Microtubule-independent and protein kinase A-mediated function of kinesin KIF17b controls the intracellular transport of activator of CREM in testis (ACT).

Kinesins are motor proteins that transport their cargos along microtubules in an ATP-dependent manner. The testis-specific kinesin KIF17b was shown to directly regulate cAMP-response element modulator (CREM)-dependent transcription by determining the subcellular localization of the activator of CREM in testis (ACT), the testis-specific coactivator of CREM in postmeiotic male germ cells. CREM is a crucial transcriptional regulator of many important genes required for spermatid maturation, as demonstrated by the complete block of sperm development at the first steps of spermiogenesis in crem-null mice. To better understand the complex regulation of postmeiotic germ cell differentiation, we further characterized the ACT-KIF17b interaction, the function of KIF17b, and the signaling pathways governing its action. In this study, we demonstrated that the abilities of KIF17b to shuttle between the nuclear and the cytoplasmic compartments and to transport ACT are neither dependent on its motor domain nor on microtubules, thus revealing a novel microtubule-independent function for kinesins. We also showed that the cyclic AMP-dependent protein kinase A mediates the phosphorylation of KIF17b, and this modification is important for its subcellular localization. These results indicate that cyclic AMP signaling controls CREM-mediated transcription in male germ cells through modification of KIF17b function.

The complex differentiation of spermatogenic stem cells into mature sperm is governed by a highly specialized and strictly controlled program of gene expression. In spermatogonia and spermatocytes, the genes required for accurate mitotic and meiotic functions, respectively, are transcribed. Transcription ceases during the homologous chromosome pairing and meiotic recombination, but is again activated in postmeiotic germ cells. In the course of the differentiation of haploid spermatids into mature sperm, the genes required for development of acrosome, tail formation, removal of cytoplasm, and compaction of DNA are expressed (1)(2)(3). Upon compaction of chromatin by replacement of histones with transition proteins and protamines, the transcriptional activity ceases, and the mature sperm is transcriptionally inactive (4,5). The transcription factor cAMPresponse element modulator (CREM) 1 protein is highly expressed in male germ cells (6,7) and regulates the expression of many important postmeiotic genes, such as genes encoding protamines and transition proteins (8). Disruption of the crem gene in the mouse blocks the development of germ cells at the first step of spermiogenesis (9,10), indicating the crucial role of CREM in postmeiotic germ cell differentiation.
CREM-dependent transcription in testis is regulated by the testis-specific LIM-only protein activator of CREM in testis (ACT) that functions as a coactivator for CREM (11). The phenotype of act-null mice indicates that ACT is involved in the control of a specific subset of CREM target genes (12). The study of act-null mice demonstrated that ACT is required for the normal development of sperm head and tail. ACT is expressed at specific stages of germ cell development, the highest level of expression occurring in postmeiotic haploid round spermatids (11). The subcellular localization of ACT is regulated by a kinesin motor protein, KIF17b, which colocalizes with ACT in haploid spermatids and mediates the transport of ACT from the nucleus to the cytoplasm at specific stages of spermatid maturation (13). The expression of ACT in testis remarkably overlaps with CREM expression, and relocalization of ACT to the cytoplasm by the action of KIF17b correlates with the termination of transcription of the CREM-regulated genes (11,13).
Spermatogenesis is controlled by a complex network of endocrine, paracrine, and autocrine signals. In response to the hypothalamic gonadotropin releasing hormone, the pituitary gland secretes two hormones, luteinizing hormone, and folliclestimulating hormone (FSH), that are involved in the regulation of spermatogenesis. Luteinizing hormone regulates the testosterone secretion by somatic Leydig cells located in the interstitium between seminiferous tubules. FSH acts on Sertoli cells, the only somatic cells inside the seminiferous tubule, by stimulating signaling, gene expression, and secretion of peptides and other signaling molecules (14). Germ cells lack the receptors for pituitary hormones and are embedded in cytoplasmic pockets of Sertoli cells, which control spermatogenesis through strict paracrine regulation. To respond to external stimuli, germ cells utilize various transduction pathways that together form a complex signaling network governing germ cell development. The signaling mechanisms transferring information between Sertoli cells and germ cells remain so far largely unknown (15).
KIF17b is shuttling between the nuclear and cytoplasmic compartments and is able to transport ACT from the nucleus to the cytoplasm at a specific stage of spermatid differentiation (stage VIII/IX). The signals in haploid germ cells that induce KIF17b to relocate with ACT to the cytoplasm after the onset of spermatid elongation are yet unknown. In this study, we have clarified the function of KIF17b and the regulation of CREMdependent transcription by studying the mechanisms that modulate KIF17b localization. Strikingly, movement of KIF17b between nucleus and cytoplasm and its ability to transport ACT were shown to be independent of microtubules and the motor domain of the kinesin. These results indicate a novel, microtubule-independent mechanism for the function of kinesins. We also demonstrated that KIF17b movement is modulated by the cyclic AMP-dependent protein kinase A (PKA) dependent on phosphorylation, suggesting a role of cyclic AMP signaling in the regulation of CREM-dependent transcription in male germ cells.

EXPERIMENTAL PROCEDURES
Plasmid Construction-pGAD-ACT, pSG5-ACT, and pCS2-MTK-ACT, containing the N-terminal Myc tag, have been described previously (11). pSG5-KIF17b has been described previously (13). KIF17b fragments were generated by PCR using the primers containing specific restriction sites. They were cloned either in pGBT9 yeast expression vector (Clontech), containing the Gal4 DNA-binding domain (EcoRI and SalI), or in pTL1 vector, which is the pSG5 vector with a modified multiple cloning site (EcoRI and HindIII). KIF17b fragments were also transferred from pTL1 vector to pCS2-MTK vector (16), containing the N-terminal Myc tag, using EcoRI and XhoI restriction sites. The fulllength Myc-tagged KIF17b was generated by digesting the N terminus of KIF17b from the pSG5 vector by EcoRI and BstEII (there is an internal BstEII site in the KIF17b sequence) and ligating it to the pCS2-MTK-KIF17b-C terminus digested with the same enzyme. The serine residues within the putative PKA consensus sites were mutated to alanines using the site-directed mutagenesis kit (Stratagene) and pCS2-MTK-KIF17b as a template.
Yeast Interaction Assay-Yeast transformation and the ␤-galactosidase assays were performed in yeast strain Y190, as described in the Clontech Matchmaker two-hybrid system protocol. Yeast cells were cotransformed with the pGBT-KIF17b constructs pGAD-ACT and ␤-galactosidase reporter. ␤-Galactosidase activities were calculated in Miller units. The results reported are the means of three independent experiments.
In Vivo Phosphorylation-COS-1 cells grown on 6-cm culture dishes were transfected with 2 g of Myc-tagged KIF17b constructs. 24 h later, the medium was changed, and the cells were grown in the phosphatedeprived medium overnight. Next morning, the cells were incubated with 32 P-labeled orthophosphate (250 Ci/ml, MP Biomedicals, Irvine, CA) for 4 h in the absence or presence of the inhibitors. Cells were lysed in the buffer containing 20 mM Tris-HCl, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 1% deoxycholate, 5 mM EDTA, 20 mM NaF, 100 M NaVO 3 , 10 mM Na 2 MoO 4 , 0.5 mM dithiothreitol, and protease inhibitor mixture. After clearing of the lysate, immunoprecipitation was performed using anti-Myc 9E10 antibody and protein G-Sepharose (Amer-sham Biosciences), the beads were washed four times with the lysis buffer, and immunoprecipitated proteins were released in Laemmli sample buffer. Samples were run into a polyacrylamide gel, stained with Coomassie Blue stain to confirm equal amounts of protein in each lane, and phosphorylation of proteins was detected by autoradiography. The rest of the samples were run into a polyacrylamide gel and immunoblotted with anti-Myc 9E10 antibody to confirm equal loading of KIF17b in each lane. Horseradish peroxidase-conjugated goat antirabbit IgG (Jackson ImmunoResearch Laboratories) was used as a secondary antibody. Immunocomplexes were detected by enhanced chemiluminescence (Pierce).
In Vitro Phosphorylation-Overexpressed Myc-tagged KIF17b was immunoprecipitated from COS-1 cells using the anti-Myc 9E10 antibody and protein G-Sepharose (Amersham Biosciences). In vitro phosphorylation was performed in a buffer containing 50 mM HEPES, 100 mM KCl, 0.5% Nonidet P-40, 5 mM NaF, 1 mM dithiothreitol, 50 M cold ATP, 5 Ci of [␥-32 P]ATP (Amersham Biosciences), and 0.05 units/l catalytic subunit of PKA (Sigma) in a volume of 20 l at 37°C for 20 min. In some cases, mouse testis protein extract was used instead of PKA. The reaction was stopped by adding 20 l of Laemmli sample buffer. The samples were run into a SDS-polyacrylamide gel, and the gel was stained with Coomassie Blue dye to confirm equal loading of PKA and substrates. The gel was dried and autoradiography was performed to detect phosphorylation. A small part of the sample was run into an SDS-polyacrylamide gel and immunoblotted with anti-Myc 9E10 antibody to confirm equal amounts of KIF17b in each lane.
Preparation of Testis Extracts-To prepare testis extracts, the testes were decapsulated, and the seminiferous tubules were homogenized in a buffer containing 170 mM NaCl, 50 mM Tris-HCl, pH 8, 0.5% Nonidet P-40, 5 mM EDTA, 20 mM NaF, 1 mM dithiothreitol and 1:1000 protease inhibitor mixture. The lysate was cleared by centrifugation at 13,000 revolutions/min for 15 min. The protein concentration was measured with the Bio-Rad protein assay reagent and adjusted to 5 g/l.

The Integrity of LIM Domains of ACT Is Required for the
Interaction with KIF17b-KIF17b was originally found as an interaction partner of ACT in a yeast two-hybrid screening using full-length ACT as a bait (13). We have performed deletion analyses and yeast two-hybrid interaction assays to characterize the regions of ACT and KIF17b involved in the interaction. ACT is a LIM-only protein containing four and a half LIM domains, each consisting of the two sequential zinc fingers (11,17). Deletion analysis of ACT demonstrated that none of the LIM domains alone mediates the interaction with KIF17b, but it is rather the integrity in the organization of the domains that is required for the association (Fig. 1A). Deletion of the half-LIM domain at the N terminus of ACT or the first full LIM domain reduced the interaction to 40 or 70%, respectively. All other deletions disrupted the interaction almost completely. These results indicated that the general three-dimensional ACT structure, rather than a specific domain, was responsible for the interaction with KIF17b.
The Central Region of KIF17b Interacts with ACT-KIF17b has an N-terminal motor domain that is very well conserved between all kinesin family members (18). Located in the C terminus is a highly variable domain that confers a cargobinding function to many kinesins (19). In the central region of KIF17b, there is a long stalk region containing two coiled-coil domains. This region is known to be a regulatory region controlling, for example, kinesin dimerization (18). Interestingly, the original KIF17b clone found in the two-hybrid screen spanned this region, covering amino acids 320 -720 of the KIF17b sequence (13). The role of this region in the interaction with ACT was confirmed in the yeast interaction assays using full-length ACT and KIF17b deletion mutants. In this assay, the C-terminal region (residues 720 -1038 of KIF17b) that corresponds to the classical cargo binding region does not bind to ACT (Fig. 1B). In contrast, the interaction is mediated by the central stalk region of KIF17b (Fig. 1B). Detailed deletion analysis demonstrated that residues 520 -620 are essential for the binding of KIF17b to ACT (Fig. 1, B and C).
The Ability of KIF17b to Transport ACT Is Independent of the Motor Function-Because the traditional function of kinesins is to carry cargos along microtubules in an ATP-dependent manner, we wanted to investigate whether microtubules and the motor function of KIF17b are needed for its ability to shuttle between the nucleus and the cytoplasm and thereby to transport ACT across the nuclear membrane. First, we studied the subcellular localization of full-length KIF17b and of a deletion lacking the N-terminal motor domain (KIF17b-(320 -1038)). KIF17b constructs were cotransfected with ACT into COS-1 cells, and immunofluorescence was performed using specific antibodies. As reported previously, KIF17b is able to shuttle between the nuclear and cytoplasmic compartments and to transport ACT from the nucleus to the cytoplasm (13). Wild type KIF17b is mainly cytoplasmic, but a small percentage of the protein is either predominantly in the nucleus or equally distributed in the nucleus and cytoplasm ( Fig. 2A). When coexpressed in the same cell, ACT follows the localization pattern of KIF17b, as confirmed by labeling cells with the antibody raised against ACT (13) (data not shown). As KIF17b-(320 -1038) is found both in nuclear and cytoplasmic compartments, it is evident that deletion of the motor domain does not change the ability of KIF17b to shuttle (Fig. 2, B and C). Nuclei were stained by DAPI to highlight the different localization patterns of KIF17b. A mutant of KIF17b lacking the motor domain still colocalizes with ACT and transports ACT through the nuclear envelope (Fig. 2, B and C). These results indicate that the motor function is not required for KIF17b shuttling or for its ability to determine subcellular localization of ACT. The microtubule independency of KIF17b shuttling was further confirmed by treating cells with a microtubule-depolymerizing drug, colchicine, before immunofluorescence. After the treatment, all microtubules were disrupted, as shown by immunostaining with anti-tubulin antibody. However, KIF17b was still able to shuttle (Fig. 2, D and E) and colocalize with ACT in both nucleus and cytoplasm (Fig. 2F).
PKA Regulates the Subcellular Localization of KIF17b-To identify the signaling pathways involved in regulating KIF17b function, cultured cells overexpressing KIF17b and ACT were treated with various inhibitors to block specific signaling pathways. The percentages of cells expressing KIF17b in the cytoplasm or in the nucleus or diffusely in both compartments were calculated. Normally, 70% of cells express KIF17b predominantly in the cytoplasm. The rest of the cells express KIF17b either mainly in the nucleus or diffusely in both compartments (Fig. 3A, column 1). The protein kinase C inhibitor (Ro-31-8220), mitogen-activated protein kinase/extracellular signalregulated kinase kinase inhibitor (PD 98059), p38 kinase inhibitor (SB 203580), or the inhibitor of small Rho GTPases (Toxin B) had no effect on the localization of KIF17b (Fig. 3A). In contrast, the protein kinase A inhibitor H-89 drastically changed KIF17b localization, decreasing the cytoplasmic fraction of KIF17b and increasing the number of cells expressing KIF17b diffusely in both nuclear and cytoplasmic compartments (Fig. 3A). The ability of KIF17b to interact and colocalize with ACT was not affected by H-89, as shown by coimmunoprecipitation and immunofluorescence experiments (data not shown and Fig. 3B). These results strongly suggest that PKA is regulating either the cytoplasmic retention of KIF17b or the transport of KIF17b through the nuclear envelope. The adenylate cyclase inhibitor SQ22536 had a similar effect, further confirming the involvement of cyclic AMP and PKA in the regulation of KIF17b movements (Fig. 3C).
KIF17b Is Phosphorylated in Cells-Next we wanted to know whether the effect of H-89 on KIF17b localization is mediated by its phosphorylation. Many kinesin family members have been reported to be phosphorylated by various kinases, but whether KIF17b is a phosphoprotein remained unknown. In vivo phosphorylation experiments were performed to investigate whether KIF17b becomes phosphorylated in cultured cells. 36 h after transfection, cells were incubated in a phosphate-free medium with 32 P orthophosphate. After immunoprecipitation, phosphorylation of KIF17b was readily detected by autoradiography (Fig. 4A). KIF17b seems to be heavily phosphorylated; both the C terminus and the central region of KIF17b are independently phosphorylated in cells, suggesting the presence of several independent phosphorylation sites (Fig.  4, B and C). KIF17b lacking the N-terminal motor domain seems to be a better substrate for phosphorylation than the full-length protein, indicating a possible role for the motor domain in the control of kinesin phosphorylation.
PKA Is Required for KIF17b Phosphorylation in Cells-To study the involvement of various signaling pathways in the phosphorylation of KIF17b, cultured cells overexpressing KIF17b were treated with signaling inhibitors during the in-cubation with 32 P orthophosphate. Interestingly, treatment of the cells with H-89 (but not with the other inhibitors) blocked the phosphorylation of KIF17b (Fig. 5A). This correlates with the effect of H-89 on KIF17b localization and strongly suggests that PKA mediates the phosphorylation of KIF17b in cells. KIF17b is predominantly expressed in testis (13). To investigate whether KIF17b can also be phosphorylated in testis, we performed the kinase assay using KIF17b immunoprecipitated from cultured cells as a substrate, and testis extract as a source of kinases. KIF17b was shown to be phosphorylated after incubation with testis extract in the presence of [␥-32 P]ATP, thus demonstrating that there are active kinases capable of phosphorylating KIF17b in male germ cells (Fig. 5B). Importantly, the KIF17b phosphorylation activity present in testis extracts was also blocked when H-89 was added to the reaction, stressing the importance of the PKA pathway in the phosphorylation of KIF17b in vivo (Fig. 5B).
PKA Phosphorylates KIF17b in Vitro-The down-regulation of KIF17b phosphorylation by the PKA inhibitor H-89 suggests that PKA can directly phosphorylate KIF17b. Another possibility is that PKA is a part of a phosphorylation cascade finally leading to modification of KIF17b. PKA was shown to be able to phosphorylate KIF17b, as demonstrated by in vitro phosphorylation assays using the purified catalytic subunit of PKA (Fig. 6A). Software-supported analysis of the KIF17b sequence predicts three potential PKA phosphorylation sites in the Cterminal region (serines 729, 893, and 931), and phosphorylation of the region, including these residues (KIF17b-(720 -1038)) was also abolished by treatment with H-89 in cultured cells (Fig. 6B). However, a KIF17b mutant protein, with serines 729, 893, and 931 mutated to alanines, was still phosphorylatable in cultured cells to an extent comparable with wild type KIF17b, indicating that these residues are not the unique targets for modification of KIF17b (Fig. 6C). Although the KIF17b/S729A,S893A,S931A mutant is still phosphorylatable, these serine residues were shown to play a role in the localization of KIF17b in cultured cells. Indeed, the Ser3 Ala mutation of these sites had a similar effect on the localization of the protein as the treatment of cells expressing the wild type KIF17b with H-89 (Fig. 7). In other words, the solely cytoplasmic localization of the KIF17b/S729A,S893A,S931A mutant was decreased and the diffuse localization increased, as compared with the localization of wild type KIF17b (Fig. 7). Importantly, treatment with H-89 did not influence the localization pattern of the KIF17b/S729A,S893A,S931A mutant (Fig. 7). DISCUSSION Kinesins are classically known for their motor function and ability to transport cargos along the intracellular microtubule network in an ATP-dependent manner (18,20). Kinesins move toward the plus end of the microtubules in contrast to another group of motor proteins, the dyneins, which are minus-enddirected microtubule motors. The discovery of KIF17b as a novel interaction partner for a transcriptional coactivator (ACT) and the ability of this kinesin to transport ACT from the active site of transcription into the cytoplasm has expanded the array of possible functions of kinesins by demonstrating that KIF17b can be directly involved in the regulation of transcription (13). Here we show that this function is neither dependent on the binding of KIF17b to microtubules nor the ATP-dependent motor function, thus demonstrating a novel microtubuleindependent mechanism for KIF17b.
Even though the transport of ACT by KIF17b is not dependent on microtubules, KIF17b might have other functions in male germ cells, those mediated through microtubule binding. Indeed, KIF17b expression in male germ cells is not restricted to the stages involving ACT function, but it is also present later in elongating spermatids and mature sperm at a time when the expression of ACT has already disappeared. At these stages of sperm development, KIF17b is associated with the manchette, 2 a structure consisting of microtubules and thought to be involved in the shaping of sperm nucleus (21) and in the principal piece of the sperm tail (12). Thus, it appears that, after functioning in a microtubule-independent manner in transporting ACT from nucleus to cytoplasm, KIF17b might switch to the microtubule-dependent mode and continue by functioning as a more "traditional" kinesin. This is very likely, because both the manchette and the sperm tail consist of highly organized microtubular arrays. The cargos associated with KIF17b during these stages have not been identified yet. We further characterized the interaction between KIF17b and ACT and demonstrated that it is not a specific LIM domain in ACT but instead the integrity of the LIM domains that is needed for the interaction with KIF17b. This suggests that the general three-dimensional structure of ACT generates the interaction surface. The binding of KIF17b to ACT is not mediated by the C-terminal highly variable region of KIF17b that would correspond to the classical cargo-binding region in many different kinesins (19). On the contrary, ACT binds to the central stalk region of KIF17b, a domain generally thought to be a regulatory region, controlling processes such as kinesin oligomerization (18). The binding of ACT to the central stalk region, rather than the C-terminal region, suggests a regulatory role of ACT on KIF17b function. Because the structure of ACT consists only of the LIM domains, known to function as protein-protein interaction surfaces (17), it is tempting to speculate that ACT might function as a mediator of interactions between the kinesin and some other proteins. Thus, it is possible that ACT is not only transported by KIF17b to the cytoplasm but could also have a more active role in the KIF17b function. Further possible members of this KIF17b-ACT complex remain to be characterized in the future.
Both ACT and KIF17b are expressed in round spermatids, where ACT is always nuclear and KIF17b is shuttling between the nucleus and the cytoplasm. Only in stage IX spermatids, when CREM-dependent transcription ceases and the spermatid nucleus starts to elongate, does ACT relocate to the cytoplasm together with KIF17b (13). This movement has to be strictly regulated to avoid premature relocalization of ACT. In this study, we have clarified the signals regulating KIF17b localization by showing that cyclic AMP-dependent PKA controls the subcellular localization and phosphorylation of KIF17b. KIF17b is expressed also in other stages of germ cell development, but the only developmental phase where KIF17b is shuttling between the nucleus and cytoplasm is in round spermatids, when it transports ACT from the nucleus to the cytoplasm. Because the inhibition of PKA signaling disturbs the nucleocytoplasmic shuttling of KIF17b, it is likely that the regulation of KIF17 by PKA pathway occurs at this stage of germ cell development.
We have found that KIF17b is a phosphoprotein (Fig. 4). It is likely that various protein kinases can phosphorylate KIF17b at many separate sites, but among all the tested signaling pathway inhibitors, PKA inhibitor H-89 was the only drug able to block the phosphorylation in cultured cells. This suggests that PKA is an important regulator of KIF17b phosphorylation. As shown by in vitro phosphorylation assay using the PKA catalytic subunit, KIF17b is a substrate of PKA. However, it is unclear whether KIF17b is phosphorylated directly by PKA in vivo or whether the phosphorylation is mediated by other kinases as a result of signal transduction cascades. KIF17b is heavily phosphorylated, which makes the study of specific phosphorylation sites more difficult. KIF17b contains three consensus PKA sites in its C-terminal region, but mutation of these sites does not abolish its phosphorylation levels in cultured cells. This indicates that other yet unidentified sites can be targets for phosphorylation.
The role of PKA in KIF17b regulation may include nuclear import, export, and cytoplasmic retention. The main nuclear localization signals are located in the C terminus of KIF17b, because the fragment of KIF17b containing only the C terminus is localized predominantly in the nucleus, and the KIF17b deletion lacking the C terminus is exclusively cytoplasmic. 2 The inhibition of PKA activity changes the localization of KIF17b from only cytoplasmic to nucleocytoplasmic. Therefore, it is likely that the nuclear import is regulated by phosphorylation, and the lack of phosphorylation induces the transport of KIF17b into the nucleus. Another possibility is that the PKA pathway controls the cytoplasmic retention of KIF17b (for example, the binding of KIF17b to specific cytoplasmic structures). When the pathway is blocked, KIF17b starts leaking to the nucleus, thus causing the diffuse localization pattern. The export of KIF17b from the nucleus to the cytoplasm is known to be controlled by the chromosomal region maintenance 1 nuclear export pathway (13). Thus, a third possibility is that PKA-mediated phosphorylation stimulates the nuclear export, and in the absence of PKA signaling, the newly synthesized KIF17b is imported into the nucleus but cannot be efficiently exported anymore to the cytoplasm.
The PKA holoenzyme consists of two catalytic subunits and two regulatory subunits. The three different isoforms of somatic catalytic subunits are C␣, C␤, and C␥. C␣ and C␤ are ubiquitously expressed in most tissues, whereas C␥ is a transcribed transposon that is expressed only in primate testis (22). Interestingly, C␣ generates a sperm-specific isoform, Cs, a product of an alternative transcript of the C␣ gene (23,24,25). Both C␣ and Cs are expressed during spermatogenesis, but their expression is mutually exclusive (24). C␣ is expressed in somatic cells, spermatogonia, and preleptotene spermatocytes, whereas Cs is germ cell-specific; it first appears in midpachytene spermatocytes, and the expression continues in round and elongating spermatids and finally in the tail of mature sperm (24). Thus, the expression of KIF17b and the Cs catalytic subunit of PKA overlaps in round and elongating spermatids when KIF17b is active in shuttling between the nucleus and the cytoplasm, indicating that the Cs catalytic subunit is involved in the phosphorylation of KIF17b at this stage of spermatogenesis. PKA is activated by cyclic AMP produced by adenylate cyclase (AC). In testis, at least two categories of ACs have been described, the membrane-associated ACs that are regulated by G protein-associated receptors and the soluble ACs that are modulated by bicarbonate (26). Both the membrane-bound and -soluble AC isoforms are expressed in postmeiotic round spermatids (27), where KIF17b and ACT are active.
FSH is a pituitary hormone regulating sperm development through its adenylate cyclase-coupled receptors located at the cell surface of somatic Sertoli cells in the seminiferous epithelium. Sertoli cells are in close contact with germ cells and constantly communicate with these cells and regulate their functions through paracrine signaling and junctions between the Sertoli and germ cells (14). The communication between Sertoli cells and germ cells is not well characterized. Although no adenylate cyclase-coupled FSH receptors have been found on germ cells, PKA activity changes during germ cell differentiation, to reach a maximum level in spermatids (28). cAMP is thought to play an important role in governing the timing of postmeiotic gene activation in response to FSH signaling (29). FSH-stimulated cAMP production is stage-dependent and is particularly significant at stages I-V (30). During these same stages, KIF17b shuttles between the nucleus and the cytoplasm (13), suggesting that this increase in cAMP production could be involved in the regulation of KIF17b movements.
Together these results demonstrate a novel, microtubule-independent function for a kinesin motor protein that is regulated by PKA and cAMP, thus linking PKA signaling to the control of CREM-dependent transcription in postmeiotic germ cells. Because ACT is able to enhance CREM transcription in a phosphorylation-independent manner (11), it appears that KIF17b and its regulation by PKA offer a new model for cAMP responsiveness of germ cells without the direct phosphorylation of CREM.