WIXLIB

Crystalline layer in Drosophila eggshell: S. J. Hamodrakas et al.Figure 4 Isolated ICL, negatively stained with 27'ouranyl acetate and washed in distilled water before air-drying. Fine striations with a50 A,pseudoperiodicity are emerging in two directions (axes a and b). The inset is an optically filtered image of this micrograph showingthe arrangement of the electron transparent domains (circle). Magnification 125 000, inset x 500 000Results and discussionsurrounds a set of four electron transparent domainsA thin sectioned mature (late stage 14) follicle, after aconventional fixation- staining procedure (glutaraldehydefollowed by osmium tetroxide), reveals clearly mosteggshell layers: vitelline membrane (VM), innermostchorionic layer (ICL), inner endochorion (IE), pillars (P),outer endochorion (OE), roof network (RN), exochorion(EX) and vaguely the wax layer (wl) (Figure la). Viewedunder higher magnification, the ICL is shown to consist of4 5 bilaminar sublayers, with a total thickness of about400- 500 A (Figure lh). After impregnation withlanthanum nitrate during fixation, with no other staining,large deposits of stain are seen to exhibit a periodicity of100 A in two directions: one perpendicular and theother parallel to the sublayers (Figure lc). Therefore, itwould seem that, there are regularly spaced vacant areaswithin the ICE, the colloidal lanthanum micelles can enterand precipitate during dehydration.Whole mount face views of the layer, after isolation andnegative staining 4, reveal under low power that the ICEconsists of numerous crystallites 1 2 /lm in size (Figure2a). In each crystallite electron dense deposits of stain areseen to form a two dimensional lattice (Figure 2b).Adjacent crystallites differ from one another in thedirection of their lattice vectors. Apparently, this is aninherent property of the ICL and not a consequence of theisolation procedure or specimen preparation: mostprobably it is related to the process of the ICL edobservations) that the ICL crystallites are formed by aself-assembly procedure, from a number of separatenucleation centres during choriogenesis; these nucleationcentres may correspond to the endochorionic pillarswhich have a similar distribution.Under higher magnification, it is clear that thearrangement of the electron dense deposits of stain inthese whole mount face views of the layer (Figure 3a) isalmost identical to the cross-sectioned views (comparewith Fi,qure lc). Each tetrad of major electron dense spots(Figure 3a, circle) which presumably represent, in28Int. J. Biol. Macromol., 1982, Vol 4, Februaryprojection, spaces occupied by actual ICL molecules.Furthermore, minor electron dense spots or striations arefound between the major, 100 A spaced, electron densespots.Different ICLs or different areas of the same ICLexamined by negative staining in whole mount, showvariations in the unit cell parameters of the twodimensional lattice (Figures 3a and 3b). The unit cell axiallengths vary between 75 and 100 A and the interaxialangle is in the range 7 5 - 9 0 : the values of the cellparameters appear to be uniformly distributed in theseranges. There are several possible reasons for thesevariations, the most obvious being distortions producedduring the air-drying procedure and differences oforientation of the ICL sheets with respect to the electronbeam.Further analysis with optical diffraction and filteringand also computer reconstruction (method A) of electronmicrographs (Figures 3c,3d correspondingly) confirmedthe structural features indicated by the initialmicrographs. In each unit cell (Figure 3dl, four electrontransparent domains of nearly circular shape. - 30 A indiameter, are seen to be related by an approximatefourfold rotation axis and represent in projection thestructural elements of the ICL: these are probably proteinmolecules, since it is almost certain that protein is themain constituent of this chorionic layer s. Dimensions ofthe order of 30 A are not unexpected for proteinmolecules' .Negatively stained ICLs, washed before drying, do notshow periodicities arising from dense accumulation ofstain every 100 A, but, instead, less dense striations,spaced every 50 A approximately (Figure 4). Therefore, itwould seem that washing the specimen prior to dryingresults in staining the ICL to a lesser extent as at whole,hence the reduced overall contrast seen in Figure 4.Nevertheless, we believe that this procedure removes at

Crystalline layer in D r o s o p h i l a eggshell: S. J. Hamodrakas et al.Figure 5 (a) and (b) Representative micrographs from a tilt series of a negatively stained, isolated, ICL at 6 and - 18' respectively.Observe the periodic distribution of stain at the two crystallographic axes a and b, in two adjacent crystallites (one and two asterisks) inFigure 5a. Striations are seen to exist in a direction almost bisecting the angle of the two axes (broken lines). In Figure 5b, overlappingelectron dense spots can be seen along the a-axis of one crystallite tone asterisk) and no periodicities along the b-axis, whereas in theother crystallite (two asterisks) there is no trace of electron dense spots along the axes a and b, but, just white striations in a directionbisecting the interaxial angle [solid line, compare with broken line in Figure 5a two asterisks). The tilt axis, shown at the lower left, isalmost perpendicular to the a-axis of the crystallite marked with one asterisk. Magnification x 125 000, insets x 450000large part of the stain from the wide' vacant spaces of theICL. having a periodicity of 100 A, but to a m i n o r p a r tfrom the narrow' ones which intervene, thus resulting in a50 A pseudoperiodicity. O p t i c a l r e c o n s t r u c t i o n of thiss a m p l e (Fiqure 4, inset) confirms the existence of theelectron t r a n s p a r e n t d o m a i n s forming tetramers, whichare repeated every 100 A in two c r y s t a l l o g r a p h i cdirections.It r e m a i n e d to be clarified whether the observed twod i m e n s i o n a l periodicity on the whole m o u n t face views ofthe I C L arises from a d s o r p t i o n of stain on the surface ofthe ICL, or whether the stain (uranyl acetate) isd i s t r i b u t e d internally within the layer, as is l a n t h a n u mnitrate (Fiqure lc). To answer this question, an isolated,negatively stained I C L was observed at various angles oftilt with respect to the electron beam. M i c r o g r a p h s of thesame a r e a of the I C L were o b t a i n e d at intervals of 6 , from- 18 to 1 8 . In F i q u r e 5 a ( 6 tilt) the b o u n d a r i e s of 4crystallites, of those forming the polycrystalline structureof the ICE, can be seen clearly. The direction of the axis oftilt was chosen a l m o s t p e r p e n d i c u l a r to the a axis of thecrystallite m a r k e d with one asterisk. This m i c r o g r a p hshows the two d i m e n s i o n a l periodicity of the electrondense spots most clearly, and thus is considered torepresent a nearly p e r p e n d i c u l a r o r i e n t a t i o n of the layerwith respect to the electron beam. At a tilt angle of - 18',in the same crystallite (Figure 5b, one asterisk and inset),multiple electron dense spots a p p e a ralong acrystallographic direction (a-axis) perpendicular to theaxis of tilt, while striations along the b-axis are no longervisible. In an adjacent crystallite with different o r i e n t a t i o n(Fiyure 5h, two asterisks), the periodicity of the electrondense spots d i s a p p e a r s , while light striations are emerginga l o n g a direction almost bisecting the angle of thec r y s t a l l o g r a p h i c axes.The p a t t e r n s o b t a i n e d by optical diffraction analysis ofInt. J. Biol. M a c r o m o l . , 1982, Vol 4, F e b r u a r y29

Crystalline layer in D r o s o p h i l a eggshell. S. J. Hamodrakas et al. [1Figure 6 (a), (b), (c), (d), (el and (f) Optical diffraction patterns from the same area {solid box in the crystallite marked with one asteriskin Fi lm'e 5a) of the micrographs obtained at 1 8 , 12 . 6 , 0 , - 6 , 12 angles of tilt correspondingly. A gradual decrease isobserved in the intensity of the 1st and 2nd order spots, which correspond to periodicities perpendicular to the tilt axisFigure 7 Computer reconstructed images of mmrographs obtained at various angles of tilt, fiom an isolated, negatively stained, ICL(method B). (a) q 18' tilt: major electron dense spots (arrow) are repeated along two axes, a and b. Minor spots (crossed arrow) arelocated off-centre in respect to the major spots (compare with 6 tilt). (b) T 6' tilt: major spots {arrow) form a basic repeating unit,including a minor sPOt at the centre {crossed arrow). Observe the transparent domains (circle), representing in projection the'biomolecules, surrounding each major (and also minor) spot. Minor spots are very clearly resolved (solid box). (c) 0 tilt: major andminor electron dense spots are elongated along the a-axis which is perpendicular to the tilt axis. (d) - 6' tilt: similar arrangement as in 0tilt. (e) 18" tilt: all spots are seen double compared to the 6' tilt (compare solid boxes), an indication that the image is a projection ofthe internal structure of the layer, as outlined by negative staining through the sublayers. Magnification x 65000030Int. J. Biol. M a c r o m o l . , 1982, V o l 4, F e b r u a r y

Crystalline layer in Drosophila egyshell: S. J. H a m o d r a k a s et al.the same area (Figure 5a, parallelogram) under variousangles of tilt (Fiqure 6) reveal a gradual decrease in theintensity of the first and second order spots (Fiqure 6,arrows), in a direction perpendicular to the axis of tilt.Computer reconstructions of the images obtained at 18 6 . 0 , - 6 and - 18 were performed employinga fully automated system (method B). The reconstructions(Fiqure 7) showed that, as the specimen was tilted aboutan axis perpendicular to the a-axis, both major a n d m i n o relectron dense spots were first elongated and thenresolved into double spots.These observations indicate that the image we obtain(e.g. Fi,qure 3a) is due not simply to adsorption of negativestain on the surface of the ICE, but arises from penetrationof the stain into its sublayers.Combining the information provided by thereconstructions of the whole mount face views of the ICL(Fi 3ures 3c, 3d and 7) with the appearance of its crosssections (Fiqure lc), we suggest that an octamericarrangement of probably spherical protein molecules is aplausible basic repeating unit in the crystalline structureof the ICL. Nevertheless, this is not certain and awaitsfurther experimental evidence.It would appear that the spaces occupied by thenegative stain (Figures lc, 3c and 3d) are wide enough invivo to allow exchange of the respiratory gases to occurfreely through the ICL during embryogenesis.In conclusion, the results reported in this studydemonstrate the crystallinity of the ICL. Informationobtained from various projections of the ICL suggeststhat its constituent (protein ?) molecules are roughlyspherical in shape, with dimensions of the order of 30 :their packing arrangement leaves intervening spaceswhich may be of physiological importance.Experiments are in progress to investigate the threedimensional structure of the ICL further, by 3Dreconstruction and by techniques such as low-doseelectron diffractionS1: other experiments are directedtowards analysis of the morphogenesis of this interestingstructure.AcknowledgementsWe are very grateful to Prof. D. J. DeRosier for his help,for use of his optical diffractometer and for makingavailable to us his automated system and to Profs. F. C.Kafatos and A. C. T. North for useful comments. Wewould also like to thank the SRC Densitometer Service atthe SRC Laboratory, Daresbury, England for the use oftheir densitometer and the Leeds University ComputingLaboratory for the provision of computing facilities.This investigation was supported by grants awarded toS.J.H. by Athens University and to L.J.M. by N A T O(No. 1397), by Athens University and by the A. Onassisfoundation. Travelling expenses to the USA, for S. J. H.were provided by the Greek Ministry of Coordination.ReferencesI234567891011Petri, W. H., Wyman,A. R. and Kafatos, F. C. Dev. Biol. 1976,49,185Margaritis, L. H. Ph.D. Thesis, University of Athens (1974)Margaritis, L.H, Petri, W. HandKafatos. F.C.J. CellSci. 1980,43. 1Margaritis,L. H., Petri, W. H. and Wyman,A. R. Cell Biol. Int.Rep. 1979, 3, No. 1, 6lFourneaux,P. J. S. and Mackay, A. L. J. Ultrastruct. Res. 1972,38. 3u,3Mindrinos,M. N., Petri, W. H., Galanopoulos, V. K., Lombard,M. F. and Margaritis, L. H. Roux Archives Dev. Biol. 1980, 189,187Revel,J. P. and Karnovsky, M. J. Cell Biol. 1967, 33, C7 C12Lee, D. L., Nixon, P. E. and North, A. C. T. Proe. R. Soc. London(B) 1980, 208, 409Derosier,D. J. and Moore, P. B. J. Mol. Biol. 1970, 52, 355Schulz,G. E. and Schirmer. R. H. in 'Principles of ProteinStructure'. Springer-Verlag, New York-Heidelberg-Berlin, 1978Unwin,P. N. T. and Henderson. R. ,I. Mol. Biol. 1975, 94, 425Int. J. Biol. Macromol., 1982, Vol 4, February31

EM200 and EM301 microscopes. The latter was equipped with a goniometer stage. lmaqe processing (a) Optical diffraction and filtering were performed using a He-Ne laser diffractometer, as described . on the Leeds University ICL 1906A computer. An area of measurements, typically 120 x 120 was put into an array 200 x 200, the remaining array .