The Selective RNA-binding Protein Quaking I (QKI) Is Necessary and Sufficient for Promoting Oligodendroglia Differentiation*

Quaking I (QKI) is a selective RNA-binding protein essential for myelination of the central nervous system. Three QKI isoforms with distinct C termini and subcellular localization, namely QKI-5, QKI-6, and QKI-7, are expressed in oligodendroglia progenitor cells (OPCs) prior to the initiation of myelin formation and implicated in promoting oligodendrocyte lineage development. However, the functional requirement for each QKI isoform and the mechanisms by which QKI isoforms govern OPC development still remain elusive. We report here that exogenous expression of each QKI isoform is sufficient to enhance differentiation of OPCs with different efficiency, which is abolished by a point mutation that abrogates the RNA binding activity of QKI. Reciprocally, small interfering RNA-mediated QKI knockdown blocks OPC differentiation, which can be partly rescued by QKI-5 and QKI-6 but not by QKI-7, indicating the differential requirement of QKI isoform function in advancing OPC differentiation. Furthermore, we found that abrogation of OPC differentiation, as a result of QKI deficiency, is not due to altered proliferation capacity or cell cycle progression. These results indicate that QKI isoforms are necessary and sufficient for promoting OPC development, which must involve direct influence of QKI on differentiation/maturation of OPCs independent of cell cycle exit, likely via regulating the expression of the target mRNAs of QKI that support OPC differentiation.

Oligodendrocytes (OLs) 2 are responsible for myelination of the central nervous system. The oligodendroglia progenitor cells (OPCs) are derived from pluripotent neural stem cells in the developing brain (1,2). After fate specification for the OL lineage, OPCs keep proliferating until they are committed for terminal differentiation (3). The maturation of OPC into myelinating OLs is characterized by the extension of numerous branched processes and membrane sheets, along with the expression of myelin-specific structural proteins (4). Adequate OPC proliferation ensures the generation of sufficient numbers of myelin-producing cells, whereas OPC differentiation is a critical prerequisite for myelin formation. Thus, the transition of OPCs from proliferation to differentiation is a key step that governs central nervous system myelin development (5,6). However, forced cell cycle exit is not sufficient for inducing OPC differentiation (7,8), suggesting the existence of cell cycle-independent yet undefined mechanisms that directly promote OPC differentiation/maturation.
QKI is a selective RNA-binding protein that controls the homeostasis and subcellular localization of target mRNAs during OL and myelin development (9 -13). Deficiency of QKI in OLs results in severe hypomyelination in the quakingviable (qk v ) mutant mice (10,14), which can be rescued by a QKI transgene specifically expressed in the OL lineage (15). Three major QKI protein isoforms with distinct C termini are derived from alternative splicing, named QKI-5, QKI-6, and QKI-7 (16,17). QKI-5 is predominantly localized in the nucleus, whereas QKI-6 and QKI-7 are mainly detected in the cytoplasm (14), suggesting distinct functional influence by QKI isoforms on their mRNA targets. The common N terminus in QKI isoforms harbors an extended RNA-binding domain homologous to that in the heterogeneous nuclear ribonucleoprotein K (KH) (16), which is essential for the interaction of QKI with its mRNA ligands and the dimerization between QKI isoforms (18). Thus, it is believed that the functional interplay of QKI isoforms governs the stability and subcellular localization of their mRNA targets during central nervous system myelin development (11,19).
In addition to its essential role in myelination, QKI is expressed in cells at the earliest stage of OL lineage development before they migrate away from the subventricular zone (14,16,20), suggesting that QKI may function to advance OL lineage development before actual myelination. Consistent with this idea, QKI is up-regulated in OLs when they extend processes to contact axons (14,21). In addition, ectopic expression of exogenous QKI-6 and QKI-7 can force cell cycle exit and differentiation of OPCs (19). However, whether endogenous QKI is essential for advancing OL development, and if so, which QKI isoform is functionally required, has not been determined. Whether QKI only accelerates OPC cell cycle exit or whether it also directly promotes OPC differentiation via cell cycle-independent mechanisms is still unknown.
In this study, we report that QKI is necessary and sufficient for OPC differentiation. In addition, QKI isoforms exhibit differential abilities in promoting OPC differentiation, which depends on the RNA binding activity of QKI. Furthermore, in the CG4 OPC cell line, RNA interference-mediated QKI knockdown abrogates differentiation without affecting the proliferation capacity or cell cycle progression. These results provide the first evidence suggesting that QKI can exert direct influence on OPC differentiation via cell cycle-independent mechanisms, most likely by regulating the expression of its mRNA targets that encode key proteins for OPC differentiation.

MATERIALS AND METHODS
Cell Culture-CG4 cells were maintained for proliferation in Dulbecco's modified Eagle's medium containing 1% heat-inactivated fetal bovine serum, insulin (5 g/ml), transferrin (50 g/ml), PBPS (putrescine (100 mM), biotin (10 ng/ml), progesterone (20 nM), and selenium (30 nM)). Platelet-derived growth factor AA (Sigma) and basic fibroblast growth factor (Promega) were added to the proliferation medium at a final concentration of 10 ng/ml each. Induction of differentiation was achieved by replacing proliferation medium with differentiation medium, which contains Dulbecco's modified Eagle's medium, insulin, PBPS, transferrin, tri-iodothyronine (50 nM), and 0.5% fetal bovine serum. Primary cultures of OPCs were isolated as described previously (22). Briefly, mixed glial cultures were prepared from the brains of neonatal rats (P2; Sprague-Dawley, Taconic Farms) and allowed to reach confluence. OPCs were removed by shaking and further purified by immunopanning with A2B5. They were plated onto poly-lysine-coated coverslips and maintained in a defined serum-free medium (22) with platelet-derived growth factor and fibroblast growth factor (10 and 20 ng/ml, respectively) to prevent differentiation. Mature OLs were obtained by culturing the OPCs in a defined medium lacking growth factors.
Immunocytochemistry and Antibodies-Cells were grown on coverslips for indirect immunofluorescent staining. After being rinsed with PBS, cells were fixed with 4% paraformaldehyde (Sigma) for 15 min at room temperature and blocked with 2% normal goat serum. The anti-O1 antibody (1:1000; Chemicon, Inc.) was incubated with cells overnight at 4°C. After washing with PBS, Texas Red-conjugated secondary antibody (1:1000, Jackson ImmunoResearch Laboratories) was incubated for 1 h at room temperature. Immunolabeled cells were mounted with Vectashield mounting medium (Vector Laboratories) containing DAPI for nuclei staining. Fluorescent signals were visualized using an Olympus IX-51 fluorescent microscope, and images were captured with a Retica digital camera.
Morphological Analysis-Light-field images and fluorescent images of transfected live CG4 cells (GFP-positive) were captured from randomly selected microscopic fields using the Olympus IX-51 inverted fluorescent microscope. Processes were categorized into three groups based on where they originate and counted: the primary processes are directly from the cell body, the secondary processes are from the primary processes, and the tertiary processes are from the secondary processes. More than 30 randomly selected cells were analyzed from each transfected culture.
BrdUrd Labeling-To examine the proliferation of CG4 cells, BrdUrd (Sigma) was added to the culture medium at a final concentration of 10 M. After incubation at 37°C for 1 h, cells were fixed with 70% ethanol at Ϫ20°C for 30 min. HCl (2 N) was used to denature DNA for 1 h at 37°C. Cells were permeabilized with 0.1% Triton X-100 and blocked with 2% normal goat serum, incubated with anti-BrdUrd (1:1000; Chemicon, Inc.) at room temperature for 1 h followed by secondary antibody incubation and quantitative analysis of cell numbers as described in the legend for Fig. 6.
Flow Cytometry Analysis of Cell Cycle-Cells were harvested, washed with ice-cold PBS, and incubated with prechilled 80% ethanol at 4°C overnight. After washing, cells were treated with 500 g/ml RNase A in PBS at 37°C for 30 min followed by incubation with a propidium iodide (Invitrogen) solution (50 g/ml propidium iodide, 0.1% sodium citrate, 0.1% Triton X-100 in PBS) at room temperature for 1 h in dark. DNA content of the cells was measured with a FACSCalibur flow cytometer (Beckman Instruments), and results were analyzed with Flowjo software (Tree Star, Inc.).

Forced Expression of Exogenous QKI Isoforms Enhances OPC Differentiation, Which Is Abolished by the E48G Mutation That
Affects the RNA Binding Activity of QKI-CG4 is an OPC cell line that can be induced for differentiation, recapitulating the morphological changes and maturation marker expression profiles observed in primary cultured OPCs as well as in the developing brain (23-25). One advantage of using CG4 cells over primary cultured OPCs is the consistent high efficiency of transfection, allowing efficient manipulation of QKI expression. We first examined the functional influence of QKI isoforms on the differentiation of CG4 cells by transfecting plasmids that encode FLAG-QKI-5, FLAG-QKI-6, and FLAG-QKI-7 individually into proliferating CG4 cells. Immunofluorescent staining indicated predominant nuclear localization of FLAG-QKI-5, whereas the majority of FLAG-QKI-6 and FLAG-QKI-7 were detected in the cytoplasm, reminiscent of the subcellular localization of QKI isoforms in the developing brain (14). Upon removal of mitogens, the bipolarshaped proliferating CG4 cells started to differentiate, characterized by the increased number of primary processes and the development of sophisticated secondary and tertiary branches (Fig. 1A). Enhanced morphological differentiation was observed in cells expressing exogenous QKI isoforms, which were marked by the expression of the GFP from the co-transfected plasmid, as compared with that in control cells that were transfected by the GFP plasmid alone (Fig. 1A). Quantitative analysis indicated that the average numbers of secondary and tertiary processes were markedly increased in cells expressing exogenous QKI, whereas the numbers of primary processes were not significantly altered (Fig. 1B). Although all three QKI isoforms were capable of enhancing process development, QKI-5 and QKI-6 apparently enhanced CG4 cell differentiation more efficiently than QKI-7 (Fig. 1B), despite the fact that all three FLAG-QKI isoforms were expressed at similar levels ( Fig. 1B,  inset). In addition, we did not detect increased apoptosis in cells expressing QKI-7 (data not shown), excluding the possibility that the weaker activity for QKI-7 to advance differentiation of CG4 cells is due to increased apoptosis. Consistent with a previous report identifying p27Kip1 mRNA as a QKI target (19), we observed association p27Kip1 mRNA with FLAG-QKI-6 in CG4 cells (Fig. 1C). In addition, exogenous QKI-6 and QKI-7, but not QKI-5, enhanced p27Kip1 protein expression (Fig. 1D). These results suggest that QKI isoforms display differential activities and may act via distinct mechanisms to promote OPC differentiation.
Along with the morphological changes, OPC differentiation is also characterized by the sequential appearance of specific surface differentiation markers at various stages (26). As shown in Fig. 2, A and B, the expression of FLAG-QKI-6 significantly increased the percentage of cells that harbor the OPC differentiation marker O1 after 72 h of induced differentiation. To test whether the RNA binding activity of QKI is required for promoting OPC differentiation, we transfected CG4 cells with a mutant FLAG-QKI-6 that carries the E48G mutation in the . In each experiment, more than 30 transfected cells were analyzed from randomly selected microscopic fields, and the cell processes are categorized as described under "Materials and Methods." S.E. is indicated for each data point. The numbers of secondary and tertiary processes are significantly increased in cells expressing each QKI isoform as compared with control cells expressing GFP alone based on standard t tests (p Ͻ 0.01). Inset, exogenous QKI isoform expression in CG4 cells. The housekeeping protein eIF5a is used as a loading control. Q5, Q6, and Q7, QKI-5, QKI-6, and QKI-7, respectively. C, p27Kip1 mRNA associates with FLAG-QKI-6 in transfected CG4 cells. Immunoprecipitation assay with anti-FLAG M2 beads was performed with cells transfected with either FLAG-QKI-6 or the PCDNA vector (PC). p27Kip1 mRNA is detected by reverse transcription-PCR in the eluate (E) of immunoprecipitate derived from QKI-6-transfected cells but not from control cells. I, input lysate for immunoprecipitation. D, influence of exogenous QKI isoforms on p27Kip1 expression measured by immunoblot. Cells transfected by the PCDNA vector (PC) was processed in parallel as a negative control. extended KH domain. This mutation was previously shown to abolish the association of QKI with its RNA ligands (27). In contrast to wild type FLAG-QKI-6, the E48G mutant QKI-6 failed to promote either process branching or O1 expression, suggesting that QKI facilitates OPC differentiation by RNA binding-dependent mechanisms.
RNA Interference-mediated QKI Knockdown Abrogates OPC Differentiation, and QKI Isoforms Display Differential Abilities in Rescuing the Differentiation Defects-To determine whether QKI is functionally required for OPC differentiation, we intro-duced a small interfering RNA (siRNA) into proliferating CG4 cells specifically targeting the 5Ј-coding region that is common for all QKI isoforms (QKIsi, Fig. 3A). A negative control siRNA (Ctrlsi) that harbors no sequence homology to any mammalian mRNA (Ambion) was used in parallel transfection. The plasmid encoding GFP was co-transfected to mark cells that received the siRNA. Immunoblot analysis showed that transfection of double-stranded QKIsi eliminated greater than 80% of endogenous QKI isoforms (Fig. 3B). We also generated lentivirus that express QKIsi or Ctrlsi, respectively, and observed similar efficiency for knocking down the endogenous QKI. Process extension and branching in primary cultured rat OPCs was severely abrogated when cells were treated with the QKIsi virus, whereas cells treated with the Ctrlsi virus developed fully extended processes with vigorous branching (Fig.  3C, panels a and b). Similarly, transfection of synthetic QKIsi into CG4 cells significantly attenuated process branching during differentiation as compared with Ctrlsi-treated cells, evident by the markedly reduced secondary processes (Fig. 3C, panels c and d).
Quantitative analysis confirmed the significant decrease in the average number of secondary processes in QKIsi-treated CG4 cells (Fig. 3D), suggesting that elimination of QKI arrested OPC differentiation at early stages of morphogenesis. We also examined the effect of QKIsi on expression of the OPC differentiation marker O1 in CG4 cells (Fig. 4). After 72 h of differentiation, 40% of Ctrlsi-treated cells were O1-positive (Fig. 4A, red fluorescence), but very few (5%) of  . QKI is functionally required for OPC differentiation. A, schematic presentation for the location of the QKI siRNA in the common coding region for all QKI mRNAs. B, immunoblot analysis of protein extracts of CG4 cells detects endogenous QKI isoforms. QKI-5 co-migrates with QKI-7 above QKI-6. Diminished expression of all endogenous QKI proteins is obvious upon QKI siRNA treatment. Actin was used as a loading control. C, morphology of primary cultured rat OPCs transduced with the lentivirus either expressing control siRNA or expressing QKI siRNA (panels a and b, marked by the co-expressed GFP) and morphology of CG4 cells co-transfected with GFP and the indicated synthetic siRNA (panels c-f). Transfected CG4 cells expressing GFP were shown in panels e and f. Cells were induced to differentiate for 72 h. Scale bar, 25 m. D, quantification of the number of primary and secondary processes per cell from QKIsi-and Ctrlsi-treated CG4 cells from three independent experiments (n ϭ 3). Black bar, control siRNA; gray bar, QKI siRNA.
the QKIsi-treated CG4 cells expressed O1. Taken together, these data support the hypothesis that QKI is required for the normal differentiation of OPC.
The specific QKI isoform required for OPC differentiation can be determined based on its ability to rescue the differentiation defect in CG4 cells caused by QKIsi. Since siRNA-mediated mRNA cleavage requires perfect sequence match between the siRNA and the mRNA target, we constructed QKIsi-resistant mutant QKI isoforms in which three silent point mutations were introduced into the QKIsi target site (Fig. 5A). When co-transfected with QKIsi, the expression level of each wild type FLAG-QKI isoform was greatly reduced with similar efficiency, whereas the levels of mutant FLAG-QKI isoforms were largely unaffected as compared with that in Ctrlsi-treated cells (Fig. 5B). In our differentiation paradigm, we found that co-transfection of CG4 cells with the QKIsi-resistant QKI-5m or QKI-6m together with QKIsi significantly increased the number of O1-positive cells as compared with that in cells treated with QKIsi alone (Fig. 5C). Quantitative analysis further revealed that although each QKI isoform is expressed at similar levels, QKI-6 is more efficient than QKI-5 in rescuing CG4 cell differentiation (Fig. 5D). In contrast, QKI-7 failed to rescue dif-ferentiation of CG4 cells, regardless of the fact that ectopic overexpression of QKI-7 can promote OPC differentiation (Fig. 1, A and B [3][4][5] is due to the attenuation of cellcycle exit, which would lead to increased number of proliferating OPCs. To address this question, we pulse-labeled QKIsi-and Ctrlsitreated CG4 cells with BrdUrd and identified cells that had undergone active proliferation based on the incorporation of BrdUrd into the replicating genomic DNA as detected by immunofluorescent staining. As shown in Fig. 6, QKIsi treatment did not alter the numbers of BrdUrd-labeled proliferating cells. Quantitative analysis indicated a similar percentage of BrdUrd-labeled cells in QKIsi-and Ctrlsi-treated cultures, suggesting that endogenous QKI isoforms are not involved in governing the proliferation capacity of OPCs. To further test whether QKI knockdown may affect cell cycle progression, we performed fluorescence-activated cell sort analysis using CG4 cells that are treated with the QKIsi lentivirus or the control lentivirus. In this experiment, greater than 90% of cells were transduced, marked by the expression of GFP from the same virus. The DNA content in cells was labeled by propidium iodide, which was used to determine the distribution of cells at various steps of cell cycle. As shown in Fig. 7A, nearly identical fluorescence-activated cell sort profiles were observed in QKIsi-treated cells and the control cells. Quantitative analysis detected no significant difference between QKIsi-treated cells and control cells on the percentage distribution at the G 1 /G 0 phase (ϳ70%), the S phase (ϳ15%), and the G 2 /M phase (15%) (Fig. 7B). Although overexpression of exogenous QKI-6 and QKI-7 enhanced p27Kip1 (Fig. 1D), knocking down QKI in CG4 cells does not affect p27kip1 pro-  The percentage of cells at G 1 /G 0 , S, and G 2 /M is calculated, and the mean value from three independent experiments is displayed graphically with S.E. indicated. The black and gray bars represent cells treated with the control virus and the QKIsi virus, respectively. C, QKI knockdown does not affect p27Kip1 protein expression. Western blots were performed with anti-QKI and anti-p27Kip1 antibody using lysates derived from QKIsi-and Ctrlsi-treated CG4 cells. Actin was used as an internal loading control. tein expression (Fig. 7D), suggesting that endogenous QKI is not necessary for maintaining p27Kip1 expression. These results suggest that QKIsi-mediated differentiation arrest is not due to delayed cell cycle exit. Instead, QKI deficiency must directly affect differentiation of CG4 cells via cell cycle-independent mechanisms.

DISCUSSION
The above studies demonstrated that the selective RNAbinding protein QKI is necessary and sufficient for advancing OPC differentiation. In addition, specific QKI isoforms have differential abilities in promoting OPC differentiation, which requires the RNA binding activity of QKI. Furthermore, we show that endogenous QKI is not necessary for cell cycle control of the CG4 OPC cell line, suggesting that at least part of the function of QKI is achieved by direct influence on OPC differentiation, after proliferating OPCs exit the cell cycle.
The role of QKI in central nervous system myelination is well documented (10). QKI selectively interacts with a subgroup of mRNAs that encode essential myelin structural proteins to control their homeostasis and subcellular localization (11,12,23). The best characterized QKI targets for myelination include the myelin basic protein (MBP) mRNA (11)(12)(13) and the myelinassociated glycoprotein mRNA (28). QKI deficiency in OLs causes destabilization and mislocalization of the MBP mRNA (12) and dysregulated splicing of the myelin-associated glycoprotein mRNA (28), which are potential mechanisms for the failure of myelination in the quakingviable (qk v ) mutant mice. The functional importance of QKI in myelination is further reinforced by the fact that N-ethyl-N-nitrosourea-induced single base pair mutations in the qkI gene also cause hypomyelination (29,30). However, the detection of QKI in the earliest stage of OL lineage development (20,31) and the association of QKI with mRNAs critical for OL fate specification and differentiation (19,27) suggest that QKI must also play important roles in early OL development prior to myelin formation. Our results that siRNA-mediated QKI knockdown abrogates OPC differentiation (Figs. 3 and 4) provide the first evidence demonstrating the functional requirement of QKI in OL differentiation. The fact that the RNA binding-deficient E48G mutant QKI failed to promote OPC differentiation (Fig. 2) suggests that QKI supports OPC development via regulating its mRNA targets. Since QKI exerts its influence on OPC differentiation before the expression of myelin structural genes such as mbp, in the differentiation paradigm, QKI must regulate its mRNA targets that are critical for OPC differentiation but distinct from those that support myelination.
A previous report showed that overexpression of QKI-6 and QKI-7 can enhance OPC differentiation (19). However, the function of QKI-5 in OPC differentiation was not addressed. Our results confirmed the ability of exogenous QKI-6 and QKI-7 in promoting OPC differentiation (Fig. 1) and further indicated that overexpression of the nuclear isoform QKI-5 also promotes OPC differentiation (Fig. 1), regardless of the fact that overexpression of QKI-5 causes nuclear retention of the MBP mRNA and presumably may negatively affect myelination (11). The fact that QKI-5 promotes CG4 cell differentiation but not p27Kip1 expression (Fig. 1) suggests that the nuclear QKI isoform may advance OPC differentiation via distinct mechanisms as compared with that employed by the cytoplasmic QKI isoforms. In fact, QKI-5 is the most abundant isoform in OPCs during embryonic and neonatal development (14) and is the second most efficient QKI isoform in rescuing the differentiation defect caused by the QKI-siRNA (Fig. 5). One possible explanation is that QKI-5 may function to achieve nuclear retention of mRNAs that encode suppressors for OPC differentiation, which in turn supports OPC differentiation. QKI isoforms harbor differential abilities to support OPC differentiation, as indicated by the quantitative effect of each QKI isoform in promoting OPC differentiation (Fig. 1) and in rescuing the defect of differentiation caused by the QKI-siRNA (Fig. 5). In both experiments, QKI-6 is the most efficient isoform in supporting OPC differentiation. This is consistent with the recent finding that QKI-6 alone can significantly improve the hypomyelination phenotype of the qk v mutant mice (15). On the other hand, QKI-6 cannot completely rescue the defect of myelin gene expression at the early stage of myelin development (15) or the differentiation arrest by the QKI siRNA (Fig.  5), suggesting the functional importance of other QKI isoforms for the early development of the OL lineage. Among the QKI isoforms, QKI-7 displays weaker activity in supporting OPC differentiation (Figs. 1 and 5). Because QKI-7 is the least abundant isoform in OPCs and is up-regulated at a later stage of myelin development (14), it may perform a major role to support myelin formation and maintenance. This idea is consistent with the observation that preferential reduction of QKI-7 is associated with white matter disruption in schizophrenia patients (32,33). An alternative possibility is that QKI-7 may depend on the other QKI isoforms to promote OPC differentiation and could not function by itself when endogenous QKI isoforms are eliminated by RNA interference.
Forced expression of QKI-6 and QKI-7 can enhance cell cycle exit of OPCs (19). One proposed mechanism is that QKI binds and stabilizes the mRNA encoding the cyclin-dependent kinase inhibitor p27Kip1 (19), one of the key regulators for cell cycle arrest of OPCs (3,34). In CG4 cells, exogenous QKI-6 and QKI-7 also enhanced p27Kip1 expression (Fig. 1D). However, although overexpression of p27Kip1 can lead to forced cell cycle arrest of OPCs, it is not sufficient for inducing OPC differentiation (7,8). Reciprocally, p27Kip1 deficiency increases cell proliferation capacity but does not affect the timing of oligodendrocyte differentiation (31). Thus, the effect of QKI on promoting OPC differentiation must also involve cell cycleindependent mechanisms and is achieved by regulating other mRNA targets in addition to p27Kip1 mRNA. Consistent with this view, our results indicate that exogenous QKI-5 promotes OPC differentiation without enhancing p27Kip1 expression (Fig. 1D). Furthermore, elimination of QKI in CG4 cells does not alter proliferation capacity or cell cycle progression (Figs. 6 and 7), suggesting that the abrogation of OPC differentiation caused by siRNA-mediated QKI knockdown (Figs. [3][4][5] is not due to blocking cell cycle exit. In fact, knocking down endogenous QKI did not affect p27Kip1 protein expression (Fig, 7D), indicating that QKI is not necessary for governing p27Kip1 expression in CG4 cells. It is important to point out that p27Kip1 expression is under sophisticated regulation by multi-ple molecular mechanisms at the levels of transcription, translation, and protein stability in addition to the stabilization of p27Kip1 mRNA by QKI (8,(35)(36)(37)(38). Therefore, compensatory mechanisms may prevent changes of p27Kip1 expression even when QKI is eliminated, which provides a possible explanation why QKI is unnecessary for p27Kip1 expression and cell cycle exit of CG4 cells (Fig. 7). In this regard, QKI must exert direct influence on OPC differentiation/maturation, most likely via enhanced expression of its mRNA targets that encode proteins critical for OPC differentiation. Thus, siRNA-mediated QKI knockdown causes deficits in expression of these differentiation factors, which in turn attenuates OPC differentiation (Figs. 3-5).
One of the QKI targets that plays key roles in supporting OPC morphological differentiation is the microtubule-associated protein 1B (MAP1B) (39). QKI binds to the 3Ј-untranslated region of the MAP1B mRNA, and QKI-dependent mRNA stabilization underlies the developmentally programmed upregulation of MAP1B during OPC differentiation (23). Considering the well characterized function of MAP1B in supporting the dynamic assembly of the microtubule cytoskeleton and the critical role of microtubules in OPC differentiation (39 -42), the effect of QKI in facilitating MAP1B up-regulation is a conceivable mechanism for supporting the process extension of OPCs. Although QKI most likely regulates distinct sets of mRNA targets to support myelination and OPC differentiation, the full range of the mRNA targets of QKI at different stages of OL lineage development still remains undetermined. Identification of the downstream mRNA targets that are under control of QKI-dependent posttranscriptional regulation in future studies will provide critical insights regarding molecular mechanisms that govern OPC development in normal myelinogenesis, as well as the white matter disruption and repair in various myelin disorders, represented by multiple sclerosis and schizophrenia (33,43).