Amino Acid Residue Val362 Plays a Critical Role in Maintaining the Structure of C Terminus of Connexin 50 and in Lens Epithelial-fiber Differentiation*

We have previously shown that connexin (Cx) 50, unlike the other two lens connexins, Cx43 and Cx46, promotes chicken lens epithelial-fiber differentiation in a channel-independent manner. Here, we show that deletion of the PEST motif at the C terminus (CT) domain of Cx50 attenuates the stimulatory effect of Cx50 on lens fiber differentiation. Valine 362, a residue located within the PEST domain, is functionally involved. The structure of the Cx50 CT predicted by molecular modeling revealed four α-helices and Val362 was found to be located in the middle of the 3rd helix. Replacement of Val362 with amino acid residues that disrupt the α-helical structure predicted by molecular modeling, such as arginine, glutamate, or phenylalanine, attenuated the stimulatory effects of Cx50 on lens differentiation, whereas replacement with threonine, isoleucine, leucine, or proline, which maintain the structure preserved the function of Cx50. Circular dichroism (CD) studies supported the structural predictions and showed that the substitution with Glu, but not Thr or Pro, disrupted the α-helix, which appears to be the structural feature important for lens epithelial-fiber differentiation. Together, our results suggest that Val362 is important for maintaining the helical structure and is crucial for the role of Cx50 in promoting lens epithelial-fiber differentiation.

The vertebrate eye lens is an avascular organ consisting of an anterior epithelial cell layer and highly differentiated fiber cells in the center. Mitotically active epithelial cells at the lens equator differentiate into the lens fiber cells, which lose major intracellular organelles and accumulate high concentrations of soluble crystallins (1). The loss of cellular organelles makes the fiber cells become incapable of supporting an active cellular metabolism. Hence, the fiber cells depend upon extensive networks of gap junction communications and transport system to direct microcirculation to support their metabolic needs and homeostasis (2).
Gap junctions are membrane channels that connect neighboring cells and allow the passage of small molecules (M r Ͻ 1,000), such as small metabolites, ions, and second messengers. This type of cell-cell coupling is essential for normal cell and tissue functions. The structural components of gap junctions are membrane proteins called connexins. Connexins have four conserved transmembrane domains with the most variable cytoplasmic C terminus (CT) 2 domain (3). Three types of connexins are expressed in the lens organ, among which Cx43 is predominately expressed in lens epithelial cells. During the transition process into fibers, the expression level of Cx43 is decreased and eventually replaced by two fiber connexins, Cx50 and Cx46. Each connexin forms channels with distinctive physiological properties of gating, permeation, and selective interaction with other connexins (4).
Intercellular communication through gap junction has been demonstrated to regulate vertebrate development, where "communication compartments" were used to describe the coupled cells sharing the same fate (5). In the lens, the importance of gap junctions and connexins was demonstrated by knock-out mouse models. Both Cx50-and Cx46-deficient mice develop cataracts. In addition, ablation of Cx50 results in decreased lens size, leading to microphthalmia (6 -8). The retardation in lens growth caused by deficiency of Cx50 is likely due to the attenuation of lens epithelial-fiber differentiation (9). Concurring with in vivo observations, we have previously shown that overexpression of Cx50, but not other lens connexins, significantly stimulates lens epithelial-fiber differentiation in lens primary cell culture (10). Moreover, the CT domain of Cx50 is sufficient to promote lens fiber differentiation, suggesting the importance of this region (11).
In this study, we identified a critical amino acid residue, Val 362 within the CT domain of Cx50 that is functionally involved in lens epithelial-fiber differentiation. Furthermore, we showed that this amino acid residue is an important component in maintaining an ␣-helical structure, and this structural feature appears to be indispensable for the role of Cx50 in lens epithelial-fiber differentiation.

Preparation of Recombinant Retroviral Constructs Encoding Cx50 and Cx50 Mutants and Generation of High-titer
Retroviruses-Retroviral constructs and high-titer retroviruses were prepared based on the protocol described previously (13,14). Briefly, a cDNA fragment containing chicken Cx50 was generated by PCR and constructed into the retroviral vector RCAS(A). Cx50 single or multiple site mutants were generated with the QuikChange site-directed mutagenesis kit according to the manufacturer's instructions with the primers shown in Table 1. PCR primers were synthesized and constructs were sequenced at the University of Texas Health Science Center at San Antonio DNA Core Facility.
Primary Chicken Lens Cell Culture-Primary chicken lens cell cultures were prepared as described (11). Briefly, lenses from 10 -11-day-old chicken embryos were dissected, rinsed with TD buffer (140 mM NaCl, 5 mM KCl, 0.7 mM Na 2 HPO 4 , 5 mM glucose, and 25 mM Tris (pH 7.4)), and digested with 0.1% trypsin at 37°C for 30 min. Lens cells were obtained by breaking the lens apart and collecting them in M199 medium with 10% fetal bovine serum. Living cells were then counted and seeded at 3 ϫ 10 5 cells per well of a 12-well culture plate. The next day after the primary culture was seeded, retroviruses expressing wild-type and mutated forms of Cx50 were added to primary lens cultures. The cultures were incubated at 37°C, 5% CO 2 and fed every other day. At the start of culturing, only monolayer lens epithelial cells proliferated on the culture plates. After 3-4 days, lens epithelial cells reached confluency and differentiated to form fiber-like "lentoid" structures. The number of lentoid was counted everyday during the culturing.
Immunofluorescence and Confocal Laser Microscopy-For immunolabeling of AQP0 expressed in the primary lens culture, a glass coverslip was placed into each well of a 12-well plate before seeding the cell. After 6 -8 days of culturing, cells were fixed with 2% paraformaldehyde for 30 min and then incubated with blocking solution (2% goat serum, 2% fish skin gelatin, 0.25% Triton X-100, and 1% bovine serum albumin in Hanks' balanced salt solution) for another 30 min. AQP0 was labeled with anti-AQP0 polyclonal antibody (1:200) followed by rhodamine-conjugated anti-rabbit IgG (1:400). The sections were mounted on a glass coverslip using Vectashield. The specimens were analyzed under a confocal laser scanning microscope (Fluoview; Olympus Optical, Tokyo, Japan). Image acquisition parameters were maintained consistent for each sample. Representative fluorescent images from random regions within each culture well were taken and the fluorescence signals of the AQP0 staining area versus the whole image area were determined using NIH ImageJ software (10). Ten to fifteen images for each sample were used to examine the AQP0 expression per measurement.
Sequence Analysis and Molecular Modeling of Cx50 CT-Multiple rounds of Psi-BLAST search in the NCBI data base were performed to identify the possible sequence homology. To construct a structural model for Cx50 CT based on human cyclin A 2 (PDB code 3bht) as a template, the SWISS-MODEL online server (37) was used. The sequence alignment between Cx50 CT and a part of cyclin A (Ser 171 -Val 430 ) was performed by the T-COFFEE program (38) which was then subjected to alignment mode modeling in SWISS-MODEL. Energy minimization of this Cx50 CT model was performed with GROMOS96 in DeepView software (39). No clashes within the individual subunits or at the subunit interfaces have been observed before or after energy minimization. To further confirm the structural information from the molecular modeling methods, other secondary prediction methods including SOPM (self-optimized method) were utilized to predict the Cx50 CT secondary structure solely based on the composition of amino acids (15). In addition, Profile Hidden Markov Models (HHpred program (16)) were also employed to predict the possible secondary structure of Cx50 CT by scoring pairs of aligned secondary structure states in a way analogous to the classical amino acid substitution matrices.
Scrape Loading Dye Transfer Assay and Fluorescence Microscopy-Chicken embryonic fibroblast cells were grown to confluence to maximize cell-cell contact. Scrape loading dye transfer was performed based on a modified protocol (17). Briefly, cells were scratched in the presence of two fluorescent dyes: LY (457 Da), which can pass through gap junction channels, and RD (10 kDa), which is too large to pass through gap junction channels. The presence of LY indicates those cells that participate in gap junction-mediated communication and RD serves as a tracer dye for cells originally receiving the dye. Cells were washed three times with Hanks' balanced salt solution plus 1% bovine serum albumin for 5 min each and then a mix- GAAAGCAGAGGAGGAGTTCGTGAGCGATGAAGTGG CCACTTCATCGCTCACGAACTCCTCCTCTGCTTTC ture solution containing 1% LY, 1% RD in phosphate-buffered saline was applied and the plates were scraped lightly with a surgical blade. After 15 min of incubation with the dyes, the cells were washed with Hanks' balanced salt solution three times, washed twice with phosphate-buffered saline, and then fixed in fresh 2% paraformaldehyde for 30 min. Dye transfer results were examined using a fluorescence microscope (Zeiss), where LY and RD could be detected by using fluorescein and RD filters, respectively. Image acquisition was kept consistent for all measurements and no threshold adjustments were used. Using RD staining as the reference for the original dye-loaded cells, the extent of dye transfer was measured as the distance from the center of the scrape line to the farthest extent of RDor LY-stained cells with the appropriate scale bar. At least five images per condition tested with six measurements per image were used to assess the degree of dye transfer.
Circular Dichroism Spectroscopy-Circular dichroism spectra were acquired in the far UV (185-250 nm) on a Jasco-815 CD spectropolarimeter at 6°C using a 1-cm path length quartz cuvette. All spectra presented in graphs were accumulated and averaged from 5 scans. The peptides containing the second and third helices were generated with a tyrosine residue added to the C termini of each peptide to facilitate concentration determination ( Table 2). The capability of the peptide to form a helical structure was evaluated by quantifying secondary structures using CD spectroscopy in the presence of varying concentrations of 2,2,2-trifluoroethanol (TFE) in 1 mM sodium phosphate buffer (pH 4.2, 7.2, or 8.8). Peptide concentration was maintained at 4 M or 0.16 mg/ml. The secondary structure contents of the peptide were calculated using an online server DICHROWEB, which integrates SELCON, CDSSTR, and K2D analysis algorithms (18).
Statistical Analysis-Data were analyzed with one-way analysis of variance and Newman-Keuls multiple comparison test along with GraphPad Prism software (GraphPad Software, La Jolla, CA). Data are presented as the mean Ϯ S.E. of at least three measurements. Asterisks represent the degree of signifi-cance in comparison with controls (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001).

Cx50 PEST Motif Is Functionally Involved in Lens Fiber
Cell Differentiation-We have previously shown that the CT domain of Cx50 is involved in lens epithelial-fiber differentiation (11). Several residues within a putative PEST motif (rich in amino acids Pro, Glu, Ser, and Thr) of Cx50 CT that we have previously identified are associated with Cx50 phosphorylation, protein stability, and caspase 3 cleavage (19,20). To test if the removal of this motif has any impact on the role of Cx50 in lens fiber differentiation, a truncated form of Cx50 without the PEST motif (⌬PEST-Cx50) was generated (Fig. 1A). The expression of exogenous Cx50 and ⌬PEST-Cx50 was achieved by infecting primary lens cells with recombinant retroviruses. Pulse-chase experiments showed that deletion of the PEST domain did not affect Cx50 protein turnover and stability. The number of lentoids, which are an indication of lens epithelialfiber differentiation, was quantified each day during 8 days of culturing. The result showed that deletion of the PEST domain (⌬PEST-Cx50) significantly abolished the effect of Cx50 on promoting lens fiber differentiation as compared with wildtype Cx50 (Fig. 1B). The attenuated effect due to PEST motif deletion was further confirmed by the reduced expression level of AQP0, a marker for lens fiber differentiation by quantifying fluorescence intensity of immuno-labeled AQP0 (11,21) (Fig.  1C). Within this PEST motif, we have shown that Ser 368 is a caspase-3 cleavage site (20). Ser 364 and Ser 371 are phosphorylated residues (19) and phosphorylation of Ser 364 regulates Cx50 turnover and caspase 3-mediated degradation (20). We generated single (E368A, E368D, S371A, S371D), double (S364A/S371A), and triple (S364A/E368A/S371A) mutants of these three residues by site-directed mutagenesis. However, none of the mutants compromised the influence of Cx50 on promoting lens fiber differentiation as indicated by the level of AQP0 expression (Fig. 1D).
Val 362 within the PEST Domain Is a Critical Residue for the Role of Cx50 in the Lens Fiber Differentiation-To identify possible amino acid residue(s) of Cx50 involved in lens fiber differentiation, a multiple Psi-BLAST search in the NCBI data base was performed. Five highly conserved residues (Glu 328 , Glu 331 , Ala 348 , Val 362 , and Pro 370 ) were identified between sequences of Cx50 and cyclin A. Among these residues, only two, Val 362 and Pro 370 , are located within the PEST domain. Three recombinant retroviruses expressing the mutated forms of Cx50 were generated ( Fig. 2A): M5 contained mutations of all five residues, E328A, E331A, A348P, V362E, and P370A; M4 contained mutations on four resides, E328A, E331A, A348P, and P370A; and M3 contained deletion of the PEST motif (⌬PEST-Cx50) in conjunction with three additional mutations, E328A, E331A, A348P. By expressing wild-type and mutants of Cx50 in the primary lens culture using retroviral infection, we showed that only M4 had a similar stimulatory effect on lentoid formation as wild-type Cx50, whereas ⌬PEST-Cx50, M3, and M5 mutants lack such effects when compared with wild-type Cx50 (Fig. 2B). These results were further verified by the expression level of AQP0 determined by immunofluorescence (Fig. 2C). M4 and M5 have significantly different roles on lens cell differentiation and the only difference between these two mutants is the mutation on Val 362 . These data imply that Val 362 is likely a critical residue in Cx50 CT that is functionally involved in lens fiber differentiation.

Secondary Structural Prediction Suggests that Val 362 Is Located in the Middle of an ␣-Helix
Structure-A homology model of structure for Cx50 CT was constructed using the atomic coordinates of cyclin A 2 as a template in the SWISS-MODEL repository. This molecular modeling result suggested that the predicted structure of Cx50 CT contains four ␣-helical loops where all five highly conserved amino acid residues are located (Fig. 3A). Notably, only Val 362 is located in the center of the 3rd helix, whereas all other conserved residues are located outside the helical structures. This finding implies that Val 362 might be a critical residue in maintaining the helical structures at Cx50 CT.
As an aid to the homology modeling of Cx50 CT, other secondary prediction methods including SOPM were used to predict the Cx50 CT secondary structure primarily based on the composition of amino acids (15). These predictions suggested that about 60 -70% of Cx50 CT contained ␣-helical structures, with four major helices (Fig. 3B, h in blue), which is consistent with the above molecular modeling analysis. Particularly, in all these predictions, Val 362 is also positioned in the middle of the 3rd helical motif. In addition, Profile Hidden Markov Models were also employed to predict the possible secondary structure of Cx50 CT by scoring pairs of aligned secondary structure states in a way analogous to the classical amino acid substitution matrices. The predicted secondary structure of Cx50 CT by the HHpred program (16) shows good alignment between Cx50 CT and rabbit Cx43 CT with an E-value of 1.9E-12. The predicted helical motif containing Val 362 is aligned within the 2nd helix of Cx43 CT (amino acids 342-348) (22) (Fig. 3C). The molecular modeling in conjunction with the structural prediction and alignment analysis suggested that Val 362 is likely to be positioned within the putative 3rd helix of Cx50CT.

Val 362 Plays an Important Role in Maintaining Helical Structure and Lens Fiber Differentiation-To test if Val 362 is critical in maintaining an
␣-helical structure and is crucial for the role of Cx50 in lens fiber differentiation, we generated single-site mutants by substituting Val 362 with representative amino acid residues. Prior to functional assay, the mutation tool in the UCSF Chimera software package (36) was used to analyze the possible structural outcomes of these substitutions. The Rotamers tool allows amino acid side chain rotamers to be viewed and evaluated. "Best" rotamer is selected based on side chain torsion () and probability value, taken from the rotamer library, as well as in the context of the structural environment. The properties of amino acids we substituted are summarized in Table 3. Replacing Val 362 with Arg, Lys, or Glu shows much lower possibilities (0.1-0.2) in maintaining the ␣-helix of Cx50 CT as compared with Leu, Thr, or Ile (0.6 -0.8) (Fig. 4A). The changes of Val 362 to a negatively charged residue, Glu, would disrupt the helical structure presumably due to an i, i ϩ 4 interaction between two negatively charged residues. Interestingly, this modeling program predicts lower probability of ␣-helical content with either Arg or Lys substitution, even though both residues have relatively high ␣-helical propensity and would create a potentially favorable i, i ϩ 4 electrostatic interaction (23). Conversely, changing Val 362 to similar hydrophobic side chain residues, Val, Thr, or Ile, would not cause any structural alteration because two non-hydrogen substituents attached to their C-␤ carbon provides bulkiness near the protein backbone, thereby restricting the presence of these amino acids in the conformations adopted by the main chain. Interestingly, substituting Val 362 with known ␣-helix breakers, Phe or Pro, has the probability value in the middle between V362R/V362K/ V362E) and V362L/V362T/V362I, suggesting that the replacement of Val 362 with Phe or Pro may or may not cause the disruption of ␣-helix in Cx50 CT.
The functional importance of Val 362 and the ␣-helical structure in lens fiber differentiation was tested in chicken lens primary cultures infected with recombinant retroviruses expressing various wild-type and Cx50 mutants. Immunofluorescence analysis showed that these mutants are partially co-localized with endogenous Cx50. Substitution of Val 362 with polar amino acid residues, Arg, Lys, and Glu significantly abolished the effect of Cx50 on increasing the number of lentoids (Fig. 4B) and the expression of AQP0 was detected by Western blotting and immunofluorescence (Fig. 4, C and D). The level of CP49, another lens fiber differentiation marker, is also similarly stimulated (Fig. 4D). The chicken cp49 gene gives rise to two transcripts, CP49 and CP49ins (24), presenting as two separate bands on Western blots.
Interestingly, the V362P mutant has similar capability in promoting lentoid formation and expression level of AQP0 and CP49 as wild-type Cx50 and V362L/V362T/V362I mutants. This could be explained by the fact that proline is present in the human and bovine orthologs of Cx50 at the same conserved site as Val 362 in chicken and Leu 362 in rodents (Fig. 4E). Together, the data suggest that Val 362 of Cx50 or the corresponding residue at the same position from other species plays a crucial role in lens fiber differentiation, and this role is likely to be mediated through the structural integrity of an ␣-helix. This suggests that sequence conservation produces consistent secondary structures with varying primary structures, further illustrating the importance of conserved three-dimensional structures and functions of Cx50 across the animal species.
Scrape loading dye transfer was conducted to determine whether these mutations had any impact on gap junction intercellular communication (Fig. 5). The extent of dye transfer was measured as the distance from the center of scrape line to the farthest extent of RD-or LYstained cells (Fig. 5B). The data suggested that all these mutants form functional gap junctions to a similar degree as wild-type Cx50, albeit their differences in promoting lens fiber differentiation. This result offers further support for the gap junction-independent role of Cx50 in lens epithelial-fiber differentiation as reported (11).
Mutations V362E, but Not V362T and V362P, Disrupted the ␣-Helical Structure Detected by CD Spectroscopy-We tested if Val 362 is crucial for forming ␣-helical structures by examining the secondary structure contents of the peptides derived from putative 2nd and 3rd helices of Cx50CT using far UV CD spectroscopy. In aqueous solution, this peptide was largely unstructured. TFE induced the formation of ␣-helical structures with two major troughs at 208 and 222 nm observed in the CD spectra. The ␣-helical content increased to 34.4 in 30% TFE and to 45.7 in 50% TFE but remained the same in 70% TFE solution (Fig. 6A). TFE has been shown to induce and stabilize the intrinsic secondary structures in the peptides derived from connexins (25,26), possibly by mimicking the plasma membrane-like hydrophobic environment. Sodium phosphate buffers with various pH were also tested in the CD assay and the data suggested that the wild-type peptide adopts the highest content of ␣-helical structures at low pH (pH 4.8) instead of pH 7.2 or 8.2 (Fig. 6B), presumably due to minimization of charge repulsion by the stretches of negatively charged glutamic acid residues found in the helical regions. In addition, the thermal stability of the peptide structures was determined at 0, 6, 25, and 37°C and the peptide exhibited the highest ␣-helical content at low temperature (6°C) (data not shown). The full CD spectrum of the wild-type and the other three mutant peptides, V362E, V362P, and V362T (Table 2), were obtained under the above optimized condition (50% TFE in 1 mM sodium phosphate buffer, pH 4.8, 6°C) (Fig. 6C). Consistent with the molecular modeling studies, the peptides con- A, based on the conserved residues between cyclin A and Cx50, three mutants of Cx50 were generated: M4, E328A, E331A, A348P, and P370A; but no V362E; ⌬PEST-M3, ⌬PEST-Cx50 with E328A, E331A, and A348P; and M5, all five highly conserved mutated residues, E328A, E331A, A348P, V362E, and P370A. B, recombinant retroviruses containing the wild-type and mutants were generated and used to infect lens primary culture, and the formation of lentoids was determined during 8-day culturing and quantified as the mean Ϯ S.E. n ϭ 3. C, primary lens cells expressing exogenous wild-type or Cx50 mutants were stained with 4Ј,6diamidino-2-phenylindole and immunolabeled with anti-AQP0 antibody, followed by fluorescein isothiocyanate-conjugated secondary antibody (upper panel). The expression of AQP0 was quantified (NIH ImageJ software) and presented as a ratio to vehicle RCAS(A) control (lower panel). The data are presented as the mean Ϯ S.E., n ϭ 10. **, p Ͻ 0.01 and ***, p Ͻ 0.001, as compared with wild-type (WT). Bar, 50 m. taining V362T and V362P exhibited similar structures as wildtype peptide with ␣-helical contents of about 42%. However, the V362E mutant reduced the ␣-helical content from 42 to 24.4%. The data demonstrate that substitution of Val 362 to Glu destabilizes the ␣-helical structures, whereas the replacement of Val with Pro or Thr preserves the secondary structures.

DISCUSSION
In this study, we first demonstrate that the PEST domain within the Cx50 CT is indispensable for the role of Cx50 in promoting lens fiber differentiation. By sequence analysis and molecular modeling, the amino acid residue Val 362 was identified and mutation of this residue to charged residues abolished the effect of Cx50 on differentiation. Structural modeling and analysis of CD spectra indicated that Val 362 is crucial in maintaining an ␣-helical structure at Cx50 CT and disruption of this structure by altering the Val 362 residue attenuated the role of Cx50 in differentiation. This study showed the importance of specific structural features of connexins in cellular function.
We have previously shown that only Cx50, but not Cx46 or Cx43, promotes chicken lens fiber differentiation (10). Further studies revealed that the C terminus of Cx50 is responsible for the role of Cx50 in promoting lens cell differentiation (11). However, the mechanism underlying this regulatory function was not clear. Here, we not only identified the involvement of specific motif and amino acid residues, but also demonstrated the importance of the integrity of a unique ␣-helical structure in the lens fiber differentiation process. This specific structural feature could be critical for binding of certain protein(s) important for differentiation. The C termini of connexin molecules have been shown to be involved in various cellular functions. The C terminus of Cx43, an ubiquitously expressed connexin, has been shown to interact with several cellular proteins, such as ZO-1, p38 MAPK, calmodulin, tubulin, etc. (27). A previous study also showed that the C terminus of Cx50 directly interacts with ZO-1 although the function of this interaction is unclear (28). There is increasing evidence suggesting the role of connexins in regulating cell cycle regulators. Zhang et al. (29,30) reported that the expression of Cx43 reduced the level of S phase kinase-associated protein 2 (Skp2), a factor important for cell cycle progression, and that this effect was independent of gap junction function. This is a likely mechanism for tumor suppressing functions of Cx43 on various tumor cell lines. We observed that culturing at days 4 -5 is a critical time point for the stimulatory effect of Cx50 on lens epithelial-fiber differentiation. This implies that Cx50 is likely to be involved in the early stages of differentiation. One possibility is that Cx50 may be involved in the inhibition of fiber progenitor cell proliferation through cell cycle control, which in turn, leads to initiation and promotion of lens cell differentiation.
At the early stages of lens epithelial cell culture, both Cx43 and Cx50 are expressed. Due to technical reasons, we could not determine whether mutants associate with endogenous Cx50 to form heteromeric connexons, which may affect endogenous Cx50 function. However, this is unlikely because expression of exogenous wild-type Cx50 promotes differentiation. By co-immunostaining, we found that mutants had similar distribution patterns as wild-type and partially co-localized with endogenous Cx50. We observed that expression of wild-type, ⌬PEST, and other mutants does not lead to a further increase in gap junction function, consistent with our previous observation (10). One possible explanation is that endogenous connexins are sufficient to form gap junction channels, thus there is no additional increase of gap junction communication due to exogenous expression. Alternatively, Cx50 may not be a major connexin forming gap junction in lens epithelial cells. Cx50 lacking PEST or the non-conserved mutant of Val 362 compromises  (35) were used to predict the structure of Cx50 CT based on the composition of amino acids. All these methods yield similar structural predications as that predicted by molecular modeling based on cyclin A. C, a Profile Hidden Markov model-based algorithm HHpred was used to predict the secondary structure of Cx50 CT based on previously determined Cx43CT (PDB code 1r5s) (22). The Val 362 residue is located in a region of Cx50 that is homologous to an ␣-helix of rCx43.

Amino acids Side chain property
Arginine or lysine Basic Glutamic acid Acidic Leucine or threonine ␤-Chains Isoleucine ␥-Chain Phenylalanine or proline ␣-Helix breaker its capability to promote lens epithelial-fiber differentiation, but has no effect on gap junctions. This is consistent with our previous studies that promotion of lens differentiation is independent of gap junctions using cataract-related mutants lacking gap junction function (11). The three-dimensional structure of Cx43 CT has been reported using NMR spectroscopy, indicating two short helices (31). Based on Profile Hidden Markov Model alignment, we identified one short putative helical structure on Cx50 CT corresponding to the 2nd helical motif on Cx43 CT. Interestingly, this motif contains Val 362 . Our CD studies further confirm the importance of Val 362 in sustaining a helical structure. Furthermore, we found that the integrity of this helical structure on Cx50 CT is directly correlated with the effect of Cx50 on lens fiber differentiation. Substitution of Val 362 with charged amino acids reduces the ␣-helical content and consequently, abolishes the stimulatory effect of Cx50. One interesting observation here is that the substitution of Val 362 with proline retains the helical structure although proline is typically known as a helix breaker. Due to the absence of the backbone amide hydrogen important for stabilizing an ␣-helix, proline is less favorable among 20 amino acids to form normal helical conformation. However, NMR studies verified the incorporation of a proline residue in the middle of an ␣-helix in the Oct-1 transcription factor. Modeling showed an altered H-bonding pattern, but the resulting helix was remarkably stable (32). Instead of valine at residue 362 in chicken, the proline residue is present in human and bovine ortholog forms of Cx50. Therefore, the helical structure is likely to be conserved across different animal species. Cellular function of proteins generally is fulfilled as a part of a complex with other protein molecules. The maintenance of certain structures of Cx50 could be critical to maintain the stability of the protein complex. Identi- FIGURE 4. Comparable change of Val 362 residue retains, but non-comparable change disrupts, the function of Cx50 in lens epithelial-fiber differentiation. A, the probability values of the substitutions of amino acid residues on Val 362 suggest that substitutions to charged amino acid residues Arg, Glu, or Lys are most likely to alter the secondary structure of Cx50 CT (lower values), whereas the substitutions to Ile, Leu, and Thr preserves the structure (higher values) (Dunbrack backbone-dependent rotamer library). B-D, recombinant retroviruses containing various mutants of Cx50, V362R, V362E, V362K, V362L, V362I, V362T, V362F, V362P, and ⌬PEST were generated and used to infect the lens primary cell culture. Substitutions of Val 362 to Ile/Leu/Thr/Pro retained the promoting role of Cx50 in lens epithelial-fiber differentiation, whereas mutants of V362R/V262E/V362K/V362F abolished such a function of Cx50. The degree of differentiation was assessed by quantifying lentoid formation during 8 days of culturing (B) and the expression levels of differentiation marker proteins, AQP0 (C and D) and CP49 (D). In C, the AQP0 level was determined by quantifying the AQP0-immunostained area versus total area. The data are presented as the mean Ϯ S.E. n ϭ 3. **, p Ͻ 0.01; ***, p Ͻ 0.001, as compared with WT. E, the sequence alignment of Cx50 across animal species. fication of potential interacting protein(s) that facilitates the role of Cx50 in lens epithelial-fiber differentiation warrants further investigation.