Ethics statement

All mouse studies were performed using protocol ASP-K001-KDB-19, approved by the Animal Research Advisory Committee (ARAC) at the NIH Intramural Research Program. Mice were kept and cared in pathogen-free animal facilities accredited by the American Association for the Accreditation of Laboratory Animal Care and meet federal (NIH) guidelines for the humane and appropriate care of laboratory animals.

Pkd1 knock-in model

Pkd1-GFP knock-in mice were generated using the CRISPR/Cas9 technology. Briefly, one sgRNA (GGTCCACCCCAGCAGCACTT) targeting the last exon of Pkd1 was designed to cut near the stop codon. The donor DNA construct was made by flanking the eGFP coding region, followed by 3-HA tags, with 5’ (2992 bp) and 3’ (3037bp) homologous arms of the cut site, introducing a silent mutation in the last threonine codon from ACT to ACA (to prevent the donor plasmid from being targeted by Cas9) and then inserting the GFP-3HA between the last codon and the stop codon (TAG). The donor DNA construct (10ng/ul) was co-microinjected with Cas9 mRNA (50ng/ul) and sgRNA (20ng/ug) into the pronuclei of fertilized eggs collected from B6D2F1/J mice (JAX #100006). The injected embryos were cultured overnight in M16 medium overnight, and those embryos which reached 2-cell stage of development were implanted into the oviducts of pseudo-pregnant foster mothers (CD-1 mice from Charles River Laboratory). Offspring born to the foster mothers were genotyped by PCR. Two male mice with confirmed insertion were bred: one was a chimeric mouse with likely no germline insertion, and the other produced pups carrying the knock-in and was therefore our founder mouse. Genomic DNA of the founder mouse was used for long-range PCR of regions flanking the homology arms, GFP-3HA sequence, and all exons in the region, which were sequenced and confirmed to be without mutations (file with sequence and chromatograms available upon request). This animal was bred with C57BL/6 mice to establish the knock-in mouse line. F1 mice were either back-crossed to C57BL/6 mice (up to F4) or interbred. Subsequent genotyping of this line was done with the primers: E46-F: 5’-TGCTTGTCCAGTTTGACCGA with eGFP-R: 5’-GCTGAACTTGTGGCCGTTTA and 3UTR-R: 5’-ATGGCCACCTAGGGGTAGAG (wild type product: 614 bp; knock-in product: 332 bp). Both C57BL6J and C57BL6NJ were inadvertently used interchangeably through the initial phases of this work. C57BL/6J (JAX strain #000664) and C57BL/6NJ (JAX strain #005304) strains carry nicotinamide nucleotide transhydrogenase (Nnt) mutant and wild type alleles, respectively. Nnt genotype was determined by PCR as described in [24] and except when specifically noted, only Nnt wild type mice were used. Briefly, the Nnt PCR reaction identifies a 312bp product only in Nnt wild type mice using primers: Exon6_L1: CAATTCTGCCAACAACTGGA and Exon6_R4: GGTCACTCTGGGCACTGTTT; and a 547bp product only in mutants using Exon12_L1: GTAGGGCCAACTGTTTCTGC and Exon6_L4: TCCCCTCCCTTCCATTTAGT.

Mouse studies

Pkd1^ko mice were previously described [25]. Mice homozygous for the Pkd1^ko allele lack exons 2 to 4 and die in utero. Pkd1^cko mice have loxP inserted into introns 1 and 4, and generates the deleted allele Pkd1^del2-4 upon cre-expression [25]. Tamoxifen-inducible Cre-ER transgenic mice (JAX strain #004682 [26]) and Ksp-cre transgenic mice (JAX strain #012237 [27]) expressing cre recombinase in kidney tubular epithelial cells were used to induce Pkd1 inactivation in the kidney. In the Cre-ER line, mice were induced at post-natal day 40 (P40) with 0.4mg/g tamoxifen (Sigma, T5648) diluted 20mg/ml in corn oil and injected intra-peritoneally under brief anesthesia with 3–5% isoflurane administered within a chamber. At the end of the studies, mice were euthanized with CO₂ administration in a chamber, followed by cervical dislocation. Throughout the study, mice were monitored and those showing signs of pain/suffering (such as hunched body posture and decreased food or water intake) were treated in consultation with the veterinary staff or euthanized as above.

Primary cell lines

Primary tubular epithelial cells (PTEC) were obtained from kidneys removed from P2 pups. Briefly, kidneys were quickly dissected after mouse euthanization and transferred to gentleMACS C Tubes (Miltenyi Biotec Inc., cat. no. # 130-093-237) in 2ml of digestion buffer (13.5 ml of Dispase (STEMCELL, cat. # 07912) with 1.5 ml of Collagenase/Hyaluronidase 10X (STEMCELL cat. # 0793), 150 μl of GlutaMax, 150 μl of 1M HEPES and 0.75 μl of 10 mg/ml DNAse I) and dissociated in gentleMACS Dissociator (Miltenyi Biotec Inc., cat. no. 130-093-235) using the Multi_E_01.01 program, followed by incubation at 37°C for 30 min. under gentle rotation. Dissociated cells were filtered through pre-separation filters (Miltenyi Biotec Inc. cat. no. # 130-095-823) and centrifuged at 150xg for 5 min. and either directly seeded on 10cm cell culture dishes or further enriched with Lotus Tetragonolobus Lectin (LTL; Vector labs biotynilated beads, cat. no. B-1325-2) LTL or Dolichos Biflorus Agglutinin (DBA; Vector labs biotinylated beads, cat. no. B-1035-5) beads and CELLection Biotin Binder Kit (Invitrogen, cat. no. 11533D). Cells were grown at 37°C with 5% CO₂ in low serum media DMEM/F12 media (Life cat. no. 21041–025) with 2% FBS (GEMINI Bio-Products cat. no. 100–106), 1 x Insulin-Transferrin-Selenium (Thermo Fisher Scientific, cat. no. 41400–045), 5 μM dexamethasone (SIGMA, cat. no. D1756), 10 ng/ml EGF (SIGMA, cat. no. SRP3196), 1 nM 3,3′,5-Triiodo-L-thyronine (SIGMA, cat. no. T6397) and 10 mM HEPES (CORNING, cat. no. 25-060-CI).

Mouse embryonic fibroblasts (MEF) were obtained from E13.5 embryos which were chopped coarsely (10–20 times) using a sterile razor blade and digested in 0.25% Trypsin-EDTA for 10 min. in a cell culture incubator (37°C; 5% CO₂) followed by dissociation in gentleMACS C Tubes (Miltenyi Biotec Inc., cat. no. #130-093-237) in a total of 8 ml DMEM media with 10% FBS using gentleMACS Dissociator Program A twice and seeded onto 15 cm tissue culture dishes and grown in DMEM with 10% FBS at 37°C with 5% CO₂.

Immunoblot and immunoprecipitation

Mouse tissue was dissociated using gentleMACS Dissociator (Miltenyi Biotec Inc., cat. no. 130-093-235) and gentleMACS M Tubes (Miltenyi Biotec Inc., cat. no. # 130-093-236) in lysis buffer (1% Triton X-100; 150 mM NaCl; 50 mM Tris HCl pH8.0, 0.25% sodium deoxycholate, 1 × protease inhibitor with EDTA (Roche Complete, cat. no. 11836153001), 1 x phosphatase inhibitor (Roche PhosSTOP, cat. no. 4906845001), 1 μl/ml Nuclease (Pierce, cat. no. 88700)). Immunoprecipitation of tagged PC1 was done using anti-GFP nanobodies coupled to magnetic agarose beads (GFP-trap, Chromotek cat. no. gtma-10) following the manufacturer’s protocol and Kohli et al. [28] with the following modifications. After lysates and GFP-trap were rotating end-over-end for 1 hr at 4°C, the GFP-trap/proteins complexes were separated from the lysates by a magnetic field. Non-specific binding proteins were washed away very gently and quickly (within seconds) in the cold room, twice with 500 μl of wash buffer 1 (50mM Tris (pH7.5), 150mM NaCl, 5% Glycerol, 0.05% Triton-X 100, 1 × protease inhibitor with EDTA, 1 x phosphatase inhibitor) and wash buffer 2(50mM Tris (pH7.5), 150mM NaCl). For mass spectrometry, the GFP-Trap products were eluted in 160 μl of pre-heated (60°C) 1% sodium dodecanoate (DODEC buffer) [29], transferred to the pre-heated glass vials, then heated at 60°C for 20 minutes. For immunoblotting, samples were denatured in 1:1 with 2 x SDS buffer with 1x reducing agent (Invitrogen, NP0004) for 10 min at 95°C, loaded into 4–12% bis tris 1.5 mm gels (Invitrogen, NP0336BOX), run in MES buffer (Invitrogen, NP0002) at 200V for 45 min., transferred to nitrocellulose membrane (Invitrogen, IB301002) using iBlot dry blotting transfer system (Invitrogen) program 3 and blocked with Intercept Blocking Buffer (Licor, cat. no. 927–70001), incubated with antibodies and imaged using Li-Cor Odyssey Imaging System following the manufacturer’s instructions and the following antibodies: anti-eGFP (Aves, cat. no. GFP-1020), anti-HA (MBL, cat. no. M180-3), anti-NNT (Thermo Fisher Scientific, cat. no. 13442-2-AP or Santa Cruz Biotechnology, cat. no. sc-390215), anti-Polycystin-2 (PC2-CT [30], PKD-RCC (https://www.pkd-rrc.org/) and anti-Polycystin-1 (7E12, Santa Cruz cat. no. sc-130554), anti-β -actin (CST, cat. no. 4967S) and secondary labeled antibodies (Licor, cat. no. 925–68072 or 926–32211).

Mass spectrometry

For each mass spectrometry experiment, 8 heads of P1 newborn mice of each genotype (Pkd1^eGFP/eGFP and Pkd1^wt/wt) were used as input for immunoprecipitation (approximately 80mg of protein crude lysate/genotype). Briefly, two heads were combined, dissociated and lysed as described above. Protein concentration was measured using BCA assay (Pierce, cat. no. 23225) and 20mg were used for immunoprecipation (IP) with GFP-trap. The eluates of the four independent IPs for each genotype were then pooled and immediately processed for mass spectrometry. This procedure was repeated three times (using a total of 24 heads). A bottom-up proteomic analysis used a soap to assist in rendering and digestion of sample proteins [29, 31], post-digestion reductive demethylation [32, 33], neutral pH off line reversed phase separation using fraction concatenation [34] conventional High-Low microscale HPLC/mass spectrometry, and processed with MaxQuant [35]. Given the expectation that there should be enrichment of pulled-down material in the samples containing the bait, we used non-normalized ratios for the analyses [36]. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [37] partner repository with the dataset identifier PXD035250 and 10.6019/PXD035250.

Interactome analyses

Enrichment ratios of protein abundance in PC1-eGFP vs. control samples were obtained in MaxQuant [35] and imported into R [38] for further analyses. Gene set enrichment analyses using fgsea [39] was performed on the ordered values of the log₂ of the mean ratios (across replicates). Pathway enrichment analyses were performed on lists of proteins with ratios above 1 in all experiments and ratio above 2 in at least two experiments (considered part of the PC1 extended interactome) and proteins sometimes enriched in PC1-eGFP, sometimes in the control groups, with ratios above ½ and below 2 (these were considered background proteins). These lists of interactome and background proteins were analyzed in Cytoscape [40] using stringApp [41] using default conditions. Networks were visualized in Cytoscape using yFiles layouts. Graphs were plotted in R using the ggplot2 package [42]. For color-coding and sub-setting mitochondrial and ciliary proteins, we used goseq [43] and the org.Mm.eg.db database to extract genes in all gene ontology cellular component categories containing the keyword mitochondria or cilia. A list of previously published Polycystin 1 (PC1) interactors was obtained from publications [14–16, 44–49] and by querying PC1 in the STRING [50] 11.5 online database (http://string-db.org/) as an output that represents the associations in the form of a physical subnetwork with nodes (proteins/genes) and edges (interactions) to extract the 1st order interactomes. The 1st order interactome contained a max number of 50 interactors. The edges are weighted and integrated, and a confidence score is assigned to each of them based upon the evidence of the association obtained from text mining, experimental data and public databases. The high (0.70)/medium (0.40)/low (0.15) confidence scores in the database, which define and compare the significance of interactions between various queried proteins, were chosen respectively to extract STRING-based PC1 interactomes.

Histology and immunostaining

Cells were seeded at a density of 100,000 cells/well in chamber slides (Ibidi, cat. no. 80822). After 24h, cells were serum-starved for 48h, fixed for 10 min. in 10% formalin, rinsed with PBS and permeabilized with 0.5% Triton X-100 in PBS for 5 min., blocked with 1% fish gelatin in PBS, stained with antibodies to detect Arl13 (ciliary marker; ProteinTech cat. no. 17711-1-AP; secondary antibody: Invitrogen A27034 labeled with Alexa Fluor 488), γ-tubulin (ciliary basal body marker; SIGMA cat. no. T5192; secondary antibody Invitrogen A27034 labeled with Alexa Fluor 488) or eGFP (Aves cat. no. GFP-1020; secondary antibody: Invitrogen A21449 labeled with Alexa Fluor 647), mounted (Invitrogen ProLong cat. no. P36981) and imaged using Zeiss LSM700 confocal microscope. Freshly dissected embryos or kidneys were frozen in OCT (Tissue-Tek; Sakura Finetek USA) and frozen sections were processed as above; or embedded in paraffin, sectioned, and stained with trichrome staining or used for immunohistochemistry. Images were visualized using Imaris (Bitplane). Cystic index was calculated by delineating kidneys in trichrome stained slides using the magnetic lasso feature in Photoshop, and measuring total kidney area, thresholding cystic area and measuring total kidney area and cystic area using Fiji [51].

Statistics

The R environment [38] was used for statistical analysis. To minimize the effect of outliers, robust statistical estimators using the WRS2 package were used [52]. In particular, comparisons of two groups were done using the function pb2gen for t-test based on medians. Plots were made using the ggplot2 package [42]. The pedigree tree network was generated in Cytoscape [40].

Generation and characterization of Pkd1 knock-in model

Using CRISPR/Cas9 technology, we inserted eGFP followed by three human influenza hemagglutinin (HA) tags in frame with PC1 C-terminus (Fig 1A–1C). Screening of targeted mice revealed one mouse with germline knock-in carrying this expected Pkd1^eGFP allele, confirmed by long-range PCR and RT-PCR (Fig 1D and 1E). This animal was the founder of the Pkd1^eGFP knock-in line (Fig 1F). Immunoblot analyses showed that tagged polycystin-1 (PC1-eGFP) is expressed in kidneys of young pups, with detection using both the HA-tag and antibodies to the endogenous protein (Fig 1G). Using nanobodies against eGFP, PC1-eGFP can be efficiently enriched in pull-down experiments (Fig 1H–1J) and can be detected in multiple mouse tissues, including brain, lung, heart and kidney; and in primary cell lines obtained from knock-in mice.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Pkd1^eGFP knock-in mouse generation.

A) Schematics of Pkd1 knock-in allele, showing sgRNA guide sequence and location of eGFP-3xHA tag. Created with BioRender.com. B) Diagram showing steps to generate the knock-in mouse. C) Structure of the knock-in Pkd1 allele. Red arrows on the light yellow rectangle mark regions confirmed by sequencing. Small arrows in the gene diagram are the locations of primers used in the reactions in panels D-F. D) Genomic long-range PCR using primers Lr1 and the eGFP-specific Lr2 showing the predicted band of ~4kb in Pkd1^eGFP founder mouse. No bands are amplified in the control tissue (wt). E) Reverse-transcriptase PCR using primers rt1 and rt2 showing expression of the Pkd1 transcript with eGFP sequence in knock-in mouse tissue. The vector positive control is larger because it includes introns. F) Genomic PCR using primers E46-F, eGFP-R and 3UTR-R to genotype knock-in mice. As expected, eGFP homozygotes only have the ~300bp E46-F:eGFP-R product, control specimens (wt) have only the 600bp E46-F:3UTR-R product, and heterozygotes have both. G) Detection of PC-1 in kidney of wild type and knock-in mice at different post-natal days; left: detected with anti-HA antibody; right: detected with anti-PC1 (7e12) antibody. Full length, uncleaved PC1 is detected with the anti-HA antibody (left) whereas 7e12 only detected NTF under these conditions (right). β−actin was used as a loading control; 7e12 specificity is shown in S1 Raw images). H) Kidney of control and knock-in P1 mice showing tagged PC1 detection with anti-HA antibody; left: cell lysate; right: protein immunoprecipated with anti-eGFP conjugated agarose beads. Note the enrichment of CTT after IP. PC1 full-length (PC1_FL), cleavage products (PC1_CT; P100), and multiple CTT-related cleavage products (presumed CTT marked with an arrow; dotted line represents range of possible CTT products). I) Immunoprecipitation of PC1 from different tissues of P1 control and knock-in mice using anti-eGFP conjugated agarose beads and detected with anti-HA antibodies. J) Mouse primary tubular epithelial cells (PTEC) and embryonic fibroblast (MEF) cells from control and knock-in mice showing tagged PC1 in the lysate (left) and immunoprecipitated by anti-eGFP and detected with anti-HA (right).

https://doi.org/10.1371/journal.pone.0289778.g001

The founder mouse was bred to wild type and to heterozygous knock-out (Pkd1^ko/wt) mice (Fig 2A). Pkd1^eGFP/wt bred to Pkd1^ko/wt produced pups at the expected mendelian rates (Fig 2B). A cohort of 6 Pkd1^eGFP/ko mice was aged to 15 months and showed no kidney pathology, though one animal developed small liver cysts (Fig 2C–2E). Pkd1^eGFP/eGFP mice euthanized at multiple ages between birth and 450 days of age had no obvious phenotypes (>10 mice/30-day brackets). These data suggest that a C-terminal tag added to PC1 does not significantly impair its function.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Pkd1^eGFP knock-in mouse.

A) Pedigree tree of mouse colony derived from the single Pkd1^eGFP knock-in mouse (red dot at the top). Genotypes are color-coded. The blue dots represent wild type mice used for back-crossing. B) Table showing that Pkd1^eGFP/ko pups born to Pkd1^eGFP/wt vs. Pkd1^ko/wt breedings are viable. C) Kidney/body weight ratios of a subset of animals in panel B allowed to age to 15 months showing normal kidney size. D) Representative Pkd1^eGFP/ko kidney has normal morphology at 15 months of age (lower panels: trichrome staining; size bar: 1mm). E) While the majority of the Pkd1^eGFP/eGFP and Pkd1^eGFP/ko livers had normal morphology, one 15 months-old Pkd1^eGFP/ko mouse had liver cysts (middle panel; trichrome staining; size bar: 200 μm).

https://doi.org/10.1371/journal.pone.0289778.g002

One of the goals of creating a Pkd1 knock-in mouse was to detect PC1 in vivo using microscopy. So far, we have been unable to unambiguously identify PC1 in live cells/tissue or in fetal or adult tissues using antibodies that detect the HA or eGFP tags. However, we have consistently detected PC1 using antibodies to GFP in the primary cilium of kidney epithelial cells (Fig 3A and 3B) and some cilia in mouse embryonic fibroblasts (Fig 3C). We have been unable to detect PC1 unambiguously above background in other subcellular structures.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Tagged PC1 can be detected in cilia of primary cells and tissues of knock-in mice.

A) Detection of PC1 with anti-eGFP antibody in cilia of most primary tubular epithelial cells (PTEC), co-localized with Arl13; B) PC1 localizes adjacent to ciliary basal bodies, marked with anti-γ-tubulin. C) PC1 is detected in cilia of some, but not all, mouse embryonic fibroblasts (MEF) cells.

https://doi.org/10.1371/journal.pone.0289778.g003

In vivo PC1-interactome

Another goal of the Pkd1 knock-in model was to identify in vivo PC1 binding partners and to establish a PC1-interactome. Towards this goal, we optimized conditions to immunoprecipitate PC1 and detect both PC1 and PC2, a known PC1 binding partner [53]. Having confirmed this in mouse tissue (S1 Fig), we tested multiple protocols and found that reliable detection of PC1 and binding partners by mass spectrometry in IP isolates required approximately 80mg of mouse tissue. We then performed 3 sets of affinity purification mass spectrometry experiments (Fig 4A), each analyzing immunoprecipitates from the lysate of a combination of eight P1 mice heads. We selected P1 mice head because brain had one of the highest PC1 expressions at this age in both knock-in and control mice. As expected, under these conditions, both PC1 and PC2 were the proteins with highest enrichment and intensity in knock-in mouse samples compared to controls (up to 36 times) (Fig 4B, S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Identification of PC1 in vivo binding partners.

A) Schematics showing steps for proteomics experiments. Briefly, for knock-in and control samples, 4 sets of 2 heads of P1 mice per genotype were independently immunoprecipitated using eGFP beads then pooled and processed for one mass spectrometry experiment. Each mass spectrometry experiment was done three independent times. Modified from template created with BioRender.com. B) Plot showing the log₁₀(Intensity) (x-axis) and log₂(intensity ratio between knock-in and control samples) (y-axis) prepared from P1 mouse heads. Each dot corresponds to one detected protein; shape (circle, square, triangle) represents experimental batch; the colors summarize the number of times the enrichment ratio was above 2 for the protein. Proteins that were never enriched in the knock-in IP are orange; proteins enriched in all three experiments are colored purple and are identified by purple text. PC1 and its primary binding partner PC2 were reproducibly the top hits in each IP study.

https://doi.org/10.1371/journal.pone.0289778.g004

The protocol was optimized with low-stringency, fast washes, with the goal of immunoprecipitating (IP) weak interactors [54]. Consequently, both experimental (PC1-eGFP) and control IPs from wild type mice lacking the eGFP-HA tag identified similar proteins, reflecting a somewhat high background (S1 Table). A standard way to identify true interactors under these conditions is to focus on consistent enrichment across replicates in experimental vs. control conditions [54]. Using this method, we identified 83 proteins with a ratio >2 in all three experiments (Fig 4B). For this study, however, we considered part of the PC1 extended interactome all proteins that were enriched (ratio > 1) in all PC1-eGFP IP mass spectrometry experiments (n = 3) and had a ratio of at least 2 in at least 2 of those (S1 Table). With these criteria, 823 proteins were identified (S2 Table).

PC1-interactome is enriched in metabolic pathways and mitochondrial proteins

Using stringApp [41] in Cytoscape [40], we obtained the protein network of reported physical interactions within these proteins in the STRING database [50] and identified a large cluster consisting of 522 proteins linked by over one thousand interactions, consistent with our expectation that the immunoprecipitated proteins were part of an extended network of interactors (Fig 5A). Pathway analyses showed enrichment of cytosolic (p<10⁻⁴⁸) and mitochondrial proteins (p<10⁻²⁵) (S3 Table). Using a set of likely background proteins (the 991 proteins with no ratio above 2 or below ½ that were sometimes enriched in PC1-eGFP and sometimes in control samples), similar analyses showed that cytosolic and mitochondrial proteins were among those with highest change in significance between the background and interactome sets (Fig 5B), whereas nuclear proteins were enriched in the background, but depleted in the interactome. While ciliary and cytoskeletal proteins were detected in the interactome (S2 Fig; S2 Table), neither category was enriched. Functionally, the PC1 interactome was enriched in oxidation-reduction pathways (p<10⁻¹⁸) and several of the metabolic pathways described in Wikipathways [55] (Fig 5C and 5D), including proteins involved in TCA cycle, glycolysis and fatty acid metabolism (Fig 5E), consistent with PC1 being a part of a network of proteins involved in metabolic and mitochondrial functions. Among signaling cascades, MAPK signaling was significantly enriched (Fig 5E). Gene set enrichment analysis (GSEA) of the interactome identified NADP metabolism as one of the top pathways (Fig 5F) and NAD(P) binding domains (Fig 5G) as the top enriched InterPro domain [56].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. PC1 interactome.

A) Physical interactions described in the STRING database connecting a large cluster (522 nodes) of the interactome. B and C) Pathway analyses of the extended PC1 interactome compared to background proteins using the Gene Ontology Cellular Component (D) and Gene Ontology Biological Processes (E). D) Network showing Wikipathways (green nodes) significantly enriched (p<0.05) in the interactome and the corresponding proteins (clear nodes). E) Physical interactions described in the STRING database connecting proteins in the PC1 extended interactome that were in some of the enriched pathways. F) Plot of Gene Ontology Biological Processes categories identified in gene set enrichment analysis (GSEA) of the extended PC1 interactome. Each dot corresponds to a category and its size correlates with the number of proteins in the category. The axes show -log₁₀(FDR-adjusted p-value) and normalized enrichment score (positive for categories enriched in the interactome). Note that NADP metabolic processes is one of the top enriched categories. G) Pathway analyses of the extended PC1 interactome compared to background proteins using the InterPro domains database. Note the enrichment of NAD(P) binding domains.

https://doi.org/10.1371/journal.pone.0289778.g005

Previous reports have suggested mitochondrial dysfunction and reactive oxygen species (ROS) damage in PKD [18, 57–60]. GSH is a scavenging antioxidant that forms one of the main lines of defense against reactive species in most cells and NADPH supply is reported to be limiting in GSH synthesis [61, 62] in certain conditions. One of the proteins in the interactome involved in NADP metabolism is nicotinamide nucleotide transhydrogenase (NNT), an inner mitochondrial membrane protein that uses H⁺ re-entry into the mitochondrial matrix and NADH to reduce NADP⁺ into NADPH. In fact, among the 161 proteins in the PC1 extended interactome that were in mitochondrial gene ontology categories (S2 Table; Fig 6A), NNT was one of the top hits. It was enriched in all replicates, represented by peptides spanning the length of the protein (S3B Fig) and confirmed in PC1-immunoprecipitates of P1 heads and P2 kidneys of the PC1-eGFP knock-in mice (Fig 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. NNT is likely a PC1 interactor.

A) Plot of immunoprecipitated proteins belonging to mitochondrial Gene Ontology Cellular Component categories. The axes show the log₁₀(intensity) and log₂(intensity ratio between knock-in and control samples). Each dot corresponds to one detected protein; shape (circle, square, triangle) represents experimental batch; the colors summarize the number of times the enrichment ratio was above 2 for the protein. Proteins that were never enriched in the knock-in IP are orange; proteins enriched in all three experiments are colored purple. Note NNT is one of the top reproducibly immunoprecipitated mitochondrial proteins. B) Immunoblot showing protein immunoprecipitated using anti-eGFP nanobody in control (wt) and Pkd1^eGFP/eGFP P1 heads and P2 kidneys and detected with anti-NNT antibodies. The doted box indicates the region with NNT bands. Note that immunoprecipitation conditions for head and left kidney panels followed low-stringency washing conditions used for the mass spectrometry studies, resulting in variable levels of NNT detection in control samples, albeit with overall enrichment in Pkd1^eGFP/eGFP samples compared to control samples. The washing and blocking conditions for the P2 kidney sample on the right panels were more stringent.

https://doi.org/10.1371/journal.pone.0289778.g006

There are two commonly used C57BL/6 mouse lines that differ with respect to their Nnt status. The C57BL/6NJ line carries wild type Nnt alleles while the C57BL/6J lines has an exon 1 missense mutation and a multiexon deletion that severely compromises Nnt function [24, 63]. We had the Pkd1^cond allele in both the C57BL/6J and C57BL/6NJ strains in our colony so we investigated if a PC1-NNT interaction could have functional consequences by comparing kidney disease severity in an early-onset (Ksp-Cre) and late-onset (inducible tamoxifen-cre) disease model. In the Ksp-cre model, at age 8 days (P8) Pkd1 mutants have highly cystic kidneys but there was no significant difference between the Nnt wild type and mutant strains, though disease severity was more variable in the Nnt wild type background (Fig 7A and S4 Table; n = 10 Nnt mutants and 34 controls). We also saw no effect of Nnt status on disease severity in mice induced at P40 and harvested between P180 and P200 (Fig 7B and S5 Table; n = 6–7 per group).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Nnt status and PKD progression.

A and B) Nnt status does not correlate with kidney/body weight. Kidney/body weight in Pkd1 mutants (blue) and controls (red) by Nnt status in P8 mice induced with nephron-specific Ksp-cre (B; n = 10 Pkd1/Nnt mutants and 34 Pkd1 mutant/Nnt controls) or induced with tamoxifen at P40 and harvested between P180-P200 (n = 6–7 of each sex/group). Sex of each mouse is identified by its shape (triangles: males; circles: females).

https://doi.org/10.1371/journal.pone.0289778.g007

These data suggest that interaction between PC1 and Nnt, also recently reported by another group [64], is not per se sufficient to significantly alter the cystic phenotype in Pkd1^-/- between Nnt wild type and mutant strains.

Previously reported PC1 interactors in the PC1 interactome

We also compared the extended PC1 interactome with a list of proteins previously reported to be binding partners of PC1. We used two approaches to identify targets: a) the STRING database, selecting targets either experimentally validated, present in curated databases or identified through text mining; b) targets reported in the literature but not necessarily present in STRING (S6 Table). We identified a total of 7 proteins present in both our dataset and other sources. Two were in the mouse proteome, 1 in human, and three in both mouse and human (Fig 8). The seventh was solely in Xenopus. Except for PC2, none of the hits was in the set of 84 proteins enriched >2 in all experiments.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Overlap of PC1-interactome with previously identified PC1-interactors.

Venn diagram showing the overlap between the PC1-interactome with targets previously reported for mouse and human (one interactor identified in Xenopus is not shown).

https://doi.org/10.1371/journal.pone.0289778.g008

Given that the extended PC1-interactome was greatly enriched for mitochondrial and cytoplasmic targets, we reviewed the pattern of PC1 peptides isolated by our IP method to exclude a CTT bias. As expected, given that full-length PC1 was detected by immunoblot in the IP eluent (S1 Fig), peptides in our mass spectrometry studies mapped along the length of PC1 (S3 Fig). These data suggest that our strategy should have captured targets for any of the domains of PC1, excluding a bias towards enrichment of just CTT targets.