Phenotypic Mutation 'phoebus' (pdf version)
Mutation Type missense
Coordinate41,095,252 bp (GRCm38)
Base Change A ⇒ G (forward strand)
Gene Aqp3
Gene Name aquaporin 3
Synonym(s) RP23-28I8.7, AQP-2, OTTMUSP00000006982, GIL, Gill blood group
Chromosomal Location 41,092,722-41,098,183 bp (-)
MGI Phenotype Animals homozygous for a mutation in this gene display increased drinking behavior, increased urination, and decreased urine osmolality.
Accession Number

NCBI RefSeq: NM_016689; MGI: 1333777

Mapped Yes 
Amino Acid Change Valine changed to Alanine
Institutional SourceBeutler Lab
Ref Sequences
V43A in Ensembl: ENSMUSP00000055110 (fasta)
Gene Model not available
SMART Domains

Pfam:MIP 16 261 1.2e-48 PFAM
Predicted Effect probably benign

PolyPhen 2 Score 0.039 (Sensitivity: 0.95; Specificity: 0.82)
(Using Ensembl: ENSMUSP00000055110)
Phenotypic Category DSS: sensitive, immune system
Penetrance 100% 
Alleles Listed at MGI

All alleles(15) : Targeted, knock-out(1) Gene trapped(14)

Lab Alleles
AlleleSourceChrCoordTypePredicted EffectPPH Score
IGL00481:Aqp3 APN 4 41093632 missense probably damaging 0.99
IGL02994:Aqp3 APN 4 41093614 missense probably benign 0.00
R0138:Aqp3 UTSW 4 41094843 splice acceptor site probably benign
R2097:Aqp3 UTSW 4 41098004 missense possibly damaging 0.95
R2128:Aqp3 UTSW 4 41098061 missense probably benign 0.00
R2129:Aqp3 UTSW 4 41098061 missense probably benign 0.00
R2278:Aqp3 UTSW 4 41093836 missense probably damaging 1.00
R5013:Aqp3 UTSW 4 41093819 missense probably damaging 1.00
Mode of Inheritance Autosomal Recessive
Local Stock Sperm, gDNA
MMRRC Submission 031872-UCD
Last Updated 11/10/2016 11:56 AM by Anne Murray
Record Created unknown
Record Posted 10/12/2009
Other Mutations in This Stock Stock #: A2778 Run Code:
Validation Efficiency: 57/67

GeneSubstitutionChr/LocMutationPredicted EffectZygosity
Capn9 T to C 8: 124,605,478 F396S possibly damaging Homo
Phenotypic Description


The phoebus phenotype was identified in a screen of ENU-mutagenized G3 mice for mutants susceptible to dextran sulfate sodium (DSS)-induced colitis (DSS-induced Colitis Screen).  The index mouse showed continuous weight loss and bleeding from day 3 until day 5 after initial exposure to DSS in the drinking water (Figure 1A).  The mouse lost a total of 15% of its initial weight, and died on day 5 of the experiment.  Histological sections with H&E staining revealed the development of tertiary lymphoid tissue within the wall of the phoebus colon by day 3 of DSS treatment, while wild type colon appears indistinguishable from that of untreated mice (Figure 1B, C).

Nature of Mutation
The phoebus mutation was mapped to Chromosome 4, and corresponds to a T to C transition at position 202 of the Aqp3 transcript, in exon 2 of 6 total exons.
38  -F--G--C--G--S--V--A--Q--V--V--L-
The mutated nucleotide is indicated in red lettering, and results in a valine to alanine substitution at amino acid 43 of the aquaporin 3 protein.
Protein Prediction

Aquaporins (AQPs) are members of the hydrophobic, integral membrane protein superfamily, major intrinisic proteins (MIPs) [(1;2); reviewed in (3)].  AQPs transport water and in some cases solutes such as glycerol, but not protons or hydroxyls, across membranes (4). Water permeation through AQPs occurs bidirectionally, with net flux determined by osmotic gradients. AQPs are hydrophobic integral membrane proteins of about 30 kDa. They are found in virtually every form of life, from bacteria to higher mammals. In mammals, the family consists of 13 members (AQP0-AQP12) expressed in both fluid-transporting tissues (e.g. kidney tubules) and in non-fluid-transporting tissues (e.g. skin). AQP1, AQP2, AQP4, AQP5, and AQP8 primarily transport water, whereas AQP3, AQP7, AQP9, and AQP10 also transport glycerol and are therefore named the ‘aquaglyceroporins.’ In addition, AQP9 may transport other small solutes such as urea (5) [reviewed in (6)], AQP1 is reported to transport cations (7;8) and gases (9-12), AQP7 and AQP9 to transport heavy metal salts such as arsenite (13), and AQP6 to transport chloride at low pH (14). The certainty and biological significance of these findings is unknown. The substrate selectivities of AQP0, AQP11, and AQP12 are unknown.

Although the 3D atomic structure of AQP3 has not been determined, those of several aquaporins, solved by X-ray or electron crystallography, demonstrate a conserved architecture within the protein family [reviewed in (15;16)]. AQP structures possess a pseudo two-fold symmetry in the plane of the membrane created by similarity between the first and second halves of each protein sequence (17). All members studied to date form homotetramers (18;19) of four functionally independent water channels (20). As revealed by the extensively studied AQP1 (18;21-25) and other AQPs (26), each channel is composed of a right handed bundle of six membrane-spanning (H1-H6) and two membrane-inserted but non-membrane-spanning (HB, HE) α-helices, connected by five loops (loops A-E) (Figure 2; PDB ID 1J4N). Both N- and C-termini of AQPs are located in the cytoplasm. Two loops (loops B and E) each contain the highly conserved asparagine-proline-alanine (NPA) motif (residues 76-78 and 192-194 in hAQP1, residues 83-85 and 215-217 in hAQP3), and project from opposite sides of the membrane into the pore to meet in the middle, where the prolines of the NPA motifs stack by van der Walls interactions. Helices HB and HE originate from these prolines. Thus, each half of the protein, joined with inverted topology, contains two membrane-spanning helices followed by a membrane inserted loop containing an NPA motif that leads into the non-membrane spanning helix, followed by a cytoplasmic loop and another transmembrane helix. The six transmembrane helices, tilted at an angle of approximately 30° with respect to the membrane, are packed against each other creating an hourglass-shaped pore with a central selectivity filter.  The overall structure of the molecule forms an extracellular vestibule connected by a narrow pore to a cytoplasmic vestibule.  Interestingly, most of the residues within the pore are hydrophobic, while three hydrophilic nodes are exposed at strategic sites (21). These few water-binding nodes reduce the energy barrier to water transport while keeping substrate-pore interactions to a minimum to facilitate rapid transport.
The key mechanism underlying substrate specificity of AQPs is thought to be the physical constraint imposed by the pore diameter. Four amino acids (Arg195, His180, Phe56, and Cys189 in human AQP1) define the narrowest region of AQP pores (2.8 Å in diameter), three of which (Arg, His, Phe) are conserved across the water-specific AQPs (21). These amino acids constitute the aromatic/arginine (Ar/R) motif, which is located near the center of the channel just above (extracellular to) the NPA motifs, and prevents molecules larger than a single water molecule from permeating the channel. The Cys residue is the binding site for the AQP1 inhibitor HgCl2 (27). In AQPs that also transport other solutes, the constriction region of the pore is defined by different amino acids, resulting in a larger opening with distinct chemical properties (Table 1). For example, in the E. coli AQP GlpF, which is optimized for the rapid transport of glycerol, the His is replaced by Gly, and Cys is replaced by Phe, increasing the size and hydrophobicity of the constriction region (28). In AQP3, which transports both water and glycerol at moderate rates, His is replaced by Gly, and Cys is replaced by Tyr, resulting in a constriction region similar in size to that of GlpF, but with more polar properties. Tyrosine may provide a location for water to form a hydrogen bond, allowing water to be transported more rapidly by AQP3 than GlpF. The two NPA motifs are located opposite each other at the center of the channel and form the second narrowest region of the channel. There, water-specific AQPs possess Phe and Leu residues (Phe24 and Leu149 in hAQP1) opposite the Asn residues of the NPA motif, whereas almost all aquaglyceroporins have two Leu residues (Leu35 and Leu171 in hAQP3), resulting in a pore large enough for the passage of solutes such as glycerol (17).  Most AQPs also prevent the passage of ions because hydrated ions are too large to fit through the constriction region without shedding their hydrating water molecules, which they cannot do because there are too few available carbonyl groups and water-coordinating residues along the pore of the channel (21).
Table 1. Amino acids at constriction region of pore
hAQP1 (water)
hAQP3 (water, glycerol)
GlpF (glycerol)

In order to maintain the proton gradient across the lipid bilayer, the primary energy source for the synthesis of ATP, AQPs cannot permit the translocation of protons. It is well known that protons have high mobility in bulk water, and can move along a continuous linear network of hydrogen-bonded waters through the proton-wire transfer mechanism (29;30). Using computer-aided molecular dynamic simulations of water translocation through AQPs, it was demonstrated that the hydrogen bonding of water molecules is highly choreographed along the path through the channel (31-33). When bulk water-water hydrogen bonds must be broken to allow water molecules to fit through the narrow constriction region, AQPs provide replacement interactions, compensating for the energetic cost of water-water bond rupture and permitting rapid permeation. Proton (hydronium ion, H3O+) movement via proton-wire transfer occurs throughout the pore (34-36). Proton exclusion by AQPs is attributed principally to the electrostatic field created by the dipoles of helices B and E in the region of the NPA motif (34-37). This strong free energy barrier prevents protons from entering the central region of the channel. The arginine and histidine residues of the Ar/R motif are also involved in proton exclusion, since mutation to valine or alanine, respectively, results in a small proton leakage (38;39).


The AQP3 mutation in phoebus mice replaces valine 43 with an alanine. Based on the alignment of human AQP3 and AQP1 (21), the predicted location of valine 43 is in transmembrane helix H1, close to the extracellular border of the membrane bilayer.


By Northern blot analysis and in situ hybridization, AQP3 transcript is detected in multiple organs including kidney, colon, small intestine, stomach, spleen, brain, and lung (40-44). It is found in skin and eye (45;46). AQP3 resides constitutively at the plasma membrane.


AQP3 is also expressed, together with AQP1, on the surface of red blood cells (47). Both proteins define blood group systems. The two Colton blood group antigens (Coa and Cob) arise from a polymorphism in AQP1 (48), and three women lacking Coa and Cob were found to lack AQP1 as a result of mutations that delete or frame shift exon 1, or a point mutation in the first transmembrane domain (49).  Similarly, AQP3 is the GIL blood group antigen, and the GIL-negative blood group phenotype was identified in two women deficient for AQP3 expression due to a mutation in the AQP3 intron 5 splice donor site (50;51). The rare AQP1 and AQP3 deficient individuals, all normal and healthy, were found when they developed alloantibodies against the Co and GIL antigens following pregnancy or blood transfusion. In mice, deletion of AQP3 does not appear to affect erythrocyte water or glycerol permeability (52).

Before discovery of the aquaporin-1 water channel in 1991, the transport of water across the plasma membrane lipid bilayer (discovered in the 1920s) was assumed to occur through simple diffusion (6). This assumption persisted for decades despite the discoveries of ion channels and transporters, and the observations in certain epithelial membranes of water permeation rates 10- to 100-fold higher than diffusion rates, because no water channel proteins could be identified or purified even as late as 1987. In the course of studying Rh blood group antigens, Peter Agre and colleagues at Johns Hopkins University purified an abundant contaminating 28 kDa peptide from erythrocytes (53), later also found to be highly expressed in renal proximal tubules (54), which behaved as an oligomeric integral membrane protein. The cDNA sequence was cloned in 1991 (55). In 1992, it was reported that Xenopus oocytes expressing the 269 amino acid protein, then designated channel-like integral protein of 28 kDa (CHIP28), possessed greatly increased osmotic water permeability and exploded when placed in distilled water (56). Expression of the protein lowered the activation energy of water transfer across the cell membrane from >>10 kcal/mol to <3 kcal/mol, comparable to the activation energy for diffusion of water in bulk solution. In 1997, the name ‘aquaporin’ was officially adopted for CHIP28 and for the growing number of homologous water channel proteins (57). In recognition of the discovery of water channels, one half of the Nobel Prize in Chemistry was awarded to Peter Agre in 2003.


Water is the major component of all life forms, and its organization within biological compartments is essential to life. AQPs mediate the rapid (~2x109 water molecules per subunit per second (58)) and selective transport of water across membrane lipid bilayers, allowing cells to regulate volume and internal osmotic pressure. Thus, AQPs have fundamental roles in all tissues, although their functions are particularly conspicuous in organs that transport fluids. For example, to concentrate urine the kidney resorbs 150-200 liters of water from primary urine each day, a task mediated by AQPs. AQP1 is the primary water channel expressed in the proximal tubules, descending thin limbs of the loop of Henle, and in the descending vasa recta (59;60). Deficiency of AQP1 in humans and mice results in an inability to concentrate urine (61;62) due to impaired transepithelial proximal tubule water permeability and defective fluid absorption (63). AQP2, the vasopressin-regulated water channel (64), is expressed in collecting duct apical membrane and intracellular vesicles (65), and AQP3 and AQP4 are expressed at the basolateral membrane of collecting duct epithelia (42;66). Deletion of AQPs 2-4 in mice reduces collecting duct water permeability, impairing osmotic equilibration between collecting duct lumen and the renal interstitium, and leading to a urinary concentrating defect (67-69). In humans with mutations in AQP2, inability of collecting ducts to absorb water in response to vasopressin results in the autosomal recessive form of nephrogenic diabetes insipidus (OMIM #125800) (70).


Understanding of the functions of the 13 mammalian AQPs in basic physiology and disease has come largely from phenotypic analysis of mice lacking each protein. AQPs are known to participate in a wide variety of processes, such as fluid secretion/absorption from various exocrine glands (salivary, sweat, lacrimal), absorptive epithelia (lung, airways), and secretory epithelia (choroid plexus, ciliary body); pressure regulation, lens transparency, and signal transduction in the eye; water movement into and out of the brain (e.g. during stroke); cell migration (e.g. during angiogenesis, tumor metastasis, wound healing); neuronal signal transduction; skin hydration; cell proliferation; and fat metabolism [reviewed in (4;71)]. The known functions of AQP3 will be discussed here.


AQP3 is an aquaglyceroporin, transporting both water and glycerol at moderate rates (42-44;72). As mentioned above, AQP3 is expressed at the basolateral membrane of renal collecting duct epithelia, and AQP3 null mice exhibit polyuria (increased urination) and polydipsia (increased thirst), although they are able to generate partially concentrated urine in response to water deprivation (68). The polyuria results from a reduction in the rate of water transport across cortical collecting duct basolateral membrane. Renal AQP2 and AQP4 expression is reduced in AQP3 null mice, apparently a maladaptive renal response observed in various forms of acquired polyuria that probably contributes to the urinary concentrating defect in AQP3 null mice (73).
The function of AQP3 in skin hydration, wound healing, and homeostasis has recently been established (71;74). AQP1, AQP5, and AQP7 are expressed in the skin, but AQP3 is the most abundant skin aquaporin (45). It is localized to keratinocyte plasma membranes in the basal layer of the epidermis, the stratum basale, and in the stratum spinosum, which is superficial to stratum basale (Figure 3). AQP3 is not observed in stratum granulosum (superficial to stratum spinosum), nor in the stratum corneum, the most superficial layer of the skin and the barrier to evaporative water loss. In humans, decreased expression of AQP3 is associated with eczema (75). AQP3 null mice display reduced stratum corneum hydration, which is not rescued by exposure to high humidity or occlusion, indicating that defective transepithelial water transport through AQP3 is not the main cause of stratum corneum dehydration (76)Aqp3-/- mice had no defects in stratum corneum structure (77)). Instead, Aqp3-/- mice were found to have reduced glycerol content in the epidermis and in the stratum corneum, but normal glycerol content in the dermis and serum, suggesting an impairment of glycerol transport from blood and dermis into epidermal keratinocytes (77). Systemic or topical administration of glycerol rescued the defects in skin hydration, and in barrier recovery after stratum corneum removal by tape-stripping in AQP3 null mice (78). The mechanism by which glycerol rescues skin defects in AQP3 null mice remains unknown. The glycerol transport function of AQP3 may also be important for epidermal cell proliferation, as Apq3-/- mice show impaired epidermal cell proliferation and wound healing that are rescued by glycerol administration (79). Furthermore, Aqp3-/- mice are resistant to the development of skin tumors, potentially through reduced keratinocyte metabolism due to impairment of ATP generation from glycerol intermediates (80).


 Another role for AQP3 in proliferation has been reported for colonic epithelial cells, where it is expressed at the basolateral membrane (81). Induction of colitis in Aqp3-/- mice by oral DSS or intracolonic acetic acid administration resulted in severe colitis after 3 days leading to death (81). Colonic hemorrhage and epithelial cell loss were observed. In contrast, wild type mice had comparable initial colonic damage, but developed less severe colitis and survived for 8 or more days. Note that because AQP3 null mice have a greater daily fluid consumption than wild type mice (approximately 3-fold higher), the dosage of DSS in these experiments was normalized according to the average daily water intake of each genotype group. Enterocyte proliferation was greatly reduced in AQP3 null mice, presumably contributing to the reduced ability of these mice to cope with the damage induced by DSS. Oral glycerol treatment improved the survival and reduced the severity of colitis in AQP3 null mice, suggesting that glycerol transport may be important for enterocyte proliferation during colitis. Whether glycerol contributes to ATP generation or to another process required for enterocyte proliferation remains unknown.


AQP3 is also expressed in the conjunctival epithelium, where it participates in the regulation of tear film homeostasis and conjunctival barrier function (82). The water permeability of the conjunctiva is significantly slowed in Aqp3-/- mice. AQP3 also facilitates migration and proliferation of corneal epithelial cells (83).


Osmosensing and osmotic regulation are critical cellular activities. In vertebrates, systemic osmotic homeostasis is maintained by osmosensory and effector neurons, which control the release of hormones, such as vasopressin (also called antidiuretic hormone, ADH), that regulate thirst, diuresis, salt appetite, and natriuresis (84). Vasopressin regulates endo- and exocytosis of AQP2 from apical membranes of renal collecting duct cells, increasing cellular water permeability by inducing exocytosis of intracellular vesicles containing AQP2, thus transferring the water channels to the cell surface (64). In addition to systemic osmotic regulation, individual cells can sense and regulate their own volume using a very wide variety of molecules including (for sensing) integrin receptors, growth factor receptors, stretch-activated channels, phospholipases, lipid kinases, and cytoskeletal components; and (for signaling) calcium regulators, kinases (MAPK, Ste-20-related kinases, WNKs), and phosphatases; and (for volume control) channels and transporters (85).  In general, the regulatory mechanisms that control AQP function are thought to act primarily at the transcriptional level rather than through protein modifications such as phosphorylation. Those that specifically control AQP3 are unknown, although extracellular pH (86) and copper (87) have been shown to affect water and glycerol transport through AQP3.


Silencing of AQP3 in human gastric carcinoma SGC7901 cells results in a decrease in several matrix metalloproteinases (MMPs) including MT1-MMP, MMP-2, and MMP-9 (3).  Similarly, increases in the these MMPs was observed upon over-expression of AQP3.  MT1-MMP, MMP-2, and MMP-9 are involved in degrading types I and IV collagen and the extracellular matrix (88). Examination of the PI3K/AKT pathway in cells where AQP3 were altered found that changes in AQP3 levels resulted in changes in the phosphorylation state of AKT.  Xu et al. indicate that AQP3 may upregulate MMPs via PI3K/AKT signaling (3).


AQP3 function in immune cells has been recently studied.  The role of AQP3 has been recently studied in the development and activation of dendritic cells (DCs) (89).  Expression analysis determined that AQP3 is highly expressed to certain regions of the plasma membrane of splenic DCs (89).  In addition, in AQP3 deficient mice, CD4+CD8- conventional DCs (cDCs) were reduced compared to wild-type mice; CD4-CD8+ and CD4-CD8- cDCs were comparable in the AQP3 deficient and wild-type mice (89).  The findings of this study indicate that AQP3 plays a role in altering the subset composition and migratory function of DCs.  AQP3 expression has also been examined in mouse resident peritoneal macrophages (mRPMs) (90).  AQP3 was detected at the plasma membrane of mRPMs (90).  Following exposure to a bacterial peritonitis model, mice deficient in AQP3 expression had reduced survival than wild-type animals, indicating that AQP3 is involved in the host immune response (90).  Examination of bacterial phagocytosis and intracellular digestion by mRPMs found that both phagocytosis and digestion of bacteria is impaired in the mRPMs from animals deficient in AQP3 (90).  Futhermore, cell migration of the mRPMs in the AQP3 deficient mice was impaired.  Zhu et al. speculate that defects in the water- and glycerol-transporting functions of AQP3 could impair the phagocytosis and digestive functions of the AQP3 deficient mRPMs (90)

Putative Mechanism

The phoebus mutation is a Val to Ala substitution of amino acid 43 in AQP3. This amino acid is predicted to align with Ala32 of AQP1 [EMBOSS-Align tool (EMBL-EBI) and (21)], which lies near the end of helix H1, close to the extracellular side of the membrane, in the X-ray crystal structure of bovine AQP1. Importantly, Ala32 was identified as one of the hydrophobic residues lining the AQP1 channel pore (21). The apparent functional defect of AQP3 in phoebus mice is surprising based on the conservative nature of the phoebus mutation. Not only is the amino acid substitution a mild change of one hydrophobic residue for another, but the corresponding position in AQP1 sequences of multiple species is occupied by alanine, the residue encoded by the phoebus mutation. How the mutation impairs AQP3 remains unknown.


Aqp3-/- mice are more susceptible to DSS-induced colitis even when given a dose of DSS equivalent to that given to control animals, suggesting that a deficiency of AQP3 in colonic epithelial cells themselves underlies susceptibility (81). Since DSS dose was not normalized to the daily water intake of phoebus mice, they probably ingested an increased DSS load relative to wild type animals. The relative importance of the phoebus mutation in perturbing renal versus intestinal physiology was not examined, so far as its net ability to produce colitis is concerned. Effects in both tissues may contribute to the colitis susceptibility.


Primers Primers cannot be located by automatic search.
Phoebus genotyping is performed by amplifying the region containing the mutation using PCR, followed by sequencing of the amplified region to detect the single nucleotide transition.
PCR program
1) 95°C             2:00
2) 95°C             0:30
3) 56°C             0:30
4) 72°C             1:00
5) repeat steps (2-4) 29X
6) 72°C             7:00
7) 4°C              ∞
Primers for sequencing
phoebus _seq(R): 5’- CCAAAGCTGAGGTGCTGTTAG -3’
The following sequence of 535 nucleotides (from Genbank genomic region NC_­­­­­000070 for linear genomic sequence of Aqp3, sense strand) is amplified:
2671                                  tgtagtgata gctcccttct gtgcccttcc
2701 cggattcaag aaaaaccagt ggatgtcatt taaccacctg ggtccccagt catgtaccca
2761 ttactgattc ccccccatcc ccccatgagg ctctgcacgt ctcctctttc tgggacttaa
2821 ggagggattt gggttatatc ccaggaagca ccttcagtgg tgtcaggcat gtctgatctc
2881 agtgggactt cacttgcttt gttttccgac agatgtttgg ctgtggctcc gtggctcagg
2941 tggtgctcag ccgtggcacc catggtggct tcctcaccat caacttggct tttggcttcg
3001 ctgtcaccct tggcatcttg gtggctggcc aggtgtctgg taaggcctca accccagctt
3061 cagttttcag ccctcaccag catttccaac aagtatctgc ctagagagca gagggggagg
3121 aacaactcca accaaggacg cacactaaca gcacctcagc tttgggccct ttggaggcaa
3181 gggtggaaga gacttcacag aacct
Primer binding sites are underlined; sequencing primer binding sites are highlighted in gray; the mutated T is indicated in red.
Science Writers Eva Marie Y. Moresco, Anne Murray
Illustrators Diantha La Vine
AuthorsKatharina Brandl, Bruce Beutler
Edit History
06/15/2011 5:29 PM (current)
12/23/2010 11:07 AM
12/21/2010 11:12 AM
12/21/2010 11:12 AM
12/20/2010 2:18 PM
02/03/2010 4:52 PM