INTRODUCTION

Starfish oocytes resume their maturation by 1-methyladenine(1-MA) treatment, and complete it without any arrest, followedby formation of the female pronucleus (Fig. 1A). The absenceof ‘metaphase arrest’ is a characteristic feature of starfishoocytes. After the germinal vesicle breakdown (GVBD), theycan be fertilized at any stage of their maturation (Fig. 1B,C).Thus, both oocytes and eggs are fertilizable. Irrespective of thetime of fertilization, later development is apparently normal. Ifthe eggs are fertilized after the completion of maturation andthe formation of female pronuclei, DNA replication is initiated30-50 minutes after fertilization (Nomura et al., 1991). Thisindicates that the DNA replication system of starfish will beestablished during this period of 30-50 minutes beginning fromfertilization. The process for preparing DNA replication wascalled the ‘postactivation process’ by Nomura et al. (1991)(Fig. 1C). In contrast, if eggs are fertilized before polar bodyformation, DNA replication immediately follows pronucleusformation (Nomura et al., 1991). This means that the postacti-vation process must be going on and completed in parallel withthe maturation process (Fig. 1B). From now on, eggs fertilized

before polar body formation will be called ‘early fertilizedeggs,’ and the eggs fertilized after the completion of the mat-uration process will be called ‘late fertilized eggs.’

Proliferating cell nuclear antigen (PCNA) has been a well-known protein essential for DNA replication in many types ofcells (Tan et al., 1986; Bravo et al., 1987; Prelich et al., 1987;see also Fairman, 1990, for review). Association of PCNA withnuclei during the DNA replication period (S phase) was firstfound in mouse NIH 3T3 cells fixed with methanol: PCNA wasdetected in the nuclei only during S phase, and it was easilyextracted from the nuclei during phases other than S phase(Bravo and Macdonald-Bravo, 1987). Extensive studies haveshown important roles for PCNA in DNA replication (Bravo,1986; Nakamura et al., 1986; Fairman et al., 1988). PCNA wasfinally identified as an auxiliary protein of DNA polymerase

δ(Bravo et al., 1987; Prelich et al., 1987). It is now certain thatPCNA binds to the replication fork, and is an indispensableprotein molecule for the DNA replication system (see van derVliet, 1989; Laskey et al., 1989, for review).

In 1989, Takasaki and his colleagues found that a humanantiserum from an auto immune patient recognized not onlyhuman PCNA but also higher plant PCNA (Suzuka et al.,

3291

Previous studies (Nomura et al. (1991)

Dev. Biol. 143, 289-296 (1993) Dev. Biol. 159, 288-297) determined the time ofDNA replication period (S phase) in starfish eggs fertilizedeither during or after oocyte maturation. Here proliferat-ing cell nuclear antigen (PCNA) localized within nuclei ofstarfish eggs was detected with an anti-PCNA humanantiserum. Using a confocal laser scanning microscope, athree-dimensional structure of the PCNA region wasanalyzed.

In eggs fertilized during maturation, PCNA started tolocalize within the nuclei at the same time as the initiationof the first S phase. During the S phase, the distribution oflocalized PCNA in a three-dimensional view coincided withthe chromatin distribution. After the S phase, PCNAremained localized within the nuclei, but its distribution nolonger coincided with the chromatin distribution.

In eggs fertilized after maturation, however, PCNAstarted to localize within the female pronuclei about 10minutes ahead of the first S phase. Localized PCNA

occupied only a limited region of the nuclei withoutdiffusing over the whole nuclear area. Chromatin distrib-uted around the peripheral region of the nuclei mostlyoutside the PCNA region. When the first S phase wasinitiated, the chromatin distribution became coincidentwith the PCNA region. Later behavior of PCNA was thesame as that of the eggs fertilized during maturation. Theprecocious localization of PCNA in those eggs fertilizedafter maturation simply demonstrates that the ‘postactiva-tion process’ for preparing DNA replication is triggered byfertilization and PCNA localization and S phase aresequentially initiated with a time-lapse. On the other hand,the simultaneous occurrence of them seen in those eggs fer-tilized during maturation indicates that the postactivationprocess must be going on in parallel with the maturationprocess.

Key words: starfish embryo, cell cycle, S phase, BrdU, PCNA,CLSM, three-dimensional reconstruction

SUMMARY

Nuclear distribution of proliferating cell nuclear antigen (PCNA) in fertilized

eggs of the starfish

Asterina pectinifera

Akira Nomura

Department of Zoology, Faculty of Science, Kyoto University, Kyoto 606, Japan

Present address: Tateyama Marine Laboratory, Ochanomizu University, Koh-yatsu, Umi-no-Hoshi, Tateyama, Chiba, 294-03, Japan

Journal of Cell Science 107, 3291-3300 (1994)Printed in Great Britain © The Company of Biologists Limited 1994

3292

1989). Dr Yoshinari Takasaki kindly gave this anti-PCNAhuman antiserum for me to test whether it recognizes starfishPCNA. In the previous study, I have already found a spatialcorrelation between DNA replication sites and chromatinstructure (Nomura et al., 1993), and my present question is:how does PCNA co-distribute with DNA replication sites orchromatin? Methanol-fixed nuclei were exclusively observedto study only those PCNA molecules incorporated into thereplication fork. The three-dimensional structure of the regionwhere PCNA and chromatin co-distribute was also observed.

MATERIALS AND METHODS

Cell culture and BrdU labelingThe starfish

Asterina pectinifera were collected on the coast ofWakasa Bay and Bohso Peninsula, Japan, and stored in aquaria inTateyama Marine Laboratory, Tateyama.

As described in a previous paper (Nomura et al., 1991), the oocyteswere: (i) induced to mature with 1 µM 1-methyladenine (1-MA); (ii)inseminated; and (iii) timed visually for their developmental events.Germinal vesicle break down (GVBD), and first and second polarbody (PB1 and PB2) formation occurred, respectively, about 20minutes, 60-70 minutes and 90-100 minutes after 1-MA treatment at20°C (cf. Figs 2, 3). They were inseminated at 40 minutes or 2 hoursafter the 1-MA treatment. Eggs inseminated 40 minutes after the 1-MA treatment are the ‘early fertilized eggs’ and eggs inseminated 2hours after 1-MA are the ‘late fertilized eggs.’ Oocytes, eggs, andembryos were cultured in a thermostatic bath at 20°C.

The fertilization membrane of the eggs was removed 5-10 minutesafter the insemination as described in a previous paper (Nomura etal., 1993). Samples of eggs taken from the culture dish every 4minutes were labeled with bromodeoxyuridine (BrdU; SigmaChemical Co., St Louis, MO) for 3 minutes in seawater containing 10mM BrdU and 0.05% Triton X-100 (Sigma). After the 3 minutelabeling, they were fixed with methanol for 30 minutes at room tem-perature. They were then rinsed with phosphate-buffered saline (PBS)at pH 7.3, three times. Some eggs in each sample were processed fordetection of the BrdU, and the other eggs were processed for stainingfor PCNA.

Immunocytological procedures BrdU incorporated into the nuclei of eggs was detected with an anti-

BrdU monoclonal mouse antibody to time the length of S phase.About 100 eggs were attached to a glass slide coated with poly-L-lysine (Sigma). The eggs were incubated in a drop of a reagent or anantibody in a moist chamber according to the following order: (i) 2M HCl for 1 hour to denature the DNA; (ii) monoclonal mouseantibody against BrdU (Becton Dickinson ImmunocytometrySystems, San Jose, CA) diluted to 1:6 with PBS for 2 hours; (iii)biotinylated sheep monoclonal antibody against mouse Ig(Amersham, Buckinghamshire, England) diluted to 1:50 with PBS for1 hour; and (iv) FITC-conjugated streptavidin (Amersham) diluted to1:100 with PBS for 1 hour.

PCNA was detected with an anti-PCNA human antiserum, andDNA was counterstained with 4′-6-diamidino-2-phenylindole (DAPI;Sigma), or propidium iodide (PI; Sigma). The eggs, attached to a glassslide, were incubated according to the following order: (i) humanantiserum against PCNA (a gift from Dr Yoshinari Takasaki; Suzukaet al., 1989) diluted to 1:100 with PBS for 2 hours; (ii) biotinylatedsheep monoclonal antibody against mouse Ig (Amersham, Bucking-hamshire, England) diluted to 1:50 with PBS for 1 hour; (iii) FITC-conjugated streptavidin (Amersham) diluted to 1:100 with PBS for 1hour; and (iv) 0.2 µg/ml DAPI in PBS or 25 mg/ml PI in PBS for 1hour. To use PI, eggs were preincubated in 2 mg/ml RNase A (Sigma)at 37°C to digest the endogenous RNA that is stainable with PI. WhenPBS was used for antibody dilution, it routinely contained 1% bovineserum albumin (Sigma).

In western blot experiments, a 36×103 Mr band, consistent withhuman PCNA (Mathews et al., 1984), was detected by the antiserumagainst PCNA. Here the molecules recognized by this antiserum willbe tentatively regarded as ‘starfish PCNA.’

Finally, a drop of glycerol containing 10% 10mM Tris-HCladjusted to pH 8.0 and 2.3% 1,4-diazabicyclo[2.2.2.]octane (DABCO;Sigma) was added to the eggs as an anti-fluorescence bleaching agentand covered with a coverslip. The specimens were examined usingeither a fluorescence microscope (Nikon Optiphoto with EFD2,Tokyo, Japan) or a confocal laser scanning microscope (CLSM;MRC600, Bio-Rad, Tokyo, Japan) equipped to an Axioplan epifluo-rescence microscope (Carl Zeiss, Germany).

Processing the images obtained by conventionalfluorescence microscopyImages obtained by conventional fluorescence microscopy wererecorded on photographic film and were digitized by a filmscanner(LS-3510AF, Nikon, Tokyo, Japan), and stored in a storage device of

A. Nomura

Maturation processPostactivation process

Maturation process

1S 2S

Postactivation process

Egg

Egg

1S

GVBD PB1 PB2 pronuclei

GVBD PB1 PB2 pronucleus

1-MA 1hr 2hr 3hr

Fertilization

Maturation process Egg

GVBD PB1 PB2 pronucleus

Fertilization

A

B

C

Fig. 1. Schematic illustration of temporalcorrelation between the maturation processand the postactivation process. (A) Unfertilized eggs. Progression of thematuration process is represented by the whiteto gray gradation. The subsequent gray zoneindicates that the maturation process has beencompleted. ‘Egg’ means a mature oocyte. (B) Eggs fertilized during the maturationprocess. The progression of the postactivationprocess is illustrated by the inner rectanglewith the gradient of white to dark gray. Sincethe postactivation process has been completed,the first S phase (1S) is initiated when thematuration process is completed. 2S, thesecond S phase. (C) Eggs fertilized aftermaturation. Since the maturation process hasbeen completed, 1S is immediately initiatedwhen the postactivation process is completed.GBVD, germinal vesicle breakdown; PB1 and2, first and second polar body formation.

3293Nuclear distribution of PCNA in starfish eggs

a personal computer (Macintosh Quadra 950, Apple Computer Inc.,Tokyo, Japan). Several kinds of digital image processing, such asnoise reduction and enhancement of the contrast, were made on theoriginal images with either an application software (Photoshop 2.5,Adobe Systems Inc., Mountain View, CA) for digitally processingphotographic images or a specific filtering software (Kai’s power tool2.0, Harvard Systems Corp., Santa Monica, CA).

Analyzing the images obtained with CLSMOptical sections of stained PCNA and chromatin were obtained withCLSM serially at a regular step of 0.48 µm. Each image was digitizedto numerical values (0-255) of 16×16 pixels for an area of 1 µm ×1 µm on an optical section. The digitized images were stored in astorage device of a computer (AX/2, Nimbus, Research MachinesLtd, Oxford, UK) attached to the CLSM. The numerical datawere exported to a Macintosh Quadra 950 using a data-conversionutility software (Access PC 2.1, Insignia Solutions Inc.). Severalkinds of digital image processing were made on the digitized imageswith either a digital built-in filter within Photoshop or Kai’s powertool.

The digitized images of the serial section at 0.48 µm step are toofar apart for three-dimensional reconstruction. Seven interpolatedimages were then computed for each pair of neighboring images, andinserted between them. This interpolation was done with an applica-tion software (Morph 2.0, Gryphon Software Corp., San Diego, CA)for transforming digital graphic images. As a result, a single pixelcomes to represent the brightness of a cube of 1/16 µm × 1/16 µm× 0.48/8 µm. All the digitized images including the interpolatedimages were integrated into a three-dimensional image with an appli-cation software (VoxelView/Mac 1.0, Vital Images, Inc., Fairfield,IA).

RESULTS

Timing the nuclear localization of PCNAStarfish oocytes or eggs were fertilized either 40 minutes (earlyfertilized eggs) or 2 hours (late fertilized eggs) after 1-MAtreatment. Their developmental events such as GVBD, polarbody formation, and cleavages were determined visually. BrdUincorporated into the nuclei was detected with the anti-BrdUantibody. In this study, time of S phase is routinely defined asthe time when more than 50% of eggs are labeled with BrdU.PCNA localized within the nuclei was also detected with theanti-PCNA antiserum. The percentages of PCNA-positive cellsand BrdU-positive cells are shown in Figs 2 and 3.

In the early fertilized eggs, both the nuclear localization ofPCNA and the first S phase were simultaneously initiated (Fig.2A). In the late fertilized eggs, however, PCNA started tolocalize about 10 minutes before the initiation of the first Sphase (Fig. 2B). The marked contrast in the time of PCNAlocalization between the two types of eggs will be referred toagain. PCNA remained localized within the nuclei for 10-15minutes after the end of the first S phase.

Localization of PCNA during the second and the third cellcycle was also examined (Fig. 3). In the early fertilized eggs,the behavior of PCNA was the same in each cell cycle as far asexamined: it started to localize at the same time as the initiationof S phase, remained localized within the nuclei, and disap-peared before the start of each metaphase (Fig. 3A). In the latefertilized eggs, PCNA had already localized within the nuclei

0

50

100

1 2 3 4 5Time after 1-MA treatment (hr)

Egg

s (%

)

Fertilization

GVBD PB1 PB2 CL4CL2CL1 CL3

BrdU

PCNA

0

50

100

1 2 3 4 5Time after 1-MA treatment (hr)

Egg

s (%

)

Fertilization

GVBD PB1 PB2 CL2CL1 CL4CL3

BrdU

PCNAB

A

Fig. 2. Time course of early four cell cyclesshowing a period of PCNA localization withinpronuclei at the first S phase in the earlyfertilized eggs (A) and in the late fertilized eggs(B). Solid circles and curves show thepercentages of PCNA-positive eggs. Graycircles and curves are the percentages of eggslabeled with BrdU. Crosses and broken curvesshow germinal vesicle breakdown (GVBD), firstand second polar body formation (PB1 and PB2)and the first through fourth cleavages (CL1-CL4).

3294

when the first S phase was initiated. In the second and the thirdcell cycle, the behavior of PCNA was the same as in the earlyfertilized eggs: it started to localize at the same time as theinitiation of S phase, remained localized within the nuclei, anddisappeared before the start of each metaphase (Fig. 3B).

If not fertilized, formed pronuclei remain quiescent for hourswithout visible change. PCNA usually does not becomelocalized to such pronuclei. Occasionally, however, I foundsome batches whose eggs were stained with anti-PCNAantiserum at the pronucleus. The batch shown in Fig. 3 is anexample. Yet the time of the localization was widely scatteredamong eggs (Fig. 3C), and I have no explanation of such anunusual PCNA localization.

The distribution of localized PCNA and thechromatin structurePCNA localized within nuclei was detected with the anti-PCNA antiserum and stained with FITC. Nuclear DNA wascounterstained with DAPI. Fig. 4 illustrates a set of pictures ofeggs, derived from the batch shown in Fig. 2, taken with a con-ventional fluorescence microscope. The first S phases areindicated by red frames surrounding the picture.

The results from the early fertilized eggs are shown incolumns A (chromatin structure) and B (localized PCNA).Before the first S phase (one hour and 44 minutes after 1-MAtreatment, cf. Fig. 2A), both egg and sperm nuclei had formedseveral small regions, indicating the formation of karyomeres

A. Nomura

0

50

100

1 2 3 4 5Time after 1-MA treatment (hr)

Egg

s (%

)

Fertilization

GVBD PB1 PB2 CL4CL2m2 a2CL1m1a1 CL3m3 a3

BrdU

PCNA

0

50

100

1 2 3 4 5Time after 1-MA treatment (hr)

Egg

s (%

)

Fertilization

GVBD PB1 PB2 CL2m2 a2CL1m1a1 CL3m3 a3

BrdU

PCNA

0

50

100

1 2 3 4 5

GVBD PB1 PB2 PCNA

Time after 1-MA treatment (hr)

Egg

s (%

)

C

B

A

Fig. 3. Time course of early cell cycles showingperiods of PCNA localization within nuclei and Sphases in early fertilized eggs (A), in late fertilizedeggs (B), and in unfertilized eggs (C). Opencircles and thin curves are first through thirdmetaphases (m1-m3) and anaphases (a1-a3). Othersymbols and curves are the same as in Fig. 2.

3295Nuclear distribution of PCNA in starfish eggs

Fig. 4. Localized PCNA within nuclei in the fertilized eggs of starfish. PCNA was detected by the anti-PCNA antiserum, and stained with FITC(columns B, D). DNA was also stained with DAPI (columns A, C). A series of images obtained from the early fertilized eggs are shown incolumns A and B, and those obtained from the late fertilized eggs are shown in columns C and D. These two kinds of series were obtained fromthe same batch of eggs shown in Fig. 2. The numerals at the lower left corner of columns A and C indicate the time of fixation measured from1-MA treatment (hour:minutes, for example 1:44 means the image obtained from eggs fixed at 1 hour and 44 minutes after 1-MA treatment). f,female chromatin; m, male chromatin. Bar, 10 µm. The red frames surrounding the pictures at 1:56 and 2:28 indicate S phase as detected byBrdU label (cf. Fig. 2).

3296

(Fig. 4, 1:44A). PCNA was not yet localized either in femaleor male karyomeres (Fig. 4, 1:44B). During the first S phase,localized PCNA was detected in both female and malepronuclei (Fig. 4, 1:56). The distribution of localized PCNAappeared similar to that of chromatin. After the end of the Sphase, PCNA remained localized, with the distribution nearlycoincident with that of chromatin (Fig. 4, 2:04). Whenchromatin became condensed, PCNA was still detected but thedistribution was limited within a portion of nuclei (Fig. 4,2:12). When chromatin condensed further, localized PCNAwas not detected (Fig. 4, 2:16). I have not clearly quantifiedthe correlation between a condensation level of chromatin andthe localization of PCNA.

The results from the late fertilized eggs are shown in columnC (chromatin) and D (PCNA). In this type of eggs, as was pre-viously reported, female pronuclei had been formed at fertil-ization (Nomura et al., 1991). Male pronuclei derived from theincorporated sperm nuclei were much smaller than the femalepronuclei (Fig. 4, 2:12C). Soon after fertilization, localizedPCNA was not yet detected either in the female or the malepronuclei (Fig. 4, 2:12 D). As already shown in Fig. 2B, about10 minutes before the initiation of the first S phase, PCNAbegan to localize within the female pronuclei (Fig. 4, 2:24D).It should be noted that the distribution of localized PCNA wasnot coincident with chromatin that mainly distributed along theperipheral region of the nuclei (Fig. 4, 2:24C). The nuclearstructure will be three-dimensionally examined later. Duringthe first S phase (red frames), localized PCNA was detected inboth female and male pronuclei (Fig. 4, 2:28D), and its distri-bution was coincident with chromatin (Fig. 4, 2:28C). After theend of the S phase, PCNA remained localized, with the distri-bution nearly coincident with that of chromatin. Even withincondensing nuclei (Fig. 4, 2:56), PCNA was still detected, butthe distribution of PCNA was limited within a portion of thenuclei. When chromatin condensed further, localized PCNAwas not detected (Fig. 4, 3:08).

In both early and late fertilized eggs, during the second cellcycle, the distribution of localized PCNA was the same as thatof the first cell cycle in the early fertilized eggs.

Three-dimensional images of the PCNA region andthe chromatin structureUsing a computer graphics program, a structure where

localized PCNA was distributed was three-dimensionallyimaged on the computer display. Localized PCNA wasdetected with the anti-PCNA antiserum and stained with FITC.Nuclear DNA was counterstained with PI instead of DAPI,since the CLSM used in this study did not detect the DAPI-signal. In my illustration of the reconstructed structure, PCNAwas shown in green color, and the chromatin was shown in red.At first, images of the whole nuclei were reconstructed bysimply stacking the optical serial sections, and the outermostsurfaces of the nuclei were illustrated. To visualize the innerpart of the nuclear structure, two kinds of modifications weremade. One is dividing the reconstructed structure into severalpartitions along an axis. They were illustrated as they wereseparated from each other. These types of image will be called‘partitioned images.’ The other modification is to make trans-parent the green area of localized PCNA. By this modification,even the chromatin structure lying behind the green areabecame partially visible through it. These types of image willbe called ‘transparent images.’ Images of both PCNA region(green) and chromatin structure (red) were merged. Thus, thearea of chromatin where PCNA was co-distributed appearsyellow. Figs 5-9 show the results obtained from the batchshown in Fig. 2.

A reconstructed image of the male and female pronucleiduring the first S phase in the early fertilized eggs is shown inFig. 5A, and its four partitioned images are shown in A1-A4.At this stage, the distribution of localized PCNA extended tothe whole pronuclei, and entirely coincided with the chromatindistribution, as shown by a predominant yellow area in the par-titioned images (Fig. 5, A1-A4). After the end of the first Sphase, localized PCNA was distributed only within a certainportion of the nuclei (Fig. 6A,B). Most PCNA was not co-dis-tributed with the chromatin. Some chromatin assumed avesicular appearance, probably representing forming chromo-somes. Some of them were wrapped around a zone of thePCNA region (Fig. 6, B1-B4). This is well illustrated by themosaic pattern of the green, red, and yellow areas in the par-titioned images (Fig. 6, A1-A4).

Nomura et al. (1993) reported that the chromatin of thefemale pronuclei in the late fertilized eggs showed a hollowstructure. This hollow structure was also observed in the latefertilized eggs examined in this study. Chromatin was distrib-

A. Nomura

Fig. 5. Reconstructed three-dimensional images of the PCNAregion and the chromatin structure ofmale and female pronuclei. Thisimage was made from the batchshown in Fig. 2. The early fertilizedeggs were fixed at 1 hour and 56minutes after 1-MA treatmentcorresponding to the time when mosteggs are PCNA positive and BrdUpositive (cf. Fig. 2A). LocalizedPCNA was stained with FITC(green), and DNA was stained with PI(red). Since two images were mergedinto one, the yellow area representsthe co-distribution of both PCNA andchromatin. (A) Outermost surface of

the reconstructed structure. (A1-A4) Four partitioned images made from the structure (A) to illustrate the inner part of the structure. Eachimage was reconstructed from optical serial sections collected with CLSM, using the application software on the computer. Bar, 10 µm.

3297Nuclear distribution of PCNA in starfish eggs

uted mainly at the periphery and was sparsely distributed at thecentral region of the nuclei (Fig. 7, A1-A4).

Before the first S phase (Fig. 7), localized PCNA (green)clustered on one side of the inner space of the nuclei withoutdiffusing over the whole nuclear area. A part of the chromatinwas found within the PCNA region (Fig. 7, B1-B4). This iswell illustrated by the mosaic pattern of the partitioned images(Fig. 7, A1-A4). When the first S phase was initiated (Fig. 8),the mosaic pattern disappeared and the yellow area extendedwidely in the partitioned images (Fig. 8, A1-A4). After the endof the first S phase, the three-dimensional structure of thenuclei was quite similar to that of the same stage in the earlyfertilized eggs shown in Fig. 6. Localized PCNA was distrib-uted only within a certain portion of the nuclei (Fig. 9A,B).Most PCNA was not co-distributed with the chromatin. Somevesicular chromatin was wrapped around a zone of the PCNAregion (Fig. 9, B1-B4). This is well illustrated by the mosaicpattern in the partitioned images (Fig. 9, A1-A4).

Going back to Figs 7 and 8, I notice that the clustering ofPCNA distribution (green + yellow) on one side of the nucleiseen before the first S phase (Fig. 7) persists through the Sphase (Fig. 8). This persistence of the clustering PCNA will bereferred to in the Discussion. On the other hand, the chromatin(red + yellow) distributes around the peripheral area of thenuclei mostly outside the PCNA region before the S phase (Fig.7), and then becomes coincident with the PCNA region withthe initiation of the S phase (Fig. 8).

DISCUSSION

Use of the anti-PCNA human antibody for detectingthe nuclear localization of starfish PCNATo detect a starfish equivalent of PCNA, the human anti-PCNAantiserum was used. This antiserum is known to recognizePCNA of various species, from higher plant PCNA to mammal

Fig. 6. Reconstructed three-dimensionalimages of the PCNA region and the chromatinstructure of a nucleus. The early fertilized eggswere fixed at 2 hours and 12 minutes whenmost eggs are PCNA positive but BrdUnegative (cf. Fig. 2A). (A) Outermost surface.(A1-A4) Four partitioned images made fromthe structure (A). (B) Transparent image madefrom (A). The green area of (A) was madetransparent. (B1-B4) Four partitioned imagesmade from the structure (B). Otherexplanations are as for Fig. 5.

Fig. 7. Reconstructed three-dimensionalimages of the PCNA region and thechromatin structure of a female pronucleus.The late fertilized eggs were fixed at 2 hoursand 24 minutes when most eggs are PCNApositive but BrdU negative (cf. Fig. 2B). (A)Outermost surface. (A1-A4) Four partitionedimages made from the structure (A). (B)Transparent image made from (A). The greenarea of (A) was made transparent. (B1-B4)Four partitioned images made from thestructure (B). Other explanations are as forFig. 5.

3298

PCNA (Suzuka et al., 1989). In the present study, the moleculein starfish eggs recognized by this antiserum was tentativelycalled ‘starfish PCNA,’ on the basis of western blot analysis.I found that the molecules recognized by the antiserum becomelocalized to chromatin in S phase nuclei, as do genuine humanPCNA molecules. This strongly indicates that the moleculeshave such a function in DNA replication as to be properlycalled ‘starfish PCNA.’

In DNA replication, PCNA molecules firmly bind to thereplication fork and are not easily extractable (see van derVliet, 1989; Laskey et al., 1989, for review). Based on resultsobtained by Bravo and McDonald-Bravo (1987), localizedPCNA detected in this study is considered to represent thosethat tightly bind to nuclei and are not extracted from nucleiwith the methanol fixation.

It should be a very interesting subject to relate the distribu-tion of localized PCNA with that of DNA replication sites, by

simultaneous detection of PCNA and BrdU. Working prelim-inarily on starfish eggs, I realized, however, that the signal oflocalized PCNA is often weakened after HCl treatment, aprocedure essential for the anti-BrdU antibody to recognizeincorporated BrdU. In a foregoing paper (Nomura et al., 1993),I have already confirmed that the distribution of DNA replica-tion sites detected by a BrdU pulse labeling was identical tothe chromatin distribution during the first and second S phasesof starfish eggs. Based on this observation, the chromatin dis-tribution was taken to represent the sites of DNA replicationin this paper. Efforts were recently made to overcome the dif-ficulty in simultaneously detecting DNA replication sites andPCNA: Kill et al. (1991) used biotin-11-dUTP to label theDNA replication sites, and Humbert et al. (1992) adopted anenzymic procedure for denaturing the DNA instead ofchemical denaturation. I have not yet established if thesemethods apply to starfish eggs.

A. Nomura

Fig. 8. Reconstructed three-dimensionalimages of the PCNA region and the chromatinstructure of a female pronucleus. The latefertilized eggs were fixed at 2 hours and 28minutes when most eggs are PCNA positiveand BrdU positive (cf. Fig. 2B). (A) Outermost surface. (A1-A4) Fourpartitioned images made from the structure(A). (B) Transparent image made from (A).The green area of (A) was made transparent.(B1-B4) Four partitioned images made fromthe structure (B). Other explanations are as forFig. 5.

Fig. 9. Reconstructed three-dimensionalimages of the PCNA region and the chromatinstructure of a nucleus. The late fertilized eggswere fixed at 2 hours and 56 minutes whenmost eggs are PCNA positive but BrdUnegative (cf. Fig. 2B). (A) Outermost surface.(A1-A4) Four partitioned images made fromthe structure (A). (B) Transparent image madefrom (A). The green area of (A) was madetransparent. (B1-B4) Four partitioned imagesmade from the structure (B). Otherexplanations are as for Fig. 5.

3299Nuclear distribution of PCNA in starfish eggs

Temporal correlation between the nuclearlocalization of PCNA and the first S phase A delay between the nuclear localization of PCNA and theinitiation of DNA replication has been noted (Celis and Celis,1985; Hutchison and Kill, 1989; Nakane et al., 1989). Thetechnics improved by Kill et al. (1991) and Humbert et al.(1992) have led them to confirm the presence of such a delay.Kill et al. (1991) observed the delay in human diploid fibrob-lasts cells and Xenopus eggs, and proposed the idea that theentry into S phase is biphasic: assembly of the replicationcomplex including PCNA is followed by a process in whichthose complexes are used. Humbert et al. (1992) also observedthe delay and ascribed it to a time-lapse between the assemblyof replication complexes and the initiation of DNA replication.Most recently, Baptist et al. (1993) proposed an unknownprocess ‘X,’ programed after the nuclear localization of PCNA,with a function necessary for DNA replication.

As I found in the late fertilized eggs of starfish, the preco-cious localization of PCNA within female pronuclei about 10minutes before the initiation of the first S phase appears to gen-eralize those delays mentioned above. It should be noted,however, that the occurrence of the delay depends on the timeof fertilization in starfish: in the early fertilized eggs, PCNAstarted to localize at the same time as the initiation of the firstS phase. What is unique in starfish eggs is that the time of fer-tilization alone makes such a contrast in PCNA localization.

Nomura et al. (1991, 1993) proposed the term ‘postactiva-tion process’ to imply a cascade preparing for DNA replica-tion in starfish eggs. The delay observed in the late fertilizedeggs indicates that the postactivation process induces thenuclear localization of PCNA at 20 minutes after fertilizationand DNA replication at 30 minutes. Thus, localized PCNAturns out to be a marker midway in the progress of the postac-tivation process. In the early fertilized eggs, both nuclear local-ization of PCNA and initiation of the first S phase immediatelyfollowed the completion of the maturation process with theformation of male and female pronuclei. These results supportthe idea that the postactivation process must be going on andmust complete in parallel with the maturation process.

Spatial correlation between the distribution of PCNAand the chromatin structureIn addition to the delay between the nuclear localization ofPCNA and the initiation of DNA replication, Kill et al. (1991)observed that the preceding distribution pattern of PCNA oftenresembles the expected distribution pattern of DNA replicationsites. Humbert et al. (1992) reported that a few of the PCNA-positive/BrdU-negative cells show a PCNA distribution verysimilar to the distribution at an early S phase.

Present studies on the late fertilized eggs of starfishconfirmed such a similarity, or a persistence of the PCNA dis-tribution as a cluster on one side of the nucleus before andduring S phase. By merging PCNA and DNA label in the three-dimensionally reconstructed images on the computer display,I further observed that the precocious distribution of PCNA isnot identical to the chromatin distribution. This implies thatmost PCNA molecules precociously localized within femalepronuclei cannot contact with the chromatin before the first Sphase. When the S phase is initiated, the distribution of PCNAand chromatin become coincident with each other.

The persistence of the PCNA cluster before and during thefirst S phase in turn will imply a dynamic relocation of thechromatin mass towards the PCNA cluster for the onset ofDNA replication. Thus, S phase is correlated with a redistrib-ution of chromatin to the PCNA sites in the nucleus. It appearsthat such a relocation of chromatin mass has been unnoticedheretofore, but the three-dimensional image analysis made inthe present study shows that it is actually the case at the startof the first S phase at least in the late fertilized eggs of starfish.If PCNA distribution can be looked upon as sites of functionalreplication complexes, then my observation suggests that repli-cation complexes are formed not strictly at replication forksbut as discrete sites in the nucleus that then become future sitesof replication.

In the early fertilized eggs and the second cell cycle of thelate fertilized eggs, the precocious localization of PCNA wasabsent, and the distribution of PCNA was coincident with thechromatin distribution at the start of the localization. Thissuggests that PCNA properly reaches the DNA replication forkas soon as the nuclei has been formed. The direct localizationof PCNA into the DNA replication fork should result in therapid start of DNA replication, which will help to shorten thecell cycle periods in early development.

A built-in stability of the cell-cycle in starfishembryosNo matter how the S phase initiation has a delay behind thenuclear localization of PCNA, the disappearance of PCNA wasalways scheduled at the initiation of chromatin condensation(cf. Fig. 3). The second and the third cell cycle was quiteregular irrespective of the time of fertilization. In a previouspaper, Nomura et al. (1993) suggested that ‘the cell cycledriving mechanism that is perturbed by the interaction of thematurational process with the postactivation process could bestabilized when the first round of the cell cycle is completed’.Timely disappearance of localized PCNA from nuclei of thestarfish eggs as observed in the present study provides furtherevidence for the operation of a certain device to ultimatelystabilize the cell cycle driving mechanism. As discussed earlier(Nomura et al., 1993), this built-in stability of the cell-cycledriving mechanism will be an essential device for normaldevelopment of starfish eggs that are naturally fertilizable atany time of maturation without the ‘metaphase arrest’ knownin most animal eggs.

I thank Dr Mitsuki Yoneda for reading the manuscript, and givinghelpful discussions. The western blotting experiment was done by MrTatsuya Ueno of Kyoto University. I also thank Dr YoshinariTakasaki in Juntendo University for giving me the anti PCNA-antiserum. I express my gratitude to Dr Shoji Tanaka in MitsubishiKasei Institute of Life Sciences for instructing me to use the CLSMat the institute and encouraging me to complete this work. Most ofthe work was performed at Tateyama Marine Laboratory, Ochano-mizu University, Umi-no-Hoshi, Tateyama, Chiba. I am grateful toDr Sin-ichi Nemoto, the director and Mr Mamoru Yamaguchi and MrsHisae Kikuchi, staff of the laboratory for providing me with the facil-ities for research. I also thank Mr Tamio Arano in the ‘Zero-OneShop, Cannon (Esaka branch)’ for his technical support to mypersonal computer. I am grateful to Mr Toshimi Miyata, Mr Shin-ichiKajiya and other members in Risuken for giving me an opportunityto master the computer programing. This work was supported in partby a grant-in-aid for JSPS fellows.

3300

REFERENCES

Baptist, M., Dumont, J. E. and Roger, P. P. (1993). Demonstration of cellcycle kinetics in thyroid primary culture by immunostaining of proliferatingcell nuclear antigen: differences in cyclic AMP-dependent and -independentmitogenic stimulations. J. Cell Sci. 105, 69-80.

Bravo, R. (1986). Synthesis of the nuclear protein cyclin (PCNA) and itsrelationship with DNA replication. Exp. Cell Res. 163, 287-293.

Bravo, R., Frank, R., Blundell, P. A. and MacDonald-Bravo, H. (1987).Cyclin/PCNA is the auxiliary protein of DNA polymerase-δ. Nature 326,515-517.

Bravo, R. and MacDonald-Bravo, H. (1987). Existence of two populations ofcyclin/proliferating cell nuclear antigen during the cell cycle: associationwith DNA replication sites. J. Cell Biol. 105, 1549-1554.

Celis, J. E. and Celis, A. (1985). Cell cycle-dependent variations in thedistribution of the nuclear protein cyclin proliferating cell nuclear antigen incultured cells: subdivision of S phase. Proc. Nat. Acad. Sci. USA 82, 3262-3266.

Fairman, M., Prelich, G., Tsurimoto, T. and Stillman, B. (1988).Identification of cellular components required for SV40 DNA replication invitro. Biochem. Biophys. Acta 951, 382-388.

Fairman, M. P. (1990). DNA polymerase δ/PCNA: actions and interactions. J.Cell Sci. 95, 1-4.

Humbert, C., Santisteban, M. S., Usson, Y. and Robert-Nicoud, M. (1992).Intranuclear co-location of newly replicated DNA and PCNA bysimultaneous immunofluorescent labeling and confocal microscopy in MCF-7 cells. J. Cell Sci. 103, 97-103.

Hutchison, C. J. and Kill, I. R. (1989). Changes in the nuclear distribution ofDNA polymerase alpha and PCNA/cyclin during the progress of the cellcycle, in a cell-free extract of Xenopus eggs. J. Cell Sci. 93, 605-613.

Kill, I. R., Bridger, J. M., Campbell, K. H. S., Maldonado-Codina, G. andHutchison, C. J. (1991). The timing of the formation and usage of replicase

clusters in S-phase nuclei of human diploid fibroblasts. J. Cell Sci. 100, 869-876.

Laskey, R. A., Fairman, M. P. and Blow, J. J. (1989). S phase of the cellcycle. Science 246, 609-614.

Mathews, M. B., Bernstein, R. M., Franza, B. R. Jr and Garrels, J. I.(1984). Identity of the proliferating cell nuclear antigen and cyclin. Nature309, 374-376.

Nakamura, H., Morita, T. and Sato, C. (1986). Structural organization ofreplicon domains during DNA synthetic phase in the mammalian nucleus.Exp. Cell Res. 165, 291-297.

Nakane, P. K., Moriuchi, T., Koji, T., Taniguchi, Y., Izumi, S. and Hui, L.(1989). Proliferating cell nuclear antigen (PCNA/cyclin): review and somenew findings. Acta Histochem. Cytochem. 22, 105-116.

Nomura, A., Maruyama, Y. K. and Yoneda, M. (1991). Initiation of DNAreplication cycle in fertilized eggs of the starfish, Asterina pectinifera. Dev.Biol. 143, 289-296.

Nomura, A., Yoneda, M. and Tanaka, S. (1993). DNA replication infertilized eggs of the starfish Asterina pectinifera. Dev. Biol. 159, 288-297.

Prelich, G., Tan, C. K., Kostura, M., Mathews, M. B., So, A. G., Downey, K.M. and Stillman, B. (1987). Functional identity of proliferating cell nuclearantigen and a DNA polymerase-δ auxiliary protein. Nature 326, 517-520.

Suzuka, I., Daidoji, H., Matsuoka, M., Kadowaki, K., Takasaki Y., NakaneP. K. and Moriuchi, T. (1989). Gene for proliferating-cell nuclear antigen(DNA polymerase δ auxiliary protein) is present in both mammalian andhigher plant genomes. Proc. Nat. Acad. Sci. USA 86, 3189-3193.

Tan, C. K., Castillo, C., So, A. G. and Downey, K. M. (1986). An auxiliaryprotein for DNA polymerase-d from foetal calf thymus. J. Biol. Chem. 261,12310-12316

van der Vliet, P., C. (1989). DNA replication. Curr. Opin. Cell Biol. 1, 481-487.

(Received 6 June 1994 - Accepted 16 August 1994)

A. Nomura