Targeting an anchored phosphatase-deacetylase unit restores renal ciliary homeostasis

Pathophysiological defects in water homeostasis can lead to renal failure. Likewise, common genetic disorders associated with abnormal cytoskeletal dynamics in the kidney collecting ducts and perturbed calcium and cAMP signaling in the ciliary compartment contribute to chronic kidney failure. We show that collecting ducts in mice lacking the A-Kinase anchoring protein AKAP220 exhibit enhanced development of primary cilia. Mechanistic studies reveal that AKAP220-associated protein phosphatase 1 (PP1) mediates this phenotype by promoting changes in the stability of histone deacetylase 6 (HDAC6) with concomitant defects in actin dynamics. This proceeds through a previously unrecognized adaptor function for PP1 as all ciliogenesis and cytoskeletal phenotypes are recapitulated in mIMCD3 knock-in cells expressing a phosphatase-targeting defective AKAP220-ΔPP1 mutant. Pharmacological blocking of local HDAC6 activity alters cilia development and reduces cystogenesis in kidney-on-chip and organoid models. These findings identify the AKAP220-PPI-HDAC6 pathway as a key effector in primary cilia development.


Introduction
Kidneys recycle about 180 liters of fluid every day to partition nutrients and remove toxins from blood (Saborio et al., 2000). Water reabsorption from luminal fluid is triggered by the hormone arginine vasopressin via phosphorylation-dependent translocation of aquaporin-2 water pores to apical surface of kidney collecting ducts (Bankir et al., 2013;Noda et al., 2010;Yui et al., 2012).
Not surprisingly, defects in renal water homeostasis have pathophysiological consequences.
Approximately 35 million Americans suffer from chronic kidney diseases, characterized as a gradual loss of renal function (Hemmelgarn et al., 2006). Polycystic kidney diseases are disorders where the collecting ducts become enlarged with fluid filled cysts that reduce glomerular filtration rate (Wilson, 2004). Autosomal dominant polycystic kidney disease (ADPKD) with an estimated prevalence of 1 in 600 people, is a common genetic disorder associated with end-stage renal failure (Halvorson et al., 2010). Clinical evidence indicates that primary cilia function is altered in ADPKD (Ma et al., 2013(Ma et al., , 2017. Hence this chronic renal disorder is classified as a ciliopathy (Badano et al., 2006;Fliegauf et al., 2007).
Both proteins are components of a receptor-channel complex that responds to local second messenger signals (Harris and Torres, 2014). Accordingly, a ciliary hypothesis has been formulated that implicates defective calcium and cAMP signaling in the ciliary compartment as a factor in disease progression (Ma et  Here we show that AKAP220 -/mice exhibit enhanced development of primary cilia. Mechanistic studies reveal that AKAP220 associated PP1 drives this phenotype by promoting changes in actin dynamics and histone deacetylase 6 (HDAC6) stability. Pharmacological blocking of local HDAC6 activity alters cilia development and reduces cystogenesis in cellular models of ADPKD. These findings point towards the AKAP220-PPI-HDAC6 pathway as a therapeutic target for the treatment of ADPKD.

Results
Loss of AKAP220 promotes renal cilia assembly.
AKAP220 modulates cytoskeletal signaling events through its ability to recruit kinases, phosphatases and the small GTPase effector protein, IQGAP1 (Fig 1A; (Logue et al., 2011)). AKAP220 -/mice display mild defects in water homeostasis and altered aquaporin 2 (AQP2) trafficking that are linked to disruption of an apical actin barrier in the kidney collecting ducts ( Fig   1B; (Whiting et al., 2016)). Further investigation in kidney sections from wildtype and AKAP220 -/mice led to the unexpected discovery that deletion of this anchoring protein correlated with increased numbers of primary cilia decorating each collecting duct (Fig 1C-H). The GTPase Arl13b (red) served as a ciliary marker, staining of AQP2 (green) marked kidney collecting ducts and DAPI (blue) highlighted nuclei (Fig 1C & D). Analysis of tissue sections collected from several animals are presented in Fig 1I, measuring a 3.5-fold increase in primary cilia in AKAP220 -/-, as compared to wildtype. Additional measurements determined that cilia in AKAP220-/-collecting ducts are 1.68-fold longer than their wildtype counterparts (1.7 µm vs 1.02 µm; Fig 1J).
Independent validation of this observation was provided when CRISPR-Cas9 gene-editing was used to delete AKAP220 from mouse Inner Medullary Collecting Duct (mIMCD3) cells.
Immunoblot analysis confirmed the loss of AKAP220 in two independent clones ( Fig 1K, top panel, lanes 2 & 3). GAPDH was used as loading control (Fig 1K, bottom panel).
Immunofluorescent detection of primary cilia in each AKAP220KO clone measured an 8 to 10fold increase in the percentage of ciliated cells in comparison to wildtype mIMCD3 cells (Fig 1L-O). Quantification is presented in Figure 1O. Serum starvation is typically used to ciliate mIMCD3 cells in culture (Westlake et al., 2011). However, serum starvation of AKAP220KO cell lines enhanced detection of multi-ciliated cells (supp Fig 1A-G). This made single cell quantification less accurate. For this reason, all further analyzes were conducted in asynchronous cultures.
Rescue experiments allowed us to attribute the increases in cilia to loss of AKAP220 (Fig 1P-U).
Immunofluorescent detection of Arl13b (red) detected cilia and GFP-fluorescence confirmed rescue with the anchoring protein (Fig 1Q & R). GFP served as a control (Fig 1S & T).
Quantification from three independent experiments confirmed that rescue with AKAP220 dramatically reduced the number of ciliated cells as compared to rescue with GFP ( Fig 1U).
Since another anchoring protein AKAP150 has been detected in primary cilia, it was important to establish if this AKAP also contributed to primary cilia development (Choi et al., 2011). Geneediting was used to generate mIMCD3 cells lacking AKAP150. Double knockout cells lacking AKAP150 and AKAP220 were also produced. Immunofluorescent detection of ARl13b (green) and acetyl-tubulin (red) as ciliary markers revealed that the loss of AKAP150 alone had no effect on cilia development (Supp Fig 1H & I). Yet, a pronounced increase in the ciliated population was observed in AKAP220/150 double knockout cells (Supp Fig 1J & K). These additional studies indicate that AKAP150 signaling does not support renal ciliogenesis.
Primary cilia formation requires tubulin heterodimers that are stabilized by acetylation on lysine 40 (Portran et al., 2017). This prompted us to consider whether acetylated tubulin was abundant in cells lacking AKAP220. Experiments were conducted in two phases. First, immunofluorescent detection of acetylated alpha tubulin confirmed that this modified form was prevalent in AKAP220KO cells (Fig 2A & B). Second, immunoblot analysis showed that acetyl tubulin was more prominent in AKAP220KO cell lysates as compared to wildtype (Fig 2C, top panel). Alpha tubulin was consistent in both cell lines (Fig 2C, mid panel). GAPDH was the loading control Amalgamated data from five independent experiments indicated that deletion of AKAP220 correlated with a 50% reduction of total HDAC6 in mIMCD3 cells (Fig 2J). This argues that interface with AKAP220 serves to stabilize HDAC6.
PLA puncta indicative of AKAP220/HDAC6 complexes were readily detected in cells transfected with V5-tagged AKAP220 (Fig 2K). Grey scale image emphasizes the distribution of AKAP220/HDAC6 complexes (Fig 2L). In contrast, detection of PLA puncta was dramatically reduced in control experiments performed in untransfected cells or in the absence of HDAC6 or V5 antibodies (Fig 2M & N, Supp Fig 2E-I). GFP was used as a transfection marker. Quantification from three independent experiments is presented ( Fig 2O).
Fourth, we performed cycloheximide-chase experiments. Cells lysates were prepared at selected time points (0-4 hours; Fig 2P). Immunoblot detection of HDAC6 (top panel) and acetylated tubulin (mid panel) monitored the stability and activity of the de-acetylase over time. GAPDH served as a loading control (bottom panel). In wildtype mIMCD3 cells, HDAC6 protein was relatively constant and acetylated tubulin levels were low ( Fig 2P, lanes 1-3). Conversely, upon loss of the anchoring protein, HDAC6 levels were markedly reduced over time and acetylated tubulin levels were elevated (Fig 2P, lanes 4-6). Densitometric analysis of HDAC6 levels in wildtype (grey) and AKAP220KO (green) from three independent experiments are presented ( Fig 2Q). These results show a significant decrease in HDAC6 stability in the absence of the anchoring protein (Fig 2P &   Q). Finally, overexpression of murine HDAC6 (cyan) abrogated primary cilia formation in AKAP220KO cells (Fig 2R & S). In contrast, expression of a catalytically inactive HDAC6 mutant (H216A, H611A) had no effect on cilia formation in the AKAP220 null background (Fig 2T & U). Amalgamated data from three experiments are presented ( Fig 2V). Collectively, these studies implicate HDAC6 activity in AKAP220-mediated control of primary cilia development.
Mechanistically, dephosphorylation of HDAC6 favors its degradation (Ran et al., 2020). Protein phosphatase 1(PP1) is a well-characterized binding partner of AKAP220 that also interacts with  Co-immunoprecipitation assays confirmed that AKAP220-ΔPP1 is unable to anchor PP1 (Supp Results from at least three independent experiments are quantified ( Fig 3N). Parallel analyses established that acetyl-tubulin was elevated in AKAP220-ΔPP1 (Fig 3J & L). This is highlighted in insets and 3D surface plots of acetylated tubulin (Fig 3K & M). Collectively, these results indicate that PP1 anchoring to AKAP220 is necessary for regulation of ciliary development and its loss is sufficient to drive enhanced ciliation. We reasoned that PP1 may serve as an adaptor protein that incorporates histone deacetylase 6 into AKAP220 signaling complexes.
A fluorometric assay monitored HDAC6 activity in each cell type ( Fig 3O). As expected, HDAC6 activity was reduced in AKAP220KO and AKAP220-ΔPP1, as compared to the wildtype mIMCD3 cells. Amalgamated data from four experiments are presented ( Fig 3O). These findings suggest that the loss of HDAC6 activity and elevated acetylated tubulin correlate with persistence of primary cilia.
This led to a working hypothesis that AKAP220-targeted HDAC6 modulates primary cilia . Co-staining with both cilia markers was necessary to delineate between cilia and acetyl-tubulin at the midbodies of dividing cells ( Fig 3Q). Importantly, no change in cilia number occurred when AKAP220-ΔPP1 cells were treated with tubacin ( Fig 3T-V). Similarly, AKAP220KO cells were unresponsive to the drug (Supp Fig 3H-K). This raises the intriguing possibility that an HDAC6-PP1-AKAP220 signaling axis is the intracellular target for the drug tubacin ( Fig 3Y).  Immunofluorescent staining with acetyl tubulin (green), Arl13b (red) and DAPI (blue). S) Higher magnification grey scale image of Arl13b staining. W) Quantification (% ciliated cells) in DMSO (white) and tubacin-treated (grey) wildtype cells. *p<0.05, ns=non-significant; N=3. T-V) Tubacin has no effect on AKAP220-ΔPP1 cells. T) DMSO or U) and V) tubacin treated AKAP220-ΔPP1 cells. X) Quantification (% ciliated cells) and analysis as described above in DMSO (white) and tubacin-treated (blue). Y) Schematic of proposed tubacin mechanism of action on AKAP220signaling complex. All error bars are s.e.m. P values were calculated by unpaired two-tailed Student's t-test. Scale bars (10µm). Number of cells analyzed indicated below each column.

AKAP220 signaling impacts cortical actin dynamics.
The cytoskeleton maintains cell shape and structure by synchronizing the assembly and disassembly of actin, intermediate and tubulin filaments (Janke and Magiera, 2020; Klymkowsky, 1999). Covalent modification of these elements is an important facet of cytoskeletal regulation (Portran et al., 2017). For example HDAC6 deacetylates the actin-binding protein cortactin to relocate it from the nucleus to its sites of action (Ran et al., 2015). Since HDAC6 activity is impaired in AKAP220KO and AKAP220-ΔPP1 cells, it was important to evaluate if nuclear accumulation of acetylated cortactin was enhanced. Immunofluorescent detection of acetylcortactin (green) was more prominent in the nuclei of AKAP220KO and AKAP220-ΔPP1 mIMCD3 cells, as compared to wildtype ( Next, we investigated if AKAP220 signaling influences actin filament morphology (Fig 4H-M). To further explore this concept, we performed Fluorescence Recovery after Photobleaching (FRAP) at cell-cell junctional actin fibers (Mov 1-3). Actin-GFP recovery upon photobleaching was monitored over a time course of 20 seconds in wildtype (grey; Fig 4N); AKAP220KO (green; Fig   4O) and AKAP220-ΔPP1(blue; Fig 4P) mIMCD3 cells. Recovery curves depict the half-life and mobile fraction (Fig 4Q). Expanded section accentuates the expedited recovery of actin in wildtype cells (Fig 4Q, inset). The t1/2 for photo-recovery is 1.6 sec; n=106 (Fig 4R, grey). In contrast, the t1/2 was 0.9 sec; n=144 for AKAP220KO cells (Fig 4R, green). This is a 0.72 sec +0.09 reduction in the rate of photo-recovery when compared to wildtype. Parallel FRAP experiments in AKAP220-ΔPP1 cells calculated a t1/2 of 1.75 sec; n=128, representing a 0.85 + 0.10 sec reduction in the rate of photo-recovery as compared to wildtype (Fig 4R, blue). The availability of actin-GFP was similar in mobile fractions from each cell type (wildtype 95% (grey); AKAP220KO 98% (green); AKAP220-ΔPP1 98% (blue); Figure 4S). Thus, disruption of AKAP220-signaling impacts the distribution and dynamics of actin filaments.

Actin polymerization dictates cilium biogenesis and length.
Actin is a key regulator of cilia formation and elongation (Kim et al., 2015). AKAP220 -/mice exhibit reduced accumulation of apical actin through the diminished GTP-loading of RhoA (Whiting et al., 2016). This key regulator of cytoskeletal reorganization contributes to the assembly of actin barriers in renal cells (Blattner et al., 2013). HDAC6-mediated stimulation of actin polymerization is a prerequisite for the disassembly of primary cilia (Ran et al., 2015). A convergence of these ideas hypothesizes that defective AKAP220-signaling alters the dynamics of the actin barrier assembly to enhance ciliation (Fig 5A).
To test this hypothesis, we treated wildtype cells with the actin-depolymerizing drug Cytochalasin D (Fig 5B-E). Drug application (200 nM) for 4 h favored disassembly of actin (white) as monitored by immunofluorescence (Fig 5B & D). Cytochalasin D promoted a 2.95-fold increase in the percentage of ciliated cells (Fig 5F, pink column). Similar effects were observed in AKAP220KO cells (Fig 5G-L). These pharmacological effects were scored by measuring the length of cilia ( Fig   5M, pink). Insets feature representative cilia at higher magnification (

AKAP220 signaling influences cilia morphology.
We reasoned that a consequence of altered actin dynamics could be changes in cilia morphology.
Super-resolution immunofluorescence imaging of fixed cells was performed using Arl13b as a ciliary marker. Cilia appeared retracted in AKAP220-null and ΔPP1-knock in cells in comparison to cylindrical and symmetrical architecture of cilia in wildtype cells (Fig 6A-D). A secondary feature was bulbous tips at the distal end of mutant cilia (Fig 6C & D). Analyses from three independent experiments show that the occurrence of bulbous tips was prevalent in cilia from AKAP220KO and AKAP220-ΔPP1 cells (Fig 6E).
To investigate this phenomenon further in living cells, we infected our mIMCD3 cell lines with a lentiviral vector encoding Arl13b-GFP (Mov 4-6; Fig 6F-H). Quantitative imaging by live-cell superresolution microscopy revealed that cilia in AKAP220KO (green) and AKAP220-ΔPP1 (blue) cells were 1.6 and 1.7-fold longer than wildtype counterparts (grey; Fig 6H). This led us to the conclusion that the bulbous tips presented in figures 6C & D were elongated, flexible cilia that partially retracted (coiled back) on themselves (Supp Fig 6).
Kidney-on-a-chip technology offers a pseudo-physiological environment that simulates kidney tubules (Weber et al., 2016). Microfluidic delivery of nutrients through the lumen of these tubules 24 recapitulates fluid-flow (Freedman et al., 2013). This sophisticated tissue-engineering approach was used to evaluate cilia development (Fig 6I). Culturing of wildtype mIMCD3 cells formed columnar organoids with few cilia protruding into the lumen (Fig 6J & K). In contrast, longer cilia were evident in AKAP220-ΔPP1 organoids (Fig 6L & M). Immunofluorescent staining of Arl13b (red) and acetyl tubulin (green) marked cilia and DAPI (blue) detected nuclei. The average cilia length was 3.95-fold greater in AKAP220-ΔPP1 pseudo-tubules (n=73 cilia) as compared to wildtype (n=53 cilia). These effects are more visible in the grey scale images of Arl13b alone ( Fig   6K&M). Amalgamated data from three independent experiments are presented in figure 6N.
Cilia assembly requires the passage of materials through an actin barrier formed across the basal body (Farina et al., 2016). We reasoned that the dynamics of this process may be altered upon manipulation of AKAP220 signaling. Therefore, we combined super-resolution microscopy with Fluorescence Recovery after Photobleaching (FRAP) to visualize GFP-Arl13b trafficking into individual cilia (Fig 6O). Arl13b recovery upon photobleaching was monitored over a time course of 2.5 seconds in wildtype (grey; Fig 6P); AKAP220KO (green; Fig 6Q) and AKAP220-ΔPP1(blue; Fig 6R) mIMCD3 cells. In wildtype cells, the rate of GFP-Arl13b recovery was steady over this time-course. FRAP was more rapid and robust in AKAP220KO and AKAP220-ΔPP1 cilia (Fig 6S).
Total Arl13b recovered in AKAP220KO is increased 22.35 +3% (n= 120 cilia; green; Fig 6T) over wildtype (n=60 cilia; grey; Fig 6T). Similarly, recovery of Arl13b in AKAP220-ΔPP1 cilia is 19.04 +3% increased (n=126 cilia; blue; Fig 6T) over wildtype. The t1/2 of Arl13b recovery was 178.9 milliseconds in wildtype cilia (grey), as compared to 344.7 and 361.6 milliseconds in AKAP220KO and AKAP220-ΔPP1 cilia respectively (Fig 6U; green and blue). These studies suggest that there is a larger mobile pool of Arl13b trafficking into the mutant cilia. One plausible mechanism is that cilia in AKAP220KO and AKAP220-ΔPP1 cells lack an architectural checkpoint that gates protein movement into this organelle. A likely candidate is an f-actin barrier at the base of the cilium ( Fig   6V).

Autosomal dominant polycystic kidney disease (ADPKD) is associated with mutations in PKD1
and PKD2, changes in apical actin and cilia dysfunction (Halvorson et al., 2010). This pathology is characterized by fluid-filled cysts that replace normal renal parenchyma (Fig 7A). Aberrant HDAC6 activity has been implicated in cyst growth (Sun et al., 2019). Therefore, we reasoned that pharmacologically targeting HDAC6 may have therapeutic benefit in the reduction of cyst formation in cellular models of ADPKD (Fig 7B). Human pluripotent stem cells (hPSCs) with a targeted disruption of PKD2 were differentiated into kidney organoids (Freedman et al., 2015).
Cysts were identified as large, translucent structures that swayed in response to agitation (Fig 7C   & D). PKD2 -/kidney organoids and matched isogenic controls were treated with tubacin (0.2-1µM) for 48 h and cyst size was evaluated. At low concentrations of tubacin (0.2 µM), renal cyst size was markedly reduced in PKD2 -/organoids as compared to the control (Fig 7E & F). Similar results were obtained at a higher dose of 1µM (Fig 7G & H). Amalgamated data from five experiments are presented (Fig 7I). Drug toxicity as assessed by luminescence assay was evident at higher doses of tubacin ( Fig 7J). Thus, pharmacologically targeting the AKAP220-binding partner HDAC6 reduces cystogenesis in a disease relevant model of ADPKD.

Discussion
The primary cilium is a highly organized mechanosensory transduction unit that responds to environmental cues (Goetz and Anderson, 2010;Wheway et al., 2018). We have discovered signaling events proceeding through AKAP220 that temper cilia development in kidney collecting ducts. While it may seem paradoxical that disruption of local signaling can positively impact organellar development, it is important to note that a few AKAPs mitigate signaling events in other cellular contexts (Bucko and Scott, 2021;Langeberg and Scott, 2015). At neuronal synapses,  Fig 1). In contrast, our in vivo and in vitro data strongly implicate AKAP220 signaling in the modulation of aquaporin-2 trafficking and ciliogenesis. This could occur through one of two mechanisms-either the cellular signals that attenuate aquaporin-2 trafficking and diminish cilia development are processed through the same macromolecular complex, or, distinct signaling islands of AKAP220-binding partners are assembled to control each process.
A unifying principle of our study is that AKAP220-binding partners affect the development of primary cilia by enacting cytoskeletal changes at the level of actin polymerization and tubulin acetylation. This implicates deacetylation as a fundamental signal termination process that checks the rate of cilia development. Histone deacetylase 6 (HDAC6) that targets the cytoskeletal elements alpha tubulin and cortactin has been identified as a driver of cilia disassembly (Ran et al., 2015). In concordance with this notion, data in figure 2Q show that rescue upon overexpression of HDAC6 in an AKAP220KO background restores control of ciliation. Several lines of evidence suggest a role for AKAP220 in this process. Data in figure 2 indicate that this deacetylase is more labile in the absence of the anchoring protein. While phosphorylation protects HDAC6 from ubiquitination, and dephosphorylation favors proteasomal degradation, it has been unclear how this enzyme is maintained in proximity of developing cilia (Ran et al., 2020). Our findings point towards a previously unrecognized adaptor function for protein phosphatase 1 (PP1). This reasoning is predicated on evidence that AKAP220 is a conventional PP1-targeting subunit that utilizes a KVxF motif to contact the phosphatase (Bollen et al., 2010).
While the mechanism of HDAC6 interaction with PP1 is less clear (Brush et al., 2004), it appears that the anchored phosphatase retains the capacity to function as an adaptor protein and recruit this additional binding partner. This is inferred by pulse-chase data showing that HDAC6 protein stability is compromised in AKAP220KO cells and supported by proximity-ligation data detecting HDAC6 in AKAP220-signaling islands (Fig 2I-O). Thus, AKAP220-PP1 subcomplexes create a platform for the targeting of HDAC6 to repress tubulin acetylation during ciliogenesis. This may represent a homeostatic mechanism that enables cells to enter mitosis. . This signals clearing of the local actin barrier to commence microtubule nucleation. In keeping with this notion, data in figure 5 show that the actin depolymerizing drug Cytochalasin D enhances cilia persistence and length, whereas the actin-stabilizing compound Jasplakinolide has the opposite effect (Casella et al., 1981;Holzinger, 2009). These pharmacological tools highlight the action of AKAP220associated enzymes at the level of the actin barrier. This substructure is part of a "ciliary necklace" that acts as a physical checkpoint for proteins moving into the cilium (Long and Huang, 2020). This is supported by photobleaching studies showing that Arl13b moves more readily in AKAP220-ΔPP1 mutant cilia and these organelles are approximately 2-fold longer than wildtype ( Fig 6H). Thus, it is plausible that AKAP220-associated PP1 participates in maintenance of an actin barrier close to the basal body of the cilium, the loss of which leads to unhindered protein movement into the developing organelle.
Understanding the signaling mechanisms that govern primary cilia biogenesis in collecting ducts has potential for the treatment of autosomal dominant polycystic kidney disease (ADPKD). Our studies provide the first demonstration of tubacin drug action in a human organoid model system for ADPKD. Our work points towards the AKAP220-PP1-HDAC6 signaling axis as a therapeutic target for this ciliopathy. Two lines of evidence converge on this notion. First, HDAC6 mediated deacetylation of tubulin controls cilia depolymerization. HDAC6 inhibitors including tubastatin A, ACY-1215 and tubacin are in early phase clinical trials to manage cancers and ciliopathies (Dong et al., 2018;Haggarty et al., 2003;Song et al., 2020). Our pharmacological studies now utilize tubacin to enhance ciliogenesis in wildtype mIMCD3 cells. Importantly, we show in figure 3Y that this selective-HDAC6 inhibitor has no effect in AKAP220-ΔPP1 cells. This argues that the AKAP220-PP1-HDAC6 axis is a molecular target for this drug. Consequently, tubacin could be exploited as a precision pharmaceutical to enhance ciliogenesis in disorders that arise from loss of primary cilia. Second, since HDAC6 activity is elevated in PKD1 mutant renal epithelial cells (Li et al., 2016), we examined tubacin action in human cellular models of ADPKD. In figure 7, we show that tubacin reduced cystogenesis in PKD2 -/precystic organoids. Hence, targeting the AKAP220-PP1-HDAC6 subcomplex may not only enhance cilia development, but also interfere with downstream signaling events that contribute to cystogenesis. The AKAP-targeting concept has recently been used to restrain kinase inhibitor drugs at defined subcellular locations (Bucko et al., 2019(Bucko et al., , 2020. We will expand this approach towards developing a precision pharmacology strategy to selectively deliver HDAC6 inhibitors to primary cilia for the treatment of chronic disorders such as polycystic kidney disease.

Tissue section immunofluorescent staining
Kidneys were fixed in 10% (vol/vol) buffered formalin (4 °C), embedded in paraffin and 4-μm-thick sections collected. Sections were deparaffinized using Citrasolv (Fisher) and antigen retrieved in buffer A using a Retriever 2100 pressure cooker (Electron Microscopy Sciences). Tissue sections were blocked in 10% (vol/vol) donkey serum in PBS solution before overnight incubation with the respective primary antibodies.

Virus generation
Constitutively active lentiviral plasmids expressing Arl13b-GFP were transfected into HEK cells along with viral packaging and envelope plasmids. The viral particles generated are filtered and added to mIMCD3 cells. After 24 hours of incubation, the cells were selected in zeocin (400ug/ml) for the next 2 weeks. After selection, the cells were trypsinized and plated at a low density of about 1 cell/well in 96 well plate, expanded and tested for expression of Arl13b-GFP by western blotting and immunofluorescence.

Immunoblotting and blot analysis
Cells were grown to the desired confluence and washed once with PBS at room temperature.
Cold lysis buffer (20 mM HEPES, (pH 7.4),150 mM NaCl,1 mM EDTA,1% triton X-100 in water) was added along with protease and phosphatase inhibitors and the plate was rocked gently at 4 degrees for 10 mins. The cell lysate was then scarped into a pre-chilled tube and cleared at 12,000*g for 10 min at 4 degrees. A BCA assay (Pierce) was used to determine protein concentrations, and 30µg of protein was loaded onto a Bolt 4-12% bis-Tris gel (Life Technologies). The cleared lysate was boiled in 2X SDS loading buffer for 10 min before loading.
Proteins were transferred to nitrocellulose membrane and blocked in either 5% milk. The blot was incubated in primary antibody at 1:1000 or as specified by the manufacturer overnight at 4 °C.
Immunoblots were washed (3 times, 10 mins each) in TBST before incubation in a 1:10,000 secondary antibody for 1 hr at room temperature. Immunoblots were washed again in TBST (3 times, 10 mins each) before imaging on an iBright FL1000 (Thermo Fisher Scientific) with SuperSignal Dura ECL reagent (Thermo Fisher Scientific). Densitometry for blot quantification was done using thermo fisher's software.

Immunofluorescence and microscopy
Sample preparation: Cells were seeded on acid-washed coverslips in 12-well plates. After they achieve the desired confluence, the wells were rinsed thrice with PBS and fixed with 4% paraformaldehyde in PBS or 10% ice cold methanol (based on the antibody specification) for 12 mins. After fixation, the cells were permeabilized in PBS+0.1%Triton x-100+1%BSA for 12 mins, blocked with 2%BSA for 30 mins and treated with the respective primary antibodies overnight.
After thorough washing in PBS the next day (3 times, 5mins each), secondary antibodies conjugated to Alexa fluor dyes were added for 2 hours. After 2 washes in PBS, DAPI along with or without the actin probe (based on the experiment) was added for 10 mins. The coverslips were washed 2 more times in PBS and mounted on slides using ProLong Diamond Antifade Mountant (Life Technologies).
Imaging and analysis: Cells were imaged on a Keyence BZ-X710 microscope (Keyence, Itasca, IL) using the relevant filter cubes for DAPI (blue filter), Actin (red filter), Cortactin (green filter). All images were acquired with the same magnification (100X, oil immersion), exposure time, and illumination intensity. Images were quantified and processed using ImageJ software.

FRAP (Fluorescence recovery after photobleaching)
Sample preparation: Cells were reverse transfected with lifeact-GFP (Addgene plasmid no.58470) and seeded into 60mm glass-bottom plates for 24 hours. The transfected plates are rinsed in PBS and incubated in Fluorobrite medium containing NucBlue Hoescht 33342 stain (R37605, Invitrogen, 1 drop/ml). For Arl13b FRAP, mIMCD3 lines were transduced with L13-Arl13bGFP lentiviral construct and selected as outlined above. These cells were seeded in 60mm glass-bottom plates. They were treated with 0.1%FBS for 24 hours and then the experiment was performed.
Imaging technique and analysis: Cells were imaged on a GE Deltavision OMX SR microscope (GE Life Healthcare Sciences). After loading (as above), cells were placed in a humidified chamber with 5% CO2 at 37 °C and imaged using a 60X oil immersion objective (Olympus, Shinjuku, Tokyo, Japan). The actin signal was photobleached in the green channel (488nm) using a 15% laser power for 0.05 seconds. A total of 5 events were captured before photobleaching.
Images after photobleaching were captured every 100 milliseconds for 10 seconds. For Arl13b FRAP, Lentiviral Arl13b GFP mIMCD3 cells. The Laser pulse at 488 was at 30% T, in a spot at the tip of the cilium to bleach it for 0.1 seconds. 5 events before the bleach and a total of 350 time points were taken at 10msec increments. All images were acquired with the same settings.
Images were quantified and processed using ImageJ software.

Line-plot analysis in ImageJ
For the actin and cortactin distribution experiments, lines were drawn from the outer edge of the nucleus to beyond the lamellipodia of the cell and line plots were generated using ImageJ to give the pixel intensity of each of the proteins along the line.

HDAC6 activity assay
The experiment was performed using BioVision's HDAC6 activity assay kit. mIMCD3 cells were seeded in 10cm dishes and allowed to grow for the desired confluence. They were then lysed using the lysis buffer provided in the kit. The protein concentration was measured using BCA assay kit (Pierce) and 20µg of protein was loaded per well. 5µM tubacin was used in the experiment and the fluorescence intensity was measured at 380/490 nm in the plate reader.

Drug treatment experiments
mIMCD3 cells were seeded on cover slips in 12-well plates and allowed to grow to the desired confluence. They were treated with 2uM tubacin for 4 hours, 200nM cytochalasin D for 4 hours or 500nM Jasplakinolide for 1.5 hours and fixed immediately after treatment (as previously described). The cells were stained for acetyl tubulin and Arl13b to mark primary cilia and DAPI to stain DNA.

Organoid differentiation
Organoids were differentiated in 384 well plates from human pluripotent stem cells (WTC11 iPS cells, Conklin lab, Gladstone Institute) that had been modified to disrupt PKD2 (Cruz et al., 2017). critical advice on experimental design, data analyses, and data interpretations. J.G. and J.D.S. prepared figures and wrote the manuscript.