.jp). Database sequences and alignment files were manipulated using siseq[52]. Amino acid sequences were aligned with muscle version 6.5 [53] using the default settings. After alignment, highly variable sites (> 20%) as well as both termini were removed. The phylogenetic tree was constructed by the maximum likelihood method using treefinder (version May 2006) [54]. A neighbor-joining tree was calculated using mega (version 3.1) [55] with the jtt model and gamma value of 2.0. A maximum parsimony tree was calculated using paup (version 4b10, Sinauer Associates, Sunderland, Massachusetts, USA) with a heuristic search. A graphical representation of the phylogenetic tree was made using the njplot[56].

DNA and RNA blot hybridization

Genomic DNA (4 µg) was digested with EcoRI and HindIII and the DNA fragments were resolved on 0.8% agarose gel. The restriction fragments were transferred to a nylon membrane (Hybond-N+ GE Healthcare, Piscataway, NJ, USA) according to the protocol recommended by the manufacturer. After prehybridization, hybridization was carried out in a hybridization mixture containing 50% formamide with digoxigenin (DIG)-labeled probes for the second exon of the pea SiR gene. Other details of DNA blot hybridization, including chemiluminescent detection, have been described previously [19].

RNA blot analysis was performed essentially as described previously [19]. Briefly, 10 µg of total RNA or poly (A) + RNA was electrophoresed in 1.2% agarose gel and then transferred to a nylon membrane (Biodyne A; Pall, New York, USA). The probe as described for DNA blotting was used for hybridization.

Immunoblot analysis

Various plant tissues stored at −80 °C were ground with a mortar and pestle under liquid nitrogen. The powder was transferred into a micro tube and homogenized with a pestle in two volumes of 50 mm Tris/HCl (pH 7.5), 50 mm NaCl, 1 mm MgCl2, 1 mm EDTA, and 1 mm phenylmethylsulfonyl fluoride on ice. The homogenate was centrifuged at 2000 g for 1 min. The crude extract was mixed with SDS to a final concentration of 1% and boiled for 5 min. Electrophoresis and immunoblotting were performed with a 10% polyacrylamide gel as described previously [18].


The total proteins associated with chloroplast nucleoids isolated from pea leaves were reduced and carboxymethylated essentially as described previously [57]. Modified samples were dialyzed against 50 mm NH4HCO3 and lyophilized, then dissolved in 2 mm NaOH and separated by SDS/PAGE. Proteins were stained with a Gel-Negative Stain Kit (Nacalai Tesque, Kyoto, Japan). The negatively stained bands were manually excised from the gel and destained according to the manufacturer's instructions. Gel pieces were ground, dried in vacuo, and resuspended in 100 µL MilliQ water containing 1 milliunit activity Achromobacter lysyl endopeptidase (Wako, Osaka, Japan) or endoproteinase Asp-N (Roche, Basel, Switzerland) for digestion over 12 h at room temperature. Peptides were sequentially extracted from the gel in 0, 10, 50, and 100% acetonitrile, all in 0.1% trifluoroacetic acid. Mass spectrometry was performed as described previously [58].

Preparation of recombinant SiR

Recombinant PsSiR and ZmSiR were over-expressed in E. coli cells and purified with ferredoxin-affinity column chromatography as described in Ideguchi et al. [32]. To facilitate the production of enzymatically active recombinant SiRs, coexpression of the E. coli cysG gene (GenBank accession number X14202) encoding siroheme synthase was attempted. In the purification of PsSiR, the three column-chromatography steps in [32] (anion exchange, gel filtration, and hydrophobic chromatographies) were not used. Instead, desalting after ammonium sulfate fractionation was performed by dialysis, and the precipitates were removed by centrifugation. This step was effective in removing major contaminants.

Gel-mobility shift assay

A double-stranded 40-mer DNA probe was prepared by annealing the oligonucleotide 5′-AGTCTAGACTGCAGTTGAGTCCTTGCTAGGACGGATCCCT-3′ and its complementary strand. A 20-mer dsDNA probe was similarly prepared by annealing 5′-AGTCTAGACTGCAGTTGAGT-3′ and its complementary strand. The oligonucleotide described above for 40-mer dsDNA was also used as a single-stranded DNA probe. These probes were 5′-end labeled. The labeled DNA probes were incubated with recombinant SiR in binding buffer (4 mm Tris/HCl, pH 8.0, 5% glycerol) on ice for 30 min, and then the binding mixtures were electrophoresed in TEA buffer (6.7 mm Tris/HCl, pH 7.9, 1 mm EDTA, 3.3 mm sodium acetate). The gel was fixed with 10% acetic acid, 10% methanol and dried under vacuum. The radioactive bands were detected by autoradiography with X-ray film (BioMax; Kodak, Rochester, NY, USA). The SiR concentration at one-half of the maximum concentration of DNA associated with SiR was taken as the apparent dissociation constant. The experiments in Fig. 6(B) were performed according to the method of Sasaki et al. [59], using XbaI-digested chloroplast DNA or StyI-digested λ phage DNA (Toyobo, Osaka, Japan) as probes. In this assay, the mixtures were electrophoresed in 1% agarose gel in TEA buffer. After electrophoresis, the gel was stained with ethidium bromide for the detection of DNA bands.

Measurement of SiR enzymatic activity

The enzymatic activity of SiR was assayed by measuring the amount of cysteine formed from S2– by cysteine synthase added to the system [60], as described previously [29]. O-Acetylserine was added as the sulfide acceptor. The reaction was started by the addition of sodium dithionite to a mixture containing recombinant maize ferredoxin I (laboratory stock [61]) as an electron donor for SiR. Recombinant PsSiR was incubated with or without HindIII-digested λ DNA (New England Biolabs, Beverly, MA, USA) at room temperature for 30 min to form DNA–SiR complexes before the reductase activity measurement. The SiR alone or DNA–SiR complex was added to the SiR assay mixture to a final concentration of 0.5 µm SiR.

In vitro compaction of DNA

Pea chloroplast DNA was prepared as described previously [18]. Chloroplast DNA (2 µg) was mixed with PsSiR or ZmSiR (35 pmol) in 50 µL of 50 mm Tris/HCl, pH 7.5, and incubated at 4 °C. Aliquots of 4 µL were taken from the mixture, mixed with 4 µL of 1% glutaraldehyde, and then stained with 4 µL of 1 µg·mL−1 4′,6-diamidino-2-phenylindole. All reagents contained 50 mm Tris/HCl, pH 7.5. The specimens were examined under a fluorescence microscope (BX-60, Olympus, Tokyo, Japan) equipped with a WU cube.

In vitro transcription assay

The in vitro transcription assay was performed essentially according to the method of Sakai et al. [62], with some modification. The reaction mixture contained 40 mm Tris/HCl (pH 7.6), 7 mm MgCl2, 24 µm (NH4)2SO4, 0.01% (w/v) Nonidet P40, 180 µm ATP, 180 µm GTP, 180 µm CTP, 5 µm[5,6-3H]UTP (about 0.16 TBq·mmol−1), and chloroplast nucleoids (30 µg of protein·mL−1). Before the addition of radiolabeled UTP to start the reaction, we incubated the mixture at 25 °C for 30 min. The reaction was carried out at 25 °C. After the reaction, an aliquot (10 µL) was spotted on a small disk of DEAE paper (DE81; Whatman, Maidstone, UK). The paper was washed successively in 5% Na2HPO4, water, and ethanol, and then finally dried. The incorporation of [5,6-3H]UTP radioactivity into the DEAE paper-bound fraction was determined by liquid scintillation counting.

Indirect immunofluorescence microscopy

Isolated chloroplasts were fixed by resuspension in grinding buffer containing 2% (w/v) paraformaldehyde. Fixed chloroplasts were washed with NaCl/Pi (137 mm NaCl, 2.7 mm KCl, 7.9 mm Na2HPO4, 1.3 mm NaH2PO4, pH 7.2) and then incubated in NaCl/Pi containing 0.05% (v/v) Triton X-100 at room temperature for 10 min. Permeabilized chloroplasts were washed with NaCl/Pi, incubated with diluted anti-PsSiR antibody in NaCl/Pi containing 1% (w/v) blocking reagent (Roche) for 1 h at 37 °C, and then washed with NaCl/Pi. After the reaction with the primary antibody, chloroplasts were incubated with diluted AlexaFluor 488 goat antiguinea pig IgG (Invitrogen) in NaCl/Pi containing 1% blocking reagent for 1 h at 37 °C, and then washed with NaCl/Pi. Chloroplasts were incubated in NaCl/Pi containing 4 µg·mL−1 4′,6-diamidino-2-phenylindole at room temperature for 15 min and then washed with NaCl/Pi. The doubly stained chloroplasts were examined under a fluorescence microscope with an NIB cube for AlexaFluor 488 and a WU cube for 4′,6-diamidino-2-phenylindole.


This work was supported in part by a Grant-in-Aid for JSPS Fellows (No. 1505869) to KS and a Grant-in-Aid for Scientific Research (No. 15370017, 18017005) to NS from the Japanese Society of for the Promotion of Science, and by the Cooperative Research Program of Institute for Protein Research, Osaka University (2003–06).