RESEARCH ARTICLE

Yohsuke Moriyama Æ Tomokazu Yamazaki

Hideo Nomura Æ Narie Sasaki Æ Shigeyuki Kawano

Early zygote-specific nuclease in mitochondria of the true slime moldPhysarum polycephalum

Received: 1 June 2005 / Revised: 22 August 2005 / Accepted: 31 August 2005� Springer-Verlag 2005

Abstract The active, selective digestion of mtDNA fromone parent is a possible molecular mechanism for theuniparental inheritance of mtDNA. In Physarum poly-cephalum, mtDNA is packed by DNA-binding proteinGlom, which packs mtDNA into rod-shaped mt-nucle-oids. After the mating, mtDNA from one parent isselectively digested, and the Glom began to disperse.Dispersed Glom was retained for at least 6 h aftermtDNA digestion, but disappeared completely by about12 h after mixing two strains. We identified two novelnucleases using DNA zymography with native-PAGEand SDS-PAGE. One is a Ca2+-dependent, high-molecular-weight nuclease complex (about 670 kDa),and the other is a Mn2+-dependent, high-molecular-weight nuclease complex (440–670 kDa); the activity ofthe latter was detected as a Mn2+-dependent, 13-kDaDNase band on SDS-PAGE. All mitochondria isolatedfrom myxamoebae had mt-nucleoids, whereas half of themitochondria isolated from the zygotes at 12 h aftermixing had lost the mt-nucleoids. The activity of theMn2+-dependent nuclease in the isolated mitochondriawas detected at least 8 h after mixing of two strains. Thetiming and localization of the Mn2+-dependent DNaseactivity matched the selective digestion of mtDNA.

Key words Isogamous mating Æ Uniparentalinheritance Æ Mitochondrial transmission genetics ÆMyxomycetes

Introduction

Mitochondria contain their own genomes, which aretransmitted either uniparentally or maternally in diversetaxa of sexual eukaryotes, including higher plants,mosses, ferns, algae, fungi, and animals, including hu-mans (reviewed by Birky 1995). However, the molecularmechanism underlying uniparental inheritance remainsunclear.

Recently, the selective destruction of sperm mito-chondria in zygotes was reported for mammals, and thepossible involvement of ubiquitin in the destruction ofsperm mitochondria in fertilized cow and monkey eggswas suggested (Sutovsky et al. 1999, 2000). Previously,we showed that the selective digestion of mtDNA occursbefore the destruction of mitochondria in the zygote ofthe true slime mold Physarum polycephalum (Moriyamaand Kawano 2003). In a cross between myxamoebastrains AI35 and DP246, AI35 is consistently the mito-chondrial donor strain and DP246 is the recipient strain.About 5 h after mating, the mtDNA in the mitochon-dria from the recipient strain was digested synchro-nously. The mitochondrial inner and outer membranesfrom the recipient strain were destroyed long afterdigestion of the mtDNA, at about 36 h after mating.

The active digestion of organelle DNA has also beenreported in chloroplast inheritance. Kuroiwa et al.(1982) found that DAPI-stained chloroplast DNA(cpDNA) from mating type minus strains disappearedpreferentially in young zygotes of the isogamous greenalgae Chlamydomonas reinhardtii within 50 min ofmating. Recently, a nuclease that is potentially involvedin chloroplast inheritance has been investigated in termsof a mating type plus (mt+)-specific DNase (MDN,Nishimura et al. 2002). This nuclease accumulates in thecytoplasm of mt+ cells during gamete maturation. Aftermating, it is imported into mating type minus (mt�)chloroplasts. Based on these results, MDN has beenproposed as a possible force driving the uniparentalinheritance of cpDNA in C. reinhardtii.

Communicated by L. Tomaska

Y. Moriyama Æ T. Yamazaki Æ H. Nomura Æ S. Kawano (&)Department of Integrated Biosciences,Graduate School of Frontier Sciences, University of Tokyo,Bldg. FSB-601, 5-1-5 Kashiwanoha, Kashiwa,277-8562 Chiba, JapanE-mail: kawano@k.u-tokyo.ac.jp

N. SasakiDepartment of Biology, Faculty of Science,Ochanomizu University, 2-1-1 Otsuka, Bunkyo-ku,112-8610 Tokyo, Japan

Curr Genet (2005)DOI 10.1007/s00294-005-0025-2

We speculated that a nuclease similar to MDN isinvolved in the uniparental inheritance of mitochondriain P. polycephalum. It is possible that its activity is highin P. polycephalum, as the mt-nucleoids of P. polyceph-alum contain an extraordinarily large amount ofmtDNA and form a simple rod shape (Kuroiwa 1982).The mtDNA of P. polycephalum is a 62,862-bp circularmolecule (Takano et al. 2001). Each mt-nucleoid con-tains 40 or 80 copies of the mtDNA molecule at themitochondrial G1 or G2 phase, respectively (Kuroiwaand Kuroiwa 1980). Myxamoebae have about 15 mito-chondria, and thus each cell is estimated to haveapproximately 63 Mb of mtDNA. The digestion ofmtDNA from the recipient strain is limited to a veryshort time period, which has been estimated to be30 min (Moriyama and Kawano 2003). This pattern ofdigestion cannot be explained by the action of restrictionendonucleases because they simply cleave DNA to pro-duce fragments. In P. polycephalum, mtDNA is packedinto the rod-shaped mt-nucleoid with DNA-bindingproteins, such as Glom, which causes intense chromatincondensation without suppressing DNA functions(Sasaki et al. 2003). As the nucleoids kept their shapesduring the active digestion of mtDNA, the selectivedigestion of mtDNA is not influenced by such anmtDNA-binding protein. This is also why we emphasizea non-specific nuclease, such as DNase I, rather than asequence-specific restriction endonuclease for the elimi-nation of the mtDNA from one parent in the zygotes.

In this paper, to examine the behavior of mitochon-drial nucleoids and mtDNA in young zygotes, we usedan anti-Glom antibody and the DNA-specific fluoro-chrome DAPI. We investigated the DNase activity inzygotes using DNA zymography, which is a highlysensitive technique for detecting enzyme activity (re-viewed by Bischoff et al. 1998). This method can alsodetect DNases of various molecular sizes by incorpo-rating substrate DNA into the electrophoresis gel. TheDNase activity changes in young zygotes were analyzedusing DNA zymography in native-PAGE and SDS-PAGE.

Materials and methods

Strains, culture, and zygote formation

The myxamoebae P. polycephalum strains AI35 andDP246 were used in this study. The myxamoebae werecultured on PGY plates (0.5% glucose, 0.05% yeastextract, 2 mM MgSO4, and 1.5% agar in 25 mMpotassium phosphate buffer (pH 6.6)) at 23�C in thepresence of live bacteria (Klebsiella aerogenes) as a foodsource. Zygote formation was induced by mixing thetwo myxamoeba strains at 23�C on SM-30 mating plates(30 mM citrate buffer (pH 4.5), 10 mM MgSO4, 1.5%agar). For efficient crossing, the myxamoebae shouldcarry different matA alleles, and one strain in eachmating pair should act consistently as the mitochondria

donor. However, the dominant strain does not neces-sarily act as the mitochondria donor in other combina-tions; the donor in each pair is determined by therespective matA alleles. In the cross of AI35 (matA2) andDP246 (matA16), the donor strain was always AI35.

Cell preparation

The myxamoeba strains grown on PGY plates weresuspended in distilled water (DW) using a spreader andcentrifuged at 1,400·g for 1 s. The myxamoebae wereresuspended in DW and then centrifuged again to re-move the supernatant. Zygote cells were obtained fromSM-30 plates in the same manner.

Immunofluorescence microscopy

For immunofluorescence staining, the myxamoeba cellswere fixed on a coverslip with 3.7% formaldehyde in10 mM KPB for 15 min and then permeabilized for 1 hwith 1% Triton X-100 in 10 mM KPB. The anti-Glomantibody was used as the primary antibody at a dilutionof 1:1,000 (v/v) in blocking buffer (5% bovine serumalbumin in PBS), and the secondary antibody was a1:500 (v/v) dilution in blocking buffer of goat anti-rabbitIgG, which was labeled with TRITC (Sigma). DNA wasstained with 4¢,6-diamidino-2-phenylindole (DAPI), anda coverslip was placed over the stained sample. Photo-graphs were taken with a BX62 Olympus epifluorescencemicroscope (Tokyo, Japan), which was equipped with aC4742 CCD camera (Hamamatsu Photonics, Japan) andthe Aquacosmos system (Hamamatsu Photonics).

Isolation of mitochondria

Mitochondria were isolated from myxamoebae andzygotes using an airbrush, as described previously(Nishimura et al. 2002). The cells were suspended in DW,and centrifuged at 1,400·g for 1 s. The supernatant fluidwas removed, and the precipitated cells were resuspendedin isolation buffer (0.3 M sucrose, 20 mM Tris–HCl(pH 7.7), 1 mM EDTA). The cells were then broken by asingle passage at 0.35 MPa through an airbrush(Evolution A; Harder & Steenbeck, Oststeinbek,Germany), which was connected to an air regulator andair compressor (APC-001, AiRTEX, Japan). The celllysates were collected in a pre-cooled 500-ml beaker.Aftercentrifugation at 300·g at 4�C for 5 min, the supernatantwas centrifuged at 15,000·g at 4�C for 10 min. The pre-cipitated mitochondria were resuspended in isolationbuffer. For further purification, the cell lysates werelayered on top of a discontinuous Percoll gradient(10–80% Percoll in isolation buffer). The mitochondrialfraction was collected by centrifugation at 10,000·g at4�C for 30 min, diluted fourfold in the suspension buffer,and then centrifuged at 15,000·g at 4�C for 3 min.

Preparation of soluble protein extracts

Cells or isolated mitochondria were suspended in theextraction buffer (50 mM Tris–HCl (pH 8.5), 0.1 mMEDTA) in the presence of a protease inhibitor cocktail(final concentrations: AEBSF, 1 mM; aprotinin,0.8 lM; leupeptin, 20 lM; bestatin, 40 lM; pepstatin A,15 lM; E-64, 14 lM). The cells were extracted andground into a fine powder using liquid N2 in a prechilledmortar and pestle, and the powder was collected insampling tubes. After centrifugation at 20,300·g at 4�Cfor 15 min, the supernatants were collected and subse-quently used as crude soluble protein extracts in DNAzymography. The isolated mitochondria were treatedsimilarly. The soluble protein concentrations of the ex-tracts were determined using a protein assay kit (Bio-Rad) with BSA as the standard.

Preparation of mtDNA

Mitochondria from myxamoebae or zygotes at 12 h aftermixing were suspended in 10x Tris/saline EDTA(100 mM Tris–HCl (pH 8.0), 150 mM NaCl, 100 mMEDTA) that contained 2% SDS and 0.5 mg/ml pro-teinase K, and were incubated for 2 h at 37�C. The nu-cleic acids were extracted twice with phenol/chloroformand treated with 0.1 mg/ml DNase-free RNase A for 1 hat 37�C. Following extraction with phenol/chloroform,the mtDNA was precipitated by adding an equal volumeof 2-propanol and was dissolved in sterile water.

DNA zymography

DNase activity was analyzed by DNA zymography innative-PAGE and SDS-PAGE. The soluble proteinextracts were normalized for protein concentration andmixed with an equal volume of either native samplebuffer (63 mM Tris–HCl (pH 6.8), 30% glycerol, 0.01%bromophenol blue) or SDS sample buffer (63 mMTris–HCl (pH 6.8), 30% glycerol, 0.01% bromophenolblue, 0.1% SDS). In native-PAGE, the proteins wereseparated with 300 V at 4�C for 60 min using a stackinggel (3.3% (w/v) acrylamide, 63 mM Tris–HCl (pH 6.8),2 mM EDTA) and a separating gel (8.0% (w/v) acryl-amide, 375 mM Tris–HCl (pH 8.8), 2 mM EDTA, and10 lg/ml salmon testes DNA or 2.5 lg/ml mtDNA). InSDS-PAGE, the proteins were separated with 300 V at4�C for 30 min using a stacking gel (3.3% (w/v) acryl-amide, 63 mM Tris–HCl (pH 6.8), 2 mM EDTA, 0.1%SDS) and a separating gel (11.0% (w/v) acrylamide,375 mMTris–HCl (pH 8.8), 2 mM EDTA, and 10 lg/mlsalmon testes DNA or 2.5 lg/ml mtDNA). Afterelectrophoresis in native-PAGE buffer (25 mM Tris (pH8.3), 192 mM glycine) or SDS-PAGE buffer (25 mMTris(pH 8.3), 192 mM glycine, 0.1% SDS), the gels werewashed with three changes of renaturation buffer(10 mM Tris–HCl, pH 8.5) at 4�C for 20 min and then

incubated overnight at room temperature in the reactionbuffer for each pH condition, in the presence of 0.2 mMPMSF and 0.1 mM DTT. For each reaction, 10 mM ofone of the following solutions was used: citric acid buffer(pH 3.5, 4.5), MES-HCl (pH 5.5), bis-Tris propane–HCl(pH 6.5), or Tris–HCl (pH 7.5, 8.5, 9.5). The chloride saltof one of Ca2+, Mg2+, Zn2+, or Mn2+ was added to thereaction buffers at a concentration of 250 mM, 250 mM,200 lM, or 100 lM, respectively. DNase activity wasvisualized by staining the gels with ethidium bromide andviewing under UV light. A high-molecular-weightcalibration kit for native electrophoresis (AmershamPharmacia Biotech) and the Precision Plus ProteinStandards, Dual Color (Bio-Rad) were used as themolecular weight markers for native-PAGE and SDS-PAGE, respectively.

Gel filtration

The soluble protein extracts (700-ll aliquots) were loa-ded on HiPrep 16/60 Sephacryl S-300 gel filtration col-umns (Amersham Pharmacia Biotech), which wereequilibrated with gel filtration buffer (50 mM Tris–HCl(pH 7.5), 0.25 mM EDTA, 0.05% Tween 20). Theproteins were eluted in the same buffer at 0.3 ml/min,and each 5-ml fraction was tested for DNase activityusing the routine DNase assay, as described above. Themolecular sizes of the proteins were calibrated with gelfiltration standards (Bio-Rad).

Results

The behaviors of mtDNA and the mtDNA-bindingprotein in young zygotes

Zygote formation was induced on a mating plate bymixing two strains of different mating types. Nuclear fu-sion occurred in the zygote at about 4 h after mixing thetwo strains. After DAPI staining, the cell nuclei andmtDNA in the myxamoebae emitted bright blue-whitefluorescence (Fig. 1a). Eachmyxamoeba contained 15–25mitochondria before mating, and thus each young zygotecell contained 30–50 mitochondria. Immunofluorescentstaining with the anti-Glom antibody revealed that Glomwas associated with mtDNA and distributed uniformlythroughout the mtDNA in all mitochondrial nucleoids(mt-nucleoids) (Fig. 1a–c). Glom would interact with allthe mtDNA molecules contained in the mt-nucleoid.

At about 5 h after mixing, the mtDNA fluorescencein approximately 50% of the mitochondria in the zygotedwindled simultaneously to small spots and disap-peared, regardless of the position of the mitochondria inthe zygote (Fig. 1d). However, the size of the mt-nu-cleoids did not change even if they were losing mtDNA(Fig. 1e, f). By about 11 h after mixing, the Glom haddispersed but was retained inside the mitochondria

(Fig. 1g–i). One hour later, the Glom was localizedexclusively with the mtDNA; dispersed Glom was notobserved in the zygote (Fig. 1j–l). These observationsshow that mtDNA digestion precedes the dispersal ofGlom, which associates with the mtDNA. DispersedGlom was retained for at least 6 h after mtDNA diges-tion and disappeared completely by about 12 h aftermixing the two strains.

Identification of DNase activity in the zygote

To reveal the molecular mechanism of mtDNA diges-tion, we surveyed the DNase activity in the pH 3.5–9.5

range, using DNA zymography. As we could not predictif DNase activity would be retained under denaturingconditions, we examined it using both SDS-PAGE andnative-PAGE (Fig. 2). Soluble proteins were extractedfrom zygotes at 12 h after mixing myxamoeba stainsAI35 and DP246. Five-microgram aliquots of the pro-teins were electrophoresed, and the ion-dependency ofthe DNase activity was analyzed using four differentdivalent ions (Ca2+, Mg2+, Zn2+, or Mn2+). DNaseactivity was detected at pH 6.5–7.5 in the presence ofCa2+ or Mg2+ after native-PAGE, but the molecularsize was ambiguous (Fig. 2a). Additional DNase activitywas detected at pH 7.5–9.5 in the presence of Mn2+. Itsmolecular size exceeded 440 kDa. However, the precise

Fig. 1 Behavior of mtDNA andthe Physarum mtDNA-bindingprotein Glom. a, d, g, andj DAPI-fluorescencemicrographs. b, e, h, andk Immunofluorescencemicrographs showing thelocalization of Glom in the mt-nucleoids. c, f, i, and l Mergedimages of the DAPI-stained andimmunofluorescencemicrographs. a–c Uninucleatezygote 4 h after mixing the twostrains. d–f Uninucleate zygote5 h after mixing. The mtDNA isbeginning to disappear in abouthalf of the mitochondria(arrows and inset).g–i Uninucleate zygote 11 hafter mixing. The mtDNA hasdisappeared completely, butGlom is retained in about halfof the mitochondria (arrows andinset). j–l Uninucleate zygote12 h after mixing. DispersedGlom is not observed. Scale bar10 lm

molecular weight could not be determined using native-PAGE, because protein mobility in native-PAGE de-pends not only on molecular size but also on nativeelectric charge. By contrast, DNA zymography in SDS-PAGE clearly showed the presence of strong DNaseactivity in the zygotes (Fig. 2b). The DNase activityoccurred at a 50-kDa band regardless of the ionic con-ditions. In particular, in the presence of Ca2+, DNaseactivity was detected across the broadest pH range of5.5–9.5. At pH 6.5–8.5, additional DNase activity wasfound in a band slightly larger than 50-kDa in thepresence of Ca2+ or Mg2+. The two bands near 50-kDawere also found at pH 6.5 in the presence of Zn2+ orMn2+. Furthermore, at pH 7.5–8.5 in the presence ofMn2+, additional DNase activity was found as clear,strong 13-kDa and faint 37-kDa band.

Changes in the DNase activity of young zygotes

We analyzed the change in DNase activity over timefollowing mating using native-PAGE and SDS-PAGE,at pH 6.5 in the presence of Ca2+ and at pH 8.5 in thepresence of Mn2+ (Fig. 3). Myxamoeba strains AI35and DP246 were used to prepare zygotes. In this cross,AI35 always transmits its mitochondria to the progeny,whereas the mtDNA from DP246 is digested selectivelyin the zygote about 5 h after mixing the two strains. Thezygote and plasmodia were obtained from mating platesat each time point after mixing the two strains. TheirDNase activity was detected at pH 6.5 in the presence ofCa2+ in both native-PAGE (Fig. 3a) and SDS-PAGE(Fig. 3b). Both AI35 and DP246 had indistinct bands ofDNase activity in the presence of Ca2+ under non-

Fig. 2 DNase activity ofzygotes detected using DNAzymography with native-PAGE(a) and SDS-PAGE (b). In eachlane, 5-lg aliquots of solubleproteins from the 12-h zygoteswere electrophoresed in native-PAGE. Salmon sperm DNAwas used as the substrate.DNase activity was surveyed inthe pH 3.5–9.5 range, usingDNA zymography in thepresence of each cation (Ca2+,Mg2+, Zn2+, or Mn2+). Innative-PAGE (a), DNaseactivity is detected at pH 6.5–7.5 in the presence of Ca2+ orMg2+. In SDS-PAGE (b), 50-and 13-kDa nuclease activity isclearly observed

denaturing conditions. The DNase activity was retainedin zygotes for at least 48 h but was absent in the matureplasmodia (Fig. 3a). Although the molecular weight ofthis Ca2+-dependent nuclease could not be estimated innative-PAGE, it produced a clear, strong 50-kDa bandwith SDS-PAGE. The 50-kDa band that existed in AI35and DP246 persisted during the mating process for atleast 48 h after mixing the two strains, but it was lost inthe mature plasmodia.

DNA zymography in native-PAGE (Fig. 3c) andSDS-PAGE (Fig. 3d) revealed that neither strain AI35nor strain DP246 had DNase activity in the myx-amoeba cells in the presence of Mn2+ at pH 8.5.However, singular Mn2+-dependent DNase activitywas detected 6 h after mixing the two strains, and thisactivity gradually increased until 48 h after mixing.This Mn2+-dependent nuclease complex was estimatedto be larger than 440 kDa in native-PAGE (Fig. 3c). InSDS-PAGE, the Mn2+-dependent DNase activity wasdetected as 13-kDa and faint 37-kDa bands. Theappearance of these two bands corresponded preciselyto the timing of mtDNA digestion (Fig. 3d). No otherzygote-stage-specific DNase activity similar to theMn2+-dependent nuclease described above was de-tected in this study.

Gel filtration of the DNase activity

The precise molecular weights of these nucleases couldnot be determined in native-PAGE DNA zymography.

To confirm whether the mitochondrial nucleases exist ashigh-molecular-weight complexes, soluble proteins fromthe zygote were first separated on a Sephacryl S-300 gelfiltration column. Each eluted fraction was checkedusing both native-PAGE and SDS-PAGE DNA zy-mography. Native-PAGE did not result in any clearbands. By contrast, SDS-PAGE showed that the frac-tions eluting near the 670-kDa marker protein containedtwo bands of 50-kDa DNase activity at pH 6.5 in thepresence of Ca2+ (Fig. 4a). Moreover, at pH 8.5 in thepresence of Mn2+, several fractions around 670 kDashowed comparatively clear bands of a high-molecular-weight nuclease complex of about 440–670 kDa in na-tive PAGE. DNA zymography with SDS-PAGE showedthat these fractions contained Mn2+-dependent 13-kDaDNase activity at pH 8.5 in the presence of Mn2+

(Fig. 4b). Gel filtration and DNA zymography withSDS-PAGE revealed that both the Ca2+-dependent 50-kDa DNase activity and the Mn2+-dependent 13-kDaDNase activity were components of the high-molecular-weight nuclease complex.

DNase activity of isolated mitochondria

To investigate whether the high-molecular-weightnuclease complex existed in the mitochondria duringmtDNA digestion, we isolated mitochondria frommyxamoebae and zygotes at 0, 4, 8, and 12 h aftermixing strains AI35 and DP246 (Fig. 5a). All the mito-chondria isolated from myxamoebae AI35 and DP246

Fig. 3 Changes in the DNaseactivity during mating andzygote development. TheDNase activity was detectedusing DNA zymography withnative-PAGE (a, c) and SDS-PAGE (b, d). The dependencyon Ca2+ and Mn2+ waschecked at pH 6.5 and 8.5,respectively. Myxamoebastrains AI35 and DP246 weremixed at time 0. Solubleproteins were extracted at eachtime point after mixing. PlsMature plasmodium

contained mt-nucleoids that were observed as brightblue-white fluorescent spots by DAPI straining (left inFig. 5a). By contrast, half of the mitochondria isolatedfrom the zygotes at 12 h after mixing had lost the mt-nucleoids (right in Fig. 5a).

Using these isolated mitochondria, we carried outDNA zymography with native-PAGE and SDS-PAGEat pH 6.5 in the presence of Ca2+ (Fig. 5b) and at pH8.5 in the presence of Mn2+ (Fig. 5c). Soluble proteinswere extracted from the isolated mitochondria, and then2-lg aliquots of the proteins were electrophoresed. AtpH 6.5 in the presence of Ca2+, DNA zymography innative-PAGE showed that highly active, indistinctDNase activity was present consistently in the mito-chondria whenever isolated. No obvious change inDNase activity was detected under this condition. InSDS-PAGE, the 50-kDa DNase activity was detectedconsistently within mitochondria isolated from eithermyxamoebae or from zygotes (right in Fig. 5b). At pH8.5 in the presence of Mn2+, DNA zymography with

native-PAGE detected 200-kDa DNase activity in themitochondria isolated from myxamoebae (left inFig. 5c). This activity was not detected 4 h after mixingor later. By contrast, the high-molecular-weight nucleasecomplex was detected as 440–670 kDa bands in thezygotes at 8 and 12 h after mixing. In SDS-PAGE,although the 13-kDa DNase activity was not detected inmitochondria isolated from the myxamoebae or zygotesuntil 4 h after mixing, this activity was detected in iso-lated mitochondria beginning 8 h after mixing (right inFig. 5c). This timing and localization of the activation ofthe Mn2+-dependent DNase activity coincided with theselective digestion of mtDNA. The 37-kDa band, whichhad detected in the SDS-PAGE of whole zygotes(Fig. 3d), was not detected from isolated mitochondria(right in Fig. 5c).

mtDNA as a substrate for DNA zymography

To investigate whether the Ca2+- and Mn2+-dependentDNase activities actually digest mtDNA, we usedmtDNA as the substrate for DNA zymography. Mito-chondria were isolated from myxamoeba strains AI35 orDP246 and from zygotes at 12 h after mixing. DP246 isthe recessive strain in mitochondrial inheritance. ThemtDNA from DP246 was digested selectively, while themtDNA from AI35 remained in the zygotes at 12 h aftermixing. If the mtDNA from AI35 were protected fromdigestion, no DNase activity would have been detectedwhen it was used as the substrate for DNA zymography.The result of native-PAGE showed that the Ca2+-dependent nuclease complex could digest mtDNAindependent of its origin under non-denaturing condi-tions (left in Fig. 5d). Conversely, in SDS-PAGE, theCa2+-dependent 50-kDa DNase activity did not digestthe mtDNA (right in Fig. 5d). By contrast, the Mn2+-dependent high-molecular-weight nuclease complex di-gested substrate mtDNA in the non-denaturing condi-tion (left in Fig. 5e), and the Mn2+-dependent 13-kDaDNase activity digested mtDNA irrespective of theorigin of the mtDNA, even under denaturing conditions(right in Fig. 5e). These results show that mtDNA wouldnot be protected from any nucleases.

Discussion

Glom is a DNA-binding protein that causes intensechromatin condensation without suppressing DNAfunctions and maintains the mt-nucleoid shape in themitochondria of P. polycephalum (Sasaki et al. 2003).The mtDNA was digested about 5 h after mixing thetwo strains. The mt-nucleoid shape did not changeduring this mtDNA digestion. However, 6 h after themtDNA was digested, the Glom gradually dispersed,although it was retained in the mitochondria (Fig. 1g–i),and the dispersed Glom signals disappeared 1 h later(Fig. 1j–l). Glom should leak into the cytosol from the

Fig. 4 Gel filtration and detection of DNase activity using DNAzymography. Soluble proteins from zygotes were separated usingSephacryl S-300 gel filtration. Molecular sizes were calibrated usinggel filtration standards (Bio-Rad). DNA zymography of native-PAGE and SDS-PAGE at pH 6.5 in the presence of Ca2+ ions (a)and at pH 8.5 in the presence of Mn2+ ions (b)

destroyed mitochondria, and, using electron micros-copy, we observed that the mitochondria that had lostmtDNA were destroyed and completely excluded fromthe cytosol. The mitochondrial outer and inner mem-branes retain their integrity during mtDNA digestion(Moriyama and Kawano 2003). If selective mtDNAdigestion were caused by an influx of cytosolic nucleasefollowing the destruction of the mitochondrial mem-branes, then the dispersed Glom after mtDNA digestionwould not be retained in the mitochondria. However, asGlom, a 41-kDa protein (Sasaki et al. 2003), does notleak into the cytosol, it would not be possible for anysubunits of high-molecular-weight nuclease complexesto move into the mitochondria by mitochondrial mem-brane-destruction. Selective mtDNA digestion is aninitial, crucial event for the uniparental inheritance ofmitochondria (Moriyama and Kawano 2003). Ourpresent results suggest that this event is initiated by notsimple mitochondrial-membrane-destruction but somenuclease-import-system.

Major mitochondrial nucleases have been identifiedin some species. For example, rat MtTGendo, human

Ape2, and mouse NTH1 prevent the accumulation ofpersistent mutations in mtDNA (Stierum et al. 1999;Tsuchimoto et al. 2001; Ikeda et al. 2002). Mouse NTH1is an ortholog of Escherichia coli endonuclease III,which functions in the first step of base excision repair.By contrast, various low-molecular-weight nucleases,which range in size from 29 to 38 kDa, have been de-scribed in the mitochondria of several organisms (Chowand Fraser 1983; Zassenhaus et al. 1988; Engel and Ray1998; Fikus et al. 2000). Most of these nucleases areactive as dimers, require Mg2+ for activity, and haveoptimum activity at around the physiological pH. ACa2+-dependent mitochondrial matrix nuclease with apH optimum of 6.4 has been identified in the rat (Davieset al. 2003). These DNases are involved in mtDNAmaintenance and nucleotide metabolism. Recently,mitochondrial endonuclease G (EndoG) has been de-scribed as a key enzyme for apoptosis in mammals(Li et al. 2001).

The stage specificity is very important in the searchfor a possible mitochondrial nuclease that causes selec-tive digestion of mtDNA from one parent during the

Fig. 5 Changes in the activitiesand substrate specificities of themitochondrial nucleases.DAPI-staining fluorescencemicrographs of themitochondria isolated frommyxamoebae (left) and zygotes(right) at 12 h after mixingstrains AI35 and DP246 (a).Changes in the mitochondrialDNase activity in myxamoebae(AI35 and DP246) and at 0, 4, 8and 12 h after mixing aredetectable using DNAzymography. DNAzymography of native-PAGEand SDS-PAGE at pH 6.5 inthe presence of Ca2+ (b) and atpH 8.5 in the presence of Mn2+

(c). DNA zymography usingmtDNA from myxamoebae(AI35 and DP246) and zygotes(Zyg) at 12 h after mating asthe substrates. DNAzymography with native-PAGEand SDS-PAGE at pH 6.5 inthe presence of Ca2+ (d) and atpH 8.5 in the presence of Mn2+

(e)

uniparental inheritance of mitochondria. In this work,we identified two mitochondrial nucleases that arepotentially involved in the active digestion of mtDNA.One is a Ca2+-dependent high-molecular-weight nucle-ase complex (about 670 kDa), and its activity was de-tected as a Ca2+-dependent 50-kDa DNase band inSDS-PAGE DNA zymography. This nuclease activitywas detected in myxamoebae and zygotes but not inmature plasmodia. The other is a Mn2+-dependenthigh-molecular-weight nuclease complex (440–670 kDa), and its activity was detected as a Mn2+-dependent 13-kDa DNase band in SDS-PAGE. Itsactivity was elevated beginning at about 6 h after mixingthe two myxamoeba strains. This timing correspondedprecisely to that of the selective mtDNA digestion.

We previously described the mechanism underlyingbiparental inheritance (Moriyama and Kawano 2003).In certain crosses of P. polycephalum strains, althoughthe mtDNA from one parent is digested in the youngzygote, several mtDNA molecules survive and subse-quently replicate during zygote maturation. As a result,mtDNA from both parents is transmitted to the plas-modium. In these crosses, mtDNA digestion does occur,although the mitochondria maintain their capacity forsemi-autonomous multiplication. If selective mtDNAdigestion were caused by the destruction of the mito-chondrial outer and inner membranes, followed by theinflux of a cytosolic nuclease, we would expect that thedispersed Glom would not be retained in the mito-chondria. If the inner membrane were destroyed, themitochondria could not replicate their DNA, as seenwith biparental inheritance. It is possible that a nucleaseis selectively incorporated into the mitochondria derivedfrom one parent without membrane destruction. Ourwork showed that Ca2+-dependent 50-kDa DNaseactivity exists in both myxamoebae and zygotes and thatMn2+-dependent 13-kDa DNase activity exists in themature plasmodium that contains mitochondria fromone parent only. These nucleases would not be selec-tively incorporated into the mitochondria derived fromone parent.

In the study of organelle inheritance, it is a favoredidea that the organelle DNA derived from one parent isprotected from the nuclease by methylation or othermodification of nucleotide. Methylation of cpDNA fromone parent has been proposed to protect it against DNAdigestion in the maternal inheritance of the chloroplast inChlamydomonas (Burton et al. 1979; Umen andGoodenough 2001; Nishiyama et al. 2004). In our study,however, there was no difference in terms of resistance toDNase digestion between the mtDNA from myxamoe-bae and zygotes (Fig. 5d, e). The Ca2+- and Mn2+-dependent high-molecular-weight nuclease complexeswere able to digest mtDNA under non-denaturing con-ditions in a manner that was independent of the mtDNAorigin. The Mn2+-dependent 13-kDa DNase activityeven digested mtDNA under denaturing conditions,whereas the Ca2+-dependent 50-kDa DNase activity wasunable to do so. If the Ca2+-dependent 50-kDa nuclease

is an exonuclease, it should not be able to digest thecircular mtDNA of P. polycephalum (Sakurai et al. 2000;Takano et al. 2001). However, as the Ca2+-dependent50-kDa DNase activity is able to digest the mtDNA as ahigh-molecular-weight complex, some component nee-ded for the digestion of mtDNA must be lacking underthe DNA zymography conditions (Fig. 5d).

The molecular weights of the Ca2+- and Mn2+-dependent nuclease complexes appear to be large. Someother high-molecular-weight nucleases have been re-ported recently, e.g., a RecBCD nuclease of E. coli andexosomes in eukaryotes. The RecBCD nuclease is a 330-kDa heterotrimer composed of RecB (134 kDa), RecC(129 kDa), and RecD (67 kDa; Kuzminov 1999). Theexosome is a 300–400-kDa complex in yeast and isresponsible for the turnover and processing of mRNAand rRNA (Van Hoof and Parker 1999; Mitchell andTollervey 2000). Its cytoplasmic complex contains tendifferent components, while its nuclear complex has anadditional component. The complete mechanisms regu-lating theses nucleases in degradation and processinghave not yet been elucidated. As with mtDNA digestion,the selectivity is also unclear. Nevertheless, one possi-bility is that the subunits of the nuclease complex controlits activity. We could not identify which nucleases playthe key role in the uniparental inheritance of mtDNA.Given that both nuclease complexes are active at pH 7.5,it is possible that they work in concert. They require atleast Ca2+ and Mn2+ ions for activation, and thus theregulation of the levels of these ions in the mitochondriamay also be important for this process. It is possible thatthe timing of the activation of the Ca2+- and Mn2+-dependent nucleases and their localizations inmitochondria correspond precisely to the preferentialdigestion of the mtDNA in P. polycephalum.

Acknowledgements We thank Prof. T. Kuroiwa (Graduate Schoolof Science, University of Tokyo) for helpful discussions. We alsothank S. Matsunaga (Graduate School of Engineering, Universityof Osaka) for helpful technical advice. This study was supported bya grant for Scientific Research in Priority Areas (no. 15440246 to S.K.) from the Ministry of Education, Culture, Sports, Science, andTechnology of Japan.

References

Burton WG, Grabowy CT, Sager R (1979) Role of methylation inthe modification and restriction of chloroplast DNA in Chla-mydomonas. Proc Natl Acad Sci USA 76:1390–1394

Birky CW Jr (1995) Uniparental inheritance of mitochondrial andchloroplast genes: mechanisms and evolution. Proc Natl AcadSci USA 92:11331–11338

Bischoff KM, Shi L, Kennelly PJ (1998) The detection of enzymeactivity following sodium dodecyl sulfate-polyacrylamide gelelectrophoresis. Anal Biochem 260:1–17

Chow TY, Fraser MJ (1983) Purification and properties of singlestrand DNA-binding endo-exonuclease of Neurospora crassa.J Biol Chem 258:12010–12018

Davies AM, Hershman S, Stabley GL, Hoek JB, Peterson J, CahillA (2003) A Ca2+-induced mitochondrial permeability transi-tion causes complete release of rat liver endonuclease G activityfrom its exclusive location within the mitochondrial inter-

membrane space Identification of a novel endo-exonucleaseactivity residing within the mitochondrial matrix. Nucleic AcidsRes 31:1364–1373

Engel ML, Ray DS (1998) A structure-specific DNA endonucleaseis enriched in kinetoplasts purified from Crithidia fasciculate.Nucleic Acids Res 26:4733–4738

Fikus MU, Mieczkowski PA, Koprowski P, Rytka J, Sledziewska-Gojska E, Ciesla Z (2000) The product of the DNA damage-inducible gene of Saccharomyces cerevisiae, DIN7, specificallyfunctions in mitochondria. Genetics 154:73–81

Ikeda S, Kohmoto T, Tabata R, Seki Y (2002) Differential intra-cellular localization of the human and mouse endonuclease IIIhomologs and analysis of the sorting signals. DNA Repair1:847–854

Kuroiwa T, Kuroiwa H (1980) Inhibition of Physarum mitochon-drial division by cytochalasin B. Experientia 36:193–194

Kuroiwa T (1982) Mitochondrial nuclei. Int Rev Cytol 75:1–59Kuroiwa T, Kawano S, Nishibayashi S, Sato C (1982) Epifluo-

rescent microscopic evidence for maternal inheritance of chlo-roplast DNA. Nature 298:481–483

Kuzminov A (1999) Recombinational repair of DNA damage inEscherichia coli and bacteriophage. Microbiol Mol Biol Rev63:751–813

Li LY, Luo X, Wang X (2001) Endonuclease G is an apoptoticDNase when released from mitochondria. Nature 412:95–99

Mitchell P, Tollervey D (2000) mRNA stability in eukaryotes. CurrOpin Genet Dev 10:193–198

Moriyama Y, Kawano S (2003) Rapid, selective digestion ofmitochondrial DNA in accordance with the matA hierarchy ofmultiallelic mating types in the mitochondrial inheritance ofPhysarum polycephalum. Genetics 164:963–975

Nishimura Y, Misumi O, Kato K, Inada N, Higashiyama T, Mo-moyama Y, Kuroiwa T (2002) An mt(+) gamete-specificnuclease that targets mt(�) chloroplasts during sexual repro-duction in C reinhardtii. Genes Dev 16:1116–1128

Nishiyama R, Wada Y, Mibu M, Yamaguchi Y, Shimogawara K,Sano H (2004) Role of a nonselective de novo DNAmethyltransferase in maternal inheritance of chloroplast genes

in the green alga, Chlamydomonas reinhardtii. Genetics168:809–816

Sakurai R, Sasaki N, Takano H, Abe T, Kawano S (2000) In vivoconformation of mitochondrial DNA revealed by pulsed-fieldgel electrophoresis in the true slime mold, Physarum polyceph-alum. DNA Res 7:83–91

Sasaki N, Kuroiwa H, Nishitani C, Takano H, Higashiyama T,Kobayashi T, Shirai Y, Sakai A, Kawano S, Murakami-Mur-ofushi K, Kuroiwa T (2003) Glom is a novel mitochondrialDNA packaging protein in Physarum polycephalum and causesintense chromatin condensation without suppressing DNAfunctions. Mol Biol Cell 12:4758–4769

Stierum RH, Croteau DL, Bohr VA (1999) Purification and char-acterization of a mitochondrial thymine glycol endonucleasefrom rat liver. J Biol Chem 274:7128–7136

Sutovsky P, Moreno RD, Romalho-Santos J, Dominko T, SimerlyC, Schatten G (1999) Ubiquitin tag for sperm mitochondria.Nature 402:371–372

Sutovsky P, Moreno RD, Romalho-Santos J, Dominko T, SimerlyC, Schatten G (2000) Ubiquitinated sperm mitochondria,selective proteolysis, and the regulation of mitochondrialinheritance in mammalian embryos. Biol Reprod 63:582–590

Takano H, Abe T, Sakurai R, Moriyama Y, Miyazawa Y, NozakiH, Kawano S, Sasaki N, Kuroiwa T (2001) The complete DNAsequence of the mitochondrial genome of Physarum polyceph-alum. Mol Gen Genet 264:539–545

Tsuchimoto D, Sakai Y, Sakumi K, Nishioka K, Sasaki M,Fujiwara T, Nakabeppu Y (2001) Human APE2 protein ismostly localized in the nuclei and to some extent in the mito-chondria, while nuclear APE2 is partly associated with prolif-erating cell nuclear antigen. Nucleic Acids Res 29:2349–2360

Umen JG, Goodenough UW (2001) Chloroplast DNA methylationand inheritance in Chlamydomonas. Genes Dev 15:2585–2597

van Hoof A, Parker R (1999) The exosome: a proteasome forRNA? Cell 99:347–350

Zassenhaus HP, Hofmann TJ, Uthayashanker R, Vincent RD,Zona M (1988) Construction of a yeast mutant lacking themitochondrial nuclease. Nucleic Acids Res 16:3283–3296