April 2005
Volume 46, Issue 4
Free
Lens  |   April 2005
Regulation of Aquaporin Water Permeability in the Lens
Author Affiliations
  • Kulandaiappan Varadaraj
    From the Department of Physiology and Biophysics, State University of New York at Stony Brook, Stony Brook, New York; and the
  • Sindhu Kumari
    From the Department of Physiology and Biophysics, State University of New York at Stony Brook, Stony Brook, New York; and the
  • Alan Shiels
    Departments of Ophthalmology and Visual Sciences and
    Genetics, Washington University, St. Louis, Missouri.
  • Richard T. Mathias
    From the Department of Physiology and Biophysics, State University of New York at Stony Brook, Stony Brook, New York; and the
Investigative Ophthalmology & Visual Science April 2005, Vol.46, 1393-1402. doi:https://doi.org/10.1167/iovs.04-1217
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Kulandaiappan Varadaraj, Sindhu Kumari, Alan Shiels, Richard T. Mathias; Regulation of Aquaporin Water Permeability in the Lens. Invest. Ophthalmol. Vis. Sci. 2005;46(4):1393-1402. https://doi.org/10.1167/iovs.04-1217.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

purpose. To examine Ca2+- and pH-mediated regulation of water permeability of endogenously expressed aquaporin (AQP)0 in lens fiber cells and AQP1 in lens epithelial cells.

methods. Large, right-side-out membrane vesicles were formed from freshly isolated groups of lens fiber cells. Osmotic shrinking or swelling of these vesicles was used to determine the water permeability of endogenously expressed AQP0. The results were compared with those in similar studies of freshly isolated lens epithelial cells, which endogenously expressed AQP1, and of oocytes, which exogenously expressed AQP0.

results. In the lens or in oocytes, decreasing external pH from 7.5 to 6.5 caused a two- to fourfold increase in the water permeability of mammalian AQP0. Several lines of evidence suggest that this effect is mediated by the binding of H+ to a histidine in the first extracellular loop (His40). Lens AQP1 lacks His40 and also lacks pH sensitivity. Increasing Ca2+ caused a two- to fourfold increase in the water permeability of endogenous AQP0. The Ca2+ effect on mouse AQP0 was a 2.5-fold increase in the lens, whereas in oocytes, it was a 4-fold decrease. In either environment, the effect was mediated through calmodulin, most likely through its binding to the proximal domain of the C terminus. Lens AQP1 does not have a similar domain and does not have calcium sensitivity.

conclusions. In either the lens or oocytes, Ca2+ and H+ appear to affect the same mechanism, probably either the open probability of the water channel, or open-channel permeability. The difference between calcium’s effects in lens versus oocytes was remarkable and is not understood. However, in the lens, Ca2+ and H+ are both increased in inner fiber cells, and so in the physiologically relevant environment, both may act to increase the water permeability of AQP0.

The major intrinsic protein (MIP) of lens fiber cell membranes is now known to be a member of the aquaporin (AQP) family of membrane water channels. In modern nomenclature, MIP is known as AQP0. However, before its role as a lens water channel was determined, MIP had a long history of misleading those who investigated its function. 
Early studies found MIP localized in gap junction plaques, 1 2 so the initial hypothesis was that it was a gap junction protein. This view was supported by studies of its functional properties when reconstituted into planar lipid bilayers 3 or liposomes, 4 5 where it formed a nonselective, large-conductance ion channel, just as expected for a gap junction protein. However, when gap junction proteins and MIP were sequenced, MIP was found to not be a member of the gap junction family. 6 Preston and Agre et al. 7 and Preston et al. 8 first discovered and sequenced a member of the AQP family (AQP1), and soon after, other members were found (reviewed in Chepelinsky 9 ) and MIP was determined to be AQP0. 
Exogenous expression of AQP0 in oocytes 10 11 showed that it is indeed a water channel, but unlike the reconstitution results, it conferred no ion permeability to the oocyte membrane. Because exogenous expression is a much more natural process than reconstitution, it was generally assumed that the ion conductance observed in the reconstitution studies was an artifact of that system. In addition to showing that AQP0 is a water channel, exogenous expression studies demonstrated a glycerol transport function for AQP0. 11 12 All these hypotheses were tested by Varadaraj et al. 13 in studies of Cataract Fraser (Catfr) mouse lenses, which express a mutant form of AQP0 that does not traffic to the plasma membrane. In lenses heterozygous for the mutant AQP0, there was reduced fiber cell membrane water permeability, but no change in fiber cell membrane conductance relative to wild-type lenses. In lenses homozygous for the Catfr mutation (i.e., no AQP0 present in the fiber cell membranes), water permeability was approximately 20% of normal, but there was no effect on glycerol transport. 
To summarize, there is good evidence that MIP/AQP0 is the primary source of water permeability in lens fiber cell membranes. Other proposed functions, such as a cell-to-cell ion channel, a transmembrane ion channel, or a glycerol transporter appear to be present only when AQP0 is expressed in a system other than the lens. Michea et al. 14 suggested AQP0 might also be an adhesion protein, based on data showing liposomes in which AQP0 was reconstituted tend to aggregate; however, there has been no method to test this hypothesis in the lens. 
These data indicate that some properties of the lens environment confer different transport properties to AQP0 than those observed by determining either reconstitution or exogenous expression. Similarly, the properties of Cx46 and -50 in the lens 15 16 are different from those determined in studies that determine exogenous expression in oocytes. 17 18 19 20 The lens expresses several unique (not yet found in other tissues) membrane proteins such as AQP0, MP20, 21 and the gap junction proteins Cx46 and -50. 22 23 24 It may also express some yet to be discovered unique regulatory proteins that modify the functions of its other proteins. Although reconstitution and exogenous expression are powerful tools for studying protein function, one needs to know the function of the protein in its native environment, particularly when that environment is the lens. 
Nemeth-Cahalan and Hall 25 reported the seminal findings that Ca2+ and pH regulated AQP0 water permeability, when bovine AQP0 was exogenously expressed in Xenopus oocytes. The purpose of the present study was to evaluate Ca2+- and pH-mediated regulation of APQ0 water permeability in the lens. We report several similarities with the oocyte data, but also one very significant difference. We also evaluated regulation of the lens’ other water channel, AQP1, which is expressed in the lens epithelium, but is also widely expressed in other tissues. A comparison of the results on regulation of these two aquaporins, each expressed in its native environment, provides information on the relation between AQP sequence and physiological regulation of water transport. 
Methods
Preparation of Lenses
The animal protocols used in this study adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Mice or rabbits (New Zealand White) were killed by injection of sodium pentobarbitone solution and frogs (Rana pipiens) were killed by pithing. Eyeballs were excised and placed in calcium-free (24°C) physiological saline solution. Mammalian lens saline contained (in mM): NaCl 150, KCl 4.7, MgCl2 1, glucose 5, and HEPES 5 (pH 7.4). Frog Ringer solution contained (in mM): NaCl 102.5, NaOH 2, KCl 2.5, MgCl2 1.5, HEPES 5, and glucose 5 (pH 7.2). Intact lenses were dissected immediately and used for immunocytochemistry or water transport studies. 
Preparation of Lens Epithelial Cells
Isolated rabbit lens capsules were incubated for 20 minutes at 32°C in dissociation buffer. Rabbit lens buffer contained (in mM): Na-aspartate 150, KCl 4.7, glucose 5, HEPES 5 (pH 7.4) with 0.125% (wt/vol) collagenase. Capsules were gently removed from the dissociation buffer and placed in 1 mL of Ca2+-free lens saline. Epithelial cells were removed from the capsules by gentle trituration with a transfer pipette, and the cells were pelleted (150g for 8 minutes) and resuspended in 250 μL of calcium-free physiological saline at room temperature. 
Preparation of Lens Fiber Cell Membrane Vesicles
Lens fiber cell vesicles were prepared as described in Varadaraj et al. 13 with slight modifications. Lens fiber cells were manually teased out from anterior to posterior sutures by using tweezers. The isolated clumps of fiber cells were incubated for 30 minutes in lens physiological saline (rabbit or mouse) or Ringer solution (frog) containing 5 mM CaCl2, then triturated gently using a rocking platform at room temperature. The vesicles were washed three times for 10 minutes in Ca2+-free lens physiological saline or Ringer solution and the vesicles were suspended in Ca2+-free lens physiological saline or Ringer solution at room temperature. 
Water Permeability of Lens Epithelial Cells and Fiber Cell Membrane Vesicles
Digital video microscopy was used to estimate volume changes and membrane surface area of the epithelial cells or fiber cell vesicles. An image of the cross-section was digitized by using a videoscope image intensifier and a digital video camera (Ikegami Electronics, USA, Maywood, NJ). The image was transferred to a PC using a video frame grabber (Meteor II; Matrox, Dorval, Québec, Canada, with Inspector 4.1 software; Matrox). Images were stored to disc at the rate of one every 3 seconds. The cross-sectional area was determined by tracing and counting pixels using image-analysis software (SigmaScan; SPSS Science, Chicago, IL). We assumed that either vesicles or isolated epithelial cells were spheres, then used the cross-sectional area to calculate their radius and volume. Water permeability was calculated from the initial rate of shrinking or swelling when the bathing solution was switched between normal saline (325 mOsM) and hypertonic lens saline (488 mOsM) by reducing the water. The time constant for the bath change was approximately 0.8 seconds. These methods have been described elsewhere. 13  
Immunostaining of AQPs in the Lens
Polyclonal antibodies against highly conserved sequences within the carboxyl terminal domains of AQP0 (17 amino acids) and -1 (19 amino acids) were purchased from Chemicon (Temecula, CA). Lenses were fixed in 4% paraformaldehyde for 4 to 5 hours at 4°C, washed and used for either paraffin sections or cryosections. Paraffin sections were cut at 4- to 6-μm thickness, using standard techniques. For cryosections, washed lenses were impregnated with 30% sucrose in 0.3% Triton X-100 PBS solution. Lens sections of 10 to 12 μm thickness were cut with a cryomicrotome (Leica, Deerfield, IL). Deparaffinated sections or cryosections were incubated for 5 minutes in 4% paraformaldehyde and 0.3% Triton X-100 in PBS and washed in PBS. Sections were incubated overnight with AQP0 or -1 antibodies at 4°C in a humidified chamber, washed, and exposed to FITC-conjugated IgG for 2 hours. Sections were viewed and photographed with an epifluorescence confocal microscope and accompanying software (AxioVision software; Carl Zeiss Meditec, Dublin, CA). 
Expression of AQP0 in Xenopus Oocytes
Mouse AQP0 expression constructs and cRNA were prepared according to published methods. 11 26 27 Briefly, the mouse AQP0 coding sequence was amplified by polymerase chain reaction with a set of two appropriate primers for 5′ and 3′ ends, and using Pfu DNA polymerase (Stratagene, La Jolla, CA). The orientation and sequences were confirmed by sequencing with fluorescent dye terminators (DNA Sequencing Facility; State University of New York at Stony Brook). Mouse AQP0 was cloned into pcDNA3 (Invitrogen, Carlsbad, CA) in between EcoRI and BamHI sites. RNA was transcribed in vitro using T7 RNA polymerase (mMESSAGE mMACHINE kit; Ambion, Austin, TX). Purified cRNA was dissolved in distilled water and stored at −80°C until use. 
Xenopus laevis oocytes were isolated as described in Kushmerick et al. 11 Briefly, a mature female toad was anesthetized by immersion for 15 to 30 minutes in a 0.1% solution of Tricane (Sigma-Aldrich, St. Louis, MO) in water, and ovary lobes containing stage V and VI oocytes were removed and defolliculated by incubation with 2 mg/mL collagenase type VIII (Sigma-Aldrich) in Ca2+-free oocyte medium for 2 hours and maintained at 18°C in oocyte medium supplemented with 10 μg/mL streptomycin and 10 μg/mL penicillin. The day after isolation, control oocytes were injected with 15 nL of distilled water, and experimental oocytes were injected with 25 ng of mouse AQP0 cRNA in 15 nL of distilled water. Injections were made into the vegetal hemisphere. After injections, the oocytes were maintained in oocyte medium for at least 48 hours at 18°C before water transport studies. 
Western Blot Analysis of AQP0 in Oocytes
Oocytes were tested for translation of injected mouse AQP0 cRNA by Western blot analysis with the anti-AQP0 antibody. Control and experimental oocytes were individually homogenized in 100 μL of lysis buffer (1.5% SDS, 20 mM Tris-HCl [pH 8], protease inhibitor, and phenylmethylsulfonyl fluoride [PMSF]) and sonicated for 2 minutes. The lysate was centrifuged at 12,000g for 20 minutes, and 30 to 50 μL was fractionated by SDS-PAGE. Western blot analysis was performed as described by Varadaraj et al. 28 Briefly, proteins from the denaturing gel were transferred to polyvinylidene difluoride (PVDF) nylon membrane, and an antibody to the C terminus of human AQP0 was used to detect the expression of AQP0. 
Water Permeability Studies of Xenopus Oocytes
Oocyte water permeability was determined by changes in oocyte volume, as described by Kushmerick et al. 11 The solution bathing the oocytes was rapidly changed between 190 mOsM normal oocyte medium and 95 mOsM of diluted oocyte medium. Digital video microscopy was used to estimate oocyte volume changes. Water permeability of the oocytes was calculated as described previously. 13  
Diethyl Pyrocarbonate Treatment
Lens membrane vesicles were incubated in freshly prepared diethyl pyrocarbonate (DEPC; Sigma-Aldrich) solution (0.8 mM at pH 6.0) for 15 minutes and washed two times for 10 minutes each with lens saline (pH 6.0) to remove the excess DEPC. Finally, the lens membrane vesicles were washed twice for 10 minutes each at pH 7.5, and the shrinking and swelling experiments were performed. 
Calmodulin Inhibitors
Lens fiber cell vesicles were incubated in 75 μM trifluoperazine (TFA) or 8 μM calmidazolium (CDZ; both Sigma-Aldrich) for 45 minutes in the dark before performing the shrinking and swelling assays. 
Results
The lens AQP0 data summarized herein were recorded in vesicles that spontaneously budded from isolated clumps of fiber cells. These vesicles were large (15–20 μm diameter) and right side out, and those from wild-type lenses stained positively for AQP0. 13 The vesicles appear to retain intracellular enzymes 29 ; however, it is doubtful that they have normal cytoskeletal structures or internal organelles, which are mostly absent, even in the intact fiber cells. 
Figure 1Ashows such a vesicle and a typical protocol for measuring membrane water permeability, which was determined by the rate of swelling or shrinking when the osmolarity of the bathing solution was changed. Lens epithelial cells, when isolated, become spheres with dimensions similar to those of vesicles; hence, epithelial cell experiments looked very much like the study shown in Figure 1A . Figure 1Bshows a similar study performed on an oocyte that expressed AQP0. The major difference between lens vesicle/epithelial cell data and oocyte data resides in the very much larger volume of the oocytes, which therefore take a much longer time to swell or shrink. The size effect was taken into account by assuming each preparation was a sphere and estimating the surface area of membrane and volume from the cross-sectional area, which is shown at the top of the figures. The specific membrane water permeability was calculated from the initial rate of volume change, the surface area and the osmotic change. The data were arbitrarily fit with a simple exponential function, which was used to obtain the initial rate of volume change. 
The vesicles must have zero calcium and neutral pH to remain stable for long periods; hence, we define this as our control solution. The control solution “Con” contained 0 Ca2+ at pH 7.5. The Hi-Ca2+ (Hi-Ca2+) solution contained 5 mM Ca2+ at pH 7.5. The low-pH (Lo-pH) solution contained 0 Ca2+ at pH 6.5. The Hi-Ca2+, Lo-pH solution contained 5 mM Ca2+ at pH 6.5. Water permeability studies were performed 5 to 10 minutes after the solution change. Data in the bar graphs in several of the figures represent the mean number of studies (n), and the hatched error bars represent standard deviations. 
Ca2+- and pH-Mediated Regulation of Endogenously Expressed Lens AQP0
Figure 2Ashows that knocking out AQP0 in the mouse lens removed approximately 80% of the fiber cell membrane water permeability. This observation is consistent with the original reports by Varadaraj et al. 13 and Shiels et al. 30 Figure 2Bshows that either increasing external Ca2+ or reducing external pH caused about a twofold reversible increase in total membrane water permeability in wild-type (+/+) lenses. Figure 2Cshows the normalized increase in water permeability in the heterozygous AQP0 (+/−) lenses, which express about half as much AQP0 protein as +/+ lenses. 30 In these lenses the effect of either Ca2+ or pH was only approximately a 1.5-fold increase in total membrane water permeability, suggesting that reducing AQP0 levels lessens the effect. In lenses lacking AQP0 protein (Fig. 2D) , neither Ca2+ nor pH had any effect on the remaining water permeability. These data suggest that the effects of either Ca2+ level or pH on fiber cell membrane water permeability are mediated by AQP0 and not by some other lens membrane protein. 
Figure 2Bdemonstrates the presence of Ca2+- and pH-mediated regulation of the water permeability of wild-type mouse AQP0, when endogenously expressed in the normal lens. The Con fiber cell membrane had a total water permeability of 35 μm/s (Fig. 2A) , which was normalized to unity in Figure 2Band is shown by the dashed line. The data in Figure 2Asuggest that in the absence of AQP0 (−/− lenses), the background water permeability was approximately 7 μm/s; hence, the Con permeability due to AQP0 in the +/+ lenses was approximately 28 μm/s. In Hi-Ca2+ lenses, the total permeability increased 2.2-fold to 77 μm/s, which represents a 42-μm/s increase in the water permeability of AQP0. Thus, increasing external Ca2+ caused the water permeability of AQP0 to increase from 28 to 70 μm/s, or a 2.5-fold increase. Reducing external pH had a slightly lesser effect, but still caused the permeability of AQP0 to increase by approximately 2.1-fold. Simultaneously increasing Ca2+ and reducing pH caused an effect that was not significantly different from the individual effects, with the increase in AQP0 permeability being approximately 2.6-fold. Thus, the effects of these relatively large changes in Ca2+ and pH do not appear to be additive, suggesting that either agent caused an essentially saturating effect on the same regulatory mechanism. 
Although the effect of reducing pH is similar to that reported by Nemeth-Cahalan and Hall 25 for exogenously expressed bovine AQP0, the effect of elevating Ca2+ is the opposite, yielding a significant increase in mouse AQP0 water permeability in lens fiber cells as opposed to a significant decrease for the bovine AQP0 in oocytes. There are two possible explanations of this difference. First, it could be a species difference, since we are studying mouse AQP0 and they studied bovine. Second, it could be due to the method of expression, either endogenous in our studies or exogenous in theirs. 
In Figure 3 , we test the hypothesis that there are species differences in the effects of Ca2+ level or pH on the water permeability of AQP0. We did not have access to bovine eyes, and so we used rabbit lenses to test the effects on a different mammalian AQP0. The Con fiber cell membrane water permeability was approximately 36 μm/s, which is quite similar to that of mouse fiber cell membranes, but the effect of Ca2+ level and pH was significantly larger, causing a 3.0- and 3.7-fold increase in total water permeability, respectively. Nevertheless, the effect of elevating Ca2+ is clearly similar to the response of endogenously expressed mouse AQP0 and opposite to the response of exogenously expressed bovine AQP0. Thus, there may be some species differences in the sensitivity of AQP0 to Ca2+ or pH, but not in the direction of the effect. 
Ca2+- and pH-Mediated Regulation of Exogenously Expressed Mouse AQP0
In Figure 4 , we test the hypothesis that exogenous expression of AQP0 in oocytes versus endogenous expression in lens fiber cells reverses the effect of external Ca2+. We cloned and expressed mouse AQP0 in Xenopus oocytes and examined its response to Ca2+ and pH. Figure 4Ashows an immunoblot of isolated membranes from oocytes injected with either AQP0 cRNA (AQP0+) or water (AQP0−). A strong band appears at 28 kDa in the immunoblot of AQP0+ membranes, suggesting good expression of AQP0. Figure 4Bshows that the basal (AQP0−) water permeability of the oocyte membrane was approximately 9 μm/s and was insensitive to either Ca2+ or pH. Figure 4Cshows that in Con conditions, the total membrane water permeability was increased to 21 μm/s by expression of AQP0; thus, the oocyte permeability due to AQP0 was approximately 12 μm/s in these conditions. 
Fig 4Cshows the effects of Ca2+ and pH on exogenously expressed mouse AQP0 in oocytes. Increasing external Ca2+ caused the total membrane water permeability to decrease. The permeability of AQP0 decreased from 12 to 3 μm/s, which is a 4-fold decrease as opposed to a 2.5-fold increase in the lens. The effect of reducing pH was to increase total water permeability. The permeability of AQP0 increased from 12 to 26 μm/s, which is a 2.2-fold increase in oocytes compared with a 2.1-fold increase in lens. Thus, the pH effect appears to be independent of the mode of expression. Simultaneously, increasing Ca2+ and decreasing pH caused offsetting effects, resulting in a slight increase in the total permeability, indicating that the permeability of AQP0 increased from 12 to 14 μm/s. Perhaps a better comparison is the effect of Ca2+ level when pH was 6.5 versus 7.5. At pH 6.5, increasing Ca2+ caused a 1.9-fold reduction in the water permeability of AQP0, whereas at pH 7.5, increasing Ca2+ caused a 4-fold reduction. Similarly, if we look at the effect of pH in Con versus Hi-Ca2+, the increase in permeability was 2.2- versus 4.7-fold, consistent with pH and Ca2+ level affecting the same regulatory mechanism (see the Discussion section), as was the case in the lens; but increasing Ca2+ decreased water permeability, which is the opposite of its effect in the lens. 
Our results from exogenously expressed mouse AQP0 in oocytes are essentially the same as those obtained by Nemeth-Cahalan and Hall 25 from exogenously expressed bovine AQP0 in oocytes. Because we have also determined that the effect of Ca2+ on endogenously expressed mouse AQP0 in the lens is the opposite of that in oocytes, the mode of expression rather than species appears to determine the Ca2+ effect. In the oocytes, the effects of Ca2+ level and pH were offsetting, but comparison of the effect on AQP0 water permeability of elevating Ca2+ at Con-pH (a 4-fold reduction) with that at Lo-pH (a 1.9-fold reduction), or comparing the effect of reducing pH at Con-Ca2+ (a 2.2-fold increase) with that at Hi-Ca2+ (a 4.7-fold increase) suggests that the Ca2+ and pH effects interact, perhaps by affecting the same regulatory mechanism. We therefore set out to learn more about the regulatory mechanism(s). 
Some Possible Sequence-Function Relationships for Lens AQP0 and -1
One of the most powerful aspects of exogenous expression studies is that the sequence of the protein of interest can be modified and the effect can usually be tested (assuming the modified protein gets to the oocyte membrane). However, in the present study we were interested in regulation of AQP0 when expressed in the lens, and so we could not easily modify its sequence, but we could take advantage of naturally occurring variations in sequence. Nemeth-Cahalan and Hall 25 used mutational analysis to identify a histidine in the first extracellular loop (His40) as the site of H+ binding at acid pH and regulation of the water permeability of bovine AQP0. Figure 5Ashows the sequences of the initial 61 amino acids of AQP0 from cow, mouse, and frog (data on rabbit were not available). Each mammalian AQP0 has His40, and the data show that each has sensitivity to external pH. However, frog AQP0 lacks His40, whereas it has His122 and -201, and therefore if His40 is indeed the site of H+ binding, frog AQP0 should lack pH sensitivity. Moreover, AQP1 typically lacks His40, as shown for rabbit in Figure 5A , or mouse in Figure 5C ; hence, AQP1 should also lack pH sensitivity. 
Previous studies have shown that calmodulin interacts with AQP0. 31 32 Moreover, Nemeth-Cahalan and Hall 25 showed that calmodulin inhibitors eliminated the Ca2+-dependent reduction in the water permeability of AQP0 in oocytes. Last, Nemeth-Cahalan et al. 33 showed that coexpression of an inert form of calmodulin with AQP0 also eliminated Ca2+-sensitivity in oocytes. The evidence that calmodulin mediates the Ca2+ effect on AQP0 in oocytes is therefore quite strong, but since the effect is the opposite of that in the lens, we were curious to know whether calmodulin has a similar role in the lens. 
The proximal region of the intracellular C terminus of AQP0 (amino acids 223-243, Fig. 5B ) is the most likely binding domain for calmodulin. Calmodulin is highly negatively charged and usually binds to a protein domain that has a net positive charge and a fair degree of hydrophobicity. 34 35 As shown in Figure 5B , the proximal region of the C terminus of AQP0 has a net of four positive charges and contains nine hydrophobic amino acids (shaded symbols). Because of these factors, this domain is also likely to bind to the inside surface of the plasma membrane. 36 Thus apo-calmodulin, Ca2+-calmodulin, and the membrane may compete for binding the proximal C terminus. Bovine, mouse, rabbit, and frog AQP0 have identical domains, and so all may be regulated by internal Ca2+-calmodulin. However, the closely related protein, AQP1, does not have a region with net positive charge in its C terminus (Fig. 5C) . If the hypothesis that the positively charged region of the C terminus is the regulatory domain for the Ca2+-calmodulin effect is correct, lens AQP1 should be insensitive to Ca2+
Ca2+-Calmodulin Regulation of Endogenously Expressed Lens AQP0
We first used two Ca2+-indicator dyes (Fluo3-AM and Calcium Green2-AM; Molecular Probes, Eugene, OR) to determine whether an increase in external Ca2+ induces an increase in internal Ca2+ in fiber cell membrane vesicles. Although the Ca2+ sensitivity of fluorescence was not calibrated, after external Ca2+ was increased, there was a time-dependent increase in fluorescence that indicated internal Ca2+ was significantly increasing (data not shown). These results were consistent with the Ca2+ effects being either intracellular or extracellular, and if the effect was internal, then it might involve calmodulin. Two calmodulin inhibitors, TFP (75 μM) and CDZ (8 μM), were used to determine whether calmodulin inhibition would eliminate the effect of Ca2+ on AQP0 in rabbit lens fiber cell membrane vesicles (Fig. 6) . As can be seen, either inhibitor greatly reduced the effect of elevated Ca2+ on AQP0 water permeability (compare these data with those in Fig. 3 ), but neither altered the response to a decrease in external pH. These data suggest that calmodulin is indeed involved in the increase in water permeability that occurs when Ca2+ is elevated, that the Ca2+ effect is intracellular, and that the sites of Ca2+ and pH regulation are physically different. Given the previous finding that Ca2+ and pH can saturate the same factor to increase water permeability, this factor appears to be regulated by two different binding sites. 
pH-Mediated Regulation of Endogenously Expressed Rabbit Lens AQP0
DEPC covalently modifies histidine, lysine, or tyrosine amino acid side chains, but preferentially modifies histidine when the reaction is performed at pH 6.0. 25 DEPC (0.8 mM at pH 6) pretreatment increased rabbit lens fiber cell membrane water permeability from 36 to 115 μm/s (Fig. 7A) , a 3-fold increase similar to the 3.7-fold increase induced by lowering pH (Fig. 3) . Reducing external pH caused a small further increase to a level of permeability that was not significantly different from the effect of DEPC alone. The effect of DEPC was reversed by incubating the DEPC-treated vesicles for 60 minutes in 25 mM hydroxylamine hydrochloride, which caused membrane water permeability to decline to 35 ± 9 μm/s. These results are consistent with the hypothesis that DEPC modifies His40 to a state that is equivalent to the binding of H+ and locks AQP0 into the high water permeability state. 
Figure 7Bprovides further support for the His40 hypothesis, as frog AQP0, which lacks His40, but has His122 and His201, was not affected by treatment with DEPC or a reduction in external pH. Nevertheless, increasing external Ca2+ caused an increase in water permeability, again suggesting that the pH and Ca2+ regulatory sites are distinct. Indeed, these data support the hypothesis that the pH effects are mediated by binding of H+ to external His40, whereas the Ca2+ effects are mediated by binding of calcium to internal calmodulin and perhaps binding of either calmodulin or the Ca2+-calmodulin complex with the domain identified in Figure 5Bon the C terminus of AQP0. This finding suggests there are two independent sites for regulation of AQP0 water permeability, but each site affects the same property (e.g., open probability of the water pore or open-pore permeability). 
Effect of Ca2+ and pH on Endogenously Expressed Rabbit Lens AQP1
The data in Figure 8Awere obtained from freshly isolated rabbit lens epithelial cells. These cells expressed AQP1, which lacks the external His40 (shown in Fig. 5 ) that is hypothesized to mediate pH effects on the water permeability of AQP0. These data provide further support for this hypothesis, since Figure 8Ashows the AQP1 also lacked pH sensitivity. The hypothesis that Ca2+ effects are mediated intracellularly through calmodulin is well supported by data shown in Figure 6 . The putative calmodulin-binding motif on the C terminus of AQP0 (Fig. 5)is a possible site of regulation. Figure 8Ashows that the water permeability of AQP1, which lacks a likely calmodulin-binding domain on its C terminus, also lacks Ca2+ sensitivity. The proximal C terminus of AQP0 therefore remains a plausible site for Ca2+-mediated regulation of water permeability. 
Figure 8Bshows immunostaining of AQP0 and -1 in the mouse lens. The same pattern was seen in frog, rat, and rabbit lenses. AQP1 is widely expressed in other tissues of the body and appears to be the generic or default water channel, whereas other AQPs are expressed when there are specialized functional requirements, probably most often for purposes of regulation of water permeability. The changeover from generic AQP1 in the lens epithelium to AQP0 in the fiber cells was strikingly abrupt. The purpose of this switchover is not known, but regulation is one possible reason. We have shown that the water permeability of AQP0 is increased by increasing intracellular Ca2+ or reducing extracellular pH, whereas the water permeability of AQP1 was not affected by these factors. Gao et al. 29 have shown that there is a gradient in calcium concentration within the fiber cells of the lens, with the highest concentration occurring in the most central fibers. Bassnet and Duncan 37 and Mathias et al. 38 have shown a similar gradient in pH, with the lowest pH occurring in the most central fibers. All these observations may be related to the physiological purpose of switching from expression of AQP1 to -0. Based on these data, the purpose may be to increase the water permeability of AQP0 in the central fiber cells (see the Discussion section). 
Figure 8Cshows an expanded view of immunostaining for AQP0 in the lens. The antibody was to the distal end of the C terminus (the last 17 amino acids). Nuclei were marked with 4′,6-diamidino-2-phenylindole (DAPI) and are shown in blue. There was an abrupt decrease in fluorescence intensity at the transition from differentiating fibers (fibers containing nuclei) to mature fibers (fibers without nuclei). This transition occurred at a distance of 10% to 20% into the lens—the site where the fibers lose all organelles 39 the C termini of Cx46 and -50 are cleaved, 29 15 and there is extensive reorganization of gap junction plaques. 40 However, the AQP0 signal did not go to zero at the transition, suggesting there are many intact proteins in the mature fibers. Schey et al. 41 reported that the C terminus of human AQP0 has multiple cleavage sites, but most are distal to the putative calmodulin-binding domain. Ball et al. 42 suggest that the primary cleavage leaves the C terminus intact up to serine 243 (see Fig. 5B ). Altogether, these results suggest that many AQP0 molecules in the mature fibers are intact and should therefore have Ca2+-mediated regulation of their water permeability. 
Discussion
We present data showing that the water permeability of lens AQP0 is increased by increasing intracellular calcium or decreasing extracellular pH. The calcium effect is mediated by calmodulin and may involve binding of calmodulin to the proximal C terminus, which is intracellularly located. The pH effect appears to be mediated by the binding of H+ to His40, which is located in the first extracellular loop. One remarkable finding was that the effect of calcium on water permeability was in opposite directions when the same protein was studied by determining exogenous expression in oocytes rather than in its native environment. Moreover, in either environment calcium’s effect is mediated by calmodulin, and because AQP0 has only one likely domain for calmodulin binding, the opposite effects may be mediated by binding to the same domain. Last, in either environment, pH and calcium appear to affect the same mechanism—possibly either the open-pore permeability or gating of the pore (discussed later). 
Regulation of AQP0 Water Permeability
There are three general methods of rapidly regulating membrane permeability due to expression of a channel-forming protein: (1) regulation of the open probability of the channel, (2) regulation of the permeability of the open channel, and (3) regulation of the number of channel-forming proteins in the plasma membrane. Essentially all ion channel-forming proteins regulate their permeability by gating between an open state, in which ions can traverse the membrane through the open channel, and a closed state, in which ion flow through the channel is completely blocked. The fraction of time spent in the open state, or the fraction of the total number of channels in the open state at any given time, is the open probability of the channel. Most ion channels do not regulate the permeability of the open channel; instead, they regulate the open probability of the channel. However, gap junction channels are one notable exception, as they usually have what are called subconductance states, in which the open-channel permeability is reduced by a discrete amount, but does not go to zero. Last, most transport proteins can be cycled into and out of the plasma membrane by exocytosis or endocytosis of vesicles that contain the relevant protein. An important example is regulation of water transport in the distal tubules of the kidney, where AQP2-mediated water permeability is regulated in this manner. 43  
In the lens, there are no data to suggest that the fiber cells have a pool of AQP0-containing, sub-plasma-membrane vesicles that could insert into the plasma membrane to increase water permeability. Moreover, in our studies, we used membrane vesicles that bleb off of fiber cells. This is a very simple preparation that is unlikely to contain those cytoskeletal elements that are needed for membrane trafficking. 44 Thus, the most likely mechanisms of regulating AQP0-mediated fiber cell membrane water permeability are either gating of the water pore or alteration of the open-pore water permeability. As mentioned earlier, there are precedents among ion channels for either mechanism. Nemeth-Cahalan et al. 33 present a model in which the binding of H+ organizes the water molecules in the pore and increases open-pore permeability. Although this is an elegant model, none of the studies presented herein or in their article can distinguish between effects on open-pore permeability and gating of the pore. Indeed, our own prejudice is that ligand binding induces a conformational change that alters the open probability of the pore. There are no data that directly support this view, but it is a more common mechanism and seems to fit more naturally with our discussion of the calmodulin effects, which will be discussed later. The data do suggest, however, that Ca2+ and pH affect the same mechanism, whether it is gating or open-channel permeation. For example, if Ca2+ causes a twofold increase in open probability, whereas pH causes a twofold increase in the permeability of each open channel, then Ca2+ and pH together should have a fourfold effect. This is not the effect that we observed. What we observed was more consistent with, for example, either Hi-Ca2+ or Lo-pH causing the open probability to increase to near unity, so there could be little incremental effect of both. 
Ca2+-Calmodulin Effects
An interesting question concerns the different Ca2+-calmodulin effects on AQP0 in oocytes versus the lens. AQP0 has only one likely calmodulin-binding domain, which is the proximal region of its C terminus. Calmodulin is a highly negatively charged molecule with several hydrophobic regions. It does not have a specific target sequence, rather it tends to bind regions where there is net positive charge and a concentration of hydrophobic amino acids. 34 35 In general, a domain like this also binds to the inner surface of the plasma membrane, where there are many negatively charged lipid head groups, and the hydrophobic residues of the domain can partially insert between the lipid head groups and be in contact with the hydrophobic interior of the membrane. 36 Thus, the plasma membrane, apo-calmodulin and Ca2+-calmodulin are likely to compete for binding the C terminus of AQP0. In several known situations, activation of calmodulin by Ca2+ causes binding of Ca2+-calmodulin to the target protein. However, the opposite can also occur (e.g., p52 binding to apo-calmodulin 45 ) and inactive apo-calmodulin can have a higher affinity than Ca2+-calmodulin for the target protein. 
These observations suggest a simple model that could produce opposite effects in oocytes and lens fiber cells. Assume the AQP0 water channel is open when the C terminus is bound to the plasma membrane but closes when it is released from the membrane. Further, assume that this intrinsic property of AQP0 is the same regardless of where it is expressed. Differences in membrane composition or posttranslational processing of AQP0 in oocytes and the lens could lead to differences in the relative affinity of the C terminus for Ca2+-calmodulin, apo-calmodulin, and the plasma membrane. The following binding affinity sequences for the C terminus could result:
  •  
    Lens: Ca2+-calmodulin ≪ membrane ≪ apo-calmodulin
  •  
    Oocytes: apo-calmodulin ≪ membrane ≪ Ca2+-calmodulin
In this model of the lens, when calmodulin is mostly in the Ca2+-calmodulin form (Hi-Ca2+), the C terminus is bound to the membrane and the water pore is open, whereas when there is no Ca2+, apo-calmodulin abolishes the C terminus from the membrane and most of the channels are closed. In oocytes, where the hypothetical affinity sequence is the opposite, the effect of Ca2+ would be the opposite. Although there are no data that directly support this model for AQP0, it is testable, and therefore it seemed worthwhile to present it as a working hypothesis for future investigation. 
Physiological Role of AQP0 Regulation
As lens epithelial cells differentiate into fiber cells, they eliminate AQP1 and express AQP0, and so it appears that AQP0 has a role in fiber cells that AQP1 cannot fulfill. We have shown that, in the lens, an increase in either intracellular calcium or extracellular hydrogen causes an increase in the water permeability of AQP0, but neither affects AQP1. Thus, calcium- and pH-mediated regulation of water permeability could be that role. Moreover, there are data showing that intracellular calcium and hydrogen are elevated in central fiber cells, 29 38 37 and it is probable that hydrogen is also elevated in the restricted extracellular spaces between central fiber cells. Altogether, these observations suggest that calcium- and pH-mediated regulation of AQP0 increases its water permeability in the central fiber cells of the lens, and this may be the purpose of replacing AQP1 with AQP0. 
The calmodulin-binding site is thought to be on the C terminus, but a significant fraction of the C terminus of AQP0 is cleaved in most of the mature fibers (Fig. 8C) . However, if the cleavage leaves the C terminus intact up to serine 243, 41 42 then the Ca2+-sensitivity may not be affected, because the putative calmodulin-binding domain is not cleaved. We need data on calcium-mediated regulation of the AQP0 water permeability in mature fibers, but we have been unable to obtain vesicles from the mature fiber cells. Ball et al. 42 expressed AQP0 1-243 in oocytes and found it had about the same water permeability as intact AQP0, but they did not examine Ca2+-mediated regulation. Thus, we lack critical data. Sensitivity to Ca2+ and pH in the normal physiological range within the lens suggests that regulation has a physiological role, and at this stage, a plausible hypothesis is that the role is to increase the membrane water permeability of mature fiber cells. 
A logical extension of this hypothesis is the question of why the mature fiber cells should need an increase in membrane water permeability. Mathias et al. 46 proposed a model in which fluid circulates through the avascular lens to form a microcirculatory system that carries nutrients to the interior fibers. Although this model does not require an increase in membrane water permeability of mature fibers, it does imply that these cells have relatively high water permeability. AQP0 is continuously cleaved as the fibers age, and cleavage appears to induce the formation of cell-to-cell junctions, 47 thus reducing the number of AQP0 channels contributing to membrane water permeability. Varadaraj et al. 13 were able to obtain vesicles from near the differentiating fibers to mature fibers (DF-to-MF) transition (Fig. 8C) , and these vesicles had less immuno staining for the C terminus of AQP0 and lower water permeability than those from the more peripheral DFs. Because the lower water permeability was measured at 0 Ca2+ and pH 7.4, perhaps the effects of high calcium and low pH offset this reduction in water permeability, and thus maintain relatively high water permeability for cells throughout the lens. This is speculation, but it provides one possible motivation for these effects. The actual purpose is an interesting question that calls for extensive further work. 
 
Figure 1.
 
Typical data on the swelling-shrinking assay of either a lens fiber cell membrane vesicle or an oocyte that expressed AQP0. Lens epithelial cells appeared very similar to the vesicles, as they became spherical when isolated. (A) A vesicle formed from lens fiber cell membrane and placed in a hypertonic solution. An exponential was arbitrarily fit to the data to obtain the initial rate of shrinking to estimate the membrane water permeability. (B) An oocyte that expressed AQP0 and was placed in a hypotonic solution. Data were analyzed as in (A).
Figure 1.
 
Typical data on the swelling-shrinking assay of either a lens fiber cell membrane vesicle or an oocyte that expressed AQP0. Lens epithelial cells appeared very similar to the vesicles, as they became spherical when isolated. (A) A vesicle formed from lens fiber cell membrane and placed in a hypertonic solution. An exponential was arbitrarily fit to the data to obtain the initial rate of shrinking to estimate the membrane water permeability. (B) An oocyte that expressed AQP0 and was placed in a hypotonic solution. Data were analyzed as in (A).
Figure 2.
 
Regulation of fiber cell membrane water permeability by pH or Ca2+ was absent in lenses from AQP0-knockout mice. (A) Reductions in AQP0 in heterozygous (+/−) and homozygous (−/−) knockout lenses caused systematic reductions in fiber cell membrane water permeability relative to that in wild-type (+/+) lenses. In lenses lacking AQP0 (−/−), fiber cell membrane water permeability was approximately 7 μm/s or approximately 20% of that in wild-type (+/+) lenses, where it was 35 μm/s. (B) Regulation of fiber cell membrane water permeability by pH or Ca2+ in wild-type mouse lenses. The data are normalized to the +/+ permeability of 35 μm/s in (A), and the increase in permeability in reduced pH, Hi-Ca2+, or both is shown relative to the control value (dashed line). (C) The same protocol as in (A), but for +/− lenses. Note the reduction in the effect of either pH or Ca2+. (D) The same protocol as in (A), but for −/− lenses. Note that neither pH nor Ca2+ had an effect in the absence of AQP0.
Figure 2.
 
Regulation of fiber cell membrane water permeability by pH or Ca2+ was absent in lenses from AQP0-knockout mice. (A) Reductions in AQP0 in heterozygous (+/−) and homozygous (−/−) knockout lenses caused systematic reductions in fiber cell membrane water permeability relative to that in wild-type (+/+) lenses. In lenses lacking AQP0 (−/−), fiber cell membrane water permeability was approximately 7 μm/s or approximately 20% of that in wild-type (+/+) lenses, where it was 35 μm/s. (B) Regulation of fiber cell membrane water permeability by pH or Ca2+ in wild-type mouse lenses. The data are normalized to the +/+ permeability of 35 μm/s in (A), and the increase in permeability in reduced pH, Hi-Ca2+, or both is shown relative to the control value (dashed line). (C) The same protocol as in (A), but for +/− lenses. Note the reduction in the effect of either pH or Ca2+. (D) The same protocol as in (A), but for −/− lenses. Note that neither pH nor Ca2+ had an effect in the absence of AQP0.
Figure 3.
 
Regulation of the water permeability of rabbit lens AQP0 by pH or Ca2+. AQP0 in fiber cell membrane vesicles from either rabbit or mouse lenses (Figure 1)showed an increase in water permeability when either pH was lowered or Ca2+ was elevated, whereas, previous reports on the water permeability of bovine AQP0 expressed in oocytes indicated elevated Ca2+ would decrease water permeability. 25
Figure 3.
 
Regulation of the water permeability of rabbit lens AQP0 by pH or Ca2+. AQP0 in fiber cell membrane vesicles from either rabbit or mouse lenses (Figure 1)showed an increase in water permeability when either pH was lowered or Ca2+ was elevated, whereas, previous reports on the water permeability of bovine AQP0 expressed in oocytes indicated elevated Ca2+ would decrease water permeability. 25
Figure 4.
 
Regulation of the water permeability of mouse lens AQP0 when exogenously expressed in Xenopus oocytes. (A) Western blot analysis for expression of mouse AQP0 in oocytes. Oocytes injected with cRNA for mouse lens AQP0 (AQP0+) showed a band at 28-kDa that was lacking in water-injected oocytes (AQP0−). The 28-kDa band suggests good expression of AQP0. (B) The membrane water permeability of oocytes not expressing AQP0 was not regulated by either pH or Ca2+. (C) The membrane water permeability of oocytes expressing AQP0 was increased by a decrease in pH, but was decreased by an increase in Ca2+, which is the opposite of the effect of Ca2+ on AQP0 in the mouse lens.
Figure 4.
 
Regulation of the water permeability of mouse lens AQP0 when exogenously expressed in Xenopus oocytes. (A) Western blot analysis for expression of mouse AQP0 in oocytes. Oocytes injected with cRNA for mouse lens AQP0 (AQP0+) showed a band at 28-kDa that was lacking in water-injected oocytes (AQP0−). The 28-kDa band suggests good expression of AQP0. (B) The membrane water permeability of oocytes not expressing AQP0 was not regulated by either pH or Ca2+. (C) The membrane water permeability of oocytes expressing AQP0 was increased by a decrease in pH, but was decreased by an increase in Ca2+, which is the opposite of the effect of Ca2+ on AQP0 in the mouse lens.
Figure 5.
 
The amino acid sequence and membrane topology of lens aquaporins. (A) A comparison of the initial amino acid sequence of AQP0 from several species and of AQP1. The histidine located at position 40 (His40) is the likely site of H+ binding and regulation by external pH. His40 is absent from frog AQP0 and -1. (B) The membrane topology of AQP0 and -1. The proximal region of the C terminus of AQP0 (223-243) has a net charge of +4 and many hydrophobic amino acids (shaded circles) and is therefore a likely domain for the binding of calmodulin. These properties also make it likely to bind to the inner surface of the plasma membrane. (C) The membrane topology of AQP1. The external His40 is absent, and the C terminus does not have the same positively charged region as AQP0. Thus, both pH and Ca2+-calmodulin regulation are expected to be absent from AQP1.
Figure 5.
 
The amino acid sequence and membrane topology of lens aquaporins. (A) A comparison of the initial amino acid sequence of AQP0 from several species and of AQP1. The histidine located at position 40 (His40) is the likely site of H+ binding and regulation by external pH. His40 is absent from frog AQP0 and -1. (B) The membrane topology of AQP0 and -1. The proximal region of the C terminus of AQP0 (223-243) has a net charge of +4 and many hydrophobic amino acids (shaded circles) and is therefore a likely domain for the binding of calmodulin. These properties also make it likely to bind to the inner surface of the plasma membrane. (C) The membrane topology of AQP1. The external His40 is absent, and the C terminus does not have the same positively charged region as AQP0. Thus, both pH and Ca2+-calmodulin regulation are expected to be absent from AQP1.
Figure 6.
 
Calmodulin inhibitors blocked Ca2+-mediated regulation of lens AQP0 but not the effect of reduced pH. (A) The effect of the calmodulin inhibitor TFP (75 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation. (B) The effect of the calmodulin inhibitor CDZ (8 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation.
Figure 6.
 
Calmodulin inhibitors blocked Ca2+-mediated regulation of lens AQP0 but not the effect of reduced pH. (A) The effect of the calmodulin inhibitor TFP (75 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation. (B) The effect of the calmodulin inhibitor CDZ (8 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation.
Figure 7.
 
Regulation of lens AQP0 by pH was mediated by H+ binding to His40. (A) Modification of external histidines by DEPC locked mammalian AQP0 in the high water-permeability state. This result is the same as found when mammalian AQP0 was expressed in oocytes, where mutations showed that it was specifically a modification of His40 that caused the effect. (B) Frog lens AQP0, which lacks His40, was not affected by DEPC, was not sensitive to pH, and yet retained Ca2+ sensitivity.
Figure 7.
 
Regulation of lens AQP0 by pH was mediated by H+ binding to His40. (A) Modification of external histidines by DEPC locked mammalian AQP0 in the high water-permeability state. This result is the same as found when mammalian AQP0 was expressed in oocytes, where mutations showed that it was specifically a modification of His40 that caused the effect. (B) Frog lens AQP0, which lacks His40, was not affected by DEPC, was not sensitive to pH, and yet retained Ca2+ sensitivity.
Figure 8.
 
Regulation of water permeability and expression of lens AQP0 and -1. (A) The water permeability of lens AQP1, which lacks His40 and the calmodulin-binding domain on its C terminus, also lacked pH and Ca2+ sensitivity. (B) Typical expression patterns for AQP1 and -0 in the lens. The images are from mouse lens but frog, rat, and rabbit lenses all showed the same pattern of expression. Note that AQP1 was exclusively expressed in the lens epithelium and was absent in the fiber cells. The pattern of AQP0 expression was complimentary to that of AQP1. It was not expressed in the epithelium but was expressed in the membranes of fiber cells. The transformation was abrupt and suggests that AQP0 has some functional role that is absent in AQP1, perhaps pH- and Ca2+-medinated regulation of water permeability. (C) Immunostaining pattern for the distal C terminus of AQP0. Blue: nuclei. There was an abrupt decrease in intensity at the transition from differentiating fibers to mature fibers, suggesting significant cleavage of the distal C termini. Epi, lens epithelium; DF, peripheral differentiating fiber cells; MF, the organelle-free mature fiber cells.
Figure 8.
 
Regulation of water permeability and expression of lens AQP0 and -1. (A) The water permeability of lens AQP1, which lacks His40 and the calmodulin-binding domain on its C terminus, also lacked pH and Ca2+ sensitivity. (B) Typical expression patterns for AQP1 and -0 in the lens. The images are from mouse lens but frog, rat, and rabbit lenses all showed the same pattern of expression. Note that AQP1 was exclusively expressed in the lens epithelium and was absent in the fiber cells. The pattern of AQP0 expression was complimentary to that of AQP1. It was not expressed in the epithelium but was expressed in the membranes of fiber cells. The transformation was abrupt and suggests that AQP0 has some functional role that is absent in AQP1, perhaps pH- and Ca2+-medinated regulation of water permeability. (C) Immunostaining pattern for the distal C terminus of AQP0. Blue: nuclei. There was an abrupt decrease in intensity at the transition from differentiating fibers to mature fibers, suggesting significant cleavage of the distal C termini. Epi, lens epithelium; DF, peripheral differentiating fiber cells; MF, the organelle-free mature fiber cells.
BokD, DockstaderJ, HorwitzJ. Immunocytochemical localization of the lens main intrinsic polypeptide (MIP26) in communicating junctions. J Cell Biol. 1982;92:213–220. [CrossRef] [PubMed]
FitzGeraldPG, BokD, HorwitzJ. Immunocytochemical localization of the main intrinsic polypeptide (MIP) in ultrathin frozen sections of rat lens. J Cell Biol. 1983;97:1491–1499. [CrossRef] [PubMed]
EhringGR, ZampighiG, HorwitzJ, BokD, HallJE. Properties of channels reconstituted from the major intrinsic protein of lens fiber membranes. J Gen Physiol. 1990;96:631–664. [CrossRef] [PubMed]
GirschSJ, PeracchiaC. Lens cell-to-cell channel protein: II. Conformational change in the presence of calmodulin. J Cell Biol. 1985;83:227–335.
PeracchiaC, GirschSJ. Calmodulin site at the C-terminus of the putative lens gap junction protein MIP26. Lens Eye Toxic Res. 1989;6:613–621. [PubMed]
NicholsonBJ, TakemotoLJ, HunkapillerMW, HoodLE, RevelJP. Differences between liver gap junction protein and lens MIP 26 from rat: implications for tissue specificity of gap junctions. Cell. 1983;32:967–978. [CrossRef] [PubMed]
PrestonGM, AgreP. Isolation of the cDNA for erythrocyte integral membrane protein of 28 kilodaltons: member of an ancient channel family. Proc Natl Acad Sci USA. 1991;88:11110–11114. [CrossRef] [PubMed]
PrestonGM, CarrollTP, GugginoWB, AgreP. Appearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science. 1992;256:385–387. [CrossRef] [PubMed]
ChepelinskyAB. The MIP transmembrane channel gene family. Handbook of Membrane Channels. 1994;413–432.Academic Press Orlando, FL.
MuldersSM, PrestonGM, DeenPM, GugginoWB, van OsCH, AgreP. Water channel properties of major intrinsic protein of lens. J Biol Chem. 1995;270:9010–9016. [CrossRef] [PubMed]
KushmerickC, RiceSJ, BaldoGJ, HaspelHC, MathiasRT. Ion, water and neutral solute transport in Xenopus oocytes expressing frog lens MIP. Exp Eye Res. 1995;61:351–362. [CrossRef] [PubMed]
KushmerickC, VaradarajK, MathiasRT. Effects of lens major intrinsic protein on glycerol permeability and metabolism. J Membr Biol. 1998;161:9–19. [CrossRef] [PubMed]
VaradarajK, KushmerickC, BaldoGJ, BassnettS, ShielsA, MathiasRT. The role of MIP in lens fiber cell membrane transport. J Membr Biol. 1999;170:191–203. [CrossRef] [PubMed]
MicheaLF, de la FuenteM, LagosN. Lens major intrinsic protein (MIP) promotes adhesion when reconstituted into large unilamellar liposomes. Biochemistry. 1994;33:7663–7669. [CrossRef] [PubMed]
GongX, BaldoGJ, KumarNM, GilulaNB, MathiasRT. Gap junctional coupling in lenses lacking alpha3 connexin. Proc Natl Acad Sci USA. 1998;95:15303–15308. [CrossRef] [PubMed]
BaldoGJ, GongX, Martinez-WittinghanFJ, KumarNM, GilulaNB, MathiasRT. Gap junctional coupling in lenses from alpha connexin knockout mice. J Gen Physiol. 2001;118:447–456. [CrossRef] [PubMed]
LinJS, EckertR, KistlerJ, DonaldsonP. Spatial differences in gap junction gating in the lens are a consequence of connexin cleavage. Eur J Cell Biol. 1998;76:246–250. [CrossRef] [PubMed]
WhiteTW, BruzzoneR, GoodenoughDA, PaulDL. Voltage gating of connexins. Nature. 1994;371:208–209. [CrossRef] [PubMed]
StergiopoulosK, AlvaradoJL, MastroianniM, Ek-VitorinJF, TaffetSM, DelmarM. Hetero-domain interactions as a mechanism for the regulation of connexin channels. Circ Res. 1999;84:1144–1155. [CrossRef] [PubMed]
EckertR. pH gating of lens fibre connexins. Pflugers Arch. 2002;443:843–851. [CrossRef] [PubMed]
LouisCF, HurKC, GalvanAC, et al. Identification of an 18,000-dalton protein in mammalian lens fiber cell membranes. J Biol Chem. 1989;264:19967–19973. [PubMed]
KistlerJ, BullivantS. Protein processing in lens intercellular junctions: cleavage of MP70 to MP38. Invest Ophthalmol Vis Sci. 1987;28:1687–1692. [PubMed]
YangDI, LouisCF. Molecular cloning of sheep connexin49 and its identity with MP70. Curr Eye Res. 1996;15:307–314. [CrossRef] [PubMed]
PaulDL, EbiharaL, TakemotoLJ, SwensonKI, GoodenoughDA. Connexin46, a novel lens gap junction protein, induces voltage-gated currents in nonjunctional plasma membrane of Xenopus oocytes. J Cell Biol. 1991;115:1077–1089. [CrossRef] [PubMed]
Nemeth-CahalanKL, HallJE. pH and calcium regulate the water permeability of aquaporin 0. J Biol Chem. 2000;275:6777–6782. [CrossRef] [PubMed]
VaradarajK, SkinnerDM. Cytoplasmic localization of transcripts of a complex G+C-rich crab satellite DNA. Chromosoma. 1994;103:423–431. [CrossRef] [PubMed]
KumariSS, VaradarajK, ValiunasV, et al. Functional expression and biophysical properties of polymorphic variants of the human gap junction protein connexin37. Biochem Biophys Res Commun. 2000;274:216–224. [CrossRef] [PubMed]
VaradarajK, KumariSS, SkinnerDM. Actin-encoding cDNAs and gene expression during the intermolt cycle of the Bermuda land crab Gecarcinus lateralis. Gene. 1996;171:177–184. [CrossRef] [PubMed]
GaoJ, SunX, Martinez-WittinghanFJ, GongX, WhiteTW, MathiasRT. Connections between connexins, calcium and cataracts in the lens. J Gen Physiol. 2004;124:289–300. [CrossRef] [PubMed]
ShielsA, BassnettS, VaradarajK, et al. Optical dysfunction of the crystalline lens in aquaporin-0-deficient mice. Physiol Genom. 2001;7:179–186. [CrossRef]
LouisCF, HoganP, ViscoL, StrasburgG. Identity of the calmodulin-binding proteins in bovine lens plasma membranes. Exp Eye Res. 1990;50:495–503. [CrossRef] [PubMed]
GirschSJ, PeracchiaC. Calmodulin interacts with a C-terminus peptide from the lens membrane protein MIP26. Curr Eye Res. 1991;10:839–849. [CrossRef] [PubMed]
Nemeth-CahalanKL, KalmanK, HallJE. Molecular basis of pH and Ca2+ regulation of aquaporin water permeability. J Gen Physiol. 2004;123:573–580. [CrossRef] [PubMed]
ChinD, MeansAR. Calmodulin: a prototypical calcium sensor. Trends Cell Biol. 2000;10:322–328. [CrossRef] [PubMed]
HoeflichKP, IkuraM. Calmodulin in action: diversity in target recognition and activation mechanisms. Cell. 2002;108:739–742. [CrossRef] [PubMed]
ArbuzovaA, WangL, WangJ, et al. Membrane binding of peptides containing both basic and aromatic residues: experimental studies with peptides corresponding to the scaffolding region of caveolin and the effector region of MARCKS. Biochemistry. 2000;39:10330–10339. [CrossRef] [PubMed]
BassnettS, DuncanG. Direct measurement of pH in the rat lens by ion-sensitive microelectrodes. Exp Eye Res. 1985;40:585–590. [CrossRef] [PubMed]
MathiasRT, RiquelmeG, RaeJL. Cell to cell communication and pH in the frog lens. J Gen Physiol. 1991;98:1085–1103. [CrossRef] [PubMed]
BassnettS. Lens organelle degradation. Exp Eye Res. 2002;74:1–6. [CrossRef] [PubMed]
JacobsMD, SoellerC, SisleyAM, CannellMB, DonaldsonPJ. Gap junction processing and redistribution revealed by quantitative optical measurements of connexin46 epitopes in the lens. Invest Ophthalmol Vis Sci. 2004;45:191–199. [CrossRef] [PubMed]
ScheyKL, LittleM, FowlerJG, CrouchRK. Characterization of human lens major intrinsic protein structure. Invest Ophthalmol Vis Sci. 2000;41:175–182. [PubMed]
BallLE, LittleM, NovakMW, GarlandDL, CrouchRK, ScheyKL. Water permeability of C-terminally truncated aquaporin 0 (AQP0 1-243) observed in the aging human lens. Invest Ophthalmol Vis Sci. 2003;44:4820–4828. [CrossRef] [PubMed]
BorgniaM, NielsenS, EngelA, AgreP. Cellular and molecular biology of the aquaporin water channels. Annu Rev Biochem. 1999;68:425–458. [CrossRef] [PubMed]
RidleyAJ. Rho proteins: linking signaling with membrane trafficking. Traffic. ;2:303–310. [CrossRef] [PubMed]
AndreasenTJ, LuetjeCW, HeidemanW, StormDR. Purification of a novel calmodulin binding protein from bovine cerebral cortex membranes. Biochemistry. 1983;22:4615–4618. [CrossRef] [PubMed]
MathiasRT, RaeJL, BaldoGJ. Physiological properties of the normal lens. Physiol Rev. 1997;77:21–50. [PubMed]
GonenT, SlizP, KistlerJ, ChengY, WalzT. Aquaporin-0 membrane junctions reveal the structure of a closed water pore. Nature. 2004;429:193–197. [CrossRef] [PubMed]
Figure 1.
 
Typical data on the swelling-shrinking assay of either a lens fiber cell membrane vesicle or an oocyte that expressed AQP0. Lens epithelial cells appeared very similar to the vesicles, as they became spherical when isolated. (A) A vesicle formed from lens fiber cell membrane and placed in a hypertonic solution. An exponential was arbitrarily fit to the data to obtain the initial rate of shrinking to estimate the membrane water permeability. (B) An oocyte that expressed AQP0 and was placed in a hypotonic solution. Data were analyzed as in (A).
Figure 1.
 
Typical data on the swelling-shrinking assay of either a lens fiber cell membrane vesicle or an oocyte that expressed AQP0. Lens epithelial cells appeared very similar to the vesicles, as they became spherical when isolated. (A) A vesicle formed from lens fiber cell membrane and placed in a hypertonic solution. An exponential was arbitrarily fit to the data to obtain the initial rate of shrinking to estimate the membrane water permeability. (B) An oocyte that expressed AQP0 and was placed in a hypotonic solution. Data were analyzed as in (A).
Figure 2.
 
Regulation of fiber cell membrane water permeability by pH or Ca2+ was absent in lenses from AQP0-knockout mice. (A) Reductions in AQP0 in heterozygous (+/−) and homozygous (−/−) knockout lenses caused systematic reductions in fiber cell membrane water permeability relative to that in wild-type (+/+) lenses. In lenses lacking AQP0 (−/−), fiber cell membrane water permeability was approximately 7 μm/s or approximately 20% of that in wild-type (+/+) lenses, where it was 35 μm/s. (B) Regulation of fiber cell membrane water permeability by pH or Ca2+ in wild-type mouse lenses. The data are normalized to the +/+ permeability of 35 μm/s in (A), and the increase in permeability in reduced pH, Hi-Ca2+, or both is shown relative to the control value (dashed line). (C) The same protocol as in (A), but for +/− lenses. Note the reduction in the effect of either pH or Ca2+. (D) The same protocol as in (A), but for −/− lenses. Note that neither pH nor Ca2+ had an effect in the absence of AQP0.
Figure 2.
 
Regulation of fiber cell membrane water permeability by pH or Ca2+ was absent in lenses from AQP0-knockout mice. (A) Reductions in AQP0 in heterozygous (+/−) and homozygous (−/−) knockout lenses caused systematic reductions in fiber cell membrane water permeability relative to that in wild-type (+/+) lenses. In lenses lacking AQP0 (−/−), fiber cell membrane water permeability was approximately 7 μm/s or approximately 20% of that in wild-type (+/+) lenses, where it was 35 μm/s. (B) Regulation of fiber cell membrane water permeability by pH or Ca2+ in wild-type mouse lenses. The data are normalized to the +/+ permeability of 35 μm/s in (A), and the increase in permeability in reduced pH, Hi-Ca2+, or both is shown relative to the control value (dashed line). (C) The same protocol as in (A), but for +/− lenses. Note the reduction in the effect of either pH or Ca2+. (D) The same protocol as in (A), but for −/− lenses. Note that neither pH nor Ca2+ had an effect in the absence of AQP0.
Figure 3.
 
Regulation of the water permeability of rabbit lens AQP0 by pH or Ca2+. AQP0 in fiber cell membrane vesicles from either rabbit or mouse lenses (Figure 1)showed an increase in water permeability when either pH was lowered or Ca2+ was elevated, whereas, previous reports on the water permeability of bovine AQP0 expressed in oocytes indicated elevated Ca2+ would decrease water permeability. 25
Figure 3.
 
Regulation of the water permeability of rabbit lens AQP0 by pH or Ca2+. AQP0 in fiber cell membrane vesicles from either rabbit or mouse lenses (Figure 1)showed an increase in water permeability when either pH was lowered or Ca2+ was elevated, whereas, previous reports on the water permeability of bovine AQP0 expressed in oocytes indicated elevated Ca2+ would decrease water permeability. 25
Figure 4.
 
Regulation of the water permeability of mouse lens AQP0 when exogenously expressed in Xenopus oocytes. (A) Western blot analysis for expression of mouse AQP0 in oocytes. Oocytes injected with cRNA for mouse lens AQP0 (AQP0+) showed a band at 28-kDa that was lacking in water-injected oocytes (AQP0−). The 28-kDa band suggests good expression of AQP0. (B) The membrane water permeability of oocytes not expressing AQP0 was not regulated by either pH or Ca2+. (C) The membrane water permeability of oocytes expressing AQP0 was increased by a decrease in pH, but was decreased by an increase in Ca2+, which is the opposite of the effect of Ca2+ on AQP0 in the mouse lens.
Figure 4.
 
Regulation of the water permeability of mouse lens AQP0 when exogenously expressed in Xenopus oocytes. (A) Western blot analysis for expression of mouse AQP0 in oocytes. Oocytes injected with cRNA for mouse lens AQP0 (AQP0+) showed a band at 28-kDa that was lacking in water-injected oocytes (AQP0−). The 28-kDa band suggests good expression of AQP0. (B) The membrane water permeability of oocytes not expressing AQP0 was not regulated by either pH or Ca2+. (C) The membrane water permeability of oocytes expressing AQP0 was increased by a decrease in pH, but was decreased by an increase in Ca2+, which is the opposite of the effect of Ca2+ on AQP0 in the mouse lens.
Figure 5.
 
The amino acid sequence and membrane topology of lens aquaporins. (A) A comparison of the initial amino acid sequence of AQP0 from several species and of AQP1. The histidine located at position 40 (His40) is the likely site of H+ binding and regulation by external pH. His40 is absent from frog AQP0 and -1. (B) The membrane topology of AQP0 and -1. The proximal region of the C terminus of AQP0 (223-243) has a net charge of +4 and many hydrophobic amino acids (shaded circles) and is therefore a likely domain for the binding of calmodulin. These properties also make it likely to bind to the inner surface of the plasma membrane. (C) The membrane topology of AQP1. The external His40 is absent, and the C terminus does not have the same positively charged region as AQP0. Thus, both pH and Ca2+-calmodulin regulation are expected to be absent from AQP1.
Figure 5.
 
The amino acid sequence and membrane topology of lens aquaporins. (A) A comparison of the initial amino acid sequence of AQP0 from several species and of AQP1. The histidine located at position 40 (His40) is the likely site of H+ binding and regulation by external pH. His40 is absent from frog AQP0 and -1. (B) The membrane topology of AQP0 and -1. The proximal region of the C terminus of AQP0 (223-243) has a net charge of +4 and many hydrophobic amino acids (shaded circles) and is therefore a likely domain for the binding of calmodulin. These properties also make it likely to bind to the inner surface of the plasma membrane. (C) The membrane topology of AQP1. The external His40 is absent, and the C terminus does not have the same positively charged region as AQP0. Thus, both pH and Ca2+-calmodulin regulation are expected to be absent from AQP1.
Figure 6.
 
Calmodulin inhibitors blocked Ca2+-mediated regulation of lens AQP0 but not the effect of reduced pH. (A) The effect of the calmodulin inhibitor TFP (75 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation. (B) The effect of the calmodulin inhibitor CDZ (8 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation.
Figure 6.
 
Calmodulin inhibitors blocked Ca2+-mediated regulation of lens AQP0 but not the effect of reduced pH. (A) The effect of the calmodulin inhibitor TFP (75 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation. (B) The effect of the calmodulin inhibitor CDZ (8 μM) on the regulation of AQP0 water permeability. The effect of elevated Ca2+ is largely blocked with no effect on pH-mediated regulation.
Figure 7.
 
Regulation of lens AQP0 by pH was mediated by H+ binding to His40. (A) Modification of external histidines by DEPC locked mammalian AQP0 in the high water-permeability state. This result is the same as found when mammalian AQP0 was expressed in oocytes, where mutations showed that it was specifically a modification of His40 that caused the effect. (B) Frog lens AQP0, which lacks His40, was not affected by DEPC, was not sensitive to pH, and yet retained Ca2+ sensitivity.
Figure 7.
 
Regulation of lens AQP0 by pH was mediated by H+ binding to His40. (A) Modification of external histidines by DEPC locked mammalian AQP0 in the high water-permeability state. This result is the same as found when mammalian AQP0 was expressed in oocytes, where mutations showed that it was specifically a modification of His40 that caused the effect. (B) Frog lens AQP0, which lacks His40, was not affected by DEPC, was not sensitive to pH, and yet retained Ca2+ sensitivity.
Figure 8.
 
Regulation of water permeability and expression of lens AQP0 and -1. (A) The water permeability of lens AQP1, which lacks His40 and the calmodulin-binding domain on its C terminus, also lacked pH and Ca2+ sensitivity. (B) Typical expression patterns for AQP1 and -0 in the lens. The images are from mouse lens but frog, rat, and rabbit lenses all showed the same pattern of expression. Note that AQP1 was exclusively expressed in the lens epithelium and was absent in the fiber cells. The pattern of AQP0 expression was complimentary to that of AQP1. It was not expressed in the epithelium but was expressed in the membranes of fiber cells. The transformation was abrupt and suggests that AQP0 has some functional role that is absent in AQP1, perhaps pH- and Ca2+-medinated regulation of water permeability. (C) Immunostaining pattern for the distal C terminus of AQP0. Blue: nuclei. There was an abrupt decrease in intensity at the transition from differentiating fibers to mature fibers, suggesting significant cleavage of the distal C termini. Epi, lens epithelium; DF, peripheral differentiating fiber cells; MF, the organelle-free mature fiber cells.
Figure 8.
 
Regulation of water permeability and expression of lens AQP0 and -1. (A) The water permeability of lens AQP1, which lacks His40 and the calmodulin-binding domain on its C terminus, also lacked pH and Ca2+ sensitivity. (B) Typical expression patterns for AQP1 and -0 in the lens. The images are from mouse lens but frog, rat, and rabbit lenses all showed the same pattern of expression. Note that AQP1 was exclusively expressed in the lens epithelium and was absent in the fiber cells. The pattern of AQP0 expression was complimentary to that of AQP1. It was not expressed in the epithelium but was expressed in the membranes of fiber cells. The transformation was abrupt and suggests that AQP0 has some functional role that is absent in AQP1, perhaps pH- and Ca2+-medinated regulation of water permeability. (C) Immunostaining pattern for the distal C terminus of AQP0. Blue: nuclei. There was an abrupt decrease in intensity at the transition from differentiating fibers to mature fibers, suggesting significant cleavage of the distal C termini. Epi, lens epithelium; DF, peripheral differentiating fiber cells; MF, the organelle-free mature fiber cells.
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×