| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Hampl2,,a,b
í Pacherníkcáka,b
a Laboratory of Molecular Embryology, Mendel University Brno, 613 00 Brno, Czech Republic
b Developmental Biology Unit, Institute of Animal Physiology and Genetics, Academy of Sciences of the Czech Republic, 613 00 Brno, Czech Republic
c Institute of Biophysics, Academy of Sciences of the Czech Republic, 612 65 Brno, Czech Republic
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The proper coordination of growth- and differentiation-related intracellular processes is ensured by the concerted action of cyclin-dependent kinases (CDKs) in all eukaryotes. The mechanisms that evolved to regulate the activity of CDKs include heterodimerization of CDK with a cyclin subunit to form a holoenzyme, changes in phosphorylation, and interactions with members of a group of regulatory molecules, commonly called CDK inhibitors (CKIs) [4–7]. To achieve differentiation-associated G1/G0 arrest, cells can generally utilize any of several mechanism(s) including down-regulation of G1 cyclins, up-regulation of CKIs, and down-regulation of CDKs itself. However, current data indicate that the significance of the particular regulatory event for establishment and/or maintenance of the nonproliferative/differentiated state is dependent on the actual combination of molecules present in the particular cell type [8]. Furthermore, one cell type can achieve the same outcome by using different changes in cell cycle regulators according to the nature of the inducing signal [9].
In luteinizing cells, the four most likely interdependent molecular changes have now been assigned to regulation of the cell cycle exit. These include down-regulation of cyclin D2, down-regulation of cyclin E, and up-regulation of p21 and p27 CKIs [1, 2]. In mice, the majority of follicular granulosa cells stop synthesizing DNA within 4 h after their stimulation by LH. The down-regulation of cyclin D2 is the only change that is detectable at this time point, before significant changes in the amount of all three other regulators. The total amount of cyclin E does not decrease for at least another 20 h. Originally, p27 had been reported to be indispensable for luteinization, using p27-deficient mice [10–12]. However, Tong et al. [3] established that the differentiation process, as judged by the capability of luteal cells to support progesterone synthesis, resumes even in p27-null mutant mouse females. Nevertheless, p27-deficient luteal cell withdrawal from the cell cycle is temporally abnormal, being delayed for at least 1 day compared to controls. Moreover, the total amount of p27 protein is decreased at 4 h and does not start to increase significantly earlier than 12 h after the LH surge [1, 2]. As shown by in situ hybridization, the expression of p21 is up-regulated during the luteinization process with dynamics similar to that of p27 [1]. However, no alteration in ovarian function was found in p21-deficient mice [13].
Although current data point to cyclin D2 as the primary target for the signals that drive granulosa cells to stop dividing, several molecular mechanisms clearly work in concert in order to drive differentiation of luteal cells appropriately. In order to gain further insight into these mechanisms, in this study we systematically analyzed changes in the expression and activity of other potentially related cell cycle regulators throughout the life span of the mouse corpus luteum. Moreover, since p27 seems to operate relatively late in the differentiation process, we also addressed the question of how this cell cycle regulator relates to the establishment and/or maintenance of growth arrest, and other possible specific aspects of the luteinization process.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(C57BL/6 x BALB/c)F1 hybrid mice were used throughout this study. Parental inbred animals were originally purchased from Charles River Deutschland (Niederlassung Sultzfeld, Germany). They were maintained and crossed to obtain F1 hybrids in the animal facility of the Laboratory of Molecular Embryology, Mendel University Brno (Brno, Czech Republic). Animals were supplied with water and food ad libitum. Ovaries were removed from animals after they were killed by cervical dislocation. This method is consistent with the recommendations of the Panel of Euthanasia of the American Veterinary Medical Association. To obtain nonluteinized granulosa cells, prepubertal females (22–24 days of age) were stimulated for 44–46 h by 5 IU eCG applied i.p. To obtain cells at early stages of luteinization, females that were stimulated by eCG for 48 h (as described above) received i.p. injections of 5 IU hCG to synchronously induce ovulation/luteinization. The time of hCG application represents 0 h of luteinization. To obtain corpora lutea of pregnancy, females were naturally mated at the age of 4–6 mo. The day of vaginal plug represents Day 1 of pregnancy.
Antibodies and Reagents
Rabbit polyclonal antibody to human CDK2, which cross-reacts with the mouse homolog (sc-163); rabbit polyclonal antibody to mouse CDK4 (sc-260); rabbit polyclonal antibody to human CDK6, which cross-reacts with the mouse homolog (sc-177); mouse monoclonal antibody to mouse cyclin D1 (sc-450); and rabbit polyclonal antibody to human p27, which cross-reacts with the mouse homolog (sc-528), were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse monoclonal antibody to mouse p27 was purchased from Transduction Laboratories (Lexington, KY). Mouse monoclonal antibody to human retinoblastoma protein (pRb), which cross-reacts with the mouse homolog (14001A), was purchased from Pharmingen (San Diego, CA). Mouse monoclonal antibody against a C-terminal part of human cyclin D3, which cross-reacts with the mouse homolog, and a substrate for the kinase assay, glutathione-S-transferase (GST)-pRb(773–928), were generously provided by Dr. Jiri Lukas (Danish Cancer Society, Copenhagen, Denmark). All chemicals were purchased from Sigma (St. Louis, MO) or Fluka (Buchs, Switzerland). Anti-immunoglobulins and Protein G agarose beads were from Sigma; polyvinylidene difluoride (PVDF) membrane Hybond-P, [32P]ATP, and chemiluminescence detection reagents (ECL+Plus) were purchased from Amersham (Amersham, Aylesbury, UK). eCG was from Bioveta (Ivanovice, Czech Republic), and hCG was purchased from Organon (Oss, Holland).
Tissue Extracts and Immunochemical Analyses
For all analyses, each sample was made by pooling nonluteinized/luteinized tissue recovered from 2 to 4 females. To isolate ovarian cells, ovaries were extirpated and placed into minimal essential medium with Earle's salts supplemented with penicillin-G (75 mg/L), streptomycin sulphate (50 mg/L), and BSA (3 g/L). Granulosa cells were released by puncturing large follicles using 27-gauge needles; then the cells were aspirated using glass pipettes under a microscope (SZH10, Olympus, Prague, Czech Republic) and transferred into an Eppendorf tube. Undamaged, easily recognizable corpora lutea at different stages of development were microdissected according to their morphology using 27-gauge needles under the microscope and then transferred into an Eppendorf tube. Immediately after their isolation, cells/tissues were washed three times with PBS, and mechanically disintegrated and lysed in ice-cold lysis buffer containing 50 mM Tris/HCl (pH 7.4), 150 mM sodium chloride, 0.5% Nonidet P-40, 1 mM EDTA, 0.1 mM dithiothreitol, 50 mM sodium fluoride, and 8 mM ß-glycerophosphate. The following protease inhibitors were included: PMSF (100 µM), leupeptin (1 µg/ml), aprotinin (1 µg/ml), soybean trypsin inhibitor (10 µg/ml), and tosylphenylalanine chloromethane (10 µg/ml). After 30 min of extraction on ice, lysates were stored at -80°C until use. After thawing, lysates were cleared by centrifugation at 15 000 x g for 20 min at 4°C, and the concentrations of total protein in supernatants were determined using a DC Protein Assay Kit (Bio-Rad, Hercules, CA). Extracts were equalized for total protein and then used for Western blot analyses, immunoprecipitations, and kinase assays. For Western analysis, samples were mixed with double-strength Laemmli sample buffer and subjected to 10% SDS/PAGE (7% SDS/PAGE for the pRb detection). After being electrotransferred onto Hybond-P, proteins were immunodetected using appropriate primary and secondary antibodies, and visualized by ECL reagent according to the manufacturer's instructions. When required, membranes were stripped in 62.5 mM Tris/HCl pH 6.8, 2% SDS, and 100 mM mercaptoethanol, washed, and reblotted with another antibody from this selection. For immunoprecipitation, extracts were first subjected to an initial absorption with Protein G agarose and then immunoprecipitated with appropriate antibodies for 1 h in an ice bath. Immunoprecipitates were collected on Protein G agarose by overnight rocking, washed five times with lysis buffer, resuspended in double-strength Laemmli sample buffer, and subjected to 10% SDS/PAGE.
Kinase Assays
Histone H1 kinase and pRb kinase activities were used as measures of kinase activities of CDK2 and CDK4/6, respectively. For both types of kinase assay, CDKs were first immunopurified from cell extracts using appropriate antibodies as described for immunoprecipitation, except that the last two washes were done using kinase buffer containing 50 mM Hepes (pH 7.5), 10 mM magnesium chloride, 10 mM manganese chloride, 8 mM ß-glycerophosphate, and 1 mM dithiothreitol. Immune complexes were collected by centrifugation and used directly for the assay. For CDK2, kinase reactions were carried out for 30 min at 37°C in a total volume of 25 µl in kinase buffer supplemented with 100 µg/ml histone H1 (type III-S; Sigma) and 40 µCi/ml [32P]ATP. For CDK4/6, kinase reactions were carried out for 30 min at 30°C in a total volume of 25 µl in kinase buffer supplemented with 80 µg/ml GST-pRb and 40 µCi/ml [32P]ATP. Reactions were terminated by mixing with double-strength Laemmli sample buffer, and each total reaction mix was subjected to 10% SDS/PAGE followed by autoradiography.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The fates of p27, cyclin D2, and cyclin E proteins were followed in detail in luteinizing cells up to 48 h after LH peak by Robker and Richards [1, 2]. Since, at this time point, the differentiation processes might still not be completed, the total amounts of the selected cell cycle regulators were determined throughout the life span of the corpus luteum. Both hormonally stimulated prepubertal females and females at different stages of pregnancy were used. Specifically, nonluteinized granulosa cells were obtained from females at 22–24 days of age that were stimulated by 5 IU eCG for 44–46 h. Prepubertal noncycling mice of this age were used in order to ascertain the maximum numbers of fully developed follicles and to prevent contamination by luteinized tissue. In order to synchronously induce luteinization, eCG-stimulated females received injections of 5 IU hCG. First, the peak level of p27 reached during the initial phase of the luteinization process was maintained unchanged until about Day 18 of pregnancy (Fig. 1A). Then p27 became down-regulated to a level similar to that in nonluteinized granulosa cells (Fig. 1A). p27 is able to regulate activity of any CDKs that drive the progression through the G1 and S phases [5]. Therefore, the amounts of CDK2, CDK4, and CDK6 were determined in luteinized cells in order to resolve potential p27 partners. A major difference between the metabolism of CDK2 and that of CDK4/6 was revealed. During luteinization, CDK2 underwent dramatic down-regulation, leaving only a very small amount of CDK2 present in fully luteinized cells (Fig. 1B). On the other hand, although the amounts of both CDK4 and CDK6 also partially decreased relative to their levels in nonluteinized granulosa cells, they still remained at significant levels until about Day 18 of pregnancy (Fig. 1, C and D). The obvious similarity in timing of the maintenance of CDK4/6 and the up-regulation of p27 suggested that those CDKs might be targets for regulation by p27.
|
Further analyses addressed which D type cyclin(s) might serve as a partner(s) of CDK4/6 in luteinized cells. Cyclin D2 vanishes completely in the first hours after stimulation of granulosa cells by LH [1, 2]. Given this, the luteinizing/luteinized cells were subjected to Western analysis for only cyclins D1 and D3. Despite the observation, by in situ hybridization, that cyclin D1 mRNA predominates in theca and interstitial cells of the mouse ovary [1], a significant amount of cyclin D1 protein was detectable in nonluteinized granulosa cells (Fig. 1E). However, as shown in Figure 1E, there was no cyclin D1 detectable in luteal cells at 72 h after hCG injection, and a significant decrease was evident by 48 h (data not shown). Several studies have previously shown that cyclin D3 is abundantly expressed in many differentiating/differentiated cells and/or tissues, including corpus luteum [14–17]. Here, the finding of cyclin D3 in corpora lutea, made by Bartkova and others [14] using a immunohistochemical approach, was extended by describing the dynamics of cyclin D3 accumulation during the luteinization process (Fig. 1F). When compared to its level in nonluteinized granulosa cells, the amount of cyclin D3 was elevated only moderately at 72 h after hCG (Fig. 1F). In fully luteinized cells, however, cyclin D3 became dramatically up-regulated and remained in this state until about Day 18 of pregnancy (Fig. 1F).
Loss of p27 from CDK2 and the Formation of Complex p27/CDK4 Occurred Concomitantly with the Accumulation of p27
On the basis of Western blot data described above, further attention was focused on the cell cycle regulators p27, CDK4/6, and cyclin D3, which were coexpressed in luteinized cells. The total amount of CDK2 decreased simultaneously with the accumulation of p27 in luteinizing cells and was stabilized in this striking ratio of "low CDK2/high p27" throughout the life span of corpus luteum. Thus, only CDK4 and CDK6 seemed to remain as potential partners of p27. To investigate this possibility, CDK2, CDK4, and CDK6 were each immunoprecipitated, and the precipitates were probed for the presence of p27 protein. As expected, very little p27 was coprecipitable by anti-CDK2 antibodies in fully luteinized cells (Fig. 2A). On the other hand, a large amount of p27 was complexed with CDK2 in nonluteinized granulosa cells, and some p27 still remained in this complex even at 72 h after hCG injection. Altogether, p27 was progressively lost from CDK2 during the luteinization, most likely because of the concomitant loss of CDK2 itself. In contrast, while only a very small amount of p27 was associated with CDK4 in nonluteinized cells, the luteinization process itself was accompanied by a dramatic increase of p27 protein complexed with CDK4 (Fig. 2B). The least dramatic changes were found in the association of p27 with CDK6. As demonstrated in Figure 2C, although the amount of p27 coprecipitated with CDK6 decreased during luteinization, it never fell below its detectability before Day 18 of pregnancy. When compared with the relative changes in the total amount of CDK6 (Fig. 1D), the CDK6-associated p27 shown in Figure 2C followed the same pattern. The behaviors of CDK2- and CDK6-associated p27 were similar to each other with regard to the proportionality between the total amount of the particular CDK and the amount of p27 complexed with it. Finally, the amount of p27 that was collectively associated with all three CDKs was much less than the total p27 that was accumulated in luteinized cells. This became clear when the amounts of CDK-associated p27 were compared to the amount of p27 directly precipitated by anti-p27 antibody (Fig. 2D). These data, however, could not reveal to what extent the release of p27 from CDK2 and/or its new synthesis participated in the formation of complexes between CDK4 and p27.
|
In Luteinized Cells, Cyclin D3 Was Complexed with Both CDK4 and p27, Creating a Trimer, but Was Not Associated with CDK2 and CDK6
As documented here by Western analysis, the amount of cyclin D3 in nonproliferating luteal cells was increased by several-fold compared to that in nonluteinized granulosa cells. Recently, the existence of ternary p27/cyclin D3/CDK4 complexes with kinase activity independent of the progression of the cell cycle was shown in BALB/c 3T3 fibroblasts [18]. These two facts led us to hypothesize that newly synthesized cyclin D3 is also attracted primarily into complexes with CDK4 in noncycling luteinized cells. To investigate this possibility, CDK2, CDK4, and CDK6 were each precipitated, and coprecipitated cyclin D3 was visualized by Western analysis. Generally, of all three kinases in this assay, only CDK4 coprecipitated cyclin D3 from cell lysates of luteinizing cells (Fig. 3, A–C). In good agreement with its total amount, shown in Figure 1F, cyclin D3 physically associated with CDK4 was detected in luteinized cells but not in granulosa cells (Fig. 3B). Taken together, coprecipitation experiments shown in Figures 2B and 3B showed that CDK4 was present in luteinized cells simultaneously in complexes with p27 and cyclin D3.
|
To determine whether these complexes might be trimeric p27/cyclin D3/CDK4, we first asked whether or not there was also a physical association between p27 and cyclin D3 in luteinized cells. As shown in Figure 3D, cyclin D3 coprecipitated with p27. Notably, the amounts of cyclin D3 coprecipitating with CDK4 (Fig. 3B) were somewhat less than those coprecipitating with p27 (Fig. 3D). To determine the proportions of p27, cyclin D3, and CDK4 that were assembled together, and/or possibly document the existence of a trimeric p27/cyclin D3/CDK4 complex, sets of immunodepletion experiments were carried out. Generally, cell extracts of corpora lutea isolated at Day 16 of pregnancy were depleted of either CDK4 or p27 by immunoprecipitation using anti-CDK4 and anti-p27 antibodies. Aliquots of extracts were taken before and after the immunodepletion procedure and subjected to Western analysis for p27, cyclin D3, and CDK4. To validate this approach, the proportions of p27 and CDK4 proteins recoverable from the lysates using anti-p27 and anti-CDK4 antibodies were determined. As demonstrated in Figure 4, A and C, both p27 and CDK4 were almost completely removed using the corresponding antibodies. Despite this fact, the depletion of the assumed p27- and CDK4-partners varied significantly in its effectiveness. While p27 depletion led to removal of essentially all cyclin D3 (Fig. 4E), the same procedure removed only part of the total CDK4 (Fig. 4D). Removal of CDK4 resulted in reduction of the cyclin D3 level (Fig. 4E); however, it did not lead to a detectable change in the amount of p27 (Fig. 4B). Together, these findings unambiguously demonstrate that ternary p27/cyclin D3/CDK4 complexes exist in luteinized cells. We give the following rationale to support this conclusion: Since all the cyclin D3 is in physical contact with p27, then CDK4-associated cyclin D3 must be necessarily complexed with p27 (Fig. 4E).
|
Luteinization Was Associated with Dephosphorylation of pRb
Since the primary function of cyclins and CKIs is to regulate the kinase activities of CDKs, it was important to elucidate how the accumulation of cyclin D3 and p27, along with the prominent switch in their association with CDK2 and CDK4, translated into the kinase activities of those CDKs. First, the in vivo phosphorylation status of pRb, the major substrate of G1/S CDKs, was analyzed by separating the standard panel of the cellular extracts on 7% SDS PAGE followed by Western blotting using anti-pRb antibody. To demonstrate the spectrum of all possible phosphorylation forms of pRb, the extract made of the whole ovary was included in this analysis. As shown in Figure 5, entrance into the luteinization process was accompanied by the disappearance of phosphorylated forms of pRb. This corresponds with accumulation of noncycling luteinized cells in G0. Specifically, whereas nonluteinized granulosa cells still contained significant portions of phosphorylated forms of pRb, only nonphosphorylated forms were detectable by 72 h after hCG injection. No signs of phosphorylation occurred in fully luteinized cells. Notably, the total amount of pRb was reduced along with the loss of its phosphorylated forms. Therefore, a separate experiment was carried out using more total protein to exclude the possibility that the absence of phosphorylated pRb in luteinized cells was due only to our inability to visualize it in proportionally smaller amounts (data not shown). Together, those analyses suggested that the activity of all pRb kinases might be thoroughly extinct in luteinized cells. To test this hypothesis, the kinase activities of CDKs in both nonluteinized granulosa cells and fully luteinized cells were analyzed by in vitro kinase assays. Surprisingly, the luteinized cells were not completely devoid of in vitro measurable kinase activity associated with immunoprecipitates made using anti-CDK2, -CDK4, -CDK6, and -p27 antibodies, respectively (Fig. 6). Specifically, although the kinase activity of CDK2 toward histone H1 underwent a prominent decrease during the luteinization, significant kinase activity was still retained in fully luteinized cells (Fig. 6A). On the other hand, CDK4, CDK6, and p27 showed moderate kinase activities against GST-pRb when compared to the blanks, with only small differences between nonluteinized and luteinized cells (Fig. 6B). Unfortunately, we were not able to directly determine the kinase activity associated with cyclin D3, probably because of steric constraints imposed by binding of antibody to the epitope that is close to cyclin box [14].
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Dramatic changes in expression of four cell cycle regulators, p21, p27, cyclin D2, and cyclin E occur during the first 48 h of luteinization in vivo [1, 2]. The data of Robker and Richards [1, 2] and our results show that p27 accumulates early in the development of luteinized tissue. It is shown here that p27 is maintained at this high level until down-regulation that occurs after about Day 18 of pregnancy. Moreover, changes in cyclin D3, CDK4, and CDK6, and in their associations with p27, occur with dynamics similar to that of p27. Taken together, fully luteinized cells are devoid of any major fluctuations in cell cycle regulatory machinery until all the molecules studied here change together late in luteal life. In mice, plasma concentration of progesterone, the major product of luteal cells, declines between Days 17 and 19 of pregnancy ([19, 20] and citations therein) because of functional and structural luteolysis. Taken together, the temporal overlap suggests the existence of a functional link between the maintenance and the final switch in the composition and function of cell cycle regulators, shown here, and the persistence and termination of the progesterone production.
Proliferation ceases early in the luteinization process, and luteal cells were not shown to ever reenter the cell cycle [1–3]. In such cells, which are naturally kept in noncycling status for a long period of time and are not destined to ever resume the cell cycle, the proliferation-stimulating machinery might be completely shut down rather than be subject to transient inhibition by CKIs. The absence of the majority of positive regulators of the cell cycle, cyclins D1, D2, and E, and CDK2 supports this idea concerning luteal cells. Importantly, these findings also suggest that p27 accumulation does not occur for the purpose of inhibiting G1 kinases, thus preventing G1/S progression in luteal cells. Recently, Tong et al. [3] reported that, although the absence of p27 only slightly delays the exit from the cell cycle in luteinizing cells and is not thoroughly detrimental to the establishment of the differentiated phenotype, it still causes irregularities in the production of progesterone and nidatory estrogen. They also concluded that p27 is required for the proper functioning of differentiated luteal cells rather than for differentiation itself. Together with the increase in its total amount, p27 also exerts a cell cycle-inhibitory function through its redistribution from CDK4 to CDK2 [21–25]. Although there is great variability in the employment of these two modes of action of p27 among the different cell types, as indicated by the data of Hengst and Reed [26], Ladha et al. [27], Rogatsky et al. [8], and others, the reallocation of p27 from CDK2 to CDK4 during luteinization represents a novel pattern. Certainly, it contradicts the currently well-accepted scenario in which cyclin D/CDK4 acts as a reservoir of p27 that is released when CDK2 is to be inhibited in order to block cell proliferation. According to current understanding, the amount of p27 decreases when cells reenter the cell cycle [28, 29]. However, luteinized cells complete their life span by a process of structural luteolysis that includes cell fusion, apoptosis, and phagocytosis [19, 30, 31]. Thus, the decrease of p27 protein appears not to reactivate CDKs that would result in driving the cells into proliferation, but rather it is related to the thorough loss of cellular architecture. In this context, there is increasing evidence that phosphorylation of p27 by CDK2 is critical for lowering the amount of p27 in S phase [32–34]. Thus, our data identify at least one factor (CDK2), whose activity is probably not accessible for down-regulation of p27 in fully luteinized cells. Taken together, these findings amplify the idea that the expression of p27 in luteal cells is uncoupled from the regulation of the cell cycle itself and instead serves some other specific function(s). In this regard, associations of p27 were analyzed to gain insight into those potential functions.
At least three separate pools of p27 most likely coexist in luteinized cells. Only one of them represents p27 that is complexed with CDK, specifically with cyclin D3/CDK4. Since p27 depletion removes all the cyclin D3 from the extracts, but CDK4 depletion does not (Fig. 4E), certain p27/cyclin D3 complexes devoid of CDK4 must exist. Our inability to detect the loss of p27 upon the depletion of CDK4 (Fig. 4A) together with the large difference between the amounts of p27 that are recoverable by precipitation and coprecipitation using anti-p27 (Fig. 2D) and anti-CDK4 antibodies, respectively (Fig. 2B), documents that only a small part of the total p27 is recruited for complexing with CDK4. Thus, there are other pools of p27 in luteal cells that are clearly devoid of either CDK4 or both cyclin D3 and CDK4. The possibility that p27 can associate with proteins other than CDKs has already been addressed with respect to the potential mechanisms that make the proliferating cells refractory to the up-regulated level of this CKI [35–38]. Although the effect of complexing p27 with non-CDK molecules underlying the proliferative phenotypes in these studies is clearly different from that in luteinized cells, it still points to the existence of wide spectrum of the potential p27 partners. Correspondingly, Pedram et al. [39] found that only 10% of the p27 that accumulated in fetal rat hypothalamic astrocytes upon treatment with growth-inhibiting natriuretic peptide was associated with CDKs. On the basis of this finding, they speculated that CDK-free CKI may have additional roles in restraining growth. We hypothesize from our data that in luteal cells, p27 affects processes unrelated to growth through its association with currently unknown molecules.
As determined in this study, the transition from granulosa cells to the nonproliferative/luteinized state is correlated with the complete disappearance of phosphorylated forms of pRb. This behavior agrees with the well-accepted scenario in which inactivation of pRb via its phosphorylation by CDKs represents the critical regulatory step toward progression through the G1/S border [40, 41]. Strikingly, the absence of pRb phosphorylation in nongrowing luteinized cells coincides with the high levels of cyclin D3 that is in physical contact with its catalytic partner, CDK4. Moreover, when the activity of CDK4 was assayed in vitro using pRb as a substrate, there was only a moderate decrease in kinase activity in luteinized cells. Thus, in luteinized cells, the formation of large amounts of p27/cyclin D3/CDK4 complexes having in vitro measurable kinase activity does not result in phosphorylation of pRb in vivo. In this context, the down-regulation of pRb protein may also reflect its reduced role in luteinized cells, corresponding to the diminished significance of its kinase-mediated regulation. It is known that cells may also change the total pRb level, besides regulating the phosphorylation status of pRb [42–44]. Several other independent studies have recently pointed to the unexpected kinase activities associated with either p27, cyclin D3, or CDK4. For example, differentiation of mouse keratinocytes in vitro is accompanied by elevation of p27 that is not required for inhibition of proliferation but results in the formation of pRb kinase containing cyclin D3 [15]. Also, nonproliferating/differentiating keratinocytes contain CDK4 that is active as a kinase even when it is in complex with both cyclin D3 and p27 [43]. Dong and others [18] have identified the existence of ternary p27/cyclin D3/CDK4 complexes that are active as kinases in both growing and nongrowing BALB/c 3T3 fibroblasts. Importantly, p27/cyclin D3/CDK4 trimers show different substrate specificity when compared to the active binary cyclin D3/CDK4 complexes. During myogenic differentiation, cyclin D3 accumulates and forms complexes with CDK4 and some unknown proteins to inhibit activity of CDK4 as a pRb kinase [16]. Taken together, the results presented here suggest that, in the specific intracellular environment of luteinized cells, the kinase activity of CDK4 is directed primarily toward a yet unknown non-pRb substrate(s) by its association with cyclin D3, p27, and possibly also other modifying proteins. Importantly, the existence of p27-unbound CDK4 (Fig. 4D) indicates that CDK4 coexists simultaneously in at least two different states in luteal cells, only one of them being the participation in the p27/cyclin D3/CDK4 trimer. Correspondingly, in luteal cells, p27 may therefore serve unknown functions by complexing with the unknown partners, as suggested above, and also produce the characteristic properties of well-known cell cycle regulators, such as CDK4 and cyclin D3. Although cyclins generally contribute to progression of the cell cycle by regulating the activation of CDKs, in many cell types the induction of cyclin D3 to high levels is correlated with the process of differentiation [14–16, 45]. The large-scale immunohistochemical screen accomplished by Bartkova et al. [14] suggested that cyclin D3 expression may be associated with either the establishment or the maintenance of the mature phenotype, depending upon the particular cell type. The dynamics of cyclin D3 accumulation and maintenance shown here distinguish luteinized cells as an unambiguous example of this type of cyclin D3 function.
In conclusion, the maintenance of the fully differentiated phenotype of luteal cells in vivo is characterized mainly by up-regulated levels of p27 and cyclin D3 that function in ways other than inhibition of growth. The potential importance of p27 and cyclin D3 for the development of the differentiated phenotype, contrasted to cell cycle regulators involved rather in driving cell growth arrest, is depicted in Figure 7. Under the conditions specific to luteinized tissue, non-pRb substrate(s) seem to be targeted by CDK4, most likely because of its association with p27 and cyclin D3. Although our results do not provide information on whether such changes are specific to only the luteinization process and whether they are necessary for luteinization to occur, it will be important to identify and characterize these hypothetical substrates and to determine whether other protein-protein interactions participate in driving this function. From a general point of view, assuming that this phenomenon is likely to apply also to other terminally differentiated cell types, the luteal cell system described here seems well suited to these objectives.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 The project was supported by grants from the Ministry of Education, Youth and Sports of the Czech Republic (VS96115) and from the Grant Agency of the Czech Republic (524/96/K162 and 312/97/0393). 
2 Correspondence: Ale Hampl, Mendel University Brno, Zem
dlská 1, 613 00 Brno, Czech Republic. FAX: 420 5 45133298; hampl{at}mendelu.cz
Accepted: December 21, 1999.
Received: September 8, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
luteal transition. Cell Growth Differ 1998; 9:787–794.[Abstract]This article has been cited by other articles:
|
|
 |
|
  C. Stocco, C. Telleria, and G. Gibori The Molecular Control of Corpus Luteum Formation, Function, and Regression Endocr. Rev., February 1, 2007; 28(1): 117 - 149. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Paternot, J. E. Dumont, and P. P. Roger Differential Utilization of Cyclin D1 and Cyclin D3 in the Distinct Mitogenic Stimulations by Growth Factors and TSH of Human Thyrocytes in Primary Culture Mol. Endocrinol., December 1, 2006; 20(12): 3279 - 3292. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Bu, C. Cao, Y. Yang, C. Miao, Z. Hu, Y. Cao, Q. A. Sang, and E. Duan Localization and temporal regulation of tissue inhibitor of metalloproteinases-4 in mouse ovary. Reproduction, June 1, 2006; 131(6): 1099 - 1107. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Cannon, M. Cherian-Shaw, and C. L. Chaffin Proliferation of Rat Granulosa Cells during the Periovulatory Interval Endocrinology, January 1, 2005; 146(1): 414 - 422. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Tomic, K. P. Miller, H. A. Kenny, T. K. Woodruff, P. Hoyer, and J. A. Flaws Ovarian Follicle Development Requires Smad3 Mol. Endocrinol., September 1, 2004; 18(9): 2224 - 2240. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Kohoutek, P. Dvorak, and A. Hampl Temporal Distribution of CDK4, CDK6, D-Type Cyclins, and p27 in Developing Mouse Oocytes Biol Reprod, January 1, 2004; 70(1): 139 - 145. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Preclikova, V. Bryja, J. Pachernik, P. Krejci, P. Dvorak, and A. Hampl Early Cycling-independent Changes to p27, Cyclin D2, and Cyclin D3 in Differentiating Mouse Embryonal Carcinoma Cells Cell Growth Differ., September 1, 2002; 13(9): 421 - 430. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Moons, S. Jirawatnotai, T. Tsutsui, R. Franks, A. F. Parlow, D. B. Hales, G. Gibori, A. T. Fazleabas, and H. Kiyokawa Intact Follicular Maturation and Defective Luteal Function in Mice Deficient for Cyclin- Dependent Kinase-4 Endocrinology, February 1, 2002; 143(2): 647 - 654. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. L. Chaffin, K. M. Schwinof, and R. L. Stouffer Gonadotropin and Steroid Control of Granulosa Cell Proliferation During the Periovulatory Interval in Rhesus Monkeys Biol Reprod, September 1, 2001; 65(3): 755 - 762. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. H. Burns, C. Yan, T. R. Kumar, and M. M. Matzuk Analysis of Ovarian Gene Expression in Follicle-Stimulating Hormone {beta} Knockout Mice Endocrinology, July 1, 2001; 142(7): 2742 - 2751. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Tong and J. W. Pollard Genetic Evidence for the Interactions of Cyclin D1 and p27Kip1 in Mice Mol. Cell. Biol., February 15, 2001; 21(4): 1319 - 1328. [Abstract] [Full Text]
|
|
|
|
|
|
J. S. Richards New Signaling Pathways for Hormones and Cyclic Adenosine 3',5'-Monophosphate Action in Endocrine Cells Mol. Endocrinol., February 1, 2001; 15(2): 209 - 218. [Abstract] [Full Text]
|
|
| ||||||||
