Protein Kinase A-Mediated Septin7 Phosphorylation Disrupts Septin Filaments and Ciliogenesis

Septins are GTP-binding proteins that form heteromeric filaments for proper cell growth and migration. Among the septins, septin7 (SEPT7) is an important component of all septin filaments. Here we show that protein kinase A (PKA) phosphorylates SEPT7 at Thr197, thus disrupting septin filament dynamics and ciliogenesis. The Thr197 residue of SEPT7, a PKA phosphorylating site, was conserved among different species. Treatment with cAMP or overexpression of PKA catalytic subunit (PKACA2) induced SEPT7 phosphorylation, followed by disruption of septin filament formation. Constitutive phosphorylation of SEPT7 at Thr197 reduced SEPT7‒SEPT7 interaction, but did not affect SEPT7‒SEPT6‒SEPT2 or SEPT4 interaction. Moreover, we noted that SEPT7 interacted with PKACA2 via its GTP-binding domain. Furthermore, PKA-mediated SEPT7 phosphorylation disrupted primary cilia formation. Thus, our data uncover the novel biological function of SEPT7 phosphorylation in septin filament polymerization and primary cilia formation.

SEPT7 plays important roles during cell division, cytokinesis, and neuronal differentiation. For example, depletion of SEPT7 leads to microtubule destabilization as well as cytokinesis defects in fibroblasts [11]. SEPT7 is also required for chromosome alignment, affecting the extrusion of the second polar body during meiosis [12]. In mature neurons, depletion of SEPT7 results in decreasing dendritic branching and dendritic spine morphogenesis [13]. Thus, precise regulation of SEPT7 is crucial for development and differentiation.
SEPT7 is a key component of most types of septin filaments [8]. Septins can heterooligomerize into the hexamer core complex SEPT2-SEPT6-SEPT7-SEPT7-SEPT6-SEPT2. SEPT7 interacts with SEPT6 and SEPT7 via the N-C termini and GBD, respectively [14]. Therefore, SEPT7 assembles several septin members into a multimeric structure to carry out its biological functions.
Phosphorylation is an important mechanism in the regulation of septin function. During the epididymal transition of spermatozoa, defective SEPT4 phosphorylation results in a defective membrane diffusion barrier [15]. In the brain, SEPT4 phosphorylation is regulated by the dual-specificity tyrosine phosphorylation-regulated kinase 1A (DYRK1A) [16]. The phosphorylation of SEPT3 by PKG modulates neuronal function [17]. PLK1 interacts with and phosphorylates SEPT9, thus affecting cytokinesis [18]. Therefore, the phosphorylation of septins is important for various physiological functions.
The primary cilium is a microtubule-based organelle that protrudes from the surface of most mammalian cells. It plays important roles in development and differentiation. Disruption of cilium structure or function causes many diseases, including obesity, blindness, polycystic kidney disease, mental retardation, and polydactyly [19]. The primary cilium is composed of a central axoneme, which is the extended structure of the mother centriole, and an overlying ciliary membrane [20][21][22]. Septins have been shown to localize to the primary cilium and are involved in the formation of cilia, known as ciliogenesis [23]. For example, SEPT2 functions as diffusion barriers at the base of the ciliary membrane; the loss of ciliary membrane protein localization and inhibited ciliogenesis are observed in SEPT2-deficient cells [22]. In Xenopus embryos, knockdown of SEPT7 or SEPT2 causes defects in ciliogenesis [23]. In mammals, the SEPT2-SEPT7-SEPT9 complex localizes to the axoneme of cilia for ciliogenesis and the maintenance of ciliary length [21]. Thus, septin complexes are required for maintaining cilia growth and length [24][25][26].
In this study, we aimed to explore whether SEPT7 phosphorylation influences septin filament assembly and ciliogenesis. We found that PKA increased SEPT7 phosphorylation at the Thr. Overexpression of phospho-memetic SEPT7 reduced filament assembly, suggesting that SEPT7 phosphorylation at Thr197 is important for the assembly of SEPT7 filaments. We also found that PKA interacted with SEPT7 through the GTP-binding domain. Furthermore, we found that PKA-mediated phosphorylation of SEPT7 reduced the frequency of ciliated cells and reduced the ciliary length. These findings suggest that SEPT7 phosphorylation disrupts septin filament assembly and ciliogenesis.

DNA Constructs, Cell Culture, and Transfection
Human SEPT7 and PKA were amplified from a human RNA panel (Clontech, Mountain View, CA, USA) and cloned into pFLAG-CMV-2, pEGFP-C3, and pCDNA3-HA vectors, as described previously [27]. All constructs were verified by DNA sequencing. For transient transfection, malignant human testis pluripotent embryonic carcinoma NT2/D1 cells, REP cells, and human embryonic kidney 293 T cells were incubated in Dulbecco's minimal essential medium (DMEM), supplemented with 10% fetal bovine serum (FBS) and 1% antibiotics. Cells were transfected with plasmids using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. After 24 h, the cells were subjected to immunofluorescence staining or immunoblotting. Alternatively, cells were treated with 8-bromo-cAMP or Na 3 VO 3 (Sigma-Aldrich, Darmstadt, Germany) as indicated for analysis of protein expression.

Immunoprecipitation Assay and Western Blot Analysis
For immunoprecipitation analysis, 2 µg of indicated antibodies was incubated with Dynabeads Protein G (Thermo Fisher Scientific) at room temperature for 15 min on a rotator, and the cell lysates were immunoprecipitated with the bead-antibody complex at 4 • C overnight on a rotator. The samples were washed three times with a wash buffer (100 mM Tris, 150 mM NaCl, 2 mM EDTA, 0.5% Tween-20, and 0.01% NP-40). The precipitates were mixed with SDS sample buffer, then boiled for 10 min. For Western blotting, the proteins were adjusted to an equal amount of protein, electrophoresed through sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and blotted onto PVDF

Immunofluorescence
For the immunofluorescence analysis, the plasmids were transfected into NT2/D1 cells or 293T cells using Lipofectamine 2000 (Thermo Fisher Scientific). After 24 h, NT2/D1 cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) permeabilized with 0.1% triton X-100 in PBS, and blocked with the antibody diluent (Dako North America, Santa Clara, CA, USA) for 1 h. NT2/D1 cells or 293T cells were washed three times with PBS. The filament could then be visualized using fluorescence microscopy (Olympus, Tokyo, Japan, BX60).

ClustalW multiple Sequence Alignment
The human SEPT7 orthologous proteins were used for multiple sequence alignment by the ClustalW2 program (http://www.ebi.ac.uk/, accessed on 25 January 2020). The accession numbers for SEPT7 proteins of different species were as follows: Homo sapiens

Examining Primary Cilia Formation
Wild-type RPE1 cells were grown on glass cover slips at 37 • C overnight. For SEPT7deficient RPE1 cells, SEPT7 was depleted by transfecting RPE1 cells with siRNA against SEPT7 for 3 days. Then, cells were cultured in a serum-free medium for 24 h for inducing primary cilia formation. All cells were fixed with ice-cold methanol at −20 • C for 6 min. After blocking with 5% BSA for 1 h, the cells were incubated with primary antibodies for 24 h at 4 • C. After extensive washing with PBS, the cells were incubated with fluorescein isothiocyanate-and Cy3-conjugated secondary antibodies (Invitrogen, Carlsbad, CA, USA, A10521) for 1 h in the dark. The nuclei were stained simultaneously with 4 ,6-diamidino-2phenylindole (DAPI, 0.1 µg/mL). After extensive washing, the cover slips were mounted on glass slides in 50% glycerol. Primary cilia were imaged with an Axio Imager M2 fluorescence microscope (Zeiss, Oberkochen, Germany) and captured using ZEN Pro software (Zeiss). Primary cilia images were generated, and the length of cilia was measured from z-stacks using add-on features of the ZEN Pro software.

Statistical Analysis
Data were expressed as the mean ± SEM. Statistical significance was determined by one-way analysis of variance (ANOVA), combined with Tukey's multiple comparison test for posterior comparisons. p-values were considered significant at * p < 0.05; ** p < 0.01; and *** p < 0.001.

Conservation of the SEPT7 Phosphorylation Site among Species
Our previous study showed that PKA phosphorylates SEPT12 at Ser196, which is located within the GTP-binding domain (GBD), thus disrupting septin filaments in the sperm annulus [28]. GBD is conserved in the septin family, so thus screened the PKA recognition motif, [R/K]-X-X-[pS/T], using amino acid alignments of 28 septin sequences selected from all septin subgroups. This PKA recognition motif was identified in most septins, including SEPT2, SEPT3, SEPT4, SEPT7, SEPT9, SEPT10, SEPT11, and SEPT14 ( Figure 1A). Here, we were interested in SEPT7 as it was the irreplaceable component of all septin filaments [7]. The putative PKA phosphorylation site of human SEPT7 was located on the Thr197, so we checked whether this PKA phosphorylation site was conserved among different species. From zebrafish to Homo sapiens, the predicted phosphorylation site is highly conserved ( Figure 1B). Thus, Thr197 of SEPT7 was a putative phosphorylation site targeted by PKA. The amino acid sequences were analyzed using the ClustalW2 program at EMBL-EBI.

SEPT7 Is Phosphorylated by PKA
Next, we checked whether PKA was responsible for SEPT7 phosphorylation at Thr197. Due to lacking a specific antibody against phosphorylated SEPT7 at Thr197, FLAG-tagged SEPT7 was transfected into cells, and cell lysates were immunoprecipitated with a FLAG antibody. The immunoprecipitants were analyzed by an immunoblotting assay with an antibody against phosphothreonine. Treatment with 8-Br-cAMP, the activator of PKA, increased the phosphorylation on Thr sites of SEPT7 in 293T ( Figure 2A) and NT2/D1 ( Figure 2B) in a dose-dependent manner. Importantly, this phosphorylation signal was reduced in a phosphodeficient SEPT7 (T197A) mutant. These results suggest that the activation of PKA increased SEPT7 phosphorylation at the Thr197 site. To further confirm this finding, the catalytic subunit of PKA, PKACA2 (encoded by the gene PRKACA), was co-transfected with wild-type SEPT7 or T197A mutant followed by an examination of the phosphorylation status of SEPT7. Overexpression of PKACA2 increased the Thr phosphorylation of SEPT7, and this signal was reduced in the T197A mutant of the 293T and NT2/D1 cell lines ( Figure 2C,D). Thus, PKA phosphorylates SEPT7 on Thr197.

SEPT7 Interacts with PKA via the GTP-Binding Domain
We then checked whether SEPT7 interacted with PKACA2. FLAG-tagged SEPT7 and HA-tagged PKACA2 were co-transfected into NT2/D1 cell lines, followed by an immunoprecipitation assay. PKACA2 was detected in the precipitant of FLAG-SEPT7 and vice versa ( Figure 3A,B), suggesting that SEPT7 interacted with PKACA2. Next, the interacting domain of SEPT7 was examined. SEPT7 was dissected into the N-terminus (N'), GTP-binding domain (GBD), and C-terminus (C'). Unfortunately, the expression of N' of SEPT7 was unstable, so it could not be used for immunoprecipitation assay. We then examined whether the GBD or C' of SEPT7 interacted with PKACA2. GFP-tagged GBD or C' was co-transfected with HA-tagged PKACA2 into NT2/D1 cell lines, followed by an immunoprecipitation assay. PKACA2 was detected in the precipitant of GFP-tagged GBD ( Figure 3C), but not in the C' precipitant ( Figure 3D), and vice versa. Thus, SEPT7 interacts with PKACA2 via the GBD.

SEPT7 Phosphorylation Disrupts Septin Filament Formation
SEPT7 is crucial for septin filament oligomerization, so we checked whether SEPT7 phosphorylation affected septin filament formation. Two human SEPT7 mutant constructs, T197E and T197A, which mimic the constitutive phosphorylated and dephosphorylated status, respectively, were generated, and their effects on septin filament polymerization were examined. The SEPT7 filaments were observed throughout the cytoplasm when cells were transfected with wild-type SEPT7 (Figure 4, upper panel). This phenotype was also observed in cells transfected with the phosphodeficient construct T197A (Figure 4, middle panel). However, when cells were transfected with the phosphomimetic construct T197E, the septin filaments were hardly detected, and a smear signal was shown throughout the cytoplasm (Figure 4, lower panel), suggesting that phosphomimetic SEPT7 did not form filaments. Thus, phosphorylation of SEPT7 disrupts septin filament formation. Septin filament was orchestrated by the order of the septin hexamer core complex, arranged as SEPT2-6-7-7-6-2 or SEPT4-6-7-7-6-4 [14] ( Figure 5A). SEPT7 phosphorylation disrupted septin filaments; thus, the interactions of SEPT7 with different septins were examined.

SEPT7 Phosphorylation Disrupts SEPT7-SEPT7 Interaction
Different septin constructs were transfected into NT2/D1 cells, followed by an immunoprecipitation assay. The interactions of FLAG-tagged SEPT7, including wild-type, T197A, or T197E, with Myc-tagged SEPT6, HA-tagged SEPT4, or HA-tagged SEPT2 were then examined via an immunoblotting assay. First, we checked whether the interaction between SEPT7, which was important for the polymerization of different septin complexes, was affected by SEPT7 phosphorylation. FLAG-tagged SEPT7 and GFP-tagged SEPT7 were co-transfected into cells; FLAG-SEPT7 was then precipitated by an anti-FLAG antibody, followed by an examination of GFP-SEPT7 with a Western blotting assay. GFP-SEPT7 could be detected in either the wild-type or T197A precipitant ( Figure 5B). However, this interaction was detected to a lesser extent for T197E mutant ( Figure 5B), suggesting that SEPT7 phosphorylation impeded SEPT7-SEPT7 interaction. Next, the SEPT2-6-7 and SEPT4-6-7 complexes were checked. SEPT2 and SEPT6 were detected in wild-type SEPT7, the T197A precipitant, and T197E mutants. The data suggested that SEPT7 phosphorylation did not affect the orchestration of the SEPT2-6-7 heterocomplex ( Figure 5C). SEPT7 phosphorylation status also did not affect the SEPT4-6-7 heterocomplex (Figure 5D), suggesting that SEPT7 phosphorylation affected SEPT7-SEPT7 interaction. Our immunofluorescence data showed that overexpression of SEPT7 T197E disrupted septin filament polymerization. We also treated with phosphatase inhibitor sodium orthovanadate (Na 3 VO 4 ) to understand whether phosphatase regulated SEPT7-SEPT7 interaction. We found that Na 3 VO 4 did not interfere with the interaction between GFP-SEPT7 and FLAG-SEPT7 ( Figure S1). These data support the hypothesis that T197E disrupted the assembly of septin filament. Thus, SEPT7 phosphorylation blocks SEPT7-SEPT7 interaction, but does not affect SEPT2-6-7 or SEPT4-6-7 complex interaction.

Overexpression of PKA Disrupts SEPT7-SEPT7 Interaction
Our data showed that SEPT7 was phosphorylated by PKA, and SEPT7 phosphorylation disrupted septin filaments. Then, we checked whether overexpression of PKACA2 affected SEPT7-SEPT7 interaction. In the absence of PKACA2, SEPT7 interacted with SEPT7. However, this interaction was reduced when PKACA2 was overexpressed ( Figure 6A), supporting the hypothesis that PKACA2 blocked SEPT7-SEPT7 interaction. Then, the SEPT4-SEPT6-SEPT7 complexes were examined. SEPT7 interacted with SEPT6 and SEPT4, and these complexes were not affected when PKACA2 was overexpressed ( Figure 6B). Thus, PKACA2 reduces SEPT7-SEPT7 interaction, then disrupts septin filament formation. However, the expression of Myc-SEPT6 is different between the three different groups. We further determined whether PKA affected the septin expression. We found that SEPT7 expression increased after PKACA2 overexpression ( Figure 6C). We further decreased the amount of SEPT7 plasmid in the PKA overexpression group. We found that the expression of Myc-SEPT6 is equal between the three different groups ( Figure 6D). These data support the hypothesis that PKACA2 not only affects the SEPT7 expression but also blocks SEPT7-SEPT7 interaction.

PKA-Mediated SEPT7 Phosphorylation Affected Ciliogenesis
Septin-based ring filaments act as a diffusion barrier at the base of primary cilia [22]. We then investigated whether SEPT7 phosphorylation modulated primary cilia formation. Under serum starvation, cells start to grow primary cilia [29]; we thus checked whether serum starvation affected SEPT7 phosphorylation. Under serum starvation, the phosphorylation of SEPT7 was reduced ( Figure 7A,B), suggesting that SEPT7 phosphorylation might affect ciliogenesis. To further confirm our hypothesis, different SEPT7 mutants were transfected into immortalized human retina pigmented epithelial (RPE1) cells, a well-established model for studying ciliogenesis, and the population of ciliated cells was quantified. Transfection of T197A had no effect on ciliogenesis when compared with transfection of wild-type SEPT7. However, the frequency of ciliation was reduced when RPE1 cells were transfected with T197E ( Figure 7C). Thus, SEPT7 phosphorylation inhibits primary cilia formation. Then, the role of the cAMP-PKA cascade was examined. Treatment with cAMP reduced the frequency of ciliation and the length of cilia ( Figure 7D-F). In addition, the overexpression of PKA significantly inhibited ciliogenesis ( Figure 7G). Thus, SEPT7 phosphorylation inhibits ciliogenesis during serum starvation.

Discussion
In this study, we demonstrated that SEPT7 phosphorylation at Thr197 impeded SEPT7-SEPT7 interaction but did not affect SEPT2-6-7 or SEPT4-6-7 complexes, thus inhibiting septin filament polymerization. We also demonstrated that SEPT7 phosphorylation was mediated by the cAMP-PKA axis. Furthermore, we showed that SEPT7 phosphorylation inhibited primary cilia formation and reduced cilia length. Taken together, PKA-mediated SEPT7 phosphorylation inhibits septin filament formation, thus reducing ciliogenesis.
Post-translational modifications play an important role in the regulation of septinseptin interactions and control the formation of a high-order septin complex. These modifications include SUMOylation, acetylation, ubiquitination, and phosphorylation. Phosphorylation of the terminal subunit septin, Cdc11, disrupted neck filaments and influenced higher-order septin architecture [30]. Phosphorylation of Drosophila septin Pnut (homolog of human SEPT7) in the early stages of embryogenesis disrupts the assembly of septin filament formation, thus leading to the dissociation of the septin complex. Our previous work shows that phosphorylation of a germ-cell-specific SEPT12 leads to a complete loss of the septin ring at the sperm annulus, illustrating the important roles of septin phosphorylation in the regulation of septin assembly and the formation of higherorder structures. The phosphorylation site of SEPT12 is located in the GBD, and this domain is conserved among several septins, including SEPT7. We thus checked whether phosphorylation of SEPT7 affected the assembly of the septin filament. Indeed, SEPT7 phosphorylation in the GBD led to the defective polymerization of septin filaments. Thus, phosphorylation of septin in the GBD might affect septin filament formation. However, it is still unclear whether other septins, in addition to SEPT7 and SEPT12, also show a similar phenotype, and this hypothesis still needs to be tested in the future. It has been shown that the septin complexes regulate cilium length and ciliogenesis. Ablation of any of these septins' expression by RNA interference inhibits ciliogenesis.
Depletion of SEPT7 results in the concomitant loss of SEPT2 and SEPT9 expression of the complex [22,31]. In the Xenopus epidermis, SEPT7 forms a ring at the base of motile cilia. Knockdown of SEPT7 leads to fewer and shortened cilia [23]. Although SEPT7 is important for ciliogenesis, it is still unclear whether post-translational modifications of SEPT7 affect ciliogenesis. Here we showed that SEPT7 phosphorylation was reduced during serum deprivation, and T197E mutant inhibited ciliogenesis, suggesting that the phosphorylation status of SEPT7 was important for the growth of primary cilia. So far, it is still unclear how T197E affects ciliogenesis. Septin filaments orchestrated the transition of primary cilia, and T197E mutants disrupted septin complex formation; thus, we speculate that the T197E mutant disrupts septin filament formation and inhibits ciliogenesis. However, this hypothesis still needs to be confirmed. In primary kidney inner medullary collecting duct (IMCD3) cells, blocking the expression of SEPT2 inhibits ciliogenesis, thus reducing Sonic hedgehog (Hh) signaling [22]. Hedgehog signaling provides an essential role in cilia function. Knockdown of SEPT7 causes defective ciliogenesis and abrogated Hedgehog signaling in developing Xenopus embryos [23]. Interestingly, PKA is a conserved negative regulator of the Hh signaling transduction. Increasing the cAMP concentration leads to the activation of PKA, then induces Gli3 repressor (Gli3R). Gli3R translocates to the nucleus and represses gene expression. PKA phosphorylated SEPT7, thus inhibiting ciliogenesis; thus, PKA can inhibit Hh signaling through activation of Gli3R or the disruption of primary cilia. However, this hypothesis still needs to be confirmed. Several studies have demonstrated that septin filament might play an exclusive role in membrane diffusion barrier functions in diverse cell types, including sperm, neurons, and epithelia. In the sperm, a septin ring structure (annulus) connects the midpiece and the principal piece. The SEPT4 null mice show a disruption of the diffusion barrier phenotype at the sperm tail [32]. In the epithelia, the absence of septin filaments caused the misalignment of microtubules, loss of apicobasal polarity [33,34], and an alteration of the cell shape [2], thus leading to defective cytokinesis [35][36][37]. In the nervous system, septins form diffusion barriers at the dendritic spine's neck to restrict the diffusion of membrane proteins across the spine [38,39]. In this paper, we show that PKA-mediated SEPT7 phosphorylation is important for septin filament assembly and ciliogenesis. We speculate that the SEPT7 complex serves as a diffusion barrier at the ciliary base and participates in a variety of receptor signaling pathways. However, the downstream pathways of PKA and SEPT7 require further investigation. Our findings provide an important avenue to decipher the dynamics of SEPT complex assembly, as well as the physiological roles of the SEPT7 complex.

Conclusions
In summary, we found that SEPT7 was phosphorylated by PKA. PKA-mediated SEPT7 phosphorylation impeded SEPT7-SEPT7 interaction, but did not affect the SEPT2-6-7 or SEPT4-6-7 complexes. In addition, phosphorylation of SEPT7 was reduced upon serum starvation, during which primary cilium started to grow. The T197E mutant inhibited ciliogenesis upon serum starvation. Thus, the phosphorylation status of SEPT7 is essential for septin filament formation and ciliogenesis.