The highly secretory Clara cells play a pivotal role in protecting the lung against inflammation and oxidative stress. This study reports the positional cloning of a novel protein required for Clara cell physiology in mouse lung development. The perinatal lethal N-ethyl-N-nitrosourea-induced l7Rn64234SB allele contained a nonsense mutation in the previously hypothetical gene NM_026304 on chromosome 7. Whereas l7Rn6 mRNA levels were indistinguishable from wild type, l7Rn64234SB homozygotes exhibited decreased expression of the truncated protein, suggesting protein instability. During late gestation, l7Rn6 was widely expressed in the cytoplasm of lung epithelial cells, whereas perinatal expression was restricted to the bronchiolar epithelium. Homozygosity for the l7Rn64234SB allele did not affect early steps in lung patterning, growth, or cellular differentiation. Rather, mutant lungs demonstrated severe emphysematous enlargement of the distal respiratory sacs at birth. Clara cell pathophysiology was evident from decreased cytoplasmic CCSP and SP-B protein levels, enlargement and disorganization of the Golgi complex, and formation of aberrant vesicular structures. Additional support for a role in the secretory pathway derived from l7Rn6 localization to the endoplasmic reticulum. Thus, l7Rn6 represents a novel protein required for organization and/or function of the secretory apparatus in Clara cells in mouse lung.
THE highly secretory Clara cells represent nonciliated columnar cells, which, along with ciliated cells, line the bronchiolar epithelium (Ten Have-Opbroek 1991). Clara cell functions include cell renewal of airway epithelium following injury, for example, from ozone exposure (Evans et al. 1976, 1978). Furthermore, these cells express CCSP (also known as uteroglobin or blastokinin), a 16-kDa homodimeric protein, which protects the respiratory tract against oxidative stress, inhibits phospholipase A2, and exhibits antiinflammatory and immunomodulatory activities (Johnston et al. 1997; Mango et al. 1998; Broeckaert et al. 2000; Singh and Katyal 2000; Chen et al. 2001; Watson et al. 2001). Other proteins produced and presumably secreted by Clara cells include phospolipase A, surfactant proteins A, B, and D (SP-A, SP-B, and SP-D) (Phelps and Floros 1991; Kalina et al. 1992; Singh and Katyal 2000), CC26, a 26-kDa selenium-independent glutathione peroxidase (Power and Nicholas 1999), as well as several members of the cell surface integrin family (Blundell and Harrison 2005).
Recent studies in humans and mice have linked abnormal levels of secretory proteins with defects in lung homeostasis. In humans, surfactant deficiency resulted in respiratory distress syndrome, the most frequent cause of lethality and morbidity in infants of <1 year of age (Nogee et al. 1994; Cole et al. 2001; Shulenin et al. 2004). Likewise, targeted ablation of the gene encoding SP-B resulted in respiratory failure in newborn mice and death shortly after birth (Clark et al. 1995). Conditional deletion of winged helix transcription factor Foxa2 caused a significant reduction in the expression of CCSP, SP-A, SP-C, and SP-B as well as emphysematous enlargement of the lungs, decreased alveolar septation, goblet cell hyperplasia, and increased neutrophil infiltration (Wan et al. 2004a,b). Interestingly, premature human newborns with bronchopulmonary dysplasia demonstrated increased CCSP oxidation and decreased CCSP protein levels in tracheal aspirate, supporting an important role of Clara cells in lung homeostasis at birth (Ramsay et al. 2001).
Here we report the positional cloning of the N-ethyl-N-nitrosourea (ENU)-induced l7Rn64234SB mutation and identify the previously hypothetical gene NM_026304 as the l7Rn6 locus. Phenotypic and marker gene expression analysis revealed a pivotal role for l7Rn6 in Clara cell function in mouse lung development. In addition to a perinatal emphysematous lung phenotype, l7Rn64234SB Clara cells demonstrated decreased cytoplasmic CCSP and SP-B protein levels, significant enlargement and disorganization of the Golgi complex, as well as aberrant vesicular structures. Since early patterning, growth, or cellular differentiation appeared unaffected, we suggest that secretory defects in Clara cells might govern the perinatal emphysematous phenotype in l7Rn64234SB homozygous lungs.
MATERIALS AND METHODS
Animals and genotyping:
The l7Rn64234SB allele used in this study was isolated in an ENU mutagenesis screen to recover (pseudo-)recessive mutations in the albino (c) region (Rinchik et al. 1990). The l7Rn64234SB allele is closely linked to c on distal chromosome 7 and was maintained on a mixed genetic background opposite a wild-type chromosome marked by the chinchilla (cch) locus, a partial loss-of-function allele at the tyrosinase locus (Rinchik and Carpenter 1993, 1999). Following identification of the l7Rn64234SB point mutation in NM_026304, primers 5′-TACAAATTTGGATTTTCTTTTCTAGTTA (forward) and 5′-GGTAGGTTTCATGACCCTTTTGAC (reverse) were used for direct genotyping. The forward primer contains a thymine instead of a guanine in the -3 position compared with wild-type sequence, which, in the context of the l7Rn64234SB T → A transversion, creates a diagnostic MseI restriction enzyme site for the mutant allele.
The transgenic rescue construct encompassed an ∼6.3-kb EcoRI fragment with the putative l7Rn6 promoter, which was cloned upstream of a KCR cassette containing intronic sequences from the rabbit β-globin locus to augment transgene expression (Ma et al. 1999). In addition, a heterologous 3′-UTR containing the bovine growth hormone polyadenylation signal (Tepera et al. 2003) was cloned downstream of the full-length l7Rn6 (NM_0263040) open reading frame (Figure 1, A and D). Two founder lines (A1tg and A4tg) were obtained from injection of the linearized construct into the pronuclei of FVB/NJ zygotes. The line with the higher expression level A4tg was crossed with l7Rn64234SB/+ animals for rescue experiments. Following one backcross onto the FVB/NJ background, l7Rn64234SB/+; A4tg animals were intercrossed to assess transgenic rescue on an ∼75% FVB/NJ background. The presence of the transgene was detected by PCR using primers 5′-GAACTTCAGGCTCCTGGGCAAC (forward) and 5′-ACCAGGCAGCCAAACATG (reverse). The Institutional Animal Care and Use Committee of Baylor College of Medicine approved all animal experiments in this study.
Genomic mapping of candidate genes:
Candidate genes were mapped to the l7Rn6 critical region by PCR amplification of the first and last exon of each candidate gene from several yeast artificial chromosomes (YACs) spanning the 150-kb candidate region (Holdener et al. 1995). The following primers were used: 5′-TTAGGAAGGAAGGGTCAGAGGTCG (forward) and 5′-ACCAGGCAGCCAAACATG (reverse) for the first exon of l7Rn6 (NM_026304), 5′-CACAGAACCCCCTCTTTTGG (forward) and 5′-GCCTCCAAGAAGTATGTCACAC (reverse) for the last exon of l7Rn6 (NM_026304), 5′-CTTCTTAGTTCAGCAACCCTCC (forward) and 5′-GCACTAAGGTGACTTTCCCAC (reverse) for the first exon of AK014876, and 5′-GCTGTTTCTCTTGGATCAGTG (forward) and 5′-CTCGGGAAATAGTAGGTTCAAGCG (reverse) for the last exon of AK014876.
cDNA cloning and mRNA expression analysis:
Total RNA was isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA). Five micrograms of total RNA were reverse transcribed using oligo (dT) priming and M-MLV reverse transcriptase (Invitrogen) as described (Mok et al. 2004a). Primers used for cDNA cloning included: 5′-TTAGGAAGGAAGGGTCAGAGGTCG (forward) and 5′-GCCTCCAAGAAGTATGTCACAC (reverse) for l7Rn6 (NM_026304), as well as 5′-GTGGGAAAGTCACCTTAGTGC (forward) and 5′-CTCGGGAAATAGTAGGTTCAAGCG (reverse) for AK014876. Expression levels of the transgenic lines A1tg and A4tg and endogenous l7Rn6 (NM_026304) were analyzed by semiquantitative RT-PCR from brain, lung, heart, liver, kidney, and muscle. The following primers were used: 5′-GTTGGACAGTCTGGCTCAGCAG (forward) and 5′-GGTAGGTTTCATGACCCTTTTGAC (reverse) for l7Rn6 (NM_026304), 5′-AGCCATGTACGTAGCCATCC (forward) and 5′-TTTGATGTCACGCACGATTT (reverse) for β-actin, and 5′-TGCGGAATTCCTGCAGCC (forward) and 5′-CACCCCATTTGAATCAGGATAGGA (reverse) for the transgene.
Lungs were dissected between embryonic day 15.5 (E15.5) and postnatal day 0 (P0) and fixed overnight in 4% paraformaldehyde. P0 lungs were inflated with paraformaldehyde; lung weight was ascertained before inflation. Following fixation, lungs were dehydrated in a graded series of ethanol, embedded in paraffin, and sectioned at 5 μm. Airway dilation was quantified by determination of mean linear intercepts as described previously (Dunnill 1962). Briefly, six horizontal lines were drawn onto digital images, taken at ×200 magnification, from randomly selected, hematoxylin-eosin (H&E)-stained P0 lung sections. The number of times each line crossed a saccular wall was counted. Results are expressed as the total number of intercepts divided by the total length of the six lines. Five independent lung sections per animal, and at least six animals per genotype were analyzed. All data are reported as the means ± SD. Statistical analysis was performed using Student's unpaired t-test. P0 lungs from l7Rn64234SB/4234SB and wild-type littermates were analyzed by electron microscopy. Dissected lungs were inflated with phosphate-buffered 2.5% glutaraldehyde solution and fixed for 12 hr at 4°. Samples were processed and analyzed as described previously (Ray et al. 1996).
Immunohistochemistry and mRNA in situ hybridization:
A polyclonal αl7Rn6 antiserum was raised in rabbits against amino acids 20-34 (EDKFVFDLPDYENINH) and purified by immunoaffinity. Antibodies were used at the following dilutions: αl7Rn6 at 1:400, αSP-A (Chemicon) at 1:200, αSP-B (Chemicon) at 1:200, αSP-C (Chemicon) at 1:250, αCCSP (Santa Cruz Biotechnology) at 1:100, αCGRP (Santa Cruz Biotechnology) at 1:200, and αCC26 (Chemicon) at 1:500. E15.5 embryos as well as dissected E16.5-P0 lungs were processed for immunohistochemistry as described previously (Mok et al. 2004b). Sections were incubated with biotinylated secondary antibodies and horseradish peroxidase-conjugated streptavidin using the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA). Vector NovaRED substrate (Vector Laboratories) was used for signal detection. Sections from both l7Rn6+/+ and l7Rn64234SB/4234SB lungs were mounted side by side on the same slide to allow for comparison of staining intensity.
For mRNA in situ hybridization on lung sections, digoxigenin-11-UTP-labeled antisense cRNA probes were prepared as described (Wilkinson and Nieto 1993). A 0.5-kb ccsp mRNA in situ hybridization probe was kindly provided by F. DeMayo. Sense ccsp cRNA probe served as a negative control. mRNA in situ hybridization to sectioned lungs was performed as described previously with minor modifications (Neubüser et al. 1995). Briefly, 5-μm parasagittal sections were hybridized overnight at 65° with digoxigenin-11-UTP-labeled antisense cRNA probes in a buffer containing 50% deionized formamide, 5× SSC, 50 μg/ml heparin, 0.1% Tween-20, 100 μg/ml tRNA, 5% dextran sulfate, and 8 mm citric acid. Following hybridization, sections were washed and incubated with an anti-digoxygenin antibody (Roche). To ensure comparability, the antisense and sense probe experiments were conducted under the same conditions, including the duration of the alkaline phosphatase detection reaction.
Western blot analysis:
Lungs were homogenized in lysis buffer containing 20 mm Tris, pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Trixton X-100, and EDTA-free protease inhibitor (Complete, Roche). Protein lysates were quantified using BIORAD DC Protein Assay kit (Bio-Rad). The samples were mixed with equal volume of 2× sample buffer containing β-mercaptoethanol and boiled for 5 min. Twenty-five micrograms of protein were loaded per lane and separated on 10 or 15% SDS-PAGE, followed by transfer to PVDF membrane (Bio-Rad). Membranes were blocked by incubation in 5% BSA in TBS containing 0.1% Tween 20 for 2 hr at room temperature (RT), and then incubated with primary antibodies for 12 hr at 4°. Antibodies were used at the following dilutions: αl7Rn6 at 1:4000, αMAC3 (BD Biosciences) at 1:2000, and anti-β-Actin (Santa Cruz Biotechnology) at 1:2000. Membranes were incubated with horseradish peroxidase-conjugated secondary antibodies at 1:5000, and signals were developed using Enhanced Chemiluminescence System (Santa Cruz Biotechnology).
Indirect immunofluorescence microscopy:
A549 lung carcinoma cells were grown overnight at 37° on slides coated with poly(D)-lysine (BD Biosciences) and collagen (Sigma, St. Louis). Slides were fixed in 100% prechilled methanol for 5 min at −20° and rehydrated in PBS for 10 min at RT. Following blocking with 5% donkey or goat serum for 30 min at RT, cells were incubated overnight with sets of primary antibodies: αl7Rn6 (NM_026304) (1:50 dilution)/αCCSP (Santa Cruz; 1:50), αl7Rn6 (NM_026304) (1:50)/αCalnexin (Chemicon; 1:500), αl7Rn6 (1:50)/αGS28 (BD Biosciences; 1:50), and αl7Rn6 (NM_026304) (1:50)/anti-α-SNAP (Stressgen; 1:50). The cells were washed three times with PBS and incubated with fluorophore-conjugated secondary antibodies at 1:250 dilution for 1 hr at RT. The secondary antibodies included Cy3-conjugated donkey anti-goat-IgG, Cy3-conjugated goat-anti-mouse IgG, Alexa 488-conjugated goat-anti-rabbit IgG (all from Jackson ImmunoResearch), and Alexa 488-conjugated donkey-anti-rabbit IgG (Molecular Probes, Eugene, OR). Slides were mounted with Vectashield mounting medium (Vector). Confocal microscopy was performed on an LSM 510 laser scanning microscope system (Zeiss, Thornwood, NY) equipped with argon and HeNe laser at an excitation of 488 and 543 nm, respectively. Slides were viewed with ×63 microscope objectives. Images were collected and processed with LSM 5 system software (Zeiss) and edited with Image J, AxioVisionAC, and Adobe Photoshop software packages.
Perinatal lethality in l7Rn64234SB homozygotes:
Homozygosity for the l7Rn64234SB allele was largely compatible with embryonic development and ∼80% of the mutants survived to birth (data not shown). However, the vast majority of l7Rn64234SB homozygotes died within 24 hr after birth and all pups were dead within 48 hr. Mutant pups frequently demonstrated cyanosis shortly before death. Histological analysis did not reveal any gross abnormalities or patterning defects in vital organs and tissues, such as brain, heart, liver, kidney, and skeletal muscle (data not shown). Furthermore, metabolic screening for abnormal urine organic acids and serum amino acids did not detect any significant differences between l7Rn64234SB/4234SB and wild-type pups at birth (data not shown).
Positional cloning of l7Rn6:
Two complementation groups, l7Rn5 and l7Rn6, were previously mapped to a 150-kb interval delineated by the c3H and cAL deletion breakpoints (Rinchik and Carpenter 1993; Holdener et al. 1995) (see also Figure 1A). The l7Rn5 complementation group encompassed three alleles of the Polycomb-group gene Eed (Schumacher et al. 1996). In contrast, l7Rn6 was represented by a single allele, l7Rn64234SB. The strategy for cloning the l7Rn6 locus involved identification of candidate transcription units in the vicinity of Eed from draft versions of the mouse and human genome sequence as well as EST databases. l7Rn6 candidate genes were mapped relative to the critical 150-kb interval using YACs spanning the region (Holdener et al. 1995). Two candidate genes encoding hypothetical proteins were identified in the l7Rn6 region: AK014876 and NM_026304 (Figure 1A). Following PCR amplification from l7Rn64234SB/4234SB DNA, candidate genes were analyzed for sequence alterations absent from coisogenic l7Rn5 alleles (Rinchik and Carpenter 1993). The coding region and untranslated regions of AK014876 displayed wild-type sequences (data not shown). In contrast, a T → A transversion was detected at thymine 543 of the NM_026304 coding region in l7Rn64234SB (Figure 1B). The T → A transversion encompassed the third nucleotide of codon 181 (TAT → TAA) and substituted a tyrosine residue with a stop codon (Y181X), truncating the carboxy-terminal 17 amino acids of the l7Rn6 protein (Figure 2). T → A transversions represent the most common type of ENU-induced mutations in the mouse germline, constituting approximately one-third of all mutations identified thus far (Noveroske et al. 2000). Importantly, the three co-isogenic l7Rn5 alleles exhibited the wild-type thymine 543 in the NM_026304 locus (Figure 1B and data not shown). Furthermore, the T → A transversion cosegregated invariably with the l7Rn64234SB mutant phenotype. No other sequence alterations were found and NM_026304 is herein referred to as the l7Rn6 gene. On the basis of the University of California at Santa Cruz mouse genome assembly mm5 (build m33 NCBI) and consistent with our YAC mapping (data not shown), l7Rn6 is located ∼13 kb proximal to Eed and is composed of five exons, which span 23 kb in genomic DNA (Figure 1A). No gene sequences homologous to l7Rn6 were identified in the mouse genome (data not shown). RT-PCR analysis indicated l7Rn6 mRNA expression in all tissues analyzed, including brain, lung, heart, liver, kidney, and muscle (Figure 1C).
Partial transgenic rescue:
The l7Rn6 transgenic founders were generated on a pure FVB/NJ background and potential rescue of the l7Rn64234SB lung phenotype was assessed on a 75% FVB/NJ background. Surprisingly, homozygosity for the l7Rn64234SB mutation appeared embryonic lethal on this background since no l7Rn64234SB/4234SB pups were recovered at birth. In contrast, 24 l7Rn64234SB/4234SB;A4tg pups were recovered among 90 pups at P0 from intercrosses between l7Rn64234SB/+;A4tg animals. Ten of these l7Rn64234SB/4234SB;A4tg pups were found dead at P0 and 7 died before P2. The remaining 7 animals died between P7 and P21. They appeared slightly runted at P0 but displayed grossly normal development over the next few days. Approximately 2 days before death, l7Rn64234SB/4234SB;A4tg animals became increasingly runted and lethargic. Thus, the A4tg transgenic line partially rescued the prenatal lethality of l7Rn64234SB homozygotes on a largely FVB/NJ background, confirming NM_026304 as the l7Rn6 locus. The partial rescue is consistent with the relatively low expression level of both transgenic lines compared with endogenous l7Rn6 (Figure 1C and data not shown), suggesting that the ∼6.3-kb genomic promoter region lacked regulatory elements and/or was sensitive to repressive positional effects.
Protein sequence analysis of l7Rn6:
l7Rn6 encodes a 197-aa protein, which is highly conserved from yeast to mammals. Protein sequence identity ranges from 77 to 100% among vertebrates (Figure 2). Sequence alignment using ClustalW 1.74 and 1.82 revealed four highly conserved subsequences, mostly located in the amino-terminal half of the protein (Figure 2). The WQLLGFVTNGKPSAIFKI motif displays the highest evolutionary conservation, most prominently in the KPSAI subsequence, which exhibits 100% identity across all species, constituting a distinctive signature of this novel protein. PSORT II (Nakai and Kanehisa 1992) identified the FWKT sequence at the carboxy terminus of l7Rn6 as a KKXX-like, endoplasmic reticulum membrane retention signal. Other protein motifs, including nuclear localization signals, DNA-binding domains, or transmembrane domains, were not found (data not shown).
l7Rn64234SB homozygotes display emphysematous lungs at birth:
l7Rn64234SB/4234SB lungs were remarkable for a discrete developmental delay during the canalicular stage (E16.5) (data not shown). At the beginning of the saccular stage (E17.5), l7Rn64234SB/4234SB lung saccules appeared not fully dilated and resembled E16.5 wild-type lungs (Figure 3A and data not shown). However, as development proceeded, the delay became less evident (data not shown). At P0, the lung:body weight ratio revealed no statistically significant difference between l7Rn64234SB/4234SB and l7Rn6+/+ mice (P > 0.05) (Figure 3B). Furthermore, Periodic Acid Schiff (PAS) glycogen staining, a marker for perinatal lung maturity (Compernolle et al. 2002), detected no differences in lung glycogen content between l7Rn64234SB/4234SB and wild-type littermates at P0 (data not shown).
Serial sections of lungs from l7Rn64234SB homozygotes demonstrated a significant emphysematous phenotype at birth (Figure 3A). The extent of distal airway dilation was quantified by analysis of the mean linear intercepts (Dunnill 1962) in l7Rn6+/+ and l7Rn64234SB/4234SB lungs. At P0, l7Rn64234SB/423SB showed a statistically significant enlargement of the distal airways, with a mean linear intercept of 52.1 ± 6.2 μm compared with 39.8 ± 3.1 μm in l7Rn6+/+ pups (P < 0.01) (Figure 3B). In the context of the emphysematous lung phenotype and frequent cyanosis in l7Rn64234SB homozygotes, we attribute the perinatal lethality to respiratory failure.
As demonstrated by Western blot analysis and immunohistochemistry for MAC3, a marker for alveolar macrophages (Glasser et al. 2003), the distribution and number of macrophages appeared indistinguishable between l7Rn64234SB/4234SB and wild-type littermates (Figure 3C and data not shown). In addition, mutant lungs were devoid of inflammatory infiltrates (Figure 3A).
l7Rn6 expression and instability of the l7Rn64234SB protein:
E15.5 embryos displayed cytoplasmic l7Rn6 protein expression in multiple tissues, including heart, kidney, adrenal gland, pituitary, submandibular glands, intestine, trachea, and lung (Figure 4A and data not shown). In the developing lung, l7Rn6 expression was detected in epithelial cells between E15.5 and E17.5 (Figure 4B and data not shown). At P0, l7Rn6 expression appeared restricted to the bronchiolar epithelium, including Clara cells (Figure 4B). There was no evidence for l7Rn6 expression in type II cells at P0 (Figure 4B). Western blot analysis confirmed significant l7Rn6 expression at E14.5-E16.5, which decreased toward late gestation (Figure 4C). At P0 as well as in adult lungs, l7Rn6 expression levels ranged barely above detection limits. Strikingly, immunohistochemistry and Western blot analysis, respectively, demonstrated a significant reduction and absence of l7Rn6 expression in l7Rn64234SB homozygous lungs (Figure 4, B and C). l7Rn6 mRNA levels were indistinguishable between P0 wild-type and l7Rn64234SB/4234SB lungs by semiquantitative RT-PCR (Figure 4D). Thus, the significant decrease in l7Rn6 protein levels in l7Rn64234SB homozygous lungs did not appear to result from nonsense-mediated mRNA decay, which frequently eliminates mRNAs with premature termination codons (Baker and Parker 2004). Additional evidence for l7Rn64234SB protein instability derived from transfection experiments in COS7 and HEK293 cells. Whereas tagged wild-type protein displayed abundant expression, a construct containing a cDNA truncated at the mutant stop codon consistently yielded significantly lower levels of l7Rn64234SB expression (data not shown). Therefore, the ENU-induced point mutation appeared to render the mutant protein unstable.
Normal cellular composition in l7Rn64234SB homozygous lungs:
To characterize the various lung cell types between E15.5 and P0, we studied marker gene expression by immunohistochemistry on sections from wild-type and l7Rn64234SB/4234SB lungs at P0. No differences in expression of SP-C and CGRP were detected between wild-type and l7Rn64234SB homozygotes, suggesting that type II cells and neuroendocrine cells were present in normal numbers and distribution at P0 (Figure 5). Immunohistochemistry for CC26, a selenium-independent glutathione peroxidase expressed specifically in Clara cells (Power and Nicholas 1999), detected no difference in expression between wild-type and mutant lungs (Figure 5). Furthermore, there was no difference in the number of Clara cells expressing SP-A (Figure 5). On the basis of expression of MAC-3 and α-smooth muscle actin, the macrophage and alveolar myofibroblast cell population, respectively, also appeared indistinguishable between l7Rn64234SB/4234SB and wild-type littermates (Figure 3C and data not shown). Thus, these results suggest that the l7Rn64234SB mutation did not affect the cellular composition of the lung.
A defect in Clara cell function in l7Rn64234SB homozygous lungs:
Strikingly, expression levels of cytoplasmic CCSP and SP-B were significantly decreased in Clara cells from l7Rn64234SB/4234SB lungs compared with wild-type littermates (Figure 5). The decrease in CCSP and SP-B expression was consistently detected in sections from l7Rn64234SB/4234SB lungs compared with l7Rn6+/+ sections mounted side by side on the same slide. To ascertain a potential mechanism of decreased cytoplasmic CCSP expression, we performed mRNA in situ hybridization with a ccsp antisense probe. As demonstrated in Figure 5, ccsp mRNA levels and distribution in wild-type and mutant lungs were indistinguishable at P0. In addition, semiquantitative RT-PCR revealed no differences in ccsp mRNA levels between wild-type and mutant lungs (Figure 4D). These data implicate an unknown post-transcriptional, presumably post-translational mechanism in decreased expression of cytoplasmic CCSP.
Taken together, the l7Rn64234SB mutation disrupted Clara cell function as evidenced by decreased levels of CCSP and SP-B expression. However, a general defect in protein expression in Clara cells appeared less likely since CC26 and SP-A expression levels were indistinguishable between wild-type and l7Rn64234SB mutant lungs (Figure 5).
Subcellular localization of l7Rn6:
Indirect immunofluorescence in A549 lung carcinoma cells detected largely punctate l7Rn6 signal throughout the cytoplasm. l7Rn6 did not colocalize with GS28, which marks the cis-Golgi complex (Subramaniam et al. 1996), in A549, HEK293, and COS7 cells (Figure 6 and data not shown). However, l7Rn6 exhibited significant colocalization in the perinuclear region with Calnexin, a marker for the endoplasmic reticulum (Wada et al. 1991) (Figure 6). Interestingly, l7Rn6 and CCSP did not colocalize in A549 cells (Figure 6), despite concomitant reduction of their protein levels in l7Rn64234SB/4234SB lungs (Figure 4B and C).
Clara cell abnormalities in l7Rn64234SB homozygotes:
Electron microscopy revealed an overall normal cellular organization and glycogen content of Clara cells in l7Rn64234SB homozygotes compared with wild type at P0 (Figure 7A). However, mutant Clara cells were remarkable for defects in the secretory apparatus, including enlargement and disorganization of the Golgi complex (Figure 7, A and B). Furthermore, abnormal vesicular structures of varying size and number were frequently detected in the cytoplasm of l7Rn64234SB homozygous Clara cells (Figure 7, A and C). These structures contained circular membranes and glycogen-like particles and were commonly found in close apposition to the Golgi complex (Figure 7, A and C). Mutant Clara cells presented no other defects in the cytoplasm or the nucleus. Other lung epithelial cells, such as type II cells, were devoid of ultrastructural abnormalities, including the secretory apparatus (data not shown).
In recent years, significant progress has been made in dissecting the genetic control of mammalian lung development. Many transcription factors, peptide growth factors and their receptors, as well as extracellular matrix components have been identified as important regulators of lung morphogenesis in reverse genetics approaches (Warburton et al. 2000; Groenman et al. 2004; Kumar and Ryan 2004). Chemical mutagenesis using the potent germline mutagen ENU represents a powerful means to complement these gene-driven approaches by phenotype-based screens in mice (De Angelis et al. 2000; Nolan et al. 2000; Kile et al. 2003). This report describes the positional cloning of the perinatal lethal, ENU-induced l7Rn64234SB allele and identifies a pivotal role for the previously hypothetical gene NM_026304 in Clara cell function during mouse lung development. The ENU-induced nonsense mutation truncated the carboxy-terminal 17 aa, rendering the l7Rn64234SB protein unstable. However, detection of residual amounts of the mutant protein by immunohistochemistry suggested that l7Rn64234SB constitutes a strong hypomorphic rather than a null allele. Partial transgenic rescue and absence of nucleotide alterations in all other known or hypothetical genes in the 150-kb critical region revealed NM_026304 as the causative gene for the l7Rn64234SB phenotype.
l7Rn6 protein expression was detected in several epithelial and secretory tissues during embryonic development, including kidney, adrenal gland, intestine, trachea, and lung. Evidence for abnormal Clara cell function derived from a significant decrease in intracellular CCSP and SP-B levels. Since expression levels of CC26 and SP-A were indistinguishable between wild-type and l7Rn64234SB homozygotes, a general decrease in proteins secreted by Clara cells appeared less likely. Interestingly, previous yeast two-hybrid screening identified physical interaction of YOL032W and Sec17, the Saccharomyces homolog of l7Rn6 and α-SNAP, respectively (Uetz et al. 2000; Hazbun et al. 2003). α-SNAP governs a crucial step in vesicle fusion (Bonifacino and Glick 2004) and has recently been implicated in surfactant secretion in type II pneumocytes (Abonyo et al. 2003). However, we could not detect colocalization of l7Rn6 and α-SNAP in COS7, HEK293, and A549 cells (data not shown).
KKXX motif-containing proteins typically reside in the endoplasmic reticulum (ER) and participate in the retrograde transport of COPI vesicles from the Golgi complex to the ER (Lee et al. 2003; Stornaiuolo et al. 2003; Bonifacino and Glick 2004). A carboxy-terminal KKXX-like motif in the l7Rn6 protein, which is ablated by the l7Rn64234SB nonsense mutation, partial colocalization with Calnexin, and decreased levels of secreted proteins such as CCSP and SP-B, portend a role for l7Rn6 in the secretory apparatus. Indeed, aberrant vesicular structures, filled with membranous and glycogen-like material, were frequently detected in close apposition to an enlarged and disorganized Golgi complex in Clara cells from l7Rn64234SB homozygotes. Thus, this study established l7Rn6 as a novel protein with a potential function in vesicular trafficking between the ER and Golgi complex.
Phenotypic and marker gene expression analysis revealed that homozygosity for the l7Rn64234SB allele did not appear to affect early lung patterning, growth, or cellular differentiation. Rather, l7Rn64234SB/4234SB lungs demonstrated an emphysematous enlargement of the terminal respiratory sacs. Emphysematous phenotypes have been described in multiple mouse models and various disease mechanisms have been identified. For example, mice doubly homozygous for mutant alleles of Fgfr3 and Fgfr4 showed an emphysematous phenotype and lack of alveologenesis due to failure in secondary septation and increased elastin deposition (Weinstein et al. 1998). Pdgfa-deficient mice exhibited atelectasis and emphysema secondary due to a failure in alveolar septation, which, in turn, resulted from a loss of alveolar myofibroblasts and parenchymal elastin deposition (Boström et al. 1996, 2002). Likewise, loss of function of the tissue inhibitor of metalloproteinases-3 (Timp3) led to an emphysematous phenotype by the second week of life, enhanced degradation of collagen in the peribronchiolar space, and decreased bronchiolar branching (Leco et al. 2001; Gill et al. 2003). Whereas most of these mouse mutants developed emphysematous phenotypes with age, loss of function of Fibrillin-1 caused emphysema formation as early as P1 (Neptune et al. 2003). Given similar early-onset emphysema in l7Rn64234SB homozygotes, we considered potential defects in extracellular matrix components as a contributing factor in disease pathogenesis. However, preliminary analysis revealed no differences in parenchymal elastin content and tissue distribution of myofibroblasts between wild-type and l7Rn64234SB homozygous lungs (data not shown). Furthermore, the perinatal death of l7Rn64234SB homozygotes preceeds alveolar septation, which occurs between P5 and P30 in mice (Ten Have-Opbroek 1991; Warburton et al. 2000). This renders a septation defect as a contributing factor in the pathogenesis of the emphysematous lung phenotype less likely.
Several mouse mutants exhibited pulmonary emphysema in the presence of inflammation. For example, overexpression of Pdgfb caused emphysematous lesions, inflammation, as well as lung fibrosis (Hoyle et al. 1999). Loss of integrin αvβ6 (Itgb6)-mediated TGF-β activation caused age-related emphysema, inflammation, and upregulation of the macrophage metalloelastase Mmp12 (Morris et al. 2003). Mice deficient for SP-C developed progressive emphysema, type-II cell hyperplasia, pneumonitis, and upregulation of several metalloproteinases, including Mmp12 (Glasser et al. 2003). However, analysis of Timp3-mutant mice and double homozygotes for null alleles of Itgb6 and Mmp12 demonstrated that inflammation is not required for the development of emphysema (Leco et al. 2001; Gill et al. 2003; Morris et al. 2003). Indeed, the emphysematous enlargement of the respiratory sacs in l7Rn64234SB/4234SB lungs was devoid of inflammatory infiltrates and changes in the tissue distribution or number of alveolar macrophages.
This raises the question as to whether the emphysematous enlargement of the terminal respiratory sacs and perinatal death of l7Rn64234SB homozygotes could result from a secretory defect in Clara cells. In support of this notion, deletion of Foxa2 resulted in emphysematous enlargement of the lungs in the context of decreased expression levels of several secreted proteins, including CCSP, SP-A, SP-C, and SP-B (Wan et al. 2004a,b). SP-B deficiency in mice caused respiratory failure and death shortly after birth (Clark et al. 1995). Interestingly, these mutants also demonstrated reduced levels of SP-C, an abnormal Golgi apparatus and large, atypical intracellular membranous inclusions in cells of the respiratory epithelium. Likewise, Clara cells in ccsp-deficient mice revealed an enlarged Golgi complex and concentric whorls of endoplasmic reticulum (Stripp et al. 2002), although animal viability was unaffected in two different mutant lines (Stripp et al. 1996; Zhang et al. 1997). Thus, l7Rn64234SB homozygotes share certain phenotypic features with these mutants and decreased levels of CCSP and SP-B expression might indeed play a significant role in the development of the perinatal emphysematous enlargement and lung failure.
We are grateful to Terry Magnuson and Gene Rinchik for providing the l7Rn64234SB mutant strain. We thank James Barrish (Department of Pathology, Texas Children's Hospital) for electron microscopy, Richard Atkinson for help with confocal microscopy, Gerard Karsenty for generous microscope access, and Miriam Mendoza for technical assistance. Francesco DeMayo as well as members of our laboratory are acknowledged for helpful discussions and critical reading of the manuscript. This work was supported by a research grant from the National Institutes of Health to A.S.
↵1 Present address: Department of Molecular and Cell Biology, Baylor College of Medicine, Houston, TX 77030.
Communicating editor: K. V. Anderson
- Received July 24, 2005.
- Accepted September 6, 2005.
- Copyright © 2006 by the Genetics Society of America