Báo cáo khoa học: DNA binding and partial nucleoid localization of the chloroplast stromal enzyme ferredoxin:sulfite reductase pptx

Here, we report the DNA-binding properties of SiRs from pea PsSiR and maize ZmSiR using an enzymatically active holoenzyme with prosthetic groups.. Comparison of PsSiR and ZmSiR suggests

chloroplast stromal enzyme ferredoxin:sulfite reductase Kohsuke Sekine1,2, Makoto Fujiwara2, Masato Nakayama3, Toshifumi Takao3, Toshiharu Hase3 and Naoki Sato1,2

1 Department of Molecular Biology, Faculty of Science, Saitama University, Japan

2 Department of Life Sciences, Graduate School of Arts and Sciences, The University of Tokyo, Japan

3 Institute for Protein Research, Osaka University, Japan

The assimilation of sulfur is an important process for

the synthesis of various sulfur compounds such as

amino acids, sulfolipids, and coenzymes Sulfite

reduc-tase (SiR) is a central enzyme within the sulfur

assi-milation 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

phosphoryl-ated by adenosine-5¢-phosphosulfate kinase, forming

3¢-phosphoadenosine-5¢-phosphosulphate

3¢-Phosphoa-denosine-5¢-phosphosulfate is reduced to sulfite by

3¢-phosphoadenosine-5¢-phosphosulfate reductase, and

sulfite is further reduced to sulfide by SiR The

resul-tant 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 previ-ously 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

Keywords

bifunctional protein; chloroplast nucleoid;

DNA-binding protein; ferredoxin:sulfite

reductase

Correspondence

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

Note

Nucleotide sequence data for PsSiR are

available in the DDBJ ⁄ EMBL ⁄ GenBank

databases under accession number

AB168112

(Received 1 December 2006, revised 15

February 2007, accepted 19 February 2007)

doi:10.1111/j.1742-4658.2007.05748.x

Sulfite reductase (SiR) is an important enzyme catalyzing the reduction of sulfite to sulfide during sulfur assimilation in plants This enzyme is locali-zed 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 abil-ity 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 parti-cular 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

Abbreviations

PsSiR, pea (Pisum sativum) sulfite reductase; SiR, sulfite reductase; ZmSiR, maize (Zea mays) sulfite reductase.

stromal protein [5–7], but localization to plastid

nucle-oids provides a new aspect of this enzyme and

chloro-plast 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

plas-tid 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

var-ious enzymes were reported in addition to those

involved in transcription, DNA replication, DNA

topology, and DNA binding These were iron

super-oxide dismutase, putative thioredoxin, pfkB-type

car-bohydrate 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]

Dif-ferent researchers use difDif-ferent 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

com-paction 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,

pro-teins lacking prosthetic groups However, we have

been using pea to isolate chloroplast nucleoids because

a large amount of chloroplast nucleoid is obtained

effi-ciently 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

Results Isolation and molecular characterization

of a cDNA clone encoding pea SiR The k gt10 pea cDNA library was screened One posi-tive 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 nuc-leoid, ‘VSTPAKS’ [1], was found within the deduced amino acid sequence (Fig 1) To determine the missing region, the k 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)

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 Synecho-cystis 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 lig-ands, Cys434, Cys440, Cys479, and Cys483, and that four basic residues, Arg83, Arg153, Lys215, and Lys217, are involved in the substrate coordination to

Fig.

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

sim-ilar trees were obtained by the neighbor-joining and

maximum parsimony methods All SiRs of flowering

plants were monophyletic and the SiRs of plants

ori-ginated from cyanobacterial SiR (Fig 2) The

reliabil-ity of the tree was confirmed by bootstrapping The

important nodes for the above statement were

suppor-ted with high confidence levels

Basic characterization of the PsSiR gene

cDNA gel blot analysis was carried out with pea

genomic DNA restricted with EcoRI or HindIII A

sin-gle band was detected in both of the digests (Fig 3D),

suggesting a single PsSiR gene per genome To

exam-ine 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

expres-sion 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

agree-ment with previous studies in tobacco [29], maize [30],

and A thaliana [31]

The presence of SiR in chloroplast nucleoids was

detected previously using immunological and

enzymo-logical 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 pep-tides were detected with 13.9% sequence coverage Importantly, the AspN digestion gave peptides corres-ponding 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 predic-ted transit peptide sequence was detecpredic-ted 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 co-expression of siroheme synthase and purified by a com-bination 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

Fig 2 Phylogenetic tree of SiRs based on

amino acids sequences The tree was

con-structed using the maximum likelihood

method The three numbers on each branch

show confidence levels for the maximum

likelihood ⁄ neighbor-joining ⁄ maximum

parsi-mony analyses, estimated by bootstrap

ana-lysis with 1000 replicates.

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)

This preparation was used to examine the

DNA-bind-ing activity of SiR A gel-mobility shift assay was

car-ried 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)

The sequence specificity of SiR during DNA binding was examined Non-labeled poly(dI-dC)Æpoly(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 com-petitor were 34, 17 and 12%, respectively, of that with-out a competitor, indicating comparable affinities of SiR for the 20-mer dsDNA and the poly(dI-dC) The

k DNA digested with StyI and pea chloroplast DNA digested with XbaI were mixed with PsSiR and electro-phoresed (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 lm,

300 400 500 600

0.4 0.3 0.2 0.5

0.1 0

nm

175

47.5 62 83 KDa

1 2

Fig 4 (A) UV-visible absorption spectrum and (B) SDS ⁄ PAGE ana-lysis of purified recombinant PsSiR (A) Absorption maxima at 389 and 580 nm (arrows) indicate the presence of a siroheme-contain-ing prosthetic group (B) A ssiroheme-contain-ingle band (arrowhead) was observed

by staining with Coomassie brilliant blue.

Leaves Stems Roots

PsSiR

Light Dark

PsSiR

Leaves Stems Roots

5

2

kbp

95 kDa

Fig 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 lg) prepared from

green leaves, stems, or roots, and polyA + RNAs (3 lg) prepared

from leaves grown in light or dark were electrophoresed on 1.2%

agarose gel The blots were probed with a DIG-labeled DNA

frag-ment corresponding to the second exon of PsSiR Lower panels

indicate the staining of (A) 28S rRNA and (B) blotted mRNA of pea

b-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 ⁄ P i 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

diges-ted with EcoRI and HindIII and separadiges-ted on 0.8% agarose gel The

blots were probed with the same probe used for RNA blotting.

higher than that of ZmSiR [33], which is about 4 lm.

Unlike PsSiR, the activity of ZmSiR was assayed by

measuring the increase in NADP+oxidized by

reduc-tion of ferredoxin III donating electrons to SiR as

des-cribed by Yonekura-Sakakibara et al [30] The

difference in Km may be due to the differences in the

methods used to measure 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

enzy-matic activity was measured (Fig 7) There was no

sig-nificant 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

experi-ment 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

mix-ing DNA with PsSiR or ZmSiR Immediately after

mixing, a number of blurred particles were formed

upon addition of PsSiR, whereas either ZmSiR or

9 8 7 6 5 4 3 2

FD

FD

9 8 7 6 5 4 3 2

FD

9 8 7 6 5 4 3 2

Fig 5 DNA binding activity of SiRs The 32 P-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 n M PsSiR

(lanes 2–5) or ZmSiR (lanes 6–9) before electrophoresis in 6%

poly-acrylamide gel FD indicates free DNA The bands very close to

wells, which suggest large complexes, are indicated by

arrow-heads 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.

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

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 transcrip-tional activity of chloroplast nucleoids Isolated pea chloroplast nucleoids were mixed with various con-centrations 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 lgÆmL)1 heparin In a previous paper [27], we showed that addition of heparin releases the endogenous SiR from nucleoids, and the full transcrip-tional 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 lm (not shown) This result is consistent with the results

of our previous experiment [27] When PsSiR was added at a concentration of 0.2 lm, 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

0

100

200

300

400

Ferredoxin (µM)

Fig 7 Measurement of enzymatic activity of DNA-bound PsSiR.

PsSiR was incubated without (open circles with solid thin line) or

with 10 lgÆmL)1(filled squares with bold solid line) or 20 lgÆmL)1

(filled triangles with dashed line) HindIII digested k DNA before

measurement Various concentrations of recombinant maize

ferre-doxin I were added as an electron donor to SiR The amount of

cys-teine produced per minute per mole of SiR was used as a measure

of enzymatic activity.

5 4 3 2 1

poly(dI-dC)

A

2x 5x 10x

9 8 7 6 5 4 3 2

B

Fig 6 Sequence specificity for DNA-binding by SiRs (A) Competition for DNA-SiR complex formation with poly(dI-dC)Æpoly(dI-dC) The

32 P-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 n M 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 k 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 n M PsSiR (lanes 2–4 and 9–11) or ZmSiR (lanes 5–7 and 12–14) prior to electrophoresis in 1% agarose gel.

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

act-ive in transcription than those from mature leaves

(closed circle) due to stronger repression by

endo-genous 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,

indi-rect immunofluorescence microscopy of isolated

chloro-plasts 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

chloro-plasts were also counterstained with

4¢,6-diamidino-2-phenylindole to visualize chloroplast DNA In pea

chloroplasts, the AlexaFluor signal was detected

non-uniformly 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

Alexa-Fluor 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

concen-trated in nucleoids

Discussion

Comparative aspects of SiR

The sequence alignment (Fig 1) indicates that PsSiR

has the archetypal molecular structural characteristics

0

1

6 2

20

h

Fig 8 Compaction of chloroplast DNA by SiRs Isolated pea

chloro-plast 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-di-amidino-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.

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 reta-ins this reta-insertion from Ala234 to Phe261 with high homology, as expected However, region A sequences

in C merolae SiRA and SiRB are poorly conserved,

SiR (µM)

40

20

0

0.4 0.2

0

A

B

SiR DNA

Fig 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, 3 H-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

per-centage of the activity in the presence of 100 lgÆmL)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

act-ual UTP incorporation in the nucleoids from mature and developing

chloroplasts was 601 ± 83 and 824 ± 113 cpmÆng)1 cpDNA,

respectively s,d, nucleoids from mature pea chloroplasts

(14-day-old leaves); h,j, nucleoids from developing pea chloroplasts

(6-day-old leaf buds).s,h, addition of maize SiR; d,j, 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.

Fig 10 Intrachloroplast localization of SiR Isolated pea (A, B, C, D)

or maize (E, F, G, H) chloroplasts were fixed with paraformalde-hyde, probed with antibodies raised against PsSiR (A, B, E, F) or preimmunized serum (C, D, G, H) and AlexaFluor488-tagged secon-dary 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 lm.

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]

Phylo-genetic analysis with plastid-encoded protein genes

suggested that plastids originate from the Anabaena–

Synechocystis clade However, plant SiRs are

asso-ciated with the Synechococcus–Prochlorococcus 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

sec-ondary endosymbiosis In C merolae, an ORF (CMR

440 C) homologous to the a-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

locali-zed to plastid nucleoids [1,2] The nucleoid localization

of SiR could have been explained by an isozyme

enco-ded 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

act-ive 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 repres-sion of DNA synthesis by SiR in isolated nucleoids [26] Radiolabeled DNA and poly(dI-dC)Æpoly(dI-dC) competed comparably for binding to SiR, and all the restriction fragments of both k 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 regula-tor 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-phen-ylindole 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 experi-ments made the localization of SiR slightly obscure, but this indicates that SiR is not confined to the nucle-oids 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, indica-ting that the DNA-compacindica-ting 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