Essential role of cleavage of Polycystin-1 at G protein-coupled receptor proteolytic site for kidney tubular structure

Abstract

Polycystin-1 (PC1) has an essential function in renal tubular morphogenesis and disruption of its function causes cystogenesis in human autosomal dominant polycystic kidney disease. We have previously shown that recombinant human PC1 is cis-autoproteolytically cleaved at the G protein-coupled receptor proteolytic site domain. To investigate the role of cleavage in vivo, we generated by gene targeting a Pkd1 knockin mouse (Pkd1^V/V) that expresses noncleavable PC1. The Pkd1^V/V mice show a hypomorphic phenotype, characterized by a delayed onset and distal nephron segment involvement of cystogenesis at postnatal maturation stage. We show that PC1 is ubiquitously and incompletely cleaved in wild-type mice, so that uncleaved and cleaved PC1 molecules coexist. Our study establishes a critical but restricted role of cleavage for PC1 function and suggests a differential function of the two types of PC1 molecules in vivo.

Polycystin-1 (PC1) is the product of PKD1, the principal gene mutated in human autosomal dominant polycystic kidney disease (ADPKD) affecting 1 of 1,000 people (1). The disease is characterized by the progressive development of multiple bilateral cysts in the kidney, resulting in renal failure in 50% of patients by their sixth decade of age (2). ADPKD is a systemic disease with many extrarenal manifestations including hepatic and pancreatic cysts and intracranial aneurysms.

Pkd1 knockouts exhibit a variable degree of early extrarenal developmental abnormalities including polyhydramnios, s.c. edema, and hemorrhage, but invariably develop severe polycystic kidneys, culminating in embryonic lethality (3–6). This finding indicates wider-ranging functions of PC1 in embryonic development. PC1 (mouse) is a 4,293-aa 11-transmembrane (TM) glycoprotein with a 3,066-aa-long N-terminal ectodomain that contains a unique combination of motifs involved in protein–protein interactions, and a ≈200-aa cytoplasmic C terminus (7, 8) that can activate a number of signaling pathways (9). The ectodomain is separated from the 11-TM segment by the G protein-coupled receptor (GPCR) proteolytic site (GPS) domain of ≈50 aa (10). The GPS domain was first identified as an internal cleavage site in a neuronal GPCR protein CL and was later found at a similar position in >30 adhesion GPCRs that are extraordinary for having an unusually large and complex ectodomain (reviewed in ref. 11). We have shown that recombinant human PC1 is cleaved at HL̂T (where “^” identifies the position of cleavage) within the GPS domain in early secretory pathway through cis-autoproteolysis (12, 13). The cleavage results in the N-terminal fragment (NTF) and the 11-TM C-terminal fragment (CTF) that remain noncovalently tethered to each other (12). Cleavage is incomplete, with similar ratios of the uncleaved full-length uFL-PC1 and the cleaved PC1 products. PKD1-associated missense mutations in the vicinity of the cleavage site disrupt cleavage and abolish PC1's ability to activate the Jak2-Stat pathway and to induce in vitro tubulogenesis in MDCK cells (12). The uFL-PC1 and cleaved PC1 products therefore may have different functions in vivo.

Proteolytic cleavage is a common regulatory mechanism for protein functions (14). It has recently been reported that PC1 undergoes cleavage by a non-GPS-mediated process to release a fragment from its cytoplasmic C terminus from the membrane in association with mechanical stimuli, and this fragment enters the nucleus to initiate signaling (15, 16). However, the mechanism by which GPS cleavage affects PC1 function is unknown. The GPS sequence is present in the entire PC1 family (13), but some members such as PKDREJ (17) and suREJ2 (18) are, in fact, not cleaved in vivo. In addition, most previous studies have reported that endogenous PC1 is not cleaved at GPS in various cell types and tissues including the kidney (e.g., refs. 19–21). These findings raise a question about the functional significance of GPS cleavage for PC1 in vivo.

In the present study, we generated and characterized a Pkd1 knockin mouse without GPS cleavage of PC1 to examine the function of the cleavage in vivo. We analyzed the cleavage patterns of endogenous PC1 in WT and mutant mice. Our study establishes a critical but restricted role of cleavage for PC1 function in vivo.

Results

Generation of Pkd1 Knockin Mouse with Absent GPS Cleavage of PC1.

We chose a reverse genetic approach, mutating the codon ACT (Thr) to a GTT (Val) at amino acid position 3041 of the cleavage site HL̂T³⁰⁴¹ in the mouse Pkd1 gene to investigate the function of PC1 cleavage in vivo. The mutation T3041V is known to prevent cleavage of recombinant PC1 (13) and is expected to disrupt presumed cleavage of PC1 in the mouse. The resulting noncleavable mutant PC1 (named PC1^V) should direct a Pkd1 null-like phenotype if cleavage is essential for PC1 function.

We generated and transfected the targeting vector (Fig. 1A i) into ES cells for homologous recombination and identified 1 of 200 selected ES clones that had proper targeting (Fig. 1B). This clone (G8) was used to generate F1 heterozygous Pkd1^NeoV/+ mice (Fig. 1A iii), which appeared normal and were fertile. Their intercross did not produce viable homozygous offspring, probably because of interference of RNA splicing by Neo (data not shown). We therefore removed Neo by crossing the Pkd1^NeoV/+ mice with FLP transgenic mice (22) (Fig. 1C), as confirmed by genomic DNA sequencing (data not shown). The resulting Pkd1^V/+ heterozygotes (Fig. 1A iv) lived as long as the WT littermates did and showed no cysts in the kidney and liver, or any other abnormalities, at 13 and 22 months (n = 6).

Fig. 1.

Generation of the Pkd1^V/V mouse by gene targeting. (A) Structure of various Pkd1 alleles. (i) Targeting vector. Neo, FRT-flanked (red triangle) PGK-neomycin selection cassette; TK, thymidine kinase gene; S, SpeI. Exons are depicted by solid boxes. The T3041V mutation is indicated. (ii) Pkd1 WT allele, with exon number shown. (iii) Pkd1^NeoV allele. The expected size of the SspI or SpeI restriction fragments for WT and Pkd1^NeoV alleles by using the 5′ probe (within exon 15 outside of the vector) is shown. The expected size of the 3′ long-range PCR product is also shown. (iv) Pkd1^V allele. A single FRT site of 47 bp is retained at the site of Neo insertion. (B) Genomic Southern blot showing correct targeting of the ES clone G8 by detection of 9.9-kb (SspI) and 10.4-kb (SpeI) Pkd1^NeoV-specific bands by the 5′ probe. The 3′ long-range genomic PCR confirmed correct targeting (data not shown). (C) Removal of Neo in the offspring of the Pkd1^NeoV/+:FLP mating as shown by the three-primer PCR-based strategy. (i) Primer positions and the expected size of the PCR product of the each allele. (ii) PCR result showing successful removal of Neo by amplification of the Pkd1^V-specific 329-bp product (Upper) only in the FLP-positive offspring (Lower). −, negative control for PCR. (D) RT-PCR from WT, Pkd1^V/+ (V/+), and Pkd1^V/V (V/V) kidneys at P9. The expected size of the PCR product using exon 23 and 27 specific primers is shown (Upper). A single product of expected size was amplified from each genotype (Lower). DNA size marker is shown on right. (E) DNA sequence of the RT-PCR products in D at the cleavage size.

Pkd1^V/+ mice exhibited normal reproduction when intercrossed, with typical litter size between 6 and 10 pups. Pkd1^V/V homozygotes were born at the predicted Mendelian frequency when examined in a total of 255 offspring (Table 1). This result shows that the Pkd1^V allele can overcome embryonic lethality usually seen in the Pkd1^−/− mice. This finding was confirmed by the normal appearance of Pkd1^V/V embryos (E15; n = 3) as the WT littermates (data not shown). We confirmed the proper mRNA expression of the Pkd1^V allele in the mouse by RT-PCR. Only a single product of predicted size for WT mRNA was amplified from the kidneys of postnatal day 9 (P9). Pkd1^V/V, Pkd1^V/+, and WT littermates with primers within exon 23 and exon 27 (Fig. 1D) and each product showed the expected sequence at codon 3041 (Fig. 1E). The same result was obtained for liver, spleen, brain, lung, and heart (data not shown).

Table 1.

Genotype ratios in a total of 255 mice obtained from 26 interbreedings of heterozygous Pkd1^V/+ mice

Genotype	+/+	V/+	V/V
Number	65	124	66
Actual ratio, %	25.5	48.6	25.8
Expected ratio, %	25	50	25

Cystic Kidney Development in Postnatal Pkd1^V/V Mice.

The Pkd1^V/V mice appeared normal in the first 8 days of postnatal life, but became slightly smaller than the WT or Pkd1^V/+ littermates from P9 onward (Fig. 2A). By P16, all animals displayed markedly smaller body statures with distended abdomens (Fig. 2B). Mice killed at P9 and P16 had pale, grossly enlarged cystic kidneys (Fig. 2 A and B Right), as well as dilated common bile ducts (data not shown). At 3 weeks or after, the Pkd1^V/V mice weighed less than half the WT or Pkd1^V/+ littermates, but had a considerably larger kidney/body ratio (26.9 ± 2.1, n = 3) than the WT or Pkd1^V/+ littermates (1.3 ± 0.1, n = 8). They died between 2 and 6 weeks after birth, with ≈50% dead by the third week of age (Fig. 2C), presumably because of renal insufficiency as indicated by the elevated blood urea nitrogen (BUN) levels (Fig. 2D). Pancreas and spleen appeared normal.

Fig. 2.

Postnatal development of cystic kidney in Pkd1^V/V mice. General appearance of the mice and their kidneys at P9 (A) and P16 (B). Note that Pkd1^V/V kidneys are enlarged, pale, and cystic compared with kidneys from WT and heterozygous (V/+) littermates. (C) Survival curve of WT (n = 33), V/+ (n = 62), and V/V (n = 29) mice. The median age of survival of Pkd1^V/V mice is 23 days. (D) Elevated BUN level in the Pkd1^V/V mouse as compared with that in Pkd1^V/+ and WT littermates at P14 and P23.

Tubular Cystic Dilation in Pkd1^V/V Mouse Kidney.

Histological examination revealed a rapid and progressive tubular dilation of the Pkd1^V/V kidneys during the postnatal maturation stage (Fig. 3). At P0, the mutant kidneys contained a few microcysts (≈10 discrete cysts) in the subcortical region, but otherwise appeared normal (Fig. 3A). From P1 through P5, both the number and size of the cysts gradually increased, initially in the cortex and then extending into the medulla. The overall kidney sizes, however, were not noticeably enlarged as compared with those of the WT littermates, and significant amounts of normal parenchyma remained (Fig. 3 A and B).

Fig. 3.

Rapid and progressive cystic dilation in Pkd1^V/V postnatal kidneys. (A) Overview of Pkd1^V/V kidneys of various postnatal stages in H&E-stained sections, with two to three animals analyzed for most of the stages. (Scale bars, 2 mm.) Note the intact papilla tip in P3–P14 kidneys. P22 and P28 kidneys are more enlarged (≈15 mm long) with further cystic expansion (data not shown). (B–D) H&E-stained sections of P5, P7, and P14 Pkd1^V/V kidneys, respectively. Note the intact glomeruli and proximal tubule (PT)-like structures. (E) Masson-trichrome (MT)-stained kidney section of P22 Pkd1^V/V mouse. Note the intact glomeruli and PT-like tubules surrounded by interstitial fibrosis. (F) H&E-stained section of P28 Pkd1^V/V kidney. Note the large size of the cysts and intact glomeruli and PT-like tubules (Inset). (G) MT-stained liver section of P14 Pkd1^V/V mouse. Note the fibrosis around the dilated biliary ducts.

Cystic dilation extended progressively into the medulla from P7 onward, and by P14, cysts had replaced most of the normal renal parenchyma with the exception of the papillae tip that remained grossly intact (Fig. 3A). Intact glomeruli and tubules with brush borders characteristic of proximal tubules (PTs) were readily identified in the cortical region at this stage (Fig. 3 C and D). Both structures were consistently observed throughout later stages, at P22 (Fig. 3E) and P28 (Fig. 3 E and F Inset), when massively cystic kidneys reached a length of ≈15 mm and a weight of ≈10 times the kidney of the WT littermate and intact papillae were no longer observed.

The Pkd1^V/V mice also showed mild dilation of biliary ducts surrounded with fibrotic areas, starting at P7 through P28 (Fig. 3G).

Segments of Origin of Renal Tubular Cysts.

We determined cyst origins by using a set of nephron segment-specific markers. At P1, the few cysts seen were of collecting duct (CD) origin (AQP2 positive), whereas PT (LTL positive) was intact (Fig. 4A). At P5, virtually all cysts were of CD origin (DBA positive) in the medulla (Fig. 4 B–D), whereas PT (Fig. 4B) and thick ascending limb (TAL) (Fig. 4D, NKCC2 positive) were intact. In the cortex, three types of cysts could be identified: (i) DCT-derived (THP positive and DBA and NKCC2 negative on serial sections) (Fig. 4 C and D, *); (ii) CD-derived (DBA positive) (Fig. 4 C and D, arrowhead); (iii) probably derived from the junction between DCT and CD (composite staining with DBA and anti-THP) (Fig. 4 C and D, rectangle). At P9, most cysts were either CD-derived (Fig. 4 E and F) or, to a lesser extent, DCT-derived (≈10%) (Fig. 4E, star), whereas TAL (Fig. 4F Inset) and PT (data not shown) were intact. At P14 (Fig. 4G) and P28 (Fig. 4H), virtually all cysts had a CD origin, whereas PT (Fig. 4 G and H), now confined to the narrow areas between adjacent cystic walls, were consistently intact, thus validating our morphological findings.

Fig. 4.

Distal nephron segment origin of cysts in Pkd1^V/V kidneys. (A–H) Pkd1^V/V kidney sections of various postnatal stages were double-stained with segment-specific markers as indicated under each section. Lectin Lotus tetragonolobus (LTL), proximal tubule (PT); lectin Dolichos biflorus (DBA), collecting duct (CD) (26); AQP2, CD (27); Tamm–Horsfall protein (THP), both the thick ascending limb (TAL) and the distal convoluted tubule (DCT) (28); and Na-K-Cl Cotransporter 2 (NKCC2), TAL only (29). (I) Summary of cyst origin in Pkd1^V/V kidneys. Cysts are derived from DCT and CD (in red) (except the papilla tip), and PT and TAL are not dilated.

Our results show that the renal tubular cystic dilation in Pkd1^V/V kidney involves primarily distal nephron segments, but not the more proximal nephron segments, throughout the postnatal stages (Fig. 4I). Our finding contrasts with the findings in Pkd1^−/− mice (3) or human ADPKD (23), in which all segments of the nephron are affected. The Pkd1^V allele, therefore, directs a hypomorphic phenotype, with a delayed onset and distal segment-specific involvement of cystic dilation at postnatal stage.

PC1 Cleavage Pattern During Embryonic Development.

To elucidate the basis of the observed hypomorphic activity of the Pkd1^V allele, we characterized the cleavage pattern of endogenous PC1 in the WT and Pkd1^V/V mice. We first analyzed PC1 cleavage in murine embryonic fibroblasts (MEFs) isolated from E12 embryos of WT, Pkd1^V/V, Pkd1^V/+, and Pkd1^−/− genotypes, by a combination of immunoprecipitation (IP) and Western blot with use of anti-CC, a new polyclonal antibody against the C-terminal 170 aa of mouse PC1 (Fig. 5A). The specificity of this antibody is demonstrated by its detection of both uFL and CTF of the C-terminally FLAG-tagged recombinant PC1 encoded by the full-length mouse Pkd1 cDNA (mPC1-F) after IP with either anti-FLAG or anti-CC (Fig. 5B). This cleavage pattern is identical to that of the equivalent human recombinant PC1 (lane hPKD1-F) by anti-CT, as found (12, 13), and endogenous PC1 of the human embryonic kidney cell line (HEK).

Fig. 5.

Cleavage patterns of endogenous PC1 in WT and Pkd1^V/V mice. Endogenous PC1 was immunoprecipitated by using anti-CC and detected on Western blot by anti-CC. The protein size marker and the position of uFL-PC1 and CTF are indicated. (A) Schematic diagram of the domain organization of mouse PC1. LRR, Leucine-rich repeat; R1 and R2–16, PKD repeats; CLD, C-type lectin domain; REJ, receptor for egg jelly domain. The cleavage site HL̂T³⁰⁴¹ within GPS and the resulting NTF (3,040 aa) and CTF (1,253 aa) are shown. The epitope recognized by anti-CC or anti-CT (the equivalent antibody against human PC1), and anti-LRR used in this study is indicated by a black bar. (B) Western blots demonstrating the specificity of anti-CC. It detects both C-terminally FLAG-tagged uFL and CTF from lysate of HEK cells with stable expression of mouse full-length Pkd1 cDNA (mPC1-F) after IP with either anti-FLAG or anti-CC (Right). These signals were not detected from untransfected HEK cells (data not shown). Note that this pattern is identical to that of exogenously expressed human recombinant (hPKD1-F) or endogenous PC1 of HEK cells (HEK). (C–I) Western blots showing PC1 cleavage pattern in MEFs of various genotypes as indicated (C) and whole E12 embryos of various genotypes as indicated (D). Bracket indicates nonspecific bands that are also present in E12.5 and E14.5 Pkd1^−/− embryos (Right). (E) WT embryos of various stages as indicated. (F) Postnatal kidneys of WT and Pkd1^V/Vmice at different stages. * indicates an unknown PC1-specific band that occurs only in Pkd1^V/V cystic kidneys. (G) Primary PT and CD cells isolated from P5 WT and Pkd1^V/Vkidneys (Left) and conditionally immortalized CD cells isolated from P14 WT and Pkd1^V/Vkidneys (Right). Same results were obtained for primary PT and CD cells isolated from P10 WT and Pkd1^V/Vkidneys, respectively (data not shown). The purity of the PT and CD cells is verified by using the segment-specific markers [AQP2 for CD; APN, aminopeptidase N, for PT (30)] on Western blot of total lysates, or DBA by immunofluorescence. (H) Various organs of P3 WT and Pkd1^V/Vmice. B, brain; L, liver; K, kidney; Lu, lung; H, heart. (I) Organs of WT mice at P19. S, spleen; P, pancreas. Note the low expression level of PC1 in kidney. Arrowhead in G and H indicates the high-molecular-weight band in Pkd1^V/V cells and organs that probably represents a modified uFL form.

We found that PC1 of WT MEF is incompletely cleaved, as evidenced by the detection of both uFL-PC1 of ≈520 kDa and CTF of ≈150 kDa (Fig. 5C), a pattern similar to that of endogenous human PC1 of HEK cells. The identity of these bands is confirmed by the finding that they were detected in the Pkd1^V/+ MEF, but not in Pkd1^−/− MEF. Cleavage of the endogenous PC1 of WT MEF is more extensive than that of the recombinant PC1. In the Pkd1^V/V MEF, only the ≈520-kDa PC1^V, and no CTF, was detected. This result indicates that noncleavable PC1^V was indeed expressed in Pkd1^V/V MEF, at a level 1.5–2 times the uFL-PC1 of WT MEF. This result is further confirmed by positive and negative detection of NTF by anti-LRR in WT and Pkd1^V/V MEF, respectively (data not shown). Similar cleavage patterns were found in E12 embryos (Fig. 5D).

PC1 cleavage also occurs in WT embryos at various later stages (Fig. 1E), when various abnormalities occur in Pkd1^−/− but not in Pkd1^V/V mice. This finding shows that PC1^V is sufficient to overcome embryonic lethality usually seen in Pkd1^−/− mice, and the GPS-cleaved PC1 products are not essential for embryonic development.

PC1 Cleavage Patterns in Postnatal and Adult Mice.

We next examined the cleavage pattern of PC1 in postnatal kidneys, when rapid tubular cystic dilation is occurring in the Pkd1^V/V kidneys. As shown in Fig. 5F, PC1 is extensively cleaved in WT kidneys at P7 and P14, whereas PC1 became undetectable at P21. PC1^V was expressed in the cystic kidneys of P7 and P14 Pkd1^V/V littermates at higher levels than uFL-PC1, but was hardly detectable at P21. We further demonstrated that cleavage of PC1 occurred in the CD cells isolated from P5, P10, and P14 WT kidneys (Fig. 5G). PC1^V was expressed in the cystic CD cells isolated from the respective Pkd1^V/V littermates at levels comparable to the WT uFL-PC1 (Fig. 5G), thus excluding a CD-specific down-regulation of the PC1^V expression as a trivial explanation for its cystic changes. This result indicates that PC1^V is not sufficient for the integrity of CD (except the papilla tip) during the postnatal maturation stage and that the cleaved PC1 products are most probably required for the process.

PC1 was also cleaved in primary PT cells isolated from P5 and P10 WT kidneys (Fig. 5G), as well as in primary human PT cells from an adult kidney (data not shown). PC1^V was expressed in the PT of the Pkd1^V/V littermates at a level similar to the uFL-PC1. This finding indicates that PC1^V is sufficient for the normal structure of PT, whereas the cleaved PC1 is not essential.

We finally analyzed the cleavage patterns in various organs of WT postnatal and adult mice. We found that PC1 was extensively cleaved at P3 (Fig. 1H, WT), P7, P14 (data not shown), and P19 (Fig. 5I, including spleen and pancreas). Similar cleavage patterns were found for the organs of adult mice at postnatal 6 weeks and 3 months (data not shown). PC1^V was not cleaved at all times at P3 (Fig. 1 H, V/V) or P7 (data not shown), and was expressed at a level two to three times the corresponding uFL-PC1. Our results demonstrate that cleavage of endogenous PC1 is a conserved and ubiquitous phenomenon.

Discussion

To investigate the function of GPS cleavage of PC1 in vivo, we generated and characterized the Pkd1^V/V mouse without GPS cleavage. The Pkd1^V/V mice show several striking features that are not seen in Pkd1^−/− mice: (i) they are born with normal Mendelian frequency and are viable; (ii) they show rapid cystic dilation in CD (except the papilla tip) and DCT, but not in the proximal portion of the nephron, during the postnatal period, and die with severe uremia, mostly at 3 weeks of age; (iii) they show dilation of the common bile duct and intrahepatic biliary ducts, but develop a normal pancreas within their life span. We conclude from these data that cleavage is required for the integrity of almost all the distal nephron segments, but not for proximal nephron segments, during the postnatal maturation period. It is also critical for the normal structure of common bile duct and intrahepatic bile tracts. It is, however, not essential for the embryonic development of various organ systems including kidney and pancreas.

To understand the basis of hypomorphic phenotype of the Pkd1^V/V mice, we characterized PC1 cleavage patterns in WT and mutant mice. We found that endogenous PC1 is ubiquitously but incompletely cleaved in WT mice at various developmental stages, as is endogenous human PC1 in HEK and renal epithelial cells. Cleavage results in two classes of PC1 molecules: a small amount of uFL-PC1 and the more abundant cleaved PC1 products. In the Pkd1^V/V mice, the mutant PC1^V is not cleaved at all times and is expressed with overall stage- and tissue-specific patterns similar to uFL-PC1, in general, at two to three times the levels uFL-PC1. This modestly higher level of PC1^V over the WT uFL-PC1 per se is unlikely to be the reason for the cystic change in Pkd1^V/V mice, because hemizygous Pkd1^V/− mice, with the dosage of PC1^V presumably reduced to half, showed a phenotype similar to the Pkd1^V/V mice (data not shown). PC1^V is most probably functionally equivalent to uFL-PC1 based on the nature of the T3041V mutation chosen: (i) it is precisely at the +1 position of the cleavage site HL̂T of the GPS domain, which is so far known only for the function as a cis-autoproteolytic signal; (ii) Thr and Val differ solely by one functional group at the very terminus of the side chain (OH vs. CH₃). Although effectively disrupting cleavage of PC1 by preventing the nucleophilic attack, the critical initial step of cis-autoproteolysis (13), this smallest possible change seems less likely to significantly alter the conformation surrounding the cleavage site or the overall conformation in PC1^V. Unchanged conformation by similar +1 mutations is in fact shown for the cis-autoproteolytic protein by structural analysis (24). Therefore, our knockin approach specifically devoids all GPS-cleaved products, while it preserves the uncleaved form that should represent the WT uFL-PC1. This outstanding feature could make this mouse model very informative in dissecting differential functions of uFL and cleaved PC1 molecules in vivo.

We can conclude from our data that PC1^V is sufficient to rescue the embryonic lethality and to prevent various abnormalities during embryonic development such as severe cystic expansion of kidney and pancreas, usually observed in Pkd1 knockouts. Our results suggest that uFL-PC1 plays a critical role during embryonic development, whereas the GPS-cleaved PC1 products are not essential in this process.

Our data also suggest a differential role of uFL-PC1 and cleaved PC1 molecules for the proximal and distal nephron segments during postnatal phase. We find that PC1 is extensively cleaved in postnatal kidneys, both in the CD and PT cells, at P3–P14 of WT mice. In the Pkd1^V/V CD, the cystic change occurs in the presence of PC1^V but in the absence of any cleaved PC1 products. Therefore, although uFL-PC1 appears sufficient for the embryonic development of the CD, it is no longer enough for its postnatal maturation and maintenance, and the cleaved PC1 molecules are required instead. Consistent with this notion, Pkd1^V/+ heterozygotes, which express slightly more uFL-PC1 but less cleaved PC1 products as compared with WT mice, appear normal. However, PC1^V is sufficient to protect the PT from cystic dilation, indicating that uFL-PC1 is fully sufficient for the normal structure of PT and the cleaved products are not essential.

There are several possible explanations for the apparent nonessential role of cleavage in the PT and possibly in the papilla tip: (i) the cleaved products may be functional but redundant to other PC1 family members; (ii) the cleavage products may be inactive by-products of a default cis-autoproteolysis, and cleavage could serve to keep the active uFL-PC1 at a low level, possibly because a high level is detrimental in proximal nephron segments and the papilla tip; and (iii) unknown GPS-independent cleavage(s) not detected in our study could occur to release active fragments that compensate the loss of GPS cleavage products in Pkd1^V/V PT. Further studies are required to resolve this issue.

Our finding of ubiquitous cleavage of endogenous PC1 is at odds with the results of most reports (19–21) using anti-PC1 antibodies directed to the CTF. These antibodies detected high-molecular-mass bands (>400 kDa) that were interpreted as being PC1, but did not recognize the far more abundant CTF band on the same Western blots. The possibility of a nonspecific detection by these reagents should be considered.

In conclusion, our results support the hypothesis that uFL-PC1 and GPS-cleaved PC1 products exert differential functions in vivo, with uFL-PC1 playing a more critical role during embryonic development, and the cleaved PC1 products being essential for distal nephron segments during the postnatal development of the mouse kidney and the common bile duct and biliary tracts. Intracellular locations and functions of each PC1 molecules remain to be determined.

Materials and Methods

Generation of Pkd1^V/V Knockin Mice.

The targeting vector (Fig. 1A i) is made by using genomic DNA fragments derived from 129Sv strain. The T3041V mutation was introduced by using the QuikChange kit (Stratagene). Electroporation, selection, and screening of ES cells (129/S6, CMTI-1) were carried out by using standard procedures. One clone (G8) with correct targeting was identified by using Southern blot analysis and long-range genomic PCR and was injected into C57Bl6 blastocysts to generate chimeric mice that produced 129S6/C57Bl6 offspring carrying the targeted Pkd1^NeoV allele. The F1 Pkd1^NeoV/+ mice were bred with FLP transgenic mice to delete Neo. The resulting Pkd1^V/+ offspring was intercrossed to generate Pkd1^V/V mice, which were backcrossed onto the C57Bl6 strain for more than five generations for our analyses.

Genotyping of Mice.

Initial genotyping was performed by using a three-primer PCR-based strategy (Fig. 1C i) with two forward primers, F24 and F2, and one reverse primer R1. The FLP transgene was identified by PCR amplification of a 399-bp product with use of primers YSQ-F2 and YSQ-R (Fig. 1C ii). For genotyping of offspring of subsequent Pkd1^V/+ matings, only F2 and R1 primers were used. Details regarding primer sequence and PCR conditions are available on request.

Reverse Transcription-PCR.

First-strand cDNA was synthesized by using SuperScript II (Invitrogen) from 100 ng of total RNA and then used as template for PCR amplification of Pkd1 transcripts with the primers F2 and R24 (Fig. 1D). The PCR products were sequenced by using HXB-F15 within exon 25 (Fig. 1E).

Histology and Immunohistochemistry.

Kidney and other specimens from different stages of animals were collected, fixed, embedded in paraffin, and sectioned at 5 μm thickness for histological analysis as described in ref. 6. H&E and MT stains were applied by using standard protocols. Sections were first microwaved for 10 min in citrate buffer solution (pH 6.0) to enhance antigen retrieval and then subjected to immunohistochemistry studies as described in ref. 6. Indirect fluorescence microscopy was performed by using a Nikon Eclipse E600, and images were captured by using a SPOT-RT monochromic camera.

Cultured Cells.

Conditionally immortalized MEFs were isolated from E12 embryos containing a thermolabile SV40 tAg transgene (25) by using standard procedures. Cells were maintained at 33°C with IFN-γ and were moved to 37°C without IFN-γ for 3 days before harvest for analysis. PT and CD cells were affinity-purified from postnatal kidneys by using LTL and DBA-conjugated Dynabeads (Invitrogen), respectively. In brief, the kidneys were minced into small pieces and digested in MEM/F12 containing 0.2% collagenase, 0.2% hyaluronidase, and 0.001% DNase I at 37°C for 2 h with gentle agitation. The digested tissue was incubated with Dynabeads at 4°C for 30 min. The cells bound to the beads were collected either directly for cell lysis (12) or, in the case of conditional immortalization, were resuspended in culture medium for propagation.

Anti-PC1 Antibodies.

To generate the polyclonal CC antibody, a cDNA fragment corresponding to amino acid residues 4123–4291 of mouse PC1 was cloned into pET28c (Novagen). After expression in Escherichia coli, the protein was affinity-purified with Ni-agarose beads and then used to immunize rabbits and chickens. Chicken IgY antibody was purified by using the EGGstract IgY Purification System (Promega). Typically, purified chicken anti-CC was used for IP and rabbit anti-CC was used for Western blot detection. Anti-CT and anti-LRR have been described (12).

IP and Western Blot Analysis.

Tissue samples taken from killed mice were immediately homogenized in lysis buffer (12) with a Polytron homogenizer (Kinematica). The homogenate was incubated for 1 h on ice and cleared of debris by centrifugation at 17,000 × g for 10 min at 4°C. Ten milligrams of protein in 1 ml was typically used for IP and Western blot analysis as described in ref. 12.