A non-connexon protein (MIP) is involved in eye lens gap-junction formation

Transcription

1 A non-connexon protein (MIP) is involved in eye lens gap-junction formation W. T. M. GRUIJTERS Department of Cellular and Molecular Biology, University of Auckland, Auckland, Xew Zealand Summary New immunolocalization data put the role of the lens MP26 (MIP) protein in a new perspective. During maturation of lens fibre cells, MIP is found to associate specifically with two structures, gap junctions and cell interlocking processes (known as ball and socket domains). It is significant that the zone in which these associations are most striking is discrete, coinciding with the zone of rapidly enlarging junctional plaques and newly forming ball and socket domains. Observation of domain-specific interactions of MIP with forming gap junctions and ball and socket domains suggests that MIP may be involved in the formation of close membrane appositions. Furthermore, previous ambiguities in the literature over the presence of MIP in gap junctions are clarified by the knowledge that, in situ, MIP associates strongly with gap junctions for only a brief period (with less than about 5 % of all lens gap junctions at any one time) during the assembly of junctional plaques. Key words: gap-junction plaque development, MIP, MP26, MP70, lens, membranes, immunofluorescence. Introduction Studies on the distribution of a non-connexon protein (MIP) in the lens with respect to gap junctions have led to apparently conflicting results. MIP has been seen exclusively in non-junctional membranes (Paul & Goodenough, 1983) and included in junctional and non-junctional membranes (Bok et al. 1982), leading to much speculation about its function. MP70, a proven lens gapjunction-specific component (Gruijters et al ), with homology toconnexon proteins (Kistlerefa/. 1988), was used to map gap-junctional development (Gruijters et al. 1987a). In contrast, MIP has no connexin homology (Gorin et al. 1984). MP70 is therefore ideal as a junctional marker to map concurrently the distribution of MIP. Mapping results show that MIP associates specifically with gap junctions for only a short period during junctional development, while no such specific association of MIP is detected in fully formed junctions. MIP distribution is also related to cellular interlocking structures on lens fibres known as ball and socket domains. MIP has been shown to be associated with these structures in the past (Paul & Goodenough, 1983). More detailed information on MlP/ball and socket associations is now presented. Materials and methods Immunoreagents Monoclonal anti-mp70 antibodies were derived from hybrid- Journal of Cell Science 93, (1989) Printed in Great Britain The Company of Biologists Limited 1989 oma 6-4-B2-C6 and polyclonal anti-mip26 antibodies from rabbit J12. Both antibody specificities had been fully characterized (Kistler et al. 1985). Fluorescein- and rhodamine-coupled secondary antibodies and the lectin Arachis liypogaea (peanut) agglutinin (PNA) were from Serotec (Bicester, Oxon, England). Dissection Eye lenses were removed from sheep immediately after death and placed directly into 2% formaldehyde in phosphatebuffered saline at room temperature. Fixed lenses were used within 30 h of extraction. Dissection of lenses began by cutting the lens into quarters along the pole-to-pole axis. The nuclear portion of one quarter was removed with tweezers and successive layers of fibres in the remaining cortex were peeled back until the desired layers of cells were reached. Small bundles of each of these layers were removed for immunolabelling. All the bundles used in this study came from less than 1 mm below the lens outer surface. A consequence of the dissection technique is that, because the bundles lie flat on the microscope slide, all micrographs show longitudinal views of the elongated cells. Immunofluorescence microscopy The distribution, size and shape of junctional plaques are characteristic of their position in the lens and the stage of cellular differentiation (Gruijters et al. 1987o). The present investigation of MIP distribution focuses on the extreme outer cortex and bow region where junctional plaque formation to completion has been seen to occur. This region is unlike mature fibre cell regions in that MIP is found in forming junctional plaques at levels detectable above those in the surrounding membranes. By selecting fibres of different lengths from the outer cortex and bow region, an age-related sequence of MIP 509

2 distribution can be produced. Antibody labelling of dissected fibres was carried out in a spotting dish using a 1 h incubation time from primary and secondary antibodies, with washing between antibody changes done by transferring the fibre bundles three times into fresh PBS. Double labelling was carried out in a similar fashion except that anti-mip and anti- MP70 antibodies were mixed and added to the fibres in a single labelling step. Antibody binding was detected with fluoresceinor rhodamine-conjugated, species-specific secondary antibodies. Control experiments included labelling with each primary antibody singly to ensure that binding of one antibody was not sterically hindered by the other, and to detect any unusual cross-reactivity of the secondary antibody between other secondary and heterologous primary antibodies. Lectin labelling was done as previously described for MIP except that the fluorescein was conjugated directly to the lectin (Arachis hypogaea (peanut) agglutinin (PNA)), negating the need for secondary labelling. Micrographs were taken on an epifluorescence microscope and recorded on Kodak Tri-X or Fuji HR1600 films. Results Studying in situ MIP distribution reveals an important characteristic of this protein with respect to gap junctions. Intense labelling of gap junction plaques (Figs 1A,B,C, 2A) only appears in a narrow zone of developing fibres where gap junctions are growing rapidly in size. For readers unfamiliar with lens structure please refer to Gruijters et al. (1987a). Because of the nature of the technique of labelling very small bundles of dissected-out fibres, it is not easy to be quantitative, but it is estimated that less than 5 % of all gap junction plaques in the lens have such a pronounced specificity of MIP association. At later stages of junctional formation (Figs ID, 2B), junctional plaques have no detectable specific association of MIP, but at these same stages MIP labelling of newly forming ball and socket domains is high compared with the level on the broad sides of the fibre (Fig. ID). Levels of MIP appear to remain elevated to some extent in more mature ball and socket domains as seen in Fig. 2B, although this effect may partially result from viewing these structures in profile rather than only en face. A number of black and white half-tone micrographs were also made (Fig. 1A,B) because they have superior resolution to the coloured ones. Resolution was further improved by labelling samples with only one primary and one secondary antibody. Photographs of fibres exposed to only anti-mip primary and fluorescein-coupled secondary antibodies are shown in Fig. 2A and B. Note the 'spotty' distribution of MIP in the higher-resolution micrograph (Fig. 2A). The single labelling was also used to exclude the possibility of interference with the second set of primary and secondary antibodies. Figs ID, 2A and 2B are all of immature full-length fibres, as indicated by their position in the lens and the fact that ball and socket formation is only complete along parts of their length. This is a frequent occurrence as ball and socket domains form on different parts of the same fibre at different times during its development. Therefore, Fig. 2B shows a slightly more advanced stage of ball and socket formation than Fig. ID, though of similar age. Figs ID and 2B demonstrate that while there is strong labelling of the ball and socket domains, the broad sides of the fibre label uniformly at a lower level. Hence, no enhanced association of MIP is detectable in the mature gap junctions, although the uniform labelling indicates that MIP levels are similar in junctional and non-junctional membrane regions. Occasionally 'negative plaques' have been observed, possibly indicating exclusion of MIP from a membrane domain. To indicate what might be expected if MIP were really excluded from gap junctions, lectin labelling does indeed demonstrate large negative plaques (Fig. 2C). Negative plaques, either these large ones against a lectin background (Fig. 2C) or the smaller ones against the MIP background (Fig. 2B), do demonstrate that this technique would show gap-junction-sized MIP negative plaques if they existed. One can thus conclude that the generally uniform MIP labelling on the broad side of the fibre indicates that the MIP levels in mature gap junctions are approximately the same as those in the surrounding non-junctional membrane. It is likely that the small, infrequent MIP negative plaques may simply result from damage during dissection. An unusual aspect of the specific association of MIP with forming gap junctions is the 'spotty' distribution of the label in such regions, seen better in the black and white micrographs (Figs 1A, 2A). Such a distribution of label could be associated with preparation artefacts such as fixation-induced protein clumping or antibody crosslinking, but there are internal controls that make such an explanation improbable: (1) layers containing fibres displaying spotty labelling (Figs 1A,C, 2A) are sandwiched between uniformly labelled less-mature and more-mature fibre cell layers. (2) A single immature cell will often have transitions along its length between a uniform MIP distribution and a spotty distribution (not shown). Discussion The main finding of this paper is that there is a domainspecific association of MIP with forming gap junction plaques and ball and socket domains. MIP specificity is transient in gap junctions, occurring only in association with the formation of junctional plaques. Such a specific association of a non-connexon protein with gap junctions during their development suggests a role for the protein in the formation of these junctions. MIP levels in ball and socket domains appear greater during their development than in the mature structures, although quantification is subjective. It is also important to consider what the resting level of MIP is in the majority of gap junctions, i.e. those gap junctions without high levels of MIP. While being aware that light intensities seen in immunofluorescence are relative, the best answer appears to be that these 'resting' levels are similar to and certainly not greater than those in the non-junctional surrounding plasma membranes. Negative plaques could be interpreted to mean that MIP levels are indeed below that of the surrounding mem- 510 W. T.M. Gruijters

3 Fig. 1. Double labelling of dissected lens fibres with anti-mip and -MP70 primary antibodies. MIP appears yellow-green in the colour micrograph (C) (fluorescein-eoupled secondary antibody) while gap junctions with bound anti-mp70 appear red (rhodamine-coupled secondary antibody). The half-tone micrographs (A and B) are of the same field as the colour one (C) as indicated by the brackets in the latter. A is exposed only to show MIP, B only for MP70, while the colour micrograph represents a double exposure of the same field using the two appropriate filter sets to show labels for both MIP and MP70. An area towards the lens pole of an approximately 8 mm length fibre is shown. Junctional plaques appear to grow rapidly at the expense of smaller ones (Gruijters et al. 1987a), an indication of possible mechanisms of plaque formation. In these highly localized areas, MIP association with junctions is at its most specific with levels in junctions well above those in the surrounding membrane. A-C, X1400. D is labelled in the same fashion as C. Full-length (over 10 mm) immature lens fibres showing an area near to the lens pole. Ball and socket domains (convoluted green bands) have raised levels of MIP, which appear to occur from the time of formation as indicated by their position in the lens (Kuszak et al. 1980). Note the relative lack of MIP within junctional plaques when compared with C. X850.

4

5 brane, but as pointed out in Results the number of negative plaques and their size were well below the number and size of the actual gap junctions. At this stage it seems likely that these small negative plaques result from mechanical damage during the dissection and processing of the fibre bundles. Hence, the secondary finding of this paper is that there is a resting level of MIP, similar to that in the surrounding membrane, in mature gap junctions of the lens in situ. This is supported by other in situ work (Fitzgerald et al. 1983; Vallon et al. 1985). However, as already mentioned, there have been contradictory results when Fig. 2. A. Labelling with only anti-mip primary and the appropriate secondary antibody improves resolution over the doublelabelled micrographs (Fig. 1). When associated with gap junctions in young fibres, MIP has a typical spotty appearance in immunofluorescence. This distribution of MIP was also noted by Bok et al. (1982) using a higher-resolution technique (immunogold label viewed by TEM). X1400. B. Full-length immature lens fibres, showing the 'non-spotty' distribution of MIP near the equator with apparently elevated levels in ball and socket domains that are typical of the majority of lens fibres. Despite extensive searches, the negative plaques shown (arrows) were the most prominent found. X1400. C. Labelling of dissected fibre bundle using fluorescein-conjugated PNA. Note the raised levels in ball and socket domains and exclusion of label from presumed gap-junction plaques. X1400. MIP distribu tion 511

6 localizing MIP in isolated lens gap junctions (Paul & Goodenough, 1983; Boketal. 1982). Recent work in our laboratory (Kistler, personal communication), using gold labelling in combination with negative stain, revealed no MIP labelling in isolated lens gap junctions. The question must be left open until these differences between results from in situ studies and some isolated preparations have been more fully explained. Conflicting results with respect to MIP's presence at high levels in gap junctions may now be reconciled with the knowledge that MIP concentrations in gap junctions in vivo vary considerably depending on the developmental state of junctional plaques, and hence on their position within the lens. Isolation of junctions does not enable the position in the lens from which the junctions came to be determined and could also lead to the loss of MIP from gap junctions or to inability to label it. For example, Johnson et al. (1988) used two techniques to isolate junctions. One involved preadsorption of total membrane protein onto polyvinyl ELISA plates prior to immunolabelling for MIP, resulting in the labelling of MIP in gap junctions. The second technique involved labelling of total membrane protein while in suspension and pelleting by centrifugation, resulting in an inability to label the isolated junctions with MIP label. If it is realized that only a small fraction of junctional plaques contains relatively high amounts of MIP then it becomes clear that the first technique, which involves adsorption of membranes onto a surface, could lead to selection for MIPcontaining junctions. The second technique allows for no such selection. Despite extensive research, a proven in vivo function for MIP has not been found. Apparent in vitro channelforming properties (Gooden et al. 1985; Peracchia & Girsch, 1985; Zampighi et al. 1985) suggest a role for MIP in communication across cell boundaries. However, the present results, which show great variability in junctional MIP levels, taken together with the fact that MIP has no homology with the connexin family of proteins (Beyer et al. 1987), whereas MP70 does (Kistler et al. 1988), makes it unlikely that MIP has a function as a gap junction channel-forming protein. What is the role of MIP? Investigation of the distribution of MIP reveals two domain-specific MIP interactions in the lens: (1) gap junction plaques during formation, particularly during rapid aggregation of MP70 into large plaques. (2) Ball and socket domains. In both cases, specific MIP interaction is first detected as these domains form during fibre differentiation. Neither structure could form without two adjacent membranes being brought into close apposition. Ball and socket domains aid intercellular adhesion and the minimization of extracellular space in the lens (Kuszak et al. 1980), while gap junctions form a network of intercellular channels. Lens gap junctions in particular indicate close apposition when compared with junctions in liver and heart. The 'gap' between the adjacent membranes forming the lens junction is reduced (Rafferty & Esson, 1974). Such a correlation of MIP with structures requiring close membrane contact could suggest that MIP plays a role in bringing this about. Further support for this idea is forthcoming when the more uniform distribution of MIP in the bulk of the lens is considered. Close membrane apposition is essential for optimal lens transparency (Kuszak et al. 1980). It would not therefore be surprising to find the most ubiquitous membrane protein in the lens (MIP) associated with this function. Sequence data suggest that MIP is well anchored in the membrane with many hydrophobic sequences (Gorin et al. 1984), a property essential for any protein carrying out such a function. It is therefore speculated that MIP is a membrane 'zipper' protein, aiding the development and maintenance of membrane domains requiring close membrane apposition during fibre differentiation, and minimizing intercellular spaces throughout the whole of the lens, so aiding lens transparency. I am grateful to Professor S. Bullivant for assistance with the script. Many thanks to the workers at the Auckland Municipal Abattoir for cooperation in the collection of lenses. I am also grateful to Philip Schofield for his patience during the production of the colour figure. This work was supported by the Auckland Medical Research Foundation. References BEYER, E. C, PAUL, D. L. & GOODENOUGH, D. A. (1987). Connexin 43: A protein from rat heart homologous to a gap junction protein from liver. J. Cell Biol. 105, BOK, D., DOCKSTADER, J. & HORWITZ, J. (1982). Immunocytochemical localization of the lens main intrinsic polypeptide (MIP26) in communicating junctions. J. Cell Biol. 92, FITZGERALD, P. G., BOK, D. & HORWITZ, J. (1983). Immunocytochemical localization of the main intrinsic polypeptide (MIP) in ultrathin frozen sections of rat lens. J. Cell Biol. 97, GOODEN, M., RINTOUL, D., TAKEHANA, M. & TAKEMOTO, L. (1985). Major intrinsic polypeptide (MIP26) from lens membrane: reconstitution into vesicles and inhibition of channel forming activity by peptide antiserum. Biochem. biophvs. Res. Commun. 128, GORIN, M. B., YANCEY, S. B., CLINE, J., REVEL, J.-P. & HORWITZ, J. (1984). The MIP of bovine lens fiber membrane: characterization and structure based on cdna cloning. Cell 39, GRUIJTERS, W. T. M., KISTLER, J. & BULLIVANT, S. (1987O). Formation, distribution and dissociation of intercellular junctions in the lens. J. Cell Sci. 88, GRUIJTERS, W. T. M., KISTLER, J., BULLIVANT, S. & GOODENOUGH, D. A. (19876). Immunolocalization of MP70 in lens fiber nm intercellular junctions. J. Cell Biol. 104, JOHNSON, R. G., KLUKAS, K. A., TZE-HONG, L. & SPRAY, D. C. (1988). Antibodies to MP28 are localized to lens junctions, alter intercellular permeability and demonstrate increased expression during development. In Modern Cell Biology, vol. 7 (ed. E. L. Hertzberg & R. G. Johnson), pp New York: Alan R. Liss, Inc. KISTLER, J., CHRISTIE, D. & BULLIVANT, S. (1988). Homologies between gap junction proteins in lens, heart and liver. Nature, Loud. 331, KISTLER, J., KIRKLAND, B. & BULLIVANT, S. (1985). Identification of a 70,000 dalton protein in lens membrane junctional domains. J'. Cell Biol. 101, KUSZAK, J. R., ALCALA J. & MAISEL, H. (1980). The surface morphology of embryonic and adult chick lens-fibre cells. Am. J. Anal. 159, PAUL, D. L. & GOODENOUGH, D. A. (1983). Preparation, characterization and localization of antisera against bovine MP26, 512 W.T.M. Gruijters

7 an integral protein from lens fiber plasma membrane. J. Cell Biol. BENEDETTI, E.-L. (1985). MP26 in the bovine lens: a post- 96, embedding immunocytochemical study. Biol. Cell 53, PERACCHIA, C. & GlRSCH, S. J. (1985). Is the C-terminal arm of lens ZAMPIGHI, G. A., HALL, J. E. & KREMAN, M. (1985). Purified lens gap junction channel proteins the channel gate? Biochem. biophys. junctional protein forms channels in planar lipid films. Proc. natn. Res. Cowwitu. 133, Acad. Sci. U.S.A. 82, RAFFERTY, N. S. & ESSON, E. A. (1974). An electron-microscope study of adult mouse lens: some ultrastructural specializations..7. Ultrastnict. Res. 46, (Received 20 December 19SS Accepted, in revised fonn, VALLON, O., DUNIA, I., FARVARD-SERENO, C, HOEBEKE, J. & 21 April I9S9) MIP distribu tioti 513

8