|Molecular Vision 2005;
Received 6 October 2004 | Accepted 24 January 2005 | Published 25 January 2005
Specific interaction between lens MIP/Aquaporin-0 and two members of the γ-crystallin family
Jianguo Fan,1 Robert N.
Fariss,2 Andrew G. Purkiss,3 Christine Slingsby,3
Aileen Sandilands,4 Roy Quinlan,4
Ana B. Chepelinsky1
1Laboratory of Molecular and Developmental Biology, 2Biological Imaging Core, and 5Section on Molecular Structure and Function, National Eye Institute, National Institutes of Health, Bethesda, MD; 3School of Crystallography, Birkbeck College, London, UK; 4School of Biological & Biomedical Sciences, University of Durham, South Road, Durham, UK
Correspondence to: Dr. Ana B. Chepelinsky, National Eye Institute, National Institutes of Health, 7 Memorial Drive, MSC 0704, Building 7, Room 105, Bethesda, MD, 20892-0704; Phone: (301) 435-7245; FAX: (301) 480-7933; email: email@example.com
Purpose: Major Intrinsic Protein (MIP)/Aquaporin 0 is required for lens transparency and is specifically expressed in lens fiber cell membranes. We have demonstrated previously that in the rat lens MIP interacts specifically with γE-crystallin, resulting in its recruitment to the plasma membrane. Our goal was to examine the interaction or lack of interaction between MIP and all members of the γ-crystallin family and to provide evidence for a physiological role these interactions may play in γ-crystallin or MIP function.
Methods: Full length MIP was expressed as untagged, enhanced green fluorescent protein (EGFP) tagged, or myc tagged proteins. Members of the γ-crystallin family were expressed as red fluorescent protein (HcRed) tagged proteins in the rabbit kidney epithelial cell line RK13. Co-localization of tagged proteins was analyzed by confocal fluorescence microscopy.
Results: Confocal fluorescence microscopy demonstrated that γE- and γF-crystallin co-localize specifically with full length MIP in mammalian cells while other γ-crystallins, including γA-, γB-, γC-, γD-, and γS-crystallin do not. As a result of this interaction, either γE- or γF-crystallin was recruited to the plasma membrane from the cytoplasm. MIP does not interact with the Elo mutant of γE-crystallin, which has been linked to a dominant cataract phenotype in mice.
Conclusions: These experiments demonstrate that MIP interacts selectively with γE- and γF-crystallin, and not with other γ-crystallins. This raises the possibility of MIP playing a structural role in the organization of γ-crystallins in rodent lens fibers and/or that γE- and γF-crystallin may have a specific role in MIP function in the rodent lens.
Lens major intrinsic protein (MIP), also known as Aquaporin 0, is the most abundant protein of the ocular lens fiber membrane. It belongs to an ancient family of membrane channel proteins [1-3]. Mutations in the MIP gene have been linked to genetic cataracts in mice and humans, suggesting an important role for MIP in maintaining lens transparency [4-9]. MIP may play a role in reducing the interfiber space, as tightly packed fibers are required for lens transparency and vacuolated fibers are observed in the mouse mutant lenses. Four mouse MIP mutations, including a point mutation at amino acid 51 (A51P, lop), replacement of the last 61 amino acids (amino acids 203 to 263) at the MIP C-terminus by a transposon sequence (Cat Fr), deletion of MIP amino acids 121 to 175 (Hfi) and deletion of amino acids 46 to 49 (CatTohm) result in autosomal dominant cataracts and MIP trapping in the endoplasmic reticulum without being inserted into the plasma membrane [4-7]. Similarly, two different point mutations in MIP identified in two human cataract families (E134G, T138R) do not integrate into the plasma membrane when expressed in Xenopus oocytes [8,9].
Water channel activity measurements of reconstituted lens membrane vesicles derived from Cat Fr or MIP null mice, showed a marked decrease in water channel activity compared with those from the wild type lens [10,11]. Severe changes in lens fiber structure with dominant phenotype are observed in MIP mutant mice, suggesting there maybe additional functions for MIP in the lens, such as maintenance of fiber structure and arrangement required for optimal focusing of the lens [11,12].
MIP forms pH and Ca2+ dependent water channels when expressed in Xenopus oocytes [13,14] and pH dependent voltage dependent channels in mammalian and insect cells . MIP voltage dependent channels in lipid bilayer vesicles are regulated by protein kinase A dependent phosphorylation . MIP may also function as an adhesion molecule and may play a role in gap junction formation [17-21]. MIP localizes to thin junctions of lens fibers, where it appears to be in a closed water pore configuration [22,23]. Both the N-terminus and the C-terminus of MIP are located in the cytoplasmic side of the plasma membrane. The MIP C-terminal polypeptide may play an important role in its physiological function as it is cleaved in cataractogenesis and aging [24-28], is serine phosphorylated , and may interact with calmodulin [30,31].
In our previous report, we demonstrated that γE-crystallin, a soluble protein that is also specifically expressed in lens fibers, is an MIP binding protein that interacts with the MIP C-terminal domain. We also demonstrated that this specific interaction results in the recruitment of γE-crystallin to the plasma membrane in mammalian cells expressing MIP . In the current report, we demonstrate that MIP interacts only with γE- and the closely related γF-crystallin, but not with other γ-crystallins. As a result of this interaction, γE- and γF-crystallins are recruited to the plasma membrane from the cytoplasm.
The γ-crystallins are lens fiber specific soluble proteins that also play a critical role in maintaining lens transparency . Various mutations in different members of the γ-crystallin family have been linked to genetic cataracts in both mice and humans [34-45]. Major mutations in γ-crystallins can result in formation of insoluble aggregates whereas point mutations generally lower solubility , causing lens opacity. Our study raises the possibility of a functional link between two distinct classes of lens proteins required for lens transparency.
Chemicals and reagents
Turbo Pfu DNA polymerase was obtained from Stratagene (La Jolla, CA). Polymerase Chain Reaction (PCR) Purification Kit and Plasmid Midi and Maxi Kits were purchased from Qiagen (Valencia, CA). All other chemicals were reagent grade and from standard commercial sources.
The RK13 (rabbit kidney) cell line was obtained from American Type Culture Collection and maintained as monolayer cultures at 37 °C in a 5% CO2/95% air incubator in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% heat inactivated fetal bovine serum (FBS).
The plasmids pEGFP-C2, pEGFP-N2, and pHcRed1-C1 were purchased from Clontech (Palo Alto, CA). All plasmid DNAs used were propagated in E. coli strain DH5α or DH10B and purified by ion exchange chromatography using Plasmid Midi Kits or Endonuclease-Free Maxi Kits from Qiagen (Valencia, CA).
Sequencing of plasmid DNA was performed using CEQ DTCS-Quick Start Kit (Beckman-Coulter, Fullerton, CA) and an automated, Beckman-Coulter CEQ 2000XL DNA Analysis System, according to the manufacturer's instructions.
Construction of expression vectors
The expression vector for untagged MIP (pCMV-MIP) was constructed as follows. The full length MIP cDNA fragment was cut from the plasmid pmyc-MIP  by restriction enzymes BamHI and EcoRI, purified on a 2% (w/v) agarose gel and cloned into the vector pCMVScript (Stratagene) at the BamHI and EcoRI sites. The resulting plasmid is designated as pCMV-MIP. The expression plasmid for enhanced green fluorescent protein (EGFP) tagged MIP fusion protein (pEGFP-MIP) was constructed as described previously . HcRed tagged γE- and γD-crystallin fusion constructs pHcRed-γE-cry and pHcRed-γD-cry were made as described previously . To construct the HcRed/γ-B-cry fusion plasmid pHcRed-γ-B-cry, the full length rat γB-crystallin cDNA was amplified, (using PCR with turbo Pfu DNA polymerase), from pSport-γB-cry clone jd05h10, obtained from a rat whole eye library (unpublished), with the following amplimers: 5'-ATA TGG ATC CAT GGG AAA GAT CAC CTT CTT C-3' as the forward primer and 5'-ATA TGT CGA CTC AGT AAA AAT CCA TGA CTC TTC-3' as the reverse primer. After digestion with BamHI and SalI restriction enzymes and purification using a PCR Purification Kit (Qiagen), the PCR product was cloned into pHcRed1-C1 (Clontech) at the BglII and Sal I sites. To construct the HcRed/γ-S-cry fusion plasmid pHcRed-γ-B-cry, the full length rat γB-crystallin cDNA was amplified, (using PCR with turbo Pfu DNA polymerase), from pSport-γS-cry clone jd03a11, obtained from a rat whole eye library (unpublished), with the following amplimers: 5'-ATA TGG ATC CAT GTA CAT CTT ACC CCA GGG-3' as the forward primer and 5'-ATA TGT CGA CTC ACT CCA CAA TGC GGC G-3' as the reverse primer. After digestion with BamHI and SalI restriction enzymes and purification using a PCR Purification Kit (Qiagen), the PCR product was cloned into pHcRed1-C1 (Clontech) at the BglII and SalI sites. To construct the HcRed/γA-cry fusion plasmid pHcRed-γA-cry, the full length rat γA-crystallin cDNA (0.5 kb) was cut from the plasmid pAD-γA cry  with BamHI and SalI restriction enzymes and purified on an 1.2% (w/v) agarose gel. This fragment was subsequently cloned into pHcRed1-C1 (Clontech) at the BglII and SalI sites to produce the plasmid pHcRed-γA-cry. To construct the HcRed/γC-cry fusion plasmid pHcRed-γC-cry, the full length rat γC-crystallin cDNA (0.5 kb) was cut from the plasmid pAD-γC cry  with BamHI and SalI restriction enzymes and purified on an 1.2% (w/v) agarose gel. This fragment was subsequently cloned into pHcRed1-C1 (Clontech) at the BglII and SalI sites to produce the plasmid pHcRed-γC-cry. To construct the HcRed/γF-cry fusion plasmid pHcRed-γF-cry, the full length rat γF-crystallin cDNA (0.5 kb) was cut from the plasmid pAD-γF cry  with BamHI and SalI restriction enzymes and purified on an 1.2% (w/v) agarose gel. This fragment was subsequently cloned into pHcRed1-C1 (Clontech) at the BglII and SalI sites to produce the plasmid pHcRed-γF-cry.
The expression vectors for EGFP tagged mouse γE-crystallin wild type (pEGFP-γE crystallin) and the Elo mutant (pEGFP-gE Elo) were constructed as follows. The γE-crystallin was cloned by RT-PCR using RNA prepared from C57Bl6 mouse lenses. Sequences for the Ampliprimers were as follows: Forward: 5'-CAT ATG GGA AAG ATC ACC TTC-3'; Reverse: 5'-GAA TTC AAT AGA AAT CCA TGA TTC-3'. The primers included convenient NdeI/EcoRI sites flanking the coding sequence of γE-crystallin to allow subcloning from the cloning and sequencing vector pGEMeasy (Promega, Madison, WI). To clone the γE-crystallin Elo mutation a new forward primer was generated and used in conjunction with the reverse primer above. The new forward primer removed base 404 and introduced a unique ClaI site for subsequent identification. Elo forward: 5'-GGC TAC TGG GTC CTC TAT CGA TGC CCA ACT A-3'. This fragment was also cloned into pGEMeasy (Promega) for sequencing. The full length γE-crystallin Elo mutation was generated by digesting the wild type γE-crystallin construct with PpumI and SphI and replacing this fragment. These constructs were then subcloned into pET23 for bacterial expression using the engineered NdeI and EcoRI sites to pET23Cryge and pET23 Crygeelo. Subcloning into the pEGFPCI vector was achieved by excising the NdeI-EcoRI fragment from the pET23 expression vector and subcloning into the BglII-EcoRI sites of the pEGFPC1 vector. The NdeI and BglII sites were blunt-ended with Klenow enzyme to maintain the reading frame of the γE-crystallin in its fusion with EGFP.
All expression vectors constructed were under the control of a CMV promoter. DNA sequences of the insert and the junction regions flanking the insert in these plasmids were verified by DNA sequencing.
Co-transfection of mammalian cells with EGFP tagged MIP and HcRed tagged γ-crystallins
RK13 cells were grown to about 30-50% confluence in 60 mm tissue culture dishes. A transfection mixture was prepared by thoroughly mixing 5 μg of each plasmid (EGFP or HcRed vectors) in 150 μl of serum free media with 60 μl of Superfect Transfection Reagent (Qiagen), incubating the resulting mixture for 20 min at room temperature followed by addition of 1 ml growth medium containing 10% (v/v) heat inactivated fetal bovine serum. The transfection mixture was then added to each RK13 culture dish from which the medium had been removed and the dish was incubated at 37 °C in a CO2 incubator for 1.5 h. Growth medium (4 ml) containing 10% heat inactivated fetal bovine serum was added to each dish which was then incubated at 37 °C in a CO2 incubator for an additional 40 h. After washing with PBS three times and complete removal of buffer from the culture dish, cells were fixed by addition of 2 ml of PBS (pH 7.4; Digene, Gaithersburg, MD) containing 4% (w/v) paraformaldehyde and incubated for 20 min at room temperature. Cells were washed 3 times with PBS and counterstain with 2 ml of Immunocytochemistry (ICC) buffer containing 1X PBS, 0.5% (w/v) bovine serum albumin, 0.2% (v/v) Tween 20 (v/v), 0.05% (w/v) sodium azide, and 1 μg/ml 4',6'-Diamidino-2-phenylindole Dihydrochloride, hydrate (DAPI; pH 7.3) for 30 min at room temperature. Cells were washed 3 times with PBS and mounted with Gel Mount (BioMeda Corp., Foster City, CA), an aqueous mounting medium containing a proprietary anti-photo bleaching reagent. Cells were kept at 4 °C before being imaged by confocal fluorescence microscopy as described below.
Immunohistochemistry of transfected mammalian cells co-expressing untagged MIP and EGFP tagged wild type and Elo mutant of γE-crystallin
RK13 cells were grown to about 30-50% confluence in 60 mm tissue culture dishes. A transfection mixture was prepared by thoroughly mixing 5 μg of each plasmid (pCMV-MIP or pmyc-MIP along with either pEGFP-γE crystallin or pEGFP-γE Elo) in 150 μl of serum free media with 60 μl of Superfect Transfection Reagent (Qiagen) and incubating the resulting mixture for 20 min at room temperature, followed by addition of 1 ml growth medium containing 10% (v/v) heat inactivated fetal bovine serum. The transfection mixture was subsequently added to each RK13 culture dish in which the medium had been removed, and the dish was incubated at 37 °C in a CO2 incubator for 3 h. Growth medium (4 ml) containing 10% (v/v) heat inactivated fetal bovine serum was added to each dish and incubation at 37 °C in a CO2 incubator continued for an additional 40 h. After washing with PBS three times and complete removal of buffer from the culture dish, cells were fixed by addition of 2 ml of PBS (pH 7.4; Digene) containing 4% (w/v) paraformaldehyde and incubated for 20 min at room temperature. Cells were washed 3 times with PBS and blocked with 2 ml ICC buffer [0.5% BSA (w/v), 0.2% (v/v) Tween-20, 0.05% (w/v) sodium azide in PBS, pH 7.4] containing 5% (v/v) normal goat serum for 1 h at room temperature. After removal of buffer, the cells were incubated with 1.5 ml ICC buffer containing 20 μg/ml rabbit anti-human MIP (Alpha Diagnostic International, San Antonio, TX) or rabbit anti-myc antibody (Cell Signaling, Beverly, MA) for 1 h at room temperature. After being washed three times with 4 ml cold ICC buffer (10 min each wash), the cells were incubated for 1 h at room temperature with 1.5 ml ICC buffer containing 2.6 μg/ml biotin-SP conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Cells were washed three times with 4 ml cold ICC buffer (10 min each wash) and incubated for 1 h at room temperature with 1.5 ICC buffer containing 2.6 μg/ml Cy3 conjugated streptavidin (Jackson ImmunoResearch Laboratories) and 1 μg/ml 4',6'-Diamidino-2-phenylindole Dihydrochloride, hydrate (DAPI). Cells were washed three times with 4 ml cold ICC buffer (10 min each wash), followed by washing two times with 5 ml of cold PBS. Cells were mounted in Gel Mount (BioMeda Corp.) and kept at 4 °C before being imaged by confocal fluorescence microscopy as described below.
Confocal fluorescence microscopy of transfected mammalian cells
The stained and mounted cells in 60 mm culture dishes were imaged using a Leica TCS SP2 Confocal Microscope with a Leica 40x HCX Plan Apo CS 0.85 NA and a Leica 40x Oil HCX Plan Apo CS 0.75-1.25 NA objective lens. EGFP fluorescence was imaged at 488 nm (excitation) and 500-540 nm (emission); HcRed fluorescence was imaged at 568 nm (excitation) and 600-650 nm (emission); Cy3 fluorescence was imaged at 568 nm (excitation) and 580-650 nm (emission) and DAPI fluorescence was imaged at of 351/364 nm (excitation) and 400-500 nm (emission). In order to minimize cross talk between channels, fluorochromes were excited sequentially rather than simultaneously, using the sequential scan mode of the confocal microscope.
Spatial quantification of fluorescence
The confocal images obtained were subjected to a procedure called "spatial quantification" analysis as described by Fan et al.  to demonstrate plasma membrane co-localization of two or more proteins. This method measures the fluorescence intensity of individual fluorescence channel (blue, red, and green) along a path that goes across the plasma membrane. Peaked fluorescence at the vicinity of the plasma membrane for both fluorescence channels is indicative of plasma membrane co-localization of these fluorescence signals.
Protein sequences of different γ-crystallin family members were compared using DNAStar DNA/Protein sequence analysis software (Madison, WI). Stereo representations comparing the C-terminal domain (III and IV motifs) of human γD-crystallin and rat γE-crystallin (Figure 1) were produced using the program PyMOL, with secondary structure assigned with STRIDE . X-ray crystal coordinates were used from the following PDB files; 1HK0 for human γD-crystallin  and 1A5D for rat γE-crystallin .
MIP interacts with γE- and γF-crystallins but not with other γ-crystallins
We have previously demonstrated that specific interaction between MIP and γE-crystallin resulted in the recruitment of γE-crystallin to the plasma membrane from the cytoplasm, as revealed by confocal fluorescence microscopy and co-immunoprecipitation analysis . Under different experimental conditions, we observed that co-localization and lack of co-localization between MIP and γE-crystallin as demonstrated by fluorescence microscopy always corresponded with interaction or lack of interaction between the two proteins as demonstrated by co-immunoprecipitation. For this reason, in the current study, we used the co-localization assay to demonstrate the interaction between MIP and all members of the γ-crystallin family (including γA-, γB-, γC-, γD-, γE-, γF-, γD-, and γS-crystallin) by confocal fluorescence microscopy of RK13 rabbit kidney epithelial cells co-transfected with tagged MIP and γ-crystallins. Figure 2 shows the mammalian expression vectors for EGFP tagged MIP and HcRed tagged γ-crystallin family members we constructed for this purpose. The co-transfected cells were subjected to confocal fluorescence microscopy and co-localization of MIP with γ-crystallins and the ability of MIP to recruit the γ-crystallins to the plasma membrane were subsequently examined.
We first confirmed the interaction between MIP and γE-crystallin already observed in our previous study . RK13 cells were co-transfected with expression vectors for EGFP tagged MIP (pEGFP-MIP) and HcRed tagged γE crystallin (pHcRed-γE cry; Figure 3). The EGFP fluorescence (green) was concentrated to the plasma membrane, although cytoplasmic expression was also evident. The HcRed fluorescence was also concentrated on the plasma membrane along with cytoplasmic and nuclear expression (Figure 3). Membrane co-localization was confirmed by spatial quantification analysis, which showed that the EGFP and HcRed fluorescence both peaked at the vicinity of the plasma membrane (see the blue dotted line in Figure 3) when each channel of fluorescence was measured along a path crossing the plasma membrane in the merged image. Thus, these results indicated that γE-crystallin co-localized with MIP at the plasma membrane. In the absence of MIP expression, γE-crystallin showed a diffuse expression pattern throughout the cytoplasm and no membrane localization was evident (see Figure 4J; also see Figure 4M from Fan et al. ), indicating that γE-crystallin was recruited to the plasma membrane as a result of MIP expression.
We next examined the interaction between MIP and γF-crystallin, a member of the γ-crystallin family that is the most closely related to γE-crystallin (98% identity in its protein sequence). RK13 cells were co-transfected with pEGFP-MIP and pHcRed-γF cry. Both EGFP and HcRed fluorescence showed plasma membrane localization (Figure 3). Plasma membrane co-localization was confirmed by spatial quantification along a path that crosses the plasma membrane in the merged image (Figure 3). These results indicated that MIP co-localized with γF-crystallin, another member of the γ-crystallin family, at the plasma membrane.
Subsequently we examined the interaction between MIP and the other members of the γ-crystallin family; γA-, γB-, γC-, γD-, and γS-crystallin. When RK13 cells were co-transfected with pEGFP-MIP and pHcRed-γA crystallin, the EGFP fluorescence (Figure 3) was localized to the plasma membrane and cytoplasmic space as well. However, the HcRed fluorescence was distributed throughout the cytoplasm and the nucleus (i.e., there was no apparent plasma membrane localization of the HcRed fluorescence). Thus, MIP and γA-crystallin did not co-localize to the plasma membrane, which was confirmed by spatial quantification analysis (Figure 3). Therefore, these data indicated that MIP does not interact with γA-crystallin in RK13 cells. Very similar results were obtained when cells were co-transfected with pEGFP-MIP together with γB-, γC-, γD-, or γS-crystallin expression vectors (Figure 3), indicating that MIP does not interact with γB-, γC-, γD-, or γS-crystallins.
In summary, MIP interacts only with γE- and γF-crystallins but not with other γ-crystallin family members (including γA-, γB-, γC-, γD-, or γS-crystallin). As a result of this interaction, γE- and γF-crystallins but not other γ-crystallins were recruited to the plasma membrane from the cytoplasm by MIP.
Elo mutation in γE-crystallin prevents the recruitment of γE-crystallin to the plasma membrane by MIP
Several inherited cataracts in mice are linked to mutations of γ-crystallins [34,36,37,39,41,42,45,49]. In particular the Elo mutant of γE-crystallin causes severe malformation of the developing mouse lens [50,51]. The Elo mutation disrupts the normal localization of γE-crystallin to the cytoplasm of the lens fibers and localizes instead in the cell nucleus . To determine if the mutant protein is capable of interaction with MIP, RK13 cells were co-transfected with pCMV-MIP (expressing untagged MIP, Figure 2A) and either pEGFP-gE crystallin (expressing EGFP tagged γE-crystallin, Figure 2J) or pEGFP-Elo mutant (expressing EGFP tagged γE Elo mutant, Figure 2K). The untagged MIP was visualized using an anti-MIP antibody, biotinylated secondary antibody and Cy3 conjugated streptavidin in conjunction with confocal fluorescence microscopy.
MIP, as shown by the red Cy3 fluorescence, was mostly concentrated on the plasma membrane (Figure 4A) with minimal cytoplasmic expression. With pEGFP-γE, the green EGFP fluorescence (i.e., γE-crystallin expression) localized to the plasma membrane although cytoplasmic and nuclear expression was also evident (Figure 3). The appearance of yellow color along the plasma membrane in the merged image (Figure 4C) with the spatial quantification analysis (Figure 4D) confirmed that MIP co-localized with wild type γE-crystallin on the plasma membrane and recruited γE-crystallin to the plasma membrane from the cytoplasm. In contrast, in co-transfection of pCMV-MIP (expressing untagged MIP) and pEGFP-Elo mutant (expressing EGFP tagged γE Elo mutant, Figure 2K) γE-crystallin Elo mutant protein showed only cytoplasmic and nuclear localization (Figure 4G). In the merged image (Figure 4G) the Cy3 and EGFP fluorescence were completely separated from each other; no co-localization was evident, which was confirmed by spatial quantification analysis (Figure 4H). In the absence of MIP expression (Figure 4I-P), both EGFP tagged wild type γE-crystallin (Figure 4J) and the γE-crystallin Elo mutant (Figure 4N) showed only cytoplasmic and nuclear localization, without any plasma membrane localization. These results indicate that MIP recruits γE-crystallin but not the γE-crystallin Elo mutant to the plasma membrane from the cytoplasm.
Differences between γE-/γF-crystallins and other members of the γ-crystallin family
To search for candidate regions of γE-/γF-crystallin that are responsible for the specific interaction with MIP, the sequences and structures of the mammalian γ-crystallins were compared. The γ-crystallin family arose by successive gene duplications and each protein polypeptide is built up from four similar modified Greek key motifs [33,52,53]. Primary structure alignment showed that γ-crystallins are most divergent in the third Greek key structural motif of the polypeptide  (indicated as bracket III in Figure 5).
γE- and γF-Crystallins are the most evolutionarily conserved in amino acid sequence and X-ray structure  and are the only two γ-crystallins we have found that interact with MIP. Sequence alignment shows that γE- and γF-crystallin amino acids in the loop between β-strands c and d of motif III are identical and different from all the other γ-crystallins  (see also Figure 5). This difference in sequence is also reflected in the three dimensional structure. Figure 1 shows the 3D comparison of the C-terminal domains of human γD-crystallin  and rat γE-crystallin . Although the overall domain structure is well conserved and the β-sheets of both proteins superimpose, there are obvious structural differences in the c-d loops. In particular the two c-d loops are much more widely separated in γE-crystallin than in the otherwise very similar γD-crystallin. This is apparent in the ribbon diagrams (Figure 1A). In space filling representations of the same region this has the effect of producing a groove or indentation in γE-crystallin that has the potential to form a binding site (Figure 1B).
MIP interacts with γE- and γF-crystallins but not with other γ-crystallins
The γA-F-crystallins in mammals are closely related proteins, expressed from a cluster of genes that has arisen through successive duplications. Although the six proteins are very similar in structure, each has characteristic sequence differences that are conserved among mammalian species. Presumably, this reflects specific evolutionarily selected functions for each family member.
We have previously demonstrated that γE-crystallin can interact with the lens major integral membrane protein MIP, resulting in the recruitment of γE-crystallin to the plasma membrane . Here we extend this observation to the other members of the γ-crystallin family in rat. No interaction occurs between MIP and γA-, γB-, γC-, or γD-crystallins or the more distantly related γS-crystallin. However γF-crystallin, which is most similar in sequence to γE-crystallin, does interact with MIP.
The specificity of the interaction between MIP and γE-/γF-crystallin interaction suggests that these two crystallins have a common binding site for MIP not shared with other members of their family. Indeed, the six members of the group provide a set of natural mutants to map the residues required for MIP binding. In fact, as shown in Figure 5, γE- and γF-crystallin have a unique region of sequence identity, located in their C-terminal domains.
γ-Crystallins (and also the related β-crystallins) contain four repeated structural motifs (I-IV), each consisting of four β-strands (a-d). These repeated motifs are arranged in pairs to form two domains (I/II and III/IV). Throughout the βγ-crystallin superfamily, the most variable in sequence of the four motifs is motif III in the C-terminal domain. As shown in the alignment, the sequence between strands c and d of motif III is well conserved between γE- and γF-crystallin but quite different in the other γ-crystallins. This is reflected in differences in the X-ray structures of human γD and rat γE-crystallins, as shown in Figure 1. The wide separation of the c-d loops that run across the domain to link the two β-sheets in γE-crystallin creates a noticeable groove in the molecular surface that is not present in γD-crystallin. This is an intriguing candidate for a binding site and might indeed be involved in the interaction with MIP, although this will need further investigation.
There are other reasons to suspect that the MIP interaction site is in the C-terminal domains of γE- and γF-crystallin. In the experiments described here, the γ-crystallins are all tagged at the N-terminus by reporter proteins (HcRed or EGFP) comparable in size to the crystallin itself. This is likely to block possible interactions involving some regions of the N-terminal domain, but is less likely to interfere directly with potential binding sites on the C-terminal domain. In the Elo mutation, γE-crystallin is truncated, losing motif IV and thereby disrupting the whole of the C-terminal domain, although the N-terminal might be able to fold correctly. The fusion protein of the Elo mutant of γE-crystallin that lacks a folded C-terminal domain, does not interact with MIP. This could be due to loss of the proposed binding groove in the C-terminal domain, although it could also be due to decreased solubility of the truncated mutant protein.
Functional link between MIP and γE-crystallin
Both MIP and γ-crystallins are specifically expressed in the lens fibers. Five mutations resulting in genetic cataracts with a dominant phenotype have been identified in the murine γE-crystallin gene [34,36,39,55] and four in MIP [4-7]. Interestingly, both MIP and γE-crystallin associated cataracts have some similarities. When either MIP or γE-crystallin is mutated, the tight packing of lens fiber cells is disrupted and vacuolated cells are observed [4-7,36,39,55]. In addition, either in the absence of MIP  or in the presence of mutants of γE-crystallin [39,55] suture formation is perturbed in the mouse lens. There is also some correlation between γE-crystallin and MIP cataracts in time of appearance. In the Elo mouse, in which γE-crystallin is mutated, the cataract phenotype appears at embryonic day 12-13 [50,51,56], the same embryonic stage at which MIP expression begins in lens primary fibers  and at which the MIP Cat Fraser mouse mutant shows the cataract phenotype .
These observations suggest a possible functional link between γE-crystallin and MIP in the maintenance of lens transparency. MIP has at least two potential roles in the lens. It can serve as a water channel, with consequences for lens hydration, and it can also act as a lens fiber adhesion protein [10,14,22]. There is evidence that the C-terminal peptide of MIP has a role in both these functions, affecting the constriction of the water channel and the ability of MIP subunits to form intercellular junctions [22,28]. Thus, an interaction between the MIP C-terminal peptide and either γE- or γF-crystallin could have major consequences for lens organization and physiology.
Evolutionary implications of γE-/γF-crystallin interaction with MIP and lens hydration
Although both γE- and γF-crystallin are expressed in mice and rats, the CRYGE and CRYGF genes in humans are non-functional pseudogenes [59-63]. Clearly, if there is a functional role for the MIP interaction with γE-/γF-crystallin in the rodent lens, it is absent from humans and must relate to a change in the human lens. In fact there are several ways in which the human lens differs from those of rodents. Most obvious is the "softness" or hydration of the human lens. Rodent lenses have high concentrations of γ-crystallins and may achieve protein concentrations of up to 60% wet weight. Their lenses are quite rigid and spherical with a short focal length and limited accommodative ability. In contrast human lenses have lower γ-crystallin content, lower overall protein concentration, and are much more flexible and accommodative. Furthermore, the organization of fiber cells and sutures is also significantly different in rodents and primates, with functional consequences for their focusing ability . Thus, it is conceivable that the loss of γE- and γF-crystallin from the human lens reflects an evolutionary change in the regulation of organization and function of MIP, which in turn has major effects on the overall properties of the mammalian lens. Indeed, bird lenses lack the whole group of γA- to γF-crystallins and also have softer, accommodating lenses and a different organization of fiber cell packing from that seen in most mammals.
1. Wistow GJ, Pisano MM, Chepelinsky AB. Tandem sequence repeats in transmembrane channel proteins. Trends Biochem Sci 1991; 16:170-1.
2. Reizer J, Reizer A, Saier MH Jr. The MIP family of integral membrane channel proteins: sequence comparisons, evolutionary relationships, reconstructed pathway of evolution, and proposed functional differentiation of the two repeated halves of the proteins. Crit Rev Biochem Mol Biol 1993; 28:235-57.
3. Agre P, Sasaki S, Chrispeels MJ. Aquaporins: a family of water channel proteins. Am J Physiol 1993; 265:F461.
4. Shiels A, Bassnett S. Mutations in the founder of the MIP gene family underlie cataract development in the mouse. Nat Genet 1996; 12:212-5.
5. Shiels A, Mackay D, Bassnett S, Al-Ghoul K, Kuszak J. Disruption of lens fiber cell architecture in mice expressing a chimeric AQP0-LTR protein. FASEB J 2000; 14:2207-12.
6. Sidjanin DJ, Parker-Wilson DM, Neuhauser-Klaus A, Pretsch W, Favor J, Deen PM, Ohtaka-Maruyama C, Lu Y, Bragin A, Skach WR, Chepelinsky AB, Grimes PA, Stambolian DE. A 76-bp deletion in the Mip gene causes autosomal dominant cataract in Hfi mice. Genomics 2001; 74:313-9.
7. Okamura T, Miyoshi I, Takahashi K, Mototani Y, Ishigaki S, Kon Y, Kasai N. Bilateral congenital cataracts result from a gain-of-function mutation in the gene for aquaporin-0 in mice. Genomics 2003; 81:361-8.
8. Berry V, Francis P, Kaushal S, Moore A, Bhattacharya S. Missense mutations in MIP underlie autosomal dominant 'polymorphic' and lamellar cataracts linked to 12q. Nat Genet 2000; 25:15-7.
9. Francis P, Chung JJ, Yasui M, Berry V, Moore A, Wyatt MK, Wistow G, Bhattacharya SS, Agre P. Functional impairment of lens aquaporin in two families with dominantly inherited cataracts. Hum Mol Genet 2000; 9:2329-34.
10. Varadaraj K, Kushmerick C, Baldo GJ, Bassnett S, Shiels A, Mathias RT. The role of MIP in lens fiber cell membrane transport. J Membr Biol 1999; 170:191-203.
11. Shiels A, Bassnett S, Varadaraj K, Mathias R, Al-Ghoul K, Kuszak J, Donoviel D, Lilleberg S, Friedrich G, Zambrowicz B. Optical dysfunction of the crystalline lens in aquaporin-0-deficient mice. Physiol Genomics 2001; 7:179-86.
12. Al-Ghoul KJ, Kirk T, Kuszak AJ, Zoltoski RK, Shiels A, Kuszak JR. Lens structure in MIP-deficient mice. Anat Rec A Discov Mol Cell Evol Biol 2003; 273:714-30.
13. Nemeth-Cahalan KL, Hall JE. pH and calcium regulate the water permeability of aquaporin 0. J Biol Chem 2000; 275:6777-82.
14. Nemeth-Cahalan KL, Kalman K, Hall JE. Molecular basis of pH and Ca2+ regulation of aquaporin water permeability. J Gen Physiol 2004; 123:573-80.
15. Drake KD, Schuette D, Chepelinsky AB, Jacob TJ, Crabbe MJ, Jacob TJ. pH-Dependent channel activity of heterologously-expressed main intrinsic protein (MIP) from rat lens. FEBS Lett 2002; 512:199-204. Erratum in: FEBS Lett 2002; 516(1-3):287.
16. Ehring GR, Lagos N, Zampighi GA, Hall JE. Phosphorylation modulates the voltage dependence of channels reconstituted from the major intrinsic protein of lens fiber membranes. J Membr Biol 1992; 126:75-88.
17. Michea LF, de la Fuente M, Lagos N. Lens major intrinsic protein (MIP) promotes adhesion when reconstituted into large unilamellar liposomes. Biochemistry 1994; 33:7663-9.
18. Gruijters WT. A non-connexon protein (MIP) is involved in eye lens gap-junction formation. J Cell Sci 1989; 93:509-13.
19. Dunia I, Recouvreur M, Nicolas P, Kumar N, Bloemendal H, Benedetti EL. Assembly of connexins and MP26 in lens fiber plasma membranes studied by SDS-fracture immunolabeling. J Cell Sci 1998; 111:2109-20.
20. Fotiadis D, Hasler L, Muller DJ, Stahlberg H, Kistler J, Engel A. Surface tongue-and-groove contours on lens MIP facilitate cell-to-cell adherence. J Mol Biol 2000; 300:779-89.
21. Yu XS, Jiang JX. Interaction of major intrinsic protein (aquaporin-0) with fiber connexins in lens development. J Cell Sci 2004; 117:871-80.
22. Gonen T, Sliz P, Kistler J, Cheng Y, Walz T. Aquaporin-0 membrane junctions reveal the structure of a closed water pore. Nature 2004; 429:193-7.
23. Gonen T, Cheng Y, Kistler J, Walz T. Aquaporin-0 membrane junctions form upon proteolytic cleavage. J Mol Biol 2004; 342:1337-45.
24. Takemoto L, Takehana M, Horwitz J. Covalent changes in MIP26K during aging of the human lens membrane. Invest Ophthalmol Vis Sci 1986; 27:443-6.
25. Takemoto L, Kuck J, Kuck K. Changes in the major intrinsic polypeptide (MIP26K) during opacification of the Emory mouse lens. Exp Eye Res 1988; 47:329-36.
26. Schey KL, Fowler JG, Shearer TR, David L. Modifications to rat lens major intrinsic protein in selenite-induced cataract. Invest Ophthalmol Vis Sci 1999; 40:657-67.
27. Boyle DL, Takemoto LJ. Localization of MIP 26 in nuclear fiber cells from aged normal and age-related nuclear cataractous human lenses. Exp Eye Res 1999; 68:41-9.
28. Ball LE, Little M, Nowak MW, Garland DL, Crouch RK, Schey KL. Water permeability of C-terminally truncated aquaporin 0 (AQP0 1-243) observed in the aging human lens. Invest Ophthalmol Vis Sci 2003; 44:4820-8.
29. Schey KL, Little M, Fowler JG, Crouch RK. Characterization of human lens major intrinsic protein structure. Invest Ophthalmol Vis Sci 2000; 41:175-82.
30. Girsch SJ, Peracchia C. Calmodulin interacts with a C-terminus peptide from the lens membrane protein MIP26. Curr Eye Res 1991; 10:839-49.
31. Swamy-Mruthinti S. Glycation decreases calmodulin binding to lens transmembrane protein, MIP. Biochim Biophys Acta 2001; 1536:64-72.
32. Fan J, Donovan AK, Ledee DR, Zelenka PS, Fariss RN, Chepelinsky AB. gammaE-crystallin recruitment to the plasma membrane by specific interaction between lens MIP/aquaporin-0 and gammaE-crystallin. Invest Ophthalmol Vis Sci 2004; 45:863-71.
33. Wistow GJ, Piatigorsky J. Lens crystallins: the evolution and expression of proteins for a highly specialized tissue. Annu Rev Biochem 1988; 57:479-504.
34. Klopp N, Favor J, Loster J, Lutz RB, Neuhauser-Klaus A, Prescott A, Pretsch W, Quinlan RA, Sandilands A, Vrensen GF, Graw J. Three murine cataract mutants (Cat2) are defective in different gamma-crystallin genes. Genomics 1998; 52:152-8.
35. Stephan DA, Gillanders E, Vanderveen D, Freas-Lutz D, Wistow G, Baxevanis AD, Robbins CM, VanAuken A, Quesenberry MI, Bailey-Wilson J, Juo SH, Trent JM, Smith L, Brownstein MJ. Progressive juvenile-onset punctate cataracts caused by mutation of the gammaD-crystallin gene. Proc Natl Acad Sci U S A 1999; 96:1008-12.
36. Klopp N, Loster J, Graw J. Characterization of a 1-bp deletion in the gammaE-crystallin gene leading to a nuclear and zonular cataract in the mouse. Invest Ophthalmol Vis Sci 2001; 42:183-7.
37. Graw J, Neuhauser-Klaus A, Loster J, Favor J. A 6-bp deletion in the Crygc gene leading to a nuclear and radial cataract in the mouse. Invest Ophthalmol Vis Sci 2002; 43:236-40.
38. Santhiya ST, Shyam Manohar M, Rawlley D, Vijayalakshmi P, Namperumalsamy P, Gopinath PM, Loster J, Graw J. Novel mutations in the gamma-crystallin genes cause autosomal dominant congenital cataracts. J Med Genet 2002; 39:352-8.
39. Graw J. Cataract mutations and lens development. Prog Retin Eye Res 1999; 18:235-67.
40. Nandrot E, Slingsby C, Basak A, Cherif-Chefchaouni M, Benazzouz B, Hajaji Y, Boutayeb S, Gribouval O, Arbogast L, Berraho A, Abitbol M, Hilal L. Gamma-D crystallin gene (CRYGD) mutation causes autosomal dominant congenital cerulean cataracts. J Med Genet 2003; 40:262-7.
41. Smith RS, Hawes NL, Chang B, Roderick TH, Akeson EC, Heckenlively JR, Gong X, Wang X, Davisson MT. Lop12, a mutation in mouse Crygd causing lens opacity similar to human Coppock cataract. Genomics 2000; 63:314-20.
42. Graw J, Klopp N, Neuhauser-Klaus A, Favor J, Loster J. Crygf(Rop): the first mutation in the Crygf gene causing a unique radial lens opacity. Invest Ophthalmol Vis Sci 2002; 43:2998-3002.
43. Basak A, Bateman O, Slingsby C, Pande A, Asherie N, Ogun O, Benedek GB, Pande J. High-resolution X-ray crystal structures of human gammaD crystallin (1.25 A) and the R58H mutant (1.15 A) associated with aculeiform cataract. J Mol Biol 2003; 328:1137-47.
44. Kmoch S, Brynda J, Asfaw B, Bezouska K, Novak P, Rezacova P, Ondrova L, Filipec M, Sedlacek J, Elleder M. Link between a novel human gammaD-crystallin allele and a unique cataract phenotype explained by protein crystallography. Hum Mol Genet 2000; 9:1779-86.
45. Sinha D, Wyatt MK, Sarra R, Jaworski C, Slingsby C, Thaung C, Pannell L, Robison WG, Favor J, Lyon M, Wistow G. A temperature-sensitive mutation of Crygs in the murine Opj cataract. J Biol Chem 2001; 276:9308-15.
46. Pande A, Pande J, Asherie N, Lomakin A, Ogun O, King J, Benedek GB. Crystal cataracts: human genetic cataract caused by protein crystallization. Proc Natl Acad Sci U S A 2001; 98:6116-20.
47. Frishman D, Argos P. Knowledge-based protein secondary structure assignment. Proteins 1995; 23:566-79.
48. Norledge BV, Hay RE, Bateman OA, Slingsby C, Driessen HP. Towards a molecular understanding of phase separation in the lens: a comparison of the X-ray structures of two high Tc gamma-crystallins, gammaE and gammaF, with two low Tc gamma-crystallins, gammaB and gammaD. Exp Eye Res 1997; 65:609-30.
49. Sandilands A, Hutcheson AM, Long HA, Prescott AR, Vrensen G, Loster J, Klopp N, Lutz RB, Graw J, Masaki S, Dobson CM, MacPhee CE, Quinlan RA. Altered aggregation properties of mutant gamma-crystallins cause inherited cataract. EMBO J 2002; 21:6005-14.
50. Cartier M, Breitman ML, Tsui LC. A frameshift mutation in the gamma E-crystallin gene of the Elo mouse. Nat Genet 1992; 2:42-5. Erratum in: Nat Genet 1992; 2:343.
51. Oda S, Watanabe K, Fujisawa H, Kameyama Y. Impaired development of lens fibers in genetic microphthalmia, eye lens obsolescence, Elo, of the mouse. Exp Eye Res 1980; 31:673-81.
52. Breitman ML, Lok S, Wistow G, Piatigorsky J, Treton JA, Gold RJ, Tsui LC. Gamma-crystallin family of the mouse lens: structural and evolutionary relationships. Proc Natl Acad Sci U S A 1984; 81:7762-6.
53. Wistow G. Lens crystallins: gene recruitment and evolutionary dynamism. Trends Biochem Sci 1993; 18:301-6.
54. Wistow G, Turnell B, Summers L, Slingsby C, Moss D, Miller L, Lindley P, Blundell T. X-ray analysis of the eye lens protein gamma-II crystallin at 1.9 A resolution. J Mol Biol 1983; 170:175-202.
55. Graw J, Neuhauser-Klaus A, Loster J, Klopp N, Favor J. Ethylnitrosourea-induced base pair substitution affects splicing of the mouse gammaE-crystallin encoding gene leading to the expression of a hybrid protein and to a cataract. Genetics 2002; 161:1633-40.
56. Yoshiki A, Hanazono M, Oda S, Wakasugi N, Sakakura T, Kusakabe M. Developmental analysis of the eye lens obsolescence (Elo) gene in the mouse: cell proliferation and Elo gene expression in the aggregation chimera. Development 1991; 113:1293-304.
57. Yancey SB, Koh K, Chung J, Revel JP. Expression of the gene for main intrinsic polypeptide (MIP): separate spatial distributions of MIP and beta-crystallin gene transcripts in rat lens development. J Cell Biol 1988; 106:705-14.
58. Hamai Y, Kuwabara T. Early cytologic changes of Fraser cataract. An electron microscopic study. Invest Ophthalmol 1975; 14:517-27.
59. Aarts HJ, den Dunnen JT, Leunissen J, Lubsen NH, Schoenmakers JG. The gamma-crystallin gene families: sequence and evolutionary patterns. J Mol Evol 1988; 27:163-72.
60. Heon E, Priston M, Schorderet DF, Billingsley GD, Girard PO, Lubsen N, Munier FL. The gamma-crystallins and human cataracts: a puzzle made clearer. Am J Hum Genet 1999; 65:1261-7.
61. Meakin SO, Du RP, Tsui LC, Breitman ML. Gamma-crystallins of the human eye lens: expression analysis of five members of the gene family. Mol Cell Biol 1987; 7:2671-9.
62. Brakenhoff RH, Aarts HJ, Reek FH, Lubsen NH, Schoenmakers JG. Human gamma-crystallin genes. A gene family on its way to extinction. J Mol Biol 1990; 216:519-32.
63. Siezen RJ, Thomson JA, Kaplan ED, Benedek GB. Human lens gamma-crystallins: isolation, identification, and characterization of the expressed gene products. Proc Natl Acad Sci U S A 1987; 84:6088-92.
64. Kuszak JR, Zoltoski RK, Sivertson C. Fibre cell organization in crystalline lenses. Exp Eye Res 2004; 78:673-87.