|Molecular Vision 2004;
Received 16 December 2003 | Accepted 17 March 2004 | Published 26 March 2004
Sequence and peptide map of guinea pig aquaporin 0
Jun Han,1 Mark Little,1
Larry L. David,2
Frank J. Giblin,3
Kevin L. Schey1
1Department of Cell and Molecular Pharmacology, Medical University of South Carolina, Charleston, SC; 2Department of Integrative Biosciences, Oregon Health & Sciences University, Portland, OR; 3Eye Research Institute, Oakland University, Rochester, MI
Correspondence to: Dr. Kevin L Schey, Department of Cell and Molecular Pharmacology, 171 Ashley Avenue, Medical University of South Carolina, Charleston, SC, 29425; Phone: (843) 792-2471; FAX: (843) 792-2475; email: firstname.lastname@example.org
Purpose: Hyperbaric oxygen-treated guinea pigs serve as a useful animal model of nuclear cataract. To understand the structure and function of major intrinsic proteins in this model, the primary sequence and major posttranslational modifications to guinea pig aquaporin 0 (AQP0) were determined.
Methods: The cDNA encoding guinea pig AQP0 was amplified by PCR, cloned and sequenced. After protein enrichment from guinea pig lens tissue, the protein sequence and the posttranslational modifications to AQP0 were determined by using combined chemical cleavage, trypsin and pepsin digestion with matrix assisted laser desorption/ionization mass spectrometry or capillary liquid chromatography tandem mass spectrometry.
Results: The primary structure of AQP0 was determined from the DNA sequence and the translated sequence confirmed by mass spectrometry. Serine 235 was identified to be the major phosphorylation site.
Conclusions: Significant sequence homology was observed between species including putative regulatory sites of phosphorylation and pH regulation. These data form a foundation of information from which to begin assessing posttranslational modifications in cataract models.
Aquaporin 0 (AQP0), formerly known as MIP or MIP26, is the most abundant integral membrane protein in mammalian lenses and is believed to function as a water channel. Its homology with other aquaporins  and functional assays in both lens vesicles  and oocyte expression systems [1,3] support its role as a water channel. Additional roles for this protein have been proposed, including as a structural element, as an ion channel  and as an adhesion molecule [5,6].
Several mutations in AQP0 have been identified as cataractogenic in both mice and humans. Mutations in mouse Mip encoding AQP0 have been shown to be the cause of cataract formation in two mouse models, MipCat-Lop and MipCat-Fr . Two cataract phenotypes are observed in the heterozygous and homozygous knockout mice . Dominantly inherited human cataracts were found in two families carrying two different point mutations in the gene encoding lens AQP0, resulting in mutant proteins, E134G and T138R . These data demonstrate the importance of AQP0 in lens development and in maintenance of lens clarity.
Changes in AQP0 structure have been examined during lens maturation and in cataract models. MP22 was identified to be the main maturation product of AQP0 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and observed to increase with lens age, especially in nuclear fractions relative to cortical fractions[10,11]. Age-related changes in human lens AQP0 revealed N-terminal and C-terminal modifications by reduction in binding of antibodies specific for N- and C-terminal sequences.
AQP0 has been sequenced from a variety of mammalian species including: cow , rat , mouse and human  and, for many years, AQP0 was believed to be lens fiber cell specific. However, a recent study has shown the protein to be present in rat hepatocytes .
Given the usefulness of the hyperbaric oxygen model of nuclear cataract generated in guinea pigs, and the putative role of AQP0 in maintaining lens homeostasis and clarity, it is essential that we determine the sequence and modifications to AQP0 in this animal model. Previous examination of guinea pig AQP0 in this model revealed that treatment of guinea pigs with hyperbaric oxygen in vivo accelerated AQP0 loss in guinea pig lens nucleus . Ultraviolet light radiation in the wavelength range of 340-410 nm produced deleterious, oxidative effects on the nucleus of guinea pig lenses and accelerated the degradation of AQP0 . To be able to more specifically determine cataractogenic changes to AQP0, the sequence and modifications to this protein in normal animals were determined. Both cDNA sequencing and peptide mapping sequencing by tandem mass spectrometry were employed.
All the experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All the solvents and chemical reagents were of HPLC grade or analytical grade.
Eight guinea pig lenses from 3-month-old animals were homogenized in 6.0 mL of TRIsolTM reagent, total lens RNA isolated, and 5 μg reverse transcribed with an oligo dT primer using Superscript IITM reverse transcriptase according to the manufacturer's protocol (Invitrogen, Carlsbad, CA). Degenerate primers for PCR amplification of guinea pig lens aquaporin 0 cDNA were designed by alignment of mouse (31543249), human (6912505), and bovine (27806842) aquaporin 0 cDNA sequences and selection of regions tandem to both the AUG start codon and C-terminal amino acid codon. The forward primer (5'-CCA GGC RYT GTG HYC AYC CCC MCT GCC-3') and reverse primer (5'-TCT MTC CRR CTG GAG CTT TTA-3') corresponded in position to residues 23-49, and 839-859 of the bovine cDNA sequence respectively (R represents A or G; Y represents C or T; H represents A, C, or T; and M represents A or C). The PCR was performed as previously described . A faint band at the expected approximate 850 bp size was subsequently extracted from an agarose gel and reamplified. The single product of the second PCR amplification was then cloned using a TOPO TA Cloning Kit (Invitrogen), and plasmids from 2 clones purified using a Plasmid Midi Kit (QIAGEN, Valencia, CA). DNA sequence determination from both clones was performed by the OHSU-MMI Research Core Facility on a model 377 automated fluorescence sequencer (Applied Biosystems, Foster City, CA). Only two clones were sequenced because the deduced amino acid sequence was subsequently confirmed by mass spectral analysis.
One guinea pig lens (0.10 g) was decapsulated on ice and homogenized in 300 μL of cold 10 mM NaHCO3 (pH 8.0), 5 mM EDTA, 10 mM NaF using a hand held glass homogenizer. Plasma membranes were isolated. Briefly, the homogenate was initially centrifuged at 88,000 x g for 25 min at 4 °C. The crude membrane pellets were washed using the same ultracentrifugation settings with Tris buffer solution consisting of 10 mM Tris, 1 mM EDTA and 1 mM CaCl2, pH 9.0. The pellets were further washed in turn with 4 M urea and 7 M urea in the same Tris buffer, and finally washed with water. About 5% of the urea-washed pellets were dilipidated with ethanol-acetone (1:1, v/v) and used for molecular mass measurement of the intact AQP0 by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS).
The remaining urea-washed membrane pellets were suspended in 200 μL of 1:4 n-propanol-1.5 M Tris buffer (pH 7.4) followed by addition of 1 μL tributyl phosphine and 1 μL 4-vinylpyridine to carry out cysteine reduction and alkylation. The sample was carefully layered with argon and capped. The reaction was carried out for 1 h with intermittent vortexing at room temperature. The excess reagents were removed by three water washes. The pellets after ultracentrifugation were delipidated with ethanol-acetone (1:1, v/v) overnight at -20 °C.
The reduced and alkylated protein pellets equivalent to one half of a guinea pig lens were dissolved into 50 μL of 75% trifluoroacetic acid (TFA). Chemical cleavage was accomplished by adding 5 μL of 0.5 M CNBr in acetonitrile. The reaction lasted 24 h in the dark at room temperature and was quenched by diluting the sample in 5-fold volume of water. The sample was vacuum dried in a speedvac.
Trypsin (modified sequencing-grade; Boehringer-Mannhein, Indianapolis, IN) digestion was performed on the collected C-terminal containing HPLC fraction of the CNBr-cleaved products. The lyophilized fraction was dissolved in 90% 10 mM ammonium bicarbonate (pH 8.0) and 10% acetonitrile. The digestion was conducted at 37 °C with an enzyme-to-substrate ratio of 1:10 overnight. The digest was vacuum dried and then reconstituted in 2% TFA prior to analysis.
The reduced/alkylated protein pellets equivalent to 20% of a whole guinea pig lens were dissolved in 2 μL TFA, sonicated for 5 s, and then diluted dropwise with 98 μL water followed by intermittent vortexing. Peptic digestion was immediately performed by adding 1 μg pepsin. The solution was layered with argon and incubated at 37 °C for 15 h. The digest was snap frozen at -80 °C.
The lyophilized CNBr-cleaved products were solubilized in 6 μL TFA, followed by addition of 9 μL acetonitrile, and then diluted to 300 μL with water. The whole sample was loaded onto a 1 mm ID x 15 cm long Vydac C4 (Grace Vydac, Hesperia, CA) column. Reversed phase liquid chromatography was performed with an Agilent HP 1100 system. The chromatographic conditions for hydrophobic protein-peptide analysis described elsewhere  were used. In short, the mobile phase consisted of solvent A: 0.05% v/v TFA in water and solvent B: 0.05% v/v TFA in acetonitrile-isopropanol (1:2, v/v) solution. The loaded sample was held in 3% solvent B for 10 min at a flow rate of 50 μL/min, then eluted with a linear gradient of 3% to 100% solvent B over 100 min. Fractions containing major peptide fragments of AQP0 were postcolumn collected using a splitter at 45 μL/min, while the remaining 5 μL/min were transferred to the mass spectrometer. The peptic and tryptic digests were chromatographed with the same HPLC instrument on a capillary 300 μm ID x 15 cm long Vydac C18 column with the mobile phase, solvent A: 0.05% TFA and solvent B: 0.05% TFA in 98% acetonitrile. The total flow rate was set at 80 μL/min and 6 μL/min were sent into the analytical column by a pre-injector splitter. 50 μL were loaded onto the analytical column and rinsed for 15 min with solvent A. The gradient for peptic digest was 0% to 40% solvent B over 120 min, and then 40% to 60% in 30 min, while the gradient for tryptic digest separation was a linear gradient of 0% to 100% over 120 min.
Mass spectrometry and data processing
Chromatographic columns were coupled to a LCQ classic ion trap mass spectrometer (Thermo Finnnigan, San Jose, CA) equipped with an electrospray ionization interface. The mass spectra were acquired within the range of m/z 300-2000. The four most abundant peaks in each full MS scan mode were selected for tandem mass spectrometry (MS/MS) by collision induced dissociation. The collision energy level was set at 40%. Dynamic exclusion was applied after four scans for each ion over 1.5 min. Amino acid sequences of peptic and tryptic peptides were searched directly with tandem mass spectra using the SEQUEST computer algorithm (Thermo Finnigan, San Jose, CA) against the guinea pig AQP0 protein sequence translated from the cDNA sequence and saved in FASTA format. All the peptide matches were manually checked with their corresponding raw MS/MS data to ensure the correct peptide assignment.
MALDI-MS was conducted on an Applied Biosystems DE-STR time-of-flight mass spectrometer (Foster City, CA) equipped with a 337 nm nitrogen laser. Samples were solubilized in 75% TFA and spotted in 0.3 μL aliquots onto the sample target prelaid with the same volume of 20 mg/mL 2,5-dihydroxybenzoic acid matrix solution prepared in 50% acetonitrile, 0.1% TFA solution. Mass spectra were acquired in linear mode using an external calibration method. Mass accuracy was better than 0.02%.
Results & Discussion
The cDNA encoding guinea pig AQP0 was amplified by PCR, cloned and sequenced (Figure 1). Since the 3' ends of each primer spanned only the translated region of the cDNA, no untranslated 5' and 3' sequences were determined. The 789 bp sequence (Genbank accession number AY485151) encoded a 263 amino acid protein. The two clones that were sequenced conflicted at 3 sites, resulting in three questionable amino acid residues located at 54 (K or T), 74 (A or V) and 128 (G or D) sites. These ambiguities were likely due to errors introduced during polymerization instead of polymorphisms, because two rounds of amplification using non-proof-reading Taq polymerase were performed. However, these ambiguities were resolved by the subsequent mass spectral analysis described below.
As the first step to analyze the AQP0 protein, MALDI-MS was utilized for molecular weight measurement of the intact protein. Hydrophobicity plots predicted that AQP0 is an integral membrane protein with cytoplasmic amino and carboxyl termini and six layer-spanning domains . The nature of transmembrane domains made this protein extremely hydrophobic and water insoluble. In order for MALDI-MS determination of the intact AQP0, the urea-washed protein pellets were dissolved in 90% TFA and mixed on the sample target with the matrix. Desorption/ionization showed the experimental mass, calibrated by an external standard method, was 28260±2 Da as displayed in Figure 2. Assuming residues 54, 74, and 128 were T, A, and G, respectively, the calculated average mass of AQP0 was 28263.9 Da. This value was within the error of the experimentally determined mass.
To further examine the detailed primary covalent structure of guinea pig AQP0, several peptide fingerprinting methods were used. First, the protein was chemically cleaved by CNBr in 75% TFA to generate large peptide fragments that could be subsequently analyzed by HPLC. The resultant HPLC base peak chromatogram for CNBr-cleaved products is displayed in Figure 3. The major peptide fragment fractions were postcolumn collected and used for further tryptic digestion or MALDI-MS analysis. Table 1 summarizes the measured peptide masses and their assignments to AQP0 assuming the specific residues at 54, 74 and 128 being T, A and G, respectively. The peptide assignments are based on the obtained tandem mass spectra or mass deconvolution of multiply charged ion envelopes of the large peptides. In an effort to obtain the complete peptide sequence of the protein, peptic digestion was carried out on the intact protein. The protein was dissolved in TFA and the digestion was accomplished in a diluted 2% TFA in order to decrease the pepsin activity and reduce spontaneous cleavage. After digestion, the digest was directly loaded onto a capillary column without any further sample treatment. After SEQUEST database searching of the acquired MS/MS data against a FASTA-format guinea pig AQP0 database with the specificity set to no enzyme, more than 95% (250/263) sequence coverage was obtained. The three questionable residues were unambiguously identified to be threonine at 54, alanine at 74 and glycine at 128, respectively. In addition, an oxidation at methionine 183 was detected. The oxidized peptide with its (M+2H)2+ at m/z 680.3 corresponded to residues 178-190, YTGAGMNPARSFA. Its tandem mass spectrum is shown in Figure 4.
The electrospray mass spectrum of the HPLC fraction eluting at 62 to 66 min and its deconvoluted spectrum are presented in Figure 5. The most abundant signal at m/z 976.7 corresponds to the (M+9H)9+ of residues 184-263 (MW=8781). A weak signal appears at 8861 Da, 80 Da higher in molecular weight, possibly representing the phosphorylated C-terminal peptide signal. The intensity of this signal is approximately 5% of the unphosphorylated C-terminal peptide signal. From the above spectrum, the signals from the oxidized product and the incomplete cleavage product 177 to 263 were also observed. To accurately determine the C-terminal phosphorylation site, tryptic digestion of the C-terminal fragment corresponding to residues 184-263 was accomplished. The digest was then analyzed by capillary LC-MS/MS. Based on the tandem mass spectra of the tryptic peptides analyzed, the complete sequence of the carboxyl terminus was observed. Three phosphorylation sites on the carboxyl terminus of AQP0 have been reported before at serine 235 [13,15], 243 and 245 in bovine and human AQP0. We also recently found another phosphorylation at serine 229 in bovine AQP0 (data not shown). In this preparation, only phosphorylation at serine 235 was detected, no in vivo phosphorylation at serines 229 or 243 was observed. The tandem mass spectrum of phosphorylated peptide of residues 229-238 is shown in Figure 6. The phosphopeptide is a tryptic peptide with one missed cleavage site blocked by phosphorylation. As reported earlier [13,15] the major phosphorylation site in AQP0 was determined to be at serine 235 rather than the previously reported serine 243. Herein, the major phosphorylation site in guinea pig AQP0 was also confirmed to be at serine 235. Compared with 25% phosphorylation level of bovine AQP0, only about 5% of serine 235 was phosphorylated in the sample investigated. Like bovine and human AQP0, asparagine residues at 245 and 246 on the carboxyl terminus were both detected as partially deamidated in this investigation (data not shown).
Figure 7 shows an alignment of guinea pig AQP0 with other mammalian AQP0 sequences. Guinea pig AQP0 shares 98.1%, 97.7%, 93.5% and 95.1% identity with mouse, rat, human and bovine AQP0, respectively. The carboxyl terminus of AQP0 has been proposed to be the putative regulation domain acting as a gate to regulate the water channel properties of AQP0 . Furthermore, phosphorylation of the carboxyl terminus of AQP0 may modulate gating function . Comparison of AQP0 among different species indicated phosphorylation at serine 235 is highly conserved in the carboxyl terminus. This phosphorylation site has a consensus sequence R-X-S (X is a neutral amino acid) and is a typical phosphorylation motif recognized by protein kinase A. In support of this hypothesis, the phosphorylation of serine 235 in guinea pig lens AQP0 detected in this study was identical to the phosphorylation in other AQP0's.
Examination of the sequence of guinea pig lens AQP0 reveals two NPA (Asn-Pro-Ala) box domains . Two highly conserved NPA regions also exist at either side of the membrane in each mammalian AQP0. Previous studies have shown a link between NPA domain and pH dependent ion channel activity of AQP0 in Xenopus oocytes [4,24]. Furthermore, pH sensitive water permeability has been shown to be due to histidine 40 , a residue now shown to be present in the guinea pig sequence.
In summary, this paper provides the cDNA sequence encoding guinea pig lens AQP0. The deduced sequence of AQP0 was confirmed by several methods using mass spectrometric peptide mapping and sequencing. Peptide mapping used chemical cleavage and subsequent tryptic digestion on the carboxyl terminus of AQP0, as well as non-specific proteolysis of the intact protein with pepsin. Full amino acid sequence coverage has been attained. Peptide mapping revealed serine 235 in the cytoplasmic C-terminal tail as the major phosphorylation site, which probably acts as an important regulatory site of protein function. Asparagine residues at 245 and 246 were both partially deamidated. The results provided herein will be used as baseline data for the detailed analysis of changes in posttranslational modifications of guinea pig AQP0 upon treatment of the animal model with hyperbaric oxygen and ultraviolet light radiation. Results from these studies are expected to be helpful in understanding nuclear cataractogenesis.
This work was supported by NIH-NEI grants EY-13462 given to KLS, EY-007755 to LD, EY02027 and EY014803 to FG. Support was also provided by the MUSC vision core grant NIH-R24EY014793. The authors acknowledge use of the Medical University of South Carolina Mass Spectrometry Facility.
1. Mulders SM, Preston GM, Deen PM, Guggino WB, van Os CH, Agre P. Water channel properties of major intrinsic protein of lens. J Biol Chem 1995; 270:9010-16.
2. Varadaraj K, Kushmerick C, Baldo GJ, Bassnett S, Shiels A, Mathias RT. The role of MIP in lens fiber cell membrane transport. J Membr Biol 1999; 170:191-203.
3. Yang B, Verkman AS. Water and glycerol permeabilities of aquaporins 1-5 and MIP determined quantitatively by expression of epitope-tagged constructs in Xenopus oocytes. J Biol Chem 1997; 272:16140-6.
4. Kushmerick C, Rice SJ, Baldo GJ, Haspel HC, Mathias RT. Ion, water and neutral solute transport in Xenopus oocytes expressing frog lens MIP. Exp Eye Res 1995; 61:351-62.
5. Fotiadis D, Hasler L, Muller DJ, Stahlberg H, Kistler J, Engel A. Surface tongue-and-groove contours on lens MIP facilitate cell-to-cell adherence. J Mol Biol 2000; 300:779-89.
6. Michea LF, de la Fuente M, Lagos N. Lens major intrinsic protein (MIP) promotes adhesion when reconstituted into large unilamellar liposomes. Biochemistry 1994; 33:7663-9.
7. Shiels A, Bassnett S. Mutations in the founder of the MIP gene family underlie cataract development in the mouse. Nat Genet 1996; 12:212-5.
8. Shiels A, Bassnett S, Varadaraj K, Mathias R, Al-Ghoul K, Kuszak J, Donoviel D, Lilleberg S, Friedrich G, Zambrowicz B. Optical dysfunction of the crystalline lens in aquaporin-0-deficient mice. Physiol Genomics 2001; 7:179-86.
9. Francis P, Chung JJ, Yasui M, Berry V, Moore A, Wyatt MK, Wistow G, Bhattacharya SS, Agre P. Functional impairment of lens aquaporin in two families with dominantly inherited cataracts. Hum Mol Genet 2000; 9:2329-34.
10. Roy D, Spector A, Farnsworth PN. Human lens membrane: comparison of major intrinsic polypeptides from young and old lenses isolated by a new methodology. Exp Eye Res 1979; 28:353-8.
11. Horwitz J, Robertson NP, Wong MM, Zigler JS, Kinoshita JH. Some properties of lens plasma membrane polypeptides isolated from normal human lenses. Exp Eye Res 1979; 28:359-65.
12. Takemoto L, Takehana M. Major intrinsic polypeptide (MIP26K) from human lens membrane: characterization of low-molecular-weight forms in the aging human lens. Exp Eye Res 1986; 43:661-7.
13. Schey KL, Fowler JG, Schwartz JC, Busman M, Dillon J, Crouch RK. Complete map and identification of the phosphorylation site of bovine lens major intrinsic protein. Invest Ophthalmol Vis Sci 1997; 38:2508-15.
14. Kent NA, Shiels A. Nucleotide and derived amino-acid sequence of the major intrinsic protein of rat eye-lens. Nucleic Acids Res 1990; 18:4256.
15. Schey KL, Little M, Fowler JG, Crouch RK. Characterization of human lens major intrinsic protein structure. Invest Ophthalmol Vis Sci 2000; 41:175-82.
16. Huebert RC, Splinter PL, Garcia F, Marinelli RA, LaRusso NF. Expression and localization of aquaporin water channels in rat hepatocytes. Evidence for a role in canalicular bile secretion. J Biol Chem 2002; 277:22710-7.
17. Padgaonkar VA, Lin LR, Leverenz VR, Rinke A, Reddy VN, Giblin FJ. Hyperbaric oxygen in vivo accelerates the loss of cytoskeletal proteins and MIP26 in guinea pig lens nucleus. Exp Eye Res 1999; 68:493-504.
18. Giblin FJ, Leverenz VR, Padgaonkar VA, Unakar NJ, Dang L, Lin LR, Lou MF, Reddy VN, Borchman D, Dillon JP. UVA light in vivo reaches the nucleus of the guinea pig lens and produces deleterious, oxidative effects. Exp Eye Res 2002; 75:445-58.
19. David LL, Lampi KJ, Lund AL, Smith JB. The sequence of human betaB1-crystallin cDNA allows mass spectrometric detection of betaB1 protein missing portions of its N-terminal extension. J Biol Chem 1996; 271:4273-9.
20. Gorin MB, Yancey SB, Cline J, Revel JP, Horwitz J. The major intrinsic protein (MIP) of the bovine lens fiber membrane: characterization and structure based on cDNA cloning. Cell 1984; 39:49-59.
21. Lampe PD, Johnson RG. Amino acid sequence of in vivo phosphorylation sites in the main intrinsic protein (MIP) of lens membranes. Eur J Biochem 1990; 194:541-7.
22. Peracchia C, Girsch SJ. Is the C-terminal arm of lens gap junction channel protein the channel gate? Biochem Biophys Res Commun 1985; 133:688-95.
23. Ehring GR, Lagos N, Zampighi GA, Hall JE. Phosphorylation modulates the voltage dependence of channels reconstituted from the major intrinsic protein of lens fiber membranes. J Membr Biol 1992; 126:75-88.
24. Drake KD, Schuette D, Chepelinsky AB, Jacob TJ, Crabbe MJ, Jacob TJ. pH-Dependent channel activity of heterologously-expressed main intrinsic protein (MIP) from rat lens. FEBS Lett 2002; 512:199-204. Erratum in: FEBS Lett 2002; 516:287.
25. Nemeth-Cahalan KL, Hall JE. pH and calcium regulate the water permeability of aquaporin 0. J Biol Chem 2000; 275:6777-82.
26. Biemann K. Contributions of mass spectrometry to peptide and protein structure. Biomed Environ Mass Spectrom 1988; 16:99-111.