insect intercellular junctions: rapid freezing by...

14
J. Cell Sri. 62, 223-236 (1983) 223 Printed in Great Britain © The Company of Biologists Limited I9S3 INSECT INTERCELLULAR JUNCTIONS: RAPID FREEZING BY JET PROPANE LESLEY S. SWALES AND NANCY J. LANE A.R.C. Unit of Invertebrate Chemistry and Physiology, Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, U.K. SUMMARY The ventral nerve cord of the cockroach, Periplaneta americana, and that of the locust, Schis- tocerca gregaria, have been studied after rapid freezing by cryo-jet, using liquid propane. Such tissues, unfixed and uncryoprotected, have been compared with unfixed cryoprotected tissues, as well as with material fixed with glutaraldehyde and cryoprotected with glycerol or polyvinyl- pyrrolydone. The perineurial tight junctions in the cryo-jet-frozen tissues exhibit characteristic intramembranous P-face ridges, but frequently these are composed of smooth-surfaced strands, comparable to those seen with fast-freezing in mammalian tissues, rather than of bead-like fibres. The intramembranous PF ridges, characteristic of axonal and glial processes in the insect central nervous system, also display a smooth surface in rapidly frozen preparations. Prior fixation and/or cryoprotection produces a bead-like appearance in the ridges. The interglial gap junctions, after fast- freezing, exhibit both clustered connexon arrays in the E-face and loosely aggregated ones; hence the coupled state cannot be unequivocally associated with the latter configuration. The septate junctions between glial cells are unchanged after rapid freezing, exhibiting the typical rows of P-face intramembranous particles with complementary E-face pits that are found in replicas from fixed and cryoprotected tissues. The surfaces of the axons and glial processes exhibit pleiomorphic depressions and associated particles as well as PF pits with complementary EF mounds, both with associated IMPs. These structures are not usually seen after fixation or cryoprotection and may represent some kind of receptor structure, or axo-glial specialization. INTRODUCTION The technique of rapidly freezing biological tissues, or cryofixation, has emerged recently in an attempt to avoid the artefacts that may occur during fixation and cryoprotection (Staehelin, 1973; Mclntyre, Gilula&Karnovsky, 1974; Hasty & Hay, 1978; Arcancia, Valente & Crateri, 1980), one or both of which precedes freezing in conventional freeze-fracture methodology. Methods used to cryofix tissues include spray-freezing (Bachmann & Schmitt, 1971; Planner, Fischer, Schmitt & Bachmann, 1972) and surface-freezing on cold metal blocks (Heuser, Reese & Landis, 1976; Dempsey & Bullivant, 1976; Heuser et al. 1979). The former is not suitable for solid masses of tissue. The latter procedure, using a slamming device, involves tissue being brought into contact, at a considerable speed, with a copper block pre-cooled with liquid helium (Heuser et al. 1979). Although this has the advantage of enabling the investigator to stimulate a biological preparation just milliseconds before the moment of impact, it does involve the problem of exerting sudden massive pressure on the advancing tissue front and can cause disruption of critical aspects of membrane arch- itecture (Pinto da Silva & Kachar, 1980). To avoid this particular problem, jet- freezing using liquid propane has evolved as an alternative technique for freezing

Upload: vandung

Post on 04-Jul-2018

219 views

Category:

Documents


0 download

TRANSCRIPT

J. Cell Sri. 62, 223-236 (1983) 223Printed in Great Britain © The Company of Biologists Limited I9S3

INSECT INTERCELLULAR JUNCTIONS: RAPID

FREEZING BY JET PROPANE

LESLEY S. SWALES AND NANCY J. LANEA.R.C. Unit of Invertebrate Chemistry and Physiology, Department of Zoology,University of Cambridge, Downing Street, Cambridge CB2 3EJ, U.K.

SUMMARY

The ventral nerve cord of the cockroach, Periplaneta americana, and that of the locust, Schis-tocerca gregaria, have been studied after rapid freezing by cryo-jet, using liquid propane. Suchtissues, unfixed and uncryoprotected, have been compared with unfixed cryoprotected tissues, aswell as with material fixed with glutaraldehyde and cryoprotected with glycerol or polyvinyl-pyrrolydone. The perineurial tight junctions in the cryo-jet-frozen tissues exhibit characteristicintramembranous P-face ridges, but frequently these are composed of smooth-surfaced strands,comparable to those seen with fast-freezing in mammalian tissues, rather than of bead-like fibres.The intramembranous PF ridges, characteristic of axonal and glial processes in the insect centralnervous system, also display a smooth surface in rapidly frozen preparations. Prior fixation and/orcryoprotection produces a bead-like appearance in the ridges. The interglial gap junctions, after fast-freezing, exhibit both clustered connexon arrays in the E-face and loosely aggregated ones; hencethe coupled state cannot be unequivocally associated with the latter configuration. The septatejunctions between glial cells are unchanged after rapid freezing, exhibiting the typical rows of P-faceintramembranous particles with complementary E-face pits that are found in replicas from fixed andcryoprotected tissues. The surfaces of the axons and glial processes exhibit pleiomorphicdepressions and associated particles as well as PF pits with complementary EF mounds, both withassociated IMPs. These structures are not usually seen after fixation or cryoprotection and mayrepresent some kind of receptor structure, or axo-glial specialization.

INTRODUCTION

The technique of rapidly freezing biological tissues, or cryofixation, has emergedrecently in an attempt to avoid the artefacts that may occur during fixation andcryoprotection (Staehelin, 1973; Mclntyre, Gilula&Karnovsky, 1974; Hasty & Hay,1978; Arcancia, Valente & Crateri, 1980), one or both of which precedes freezing inconventional freeze-fracture methodology. Methods used to cryofix tissues includespray-freezing (Bachmann & Schmitt, 1971; Planner, Fischer, Schmitt & Bachmann,1972) and surface-freezing on cold metal blocks (Heuser, Reese & Landis, 1976;Dempsey & Bullivant, 1976; Heuser et al. 1979). The former is not suitable for solidmasses of tissue. The latter procedure, using a slamming device, involves tissue beingbrought into contact, at a considerable speed, with a copper block pre-cooled withliquid helium (Heuser et al. 1979). Although this has the advantage of enabling theinvestigator to stimulate a biological preparation just milliseconds before the momentof impact, it does involve the problem of exerting sudden massive pressure on theadvancing tissue front and can cause disruption of critical aspects of membrane arch-itecture (Pinto da Silva & Kachar, 1980). To avoid this particular problem, jet-freezing using liquid propane has evolved as an alternative technique for freezing

224 L. S. Swales and N. J. Lane

tissues rapidly. This method avoids the problems of tissue distortion due to the forceof impact and permits very rapid freezing; hence no prefixation or cryoprotection isrequired (Moor, Kistler & Miiller, 1976; Burstein & Maurice, 1978; Espevik &Elgsaeter, 1981; Muller, Meister & Moor, 1980; Giddings & Staehelin, 1980;Pscheid, Schudt & Plattner, 1981; Knoll, Gebel & Plattner, 1982).

The technique of rapid freezing has been of particular interest as applied to thevesicular release of neurotransmitters (Heuser et al. 1979) and membrane retrieval atneuromuscular junctions (Heuser & Reese, 1981), since it enables transient events tobe recorded at specified times after stimulation, but it has also been very useful in thestudy of the structure of intercellular junctions. Already it has been establishedthrough impact fast freezing of mammalian tissues that gap-junctional plaques aremore pleiomorphic (Raviola, Goodenough & Raviola, 1980) than had been predicted,while recent studies on crustacean tissues, frozen directly in a nitrogen slush withoutany preparation, reveal comparable gap-junctional pleiomorphism (van Deurs, Dant-zer & Bresciani, 1982). Hence, cell-to-cell coupling and the degree of connexonpacking in the gap junctions may not be correlated (Hanna et al. 1978, 1981; Ander-son, Atkinson, Sheridan & Johnson, 1981) in quite the way suggested by previousstudies (Peracchia & Dulhunty, 1976; Peracchia, 1977, 1978; Chuang-Tseng,Chuang, Sandri & Akert, 1982) - that is, that the closeness of connexon packing iscorrelated with the degree of intercellular uncoupling.

In addition to the studies on gap junctions, rapid freezing of vertebrate tissue,containing tight junctions, by impact onto a copper block has also led to a newinterpretation of tight-junctional structure (Kachar & Reese, 1981, 1982; Pinto daSilva & Kachar, 1982), according to which it is contended that the junctional ridgesmay not be composed of protein, but may instead be inverted cylindrical micelles oflipid; these may have protein associated peripherally for structural purposes.Although jet-freezing has been used for rapid freezing of cultured cell monolayers inwhich tight junctions may be seen (Espevik & Elgsaeter, 1981), this propane jettechnique has otherwise not been used in the study of intercellular junctions. More-over, the junctions of insects, which differ in many features from those of vertebrates(Lane, 1978, 1981a, 1982; Lane & Skaer, 1980), have not been investigated by rapid-freezing methods. This paper reports on the results obtained with cockroach andlocust central nervous system (CNS), when tight, gap and septate junctions areexamined after rapid jet-freezing without tissue fixation or cryoprotection. These dataare compared with those on similar junctions in tissues cryoprotected with glycerol orpolyvinylpyrrolydone (PVP), with or without prior fixation; intramembranousspecializations of the axonal membranes are also considered. A preliminary report ofthese results has appeared elsewhere in abstract form (Lane & Swales, 1982).

MATERIALS AND METHODS

The tissues used were interganglionic connectives from the ventral nerve cord of the centralnervous system of adult cockroach Periplaneta americana or locust, Schistocerca gregaria.

The control preparations were fixed in 2-5% glutaraldehyde in 0 05 %M-phosphate buffer(pH 7-2), with added 6 % sucrose, washed in phosphate buffer plus sucrose and then cryoprotected

Insect intercellular junctions 225

in either 20% glycerol (15-45 min) or 50% PVP (15min) in phosphate buffer. These tissues werethen frozen in liquid Freon 22 cooled in liquid nitrogen, supported by yeast paste in holders designedby Balzers for double, complementary replicas. Experimental tissues were dissected out and placedbetween two thin copper holders, which were rapidly frozen by two jets of melting propane, one oneither side, in a Balzers double propane cryo-jet device. The two plates of the holders had to be heldtogether by a dab of Tissue Tek (Lab-Tek Products) placed on two diagonally opposite corners ofthe plates. The tissue, having been dissected in Ringer, had the surplus fluid drained off by blottingwith filter paper. It was then mounted in one of the holders with the tissue touching one side of theholder and coming up the side of the plate's central depression. This arrangement was particularlyconvenient for the insect ventral nerve cord, which is highly attenuated and easily pulled out to oneside, for the fracture plane tends to go through the better-frozen large peripheral membrane faces,rather than the central region, which is more likely to have ice-crystal damage. The second plate wasplaced on top after Tissue-Tek application, without which the holders fell apart in storage. Fromeach of these, two complementary replicas were obtained. The material was fractured on a Balzers360 BM freeze-etching device at -100°C and at a pressure of about 2 x 10"6 Torr (1 Torr =133-3 Pa); the act of cleaving took place after shadowing had begun, so that etching was negligible.Shadowing took place with tungsten/tantalum, followed by backing with carbon. The replicas werecoated with collodion (Lane & Swales, 1980), cleaned with 1-10% sodium hypochlorite, washedwith distilled water and mounted on grids. The collodion was removed and the grids were examinedin a Philips EM300. Micrographs are mounted with the direction of shadow coming from the bottomor side. The conventional nomenclature of freeze-fractured faces is used here (Branton et al. 1975).

OBSERVATIONS

Rapid freezing

When cockroach nervous tissues are frozen rapidly by propane jet, we have thus farobtained only a relatively low success rate (about 12 %) in retrieving material withoutsignificant ice-crystal damage. The best preservation is in the areas of tissues in directcontact with the supporting plates during freezing. The areas that are optimallypreserved are, predictably, the most superficial tissues, which in interganglionicconnectives of the ventral nerve cord include the outermost glia layer, theperineurium, and the glial-ensheathed axons that lie immediately underneath. Vastexpanses of the perineurial bracelet cell membranes are most frequently exposed,exhibiting various membranous specializations. These include the bracelet cell inter-cellular junctions, which are of three types - tight, gap and septate - while theunderlying axons, neurites and glial processes exhibit both intramembranous ridgesand intramembrane particle (IMP) arrays, in a variety of configurations.

Even when the nervous tissue appears well-frozen, with little ice-crystal damage,the tight-junctional fibrils may appear as bead-like ridges on the membrane P-face(Fig. 1) although they may have some regions that appear to be in the form ofcontinuous strands (Fig. 2).

The axonal ridges, likewise, appear more as continuous strands in well-frozenpreparations (Figs 3, 4, 5; inset to Fig. 7), and less as bead-like arrays. Glial ridgesare less frequently observed (Fig. 7), while both glial and axonal grooves are rarelyseen (Figs 4, 6), probably due to the low angle of shadowing and lipid collapse(Bullivant, 1977).

The surfaces of both axons and glial processes in rapidly frozen tissues frequentlydisplay rosette-like small areas of intramembranous particles (IMPs) (Figs 3, 6, 7) as

226 L. S. Swales and N. J. Lane

Figs 1-5

Insect intercellular junctions 227

well as PF arrays of IMPs, irregular in outline, which lie in somewhat depressedregions of the membranes (Figs 6, 7). The former appear as raised protuberances onthe EF, circular in outline with a few IMPs (Figs 3, 6), and as depressed regions inthe PF, also with a few IMPs (Figs 5, 7). The latter are pleiomorphic in outline onboth axonal (Figs 3, 4, 6) and glial (Fig. 7) membranes.

The gap-junctional plaques in rapid jet-frozen tissues seem to freeze rather poorlyand are most frequently in the form of closely packed plaques (Fig. 8) with occasionalloosely clustered arrays (Figs 9, 10). The connexons fracture into the E-face as hasbeen found previously with fixed insect tissues (Flower, 1972; Lane, 1978; Lane &Skaer, 1980), while the complementary PF pits are less commonly seen (PF in Fig.10). The size of the connexons is difficult to ascertain, since they are very closelyassociated, often fusing with one another. In the looser arrays the connexon size rangeis from 14 nm to 16 nm, while in closely packed clusters the connexons are 13-15 nmin diameter.

Septate junctions appear little different from those in conventionally fixed andfrozen preparations (Fig. 11). The rows of junctional IMPs fracture onto the P facewith complementary rows of pits on the E face; the two are in register at fracture facetransitions from PF to EF (arrow in Fig. 11). They exhibit a range of patterns from

All the preparations in Figs 1-11 are replicas after rapid-freezing by liquid propane in thecryo-jet. These tissues were not subjected to fixation or cryoprotection.

Figs 1,2. Tight-junctional ridges (arrows) in the perineurial layer ensheathing the ventralnerve cord. Note that although in some cases, when the freezing is not optimal, they appearbead-like (Fig. 1), but they may appear as smoother fibrils (Fig. 2). Fig. 1, x77 590;Fig.2, X144700.

Figs 3-5. Low-power views of regions of axons (a) and glia (g) from the ventral nerve cordimmediately underlying the outer perineurial layer, showing the excellence of tissuepreservation with optimal conditions of cryo-jet freezing. E-face (EF) arrays of IMPs onraised protuberances (curved arrows in Fig. 3) have complementary PF depressions withIMPs (curved arrow in Fig. 5). PF ridges (arrows) occur on both axonal (Figs 3, 5) andglial processes; these have complementary grooves (double arrows in Figs 4, 5). Pleiomor-phic arrays of PF IMPs (thick arrows) are found on axonal surfaces (Figs 3,4). Note thatthe ridges are smooth in appearance, not moniliform. Fig. 3, X43 350; Fig. 4, X52450;Fig. 5, X57 600.

Figs 6, 7. Other areas of axonal and glial processes, to show the detailed features of axonal(a) and glial (g) ridges (arrows), grooves (double arrow), pleiomorphic depressions withIMPs (thick arrows) and spherical PF pits and EF protuberances (curved arrows). Notethat the ridges (inset to Fig. 7) show no bead-like substructure. Fig. 6, X55 460; Fig. 7,X40500; inset to Fig. 7, x 103 400.

Figs 8, 9, 10. Gap-junctional plaques to show the range of configurations of the EFconnexons. Many are closely packed (as in Fig. 8) while others show a looser arrangement(Fig. 9), which may spread over quite a surface area of the membrane (Fig. 10). Whereclosely associated, the particles are indistinguishable one from another (arrows). P-face(Pl'^j pits are less regularly seen, except when the fracture plane cleaves down from the E-face (EF) to patches of PF (as in Fig. 10). Fig. 8, X54600; Fig. 9, X46800; Fig. 10,X46600.

Fig. 11. Septate junction showing linear arrays of IMPs on the P-face (PF), which frac-tures up to the E-face (EF) in phase (at arrow) with the rows of EF pits. Note the largerIMPs scattered over the EF. X54900.

228 L. S. Swales and N. jf. Lane

44- • -•••3ET »•»:*• •MlZW^

Figs 6-11. For legend see p. 227.

Insect intercellular junctions 229

Figs 12-18. For legend see p. 230.

230 L. S. Swales and N. J. Lane

wide bands of IMP rows to branching, splayed rows with larger IMPs lying betweenthe branches. The IMPs making up the rows are about 8 nm in diameter and theindividual rows may be closely aligned with one another, side by side, as close as3-4 nm apart.

Unfixed, cryoprotected preparations

The tight junctions and axonal ridges in tissues prepared, without fixation but withcryoprotection in either 20—25% glycerol (Fig. 15) or 50% polyvinylpyrrolydone(Figs 12, 13, 14), tend to exhibit a bead-like structure rather than a smooth ap-pearance (Figs 14, 15). However, PVP-cryoprotected preparations can producesmooth fibrils (Figs 12, 13) when the freezing is generally good. The axonal and glialirregular particle arrays are in evidence (Figs 12, 13, 14), and appear very similar tothose seen in rapidly frozen replicas.

The connexons of the gap junctions fracture onto the E-face and are arrayed bothin loosely and more closely packed clusters. The septate junctions appear similar tothose in unfixed material with rows of PF IMPs and EF grooves.

Fixed, cryoprotected tissues

In fixed preparations the tight junctions cryoprotected with either PVP (Fig. 16)or glycerol (Fig. 17) appear as bead-like alignments, as do the axonal ridges (Fig. 18).

Figs 12-14. Tissues have not been fixed but have been cryoprotected with either PVP(Figs 12, 13, 14) or glycerol (Fig. 15). The axonal ridges shown here (at arrows) mayappear as smooth fibrils (Fig. 13) after PVP, but may also, presumably when freezing isless good, appear bead-like (Fig. 14). After glycerination they appear as discontinuousbeads (Fig. 15). The pleiomorphic IMP arrays (thick arrows) can still be seen in thesecryoprotected tissues. Fig. 12, X37200; Fig. 13, X68100; Fig. 14, X99100; Fig. 15X73 000.

Figs 16-18. Glutaraldehyde-fixed tissue, subsequently cryoprotected in either PVP (Fig.16) or glycerol (Figs 17, 18). The perineurial tight-junctionalPF ridges (arrows) are hereseen to be moniliform, so that fixatives have disrupted the putative cylindrical lipidmicelles. Grooves (Fig. 16) are also present (double arrows). Similarly, the axonal ridges(Fig. 18 and inset) show a moniliform structure (at arrow) but few axonal or glial particlearrays are to be found, in contrast with unfixed tissues; only a few scattered individualIMPs are present. Fig. 16, X55 200; Fig. 17, X113 400; Fig. 18, X44 800; inset,X119400.

Figs 19, 20. Gap junctions (jgj) in various degrees of aggregation as can also be found inconventionally fractured tissues; Fig. 20 reveals that they can also co-exist with septatejunctions (sj). Gap-junctional plaques can be seen in the EF; c, cytoplasm; Fig. 19,X35 800; inset, Fig. 19, X119000; Fig. 20, X35 70O.

These tissues were all fixed and cryoprotected in glycerol (except for Fig. 22, which isfixed and PVP-cryoprotected), and serve as controls for the rapid-frozen material as theywere all fractured in complementary replica holders and shadowed simultaneously.

Figs 21-23. Septate junctions with typical rows of P-face (PF) IMPs, which fracture atface transitions in phase (at arrows) with rows of E-face (EF) pits. PVP-cryoprotectiondoes not yield such distinct IMP rows (Fig. 22), although the angle of shadowing may alsohave an effect here. Tricellular junctions (thick arrows in Fig. 23) indicate regions wherethree cells intersect. Fig. 21, X43 200; Fig. 22, X41 300; Fig. 23, x27 100.

Insect intercellular junctions 231

Figs 19-23

232 L. 5. Swales and N. J. Lane

There is a reduction in the overall number of IMP rosettes and irregular arrays (Fig.18) in the axonal and glial membranes. There may be IMPs scattered over the mem-branes but they do not normally reside in depressions.

The gap junctions have the characteristic appearance of plaques of connexons inarthropods (Fig. 19), fracturing onto the EF with complementary pits on the PF (Fig.19). There are frequently also loose clusters of connexons to be seen (Fig. 19, inset;Fig. 20). The size range of these connexons is about 13-16nm in diameter and thejunctions are found co-existing in intimate spatial relationship with septate junctions(Fig. 20).

Septate junctions appear very like those seen in unfixed preparations with PF IMPsand complementary EF pits (Fig. 21). This appears to be the same whether thematerial is cryoprotected in PVP (Fig. 22) orglycerol (Figs 21, 23), although the rowsare less distinct in PVP-treated material (Fig. 22). The parallel rows are separated bya range of distances but the shortest appears to be about 3 nm, as intimate as those inrapidly frozen preparations. Tricellular junctions (thick arrows in Fig. 23) are alsotypical in appearance (Noirot-Timothe'e & Noirot, 1980).

DISCUSSION

Insect tissues have not hitherto been studied by rapid freezing without fixation andcryoprotection; nor, indeed have those of other invertebrates, although some work(Hanna et al. 1981) has been briefly reported in an abstract on the lower chordate,Ciona, and recently crustacean root systems have been frozen directly in nitrogenslush (van Deurs et al. 1982). The results obtained in the present study suggest thatrapid-freezing with the double cryo-jet is not as effective in achieving good tissuepreservation as that with the copper-block technique (Heuser et al. 1976, 1979). Aswith the latter method, after jet-freezing only the initial 10-15 /mi of tissue in theinsect CNS is frozen adequately, but in a high percentage of cases even this peripheralregion shows ice-crystal damage. Such poor freezing was not reported after mono-jetcryofixation (Knoll et al. 1982) or after double liquid-propane jet-freezing (Espevik& Elgsaeter, 1981) of monolayer cell cultures, subcellular fractions or cells in suspen-sion. In the latter study, however, although membrane preservation was reasonable,cytoplasmic damage, unrecognized by the authors, was not inconsiderable. Studieson the cooling rates achieved with the same double propane jet device used in ourlaboratory, compared with plunging into liquid propane (Robards & Severs, 1981),showed the former to be inferior to the latter, both with regard to mean cooling rate,reproducibility of results and ice-crystal size found in samples of the tissue studied(myocytes). In that study it was considered that warm propane gas might emergeinitially from the jet aperture; certainly in our case it seems that a steady flow ofpropane in a liquid state is not achieved. Given that some of the replicas made fromfast double jet-frozen tissue yielded reasonable freezing, the cryo-jet technique isclearly usable, but not always dependable, due to the variability in cooling ratesobtained. Moreover, studies comparing mono-jet with double-jet systems (Knoll etal. 1982) suggest that the latter device, a two-side propane jet, is inferior since it

Insect intercellular junctions 233

cannot reach both sides of the copper-tissue-copper sandwich at precisely the sametime.

The results obtained in this study after fast-freezing of both insect tight-junctionalstrands and axonal and glial ridges indicate that, with good freezing conditions, thesestructures appear smooth and so could be interpreted as being inverted cylindricallipid micelles sandwiched between linear fusions of the EF halves of the membranesof adjacent cells, as suggested for vertebrate tight junctions (see Kachar & Reese,1981, 1982; Pinto da Silva& Kachar, 1982). Although insect tight junctions occasion-ally appear as discontinuous ridges in fast-frozen tissues, as may those of mammaliantissues (Hirokawa, 1982), such preparations tend to be those that show ice-crystaldamage and other evidence of low cooling rates. These, as well as the fixed orcryoprotected tissues, may possess bead-like ridges only because the continuous cylin-ders of lipid have been disrupted, leaving IMPs as remnants of the intramembranecylinders. Studies on developing junctions in embryonic arthropod tissues will beinitially less easy to reconcile with this notion, however, since here the individual IMPs,hitherto assumed to be protein, apparently migrate translaterally, fusing to form thejunctional strands (Lane & Swales; \97Sa,b; Lane, 19816). Such events have beentaken to indicate that the IMPs have a major protein component, although they couldbe individual spherical inverted micelles of lipid. Certainly, all the glutaraldehyde-fixed insect tissues show bead-like alignments, not smooth strands, in agreement withthe interpretation that fixatives induce disruption of the putative cylindrical lipidmicelles. However, recent preliminary investigations on isolated chicken and mousemembrane fractions enriched in tight junctions reveal a prominent peptide com-ponent and are structurally resistant to certain detergents (Stevenson, Goodenough&Mooseker, 1982).

The rapidly frozen cockroach gap junctions exhibit variability in connexon pack-ing, from very loose to closely clustered, although the particles continue to cleave tothe E-face. It has been assumed that glutaraldehyde fixation causes the connexoncrystallization observed in fixed and cryoprotected tissues (Bennett, Spira & Pappas,1972; Sikerwar & Malhotra, 1981), which is thought to be characteristic of theuncoupled state (Peracchia, 1980). However, even fast-frozen junctions, althoughunfixed, may not reflect the coupled state if, for other physiological reasons, the cellsare already uncoupled. The results obtained here indicate that it is not possible todistinguish between the coupled and uncoupled state in these insect tissues; the datasuggest that both states may exist simultaneously in any given area of cockroach glialcell membrane. Certainly, in fixed tissues one also finds both closely and looselypacked connexons (see Fig. 20), as has also been reported in crustacean tissues (vanDeurs et al. 1982). Extrapolating from the results in other arthropod systems(Peracchia & Dulhunty, 1976), it might be considered that the more loosely packedjunctions must represent those that are coupled. However, other evidence from rapid-ly frozen tissues, from both lower chordates (Hanna et al. 1981) and mammaliansystems (Raviola et al. 1980), suggests that alterations in junctional permeability arenot necessarily consistent with connexon-particle packing density. It is also possiblethat junctional turnover takes place in adult tissues, so that the states observed may

234 L. S. Swales and N. J. Lane

be stages in junctional disassembly or reaggregation (see Lane & Swales, 1980). It hasbeen suggested (Peracchia, 1980) that there is a reduction in size of connexons whenthey are closely packed and, theoretically, uncoupled. It is impossible in the cock-roach system to make the appropriate measurements accurately, since the particleswhen closely packed are so closely associated with one another that they almost appearto fuse. Given this variability in appearance, no consistent reduction in size can beestablished in this study. However, in other systems, such as amphibian embryonicepithelium (Chuang-Tseng et al. 1982), the sizes of the gap-junctional particles didnot differ significantly between the coupled and uncoupled stages.

The pleated septate junctions are no different in the fast-frozen state from those infixed tissues, in that their component IMPs remain on the PF, aligned in undulatingrows with complementary rows of EF pits. Their packing is no tighter, with regardto the separation of the individual rows of IMPs. Fixation, therefore, does not in anyway radically modify their appearance as might be expected for smooth septate junc-tions, where the junctional IMPs of some tissues fracture onto the PF in fixed tissues,and onto the EF in unfixed but cryoprotected tissues (Flower & Filshie, 1975; Lane,1978; Skaer, Harrison & Lee, 1979; Lane & Skaer, 1980). The IMPs in PVP-cryoprotected tissues stand less proud than those in glycerol-treated material, perhapsbecause PVP binds water and therefore may dehydrate the membrane, producing adifferent chemical state than is produced by glycerol (Skaer, 1982).

The arrays of axonal and glial depressions and elevations with associated IMPs werenot seen so clearly in fixed and cryoprotected specimens, although some intimationas to their existence arises from earlier reports (Lane & Swales, 1978a,6). They mayonly be so apparent in these unfixed preparations because of changes induced by therapid freezing in the surrounding lipid bilayer, which may have undergone a phasechange such as the formation of inverted lipid micelles around the protein particles(Kato & Bito, 1980). Alternatively, the protein particles may have become moreapparent due to lipid collapse (Bullivant, 1977) or other changes in the degree of lipidcrystallization (Aggerbeck & Gulik-Krzywicki, 1982). These modifications may besome form of axo—glial interaction, or may represent some kind of receptor molecules.

We are grateful to Mr W. M. Lee for technical assistance and to Dr Helen le B. Skaer for helpfulcomments on the manuscript.

REFERENCESAGGERBECK, & GULIK-KRZYWICKI, T. (1982). Freeze-fracture electron microscopic and low tem-

perature X-ray scattering studies of the effect of cryofixation upon serum low density lipoproteinstructure. J . Microsc. 126, 243-252.

ANDERSON, S. K., ATKINSON, M. M., SHERIDAN, J. D. & JOHNSON, R. G. (1981). Freeze-fracture studies of gap junctions in cells infected with a temperature-sensitive mutant of aviansarcoma virus. J . Cell Biol. 91, 121A.

ARANCIA, G., VALENTE, F. R. & CRATERI, P. T. (1980). Effects of glutaraldehyde and glycerolon freeze-fractured Escherichia coli.J. Micrvsc. 118, 161-176.

BACHMANN, L. & SCHMITT, W. W. (1971). Improved cryofixation applicable to freeze etching.Proc. natn. Acad. Sci. U.SA. 68, 2149-2152.

Insect intercellular junctions 235

BENNETT, M. V. L., SPIRA, M. & PAPPAS, G. D. (1972). Effects of fixatives on properties ofelectrotonic junctions between embryonic cells. J. Cell Biol. 55, 17A.

BRANTON, D., BULLIVANT, S., GILULA, N. B., KARNOVSKY, M. J., MOOR, H., MUHLETHALER, K.,

NORTHCOTE, D. H., PACKER, L., SATIR, B., SATIR, P., SPETH, V., STAEHELIN, L. A., STEERE, R.

L. & WEINSTEIN, R. S. (1975). Freeze-etching nomenclature. Science, N.Y. 190, 54-56.BULLIVANT, S. (1977). Evaluation of membrane structure; facts and artefacts produced during

freeze-fracturing.J. Mictvsc. I l l , 101-116.BURSTEIN, N. L. & MAURICE, D. M. (1978). Cryofixation of tissue surfaces by a propane jet for

electron microscopy. Micron 9, 191-198.CHUANG-TSENG, M. P., CHUANG, H. H., SANDRI, C. & AKERT, K. (1982). Gap junctions and

impulse propagation in embryonic epithelium of Amphibia. A freeze-etching study. Cell Tiss.Res. 225, 249-258.

DEMPSEY, G. P. & BULLIVANT, S. (1976). A copperblock method for freezing non-cryoprotectedtissue to produce ice-crystal-free regions for electron microscopy. I. Evaluation using freeze-substitution.J. Microsc. 106, 251-260.

ESPEVIK, T. & ELGSAETER, A. (1981). In situ liquid propane jet-freezing and freeze-etching ofmonolayer cell cultures. J. Microsc. 123, 105-110.

FLOWER, N. E. (1972). A new junctional structure in the epithelia of insects of the order Dictyop-tera.J. Ce//S«. 10, 683-691.

FLOWER, N. E. & FILSHIE, B. K. (1975). Junctional structures in the midgut cells of lepidopterancaterpillars. J . Cell Sci. 17, 221-239.

GIDDINGS, T. H. & STAEHELIN, L. A. (1980). Ribosome binding sites visualised on freeze-fractured membranes of the rough endoplasmic reticulum. J. Cell Biol. 85, 147-152.

HANNA, R. B., REESE, T. S., ORNBERG, R. L., SPRAY, D. C. & BENNETT, M. V. L. (1981). Fresh

frozen gap junctions: Resolution of structural detail in the coupled and uncoupled states. J. CellBiol. 91, 125A.

HANNA, R. B., SPRAY, D. C , MODEL, P. G., HARRIS, A. L. & BENNETT, M. V. L. (1978).

Ultrastructure and physiology of gap junctions of an amphibian embryo, effects of CO2. Biol.Bull. mar. biol. Lab., Woods Hole 155, 442.

HASTY, D. L. & HAY, E. D. (1978). Freeze-fractured studies of the developing cell surface. II.Particle-free membrane blisters on glutaraldehyde-fixed corneal fibroblasts are artefacts. J. CellBiol. 78, 756-768.

HEUSER, J. E. & REESE, T. S. (1981). Structural changes after transmitter release at the frogneuromuscular junction. J. Cell Biol. 88, 564—580.

HEUSER, J. E., REESE, T. S. & LANDIS, D. M. D. (1976). Preservation of synaptic structure byrapid freezing. Cold Spring Harbour Symp. quant. Biol. 40, 17-24.

HEUSER, J. E., REESE, T. S., DENNIS, M. J., JAN, Y., JAN, L. & EVANS, L. (1979). Synaptic

vesicle exocytosis captured by quick freezing and correlated with quantal transmitter release, jf.Cell Biol. 81, 275-300.

HIROKAWA, N. (1982). The intramembrane structure of tight junctions: an experimental analysisof the double-fibril and two fibril-models using the quick-freeze method. J. Ultrastruct. Res. 80,288-301.

KACHAR, B. & REESE, T . S. (1981). Membrane structure at tight junctions in quick frozen tissue.J. Cell Biol. 91, 123A.

KACHAR, B. & REESE, T. S. (1982). Evidence for the lipidic nature of tight junction strands.Nature, Land. 296, 464-466.

KATO, K. & BITO, Y. (1980). Gel to liquid-crystalline phase transitions of lipids and membranesisolated from Escherichia coli cells. Microbiol. Imtnun. 24, 703-716.

KNOLL, G., GEBEL, G. & PLATTNER, H. (1982). A simple sandwich-cryogen-jet procedure withhigh cooling rates for cryofixation of biological material in the native state. Protoplasma 111,161-176.

LANE, N. J. (1978). Intercellular junctions and cell contacts in invertebrates. In Electron Micro-scopy 1978, vol. 3. State of the Art. Symposia (ed. J. M. Sturgess), Proc. 9th Int. Conf. ElectronMicrosc, Toronto, pp. 673-691. Canada: The Imperial Press.

LANE, N. J. (1981a). Tight junctions in arthropod tissues. Int. Rev. Cytol. 73, 243-318.LANE, N. J. (19816). Evidence for two separate categories of junctional particles during the concur-

rent formation of tight and gap junctions. J. Ultrastruct. Res. 77, 54—65.

236 L. 5. Swales and N. J. Lane

LANE, N. J. (1982). Studies on invertebrate cell interactions - advantages and limitations of themorphological approach. Proc. Symp. Fund. Org. Animal Tissue. Br. Soc. CellBiol., pp. 81-103.Cambridge University Press.

LANE, N. J. & SKAER, H. L E B . (1980). Intercellular junctions in insect tissues. Adv. Insect Physiol.15, 35-213.

LANE, N. J. & SWALES, L. S. (1978a). Changes in the blood-brain barrier of the central nervoussystem in the blowfly during development, with special reference to the formation and disaggrega-tion of gap and tight junctions. I. Larval development. DeviBiol. 62, 389-414.

LANE, N. J. & SWALES, L. S. (19786). Changes in the blood-brain barrier of the central nervoussystem in the blowfly during development, with special reference to the formation and disaggrega-tion of gap and tight junctions. II. Pupal development and adult flies. Devi Biol. 62, 415-431.

LANE, N. J. & SWALES, L. S. (1980). Dispersal of gap junctional particles, not internalization,during in vivo disappearance of gap junctions. Cell 19, 579-586.

LANE, N. J. & SWALES, L. S. (1982). Liquid propane rapid jet-freezing of intercellular junctionsin insect CNS. J. CellBiol. 95, 92A.

MCINTYRE, J. A., GILULA, N. B. & KARNOVSKY, M. J. (1974). Cryoprotectant-inducedredistribution of intramembranous particles in mouse lymphocytes. J. CellBiol. 60, 192-203.

MOOR, H., KISTLER, J. & MULLER, M. (1976). Freezing in a propane jet. Experientia 32, 805.MULLER, M., MEISTER, N. & MOOR, H. (1980). Freezing in a propane jet and its application in

freeze-fracturing. Mikroskopie 36, 129-140.NOIROT-TIMOTHIIE, C. & NOIROT, C. (1980). Septate and scalariform junctions in arthropods. Int.

Rev. Cytol. 63, 97-140.PERACCHLA, C. (1977). Gap junctions- structural changes after uncoupling procedures. J . Cell

Biol. 72, 628-664.PERACCHIA, C. (1978). Calcium effects on gap junction structure and cell coupling. Nature, Land.

271, 669-671.PERACCHIA, C. (1980). Structural correlates of gap junction permeation. Int. Rev. Cytol. 66,

81-146.PERACCHIA, C. & DULHUNTY, A. F. (1976). Low resistance junctions in crayfish. Structural

changes with functional uncoupling, j . Cell Biol. 70, 419-439.PINTO DA SILVA, P. & KACHAR, B. (1980). Quick freezing vs. chemical fixation: capture and

identification of membrane fusion intermediates. Cell Biol. Int. Rep. 4, 625-640.PINTO DA SILVA, P. & KACHAR, B. (1982). On tight-junction structure. Cell 28, 441-450.PLATTNER, H., FISCHER, W. M., SCHMITT, W. W. & BACHMANN, L. (1972). Freeze-etching of

cells without cryoprotectants. jf. CellBiol. 53, 116-126.PSCHEID, P., SCHUDT, C. & PLATTNER, H. (1981). Cryofixation of monolayer cell cultures for

freeze-fracturing without chemical pre-treatments._7. Microsc. 121, 149-167.RAVIOLA, E., GOODENOUGH, D. A. & RAVIOLA, G. (1980). Structure of rapidly frozen gap

junctions. J . CellBiol. 87, 273-279.ROBARDS, A. W. & SEVERS, N. J. (1981). A comparison between cooling rates achieved using a

propane jet device and plunging into liquid propane. Cryo-Letters 2, 135-144.SIKERWAR, S. & MALHOTRA, S. (1981). Structural correlates of glutaraldehyde-induced uncoup-

ling in mouse liver gap junctions. Eur.Jf. Cell Biol. 25, 319-323.SKAER, H. LE B. (1982). Chemical cryoprotection for structural studies..7. Microsc. 125, 137-147.SKAER, H. LE B., HARRISON, J. B. & LEE, W. M. (1979). Topographical variations in the structure

of the smooth septate junction. J . Cell Sci. 37, 373-389.STAEHELIN, L. A. (1973). Further observations on the fine structure of freeze-cleaved tight junc-

tions, jf. Cell Sci. 13, 763-786.STEVENSON, B. R., GOODENOUGH, D. A. & MOOSEKER, M. S. (1982). Zonulae occludentes in

plasma membrane fractions from chicken intestine and mouse liver. J. Cell Biol. 95, 92A.VAN DEURS, B., DANTZER, U. & BRESCIANI, J. (1982). Gap junction pleiomorphism in the root

system of the rhizocephalans (Arthropoda, Crustacea). Eur.J. Cell Biol. 27, 256—261.

{Received 6 October 1982-Accepted 17 January 1983)