MIP and MP70 are putative gap junction components in the plasma membranes of the mammalian lens fibre cells. We show now that MP70 can be solubilized separately from MIP in mild detergent solutions, and that this treatment results in the dissociation of the fibre gap junctions. Solubilized MP70 was isolated as 16’9 S particles by velocity gradient centrifugation and in the electron micro scope had the appearance of short double-membrane structures consistent with connexon-pairs. These observations open a new experimental avenue in which to characterize separately the two putative lens gap junction proteins structurally and functionally.

Gap junctions are clusters of transmembrane channels that connect the cytoplasms of adjacent cells. They allow small molecules to pass from cell to cell and thus play an important role in coordinating cellular signals and in tissue homeostasis. Gap junction proteins have been identified and have turned out to be tissue specific. Their apparent molecular weights are 21000 and 28 000 in the liver (Nicholson et al. 1987) and 47 000 in the heart (Page & Manjunath, 1985). On the basis of extensive sequence homology, they belong to a divergent family of gap junction proteins (Beyer et al. 1987; Nicholson et al. 1987). In lens fibres, two putative gap junction proteins with molecular weights 26000 (MIP26 or MIP; Revel et al. 1986) and 70000 (MP70; Kistler et al. 1985; Gruijters et al. 1987) have been identified. MP70 is another member of this family of gap junction proteins (Kistler et al. 1988), whereas MIP appears to be unrelated (Gorin et al. 1984). However, channel properties have been demonstrated for MIP molecules that were solubilized from fibre plasma membranes with strong detergent solutions and reconstituted into membrane vesicles (Gooden et al. 1985; Peracchia & Girsch, 1985) or into planar lipid films (Zampighi et al. 1985).

We now demonstrate that MP70 is soluble in considerably lower concentrations of detergent than those used to solubilize MIP, and that these putative junction proteins can be separated this way. Treatment with mild detergent dissociates the lens fibre gap junctions and solubilizes MP70 as 16-9 S particles. MP70 is thus an important structural protein of the lens fibre gap junctions and it will now be possible to characterize MP70 separately from MIP.

Membrane isolation

Normal lenses were from sheep generally less than 1 year old. They were extracted from the eyes within minutes of death at the local abattoir and immediately frozen in liquid nitrogen and stored at — 80°C until use.

Membranes were isolated exclusively from the lens outer cortex. Equatorial tangential 2 mm thick slices were cut from the frozen lenses and pooled. Tissue was homogenized in 5 mM-Tris-HCl, pH 8, 1 mM-EGTA, 1 mM-EDTA, 0-5 mM-diisopro-pyl-fluorophosphate, 10 mM-A’-ethyl maleimide (all reagents Sigma Chemical Co., St Louis, MO). All procedures were carried out on ice or at 4°C. Crude membranes were pelleted in an SS34 rotor (Sorvall, Wilmington, DE) at 9000 revs min−1 for 20 min, and were subsequently washed twice in the same buffer. The pellets were extracted once with 4 M-urea, 5 mM-Tris. HC1, pH9 ·5, 1 mM-EGTA, 1 mM-EDTA, and once with 7M-urea in the same buffer and pelleted at 15 000 revs min−1 for 60 min. Membranes were subsequently stripped of peripheral proteins with 20mM-NaOH in water (Hertzberg, 1984). In some cases, urea treatments were omitted and two alkali treatments were carried out to remove cytoskeletal and peripheral membrane material. After washing in phosphate-buffered saline (PBS) either membrane preparation appeared as a mixture of junctional plaques, and amorphous membrane vesicles and sheets when viewed by negative-stain electron microscopy.

Detergent treatment offibre membranes

To solubilize MP70, isolated lens fibre membranes were treated with PBS solutions containing variable concentrations of the non-ionic detergent Nonidet NP-40. For this, samples of alkali-stripped membranes (protein 4 mg ml−1) were mixed with equal volumes of PBS, and 0 ·02, 0’04, 0’08, 0 ·16, 0 ·32% Nonidet NP-40 in PBS. Mixtures were left at room temperature for 10 min and subsequently centrifuged (15000 revs min−1, SS34) to separate insoluble membrane protein and dissociated components. Proteins in supernatants and pellets were analysed by SDS-PAGE and Western blotting following the procedures and using antibodies described previously (Kistler et al. 1985). Alternatively, a dot-blot assay was used to detect solubilized MP70 and MIP in the supernatants. For this, small volumes of supernatant were spotted onto BA85 nitrocellulose membranes (Schleicher and Schuell Inc., Keene, NH) and the latter processed as described for Western blotting, except that radioactivity on the membranes was counted in a gamma counter.

Negative-stain immunocytochemistry

Lens fibre junctions in a mixture of junctional and non-junctional plasma membranes were identified by anti-MP70 immunogold labelling. Isolated membranes (protein 0 ·4-0 ·8 mgml”1), before or after detergent treatment, were adsorbed to carbon/collodion-coated grids for 1 min and nonadsorbed material was removed by floating the grids on several drops of PBS. Grids with the membranes facing down were floated on hybridoma supernatant 6-4-B2-C6 (Kistler et al. 1985) containing anti-MP70 antibodies (IgM) for 60min, washed in PBS and subsequently incubated for 60 min in a 1/10 dilution of 5 nm gold beads coated with anti-mouse IgM antibodies (Janssen Pharmaceutica, Beerse, Belgium). The Lanes j,k, Western blot of lanes d,e with anti-MlP antibodies. grids were extensively washed in several drops of water and then stained with 1% uranyl acetate. All procedures were carried out at room temperature. Specimens were viewed in a Philips 301 electron microscope and micrographs recorded on Kodak 4489 electron microscope film.

Isolation of MP70 oligomers

Velocity gradient centrifugation was used to purify and determine the sedimentation value of solubilized MP70. Typically, 50/71 alkali-stripped lens fibre membranes (protein 6 mg ml”) were mixed with an equal volume of 0 ·08% Nonidet NP-40 in PBS. This mixture was loaded onto a 5% to 20% sucrose gradient containing PBS and 04% Nonidet NP-40. The gradients were centrifuged in a Beckman SW50 rotor at 30 000 revs min−1 and 4°C for 16 h. Non-solubilized material sedimented to the bottom of the tube. Approximately 20 fractions were collected from each gradient and proteins detected by dot-blot assay or SDS-PAGE. Sedimentation values for MP70 particles in the peak fraction were calculated according to Young (1984). MP70 structures were visualized by negative-stain electron microscopy as described above.

Solubilization of MP70

MIP has previously been immunolocalized in both the lens fibre gap junctions and the non-junctional membranes (Fitzgerald et al. 1983). Membrane preparations contain such great amounts of MIP that SDS-polyacryl-amide gels have to be dramatically overloaded (80-150 μg protein per lane) in order to reveal any other membrane proteins by Coomassie staining (Fig. IB). Until now, this dominance of MIP has greatly complicated the characterization of the fibre gap junction component MP70.

The treatment of urea/alkali-stripped sheep lens fibre membranes with 0 ·04% Nonidet NP-40 solubilized MP70 (and its degradation product MP64) but only trace amounts of bulk MIP (Fig. 1D,E). Almost all MP70 was solubilized by treating fibre membranes with 0-16% Nonidet NP-40; however, this was at the expense of a slightly increased release of MIP from the membranes (Fig. 1F,G). Most M ÍP was contained in the pellet and hence insoluble under these conditions. These observations on Coomassie-stained gels were confirmed with immunoblots, using the same membrane samples and using the anti-MP70 (Fig. 1H,I) and anti-MfP (Fig. 1J,K) antibodies that had been characterized (Kistler et al. 1985). It should be noted that, although we generally refer to MP70, this protein may be mostly in the form of its cleavage product MP64. This cleavage occurs post-mortem, is variable between preparations (Kistler & Bullivant, 1987), and removes a peptide fragment from the C terminus of MP70 (Kistler et al. 1988).

The release of MP70 from the membranes was also monitored with an anti-MP70 dot-blot assay of the supernatants that had been generated by centrifugation of the detergent-treated fibre membranes (Fig. 2). Most MP70 was solubilized in Nonidet NP-40 solutions in the concentration range 0 ·02-0 ·04%, which is just above the critical micelle concentration of 0 ·3 niM or 0-018% for this detergent (Helenius & Simons, 1975). Most MIP remained in the pellet and relatively minor amounts of soluble MIP were detected with the dot-blot assay.

Electron microscopy of fibre membranes before and after detergent treatment showed that the solubilization of MP70 was paralleled by the dissociation of the fibre gap junction plaques. In untreated membrane preparations, gap junctions appeared as membrane plaques, often with the characteristic gap clearly visualized by negative staining (Fig. 3A). Anti-MP70 immunogold labelling of these preparations verified that these structures contained MP70 (Fig. 3B). After treatment with Nonidet NP-40, these gap junction plaques were no longer present (Fig. 4A). Instead, short double-membrane structures with the same characteristic gap as the intact junctional plaques were scattered on the support. Consistent with this, detergent-resistant membranes were not labelled with anti-MP70 immunogold, but the latter were bound to the short double-membrane structures scattered on the support (Fig. 4B). In control experiments using non-conditioned medium, immunogold complexes were only rarely detected, thus ensuring the specificity of anti-MP70 immunocytochemistry. Anti-MIP immunogold complexes did not label the junctional plaques, but consistently bound to the detergent-resistant membranes (pictures not shown). Untreated and detergent-treated membrane preparations have also been analysed by thin-section electron microscopy (pictures not shown). Samples of untreated membranes had abundant and long stretches of double membranes whereas, after treatment with Nonidet NP-40, single-membrane structures prevailed, thus confirming the results from negative-stain electron microscopy.

In summary, treatment of lens fibre plasma membranes with low concentrations of Nonidet NP-40 is effective in separating MP70 from bulk MIP. The solubilization of MP70 is paralleled by gap junction dissociation, whereas MIP is mostly contained in the insoluble membrane fraction.

Characterization of solubilized MP70

To characterize the oligomeric state of solubilized MP70, we treated membranes with 0-04% Nonidet NP-40 and subjected these mixtures to velocity gradient centrifugation. Non-dissociated membranes pelleted at the bottom of the tubes. Solubilized MP70 was detected in the gradient fractions using an anti-MP70 dot-blot assay (Fig. 5). MP70 ran reproducibly to a region near the middle of the 5% to 20% sucrose gradient (containing 0 ·1% Nonidet NP-40), and from this a particle sedimentation value of 16-9 was determined. As we did not know the amount of detergent bound, or the diffusion constant for these particles, we could only estimate the particle molecular weight by extrapolation from other known membrane protein oligomers. Using 250000 and 500000 as molecular weights for 8 6 S and 12’8 S acetylcholine receptor oligomers, respectively (Reynolds & Karlin, 1978), we determined the molecular weight for the 16 ·9 S particles to be in the range 700000-900000.

MIP was found accumulated in the insoluble membrane pellet at the bottom of the gradient tubes. Only trace amounts of soluble MIP were found in the gradient fractions and they could only be detected on silver-stained gels. Furthermore, these small amounts of soluble MIP were spread throughout the gradient and did not peak at the same position on the gradient as did MP70. Hence, MIP does not appear to interact with the MP70 oligomers. Soluble MIP in the gradient fractions could be further reduced by first centrifuging detergent-treated membrane mixtures and loading only the resulting supernatants onto the gradients (not shown).

The 16 ·9 S particles were characterized by negativestain electron microscopy. In average stained areas they appeared mostly as short double-membrane structures and, occasionally, in more heavily stained regions of the grid, they had the appearance of rosettes (Fig. 6). The width of the double-membrane structures was measured as 15 nm and hence was similar to the width of intact fibre gap junctions. The minimum length of the doublemembrane structures and the minimum diameter of the rosettes were both 11 nm. These aspects and dimensions are consistent with isolated connexon-pairs of lens fibre gap junctions.

These short-double membrane structures and rosettes are unique to the MP70 peak fractions and are absent from other gradient portions. It is most unlikely that these structures consist of aggregation products of lipid and detergent remnants.

In summary, solubilized MP70 has been isolated as structures that are consistent with connexon-pairs. In gradients that contained relatively small amounts of MIP, the latter did not peak at the same position as MP70 and hence these proteins do not appear to interact with each other under these conditions.

The solubility of lens fibre gap junctions in mild nonionic detergent solutions is in marked contrast to the considerably higher detergent resistance of the gap junctions from other organs (Goodenough & Stockenius, 1972; Benedetti & Emmelot, 1968). This difference in detergent resistance is paralleled by differences in the connexon order in the junctional membranes. Whereas the connexons in isolated liver and heart gap junctions interact with each other to form crystalline arrays, lens fibre gap junctions generally display unordered connexons, which do not appear to interact with each other laterally (Goodenough, 1979). This makes the lens fibre gap junctions unique in that junctional protein can be purified by way of solubilization with mild detergent.

It is important to note that our data on detergent solubility strictly concern the fibre gap junctions isolated from the lens outer cortex, which contain predominantly MP70/64. Gap junctions isolated from deeper lens regions contain MP38, which is the in vivo processed form of MP70 (Kistler & Bullivant, 1987), and these are considerably more resistant to detergent treatments.

Our data show that MP70 has an important role in the lens fibre gap junction structure. MP70 had previously been identified as a specific component of the lens fibre gap junctions and now has been isolated as 16 ·9 S particles. The estimated molecular weight range of 700 000-900 000 for MP70 oligomers is consistent with a pair of hexamers of MP70. Such hexameric symmetry has indeed been observed for lens fibre gap junctions that had been subjected to limited proteolysis (unpublished results).

MP70 belongs to a divergent family of gap junction proteins (Kistler et al. 1988) and may thus form transmembrane channels. The data reported here should bring us a step closer to the development of a functional assay for MP70 separately from MIP. The mild detergent conditions needed to solubilize MP70 would appear to permit reconstitution with lipid without denaturing the protein. On the other hand, for the reconstitution of MIP, protein purity could be improved by first removing MP70 from the fibre membranes with Nonidet NP-40 before solubilizing MIP with stronger detergents.

This work was supported by grants from the Medical Research Council, New Zealand.

Benedetti
,
E. L.
&
Emmelot
,
P.
(
1968
).
Hexagonal array of subunits in tight junctions separated from isolated rat liver plasma membranes
.
J. Cell Biol
.
38
,
15
24
.
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
,
2621
2629
.
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
,
1491
1499
.
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. biophys. Res. Commun
.
128
,
993
-
999
.
Goodenough
,
D. A.
(
1979
).
Lens gap junctions: a structural hypothesis for nonregulated low-resistance intercellular pathways
.
Invest. Ophthalmol, vis. Sci
.
18
,
1104
1122
.
Goodenough
,
D. A.
&
Stockenius
,
W.
(
1972
).
The isolation of mouse hepatocyte gap junctions
.
J. Cell Biol
.
54
,
646
656
.
Gorin
,
M. B.
,
Yancey
,
S. B.
,
Cline
,
J.
,
Revel
,
J. P.
&
Horwitz
,
J.
(
1984
).
The major intrinsic protein (MIP) of the bovine lens fiber membrane: characterization and structure based on cDNA cloning
.
Cell
39
,
49
59
.
Gruijters
,
W. T. M.
,
Kistler
,
J.
,
Bullivant
,
S.
&
Goodenough
,
D. A.
(
1987
).
Immunolocalization of MP70 in lens fiber 16-17nm intercellular junctions
.
J. Cell Biol
.
104
,
565
572
.
Helenius
,
A.
&
Simons
,
K.
(
1975
).
Solubilization of membranes by detergents
.
Biochim. biophys. Acta
415
,
29
79
.
Hertzberg
,
E. L.
(
1984
).
A detergent-independent procedure for the isolation of gap functions from rat liver. J
.
biol. Chem
.
259
,
9936
9943
.
Kistler
,
J.
&
Bullivant
,
S.
(
1987
).
Protein processing in lens intercellular junctions: cleavage of MP70 to MP38
.
Invest. Ophthalmol, vis Sci
.
28
,
1687
1692
.
Kistler
,
J.
,
Christie
,
D.
&
Bullivant
,
S.
(
1988
).
Homologies between gap junction proteins in lens, heart and liver
.
Sature, Land
.
331
,
721
723
.
Kistler
,
J.
,
Kirkland
,
B.
&
Bullivant
,
B.
(
1985
).
Identification of a 70, 000 dalton protein in lens membrane junctional domains
.
J. Cell Biol
.
101
,
28
35
.
Nicholson
,
B.
,
Dermietzel
,
R.
,
Teplow
,
R.
,
Traub
,
O.
,
Willecke
,
K.
&
Revel
,
J. P.
(
1987
).
Two homologous protein components of hepatic gap junctions
.
Sature, Land
.
329
,
732
734
.
Page
,
E.
&
Manjunath
,
C. K.
(
1985
).
Biochemistry and structure of cardiac gap functions: recent observations
. In
Cap Junctions
(ed.
M. V. L.
Bennett
&
D. C.
Spray
), pp.
49
56
.
New York
:
Cold Spring Harbor Laboratory
.
Peracchia
,
C.
&
Girsch
,
S. J.
(
1985
).
Is the C-terminal arm of lens gap junction channel protein the channel gate?
Biochem. biophys. Res. Commun
.
133
,
688
695
.
Revel
,
J. P.
,
Yancey
,
S. B.
,
Nicholson
,
B.
&
Hoh
,
J.
(
1986
).
Sequence diversity of gap junction proteins
. In
Junctional Complexes of Epithelial Cells
(ed.
G.
Bok
&
S.
Clark
), Ciba Fdn Svmp.
125
, pp.
108
-
127
. Chichester, UK: John Wiley and Sons Ltd.
Reynolds
,
J. A.
&
Karlin
,
A.
(
1978
).
Molecular weight in detergent solution of acetylcholine receptor from Torpedo californien
.
Biochemistry
17
,
2035
2038
.
Young
,
B. D.
(
1984
).
Measurement of sedimentation coefficients and computer simulation of rate-zonal separations
. In
Centrifugation, A Practical Approach
(ed.
D.
Rickwood
), pp.
127
-
159
. Oxford, England: IRL Press Ltd.
Zampighi
,
G. A.
,
Hall
,
J. E.
&
Kreman
,
M.
(
1985
).
Purified lens junctional protein forms channels in planar lipid films
.
Proc. natn. Acad. Sci. U.S.A
.
82
,
8468
8472
.