In situ preparation of the nuclear matrix of Physarum polycephalum:

ultrastructural and biochemical analysis of different matrix isolation

procedures

WERNER WAITZ

Institut fur Pctlliologische Anatomie, Miillerslr. 44, 6020 Innsbruck, Austria

and PETER LOIDL*

Institut fur Mediztmsche Chemie und Biochemie, Fritz-Pregl-Str. 3, 6020 Innsbruck, Austria

* Author for correspondence

Summary

A novel method for in situ preparation of nuclearmatrix from whole plasmodia of Physarum poly-cephalum without isolation of nuclei is presented.Plasmodia are encapsulated in agarose beads andafter solubilization of the cytoplasm the nuclearmatrix is prepared. With this quick and easytechnique nuclear matrix can be reproduciblyprepared with perfect recovery. We compared theultrastructural and biochemical properties of thematrix after three different matrix isolation pro-cedures: preparation with high salt, ammoniumsulphate and lithium diiodosalicylic acid. The

results show that the ultrastructure and proteincomposition of the three types of matrix are verysimilar or even identical. We conclude that manyof the conflicting results on nuclear matrix in theliterature are due to perturbations of nuclearintegrity during the isolation of nuclei. For thisreason the new in situ method is an importantapproach in the standardization of nuclearmatrix isolation.

Key words: nuclear matrix, nucleus, chromatin, Physanon,matrix proteins.

Introduction

The nuclear matrix is an operational term that charac-terizes residual structural entities of the nucleus aftertreatment with detergent, salt and nucleases (Berezney& Coffey, 1974). The complex processes occurring inthe nucleus suggest a structural network that directsmacromolecules to the site of their action (e.g. RNApolymerasc I must be directed to the nucleolus, RNApolymerase II and regulatory proteins to activechromatin domains) and on the other hand providesattachment sites for constituents of various processes(e.g. DNA replication, RNA splicing). The concept ofthe nuclear matrix as a basic nuclear substructure isuseful in explaining the spatial organization of thecomplex biochemical processes of the nucleus.

There are numerous reports in the literature on theultrastructure of the nuclear matrix (for review, seeJournal of Cell Science 90, 621-628 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

Hancock, 1982; Bouteille et at. 1983), but relativelylittle is known about the biochemical composition(Pogo & Procyk, 1985). It consists of non-histoncproteins, small amounts of DNA and RNA and,depending on the isolation procedure, small quantitiesof histones.

We started to investigate the nuclear matrix of thelower eukaryote Physarum polycephalum in order toelucidate possible interactions between core histonesdiffering in their acetylation state and nuclear matrixcomponents (Loidl & Grobner, 1987; Loidl, 1988).One of the major difficulties in the preparation ofnuclear matrix is that slight changes in the isolationprotocol cause major alterations in the morphology ofthe residual structures obtained (Kaufmann et al.1981). The nuclear matrix is normally prepared afterisolation of nuclei. We had difficulty in obtainingmorphologically intact nuclear matrices with high re-

621

covery when using isolated nuclei as starting material.It has been shown that even differences in the nuclearisolation procedure cause severe differences in theinternal organization of the matrix (Smith et al. 1987).For this reason we established an in situ preparation ofthe nuclear matrix, in which residual nuclear structuresare prepared from whole cells without isolation ofnuclei. The procedure is based on work of Nilsson etal. (1983) and Jackson & Cook (1985), who encapsu-lated whole cells into agarose beads in order to manipu-late the cells for different purposes. We encapsulatedwhole microplasmodia of P. polyceplialum into agarosebeads and prepared nuclear matrices in situ by differ-ent methods. The methods differ in the mode ofdeproteinization of nuclei; extraction with high salt(Berezney & Coffey, 1977), two-step extraction with0-25 M-ammonium sulphate (Capco et al. 1982) andextraction with 25 mM-lithium diiodosalicylate (Mirko-vitch et al. 1984). Moreover, we compared nuclearmatrices after hypotonic and isotonic processing ofencapsulated plasmodia. The structure and biochemi-cal composition of the three different matrix prep-arations were compared. Our in situ preparation tech-nique yields reproducible residual structures withperfect recovery after all three extraction procedures.Comparison of the morphological and biochemicalproperties reveals that the nuclear matrices obtainedexhibit similar morphological and biochemical fea-tures, regardless of the preparation method. We there-fore conclude that the more crucial step in matrixpreparation is the manipulation of cells and the iso-lation of nuclei rather than the mode of deproteiniz-ation of nuclei.

Materials and methods

Materials

These were purchased from the following sources: agarose(type IX, ultra-low gelling temperature), aprotinin, sper-mine, spermidine, 3,5-diiodosalicylic acid (lithium salt),deoxyribonuclease I, thiodiglycol, tannic acid and cacodylicacid (sodium salt) from Sigma Chem. Co., St Louis, USA;digitonin and phenylmcthylsulphonyl fluoride from ServaFeinbiochcmica, Heidelberg, West Germany; paraffin(highly liquid) and p-formaldehyde from E. Merck, Darm-stadt, West Germany; glutaraldehyde (25%) from PolaronEquipment Ltd, Watford, UK; LR-white (hard grade) fromThe London Resin Co. Ltd, Surrey, UK.

Culture techniques

P. polyceplialum, strain M3b FII (a Wis 1 isolate) was used.Microplasmodia were cultivated axenically in semi-definednutrient medium (Daniel & Baldwin, 1964). Submersedcultures were maintained in Erlenmeyer flasks on a reciprocalshaker at 24°C. All culture techniques were carried out understerile conditions.

Encapsulation of microplasmodia into agamse beadsThis technique is based on the methods of Nilsson et al.(1983) and Jackson & Cook (1985). Microplasmodia werewashed with buffer I (20mM-Tris- HC1, pH7-3, 20 mM-KC1, 0-05mM-spermine, 0125 mM-spermidine, 0-5 mM-EDTA-KOH, pH7-3, 0-2niM-phenylmethylsulphonyl flu-oride, 1% (w/v) thiodiglycol, 0 5 % (w/v) aprotinin) at25°C. One volume of pelleted microplasmodia was mixedwith one volume of 2-5 % (w/v) agarose. Then five volumesof liquid paraffin were added and the mixture was shaken for30s on a rotary shaker at 25°C. The mixture was chilled onice for 5 min with occasional mixing. All subsequent manipu-lations were carried out at 4CC. Agarose beads containingmicroplasmodia were washed twice with an excess of buffer I.Centrifugation of the beads was always done at 700 g. Torelease the cytoplasm beads were washed with buffer Icontaining 0-75 % Triton X-100 and subsequently incubatedfor 5 min in this buffer, with gentle stirring. After centrifu-gation the beads were washed twice in buffer I. The washedbeads were then used for nuclear matrix preparation bydifferent methods.

High salt (NaCl). Beads were incubated for 15 min in 2 M-NaCl (in buffer I) at 4°C. After the incubation the beadswere washed twice in buffer II (20mM-Tris- HC1, pH7-3,20mM-KCl, 50mM-NaCl, 5mM-MgCl2, 0-05 mM-spermine,0-125 mM-spermidine, 0-1% (w/v) digitonin, 0-2 mM-phenylmethylsulphonyl fluoride, 0 5 % aprotinin). Beadssuspended in buffer II were incubated for 5 min at 25°Cbefore deoxyribonuclease I (50^gml~') was added andincubation was continued for 25min. Subsequently, thepreparation was washed three times in buffer I at 4°C.

Ammonium sulphate (AS). Beads in buffer I were adjustedto 0-25 M-ammonium sulphate and incubated for 15 min at4°C, with gentle stirring. After incubation the beads werewashed twice in buffer II before they were incubated withdeoxyribonuclease I as done for the high-salt preparation.After three washes with buffer I beads were again adjusted to0-25 M-ammonium sulphate and incubated for 15 min at 4°C.Finally the beads were washed twice with buffer I.

Lithium diiodosalicylate (LIS). Beads were washed inincubation buffer (25 mM-lithium diiodosalicylate, lOmM-Tr i sHCl , pH7-3, 2mM-KCl, 2mM-EDTA-KOH, pH7-3,025 mM-spermidine, 0-1% (w/v) digitonin, 0-2mM-phenyl-methylsulphonyl fluoride). Then beads were incubated inthis buffer for 15 min at 25°C. Beads were washed three timeswith buffer II and then the samples were digested withdeoxyribonuclease I as done for high-salt preparation. Afterdigestion the beads were washed three times in buffer I.

Samples of the matrix preparations were used for analysesby electron microscopic and electrophoretic techniques.

Isotonic processing of agarose-encapsulated plasmodia

The procedure follows the isotonic conditions described byJackson & Cook (1985). Microplasmodia were washed inisotonic buffer A (lOmM-Tris- HC1, pH7-3, 100mM-KCl,25mM-EDTA-KOH, pH7-3, 0-05 mM-spermine, 0-125 mM-spermidine, 0-2mM-phenylmethylsulphonyl fluoride, 1 %(w/v) thiodiglycol, 0-5% (w/v) aprotinin) and encapsulatedin agarose as described above. Beads were washed twice withan excess of isotonic buffer A. To solubilize the cytoplasm

622 W. Waitz and P. Loidl

beads were washed with buffer A containing 0-75 % TritonX-100 and subsequently incubated for 5 min in this buffer,with gentle stirring. After centrifugation the beads werewashed twice in buffer A, before they were incubated for15 min in 2M-NaCl. After incubation the beads were washedtwice in isotonic buffer B (like buffer A, except that EDTA isreplaced by 25mM-(NH4)2SC>4 and then subjected to deoxy-ribonuclease I treatment as described above.

Electrophoretic techniques

Samples of pelleted matrix preparations were suspended inSDS sample buffer (Laemmli, 1970) and incubated for 5 minin a boiling water bath. After centrifugation for 5 min at14 000 £ the supernatant was applied to a 16 cm longSDS-15% polyacrylamide slab gel (Laemmli, 1970). Gelswere stained with Coomassie Blue and destained by dif-fusion.

Electron microscopy

Samples of pelleted matrix preparations were suspended inbuffer III (6mM-TrisHCl, pH7-3, 10mM-KCl, 10mM-MgCl2, O'l M-sucrose). This buffer III represents the finalwashing step, where matrix preparations can be kept for up to30 min without serious damage. The samples were fixed in4 % formaldehyde (freshly prepared from paraformaldehyde)and 0 2 5 % glutaraldehyde in 0-lM-sodium cacodylate

buffer, pH7-3, at 4°C overnight. The samples were dehy-drated with increasing concentrations of ethanol, and 6%tannic acid was included in the 50%, 70% and 90% ethanolsteps. Subsequently, samples were transferred to LR-white(hard grade) with the following intermediate steps: 2 partsLR-white/1 part absolute ethanol for 1 h at room tempera-ture; two changes in pure LR-white at room temperatureovernight. Finally, the samples were polymerized at 56°C for24h. Ultrathin sections were made using a Reichcrt Ultracutmicrotome. Photographs were taken at 80 kV with a 60;i<mobjective aperture in a Zeiss EM 109 electron microscopeequipped with a rollfilm camera.

Results

We initially started to prepare nuclear matrices fromisolated nuclei basically following the procedure ofMitchelson et al. (1979). However, in our hands thesepreparations were not reproducible; in particular, therecovery of nuclear matrices (in relation to the isolatednuclei) varied considerably from one experiment to theother. Normally, 5 % of the isolated nuclei could finallybe identified as a nuclear matrix; the rest of the nucleidisintegrated during the preparation, strongly contami-nating the matrix preparation (results not shown).

Fig. 1. Microplasmodia of P. polycephaliim encapsulated in agarose beads as described in Materials and methods. Beadswere photographed under phase-contrast. X500.

In situ preparation of nuclear matrix 623

r<v. v . ^ .J

Fig. 2. Nuclear matrix of P. polyceplialuiu prepared after extraction with NaCl. Microplasmodia were encapsulated inagarose beads and the nuclear matrix prepared in situ as described in Materials and methods. The peripheral lamina (/), theinterchromatin matrix (im) and remnant nucleolar material (w) are indicated. A, X6200; B,X24000.

*'if . i iffy

Fig. 3. Nuclear matrix of P. polycephalum prepared after extraction with AS. A, X4900; B, X22600.

624 W. Waitz and P. Loidt

Fig. 4. Nuclear matrix of P. polycephalum prepared after extraction with LIS. A, X4800; B, X20500.

Moreover, we found it difficult to handle these matrixpreparations during the fixation and embedding forelectron microscopy. The extremely poor recovery ofnuclear matrices was not restricted to the NaCl-cxtraction procedure, but also occurred with AS andLIS extraction. This fact made biochemical as well asultrastructural data questionable, since 5 % of thenuclei cannot ever be representative of the wholenuclear population.

For this reason we adapted a recent technique formanipulating whole cells for biochemical purposes tothe preparation of nuclear matrix in whole microplas-modia of P. polycephalum. Based on methods ofNilsson et al. (1983) and Jackson & Cook (1985) weencapsulated whole microplasmodia into agarosebeads, which can easily and quickly be handled forvarious purposes. Microplasmodia are well preservedin the agarose beads as can be checked by phase-contrast microscopy (Fig. 1). Normally, a bead con-tains one to five individual microplasmodia. Whenagarose beads are supplied with nutrient mediumunder shaking, the encapsulated plasmodia continue togrow as judged from protoplasmic shuttle streaming,which continues for at least 5h. The beads can becollected by short centrifugation (3 min) at 500-700 jf.This technique offers the advantage that even singlebeads can be processed and transferred. During ma-nipulation of beads for nuclear matrix isolation no

disintegration or loss of nuclei was observed.Figs 2, 3 and 4 show electron micrographs of in situ

matrix preparations with NaCl (Fig. 2), AS (Fig. 3)and LIS (Fig. 4). The micrographs with lower magni-fication (A) reveal the reproducible morphology of thenuclear matrices within an agarose bead. It should benoted that we did not digest the matrix preparationswith RNase. All three preparations exhibit the charac-teristic features of nuclear matrix; the complex networkof the interchromatin matrix, the nuclear lamina and aresidual nucleolar structure. The matrix preparationsclearly differ from each other with respect to theremnant nucleolar structures. The NaCl-matrix(Fig. 2) contains less residual nucleolar material incomparison with AS- and LIS-matrix preparations(Figs 3 and 4). In the case of NaCl (Fig. 2) and LIS(Fig. 4) the matrix nuclei are surrounded by a halo thatis composed of extended DNA loops that are thendigested by DNase I. Following the matrix preparationstep by step in the electron microscope, the extensionand subsequent digestion of DNA loops can be fol-lowed (results not shown).

In contrast, AS-matrix preparations do not exhibitthis halo structure but are completely embedded in thesurrounding cytoplasmic, filamentous network. Obvi-ously the relatively mild deproteinization of nuclei bytwo steps with AS does not lead to such pronouncedloop expansion.

In situ preparation of nuclear matrix 625

Mrxi0 - 3 NaCI AS LIS MrX^O~3 Nucl

H1

21

12

6-5

O

o

1

H2A

H2B/H3

H4

6

Fig. 5. Protein composition of nuclear matrix preparations. Nuclear matrices were prepared with NaCI (lane 2), AS(lane 3) and LIS (lane 4) as described in Materials and methods. Proteins were analysed on SDS-15% polyacrylamide slabgels and stained with Coomassie Blue. Lane 1 contains low molecular weight markers, lane 5 high molecular weightmarkers. To facilitate the localization of histones lane 6 contains a nuclear lysate.

In accord with this structural feature of the AS-matrix is the relatively high content of residual corehistones in the AS preparation. Obviously, DNA is notas completely stripped off the histones by AS comparedwith NaCI or LIS (Fig. 5). Fig. 3B shows that ASnuclear matrices contain numerous dense chromatinmasses that could resemble the chromocentres recentlydescribed by Matsumoto el al. (1987).

Apart from differences in core histone content, thethree matrix preparations exhibit fairly similar proteincompositions. The main protein bands appear withmolecular weights of 24, 25, 31, 37 and 44(XlO3)(Fig. 5). An additional band at 57X 103 A/r is present inthe NaCI- and LIS-matriccs. It is noteworthy that allthree matrix preparations exhibit almost the sameprotein composition (except residual histones),although the ultrastructural differences seem morepronounced.

The protein patterns of Fig. 5 are highly reproduc-ible and distinct. Additional washing steps or slightchanges in the preparation protocol do not alter thispattern. This is unlike matrix preparations from iso-

lated nuclei, where slight changes in the procedure oradditional washes caused severe alterations in theprotein pattern of the matrix (results not shown).

Since the processing of encapsulated plasmodia forour matrix preparations is done under slightly hypo-tonic conditions we also processed plasmodial agarosebeads under isotonic conditions, essentially followingthe procedure of Jackson & Cook (1985) for prep-aration of chromatin. Fig. 6 shows the electron micro-graph of an NaCl-matrix prepared from agarose-encap-sulated plasmodia that have been processed underisotonic salt conditions.

Discussion

The major advance in our experimental approach of insitu matrix preparation is that the biochemical as wellas ultrastructural properties of the final product arerelatively similar in the three different methods used inthis study. Moreover, comparison between slightlyhypotonic and isotonic processing of encapsulatedplasmodia reveals that there are no significant morpho-

626 W. Waitz and P. Loidl

Fig. 6. Nuclear matrix of P. polycephalum (NaCl) afterisotonic processing of encapsulated plasmodia. X6000.

logical differences in the final matrix preparations.This indicates that differences in salt conditions priorto matrix preparation do not dramatically affect thematrix structure. This may be partly explained by thefact that changes in the ionic environment equilibratemuch more slowly in agarose beads compared withnaked plasmodia or isolated nuclei. We assume that theresults are due to the omission of nuclear isolation,which is known to cause serious changes in the mor-phology of matrix preparations (Smith et al. 1987).Some of the conflicting results in the literature(reviewed by Kaufmann et al. 1981) may also beexplained by uncontrollable structural distortions ofnuclei arising during nuclear isolation. It may alsohappen that during nuclear isolation nuclei of distinctcell cycle stages are selected, whereas nuclei of certainother cell cycle stages may be damaged or evendisrupted during the nuclear isolation procedure. Wethink that this method of in situ preparation is the firststep in the standardization of matrix isolation pro-cedures. It is obvious that this technique is applicableto any other cell type.

We started out to prepare nuclear matrix for electronmicroscopy following conventional procedures (glutar-aldehyde and osmium tetroxide as fixatives, Durcupanor Epon as embedding plastics). The procedure wehave used for the present study (described in Materials

and methods) leads to identical results. The acrylicplastic LR-white has the advantage that it can be usedfor immunoelectron-microscopic techniques, whichwill be important for the characterization of nuclearmatrix components in further investigations.

Moreover, the method of in situ preparation togetherwith the AS method for matrix isolation and preser-vation of certain cytoplasmic elements, enables one toinvestigate interconnections between nuclear matrixand cytoskeletal elements in thin sections with trans-mission electron microscopy. This inter-relation hasbeen studied by whole-mount electron microscopy inMadin-Darby canine kidney cells (Fey et al. 1984).Such an inter-relation is destroyed in the NaCl- andLIS-matrix preparations, as long as DNA loops arcextended prior to DNase I digestion.

There have been several reports on the character-istics of the nuclear matrix in P. polycephalum. Mit-chelson et al. (1979) prepared Physarum nuclearmatrix from isolated nuclei by extraction with 2-5 M-NaCl. It is difficult to compare the protein compo-sition, but at least two prominent proteins seem to beidentical with those in our preparation. These bandshave molecular weights of 23 and 36-5(XlO3). Theirpreparation contains at least 20 additional proteins anda relatively large amount of residual core histones. In asubsequent paper, Van der Velden & Wanka (1987)described two additional abundant bands with 28 and52(XlO3)Mr. The 28xlO3Mr protein could be ident-ical with our protein of 31XlO3/V/r. Kowalska-Loth &Staron (1986) found the most abundant Physarummatrix proteins with molecular weights of 28 and36(X 103)A/r. None of these proteins has been furthercharacterized or identified as a distinct matrix proteinknown from mammalian cells. Our matrix preparationscontain less protein bands (15-18 as visualized byCoomassie Blue staining) and, except for the AS-matrix, less core histones in comparison with theresults of Mitchelson et al. (1979).

Nuclear matrix preparations from P. polycephalumhave been used to demonstrate the association of DNAreplication with the nuclear matrix (Hunt & Vogel-stein, 1981; Aelenez al. 1983). However, the biologicalsignificance of this association is still seriously ques-tioned (Djondjurov et al. 1986). To settle such contro-versies it is necessary to avoid as many manipulationsteps as possible in the preparation of nuclei andmatrices. Moreover, it is desirable to isolate the nuclearmatrix under the most favourable conditions of ionicstrength. In this sense our new preparation techniqueprovides a useful contribution to the still unfoldingproblems in nuclear matrix research.

We are indebted to Mrs M. Edlinger and Mr H. j . Steinerfor technical assistance and to Mrs U. Matthes and Mrs G.Angele for preparation of photographs and the manuscript.

In situ preparation of nuclear matrix 627

The encouragement of Professor Dr G. Mikuz and dis-cussions with Dr P. Grobner are gratefully acknowledged.

References

AELEN, J. M. A., OPSTELTEN, R. J. G. & WANKA, F.

(1983). Organization of DNA replication in Physarumpolycephalum. Attachment of origins of replicons andreplication forks to the nuclear matrix. Nucl. Acids Res.11, 1181-1195.

BEREZNEY, R. & COFFEY, D. S. (1974). Identification of a

nuclear protein matrix. Biochem. biophys. Res. Commun.60, 1410-1417.

BEREZNEY, R. & COFFEY, D. S. (1977). Nuclear matrix.Isolation and characterization of a framework structurefrom rat liver nuclei. J . Cell Biol. 73, 616-637.

BOUTEILLE, M., BOUVIER, D. & SEVE, A. P. (1983).

Heterogeneity and territorial organization of the nuclearmatrix and related structures. Int. Rev. Cytol. 83,135-182.

CAPCO, D. G., WAN, K. M. & PENMAN, S. (1982). The

nuclear matrix: three-dimensional architecture andprotein composition. Cell 29, 847-858.

DANIEL, J. W. & BALDWIN, H. H. (1964). Methods of

culture for plasmodial myxomycetes. In Methods in CellPhysiology, vol. 1 (ed. D. M. Prescott), pp. 9-41. NewYork: Academic Press.

DJONDJUROV, L., IVANOVA, E., MARKOV, D., BARDAROV,

S. & SACHSEN MAIER, W. (1986). Is the nuclear matrixthe site of DNA replication in eukaryotic cells? Expl CellRes. 164, 79-96.

FEY, E. G., WAN, K. M. & PENMAN, S. (1984). Epithelialcytoskeletal framework and nuclear matrix-intermediatefilament scaffold: three-dimensional organization andprotein composition. J. Cell Biol. 98, 1973-1984.

HANCOCK, R. (1982). Topological organization ofinterphase DNA: the nuclear matrix and other skeletalstructures. Biol. Cell 46, 105-122.

HUNT, B. F. & VOGELSTEIN, B. (1981). Association ofnewly replicated DNA with the nuclear matrix ofPhysarum polycephalum. Nucl. Acids Res. 9, 349-363.

JACKSON, D. A. & COOK, P. R. (1985). A general methodfor preparing chromatin containing intact DNA. EMBOJ. 4, 913-918.

KAUFMANN, S. H., COFFEY, D. S. & SHAPER, J. H.

(1981). Considerations in the isolation of rat liver nuclear

matrix, nuclear envelope, and pore complex lamina. ExplCell Res. 132, 105-123.

KOWALSKA-LOTH, B. & STARON, K. (1986). Nuclearmatrix proteins of Physarum polycephalum. Molec. Biol.Rep. 11, 219-223.

LAEMMLI, U. K. (1970). Cleavage of structural proteinsduring the assembly of the head of bacteriophage T4.Nature, Land. Ill, 680-685.

LOIDL, P. (1988). Towards an understanding of thebiological function of histone acetylation. FEBS Lett.211, 91-95.

LOIDL, P. & GROBNER, P. (1987). Postsynthetic acetylationof histones during the cell cycle: a general function forthe displacement of histones during chromatinrearrangements. Nucl. Acids Res. 15, 8351-8366.

MATSUMOTO, S., GUGG, S. & LAFONTAINE, J. G. (1987).

Ultrastructural investigation of the effects ofsupraoptimal temperature on late interphase nuclei ofPhysarum polycephalum plasmodia. Biol. Cell 60, 87-96.

MlRKOVITCH, J., MlRAULT, M. E. & LAEMMLI, U. K.(1984). Organization of the higher-order chromatin loop:specific DNA attachment sites on nuclear scaffold. Cell39, 223-232.

MITCHELSON, K. R., BEKERS, A. G. M. & WANKA, F.

(1979). Isolation of a residual protein structure fromnuclei of the myxomycete Phvsarum polycephalum. J.Cell Sci. 39, 247-256.

NILSSON, K., SCHEIRER, W., MERTEN, O. W., OSTBERG,

L., LIEHL, E., KATINGER, H. W. D. & MOSBACH, K.

(1983). Entrapment of animal cells for production ofmonoclonal antibodies and other biomolecules. Nature,Land. 302, 629-630.

POGO, A. O. & PROCYK, R. (1985). Nonhistone proteinsand nuclear matrix structures. In Progress in NonhistoneProtein Research, vol. II (ed. I. Bekhor et al.), pp.132-149. Boca Raton, Florida: CRC Press.

SMITH, H. C , OCHS, R. L., LIN, D. & CHINAULT, A. C.

(1987). Ultrastructural and biochemical comparisons ofnuclear matrices prepared by high salt or LIS extraction.Molec. cell. Biochem. 77, 49-61.

VAN DER VELDEN, H. M. V. & WANKA, F. (1987). The

nuclear matrix - its role in the spatial organization andreplication of eucaryotic DNA. Molec. Biol. Rep. 12,69-77.

{Received 4 January J988-Accepted 11 April 1988)

628 W. Waitz and P. Loidl