RanGAP‐mediated nucleocytoplasmic transport of Prospero regulates neural stem cell lifespan in Drosophila larval central brain

Abstract By the end of neurogenesis in Drosophila pupal brain neuroblasts (NBs), nuclear Prospero (Pros) triggers cell cycle exit and terminates NB lifespan. Here, we reveal that in larval brain NBs, an intrinsic mechanism facilitates import and export of Pros across the nuclear envelope via a Ran‐mediated nucleocytoplasmic transport system. In rangap mutants, the export of Pros from the nucleus to cytoplasm is impaired and the nucleocytoplasmic transport of Pros becomes one‐way traffic, causing an early accumulation of Pros in the nuclei of the larval central brain NBs. This nuclear Pros retention initiates NB cell cycle exit and leads to a premature decrease of total NB numbers. Our data indicate that RanGAP plays a crucial role in this intrinsic mechanism that controls NB lifespan during neurogenesis. Our study may provide insights into understanding the lifespan of neural stem cells during neurogenesis in other organisms.

eliminated through programmed cell death, whereas the cephalic and thoracic NBs enter mitotic quiescence at the embryo-larval transition (Prokop & Technau, 1991;Tsuji, Hasegawa, & Isshiki, 2008). After the hatching of the larvae, most of the remaining NBs are reactivated and resume asymmetric divisions in the late first-instar or early second-instar larval stages to contribute ∼90% of cells in the adult CNS (Sousa-Nunes, Cheng, & Gould, 2010;Srinivasan et al., 1998).
By the white prepupal stage at the end of neurogenesis, the NBs begin to exit the cell cycle and cease proliferation. The characteristic feature for the termination of NBs, both in the brain and thoracic segments, is the translocation of Pros into the nucleus (Cenci & Gould, 2005;Maurange, Cheng, & Gould, 2008). Studies have shown that as Pros enters the nucleus, the thoracic NBs precede to a slow and final cell division to produce two equal-sized daughters, resulting in NB cell cycle exit and the end of NB lifespan (Cenci & Gould, 2005;Maurange et al., 2008). The central brain NBs exit the cell cycle in a similar way (Chai, Liu, Chia, & Cai, 2013). By the end of neurogenesis at the pupal stage in Drosophila, all NBs, including those from the brain, thoracic, and abdominal segments, cease to have any proliferation potential. No actively dividing NBs have been found in adult flies (Fernandez-Hernandez, Rhiner, & Moreno, 2013;Trotha, Egger, & Brand, 2009).
For a given NB, the time to exit the cell cycle is precisely regulated and in a fixed order. It has been proposed that the basic molecular mechanism regulating NB cell cycle exit involves a stereotypical temporal series of transcription factors being sequentially expressed in the NBs (Isshiki, Pearson, Holbrook, & Doe, 2001;Maurange et al., 2008;Rossi, Fernandes, & Desplan, 2017;Syed, Mark, & Doe, 2017;Tsuji et al., 2008). This temporal transcription factor series was first identified in the proliferating embryonic NBs in the ventral nerve cord (VNC), where a serial of transient expressed transcription factors, such as Hunchback (Hb), Krüppel (Kr), Pdm, Castor (Cas), and Grainyhead (Grh), took place according to the developmental stages of the embryo (Grosskortenhaus, Pearson, Marusich, & Doe, 2005;Isshiki et al., 2001). Under the control of temporal patterning regulation, Pros was transiently detected in the NB nucleus and NBs exited cell cycle, becoming quiescent at the end of embryonic neurogenesis (Lai & Doe, 2014;Li & Vaessin, 2000).
Two transcription factors, Cas and Seven-up (Svp), have been reported to act as members of the postembryonic temporal series (Maurange et al., 2008;Syed et al., 2017). A recent study showed that the Hedgehog signaling pathway acts downstream of Cas but upstream of Grh to promote NB cell cycle exit (Chai et al., 2013).
Other members in the temporal series for the postembryonic central brain remain to be identified.
Ran is a small Ras-related GTPase that mediates the nucleocytoplasmic exchange of macromolecules across the nuclear envelope.
Extensive studies in nuclear trafficking have shown that Ran acts as a molecular switch to regulate the assembly and disassembly of nuclear transport receptor-cargo complexes (Lui & Huang, 2009;Matchett et al., 2014).
In mammalian cells, RanGAP is enriched in the cytoplasm and on the nuclear membranes (Zhang et al., 2015). In Drosophila primary spermatocytes and salivary gland cells, RanGAP is mainly localized at the outer periphery of the nuclear envelope (Kusano, Staber, & Ganetzky, 2001. In Drosophila, RanGAP is also highly expressed in many other tissues, including the CNS (https://flyba se.org/reports/FBgn0003346.html). RanGEF is a chromatin-bound guanine-nucleotide exchange factor which converts RanGDP to RanGTP in the nucleus (Bischoff & Ponstingl, 1991). Human RanGEF is the regulator of chromosome condensation 1 (RCC1).
Drosophila RanGEF is also known as Rcc1 or Bj1 (Joy, Hirono, & Doe, 2015). RanGAP and RanGEF (Rcc1) work together to maintain a functional Ran cycle. RanGTP in the cytoplasm is converted to RanGDP in the presence of RanGAP. RanGDP then enters the nucleus and exchanges GDP with GTP via Rcc1 function. RanGTP then moves out of the nucleus into the cytoplasm, completing a Ran cycle. This energy-coupled Ran cycle, together with other proteins including Importins, Exportins, and nucleopore protein complex (NPC), is responsible for nucleocytoplasmic transport (Alavian, Politz, Lewandowski, Powers, & Pederson, 2004;Matchett et al., 2014;Nagai & Yoneda, 2012;Sampathkumar et al., 2013). Dysfunction of the Ran cycle disrupts the nucleocytoplasmic trafficking that involved in a number of neurodegenerative diseases (Freibaum et al., 2015;Zhang et al., 2015). A recent study has also suggested that Rcc1 acts to promote the nuclear export of Pros in the NBs and that a mutation in Rcc1 leads to premature NB differentiation in the central brain (Joy et al., 2015).
In this study, we reveal that RanGAP is involved in an intrinsic mechanism that controls Pros nuclear translocation in larval central brain NBs. Mutations in rangap cause abnormal accumulation of Pros in the nuclei of NBs. This nuclear Pros leads to the cell cycle exit of NBs and the end of NB proliferating potential. Our data show that in the third-instar larval central brain NBs, Pros shuttles across the nuclear envelope via a nucleocytoplasmic transport system. Elimination of the RanGAP function does not affect the import of Pros but the export of Pros out of the nucleus. RanGAP or the Ranmediated nucleocytoplasmic transport system may serve as an end target for the temporal series of transcription factor cascade in controlling the NB lifespan.

| RanGAP is required for maintenance of NB number in the larval central brain
In an RNAi screen to identify the genes required for NB development, we observed the occasional mislocalization of Mira in the rangap RNAi knockdown larval central brain NBs (Supporting Information Figure S1A,B). This phenotype was confirmed in the third-instar larval central brains of rangap [EP1173] , a mutation with a P-element insertion in the 5′-UTR region of the gene (Supporting Information Figure S1C, G). All rangap [EP1173] females died at the pupal stage but some male animals of the same genotype survived to the adulthood. Based on viability results, we focused on mutant female animals. We performed immunofluorescence staining at the third-instar larval stage and found the gloss physical appearance of the brain lobes appeared to be normal (Figure 1a-c), but interestingly the mutant female animals had less total NB number as compared with wt (Figure 1a-c). One potential reason for this lower NB number in female brains could be the defective NB activation during late first-or early second-instar larval stages (Maurange & Gould, 2005;Maurange et al., 2008;Sousa-Nunes, Yee, & Gould, 2011). To explore this possibility, we quantified NB numbers of larval central brains at different developmental stages (Figure 1e,f).
The total NB numbers in the rangap [EP1173] central brains were comparable to those of the wt at the second-(48 hr ALH) and early third-instar larval stages (72 hr ALH), but dropped significantly to approximately 30% of the controls within the subsequent 24 hr. It is interesting to note that at a later stage (120 hr ALH), the NB numbers in male animals also drop notably (Figure 1e,f). This observation indicates that the reactivation of NBs is normal in the rangap [EP1173] female brains, and the disappearance of 70% of total NBs occurs in the late third-instar larval stage.
To exclude the possibility that the NB missing phenotype initially identified in rangap [EP1173] allele was due to the background mutations, we examined rangap [EP1173] /rangap 327 trans-heterozygous animals.
Interestingly, trans-heterozygous animals of both sexes were lethal at pupal stage and showed central brain NB missing phenotype (Figure 1d).
The quantification of NB numbers in this genetic background showed that the central brain NB numbers were comparable to those of wt and other two null alleles at 48 hr (ALH) but started decreasing significantly at 72 hr (ALH). The NB numbers further dropped to <30% at 96 hr and <20% at 120 hr (ALH), respectively (Figure 1e,f). The stronger phenotype observed in trans-heterozygous animals can be explained by a lower RanGAP expression levels (50% of that of rangap [EP1173] homozygous) and supports the proposal that NB missing phenotype exhibited in rangap [EP1173] allele is not due to background mutations.
We also suspected that apoptosis was the cause for the missing NBs, but the anti-Caspase 3 immunofluorescence staining failed to detect any obvious apoptotic central brain NBs in the rangap [EP1173] female animal brains. The apoptotic signals in rangap [EP1173] female brains seemed to be identical to those of the wt (Supporting Information Figure S1H,I). Based on this observation, we conclude that the loss of NB number is unlikely due to the apoptosis.
Since rangap [EP1173] is a hypomorphic allele and shows NB loss only in female animals, it would be interesting to know whether the phenotype is linked to RanGAP expression levels. An antibody against a Ran-GAP fragment from aa280 to 426 was generated in rabbits and exhibited specificity against RanGAP protein. Bands with appropriate molecular weights for the endogenous RanGAP and a tagged RanGAP protein were detected, respectively, on western blot (Supporting F I G U R E 1 RanGAP is required for maintenance of NB numbers. (a-d) Confocal images from the projection of Z-stack of the late thirdinstar larval brains (96 hr ALH) of wt (a), rangap [EP1173] male (rangap EP -m; b), and rangap [EP1173] female (rangap EP -f; c) animals, and the transheterozygous (rangap [EP1173] /rangap 327 ; d). Brain lobes were immunostained with anti-Dpn (green) to visualize the NBs in the central brain (CB). Compared with wt and rangap [EP1173] males, the NB numbers in rangap [EP1173] female and rangap [EP1173] /rangap 327 brains had decreased significantly. (e, f) Statistical analysis of the NB numbers/brain lobe at different developmental stages (48, 72, 96, 120 hr ALH) in the wt (n = 6, n = 8, n = 8, n = 10, respectively) and rangap [EP1173] brain lobes (male: n = 10, n = 10, n = 10, n = 15; female: n = 35, n = 40, n = 35, n = 45) and rangap [EP1173] /rangap 327 (n = 9, n = 9, n = 6, n = 11). According to the one-way ANOVA test and Dunnett's multiple comparison test, at 48 and 72 hr ALH, the NB numbers/lobe of both male and female animals were comparable to those of the wt brains.(at 48 hr, p = 0.6527 and 0.9683, at 72 hr, p = 0.0938 and 0.9637). The NB numbers of rangap [EP1173] female animals dropped significantly during the next 24 hr (p < 0.0001), for male, p = 0.6995. At 120 h ALH, NB numbers of rangap [EP1173] of both male and female are decreased (female, p < 0.0001; male, p = 0.0004). For trans-heterozygous rangap [EP1173] /rangap 327 animals, NB numbers decreased at 72 hr ALH (p < 0.0001). (l) The western blot analysis of the brain extracts from the wt and rangap [EP1173] males and females. The RanGAP expression levels decreased in the rangap [EP1173] mutants. (g-i)The second-instar larval brains of the wt (f), rangap 9 (g), and rangap 327 (h) groups were immunostained with Dpn (green). Both rangap 9 and rangap 327 NB number had decreased (l) as compared with the wt. (j, k) The confocal images of third-instar larval brains of the wt (i) and daughtless-Gal4>UAS-rangap;rangap 327 were double-labeled with anti-Mira (red) and anti-Dpn (green). Ectopic expression of RanGAP in rangap 327 background (daughtless-Gal4>UAS-rangap; rangap 327 ) rescued the loss of NB number phenotype. (m) Western blot data showing that RanGAP was not detected in rangap 9 and rangap 327 extracts, and these two mutants were null alleles. (n) NB number comparisons among the wt, rangap 9 , rangap 327 and the rescue with the ectopic expression of RanGAP (daughtless-Gal4>UAS-rangap;rangap 327 ) at 48 hr ALH (2nd instar). Significant differences between wt and rangap mutants were tested using one-way ANOVA test and Dunnett's multiple comparison test (p < 0.001). No significance in difference was detected between wt and UAS-rangap; rangap 327 using the same test. CB, central brain; OL, optic lobe. White dashed lines define the border between the CB and OL. Scale bar = 10 μm Information Figure S1J). Western blot data further showed that the RanGAP expression levels in rangap [EP1173] mutant brains of both male and female animals were significantly decreased as compared with the wt (Figure 1l). However, we failed to detect any expression discrepancy between male and female brains ( Figure 1l).
We generated two additional rangap alleles: rangap 9 , a frameshift mutation with five bases deleted near the N-terminal region with the CRISPR/Cas9 method and rangap 327 , a nonsense mutation due to a G to T change ("GAG" to "TAG") generated during remobilization of the P-element from rangap [EP1173] (refer to Materials and Methods; Supporting Information Figure S1G). The homozygous animals of rangap 9 and rangap 327 mutants only survived up to the second-instar larval stage. Western blot analysis showed that the RanGAP protein was not detectable for both alleles (Figure 1m). Immunofluorescent staining also failed to identify RanGAP signals in mutant NBs (Supporting Information Figure S1e,f). These data confirm that both rangap 9 and rangap 327 mutants are null alleles. We next quantified the central brain NB numbers of the second-instar larvae in both null alleles. The total NB numbers in these mutants were only about 80% of those in wt counterparts regardless of the sex of the animals (Figure 1g,h,i).
The phenotypic differences between rangap [EP1173] male and female animals observed earlier must be due to the low expression levels of RanGAP and unknown complexed sex-specific regulations.
We decided to pursue only the relationship between RanGAP and the lifespan of the NBs.
The overexpression of a rangap transgene construct (RanGAP-TAP) in the rangap 327 mutant background with daughterless-Gal4 (Supporting Information Figure S2A) effectively rescued the missing NB phenotype (Figure 1j,k,n). Based on all these results, we conclude that RanGAP is required for the maintenance of NB numbers in larval central brains.

| Lack of RanGAP causes nuclear Prosdependent NB cell cycle exit
We suspected that the loss of NBs in rangap mutants was due to the nuclear Pros-dependent NB cell cycle exit. As predicted, nuclear Pros was detected in the central brain NBs of the rangap [EP1173] , as well as in the two null allele mutants ( The MARCM clone data also indicate that in the absence of Ran-

| RanGAP modulates NB cell cycle progression
We noticed that in our MARCM clone analyses, the Group 1 NBs appeared to be less affected by the loss of RanGAP. One obvious feature that made NBs in Group 1 different from the wt counterparts was the smaller clone size (Figure 2j). This suggested that in the rangap mutant, the NB cell cycle was either delayed or stalled. To address this issue, we carried out Edu and PH3 labeling experiments to monitor S-phase and M-phase NBs. At 96 hr (ALH) in rangap [EP1173] female brains, the total numbers of NBs labeled by PH3 ( Figure 3a,b) or Edu (Figure 3c,d) were much less than those of the wt. Statistical analysis showed that only about 7% of NB populations were in the mitotic phase and 6% in or post-S-phase, while the percentages for the wt counterparts were 58% and 40%, respectively Based on these data, we conclude that RanGAP is required for correct NB cell cycle progression. This cell cycle delay or halt might be a result of the nuclear Pros-dependent cell cycle exit.

| Cytoplasmic RanGAP negatively correlates with nuclear Pros
RanGAP was widely expressed in the NBs and other cells in both the central brain (CB) and the optic lobes (OL; Figure 4a,b). We focused on the central brain NBs. At interphase, RanGAP was in the cytoplasm and restricted to the nuclear envelope in both Type I and Type II NBs

| RanGAP prevents nuclear Pros retention via the Ran cycle
It is known that RanGAP and Rcc1 work together to balance the RanGTP/RanGDP ratio in the Ran cycle (Nagai & Yoneda, 2012). When RanGAP activity is low, RanGTP will accumulate in the cytoplasm and the RanGTP/RanGDP ratio is altered. Reduction of Rcc1 expression should be able to compensate for the lowered RanGAP activity, reset the RanGTP/RanGDP ratio, and rescue the rangap phenotype. By introducing a copy of the Rcc1 NP4610 mutant gene to attenuate Rcc1 levels ( Figure 5h) in rangap [EP1173] mutant female animals (Figure 5b), we observed the restoration of NB numbers (Figure 5c,d) and the animals survived to adulthood. In addition, we did not observe any NBs (n = 368) with nuclear Pros in the third-instar larval central brains of these animals. This result indicates that the retention of nuclear Pros in the rangap mutant NBs is due to a defective Ran cycle.
F I G U R E 2 Lack of RanGAP leads to premature nuclear Pros-dependent NB cell cycle exit. (a-e) Confocal images of the third-instar larval brains of wt (a) and rangap [EP1173] female (b); and the second-instar larval brains of the wt (c), rangap 327 (d), and rangap 9 (e) labeled by anti-Pros (red) and TO-PRO3 (blue). In wt central brain NBs, no nuclear Pros was detected (6 third-instar brain lobes, 8 second-instar brain lobes). The nuclear Pros (arrowhead) was detected in every single brain lobe of all three rangap mutants (rangap [EP1173] :10 brain lobes, rangap 327 :14 brain lobes, and rangap 9 : 11 brain lobes). On average, two to three NBs with nuclear Pros were detected in each lobe. (f-i) The third-instar larval brains containing the wt (f), rangap 9 (g-i) MARCM clones double-labeled by anti-GFP (green) and anti-Pros (red). The wt clones showed large NBs with many small cells (f) and no nuclear Pros detection in the NBs. The rangap 9 clones showed three distinct phenotypes: large NB with many small cells but no nuclear Pros in the NB (g); large NB with many small cells and nuclear Pros in the NB (arrowhead, h) and only small cells without NBs (i). (j) Statistical data of NB clone size (cell numbers/clone) of the wt and rangap 9 clones. The data are plotted as mean ± SEM. Significant differences were assessed using a two-tailed unpaired t test (***p < 0.001). (k) The percentage of the three types of rangap 9 clones (with or without nuclear Pros, and clones without large NBs) at various developmental stages of 48, 72, and 96 hr ALH. Data information: Dotted lines mark the outline of the NBs. Scale bar = 10 μm A recent study has showed that the knockdown of Rcc1 in larval central brain NBs led to premature nuclear Pros phenotype identical to that of the rangap mutants ( This appeared contradictory to the observation that mutations in either rangap or Rcc1 resulted in an identical premature nuclear Pros phenotype since RanGAP and Rcc1 had opposite roles upon Ran status.
It is conceivable that in the nucleocytoplasmic transport system, energy coupled to maintain the transport system comes from the hydrolysis of GTP in the Ran cycle. Mutations in rangap or Rcc1 would then ultimately lead to malfunction in the Ran cycle and a defective nucleocytoplasmic transport system in the NBs. In this scenario, it is possible that the appearance of premature nuclear Pros in both mutants was due to a defective Ran-mediated nucleocytoplasmic transport system.

| Pros shuttles across nuclear envelope in NBs
Pros has both a nuclear localization signal (NLS) and a nuclear export signal (NES) which implies that Pros is able to be imported into and  The data are plotted as mean ± SEM. In these two groups, we analyzed seven and five brain lobes, respectively. (h) Detailed analysis of Edu + NBs per brain lobe in wt and rangap [EP1173] -f from figure f (p < 0.0001). Ten and eight brain lobes were analyzed, respectively. The error bars show the variation of the percentage. The data are plotted as mean ± SEM ( Figure 6a,b), we investigated whether the Importins play a role in Pros nuclear translocation. In the cytoplasm, the carrier protein Importin β forms complex with Importin α and the cargo proteins for the first step of protein import (Hutten et al., 2008). When Importin β was attenuated using RNAi in the interphase larval central brain NBs, Pros was largely cortical (arrowhead, Figure 6c), while Pros staining in wt NBs was quite weak, sometimes barely visible in the cytoplasm (arrowhead, Figure 6a).
This result suggests that Importin β is involved in the import of  both Importin β and RanGAP in the larval brain NBs. We reasoned that in this import and export double-blocking experiment, Pros should remain in its original compartments, and the nuclear accumulation of Pros should not occur. When Pros import/export was blocked by the attenuation of Importin β in the rangap 9 mutant background, nuclear Pros accumulation was not detected (arrowhead, Figure 6d) and the Pros immunofluorescence signal remained weak and comparable to that of wt NBs (arrowhead, Figure 6a). This result is consistent with our Pros shuttling hypothesis.
We next searched for other components involved in the protein export in nucleocytoplasmic transport system. Three nuclear pore proteins, Nup 107, Nup 160, and Nup 98-96, were of particular interest. It has been reported that mutations in these three proteins suppress G4C2 repeats-mediated neurotoxicity toxicity in the Drosophilar eyes (Freibaum et al., 2015). The mechanism of the suppression was most likely due to decreased protein export since KPT-276, a small-molecule inhibitor of nuclear export, also eased this toxicity (Zhang et al., 2015). Thus, these three nuclear pore proteins are likely involved in protein export.
When Nup 107, Nup 160, or Nup 98-96 was knocked down individually, a premature retention of nuclear Pros was observed (Figure 6e-g). Attenuation of Nup 107, Nup 160, or Nup 98-96 expression levels diminishes protein export capacity and recapitulates the rangap mutant phenotype. Based on these results, we conclude that the attenuation of nuclear export is able to cause premature accumulation of Pros in the nucleus. This conclusion is consistent with our hypothesis that rangap mutations diminish Pros export in the larval central brain NBs. (a) F I G U R E 5 Premature appearance of nuclear Pros is due to a defective Ranmediated nucleocytoplasmic transport system. (a-c, h) The confocal images of NBs of the third-instar larval brains stained with anti-Ase (green). The loss of NB number phenotype in rangap EP1173 mutant brains (b) was rescued by introducing one copy of the hypomorphic mutant Rcc1 NP4610 (h) into the rangap EP1173 mutant background (c). (d) Statistic analysis of rescue results. From the one-way ANOVA test and Dunnett's multiple comparison test, removal of one copy of RCC1 in rangap EP1173 , the NB number is greatly restored (p < 0.0001). n: number of the brain lobes examined. Scale bar = 20 µm.
(e-f) The knockdown of Rcc1 with worniu-Gal4 led to a premature nuclear Pros phenotype. Dotted circles outline NB shape. Scale bar = 10 µm. (g, i-n) Expression of Rcc1 or RanGAP was not mutually dependent. The western blot data show that neither Rcc1 knockdown (g) in the third-instar larval brain NBs affect RanGAP expression (g, i) nor the attenuation of rangap expression alter Rcc1 expression (j). In addition, the anti-Rcc1 signal (red) remained unchanged both in the control clone (k, l) and rangap 9 MARCM clones (arrow; n = 14, m, n), comparing with its neighboring wt NB (arrowhead). GFP (green). The dotted line outlines the clone size

| DISCUSSION
Pros is a homeodomain transcription factor expressed in NBs and contains both a nuclear localization signal (NLS) and a nuclear export signal (NES; Demidenko et al., 2001;Vaessin et al., 1991). Our data suggest that an intrinsic mechanism maintains Pros shuttling across the nuclear envelope in the larval brain NBs. Mutations in rangap disrupt this mechanism and lead to the impaired export of Pros, causing early accumulation of Pros in the nucleus and premature NB cell cycle exit. Our Caspase 3 staining and P35 expression data support the conclusion that apoptosis did not play a role in rangap NB missing phenotype.

| RanGAP is required for NB maintenance in late larval stage
MARCM analysis shows that with the attenuation of RanGAP, the premature nuclear accumulation of Pros only occurs after 72 hr ALH. At earlier developmental stages, the lack of RanGAP does not appear to be necessary to prevent Pros premature retention in the nucleus. This observation may explain the relatively weak phenotype of NB loss in the second-instar larvae of both homozygous null alleles.
One of the possible reasons why nuclear Pros is only detected from 72 hr (ALH) onwards might be due to the maternal effect of rangap. Since rangap null alleles only survive to the second-instar larval stage (48 hr ALH), zygotic rangap must be required after 48 hr (ALH). If it were maternal rangap that prevented the retention of nuclear Pros, we would expect to observe the phenotype at the second-instar larval stage (48 hr ALH). It is possible that other factors might have redundant functions preventing the nuclear retention of Pros. Alternatively, Pros might not shuttle across the nuclear envelope at an earlier developmental stage.

| Defective Ran cycle leads to earlier retention of nuclear Pros in the larval NBs
It is interesting to note that mutations in rangap shared the same phenotype as the Rcc1 mutant (Joy et al., 2015).  F I G U R E 6 Transcription factor Pros is constantly shuttling across the nuclear envelope in wt NBs. (a-a′′) The localization pattern of Pros in wt third-instar larval brains. Pros is not nuclear or cortical in the NBs. (brain lobes = 15). (b-b′′) In the absence of RanGAP, premature nuclear Pros accumulation was detected in the NBs (arrow) in the second-instar larval brains of rangap 9 . Nuclear Pros was detected in two or three NBs per lobe (brain lobes = 21). This low number agrees with our finding that at 48 hr ALH, few NBs contain nuclear Pros (refer to Figure 2k).(`c-c′′) Attenuation of Importin β expression in the third-instar larval brains led to cortical Pros in 140 NBs (arrowhead) from 14 lobes observed. (d-d′′) When importin β was knocked down in the rangap mutant background (arrowhead), the barely visible anti-Pros staining was detected in 287 NBs from 11 second-instar larval brain lobes. This Pros distribution was reminiscent of the pattern seen in the wt NBs (a-a′′). (e-g′′) Attenuated expression of three NPC proteins, Nup98-96, Nup160, or Nup107, led to nuclear Pros in one to two NBs per lobe (arrow) from 10, 7, and 9 third-instar larval brain lobes, respectively. In wt thirdinstar larval brain lobes, nuclear Pros is never observed in NBs. Dotted circles outline the NBs. Scale bar = 10 µm retention of nuclear Pros in the mutant NBs. This conclusion is supported by the genetic evidence that the removal of one copy of Rcc1 in the rangap [EP1173] background rescues this phenotype.

| An intrinsic mechanism maintains Pros shuttling across nuclear envelope
Our Importin experimental data are consistent with our proposal that an intrinsic mechanism maintains Pros shuttling across the nuclear envelope in larval central brain NBs.  (Freibaum et al., 2015). These three proteins function to facilitate protein export. Attenuation of any one of these proteins in the larval brain NBs recapitulates the rangap mutant phenotype. This indicates that impaired Pros export out of the nucleus is able to cause retention of Pros in the nucleus.
It is interesting to note that the transcription factor Ase is present in the nucleus of the NBs, even when Importin β expression is knocked down (Figure 6c,d). This may suggest that different transcription factors use different mechanisms to enter the nucleus. The intrinsic mechanism maintaining import/export of Pros in the NBs seems to be protein-specific. It is not clear whether other proteins containing both NLS and NES also use this mechanism to shuttle across the nuclear envelope.

| Potential end target of the temporal specification mechanism
The intrinsic Pros shuttling mechanism could serve as an effective end target of the temporal series of stereotypical transcription factors that controls the lifespan of the NBs in the central brain. It is conceivable that in the pupal brain, the last transcription factor of the temporal series downregulates RanGAP cytoplasmic levels, which leads to impaired export of Pros. The retention of nuclear Pros then triggers the signal for NB cell cycle exit and termination of NB proliferation potential. A similar mechanism has been observed in the mammalian neurogenesis when, at the very beginning of neuronal terminal differentiation, Ran-GAP levels were drastically decreased in mammalian primary cortical progenitor cells. This attenuation of RanGAP levels irreversibly triggered the cell cycle exit of progenitor cells (Fujiwara et al., 2016).
In summary, we have identified an intrinsic mechanism that maintains Pros shuttling across the nuclear envelope in the larval brain NBs ( Figure 7). This mechanism includes functions of RanGAP and other components of the nucleocytoplasmic transport system (Figure 7a). In the absence of RanGAP, only the export of Pros is compromised in the larval brain NBs (Figure 7b). Under this circumstance, the early retention of Pros in nuclei occurs. This nuclear Pros then leads to a premature NB cell cycle exit. This intrinsic mechanism could be an end target of the temporal series of transcription factors that controls NBs lifespan.

| Drosophila stocks and genetics
The fly stocks and crosses were maintained at 25°C. To generate the pUAST-rangap-CTAP construct, the full-length

| Mutant generation
The rangap 327 mutant allele was generated by imprecise mobilization of a P-element insertion P{EP}rangap [EP1173] using the standard procedure. Sequence analysis revealed the production of a point mutation (Ch2L:19441351 G→T), and this caused an early termination codon. The rangap 9 mutant allele was generated by CRISPR/Cas9 method (Yu et al., 2013). The guide-RNA target sequence was "GGGCGCCAAACTGACAGTCC" and deletion removed five bases (Ch2L:19450932-19450936; GACAG) of the coding region near the N-terminal and generated an early termination codon.

| Statistical analysis
All analyses described below were carried out using the open source softwares GraphPad Prism and ImageJ.
F I G U R E 7 Diagram depicting RanGAP function in the maintenance of Pros shuttling across the nuclear envelope in the larval central brain NBs. This mechanism includes additional components of the nucleocytoplasmic transport system (a). During the process of import, Pros binds to Imp and enters into the nucleus then GTP binds to Imp and releases Pros. During the process of export, Pros binds to GTP and Exportin. In the cytoplasm, GTP will be hydrolyzed into GDP which permits Pros release. It seems RanGTP/RanGDP plays a more important role in Pros export than import. In the absence of RanGAP (b), only the export of Pros is compromised while the import of Pros is not affected in the larval brain NBs. Exp: Exportin; Imp: Importin