Department of Cell Biology, University of Auckland, Auckland, New Zealand. (Accepted 8 February I973)

Transcription

1 J. gen. Virol. (1973), zo, 37-5o Printed in Great Britain 37 Fine Structure of Vesicles Induced in Chloroplasts of Chinese Cabbage Leaves by Infection with Turnip Yellow Mosaic Virus By T. HATTA, S. BULLIVANT AND R. E. F. MATTHEWS Department of Cell Biology, University of Auckland, Auckland, New Zealand (Accepted 8 February I973) SUMMARY Examination of chloroplasts from chinese cabbage leaves infected with turnip yellow mosaic virus (TYMV), using freeze-fracture techniques revealed, that the small peripheral vesicles induced by infection were present in clusters near the surface of the chloroplast, often being arranged in arrays that approximate to hexagonal. The small peripheral vesicles were flask shaped with their necks pointing to the chloroplast surface. The structure of the necks, which were only about I5 to 22 nm in diam., suggested but did not prove that they were open to the cytoplasmic space. The inner membrane of the peripheral vesicles is distinguished from other chloroplast membranes in showing no particles on either the A or the B fracture faces. In the regions where the peripheral vesicles were clustered, the number and kind of particles seen on the chloroplast membrane faces were changed compared with healthy chloroplasts or the non-vesicle bearing areas of chloroplasts from infected cells. In particular, the B face of the outer chloroplast membrane in the vesiclebearing regions had a reduced number of particles 1o nm in diam., but in" addition to these contained large numbers of particles about 5 nm in diam. In the chloroplasts of both healthy and infected tissue we frequently observed protuberances of the stroma corresponding to the mobile phase of chloroplasts postulated from light microscopic observations. INTRODUCTION Infection of chinese cabbage and other hosts by turnip yellow mosaic virus (TYMV) is followed at an early stage by the formation of numerous small vesicles, composed of two membranes, at the periphery of the chloroplast. These vesicles appear to be specific for the TYMV virus group rather than the host species. It has been suggested that they may be the site of virus RNA synthesis (Ushiyama & Matthews, 197o). Previous work on the effects of TYMV and the significance of the vesicles has been reviewed recently (Matthews, I97o, 1973). Electron microscopic observations on thin sections of diseased chloroplasts show clearly that the outer membrane of the vesicles is continuous with the inner chloroplast membrane; and is probably formed from it by invaginations (Lafl~che & Boy6, I969; Ushiyama & Matthews, I97o). Lafl~che & Boy6 0969) suggested that the inner membrane of the vesicles is formed by invagination from the outer chloroplast membrane. While occasional vesicles seen in thin section give the appearance of opening into the cytoplasmic space, the exact

2 38 T. HATTA~ S. BULLIVANT AND R. E. F. MATTHEWS relationship between the bounding chloroplast membranes and the vesicle membrane has not been resolved by the examination of thin sections. In the work described here we have examined the fine structure of chloroplasts from tissue infected with TYMV using freeze-fracture methods. Freeze-fracturing is particularly useful because it reveals large areas of interior faces of membranes, and gives some indication of the disposition of particulate proteins within the lipid bilayer (Branton, I97I ). It is probable that most vesicles do open to the cytoplasmic space, but it has not been possible to prove this even by the examination of complementary replicas. The main feature of the vesicles rex ' ~-~, e~eze-fracturing is that the inner membrane differs structurally from the other chloroplast m..., --,z.,, none of the particles normally seen within membranes. The vesicles are clustered over tnc... - ~.he chloroplast. In the region of these clusters the outer chloroplast membrane has an increased number of particles, indicating an increased metabolic activity. METHODS Virus and plants. TYMV was cultivated in chinese cabbage plants (Brassica pekinensis, Rupr., var Wong Bok) in pots in the glass house. Strains of virus were isolated from the stock culture in chinese cabbage as described by Chalcroft & Matthews 0967). For most experiments a severe 'white' isolate was used. Isolation of chloroplasts. Systemically infected and healthy chinese cabbage plants were kept in a dark room for about 24 h to reduce the content of starch grains before isolation of the chloroplasts. Leaf tissue (I "5 g without midribs from young leaves about ~o cm long) was put in 3o ml of cold modified Millonig phosphate buffer (Millonig, 196 I) (o. r 6 M-sodium dihydrogen phosphate, o.o8 M-glucose, o.0o5 M-MgCI~, brought to ph 7"4 with o.iz N-NaOH). The tissue was chopped with an electrically powered razor blade at o C until the tissue was cut into very small fragments (4 to 5 min cutting). The suspension was filtered through two layers of Miracloth. The filtered solution was centrifuged at ~ooo rev/min for Io min in an International refrigerated centrifuge. The green pellets so obtained were examined in the electron microscope using both the freeze-fracturing and thin-sectioning procedures outlined below. Thin sections for electron microscopy. Leaf tissue was fixed with 6 ~ glutaraldehyde in Millonig buffer for 6o min at room temperature. For isolated chloroplasts this buffer was modified as noted under the isolation procedure. After washing overnight the samples were postfixed with I ~ osmic acid in buffer for 2 or 3 h at room temperature. Dehydration, embedding, staining and sectioning procedures have been described previously (Ushiyama & Matthews, I97o ). Thin sections were collected on unfilmed 4oo-mesh grids or formvarcoated Veco large hole grids. Sections were examined in a Philips EM 2oo electron microscope. Freeze fracturing. Before freeze-fracturing, the following pretreatments were usually carried out. Isolated chloroplasts or pieces of leaf tissue were fixed in 6 ~ glutaraldehyde in the isolation medium or Millonig's buffer for about 2 h at room temperature. After washing with the medium or buffer, they were soaked in 25 ~ glycerol in the appropriate buffer for 3o rain or more in a refrigerator. The freeze-fracture method used was as described by Bullivant & Ames (I966) and Bullivant (I969). Specimens were fractured under liquid nitrogen at -~96 C. Shadowing and backing were done through tunnels in a cold block, thus protecting the cold specimen surface against contamination. Since replication was carried out with the specimen temperature at about - t4o C, there was no etching.

3 A _2~1 B Chloroplast vesicles induced by TYMV Extracellular space membrane Plasma 39 oplasm A t ~'1 I t ATAyVesicle \,\ \'\ ff/'kl inner ~k ~,,\ I[[/([~-membrane~]l/!/ ~s Vesicle outer membrane Chloroplast outer membrane Chloroplast inner membrane Fig. I. Orientation of the A and B faces of the plasma membrane, the outer and inner chloroplast membranes, and the membranes of the peripheral vesicles induced by TYMV. This schematic representation is based on the assumption that the vesicle membranes are derived from the chloroplast membranes by invagination. The fracture paths within the membranes are indicated by a dotted line. To obtain the complementary replicas of both sides of the fracture the technique of Chalcroft & Bullivant (I97O) was used, modified according to Moor Two gold London finder grids (Graticules Ltd) were stuck together with a very thin layer of vacuum grease so that the numbered and lettered arrays of holes were in register. The grids were clamped in between the two halves of the double fracture holder. The chloroplast suspension was introduced into the holder so that it penetrated through the grid holes. The specimen freeze-fractures in line with the layer of grease between the two finder grids. To find complementary areas it was necessary only to locate corresponding grid holes. RESULTS Freeze-fracture of chloroplast membranes Before describing the freeze-fracture appearance of the chloroplast limiting membranes and the virus-induced vesicles, it is necessary to describe the manner in which membranes freeze-fracture and the terminology to be used to denote the fracture faces produced. Membranes generally freeze-fracture along a unique hydrophobically bonded interior plane (Branton, I966, I97I) and this splitting thus exposes two faces. For the plasma membrane of a cell, the face of that membrane portion left frozen to the cytoplasm is referred to as an A face; and the face of that portion left frozen to the extracellular space is referred to as the B face (McNutt & Weinstein, I97o, Fig. ~). Freeze-fracturing shows particles within the membrane. In general, more remain adhering to the A face than to the B face. If one proceeds inwards from the plasma membrane, alternate membranes show reversal of the A and B face direction (Branton, I969). The plasma membrane fractures so that the A face faces out; the outer membrane of the chloroplast so the B face faces out; and the inner membrane of the chloroplast so that the A face faces out (Fig. I). It has been pointed out that this would be the case if each successive membrane system was formed by invagination of the previous one (Bullivant, I969). On the assumption that the vesicles are formed by invagination, then the vesicle inner membrane A face will face out, and the vesicle outer membrane B will face out (Fig. D-

4 40 T. HATTA, S. BULLIVANT AND R. E. F. MATTHEWS _./B Chloroplast outer membrane Chloroplast inner membrane Fig. 2. A model for the path of the fracture through the necks of the peripheral vesicles induced by TYMV. Fracture paths are indicated by dotted lines. If the fracture passes through the outer chloroplast membranes we get (all represented as a section of what would be seen from above) : For the B face For the A face For a fracture passing through the inner chloroplast membrane: For the A face For the B face The structure and distribution of the necks of the peripheral vesicles The illustrations of the membranes of chloroplasts with peripheral vesicles in Figs. 3 and 4 will be most readily understood if we consider first a model based on the examination of many replicas. In Fig. 2 we have drawn a peripheral vesicle assuming that the outer vesicle membrane is derived from the inner chloroplast membrane by invagination, that the inner vesicle membrane is similarly derived from the outer chloroplast membrane, and that the interior of the vesicle is connected to the extra-chloroplastic space by a narrow lumen in the neck. The legend to Fig. 2 illustrates the four different aspects the neck will present, depending upon which chloroplast membrane is fractured, and which face is viewed. Figs. 3 and 4 illustrate the four aspects of the membranes shown diagrammatically in Fig. 2. In particular, the complementary faces of inner chloroplast membrane (Fig. 4a, b)

5 Chloroplast vesicles induced by T Y M V Fig. 3- Aspects of the chloroplast membranes : TYMV infected leaf tissue was pretreated and then freeze-fractured as described under Methods. AO, A face of outer membrane; BO, B face of outer membrane; AI, A face of inner membrane; BI, B face of inner membrane; M, "mobile phase". Arrow indicates position of an atypical outer A face vesicle neck (see text). On all micrographs a circled arrow indicates the shadowing direction. Bar markers = I #m. 41

6 42 T. HAT'I-A, S. B U L L I V A N T AND R. E. F. M A T T H E W S Fig. 4. Aspects of the inner membrane fracture faces: chloroplasts isolated from TYMV-infected leaf tissue were pretreated and freeze-fractured as described under Methods. (a) represents the A face on the inner chloroplast membrane. (b) is from a complementary replica to that in (a) and represents the B face of the inner membrane. Complementary replicas were photographed in the electron microscope with one grid inverted with respect to the other, compared with their orientation when they were shadowed. Thus the micrographs do not appear as mirror images of each other and the corresponding structures are easily identified by their position in the micrograph. However, a consequence of this is that the shadow comes from opposite directions on the two micrographs. The complementary annuli on both may appear raised. However, if (a) is correctly orientated by turning the page upside down so that the shadow comes from below, then the annuli are seen correctly as depressions (compare with diagrams of Fig. 2). (e) shows the B face of the inner membrane at higher magnification with a cluster of vesicle necks. Bar markers = IOO nm. confirm the m o d e l for this m e m b r a n e. Similar c o m p l e m e n t a r y faces n o t illustrated confirm the m o d e l for the outer chloroplast m e m b r a n e. The chloroplasts shown in Fig. 3 illustrate the fact that in the m a j o r i t y o f chloroplasts e x a m i n e d the vesicles were n o t r a n d o m l y distributed but were arrayed in clusters on the surface, frequently in an hexagonal array. This clustering was seen b o t h in isolated chloroplasts and in freeze-fractured intact leaf tissue.

7 Chloroplast vesicles induced by TYMV 43 Fig. 5. Smooth areas on B face of outer membrane. Isolated chloroplasts from TYMV infected leaf, pretreated and freeze-fractured as described under Methods. Arrows indicate relatively smooth areas above larger vesicles which have no necks. Typical arrays of depressions for B face outer membrane necks of small vesicles can also be seen. Bar marker nm. In freeze-fracture replicas of chloroplasts we have observed protuberances of the stroma (corresponding to the 'mobile phase' illustrated by light microscopy and discussed by Wildman, 1967). There was no obvious difference in the nature or frequency of these protuberances between chloroplasts in healthy or TYMV-infected leaf. Other types of vesicle Occasionally when viewing the outer membrane B face, we have observed raised rounded areas of the chloroplast surface which are free or almost free of particles. These areas appear to lie over vesicles some or all of which do not have necks (Fig. 5). Over these raised areas the fracture plane frequently shifts to the membrane below. This is revealed as having no particles, confirming that it is a vesicle inner membrane A face rather than a chloroplast inner membrane A face (see section below). In thin sections vesicles are occasionally seen in which the outer membrane is clearly open to the cytoplasm and in which the vesicle formed by the inner membrane protrudes into the cytoplasm. Fig. 6 indicates sites where the shadowed replica has been lost during processing to remove the biological specimen. The most probable interpretation of these holes is that small vesicles were protruding from the chloroplast surface at these positions. Occasionally, in the A face of the outer chloroplast membrane, we have seen raised annuli which should not be present on the model presented in Fig. 2 (e.g. arrowed position in Fig. 3a). They may represent the aperture in the outer membrane through which such vesicles escape. Vesicle structure revealed in fractures through the chloroplast In the chloroplast illustrated in Fig. 7, the fracture plane has passed through the body of the chloroplast. For some of the vesicles the fracture has passed through the inner vesicle membrane, leaving the inner part of the vesicle itself adhering to the chloroplast. The flasklike shape of the vesicles is clearly revealed. For other vesicles, the vesicle itself has been removed revealing the B face of the inner vesicle membrane or the A face of the outer membrane. The diam. of the necks seen in the main part of Fig. 7, measured to include two half inner membranes, is approximately

8 44 T. H A T T A, S. B U L L I V A N T A N D R. E. F. M A T T H E W S Fig. 6. Freeze-fracture of part of a cluster of chloroplasts in a cell infected with TYMV. Leaf tissue prefixed and freeze-fractured as described under Methods. Arrows indicate sites at which the shadowed replica has been lost, indicating the previous position of a vesicle protruding into the cytoplasm. The aspect seen over most of the chloroplast is outer membrane B face. Bar marker = I#m. 20 rim. Some of the vesicle necks seen in Fig. 7 (and in other similar fractures) appear to extend to the chloroplast surface. However, this cannot be taken as proof of patency, since we do not know whether the fracture line representing the chloroplast surface in the replicas has passed through the inner or the outer chloroplast membrane. The membrane faces seen in Fig. 7 suggest that the A face of the inner vesicle membrane does not possess particles, contrary to what would be expected if it was an invagination of unaltered outer chloroplast membrane. This feature of the vesicles was proved by the examination of complementary replicas as is illustrated in Fig. 8. As expected the A face of the outer vesicle membrane has many particles. Both the A and B faces of the inner vesicle membranes are entirely smooth.

9 Chloroplast vesicles induced by T Y M V 45 Fig. 7- Aspects of vesicles seen when the fracture plane passes through the chloroplast. Isolated chloroplasts from TYMV-infected leaf pretreated and freeze-fractured as described under Methods. AI, A face of inner vesicle membrane; BI, B face of inner vesicle membrane; AO, A face of outer vesicle membrane. Inset: vesicle showing neck extending towards chloroplast surface. Bar markers = IOO n m. Frequency of particles in chloroplast membranefaces Inspection of m a n y micrographs indicated that besides causing the development of the smooth inner vesicle membrane, T Y M V brought about other changes in the distribution and nature of the particles seen in membrane fracture faces. Particle counts per unit area o f chloroplast fracture face were made, care being taken to select areas that were close to horizontal in the original specimen. Results are summarized in Table ~. As would be expected from work on other biological membranes, the B faces of almost all the membranes had significantly fewer I o nm particles than the corresponding A faces. The only exception was the outer membrane of infected chloroplasts in non-vesicle areas. Here the number of particles on the B face was increased to about the same number as on the A face. There were no significant differences in numbers of particles between faces of membranes of normal chloroplasts, and corresponding faces of membranes of infected chloroplasts in non-vesicle bearing areas, except that the B face of the outer membrane of diseased chloroplasts had a significantly increased number of particles (P = < o.oi). The vesicle-bearing and non-vesicle-bearing areas of infected chloroplasts differed in several respects with respect to their membrane particles. (i) Vesicle areas had about half as many particles as non-vesicle areas on the A face of the inner membrane (P = < o.oi). (ii) They also had about 66 ~ as many particles as non-vesicle areas on the inner membrane 4 VIR 20

10 46 T. HATTA, S. B U L L I V A N T AND R. E. F. MATTHEWS Fig. 8. Complementary replicas demonstrating the altered character of the vesicle inner membrane. Chloroplasts isolated from TYMV infected leaf and pretreated and freeze-fractured as described under Methods. The complementary replicas were photographed as described under Fig. 4. Thus to gain the correct conformational impression, (a) should be viewed with the page irtverted. AO, A face of outer vesicle membrane; BO, B face of outer vesicle membrane; AI, A face of inner vesicle membrane; BI, B face of inner vesicle membrane. Bar markers = 5o0 nm. B face a n d the outer m e m b r a n e A face (P = < 0.05). (iii) The n u m b e r s given in the table for the outer m e m b r a n e B face indicated that vesicle-bearing areas here also have a b o u t as m a n y ~o n m particles (P -- < o. o 0. However, in addition to the reduced n u m b e r s of m e m b r a n e particles of standard size, these areas bore very n u m e r o u s smaller particles n o t seen in a n y other chloroplast m e m b r a n e s examined (Fig. 9). Occasionally (as seen in Fig. 6) the area bearing n u m e r o u s smaller particles-extended b e y o n d the immediate vicinity of the

11 Chloroplast vesicles induced by T Y M V 47 Table I. Frequency of particles (approximately IO nm in diam.) seen on various membrane fracture faces of healthy and TYMV-infected Chinese cabbage chloroplasts M e a n * number o f particles//zm 2 ), Membrane and face Inner Inner Outer Outer A B A B Healthy ~ (342) (152) (223) (I45) Infected (non-vesicle areas) (646) (96) (477) (168) Infected (vesicle areas) IO o 43o (3o4) (I39) (I98) 0o9)t * Each n u m b e r is the mean of 6 to 8 estimates each from a different micrograph. N u m b e r s in parentheses = standard deviation o f the mean. t In these areas large numbers of particles with diameters ~ 5 n m were seen in addition to the IO n m particles. Fig. 9. Distribution of IO n m particles and smaller particles on the B face of the outer m e m b r a n e o f a chloroplast isolated from TYMV-infected tissue. Chloroplast preparation prefixed and freeze-fractured as described under Methods. A n area with an array o f small vesicles is in centre o f photograph. The A face o f the inner m e m b r a n e bearing many lo n m particies can be seen at the top of the pbotograph. Bar marker = 5oo nm. vesicle clusters, and appeared to correspond to a flattened area of chloroplast surface that had been adjacent to another chloroplast in a clump. DISCUSSION As seen in thin sections of chloroplasts the small peripheral vesicles are particularly abundant beneath those areas of the chloroplast's surface that are adjacent to other chloroplasts in a clump (Lafl6che & Boy6, 1969; Ushiyama & Matthews, 197o ). Freeze-fracture replicas clearly demonstrate the clustered distribution of the vesicles. The clusters of vesicles may usually occupy part or all of areas of contact between chloroplasts in the intact cell. 4-2

12 4 8 T. HATTA, S. BULLIVANT AND R. E. F. MATTHEWS Studies of various membrane systems have led to the generalization that there is a correlation between metabolic activity and number of particles presumably containing protein, seen in a freeze-fractured membrane (Branton, I970. Physiologically active membranes such as chloroplast lamellae have many particles, while myelin layers which function as insulators have none. The B face of the outer chloroplast membrane in the region of the vesicle clusters is rich in particles most of which appear to be smaller in diam. than the standard Io nm particles. This feature may reflect enhanced metabolic activity of the membrane related to virus synthesis. It may also reflect changed physical or chemical properties of the outer membrane that lead to the clumping of chloroplasts in infected cells. In thin sections the chloroplast membranes that lie close to adjacent chloroplasts in a clump can usually be seen to be stained more heavily than elsewhere. This increase in uptake of stain may involve the particles seen on the outer membrane B face. The clumping of the diseased chloroplasts shows organelle specificity, since the mitochondria, nucleus and the cell membranes are not included. It may be that glycoproteins or glycolipids are involved in the specific inter-organelle recognition, as has been postulated for the intercellular adhesive specificity of embryonic animal cells (Roth, McGuire & Roseman, I97I). Thin sections show occasional vesicles which are undoubtedly patent to the exterior of the chloroplast, the inner vesicle membrane being continuous with the outer chloroplast membrane (Lafl6che & Bov6, I969). These may be exceptional or abnormal vesicles. Using sections 6o0 to 9oo A thick it is usually impossible to observe detail in vesicle necks with an external diam. of about 2oo A. We had hoped that observations on complementary replicas might demonstrate whether or not the majority of the vesicles are patent. With hindsight, it is apparent that this technique cannot decide the issue. It would make no difference to the two fracture planes shown in Fig. 2 if the small lumen we have illustrated was in fact completely closed in the region between the two possible fracture planes. Various electron dense markers such as ferritin (Bruns & Palade, 1968) and lanthanum salts (Revel & Karnovsky, I967) have been used to test for the patency of vesicles, capillary vessels, etc. Ferritin, with a diam. of ~o5 A, may be too large to use for the chloroplast vesicles. In experiments not reported in detail here we have attempted to use lanthanum hydroxide, but because of technical difficulties we have not obtained any decisive results. The pinocytotic vesicles of endothelial cells of continuous capillaries are undoubtedly patent (Nickel & Grieshaber, I969). The appearance of the peripheral chloroplast vesicles and the endothelial cell vesicles seen in freeze-fracture replicas show remarkable similarities, except that the chloroplast vesicles are more complex in having two membranes rather than one. On balance it seems likely that most of the small peripheral vesicles have a very narrow channel opening to the cytoplasmic space. Some of the vesicles beneath the smooth raised areas illustrated in Fig. 5 appear to have no opening to the cytoplasm. They bear some resemblance to the surface views of budding influenza virus particles, where the membrane above the virus is free of particles (B/ichi et al ~969) to the budding Sindbis virus (Brown, Waite & Pfefferkorn, r972) and to the bulges seen on the nerve cell membranes overlying the neurosecretory granules in the posterior pituitary of the rat (G. P. Dempsey, S. Bullivant & W. B. Watkins, unpublished observations). In this last example the disappearance of the membrane-associated particles may be related to the later fusion and exocytosis of the neurosecretory granule. The interpretation of the raised vesicles without necks illustrated in Fig. 5 must await further work. They probably represent another class of vesicle, perhaps those without

13 Chloroplast vesicles induced by TYMV 49 electron dense contents which are larger than the peripheral vesicles and which may escape to the cytoplasm (Chalcroft & Matthews, I966 ). Almost all cell membranes that have been studied by freeze-fracturing methods show an A face covered with many particles and a B face with fewer particles (Branton, I97Q. The inner membrane of the chloroplast vesicles shows no particles on either its A face or its B face. Thus it is ultrastructurally distinct from the outer chloroplast membrane from which it has been considered to be derived by invagination (Lafl6che & Bov6, I969). If the inner vesicle membrane is derived from the outer chloroplast membrane, then TYMV infection must cause some change in the membrane that leads to the disappearance of the large proteins that make up the particles normally seen within membranes. Protuberances into the cytoplasm corresponding to the 'mobile phase' activity postulated by Wildman (~ 967) on the basis of light microscope observations are seen rather infrequently in thin sections. It may be that most of them disappear during the fixation and embedding procedure. Our observations on freeze-fractured chloroplasts provide independent confirmation for the reality of these protuberances. In experiments not detailed here using light microscopic examination of living cells we have attempted to measure the vol. and surface area of healthy and diseased chloroplasts assuming the healthy chloroplast to be a cap of a sphere, and the infected chloroplast a sphere. The means of measurements on I2o healthy and ~2o diseased chloroplasts were as follows: Vol. of healthy = 48 /~ma; vol. of virus infected = 75 /zm~; surface area of healthy = 88/tm 3 and of virus infected, 86/~m 3. We further calculated that the area of membrane involved in vesicle formation would be 20 to 8o ~ that of the healthy chloroplast. Taken at face value, and assuming the membranes do not stretch, these figures suggest that the change in shape of the diseased chloroplast from saucer-shaped to approximately spherical uses all the existing membrane, and that vesicle formation must require the synthesis of additional membrane. This conclusion is not valid unless we assume that the amount of chloroplast membrane involved in protuberances of the 'mobile phase' into the cytoplasm is the same in healthy and infected chloroplasts. Quantitative evidence on this point would be difficult to obtain. We thank Mr G. Grayston for skilled technical assistance. One of us (T. H.) was supported by a New Zealand Universities Post-graduate scholarship. REFERENCES B,~CHI, T., GERHARD, W., LINDENMANN, J. & M/.JHLETHALER, K. (I969). Morphogenesis of Influenza virus A in Erlich Ascites Tumour Cells as revealed by thin sectioning and freeze-etching. Journal of Virology 4, BRANTON, D. (1966). Fracture faces of frozen membranes. Proceedings of the National Academy of Sciences of the United States of America 55, Io48-1o56. BRANTON, D. (I969). Membrane structure. Annual Review of Plant Physiology 2o, 2o BRANTON, D. (I970. Freeze etching studies of membrane structure. Philosophical Transactions of the Royal Society of London. B z6x, I33-I38. BROWN, D. T., WHITE, M. g. V. & VFEEFERKORN, E. R. (X972). Morphology and morphogenesis of Sindbis virus as seen with freeze etching techniques. Journal of Virology xo, 5z BRUNS, g. g. & PALADE, O. E. (I968). Studies on blood capillaries. II. Transport of ferritin molecules across the wall of muscle capillaries. Journal of Cell Biology 37, BULLIVANT, S. & AMES, A. (I966). A simple freeze-fracture replication method for electron microscopy. Journal of Cell Biology zg, BULLIVANT, S. (I969). Freeze-fracturing of biological materials. Micron x, CHALCROFT, J. V. & BULLIVANT, S. (1970). An interpretation of liver cell membrane and junction structure based on observation of freeze-fracture replicas of both sides of the fracture. Journal of Cell Biology 47, 49-6o.

14 5 0 T. HATTA, S. BULLIVANT AND R. E. F. MATTHEWS CHALCROFX, S. & MATTHEWS, R. E. F. (I966). Cytological changes induced by turnip yellow mosaic virus in chinese cabbage leaves. Virology 28, CHALCROFT, J. & MATTHEWS, R. E. r. (I967). Role of virus strains and leaf ontogeny in the production of mosaic patterns by turnip yellow mosaic virus. Virology 33, LAFL/~CHE, D. & BOV~, J. M. (I969). Development of double membrane vesicles in chloroplasts from turnip yellow mosaic virus infected ceils. Progress in Photosynthesis Research r, McNUTT, N. S. & WEINSTEIN, R. S. (I970). The ultrastructure of the nexus. A correlated thin-section and freezecleave study. Journal of Cell Biology 47, MATTHEWS, R. E. F. (I970), In Plant Virology, 778 pp. New York and London: Academic Press. MATTHEWS, R. E. F. (I973). Induction of disease by viruses with special reference to turnip yellow mosaic virus. Annual Reviews of Phytopathology (in the press). MILLONIG, G. (196D. Advantages of a phosphate buffer for OsO~ solutions in fixation. Journal of Applied Physics 32, I637. MOOR, H. (I97D. Recent progress in the freeze-etching technique. Philosophical Transactions of the Royal Society of London. B 26I, I2I-I3I. NTCKEL, E, & GRIESHABER, E, (I969). Elektronenmikroskopische Darstellung der Muskelkapillaren im Gefriert/izbild. Zeitschrift fiir Zellforschung und mikroskopisehe Anatomic 95, L REVEL, J. P. & KARNOVSKY, M. J. (I967). Hexagonal array of subunits in intercellular junctions of the mouse heart and liver. Journal of Cell Biology 33, C7-CI2. ROTrI, S., McGUIRE, E. J, & ROSEM~,N, S, (I97D. An assay for intercellular adhesive specificity. Journal of Cell Biology 5 r, USHIYAMA~ R. & MATTHEWS, g. E. F. (I970). The significance of chloroplast abnormalities associated with infection by turnip yellow mosaic virus. Virology 42, 293-3o3. WILDMAN, S. G. (I967). The organisation of grana-containing chloroplasts in relation to location of some enzymatic systems concerned with photosynthesis, protein synthesis and ribonucleic acid synthesis. In Biochemistry of Chloroplasts 2, (Proceedings of a NATO Advanced Study Institute. Aberystwyth, I965). Edited by T. W. Goodwin. New York and London: Academic Press. (Received 4 January 1973)