Drosophila Pkd2 is haploid-insufficient for mediating optimal smooth muscle contractility.

Humans heterozygous for PKD1 or PKD2 develop autosomal dominant polycystic kidney disease, a common genetic disorder characterized by renal cyst formation and extrarenal complications such as hypertension and vascular aneurysms. Cyst formation requires the somatic inactivation of the wild type allele. However, it is unknown whether this recessive mechanism applies to life-threatening vascular aneurysms, which could involve weakening of the endothelial lining or surrounding vascular smooth muscle cells (SMCs). Drosophila Pkd2 at 33E3 (Pkd2) encodes a PKD2 family of Ca(2+)-activated Ca(2+)-permeable cation channels. We show here that loss-of-function Pkd2 mutations severely reduced the contractility of the visceral SMCs, which was restored by expressing wild type Pkd2 cDNA via a muscle-specific Gal4 driver. The effect of Pkd2 mutations alone on the skeletal muscle was minimal but was exacerbated by ryanodine-induced perturbation of intracellular Ca(2+) stores. Consistent with this, Pkd2 interacted strongly with a ryanodine receptor mutation, causing a synergistic reduction of larval body wall contraction rate that is normally regulated through Ca(2+) oscillation during excitation-contraction coupling in the skeletal muscle. These results suggest that PKD2 cooperates with the ryanodine receptor to promote optimal muscle contractility through intracellular Ca(2+) homeostasis. Further genetic analysis indicated that Pkd2 is strongly haploinsufficient for normal SMC contractility. Since Ca(2+) homeostasis is a conserved mechanism for optimal muscle performance, our results raise the possibility that inactivation of just one PKD2 copy is sufficient to compromise vascular SMC contractility and function in PKD2 heterozygous patients, thus explaining their increased susceptibility to hypertension and vascular aneurysms.

Humans heterozygous for PKD1 or PKD2 develop autosomal dominant polycystic kidney disease, a common genetic disorder characterized by renal cyst formation and extrarenal complications such as hypertension and vascular aneurysms. Cyst formation requires the somatic inactivation of the wild type allele. However, it is unknown whether this recessive mechanism applies to life-threatening vascular aneurysms, which could involve weakening of the endothelial lining or surrounding vascular smooth muscle cells (SMCs). Drosophila Pkd2 at 33E3 (Pkd2) encodes a PKD2 family of Ca 2؉ -activated Ca 2؉permeable cation channels. We show here that loss-offunction Pkd2 mutations severely reduced the contractility of the visceral SMCs, which was restored by expressing wild type Pkd2 cDNA via a muscle-specific Gal4 driver. The effect of Pkd2 mutations alone on the skeletal muscle was minimal but was exacerbated by ryanodine-induced perturbation of intracellular Ca 2؉ stores. Consistent with this, Pkd2 interacted strongly with a ryanodine receptor mutation, causing a synergistic reduction of larval body wall contraction rate that is normally regulated through Ca 2؉ oscillation during excitation-contraction coupling in the skeletal muscle. These results suggest that PKD2 cooperates with the ryanodine receptor to promote optimal muscle contractility through intracellular Ca 2؉ homeostasis. Further genetic analysis indicated that Pkd2 is strongly haploinsufficient for normal SMC contractility. Since Ca 2؉ homeostasis is a conserved mechanism for optimal muscle performance, our results raise the possibility that inactivation of just one PKD2 copy is sufficient to compromise vascular SMC contractility and function in PKD2 heterozygous patients, thus explaining their increased susceptibility to hypertension and vascular aneurysms.
Autosomal dominant polycystic kidney disease (ADPKD) 1 is a common genetic disease affecting ϳ 1 ⁄1,000 of the general population (1). Mutations in two disease-causing genes, PKD1 and PKD2, account for nearly all incidences of ADPKD. The main characteristic of ADPKD is progressive cyst formation in the kidney that results in renal failure at high frequencies (2). In addition, there are many extrarenal complications in ADPKD such as cyst formation in the liver and pancreas (3,4). ADPKD is also a disease of the blood vessels (5). Systemic cardiovascular abnormalities include hypertension, cardiac valve defects, colonic diverticulae, and vascular aneurysms of the intracranial, coronary, aortic, thoracic, and splenic blood vessels (1,2,5). Sudden intracranial aneurysms sometimes occur in asymptomatic young adults with PKD1 or PKD2 mutation, and this significantly contributes to elevated morbidity and mortality associated with ADPKD (1,5).
ADPKD patients are heterozygous individuals (PKDϩ/Ϫ) with respect to the PKD1 or PKD2 locus. Molecular analyses of cyst-forming cells in the kidney indicate that they are homozygous mutant (PKDϪ/Ϫ) clones resulting from the inactivation of the wild type gene copy by somatic mutation (6 -8). Analogous to mutations in tumor suppressor genes such as the retinoblastoma gene, ADPKD is dominantly transmitted from one generation to the next, but cyst formation involves a recessive mechanism that requires a second gene inactivation event. This mechanism of cystogenesis is supported by a transgenic mouse model (Pkd2 WS25 /Ϫ) that involves an unstable Pkd2 WS25 allele (9). In this case, Pkd2 WS25 /Ϫ mice developed kidney cysts earlier and faster than Pkd2 WS25 /Pkd2 WS25 mice. Renal cysts formed in Pkd2 WS25 /Ϫ mice invariably arise from renal tubular cells that do not produce PKD2 protein (9). Thus, somatic inactivation of PKD2 expression from both genomic copies is necessary and sufficient to trigger renal cyst formation. However, it is questionable whether this "two-hit" recessive mechanism explains many of the extrarenal phenotypes and cardiovascular pathologies of ADPKD. Some of the extrarenal phenotypes such as cardiac valve defects may originate in development early in life. Adult Pkd2ϩ/Ϫ mice show reduced long term survival in the absence of cystic disease or renal failure, indicating that a 50% reduction of gene dosage may be sufficient to cause some of the abnormalities associated with ADPKD (9).
The Drosophila genome has four Pkd2 or Pkd2-like genes (CG6504, CG13762, CG9472, and CG16793). CG6504, also known as Pkd2 (10) or almost there (11), was analyzed in this study. The protein encoded by CG6504 is most closely related to the human PKD2 family of proteins (ϳ40% amino acid identity to PKD2 and 39% identity to PKD2L). The remaining three fly PKD2-like proteins show 19 -27% amino acid identity to human PKD2 or PKD2L (11). CG6504 encodes a protein of 924 amino acids that contains essential features of the PKD2 channel family, including six transmembrane segments, a poreforming sequence (P-loop), and a C-terminal coiled-coil domain. All mammalian PKD2 proteins examined exhibit non-selective cation channel activity (12)(13)(14)(15)(16). Expression of the cloned CG6504 Pkd2 cDNA (GenBank TM accession number AY283154) in Drosophila cultured S2 cells that do not express the endogenous Pkd2 also produces a Ca 2ϩ -activated cation current, 2 suggesting that the fly PKD2 also has the capacity to function as a calcium-activated cation channel. Mammalian PKD2 has been shown to localize on primary cilia of many cell types, and cilia are involved in sensing a variety of extracellular chemical or mechanical signals (17)(18)(19). For example, ciliaassociated PKD2 of mouse kidney epithelia has been shown to mediate calcium influx in response to fluid flow (17). Similar to these, the fly Pkd2 is abundantly present on sperm flagella, a type of motile cilia (10,11). Pkd2 mutant sperm are motile but fail to congregate at the anterior uterus (10) and the storage organs in the female (10,11), which results in a severe reduction of male fertility. Thus, it appears that the fly PKD2 functions on sperm flagella to sense directional cues in the female reproductive tract.
Besides their association with cilia, mammalian PKD1 and PKD2 are expressed in vascular smooth muscle cells (SMCs) (20 -23). Vascular fragility and leakage have been reported in Pkd1Ϫ/Ϫ and Pkd2Ϫ/Ϫ mouse models (20,24,25). In Pkd2ϩ/Ϫ mouse vascular SMCs, there is a small (17%) but statistically significant reduction of intracellular calcium (26). While these phenotypes implicate the vascular smooth muscle cells, it is not known whether Pkd2 has specific functional roles in these cells.
Here we show by in vivo functional assays that the fly Pkd2 (CG6504) is haploinsufficient for optimal smooth muscle contractility. Mutations in the fly Pkd2 also weakly affected the contractility of skeletal muscles. Furthermore we show evidence that the fly Pkd2 mediates muscle contractility via calcium homoeostasis through a functional coupling with the calcium-releasing ryanodine receptor channel on the sarcoplasmic reticulum. Our results are consistent with the recent finding that Ca 2ϩ influx through ciliary PKD2 in the kidney epithelia is accompanied by Ca 2ϩ release from intracellular Ca 2ϩ storage organelles through the ryanodine receptor (17). Propagation of calcium signaling through coupling of PKD2 with the ryanodine receptor may represent an evolutionarily conserved aspect of PKD2 function. Haploid insufficiency of PKD2 for optimal smooth muscle contractility provides a plausible explanation for systemic vascular complications such as hypertension and aneurysms observed in ADPKD patients. Fig. 1A, stage-specific mRNA samples were isolated from wild type (Oregon R) embryos (0 -20 h), adult males, and adult females by using the FastTrack TM 2.0 kit (Invitrogen). First strand cDNA synthesis was done with embryonic mRNA (4 g) and adult male and female mRNAs (1 g each) by 300 ng of oligo(dT) primer at 42°C for 1.5 h by following the Smart TM protocol for rapid amplification of cDNA ends (Clontech). PCR amplification of Pkd2 cDNA used primers Pkd2.1 and Pkd2.4 (5Ј-GGGCCGCGTCTT-TCTCTATGAAAACC-3Ј and 5Ј-CTGCGTCCTTGGGTGATCTCACCC-TT-3Ј) that produce a 1,043-bp amplification product. The same pair of primers amplifies a 1,400-bp genomic fragment. For Fig. 1C, the primers used were Pkd2.1 and Pkd2.6 (5Ј-GGGCCGCGTCTTTCTCTAT-GAAAACC-3Ј and 5Ј-TAAAGCCAGCCAAGTCTTTCGAGCA-3Ј). The templates were from the Drosophila Rapid-scan TM gene expression panel (Origene Technologies, Inc.) that contains equal amounts of premade first strand cDNA from embryonic, larval, pupal, and adult stages at four 10-fold serial diluted quantities (10 ng, 1 ng, 100 pg, and 10 pg), thus allowing a rough estimation of Pkd2 mRNA abundance during development.

Reverse Transcription-PCR Analysis-For
Genetics of Pkd2-Several Pkd2 mutant alleles were generated by a targeted knock-out protocol (27). The allele used in this study is Pkd2 ko67 that was derived from the integration of a mutated Pkd2 donor into the endogenous Pkd2 locus (10). Pkd2 ko67 contains a tandem duplication of Pkd2 with one copy carrying the BamHI and the other copy carrying the MunI frameshift mutations (see Fig. 3A, *). The wild type Pkd2 gene encodes a polypeptide of 924 amino acids. The MunI muta-tion in Pkd2 ko67 is predicted to cause a frameshift at Gln 729 and translational termination at the 13th codon thereafter. This deletes the "P-loop," the sixth transmembrane domain, and the C-terminal tail that contains the coiled-coil domain. The BamHI mutation is predicted to cause a frameshift at Asp 102 and translational termination at the seventh codon thereafter. This deletes all six transmembrane domains, the P-loop, and the C-terminal tail. In addition to these designed frameshift mutations, the two Pkd2 copies of Pkd2 ko67 contain an insertion and a new MunI site due to imprecise homologous integration of the donor (see Fig. 3A). Western blotting showed that neither wild type size nor truncated Met 1 -Gln 729 PKD2 protein is produced by Pkd2 ko67 homozygote (10). This could be explained by possible instability of the truncated Met 1 -Gln 729 peptide or that the truncated peptide is never made due to the additional changes introduced into the Pkd2 ko67 allele during the targeted knock-out procedure.
Pkd2 is located on the second chromosome. To remove unrelated mutations that may be introduced into the Pkd2 ko67 genetic background during the targeted knock-out procedure, crosses were set up to first replace the X and the third chromosome of the Pkd2 ko67 stock with wild type chromosomes from a mapping stock (y w; Pin/CyO). The resulting Pkd2 ko67 stock was then backcrossed with isogenized strain w 1118 ; iso2; iso3 (28) for five generations. Muscle functional assays showed that Pkd2 ko67 homozygote is slightly weaker than Pkd2 ko67 /Df(2L)prd1.7 hemizygote, suggesting that Pkd2 ko67 is a strong loss-of-function allele.
Assays for Muscle Functions-All larval assays were repeated in triplicate to obtain standard error bars. Statistical analysis (t test) was performed using Sigma Plot version 2000. Food ingestion and waste excretion assays followed previously published protocols (29) with minor modifications. The assays were mostly done with newly hatched mutant larvae. However, for cDNA rescue experiments, second instar larvae (45 Ϯ 3 h after egg-laying) were used because the marker (Tb) for genotypic distinction is not visible in newly hatched larvae. For the food ingestion assay, ϳ150 larvae were fed with bromphenol blue-dyed yeast paste (blue yeast), and the percentage of larvae that contain blue yeast in the hindgut were scored as ingestion-positive at 0.5-or 1-h time points. For the waste excretion assay, ϳ150 larvae were fed with blue yeast for 4 h. Those that contained blue yeast in the hindgut were transferred to dye-free yeast (white yeast), and the complete loss of blue yeast inside the larvae was scored as excretion-positive at hourly intervals (blue-white pulse-chase experiment).
The rate of body wall contraction (BWC) was measured following the published protocol (29). Newly hatched larvae were fed yeast paste with or without 6 M ryanodine for 30 min (29). Then the fed larvae were placed individually on fresh agar plates. BWC was counted for a larva that was undergoing continuous forward movement for 30 s and repeated once for another 30 s. The average BWC rate/min is then calculated from three separate measurements for each larva for a total of 30 larvae per genotype.
For measuring mouth hook marks, larvae that just became wandering third instar larvae were used. An individual larva was placed on a fresh agar plate. The beginning and ending points of paths of continuous movements that took place within a total time of 2 min (by stop watch) were marked. Then the mouth hook marks along these paths were counted either as "shallow" marks as shown in Fig. 6B or "deep" marks that are any marks deeper and larger than those shown in Fig.  6B. Ten female and 10 male larvae were measured for each genotype.
Antibody, Western Blotting, and Immunocytochemistry-Anti-PKD2C was produced by immunizing rabbits with a synthetic peptide (residues 832-849) of the fly PKD2, purified by peptide-coupled affinity resin (SulfoLink gel, Pierce), and used at 1 g/ml for Western blots and tissue staining. Anti-PKD2(29/30) was produced by immunizing rabbits with a glutathione S-transferase fusion protein containing PKD2 sequence (residues 317-417), preabsorbed by passing through glutathione S-transferase-coupled glutathione resin (Sigma), and used at 1:5,000 dilutions for Western blots and immunostaining. Confocal imaging was done with a Leica TCS laser scanning spectral confocal microscope at the University of Alabama at Birmingham Imaging Facility.
Construction of Pkd2 Promoter-Gal4 Fusion Gene-The Pkd2 promoter (1.8 kb of DNA) was amplified with primers containing BamHI restriction enzyme sites (underlined): 5Ј-CGCGGATCCTGGCTT-TCAAATCGGCAGCAAG-3Ј and 5Ј-CGCGGATCCTCGTTCCAGTTG-CACGCTTTGA-3Ј. The fragment was cloned into the pGEM T , vector and the cloned sequence was verified by restriction enzymes. The cloned Pkd2 promoter was released by NotI and SacII and subcloned into the P-element transformation vector pPT-Gal4 (30) at NotI and EcoRI sites (SacII and EcoRI are blunt-ended). Visualization of Pkd2-expressing 2 Venglarik, C. J., Gao, Z., and Lu, X., J. Am. Soc. Nephrol., in press. cells was done by crossing flies carrying the Pkd2-Gal4 transgenic line to flies carrying a UAS-GFP reporter.

RESULTS
Developmental Expression of Pkd2-We determined the developmental expression pattern of Pkd2 using antibodies recognizing the fly PKD2. Tissue survey by immunostaining indicates that the male testis is the only organ where PKD2 is abundantly expressed (10). Due to low abundance of PKD2 in most somatic tissues, the protein was detectable in the male but not in the female whole body extracts (one fly/lane, Fig. 1A). However, it was detectable in more concentrated head extracts of both sexes (40 heads/lane, Fig. 1A). Consistent with this, Pkd2 mRNA was detectable in both sexes as well as in the embryo by the more sensitive reverse transcription-PCR method (Fig. 1B). The Pkd2 mRNA level in the embryo was extremely low. Coinciding with early testis development, the level began to rise in third instar male larvae and persisted at high levels in male pupae and male adults (Fig. 1C).
Immunostaining showed that PKD2 was expressed in visceral SMCs of the gut and the skeletal muscles of both sexes (Fig. 2). As in mammalian vascular SMCs (26), PKD2 in fly visceral SMCs appeared predominantly in the cytoplasm surrounding the nucleus (Fig. 2A). To facilitate the identification of other cells where Pkd2 may be expressed, we generated transgenic flies carrying the Pkd2 promoter cloned upstream of the coding region for the yeast transcriptional factor Gal4. This Pkd2-Gal4 transgene was used to drive the expression of UAS-GFP reporter to locate the cells where the endogenous Pkd2 is expressed. Several Pkd2-Gal4 insertions showed green fluorescent protein (GFP) expression in the testis, pharynx muscle, and SMCs surrounding the esophagus (Fig. 2E) and in the hindgut. The role of Pkd2 in these muscle cells is demonstrated by functional assays below.
Visceral Smooth Muscle Functions Are Severely Affected by Pkd2 Mutations-The Drosophila gut is a monolayered endothelial tube ensheathed by circular and longitudinal fibers of visceral smooth muscle cells (31). At the cellular level, the Drosophila gut is grossly analogous to mammalian blood ves-sels except that the gut lumen has no valves. Food intake into the gut is mediated by pharynx contraction. Once inside, the food is propelled along the gut via the contraction of the esophagus and midgut. Later waste is excreted out of the anus via the contraction of the hindgut. To determine whether Pkd2 has a functional role in pharyngeal and visceral smooth muscles, we determined the rates of food intake and waste excretion as these rates reflect the contractility of the pharyngeal and visceral smooth muscles (29).
The Pkd2 mutation analyzed was the Pkd2 ko67 allele we generated by the integration of a mutated Pkd2 donor sequence into the endogenous locus (10). Pkd2 ko67 contains two Pkd2 copies with one copy carrying a BamHI frameshift mutation and the other copy carrying a MunI frameshift mutation (Fig.  3, B and denoted by * in A). We chose to use Pkd2 ko67 in our assays because it carries a donor-derived white hs eye marker gene, which conveniently allowed us to isogenize Pkd2 ko67 into the w 1118 ; iso2; iso3 isogenic strain (28). Genomic Southern analyses confirmed the general structure of Pkd2 ko67 and identified two chromosomal deletion lines, Df(2L)prd1.7 (33B02-03; 34A01-02) and Df(2L)Prl (32F01-03;33F01-02) that delete the Pkd2 locus (Fig. 3C), whereas Df(2R)esc-P3-0 (33A01-02;33E) does not delete the Pkd2 locus (Fig. 3C). These analyses also showed that in the generation of Pkd2 ko67 allele, integration of the donor sequence did not occur precisely since the left Pkd2 copy contains a small insertion (Fig. 3A, denoted by ---) and the right Pkd2 copy contains a new MunI site (Fig. 3A). The combined effects of these genomic changes plus the frameshift mutations result in no detectable wild type PKD2 protein being produced in the Pkd2 ko67 homozygote (10). In the following muscle functional assays, Pkd2 ko67 appeared as a strong lossof-function allele just slightly weaker than Df(2L)prd1.7 deletion of the gene.
To determine the food intake rates, newly hatched larvae were fed with blue yeast paste (dyed with digestion-resistant bromphenol blue) for 0.5 or 1 h, and those containing blue food in the hindgut were scored. To avoid developmental age differences, newly hatched larvae were used in the assay as was done previously for measuring the reduced muscle function of Drosophila ryanodine receptor (Ryr) mutants (29). Similar to another published report (29), 83 Ϯ 3.6% of the wild type larvae were ingestion-positive by 0.5 h (Fig. 4A). In contrast, ingestion rates were reduced by over half (ϳ56%, p Ͻ 0.0001) of the wild type rate for Pkd2 ko67 /Pkd2 ko67 and Pkd2 ko67 /Df(2L)prd1.7 homozygous mutants (36 Ϯ 1.3 and 34.9 Ϯ 1.0%, respectively; Fig.  4A). Pkd2 ko67 /ϩ and Df(2L)prd1.7/ϩ heterozygous larvae also showed reduced ingestion rates (53.7 Ϯ 1.5 and 52.6 Ϯ 6.3%, respectively) by a little over one-third (ϳ35%, p Ͻ 0.0001) of the wild type rate. This indicates that Pkd2 is haploinsufficient for optimal food intake functions (Fig. 4A).
To measure the rates of waste excretion, newly hatched larvae were fed with bromphenol blue-dyed yeast paste for 4 h, and those that contained the blue food in the hindgut were then transferred to dye-free food and scored for the complete loss of the blue food in the gut in 1 h. The rate of waste excretion depends on food passage along the entire digestive tract and thus is the best measurement for overall contractility of the visceral smooth muscle cells. In this "pulse-chase" experiment, 51.6 Ϯ 1.0% of the wild type larvae had completed the blue food excretion in 1 h. In contrast, very few of the Pkd2 ko67 /Pkd2 ko67 and Pkd2 ko67 /Df(2L)prd1.7 mutant larvae (8.5 Ϯ 0.8 and 5.8 Ϯ 0.2%, respectively) had completed the blue food excretion within the same amount of time, indicating that the excretion rates were severely reduced by over 80% of the wild type rate (Fig. 5A). The excretion rate of Pkd2 ko67 /ϩ heterozygotes (24.2 Ϯ 4.9%) was nearly half of the wild type rate. The Df(2L)prd1.7/ϩ heterozygote was more severe than the Pkd2 ko67 /ϩ heterozygote (14.5 Ϯ 3.6%). This could be because Pkd2 ko67 has some residual activity or the Df(2L)prd1.7 chromosome carries background-modifying mutations since it was not isogenized. Our analyses showed that Pkd2 is strongly haploinsufficient for optimal waste excretion functions (Fig.  5A).
Since the excretion rate was most strongly affected in Pkd2 mutants, we determined whether expression of wild type Pkd2 cDNA would rescue the phenotype. Wild type Pkd2 cDNA from a Gal4-inducible UAS-F3112 rescue transgene was expressed by using muscle Gal4 driver 24B (32) that is expressed in visceral smooth muscles of the gut. One copy of 24B and UAS-F3112 restored the excretion function of Pkd2 ko67 /Pkd2 ko67 larvae back to the wild type level (Fig. 5B). The ingestion function was only partially rescued by 24B and UAS-F3112 due to the low expression level of 24B in the pharyngeal muscle (Fig. 4B). These results suggest that reduced Pkd2 activity severely reduces the contractility of visceral smooth muscles.
Skeletal Muscle Function Is Weakly Affected by Pkd2 Mutations-The skeletal muscle of the body wall mediates Drosophila larval locomotion. When larvae crawl on the surface of agar substrate, forward movement is initiated by body wall contraction at the posterior tail region. The contraction wave is propagated anteriorly until it reaches the mouth hook, a double tooth-like hook located in the larval mouth. Forward movement is completed by the extension of the larval head in a stereotypical "up, forward, and down" pattern. Head down motion is often associated with the insertion of the mouth hook into the agar substrate. This leaves mouth hook marks on the agar surface.
The locomotive behavior of Pkd2 mutant larvae was grossly normal. BWC rates of Pkd2 ko67 /Pkd2 ko67 and Pkd2 ko67 / Df(2L)prd1.7 larvae were similar to that of the wild type larva (42.5 Ϯ 1.0 BWCs/min, n ϭ 30). However, the frequency and depth of the mouth hook marks left along the traveling path, which reflect the force of skeletal muscle contraction, were reduced in the mutant. The average number of mouth hook marks made by a wild type larva was 86.5 Ϯ 2/min (n ϭ 20) of active forward movement and that made by the Pkd2 ko67 / Pkd2 ko67 larva was 65.5 Ϯ 1.9/min, a 24% reduction (n ϭ 20, p Ͻ 0.0001). In addition to reduced frequency of mouth hook marks, Pkd2 ko67 /Pkd2 ko67 or Pkd2 ko67 /Df(2R)Prd1.7 larvae made shallower mouth hook marks than wild type control larvae (Fig. 6, A versus B). For quantification purposes, any hook marks larger and deeper than those shown in Fig. 6B   FIG. 3. Genomic structure of the Pkd2 ko67 targeted disruption allele. A is a schematic drawing of the wild type allele and the Pkd2 ko67 mutant allele. Nucleotides are numbered with respect to transcription start site (ϩ1). The Pkd2 ko67 allele was produced by homologous insertion of a donor sequence into the endogenous locus. "---" indicates a small insertion and the arrow points to a new MunI site that were inadvertently introduced during the procedure. The MunI and BamHI sites in wild type genomic sequence were mutated to frameshift mutations in the donor sequence and subsequently integrated in the Pkd2 ko67 allele (* in A; see details under "Materials and Methods"). B, the MunI and BamHI frameshift mutations are verified by separate PCR amplifications of the left Pkd2 copy (cut by MunI, lanes 1 and 2) and of the right Pkd2 copy (cut by BamHI, lanes 3 and 4). C, genomic DNA samples from the indicated genotypes were cut by NgoMIV and probed with a DNA fragment (nucleotides 452-1428, filled black box in A). It shows the duplicated structure of Pkd2 ko67 and deletions of Pkd2 by Df(2L)prd1. 7 and Df(2L)Prl (referred to as Df(2)2 and Df(2)3, respectively), whereas Df(2L)esc-P3-0 (Df(2)1) does not delete Pkd2.
were counted as deep hook marks. Pkd2 ko67 /Pkd2 ko67 larvae made approximately ϳ15% fewer deep hook marks (n ϭ ϳ9,000) as compared with wild type larvae based on this criterion (10 male and 10 female were used in the assay). Thus it appears that loss of Pkd2 activity only slightly reduces the performance of the body wall skeletal muscle. Interestingly this weak effect of Pkd2 mutation was greatly augmented by a small reduction of the ryanodine receptor activity (see below).
Pkd2 Affects Ca 2ϩ Homeostasis by Cooperating with the Ryanodine Receptor-Optimal muscle performance revolves around Ca 2ϩ homeostasis. The maintenance of Ca 2ϩ homeostasis requires the concerted action of Ca 2ϩ channels on the cell surface and on the membrane of the intracellular calcium storage organelle, the sarcoplasmic reticulum. One of the important sarcoplasmic reticulum Ca 2ϩ channels is the ryanodineinhibitable ryanodine receptor (Ryr), which is involved in excitation-contraction coupling of the skeletal muscle (33,34). Ryr mutations in mouse, Caenorhabditis elegans, and Drosophila show that ryanodine receptors have conserved roles in muscle contraction (29,(33)(34)(35). Drosophila has a single Ryr gene that is involved in regulating rapid Ca 2ϩ oscillation that ultimately influences the rate of larval BWC (29).
Although the BWC rate was normal for Pkd2 ko67 /Pkd2 ko67 or Pkd2 ko67 /Df(2L)prd1.7 larvae (42.5 Ϯ 0.7 BWCs/min, n ϭ 30), we noticed that the mutant larvae showed increased sensitivity to ryanodine (Fig. 6C). After being fed with yeast paste containing a low concentration of ryanodine (6 M ryanodine for 30 min) that had no significant effect on the wild type larvae, the BWC rate of Pkd2 ko67 /Df(2L)prd1.7 larvae was reduced by 19% (34.3 Ϯ 1.4 BWCs/min, n ϭ 30, p Ͻ 0.0001). To test the genetic interaction between Pkd2 and Ryr, we used the weak Ryr 04913 allele. The BWC rate of Ryr 04913 /ϩ heterozygous larvae was nearly normal (41.0 Ϯ 0.9 BWCs/min, n ϭ 30, p Ͼ 0.02); however, that of Pkd2 ko67 Ryr 04913 /Pkd2 ko67 /ϩ was reduced by 18% (33.6 Ϯ 1.0 BWCs/min, n ϭ 50, p Ͻ 0.0001) to a level nearly as low as Ryr 04913 /Ryr 04913 homozygotes (32.0 Ϯ 0.75 BWCs/min, n ϭ 30). This is a strong synergistic reduction since the BWC rate of Ryr 16 homozygous null mutant was only reduced by 55% (19.0 Ϯ 1 BWCs/min, n ϭ 30) (29). These results strongly suggest that Pkd2 mediates optimal muscle contractility through intracellular Ca 2ϩ homeostasis and cooperates with the ryanodine receptor in this process. This is consistent with the recent finding that Ca 2ϩ influx through ciliary PKD2 in the kidney epithelia is accompanied by Ca 2ϩ release from intracellular Ca 2ϩ storage organelles through the ryanodine receptor (17). DISCUSSION This study demonstrated that Drosophila PKD2 functions in the skeletal and smooth muscle cells where it is required for optimal muscle contractility. During the course of this study, muscle fiber organization and attachment to the epidermis in Pkd2 mutants was found to be normal by visualizing GFP expressed in muscles of living Drosophila larvae (data not shown). Muscle ultrastructure determined by transmission electron microscopy was also found to be normal (data not shown). Thus, insufficient Pkd2 activity appears to compromise FIG. 4. Pharyngeal and visceral smooth muscle contractility as measured by larval food ingestion rates. A, newly hatched first instar larvae of the wild type and mutant genotypes as indicated. Notice that Pkd2 is haploid-insufficient for optimal food ingestion rate. B, rescue experiment was done using staged second instar larvae (45 Ϯ 3 h after egg-laying) to ensure the accurate recognition of larval genotypes. Notice that wild type Pkd2 cDNA transgene UAS-F3112 appears to weakly rescue the mutant phenotype possibly because the 24B Gal4 driver only weakly expresses in the pharynx. Bars represent S.E. from three separate measurements (n ϭ 150 larvae/experiment, repeated three times, p Ͻ 0.0001 at 0.5 h).
FIG. 5. Pharyngeal and visceral smooth muscle contractility as measured by larval waste excretion rates. A, newly hatched first instar larvae. B, staged second instar larvae (45 Ϯ 3 h after egg-laying). The data shown were collected at 1 h following the "blue-to-white" food pulse-chase experiment. Notice that Pkd2 is haploid-insufficient for optimal waste excretion function. The expression of wild type Pkd2 cDNA UAS-F3112 by 24B Gal4 driver fully rescues the phenotype of the homozygous mutant. UAS-F3112 or 24B alone does not rescue the mutant as expected, although 24B causes a slightly increased excretion rate possibly due to the presence of genetic modifier(s) on the 24Bcarrying chromosome. Bars represent S.E. from three separate measurements (n ϭ 150 larvae/experiment, repeated three times, p Ͻ 0.0001). muscle function rather than muscle development and structure in Drosophila. A direct role of PKD2 in muscle contractility was shown by rescue of the food excretion phenotype by musclespecific expression of wild type Pkd2 cDNA in Pkd2 ko67 / Pkd2 ko67 larvae (Fig. 5B).
The impairment of smooth muscle contractility by Pkd2 mutations was especially strong. Mutating one of the two Pkd2 gene copies was sufficient to cause ϳ53% reduction of smooth muscle contractility as measured by the food excretion assay (Fig. 5A). This is particularly interesting with respect to AD-PKD patients who are heterozygous individuals that show high frequencies (60 -100%) of hypertension and increased susceptibility to vascular aneurysms (1, 2, 5). The mammalian blood vessels are monolayered endothelial tubes ensheathed by circular and longitudinal fibers of vascular SMCs. Contractility of these cells is involved not only in regulating vessel diameters but also in maintaining vascular structural and functional integrity. It has been shown that weakened vascular SMCs of the blood vessel wall become apoptotic under fluid mechanical stress, resulting in vascular leakage and aneurysms (36). Mammalian PKD1 is expressed in the vascular SMCs throughout the circulatory system of the brain and body (20). PKD2 is also expressed in the vascular SMCs in mammals (21)(22)(23). Vascular wall integrity is severely compromised in Pkd1Ϫ/Ϫ and Pkd2Ϫ/Ϫ knock-out mouse embryos as indicated by gross edema and hemorrhage (20,25). However, evidence for a role of PKD2 in SMC contractility was previously lacking. Our study demonstrated that Drosophila PKD2 was required for optimal SMC contractility, and mutating just one of the two Pkd2 gene copies was sufficient to reduce the contractility by ϳ53%. This is consistent with a recent report that intracellular Ca 2ϩ is reduced in Pkd2ϩ/Ϫ heterozygous mouse vascular SMCs, and these mice show higher rates of vascular complications and lethality than Pkd2ϩ/ϩ control mice under an experimentally induced hypertension condition (26). Based on these data, we propose that haploinsufficiency of human PKD2 function leads to decreased vascular SMC contractility and function, and this may play a primary role in the development of a variety of vascular complications such as hypertension and aneurysms in ADPKD patients.
Our analyses also suggest that the muscle function of Pkd2 is mediated through intracellular Ca 2ϩ homeostasis that involves the synergistic actions of Pkd2 and the intracellular calciumreleasing channel, the ryanodine receptor. A recent study shows that PKD1 and PKD2 are co-localized on primary cilia of cultured renal epithelia (17). Fluid flow across the epithelial layer causes cilium bending and triggers Ca 2ϩ influx from the plasma membrane (17). Interestingly this Ca 2ϩ influx through ciliary PKD2 in the renal epithelial culture is also accompanied by intracellular Ca 2ϩ release from ryanodine-sensitive internal Ca 2ϩ stores (17). Our observation of the strong genetic interaction between Pkd2 and the ryanodine receptor mutation suggests that cooperation between PKD2 and the ryanodine receptor also occurs in the SMCs. PKD2 has been shown to localize on the endoplasmic reticulum membrane and the plasma membrane as well as on primary cilia (15,17,18,23,37,38). In both mammalian and fly SMCs, PKD2 appears to be predominantly cytoplasmic (23). Thus, propagation of Ca 2ϩ signaling through coupling of PKD2 with the ryanodine receptor may represent a common and evolutionarily conserved aspect of PKD2 function in both epithelial and non-epithelial cell types. However, it is possible that different molecular coupling mechanisms are involved between PKD2 and the ryanodine receptor depending on the subcellular location of PKD2 in different cell types. FIG. 6. Effects of Pkd2 on skeletal muscle functions. A and B are two trails of mouth hook marks made by third instar larvae of the genotypes indicated to show differences of size and depth. C shows larval BWC rates. Measurement was done with newly hatched first instar larvae of the genotypes shown (n ϭ 30). The larvae were fed for 30 min with yeast paste containing no drug (black bars) or 6 M ryanodine (gray bars). For ryanodine effects, statistical significant difference between the pairs was indicated (*, p ϭ 0.02; **, p Ͻ 0.0001). Notice that the BWC of Pkd2 ko67 is normal, but the BWC of Pkd2 ko67 / ϩ/Pkd2 ko67 Ryr 04913 is reduced by as much as that of the Ryr 04913 homozygote. This indicates a strong synergistic interaction between Pkd2 and Ryr. Ryr 16 is close to being a null allele.