Ubiquitin C-terminal hydrolase L1 (UCH-L1) is a component of the ubiquitin system, which has a fundamental role in regulating various biological activities. However, the functional role of the ubiquitin system in neurogenesis is not known. Here we show that UCH-L1 regulates the morphology of neural progenitor cells (NPCs) and mediates neurogenesis. UCH-L1 was expressed in cultured NPCs as well as in embryonic brain. Its expression pattern in the ventricular zone (VZ) changed between embryonic day (E) 14 and E16, which corresponds to the transition from neurogenesis to gliogenesis. At E14, UCH-L1 was highly expressed in the ventricular zone, where neurogenesis actively occurs; whereas its expression was prominent in the cortical plate at E16. UCH-L1 was very weakly detected in the VZ at E16, which corresponds to the start of gliogenesis. In cultured proliferating NPCs, UCH-L1 was co-expressed with nestin, a marker of undifferentiated cells. In differentiating cells, UCH-L1 was highly co-expressed with the early neuronal marker TuJ1. Furthermore, when UCH-L1 was induced in nestin-positive progenitor cells, the number and length of cellular processes of the progenitors decreased, suggesting that the progenitor cells were differentiating. In addition, NPCs derived from gad (UCH-L1-deficient) mice had longer processes compared with controls. The ability of UCH-L1 to regulate the morphology of nestin-positive progenitors was dependent on its binding affinity for ubiquitin but not on hydrolase activity; this result was also confirmed using gad-mouse-derived NPCs. These results suggest that UCH-L1 spatially mediates and enhances neurogenesis in the embryonic brain by regulating progenitor cell morphology.
Ubiquitin C-terminal hydrolase L1 (UCH-L1) is a member of the deubiquitylating enzymes and is one of the most abundant proteins in the brain. Whereas other UCH members are ubiquitously expressed, UCH-L1 is selectively expressed in neurons and testes/ovaries in the adult (Wilkinson et al., 1989). UCH-L1 is also known as PGP9.5 and is used as a neuron-specific marker in neuroanatomical and neuropathological studies (Dickson et al., 1994; McQuaid et al., 1995). Recent studies suggest that UCH-L1 is involved in neurodegeneration. The I93M mutation and the S18Y polymorphism in UCH-L1 are implicated in Parkinson's disease (Leroy et al., 1998; Satoh and Kuroda, 2001). Using gracile axonal dystrophy (gad) mice, we previously demonstrated that the dying-back type of axonal degeneration is caused by a deletion of the Uchl1 gene (Saigoh et al., 1999). UCH-L1 has an affinity for ubiquitin and ensures its stability within neurons in vivo (Osaka et al., 2003). Furthermore, UCH-L1 has ubiquitin ligase activity (Liu et al., 2002). Thus, UCH-L1 might have multiple functions and more roles in biological phenomena than previously expected.
UCH-L1 mRNA is first detected at embryonic day (E) 8.5-9 in the neural tube and in the neural epithelium (Schofield et al., 1995). In addition, UCH-L1 immunoreactivity has been observed in the neural tube at E10.5 (Sekiguchi et al., 2003). However, its functional role in embryonic neurogenesis is not well understood. CDK5 and Dab1 are involved in regulating the migratory behavior of postmitotic neurons. Both p35, which is a CDK5 kinase, and Dab1 are degraded by the ubiquitin-proteasome pathway (Arnaud et al., 2003; Bock et al., 2004; Patrick et al., 1998). Thus, the ubiquitin system might be important in the migration and differentiation of postmitotic neurons and for the lamination pattern of the cerebral cortex.
Neural progenitor cells (NPCs) differentiate into neurons, astrocytes and oligodendrocytes (Qian et al., 1998; Qian et al., 2000; Shen et al., 1998). In the embryonic brain, neuroepithelial cells and radial glia are present in the ventricular zone (VZ); neurogenesis occurs first, followed by gliogenesis. Committed progenitor cells move from the VZ to the cortical plate (CP) (Noctor et al., 2004). The differentiating cells migrate by means of radial migration, during which the migrating cells change their morphology (Kawauchi et al., 2003; Noctor et al., 2002; Tabata and Nakajima, 2003). Here, we analyzed the functional role of UCH-L1 using mouse embryonic NPCs. Our results indicate that UCH-L1 is expressed in nestin-positive NPCs and might regulate neurogenesis. The expression pattern of UCH-L1 changed in parallel with the transition from neuronal generation to glial generation. Furthermore, UCH-L1 modulated the length of nestin-positive processes in NPCs. Our results constitute the first evidence that UCH-L1 is important in neurogenesis and thus provide the basis for further investigation into the role of the ubiquitin system in neurogenesis.
UCH-L1 expression in embryonic mouse brain
We first determined the specificity of the UCH-L1 antibody using immunoblotting (data not shown) and immunostaining. Because gad mice do not express endogenous UCH-L1 (Saigoh et al., 1999), we used these mice as a negative control. Heterozygous littermates had UCH-L1 immunostaining, whereas UCH-L1 immunoreactivity was not detected in the brains of gad mice (Fig. 1). These results confirmed the specificity of the antibody against UCH-L1. Using this antibody, we further compared the distribution and expression of UCH-L1 with the neural progenitor marker nestin and the early neuronal marker TuJ1. Nestin was expressed in the VZ of brains from both gad and heterozygous mice at E13 (Fig. 1). Nestin expression was observed throughout the region, whereas TuJ1 immunoreactivity was detected at the marginal zone (MZ). In heterozygous mice, UCH-L1 and nestin immunostaining overlapped in almost all cells in the VZ, suggesting that UCH-L1 is expressed in NPCs (Fig. 1A). TuJ1 expression colocalized with that of UCH-L1 in MZ cells, indicating that UCH-L1 is expressed in embryonic neurons as well (Fig. 1B). In E13 gad mouse brain, nestin staining differed compared with that in heterozygous littermates. Nestin staining was observed in many long radial fibers in the mutant, which we believed were radial glia; by contrast, staining in the heterozygotes occurred in radial glia as well as in neuronal cells at various stages of development (Fig. 1A; arrow and arrowhead).
We then looked for developmental changes in UCH-L1 expression. In the embryonic cerebral cortex, asymmetric cell division generates one neuron and one neural progenitor (Roegiers and Jan, 2004; Zhong et al., 1996; Zhong et al., 1997). These asymmetric cell divisions begin at E11, peak around E14, and subside after E16. At E14, astrocytes and oligodendrocytes are not yet present. However, at E16, glial cell production begins. The regional expression level for both nestin and TuJ1 did not change between E14 and E16 (Fig. 2A,B). At E14 and E16, nestin immunoreactivity was stronger in the VZ (Fig. 2A) and was faintly detected only along radial glial fibers in the CP (Fig. 2A,C; arrowhead) (Malatesta et al., 2003; Malatesta et al., 2000). TuJ1 immunoreactivity was predominantly detected in the MZ, CP, intermediate zone and subventricular zone at E14 and E16 (Fig. 2B,D). In the VZ, TuJ1 immunoreactivity was detected only in migrating neurons (Fig. 2D; arrowhead).
By contrast, the pattern of UCH-L1 expression changed between E14 and E16 (Fig. 2A,B). At both stages of development, UCH-L1 was expressed in neuronal cells as well as in progenitor cells. UCH-L1 immunoreactivity was stronger in the VZ than in the CP at E14; however, the immunoreactivity was stronger in the CP than in the VZ at E16 (Fig. 2A,B). The regional change in UCH-L1 expression between E14 and E16 was further confirmed by measuring immunofluorescence intensities from confocal images of the MZ/CP and VZ. At E14, the relative UCH-L1 expression level in the VZ was 9.3 times higher than that in the MZ (Fig. 2A).
Conversely, at E16, when neuronal maturation occurs in the CP, UCH-L1 immunoreactivity in the CP was 5.0 times higher than in the VZ (Fig. 2B). UCH-L1 immunoreactivity colocalized with that of nestin in the VZ at both E14 and E16, although UCH-L1 expression in the VZ was lower at E16 (Fig. 2C). In the VZ at E14, nestin was expressed homogeneously; however, the pattern of UCH-L1 immunoreactivity was mixed, with strong and weak intensities (Fig. 2C; arrow). This expression pattern might reflect the heterogeneity of progenitor cells. Nestin-positive radial glial fibers were observed in the CP at E16 through mature neurons, which strongly expressed UCH-L1 (Fig. 2C) (Malatesta et al., 2000; Malatesta et al., 2003).
UCH-L1 and nestin expression in cultured NPCs
Because areas of nestin and UCH-L1 immunoreactivity overlapped in the VZ, where NPCs reside, we subsequently analyzed the transition of UCH-L1 expression using cultured NPCs. We performed double-labeling experiments for UCH-L1 and nestin expression in cultured NPCs. In the presence of basic fibroblast growth factor (bFGF), when NPCs are proliferating, the percentage of UCH-L1/nestin double-positive cells did not change 48 hours after plating, and almost all NPCs expressed UCH-L1 (Fig. 3A). The majority (97.5±2.2%; mean±s.d.) of cultured cells were nestin positive and most of them also stained for UCH-L1 2 hours after plating without bFGF, which triggers NPC differentiation. UCH-L1/nestin double-positive cells were detected at all time points, but as differentiation proceeded their numbers gradually decreased from 95.8±1.9% at 2 hours to 21.5±5.8% at 48 hours (Fig. 3A,B). Although UCH-L1 single-positive cells were rarely detected at 2 hours, the population increased with differentiation, and by 24 hours after bFGF removal 55.1±2.9% of cultured cells were UCH-L1 single-positive cells. Conversely, nestin single-positive cells were readily detected during the earlier phase of differentiation, especially at 6 hours (26.4±8.4% of total cells) and 12 hours (27.0±14.0% of total cells). The differentiating NPCs included nestin-positive cells in which UCH-L1 was either strongly or weakly expressed (Fig. 3A; arrow and arrowhead at 6 hours). These data indicate that UCH-L1 is expressed in progenitor cells as well as in differentiating NPCs. Nestin-positive cells can probably be categorized into at least two subgroups based on their UCH-L1 expression (Fig. 3A,B).
UCH-L1 and TuJ1 expression in cultured NPCs
We then analyzed the expression patterns of UCH-L1 and TuJ1. In the presence of bFGF, TuJ1-positive cells were rarely detected. However, in the absence of bFGF, TuJ1-positive cells were induced. In the cultures without bFGF, as the UCH-L1 single-positive cell population decreased with time, the UCH-L1/TuJ1 double-positive population increased (Fig. 4A,B). UCH-L1/TuJ1 double-negative cells were detected in the differentiating phases at 6, 12, 24 and 48 hours. UCH-L1/TuJ1 double-negative cells might be the nestin single-positive cells at 6 hours and 12 hours in Figs 3 and 4. TuJ1 single-positive cells were infrequently detected in the differentiating NPCs. Because 71.4±3.4% of NPCs differentiated into TuJ1-positive cells under our culture conditions without bFGF at 48 hours, almost all UCH-L1-positive cells are thought to differentiate into TuJ1-positive neuronal cells (Fig. 4A,B). The differentiating NPCs included TuJ1-positive cells in which UCH-L1 was either strongly or weakly expressed (Fig. 4A). These data indicate that UCH-L1-positive NPCs have a high potential for differentiating into neuronal cells and that TuJ1-positive neuronal cells are heterogeneous with regard to UCH-L1 expression.
Morphological classification of UCH-L1-positive NPCs
Nestin is a marker of undifferentiated cells, whereas UCH-L1 is a neuron-specific marker. Here, UCH-L1/nestin double-positive cells were present in cultured NPCs as well as in embryonic brain (Figs 2, 3). Cultured NPCs sequentially gave rise to neurons, then astrocytes, and finally oligodendrocytes (data not shown). Under our culture conditions, neurogenesis actively occurred in differentiating NPCs between 2 and 12 hours after plating (Fig. 4). Glial differentiation had not begun by this time. We collected differentiating NPCs at 6 hours and 12 hours after plating and then analyzed the morphology of nestin-positive cells (Fig. 5). Both UCH-L1/nestin double-positive cells and nestin single-positive cells were present in the population of differentiating NPCs. As the population of double-positive cells might represent a progression of differentiating neurons, we examined the morphology of these cells. Differentiating neurons undergo a stereotypical set of morphological changes, including length (from long to short) (Fukuda et al., 2003; Hartfuss et al., 2003; Nadarajah et al., 2001). We categorized the nestin-positive cells with respect to process length (long, short or round; Fig. 3). UCH-L1 single-positive and double-negative cells were included in the total number of cells. When the total length of processes was more than four times the diameter of the nucleus of the cell, the cell was categorized as `long', whereas cells with shorter processes were categorized as `short'. Cells that did not have processes were classified as `round'. At 6 hours, the majority of nestin single-positive cells were long (18.2±7.6% vs 4.0±0.2% short cells; mean±s.d.; Fisher's PLSD, P=0.008), whereas the majority of UCH-L1/nestin double-positive cells were short (62.0±6.3%). This population was significantly greater than that of long cells (10.3±2.0%) and round cells (5.0±1.7%; Fisher's PLSD, P<0.0001). When NPCs with processes were subcategorized as unipolar, bipolar or multipolar, the unipolar population was significantly higher (62.3±16.9%) than the bipolar population (18.2±3.9%; Fisher's PLSD, P=0.002) in UCH-L1/nestin double-positive cells. Multipolar cells were not observed at 12 hours. However, in nestin single-positive cells, more NPCs were bipolar (16.5±4.6%) than unipolar (4.5±1.9%; Fisher's PLSD, P=0.009; Fig. 6B). Similar results were obtained at 12 hours (Fig. 6). Thus, most UCH-L1/nestin double-positive cells had shorter processes and were more likely to be unipolar.
Effect of UCH-L1 on nestin-positive processes
We next examined the effect of UCH-L1 on proliferating NPC morphology using the transient transfection method. NPCs were allowed to proliferate for 48 hours after transfection and were then induced to differentiate for 12 hours. The cells were fixed, and the length of nestin-positive processes was examined. To quantify the relationship between UCH-L1 expression and process formation, we measured the total length of nestin-positive processes. Untransfected NPCs that were nestin positive had mainly long, bipolar processes (Fig. 3A, +bFGF). Cells that were transfected with a green fluorescent protein (GFP) expression vector (negative control) had a morphology that was similar to that of untransfected cells (Fig. 6A). By contrast, cells transfected with wild-type (WT) UCH-L1 cDNA had significantly shorter processes (47.6±6.4 μm, mean±s.e.m., n=81) than mock-transfected cells (69.9±7.0 μm, n=82) (Fig. 6A).
We then examined the relationship between the UCH-L1 structure and its activity with respect to morphological induction. We prepared two UCH-L1 mutants: D30A UCH-L1 lacked hydrolase activity and binding affinity for ubiquitin (Fig. 6B,C) (Osaka et al., 2003); C90S UCH-L1 lacked hydrolase activity but maintained binding affinity for ubiquitin (Fig. 6B,C) (Osaka et al., 2003). We compared the deubiquitylating activity of each UCH-L1 mutant using Ub-AMC as a substrate. The D30A mutant had little hydrolase activity, and the activity of the C90S mutant was not detectable (Fig. 6B; right). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed that there were no detectable contaminating proteins in these recombinant protein preparations (Fig. 6B; left). Co-immunoprecipitation experiments demonstrated that WT UCH-L1 and the C90S mutant physically associated with monoubiquitin. The D30A mutant (as well as GFP alone, which was used as a control) did not associate with ubiquitin (Fig. 6C). Although we did not detect a statistically significant difference, cells transfected with the D30A mutant tended to have longer nestin-positive processes (83.4±7.1 μm, n=87) as compared with cells transfected with the GFP expression vector (Fig. 6A). By contrast, cells transfected with the C90S mutant had significantly shorter fibers (39.3±4.5 μm, n=120; ANOVA: F=11.5, P<0.0001; Dunnett's multiple comparison test: GFP vs WT, P<0.05; GFP vs C90S, P<0.001; GFP vs D30A, P>0.05; Fig. 6A). We also compared the length of nestin-positive processes among UCH-L1 mutants (Bonferroni-Dunn Multiple Comparison Test: WT vs C90S, P=0.32; WT vs D30A, P<0.0001; D30A vs C90S, P<0.0001). Taken together, our data suggest that the effect of UCH-L1 expression on NPC morphology is dependent on the interaction between monoubiquitin and UCH-L1.
A comparative experiment using gad-mouse-derived NPCs
We did a comparative experiment using gad mice and heterozygous littermates. Nestin-positive NPCs from gad mice had longer processes. When we measured the length of nestin-positive fibers, NPCs from gad mice (45.0±1.4 μm, mean±s.e.m., n=366) had significantly longer nestin-positive processes compared with the control (31.4±1.3 μm, n=363) (Mann-Whitney U test: gad vs control, P<0.0001; Fig. 7A,B).
We next examined the effect of UCH-L1 on gad-mouse-derived NPCs using the transient transfection method. As observed in B6-derived cells, NPCs from gad mice that were transfected with WT UCH-L1 cDNA had significantly shorter processes (22.2±2.7 μm, mean±s.e.m., n=70) than mock-transfected cells (61.0±4.9 μm, n=88) (Bonferroni-Dunn multiple comparison test: GFP vs WT, P<0.0001) (Fig. 7C). Similarly, cells transfected with the C90S mutant had significantly shorter fibers (28.9±3.1 μm, n=71) (GFP vs C90S, P<0.0001). Although we did not detect a statistically significant difference, cells transfected with the D30A mutant tended to have longer nestin-positive processes (63.3±5.9 μm, n=80) as compared with cells transfected with the GFP expression vector (GFP vs D30A, P=0.70) (Fig. 7C). We also compared the length of nestin-positive processes among UCH-L1 mutants (Bonferroni-Dunn multiple comparison test: WT vs C90S, P=0.32; WT vs D30A, P< 0.0001; D30A vs C90S, P<0.0001). Taken together, our data suggest that the effect of UCH-L1 expression on NPC morphology is dependent on the interaction between monoubiquitin and UCH-L1.
UCH-L1 is a neuron-specific marker in the adult brain. In the present study, we provide experimental evidence that UCH-L1 is expressed in NPCs (Figs 2, 3). Using immunohistochemistry in the mouse brain, we detected UCH-L1 expression at E14 and E16. Interestingly, the expression pattern differed between E14 and E16 (Fig. 2). At E14, when the CP is forming, UCH-L1 expression was higher in the VZ than in the CP. At E14, the VZ contains progenitor cells that are generating neurons in the neocortex (Hashimoto and Mikoshiba, 2004; Malatesta et al., 2003). By contrast, UCH-L1 expression at E16 was lower in the VZ than in the CP. At E16, neuritogenesis and neuronal maturation are active in the CP, and gliogenesis is beginning in the VZ (Rice and Curran, 2001). The cerebral cortex layer becomes thicker at E16, where glial cells are not yet generated. The staining pattern for TuJ1 and nestin did not change between E14 and E16 (Fig. 2), indicating that UCH-L1 is highly expressed in the cortical layer prior to gliogenesis. The change in the expression pattern of UCH-L1 was coincident with the transition from neurogenesis to gliogenesis in the VZ. These results raise the possibility that UCH-L1 mediates not only the neuronal differentiation of NPCs but also the transition from neurogenesis to gliogenesis.
Time is a pivotal factor in the programmed sequence that produces neurons and glial cells from NPCs (Qian et al., 2000), in that the switch from neurogenesis to gliogenesis is regulated by time. The mechanism behind this progression of the progenitor cells is not well understood. Cultured NPCs generate neurons first, followed by astrocytes and then oligodendrocytes (Qian et al., 2000; Temple, 2001). This order of production for each population has been verified in vivo (Sauvageot and Stiles, 2002). The pattern of UCH-L1 immunoreactivity suggests that UCH-L1 is required for the onset of neurogenesis, which is followed by glial differentiation (Fig. 2).
We thus examined the role of UCH-L1 in neurogenesis using cultured NPCs. In UCH-L1/nestin double-staining experiments, the number of double-positive cells decreased with time in culture (Fig. 3). Conversely, UCH-L1 single-positive cells increased. In the double-staining experiments for UCH-L1 and TuJ1, the number of UCH-L1 single-positive cells decreased with time in culture, whereas the number of UCH-L1/TuJ1 double-positive cells increased (Fig. 4). These observations suggest that most UCH-L1-positive cells initially express nestin, but they express TuJ1 at a later stage. As we observed in vivo and in vitro (Figs 2, 3, 4), NPCs express UCH-L1, and its expression increases as the NPCs differentiate into neuronal cells. The number of nestin single-positive cells transiently increased before the UCH-L1 single-positive population increased (Fig. 3). The nestin single-positive population might have changed into the UCH-L1/nestin double-negative population (Fig. 3). Although the fate of the double-negative populations remains unknown, the double-negative cells might represent glial cells. Alternatively, some of the nestin single-positive cells might have changed into UCH-L1/nestin double-positive cells and then differentiated into UCH-L1 single-positive cells. A few UCH-L1-negative and TuJ1-positive cells were detected in the differentiating NPCs (Fig. 4). Thus, TuJ1-positive early neurons appear to be heterogeneous. UCH-L1/TuJ1 double-positive immunoreactivity suggested that UCH-L1 is not absolutely required for some portion of neuronal cell development (Fig. 1B and Fig. 4A). This might explain why gad mouse neurons develop despite the absence of UCH-L1.
Because UCH-L1 was expressed in nestin-positive NPCs, we further examined the role of UCH-L1 in cell morphology (Fig. 5). Differentiating NPCs change morphology (Noctor et al., 2001), but the role of UCH-L1 in differentiating neurons has not been investigated. We classified nestin-positive cells based on the length of their processes. Nestin single-positive cells were predominantly long, whereas most UCH-L1/nestin double-positive cells were predominantly short (Fig. 5). These results suggest that UCH-L1 plays a role in regulating NPC process length. We examined this possibility by inducing UCH-L1 in nestin-positive cells. Untransfected, proliferating nestin-positive NPCs had mainly long and bipolar processes [Fig. 3A, bFGF (48 hours)], but when UCH-L1 was transfected, the length of nestin-positive NPC processes shortened (Fig. 6A). The unipolar population increased following UCH-L1 expression. These results support the idea that UCH-L1 regulates NPC morphology. This idea was further confirmed by observations in NPCs from gad mice; as shown in Fig. 7B, NPCs from homozygous gad mice had longer processes than those from heterozygous controls. In addition, we observed that transfection of UCH-L1 shortened the processes of NPCs from gad mice compared with mock transfectants (Fig. 7C).
Our results also suggest that at least two populations of NPCs exist in the embryonic brain. The populations can be classified by the presence or absence of UCH-L1. In the dentate gyrus of the adult mouse brain, there are two distinct subpopulations of nestin-positive cells (Fukuda et al., 2003): those having short processes differentiate into neurons, whereas those having long processes generate late progenitors, which have short processes. The nestin staining pattern of brains from gad mice differed from that of brains from heterozygous littermates (Fig. 1). In the gad mouse brain, nestin-positive radial fibers were prominent, and almost all progenitor cells appeared to have long processes (Fig. 1). Since UCH-L1 affected NPC morphology (Fig. 6A and Fig. 7C), the difference in vivo indicates that differentiation itself was modulated by the absence of UCH-L1. Considering that neurons are present in the gad mouse even though it lacks UCH-L1 expression, further investigation into the morphological role of UCH-L1 using various approaches including the BrdU studies should provide important information about the heterogeneity of cortical neurons.
UCHs hydrolyze ubiquitin C-terminal small adducts in vitro (Larsen et al., 1998). Recently, a significant relationship was reported between UCH-L1 hydrolase activity and cell proliferation in lung cancer cell lines (Liu et al., 2003). We previously demonstrated that UCH-L1 extends ubiquitin half-life and prevents ubiquitin degradation. This function depends on the interaction between UCH-L1 and monoubiquitin but not on hydrolase activity (Osaka et al., 2003). In the present study, WT UCH-L1 and the C90S mutant both decreased the length of NPC processes. Both molecules associate with monoubiquitin, unlike another mutant, D30A, which did not affect process length (Fig. 6). Similar results were obtained from the transfection study using nestin-positive NPCs from gad mice (Fig. 7C). Thus, the effect of UCH-L1 on NPC process length is dependent on the interaction between UCH-L1 and ubiquitin but not on hydrolase activity. Although we did not examine the ligase activity of each mutant (Liu et al., 2002), the C90S mutant is unlikely to have ligase activity, because conjugation of ubiquitin to the C90S mutant forms a stable complex that prevents the release of ubiquitin (Sullivan and Vierstra, 1993). This observation suggests that the ligase activity is not related to the morphological changes that occurred in NPCs.
The ubiquitin system has an essential role in various physiological events, including cell-cycle progression, specific gene transcription, membrane protein trafficking, reversal of stress damage and intracellular signaling (Weissman, 2001). In cortical neurogenesis, the role of the ubiquitin system is not well understood. Several molecules that are important in cortical neurogenesis, including Notch, P35 and Dab1, are ubiquitylated (Arnaud et al., 2003; Bock et al., 2004; Patrick et al., 1998; Qiu et al., 2000). Here we show for the first time that UCH-L1 is expressed in NPCs and regulates their morphology. In addition, in vivo UCH-L1 expression is localized to the VZ and cortical layers that are undergoing neurogenesis. Cells undergoing gliogenesis had little UCH-L1 expression in vivo. These results suggest that UCH-L1 facilitates neurogenesis, an activity that appears to depend on the affinity of UCH-L1 for ubiquitin.
Materials and Methods
Pregnant C57BL/6J mice were purchased from CLEA Japan. The gad mouse is an autosomal recessive mutant that was obtained by crossing CBA and RFM mice (Saigoh et al., 1999). The gad line was maintained by intercrossing for more than 20 generations (Kwon et al., 2003; Saigoh et al., 1999). All animal experiments were performed in the laboratory according to the NIH Standards for Treatment of Laboratory Animals.
Antibodies and reagents
Monoclonal and polyclonal antibodies used in this study were as follows: monoclonal anti-nestin antibody (Becton Dickinson; and Rat401, Developmental Studies Hybridoma Bank, The University of Iowa, Iowa City, IA), monoclonal anti-neuronal tubulin β III antibody (TuJ1; Covance), polyclonal anti-UCH-L1 antibody (PGP9.5; RA95101, UltraClone), and polyclonal anti-FLAG antibody (Sigma). All secondary polyclonal antibodies conjugated to Alexa Fluor fluorescein were purchased from Molecular Probes.
Cortical NPC culture and differentiation conditions in C57BL/6 mice
Cortical NPCs were cultured as previously described (Nakashima et al., 1999). Briefly, embryos were removed from pregnant C57BL/6J mice (CLEA Japan) and staged according to morphological criteria to confirm the gestational day (Kaufman et al., 1998). Developing mouse cerebral cortex was dissected from E14 embryos. Cells were mechanically dissociated by trituration and plated at a concentration of 3.0×106 cells per 10 cm dish (Becton Dickinson) precoated with 10 ml of 15 μg/ml poly-l-ornithine (Sigma) and 10 ml of 1 μg/ml fibronectin (Nitta Gelatin). Cells were expanded for 5 days in serum-free neurobasal (NB) medium (Invitrogen) supplemented with B27 (Invitrogen), 0.5 mM l-glutamine (Invitrogen), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen). This medium contained 10 ng/ml bFGF (PeproTech). Cultures were maintained at 37°C in an atmosphere of 95% air and 5% CO2. For secondary culture, bFGF-expanded NPCs were washed in warm Hank's Balanced Salt Solution, detached with mechanical pipetting, and resuspended in NB medium supplemented with B27, but not bFGF. Cells were then replated in 24-well plates (Nunc; 1.8×105 cells per well) that were precoated with 500 μl of 15 μg/ml poly-l-ornithine and 500 μl of 1 μg/ml fibronectin for immunofluorescence staining at each time point.
Cortical NPC culture and differentiation conditions in gad mice
Culture of NPCs derived from gad mice was performed as with NPCs derived from B6 mice. Developing mouse cerebral cortex was dissected from embryos at E13.5 to E14.5. The precise gestational day was determined according to previously established morphological criteria (Kaufman et al., 1998). NPCs from each embryo were collected and cultured separately. Each genotype was determined later using PCR and, as a result, each pair of gad and control littermate mice from two sets of parents were used. Each culture of NPCs was replated in 24-well plates without bFGF and stained using anti-UCH-L1 24 hours after plating.
Brain sections were stained as previously described (Li et al., 2003; Osaka et al., 2003). Briefly, E14 and E16 mouse brains were removed and fixed in 4% paraformaldehyde/phosphate-buffered saline (PBS) for 2 hours at room temperature, cryoprotected in 30% sucrose in PBS and frozen in dry ice. Sections (20 μm thick) were cut on a cryostat, and mounted on aminopropylsilane (APS)-coated glass slides. They were then washed three times in PBS for 5 minutes, and blocked for 1 hour at room temperature with 3% bovine serum albumin, 2% (v/v) normal goat serum, and 0.2% (v/v) Triton X-100 in PBS (pH 7.4). Sections were incubated with primary antibodies [anti-nestin antibody (Rat401) 1:10; or anti-UCH-L1 antibody (RA95101) 1:4000; or anti-TuJ1 antibody, 1:1000] overnight at 4°C or for 2 hours at room temperature. After rinsing in PBS, the sections were incubated for 2 hours with diluted fluorescein-conjugated secondary antibody (1:200). The images were obtained with a confocal laser scanning TCS SL microscope, and detailed analyses were performed using an LSC confocal microscope system (Leica). Immunofluorescence intensities were measured from confocal images with Mac SCOPE software (version 2.59; Mitani).
Cells were stained as previously described (Aoki et al., 2002). Briefly, all incubations and washes were performed at room temperature. Cells were fixed with 3.8% formaldehyde/PBS for 10 minutes and permeabilized with 0.02% (v/v) Triton X-100/PBS for 5 minutes. Fixed cells were blocked with 3.3% goat serum for 30 minutes. Cells were incubated with a diluted primary polyclonal or monoclonal antibody (both were used for double staining) for 0.5-1 hour. The cells were then incubated with diluted secondary antibody conjugated to fluorescein for 0.5-1 hour. Antibody dilutions were as follows: anti-UCH-L1 antibody, 1:4000; anti-nestin antibody, 1:500; anti-TuJ1, 1:500. All secondary antibodies were diluted 1:200 in 1% goat serum/PBS before use. The images were obtained with fluorescence microscopy on an IX70 microscope (Olympus).
For C57BL/6 mice, cells replated in 24-well plates were cultured overnight in growth medium containing bFGF and B27. The next day, each construct was transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. NPCs were allowed to proliferate for 48 hours after transfection and then induced to differentiate for 12 hours without bFGF. For gad-mouse-derived NPCs, transfection was done in a similar manner.
Expression plasmids for human UCH-L1 variants
Mutant cDNAs encoding human UCH-L1 containing either the D30A or C90S substitution were obtained using the QuikChange site-directed mutagenesis kit (Stratagene) with the following mutagenesis oligonucleotides: 5′-CAGTGGCGCTTCGTGGCCGTGCTGGGGCTGGAAG-3′ and 5′-CTTCCAGCCCCAGCACGGCCACGAAGCGCCACTG-3′ for D30A; 5′-CCATTGGGAATTCCTCTGGCACAATCGGAC-3′ and 5′-GTCCGATTGTGCCACAGGAATTCCCAATGG-3′ for C90S. Each single-nucleotide mutation in the resulting plasmids was confirmed by sequencing. Mammalian expression plasmids containing either FLAG-tagged human WT UCH-L1 or the D30A or C90S mutants were constructed using a pCI-neo mammalian expression vector (Promega). Bacterial expression plasmids containing either 6HN-tagged human WT UCH-L1 or the D30A or C90S mutants were constructed using a tetracycline-inducible expression system. XhoI-NotI cDNA fragments of the pCI-neo WT UCH-L1 or the D30A and C90S mutants and constructs were digested, and the DNA fragments were ligated between the SalI and NotI sites in pPROtetE233 (Clontech) to generate pPROtetE233 6HN-tagged human WT, D30A and C90S UCH-L1 vectors. These expression plasmids were confirmed by sequencing.
In vitro assay for human UCH-L1 activity
Purified human UCH-L1 and the fluorogenic substrate ubiquitin-7-amino-4-methylcoumarin (Ub-AMC; Boston Biochem) were used to determine steady-state kinetic parameters as described previously (Nishikawa et al., 2003).
NIH-3T3 cells stably expressing human WT UCH-L1 or the C90S or D30A mutants, all with an HA-FLAG double tag at the N terminus, were cultured to subconfluency in a 10 cm dish, lyzed with 1 ml of modified RIPA buffer [50 mM Tris-HCl, pH 7.5, 1% (v/v) NP-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA] with EDTA-free complete protease inhibitor cocktail (Roche), sonicated and centrifuged at 18,000 g for 20 minutes at 4°C. Immunoprecipitation was performed as described previously (Ogawa et al., 2002).
Statistical analyses were performed using StatView, version 5.0 (SAS) and Prism, version 3 (GraphPad Software). Analysis of variance (ANOVA) was used to assess differences between groups. A P value of less than 0.05 was considered statistically significant. When ANOVA results were statistically significant, they were examined by Fisher's PLSD, or Dunnett's multiple comparison test, or Bonferroni-Dunn multiple comparisons post hoc test. Differences between gad mice and control mice were analyzed using the Mann-Whitney U test.
The authors thank Yuh Nung Jan and Hua-Shun Li for providing the immunohistochemistry methods; Yoshihiro Nakatani and Hidesato Ogawa for providing the retroviral expression system and immunoprecipitation methods; and Masako Shikama for the care and breeding of animals. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Health, Labour and Welfare of Japan, and Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
- Accepted September 27, 2005.
- © The Company of Biologists Limited 2006