Carbonic anhydrases (CAs) are zinc-containing metalloenzymes that catalyze the reversible conversion of carbon dioxide to bicarbonate ion and a proton. These enzymes, found in all types of organisms, are encoded by five independent gene families: the α-, β-, γ-, δ-, and ε-CAs (Supuran, 2008b; Xu et al., 2008). Only α-CA family members have been identified in mammals (Henry and Swenson, 2000; Esbaugh and Tufts, 2006). Thus far, 16 α-CA isozymes have been described, with different catalytic activities and subcellular localizations: CAI, II, III, VII, and XIII are cytoplasmic; CAIV, IX, XII, XIV, and XV are membrane-bound; CAVa and Vb are mitochondrial; and CAVI is secreted (Supuran, 2008a, 2008b). CAVIII, X, and XI are called CA-related proteins (CA-RPs) because they lack classical CA enzymatic activity and their biological functions remain unclear.
The importance of carbonic anhydrases has been shown in many physiological and pathological processes, including pH and CO2 homeostasis, electrolyte secretion, respiration, CO2/HCO3− transport, and bone resorption and calcification (Supuran, 2008a). In the respiratory system, CAs are involved in most CO2 transport and excretion from metabolically active tissues to red blood cells, and finally to gas exchanging organs, such as the lungs (Henry and Swenson, 2000). Cytoplasmic CAI and II are the most abundant CA isozymes in the red blood cells, while both cytoplasmic (CAII) and membrane-bound CAs (CAIV) are found in the lung (Henry and Swenson, 2000; Esbaugh and Tufts, 2006). The role of CAs has also been studied in detail in regulating renal physiology, such as acid/base homeostasis and bicarbonate reabsorption (Purkerson and Schwartz, 2007). CAs are expressed in most kidney segments. Cytoplasmic CAII accounts for most CA activity in the kidney, and different combinations of membrane-bound CAs, including CAIV, CAXII, CAXIV, and CAXV, are also expressed in different species (Purkerson and Schwartz, 2007). CAII is involved in bone physiology, such as osteoclast differentiation and bone resorption, and CAII-deficiency leads to osteopetrosis in humans and mice (Sly and Hu, 1995; Lehenkari et al., 1998; Margolis et al., 2008). In addition, various CAs expressed in the gastrointestinal canal and digestive glands are involved in ammonia detoxification, saliva production, gastric acid production, bile production, pancreatic juice production, and intestinal ion transport (Pan et al., 2007; Supuran, 2008a). Mitochondrial CAV has been associated with molecular signaling, such as insulin secretion in pancreatic β cells (Parkkila et al., 1998). Some CA isozymes are down- or up-regulated in various tumors and have been associated with oncogenesis and tumor progression (Supuran, 2008a, 2008b). Consistently, two hypoxia-inducible CA isozymes, CAIX and XII, were shown to promote tumor growth by regulating pH in an acidic and hypoxic microenvironment (Jarvela et al., 2008; Chiche et al., 2009). In the brain, CAIV and XIV were shown to regulate pH in extracellular fluid (Shah et al., 2005). Lastly, CAIII, which is highly expressed in many tissues, including skeletal muscle, has relatively low CA activity. Its physiological role is unclear because mice lacking CAIII do not have noticeable anatomical or physiological abnormalities (Kim et al., 2004). CAIII can, however, protect cells from oxidative damage by S-glutathiolation on two cysteine residues in response to oxidative stress, thereby functioning as an oxyradical scavenger (Raisanen et al., 1999; Mallis et al., 2002, 2000; Gailly et al., 2008).
CA activity in the inner ears has been demonstrated biochemically by visualizing CO2 hydration in guinea pigs, cats, and chinchillas (Erulkar and Maren, 1961; Lim et al., 1983; Hsu and Nomura, 1985; Okamura et al., 1996). Although the specific location of CA activity in the inner ear differs slightly depending on the species and detection method, CA activity is present in the organ of Corti, outer sulcus cells and their associated root processes, otic fibrocytes in the spiral ligament, as well as spiral ganglion neurons. Consistently, immunohistochemical studies have detected CAII and CAIII immunoreactivity in the otic fibrocytes, spiral limbus, and Reissner's membrane in gerbil, guinea pig, and human cochlea, with slight differences among species and detection methods.
The role of CA activity in the inner ear was first suggested by Erulkar and Maren (Erulkar and Maren, 1961). In cat inner ears, CA activity is distributed along the cochlear duct, with the highest concentrations in the apical turn, progressively lower concentrations toward the basal turns, and lowest concentrations in the vestibule (Erulkar and Maren, 1961). Administering intravenous acetazolamide, a specific CA inhibitor, considerably reduced the perilymph and endolymph volumes and pressures, and eliminated the high potassium concentration in the endolymph (Erulkar and Maren, 1961). Consistently, inhibiting CA activity affects the generation of endocochlear potential in guinea pigs and rats, suggesting a critical role of CA activity in the sound transduction (Prazma, 1978; Sterkers et al., 1984; Ikeda et al., 1987). Furthermore, CA activity has also been associated with normal otolith (otoconia) development in mice, chicken, and fish (Purichia and Erway, 1972; Kido et al., 1991; Tohse et al., 2004).
Despite the critical roles in the auditory and vestibular functions, no systematic study has been conducted to determine the expression patterns of all known CA isozymes in the mammalian inner ear. Although CAII and III localization have been described in some cochlear regions (Spicer and Schulte, 1991; Ichimiya et al., 1994; Weber et al., 2001), they do not account for all CA activities in the inner ear (Erulkar and Maren, 1961; Lim et al., 1983; Hsu and Nomura, 1985; Okamura et al., 1996). In this study, we examined the temporal and spatial expression of all known CA isozymes by quantitative real-time polymerase chain reaction (qRT-PCR) and in situ hybridization in mouse embryonic and postnatal inner ears. Our study provides comprehensive expression profiles for the CA isozymes and their potential role in inner ear development and function.
RESULTS AND DISCUSSION
Expression Profiles of Carbonic Anhydrases in Various Tissues in Mice
The expression levels of all known CA isozymes were analyzed by semi-quantitative reverse transcriptase PCR (RT-PCR) of RNAs isolated from various tissues, including inner ear, brain, kidney, and liver of 3-week-old mice (Fig. 1A). In the inner ear, transcripts of four cytosolic isozymes (Car1, Car2, Car3, and Car13) and a membrane-bound isozyme (Car14) were expressed at higher levels than the other isozymes. In the brain, transcripts of two cytosolic isozymes (Car2 and Car7) and two membrane-bound isozymes (Car4 and Car14) were strongly expressed. In addition, transcripts of three acatalytic CA-related proteins (CA-RPs) (Car8, Car10, and Car11) were also expressed. In the kidney, transcripts of a cytosolic isozyme (Car2), four membrane-bound isozymes (Car4, Car12, Car14, and Car15), and a mitochondrial isozyme (Car5b) were expressed relatively strongly. In the liver, two cytosolic isozymes (Car2 and Car3), a mitochondrial isozyme (Car5a), a membrane-bound isozyme (Car14), and a CA-RP (Car8), were expressed. These results showed that each tissue expresses a unique combination of CA isozymes, although some isozymes, such as Car2 and Car14, appear to be ubiquitous.
Temporal Expression of Carbonic Anhydrases During Inner Ear Development
To determine temporal expression patterns of CAs during inner ear development, we performed qRT-PCR at embryonic day (E) 15.5, postnatal day (P) 0, P5, and P21 (Fig. 1B–D). Consistent with the semi-quantitative RT-PCR (Fig. 1A), Car1, Car2, Car3, Car13, and Car14 showed relatively high expression at P21 (Fig. 1B,C). In addition, our qRT-PCR data showed that Car8, Car11, and Car12 were also expressed at moderate levels during inner ear development (Fig. 1C), while the remaining isozymes (Car4, Car5a, Car5b, Car6, Car7, Car9, Car10, and Car15) were expressed at undetectable or extremely low levels (Fig. 1D).
Car3 was strongly expressed in the inner ears throughout developmental stages, suggesting an important role of Car3 in inner ear development and function. Interestingly, acatalytic CA-RPs, such as Car8 and Car11, were also expressed at moderate levels in developing inner ears, although Car8 became undetectable in the mature inner ear at P21 (Fig. 1C). These results suggest that CA isozymes may play specific roles in different stages of inner ear development.
Spatial Expression Patterns of Carbonic Anhydrases During Inner Ear Development
The localization of expression for each CA transcript was determined by in situ hybridization in developing (E15.5, P0, P5) and mature (P21) inner ears.
Abundant cytosolic isozymes, Car1 and Car2
Consistent with qRT-PCR results (Fig. 1), Car1 transcripts were barely detectable in the inner ear during embryonic and postnatal development (E15.5, P0, and P5; data not shown) and were present in the bone marrow of the otic capsule at P21 (Fig. 2A–A″, arrows). Recently, it has been shown that CAI is associated with ankylosing spondylitis, characterized by abnormal bone formation, in humans and CAI function is involved in bone formation in vitro (Chang et al., 2012).
Car2 transcripts were expressed in mesenchymal cells adjacent to the stria vascularis, spiral limbus, and basilar membrane (Fig. 2B). Weak Car2 expression was also present in epithelial cells in the outer sulcus and greater epithelial ridge (GER) (Fig. 2B′,B″, bracket and arrowhead). This broad Car2 expression continued during neonatal development (Fig. 2E), but by P21, Car2 expression was mainly present in type II and IV otic fibrocytes (Fig. 2D′, bracket) and bone marrow of the otic capsule (Fig. 2D′, white arrows). Weaker Car2 expression was present in type I fibrocytes, the apical and basal margins of stria vascularis (arrowheads), and outer hair cells in the organ of Corti (asterisk) at P21. In the vestibules, weak and broad Car2 expression was present in mesenchymal tissues surrounding the sensory organs at E15.5 (Fig. 2B, bracket), and stronger expression in the otic capsule at P5, and bone marrows at P21 (Fig. 2D, brackets).
Previously, CAII immunoreactivity has been detected in type I, III, and IV otic fibrocytes in the spiral ligament of the gerbil cochlea (Spicer and Schulte, 1991) and in type I, II, and IV fibrocytes in guinea pigs (Ichimiya et al., 1994). These expression patterns differ slightly from our results, which show strong Car2 expression in type II and IV fibrocytes (Fig. 2). This difference could be due to different species or detection methods.
Potential role of Car3 as a free radical scavenger in otic fibrocytes
Initially, Car3 expression was closely associated with prospective type I otic fibrocytes at E15.5 (Fig. 2F,F′, asterisk), but gradually expanded to the entire spiral ligament, encompassing all types of otic fibrocytes by P5 (Fig. 2G,G′), and later restricted to type I, III, and V otic fibrocytes at P21 (Fig. 2H,H′, asterisk). Interestingly, Car3 expression in the cochlear lateral wall at P21 was complementary to Car2, that is Car3 was more abundant in type I and III fibrocytes while Car2 was stronger in type II and IV fibrocytes (Fig. 2D′,H′, asterisk and brackets). Weak Car3 expression was also detected in the otic capsule (Fig. 2G′,H′, arrows) and bony modiolus area at P5 (Fig. 2G,I, arrows), which disappeared by P21 (Fig. 2H, arrow). Car3 expression was also present in the vestibular mesenchymal regions throughout development (Fig. 2F–H, brackets). Consistent with our data, CAIII immunoreactivity was also present in type I and III fibrocytes from human and gerbil cochlea (Spicer and Schulte, 1991; Weber et al., 2001).
Unlike the other isozymes, CAIII has relatively little CA activity (Sanyal et al., 1982; Engberg et al., 1985), but mainly functions as an oxyradical scavenger and protects cells from oxidative damage by S-glutathiolation on two cysteine residues in response to oxidative stress (Raisanen et al., 1999; Mallis et al., 2002, 2000; Gailly et al., 2008). Nevertheless, introducing CAIII into hepatocellular carcinoma cells accelerated the speed at which culture medium was acidified, suggesting that CAIII can still catalyze carbon dioxide hydration, although at a reduced rate (Dai et al., 2008). The in vivo function of CAIII is unclear because Car3 null mutants do not display obvious anatomical or physiological abnormalities (Kim et al., 2004). A recent study, however, showed that a lack of CAIII function attenuates ATP synthesis by mitochondria in skeletal muscle, although its underlying mechanism is unclear (Liu et al., 2007). Therefore, CAIII may play dual roles in otic fibrocytes: facilitating mitochondrial ATP synthesis as well as detoxifying free radicals resulting from active ATP synthesis.
Possible Functional Association of Car12 and Car13 With Pendrin
Car12 was expressed at P0 in regions that would become the outer sulcus and spiral prominence (Fig. 3A,A′, bracket). By P5, Car12 expression in the outer sulcus had expanded to the associated root cell processes, continuing to P21 (Fig. 3B,I, brackets). Car12 was also expressed in the endolymphatic duct at E15.5 (Fig. 3C). These Car12 expression patterns closely overlap with expression of Pendrin, an anion exchanger that transports Cl−, I−, formate, and HCO3− (Fig. 3D–F) (Everett et al., 1999; Wangemann et al., 2004). Car12 was also weakly expressed in the spiral ganglion in the postnatal cochlea (Fig. 3A, arrows; data not shown) and also in Boettcher's cells at P21 (Fig. 3I, arrow). In addition, Car12 was strongly expressed in the vestibular hair cells of cristae and maculae (Fig. 3G,H), but not in the auditory hair cells of the organ of Corti in the cochlea (Fig. 3I, asterisk).
Car13 transcripts were observed in the lesser epithelial ridge (LER) area of E15.5 cochlea, which closely overlaps with Bmp4 expression (Fig. 3J,M, brackets) (Hwang et al., 2010). At P0, Car13 expression was restricted to the epithelial cells of the spiral prominence and outer sulcus, with faint expression in prospective type III fibrocytes (Fig. 3K, bracket, arrowheads), which was maintained until P21 (Fig. 3L, bracket and arrowheads; data not shown). In the vestibule, weak Car13 expression was observed in the mesenchymal tissues along the otic capsule at E15.5 and P5, which is similar with the Car13 expression pattern in the type III fibrocytes of the cochlea (Fig. 3K,L,N,O, arrows).
Previously, cytosolic and membrane-bound CAs have been shown to associate directly with ion transporters, forming a membrane protein complex called a biocarbonate transport metabolon. This complex can significantly enhance H+/HCO3− transport by mass CA action supplying and dissipating the reaction products (Alvarez et al., 2003; Purkerson and Schwartz, 2007). For example, in the renal tube, a cytosolic CAII and a membrane-bound CAIV form transport metabolons with bicarbonate transporters such as AE1, NBC1, NBC3, and SCL26A6 and a proton antiporter such as NHE1 (Purkerson and Schwartz, 2007).
Based on expression of the transcripts of membrane-bound Car12, cytosolic Car13, and the anion exchanger Pendrin in the outer sulcus and root cell processes, it is possible that CAXII and CAXIII may physically associate with Pendrin or other transporters to form metabolons in the outer sulcus and root cell processes and regulate ion and pH homeostasis of endolymphatic fluid. Consistently, the unique biochemical properties of CAXIII, which catalyzes CO2 hydration even in the presence of high HCO3− concentrations, suggests that CAXIII may form metabolons in tissues that require tight bicarbonate regulation. For example, the female reproductive tract requires an alkaline environment to maintain sperm motility (Innocenti et al., 2004). Further experiments should examine whether CAXII and CAXIII indeed physically or functionally associate with Pendrin to regulate endolymph homeostasis.
A membrane-bound isozyme Car14
Car14 was broadly expressed in both epithelial and mesenchymal tissues in the cochlea at P0 (Fig. 4A,A′). At P5, Car14 expression was stronger in the otic fibrocytes, stria vascularis, and supporting cells in the organ of Corti (Fig. 4B,B′). Car14 expression was maintained in the otic fibrocytes and greatly reduced in the stria vascularis at P21 (Fig. 4C). Car14 expression in Reissner's membrane at P0 decreased by P5 (Fig. 4A′,B, white arrow). Car14 expression in the organ of Corti at P5 appeared to be in pillar cells and Deiters’ cells when compared with Smpx, a hair cell-specific gene (Fig. 4B′) (Yoon et al., 2011). At P21, Car14 expression was observed in the inner phalangeal cells and inner pillar cells located at the medial side of the tunnel of Corti, and three Deiters’ cells at the lateral side but not in the outer pillar cells (Fig. 4D). Car14 was also weakly expressed in the spiral ganglion region (Fig. 4C, arrows). In the vestibule, weak Car14 expression was broadly observed in the mesenchymal tissues in the vestibule, but not in the otic capsule (Fig. 4A, bracket).
It was shown in the kidney that the membrane-bound CAXIV (in rodents; CAIV in humans) is expressed on the luminal surface of the renal tube and facilitates bicarbonate and fluid transport by ion transporters such as NBCI (Alvarez et al., 2003; Purkerson and Schwartz, 2007). Previously, it has been suggested that CO2 in the outer hair cells is hydroxylated by means of intracellular CA in the outer hair cells, and exchanged to the endolymph or diffused to the region between the outer hair cells and Deiters’ cells (Thalmann et al., 1970; Kimura, 1975; Ikeda et al., 1992; Okamura et al., 1996). Thus, it is also possible that CAII expressed in the outer hair cells may cooperate with CAXIV expressed on the extracellular surface of Deiters’ cells to facilitate the hydroxylation and transport of CO2, which is the major product of high cellular metabolism in the hair cells, and play roles in optimal pH regulation and fluid homeostasis in the organ of Corti.
Acatalytic CA-related proteins Car8 and Car11
An acatalytic CA-RP, Car8, was first detected at P0 in the otic capsule near the basal turn of the cochlear duct and of the vestibule (Fig. 5A, arrowheads). This expression overlaps with Pannexin 3 (Panx3), which is expressed in the prehypertrophic chondrocytes, perichondrium, and osteoblasts (Fig. 5C, arrowheads) (Iwamoto et al., 2010). At P5, Car8 expression in the otic capsule was down-regulated, and was expressed instead in the modiolus, which overlaps with Panx3 (Fig. 5B,D, arrows). Car8 transcripts were not detected in the inner ear at P21, consistent with the qRT-PCR results (Fig. 1). Based on the developmental stage and ossification pattern, it appears that Car8 is expressed in prehypertrophic chondrocytes in the otic capsule at P0, and later is expressed in differentiating osteoblasts in the modiolus.
It is interesting that Car8 expression in the developing otic capsule and modiolus overlaps with Panx3, which is induced in prehypertrophic chondrocytes and promotes chondrocyte and osteoblast differentiation. Panx3 performs these roles as a Ca2+ channel in the endoplasmic reticulum (ER), as a hemichannel releasing ATP into the extracellular space, and as a gap junction propagating Ca2+ waves between cells (Iwamoto et al., 2010; Ishikawa et al., 2011). Panx3 as an ER Ca2+ channel was activated similar to Inositol trisphosphate 3 receptors (IP3Rs), which increase intracellular Ca2+ levels upon IP3 binding to and activating IP3Rs. Interestingly, Car8 was shown to reduce the affinity of IP3Rs for IP3 and inhibit increases in intracellular calcium (Hirota et al., 2003). Thus, it is possible that CAVIII plays a role in forming the otic capsule and modiolus by modulating IP3-induced Ca2+ release during osteoblast differentiation. Although this hypothesis remains to be tested, Car8 expression in the differentiating otic capsule and modiolus is interesting because only Car1 and Car2, which are expressed in bone marrow cells in otic capsule, have been known as major carbonic anhydrases associated with bone development (Lehenkari et al., 1998; Rajachar et al., 2009).
Another CA-RP, Car11, was broadly expressed in the inner ear at P5, with slightly stronger expression in the spiral ganglion (Fig. 5E, arrows). At P21, Car11 was expressed in the spiral ganglion (Fig. 5F, arrows), Reissner's membrane (Fig. 5F′, black arrowhead), organ of Corti (Fig. 5F′, white arrowhead), bone marrow of the otic capsule (Fig. 5F′, arrows), spiral prominence, and root cell processes (Fig. 5F′, bracket). Car11 was also weakly expressed in type I otic fibrocytes (Fig. 5F′, asterisk). In the vestibule, Car11 expression was observed in the apical portion of cristae and maculae, most likely in the supporting cells, based on the comparison with the hair cell specific expression of Smpx (Fig. 5G–J). Although CAXI immunoreactivity has been shown in the developing and mature human brain (Taniuchi et al., 2002) and in gastrointestinal stromal tumors (Morimoto et al., 2005), the physiological roles of Car11 in the mouse inner ear remains to be elucidated.
In this study, we systemically surveyed the expression patterns of transcripts for all 16 known mammalian CA isozymes throughout mammalian inner ear development. Our data show that not all isozymes are expressed, and each isozyme has unique temporal and spatial expression patterns during inner ear development, which are summarized in Figure 6. The expression patterns for some isozymes closely overlapped with genes previously associated with inner ear development or function, such as Pendrin, Pou3f4, or Bmp4. In addition, we also profiled the expression patterns of CA isozymes in other organs, such as brain, kidney, and liver from P21 mice, showing that each organ expresses a unique combination of CA isozymes, although some isozymes, such as Car2 and Car14, appear to be expressed ubiquitously (Fig. 1). These expression patterns suggest that each CA isozyme plays a unique role in mammalian development and function depending on its temporal and spatial expression pattern.
Tissue Dissection, RNA Isolation, and cDNA Synthesis
All animals were handled according to the guidelines for the Care and Use of Laboratory Animals of Yonsei University College of Medicine. The protocol for obtaining inner ear samples was approved by the Committee on Animal Research at Yonsei University College of Medicine. The entire inner ear tissues including the vestibule and cochlea as well as surrounding otic capsule were dissected at all stages examined. Total RNA was extracted using the RNeasy Mini kit (Qiagen, Hilden, Germany) from inner ear, brain, kidney, and liver tissues dissected from C57BL/6 mice. RNase-free DNase was used to digest of genomic DNA during RNA purification (Qiagen, Hilden, Germany). The concentration and purity of the extracted RNAs were determined using both the spectrophotometric method at 260 and 280 nm and RNA electrophoresis. One microgram of total RNA was subjected to reverse transcription with oligod(T)19 primer using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA).
Synthesized cDNAs were subjected to semi-quantitative RT-PCR to analyze relative expression levels of CA isozymes among tissues. CA isozyme specific PCR primers previously reported were used for RT-PCR (Lacruz et al., 2010). PCR conditions were as follows: 28 cycles of denaturation at 94°C for 20 sec, annealing at 55°C for 40 sec, and extension at 72°C for 40 sec. The first denaturation step and the last extension step were performed at 95°C for 15 min and 72°C for 5 min, respectively. PCR products were separated and visualized on a 2% agarose gel.
Quantitative Real-Time PCR
qRT-PCR was performed to analyze the expression of each CA isozyme during inner ear development at E15.5, P0, P5, and P21. RNA purified from three replicates of each inner ear sample was used for qRT-PCR. PCR primers were designed based on mRNA sequences in the GenBank database (Table 1). Forward and reverse primers from each primer set were designed from different exons to distinguish transcripts from genomic DNA. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as an internal control for normalization. qRT-PCR for each CA isozyme was performed in triplicate using SYBR Green PCR Master Mix and the Applied Biosystems StepOnePlus Real-Time PCR Systems (Applied Biosystems). A melting curve analysis was always performed after the amplification to check PCR specificity. The results were analyzed by StepOne Software v2.1 (Applied Biosystems). The normalization and relative quantification were calculated using 2−ΔΔCt method (Livak and Schmittgen, 2001).
Table 1. Primer Sequences for Quantitative Real-Time PCR
Forward primer (5′-3′)
Reverse primer (5′-3′)
In Situ Hybridization
For E15.5 embryos and neonatal (P0 and P5) pups, animals were killed by decapitation and then hemi-sectioned. After removing the brain, the specimens were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) overnight at 4°C, dehydrated in 30% sucrose in PBS overnight at 4°C, embedded in OCT compound (Sakura, Tokyo, Japan), and stored at −80°C until use. For P21 mice, the entire inner ear tissues including the otic capsule were dissected. For better penetration of fixatives, parts of the semicircular canals were cut off, the oval and round windows were opened up, and the apex of the cochlea was punctured. After overnight fixation, the specimens were decalcified in 0.2 M EDTA in DEPC-PBS at 4°C for 2 days, followed by dehydration and mounting. Tissues were sectioned at 12 μm thickness for in situ hybridization, which was performed as previously described (Morsli et al., 1998). At least three animals were tested for each CA isozyme at each developmental stage. Sense RNA probes were also included as controls, which showed no signal anywhere in the inner ear.
Probes for Bmp4 (Morsli et al., 1998), Pou3f4 (Phippard et al., 1998), Smpx (Yoon et al., 2011), and Pendrin (Everett et al., 1999) were prepared as previously described. RNA probes for Car1 were generated from a 381 base pair (bp) mouse Car1 cDNA containing the +594 to +786 coding region and the 188 bp 3′ untraslated region (NM_009799.4); for Car2, from a 410 bp mouse Car2 cDNA containing the +654 to +783 coding region and the 280 bp 3′untranslated region (NM_009801.4); for Car3, from a 404 bp mouse Car3 cDNA containing the +687 to +783 coding region and the 334 bp 3′ untranslated region (NM_007606.3); for Car8, from a 621 bp mouse Car8 cDNA containing the +212 to +832 coding region and the 12 bp 5′ untranslated region (NM_007592.3); for Car12, from a 471 bp mouse Car12 cDNA containing the +947 to +1,065 coding region and the 352 bp 5′ untranslated region (NM_178396.4); for Car13, from a 404 bp mouse Car13 cDNA containing the +664 to +789 coding region and the 278 bp 5′ untranslated region (NM_024495.5); for Car14, from a 856 bp mouse Car14 cDNA containing the +1 to +821 coding region and the 35 bp 5′ untranslated region (NM_011797.2); for Car11, from a 916 bp mouse Car11 cDNA containing 634 bp coding region and the 282 bp 5′ untranslated region (NM_009800.4); for Osteocalcin, from a 33 bp mouse Osteocalcin cDNA containing the 47 bp 5′ untranslated region and the entire open reading frame (NM_007541.2); for Pannexin3, from a 559 bp mouse Pannexin3 cDNA containing the +565 to +1,123 coding region (NM_172454.2).