Cell Stem Cell. 2011 May 6; 8(5): 580–593.
PMCID: PMC3093620

Genome-Wide Analysis of Self-Renewal in Drosophila Neural Stem Cells by Transgenic RNAi


The balance between stem cell self-renewal and differentiation is precisely controlled to ensure tissue homeostasis and prevent tumorigenesis. Here we use genome-wide transgenic RNAi to identify 620 genes potentially involved in controlling this balance in Drosophila neuroblasts. We quantify all phenotypes and derive measurements for proliferation, lineage, cell size, and cell shape. We identify a set of transcriptional regulators essential for self-renewal and use hierarchical clustering and integration with interaction data to create functional networks for the control of neuroblast self-renewal and differentiation. Our data identify key roles for the chromatin remodeling Brm complex, the spliceosome, and the TRiC/CCT-complex and show that the alternatively spliced transcription factor Lola and the transcriptional elongation factors Ssrp and Barc control self-renewal in neuroblast lineages. As our data are strongly enriched for genes highly expressed in murine neural stem cells, they are likely to provide valuable insights into mammalian stem cell biology as well.


► Genome-wide RNAi screen finds 620 genes regulating Drosophila neural stem cells ► A set of transcriptional regulators is essential for neural stem cell self-renewal ► Brm complex, spliceosome, and TRiC/CCT-complex regulate neural differentiation ► Alternative splicing and transcriptional elongation are required in neural stem cells


Stem cells play important roles in tissue homeostasis and development. In adult organisms, they ensure continuous replacement of dying or damaged cells, while during development they generate most of the cell types in a developing organ. To fulfill this task, stem cells can maintain an undifferentiated state, but at the same time generate daughter cells that are lineage-restricted and ultimately undergo terminal differentiation. Understanding how the balance between self-renewal and differentiation is controlled within a stem cell lineage is important since defects in the control of this process can result in tissue degeneration or tumorigenesis.

Drosophila neuroblasts (Nbs) are one of the best-understood model systems for stem cell biology (Doe, 2008; Neumüller and Knoblich, 2009). In a series of asymmetric cell divisions (ACDs) that occur during embryonic, larval, and pupal stages of fly development, Nbs give rise to all neurons and glia cells in the adult Drosophila brain. Based on their lineage, two types of Nbs can be distinguished (Bello et al., 2008; Boone and Doe, 2008; Bowman et al., 2008). Type I Nbs divide asymmetrically into a large cell that retains Nb characteristics and a smaller, so-called ganglion mother cell (GMC) that undergoes just one more terminal division to create two differentiating neurons. Type II Nbs also divide asymmetrically, but in this case, the smaller daughter cell continues to divide asymmetrically. It initiates expression of the transcription factor Asense (Bowman et al., 2008) to become an intermediate neural progenitor (INP) that divides asymmetrically into one INP and one GMC, which in turn gives rise to two neurons. In both type I and type II Nbs and in the INPs, the difference between the two daughter cells is established via the asymmetric segregation of the cell-fate determinants Numb, Prospero (Pros), and Brat. In mitosis, these proteins concentrate in a cortical crescent and are inherited exclusively by the smaller daughter cell upon cytokinesis. In this cell, Numb inhibits Notch signaling while Pros represses the transcription of cell cycle genes and induces genes required for neuronal differentiation (Choksi et al., 2006). Brat can act as a translational repressor in other tissues (Sonoda and Wharton, 2001), and Brat-related proteins have been shown to regulate the transcription factor Myc and microRNAs (Neumüller et al., 2008; Schwamborn et al., 2009). The exact molecular function of Brat in Nbs, however, is currently unclear.

In the absence of the cell-fate determinants numb, pros, or brat, the balance between self-renewal and differentiation is perturbed, resulting in the formation of a brain tumor (Bello et al., 2006; Betschinger et al., 2006; Lee et al., 2006). In brat mutants, the small daughter cells of the type II Nbs fail to turn on INP markers and continue to express Nb characteristics (Bowman et al., 2008). The misspecified Nbs continue to divide asymmetrically, but no longer obey the signals that terminate Nb proliferation at the end of the larval period. When transplanted into adult host flies, they continue to proliferate indefinitely, become aneuploid, and start to metastasize (Caussinus and Gonzalez, 2005). Similar defects are observed upon inactivation of numb and pros in Nbs (Bello et al., 2006; Choksi et al., 2006; Bowman et al., 2008), although in these cases, type I Nbs are affected as well. Tumors are also formed in mutants where the asymmetric localizations of Numb, Pros, and Brat are perturbed. Experiments using a mouse breast cancer model have indicated a similar causal relationship between asymmetric stem cell division and tumorigenesis (Cicalese et al., 2009) in vertebrates. Consistently, human homologs of Numb (Pece et al., 2004), Pros (Petrova et al., 2008), and Brat (Boulay et al., 2009) have all been connected to cancer, indicating that the results obtained in Drosophila are relevant for understanding mammalian tumorigenesis.

The genetic networks downstream of Brat, Numb, and Pros that restrict self-renewal to only one daughter cell are currently poorly understood. Microarray experiments (Loop et al., 2004) and transcriptional target identification have identified lists of potential maintenance and differentiation regulators, but the functional relevance of these is largely unknown (Choksi et al., 2006; Southall and Brand, 2009). In mammalian stem cells, genome-wide RNAi studies have been performed in cell culture (Ding et al., 2009; Hu et al., 2009). Ideally, however, stem cells should be studied in their natural environment where the interactions with the surrounding niche and the tissue-specific characteristics of individual lineages are maintained. In Drosophila, this has recently become possible through the establishment of a transgenic RNAi library that can be expressed in a tissue-specific manner (Dietzl et al., 2007).

Here we use transgenic RNAi to analyze self-renewal in Drosophila Nbs on a genome-wide level. We identify 620 genes causing visible defects in Nb lineages and precisely quantify the resulting loss-of-function phenotypes. By integrating our functional data with publicly available gene- and protein-interaction data, we determine networks of functionally related genes that control cytokinesis, cell growth, and differentiation in the Drosophila brain. As our dataset is enriched for genes highly expressed in mammalian stem cells, it is likely to provide a valuable resource for mammalian stem cell biology as well.

Results and Discussion

Screen Design

To analyze self-renewal in Nbs, we combined insc-Gal4 (Betschinger et al., 2006) (expressed in type I and type II Nbs and INPs), with UAS-CD8::GFP (outlining cell membranes) to allow identification of most cells in each Nb lineage without the need for antibody staining. In a pilot screen, we could replicate the published loss-of-function phenotypes of brat, pros, and numb (Figure 1A). Since all lines causing visible phenotypes in the pilot screen are lethal when crossed to UAS-Dicer-2; insc-Gal4 (data not shown), we chose to screen for lethality first and analyze only the brains of lethal lines by confocal microscopy (Figure 1B). In total, we screened 17,362 RNAi lines from the VDRC GD library corresponding to 12,314 individual genes, approximately 89% of the annotated protein coding genes in Drosophila. 24.1% of the lines caused lethality (corresponding to 3412 or 27.7% of the analyzed genes) (Figure 1C, note that only genes that fit our quality criteria [see below] are represented in this panel). Among the 4182 lethal lines, analysis of CD8::GFP expression identified 832 lines (687 genes) that cause abnormalities in Nb, GMC, or INP number, size, or shape or cause the formation of intracellular CD8::GFP accumulations (note that only 620 of these fit our quality criteria and were included in the analysis; see below). We measured the average diameter and number of Nbs and their early daughter cells as well as the number and size of GFP aggregates within these cell types. From these measurements, we derived numbers that express phenotypic strength in 13 distinct categories on a scale from 0 to 10 (see Experimental Procedures for details). These quantitative phenotypic data are provided in Table S1 as well as an online database at http://neuroblasts.imba.oeaw.ac.at. Thus, our screen has identified and quantified putative loss-of-function Nb phenotypes for 4.5% of all protein coding genes in the Drosophila genome.

Figure 1
Transgenic RNAi Screen

Quality Control

To evaluate the quality of our dataset, we made use of a second RNAi library (KK library) generated by site-specific integration of UAS-RNAi constructs. In this library, 314 lines were available for the 687 genes that caused visible brain phenotypes, and 235 of these (75%) are also lethal when crossed to insc-Gal4 (79 nonlethal lines). We randomly selected 135 lines from the lethal set for phenotypic analysis (Figure 2A; Figure S1). We compared scores in the “GMC_less” category, the most frequent phenotype identified in the screen, and found that 121 KK lines display a phenotype identical to the corresponding GD line. For 14 lines (10.4%), only one of the two lines targeting the same gene had a phenotype in that category. To improve the overall reliability of the primary screening results, we used the S19 score that expresses the specificity of each RNAi construct on a scale from 0 (no specificity) to 1 (completely specific) (Dietzl et al., 2007). Of 79 lethal GD lines where the corresponding KK line was viable, 25.3% had an S19 score of less than 0.85. However, only 6.4% of the 235 GD lines in which the lethality could be verified by a KK line were below this score. Thus, using the S19 score to predict quality of RNAi lines significantly improves specificity, and we therefore discarded lines with a score below or equal to 0.85 from our analysis. In 226 of the 288 lines that fit this criterion, lethality could be verified by a KK line, suggesting that the reproducibility of our final dataset is 78.5% (Figure 2B) and therefore higher than in previous transgenic RNAi screens (Mummery-Widmer et al., 2009).

Figure 2
Quality Control and Transcriptional Regulators

To test the expression pattern of the identified genes, we used expression data from Flyatlas (Chintapalli et al., 2007). The set of identified genes is significantly enriched for genes expressed in the larval CNS (Figure 2C). Surprisingly, the set is also enriched for genes expressed in ovaries, whereas most other tissues are underrepresented. This is probably because the expression of insc-Gal4 in other tissues like the gut or salivary glands causes early lethality for genes generally required in all tissues. Indeed, genes upregulated in a wide variety of tissues other than the larval brain are enriched among the “early lethal” genes for which lethality before the larval third instar prevented the analysis of brain phenotypes (Figure 2C and data not shown).

Close mammalian homologs were identified for 88.23% of the genes causing Drosophila phenotypes (Table S1). To test the relevance of our dataset for mammalian stem cell biology, we compared the identified genes to previously assembled mammalian datasets that are based on expression and predicted function by searching the Molecular Signature Database (MSigDB) (Subramanian et al., 2005). We found that our gene set is significantly enriched for genes highly expressed in mouse embryonic stem cells (ESCs) and neural stem cells (Figure 2D) (Ramalho-Santos et al., 2002). Subsets of genes expressed in mammalian stem cells that cause phenotypes in Drosophila Nbs may represent valuable starting points for functional analyses and are provided in Table S2.

Transcriptional Network for Self-Renewal

To isolate regulators of self-renewal among the genes identified in the screen, we used three strategies. First, we identified all putative transcription factors and chromatin regulators that cause either loss or underproliferation of Nbs and are candidate components of a transcriptional network for self-renewal. Second, we defined genes that cause Nb phenotypes and were previously shown to interact with known regulators of ACD, the key process controlling Nb self-renewal. Third, we used hierarchical clustering of our quantitative phenotypic data to identify groups of genes causing similar phenotypes.

Thirty-three experimentally verified or computationally predicted transcription factors and chromatin regulators cause underproliferation phenotypes (Adryan and Teichmann, 2006; Pfreundt et al., 2010). Among those are 13 known transcription factors, 12 genes whose domain composition implies a role in transcriptional regulation, and eight chromatin regulators (Figure 2E, see Supplemental Information for explanation of network construction). These include the Polycomb group genes Polycomb-like (Pcl) and multiple sex combs (mxc), but also Su(z)12, which has previously been implicated in Nb self-renewal (Bello et al., 2007). In addition, we identified Su(var)2-10, the Drosophila homolog of Pias1; the Hp1 homolog Su(var)205; the Hdac1 homolog Rpd3; and domino, a gene that is also required for stem cell maintenance in Drosophila ovaries (Xi and Xie, 2005). We also found the specific transcription factors spalt related (salr) (Mollereau et al., 2001), lethal of scute (l[1]sc) (Martin-Bermudo et al., 1991), retinal homoebox (Rx) (Davis et al., 2003), and longitudinals lacking (lola) (Giniger et al., 1994), which have previously been implicated in nervous system development. Besides these known regulators, several putative transcription factors have not previously been characterized. CG9895 is homologous to the mammalian Kruppel-like factors Klf1, 2, 4, and 8 (Sur, 2009). CG9571 has homology to mammalian Foxg1 (Copley, 2005), which is expressed in forebrain progenitors and involved in the regulation of self-renewal (Shen et al., 2006; Fasano et al., 2009).

Regulatory Network for Asymmetric Cell Division

In Nbs, the ACD machinery ensures correct segregation of cell-fate determinants into the differentiating daughter cell. We used a set of 53 genes previously implicated in ACD or spindle orientation (Figure 3A) to query a database containing two-hybrid, biochemical, interolog, text-mining data, and genetic interactions between Drosophila genes (see Experimental Procedures). The resulting interaction network was reduced by only allowing connections with genes that had resulted in a phenotype in our screen (Figure 3A). To predict protein complexes and genetic pathways implicated in ACD we used clustering algorithms (MCODE, MCL, see Experimental Procedures). Three of the six protein complexes predicted in this way control cell cycle processes like kinetochore/mitotic spindle assembly, mitotic protein degradation (proteasome and anaphase promoting complex [APC]), and DNA replication. In addition, our analysis identified the RNA splicing machinery, the TRiC/CCT complex (TCP-1 ring complex or chaperonin containing T-complex 1), and a chromatin remodeling complex, which are discussed further below.

Figure 3
Regulatory Network for Asymmetric Cell Division

Alternative Splicing

One of the complexes identified contains 35 genes that regulate various aspects of RNA metabolism and transcription (Figure 3A, “splicing”). Twenty-seven of these have previously been shown to regulate RNA splicing. Interestingly, eight of these genes were previously identified in an RNAi screen for alternative splicing (Park et al., 2004). In a cellular assay, B52, Hrb87F, CG6841, Pea, and U2af50 are needed for alternative splicing of Dscam while B52, Crn, and snRNP70K are involved in alternative splicing of dAdar, and CG10418 controls splicing of paralytic. In our screen, most of these genes caused underproliferation and loss of Nbs, and this might indicate a role for alternative splicing in Nb proliferation.

To identify potential targets of alternative splicing in Nbs, we searched for genes where individual RNAi lines resulted in divergent phenotypes. Interestingly, the gene lola is targeted by three different RNAi lines, one of which causes overproliferation in type II Nb lineages (GD12573) while the other two lines (GD41415, GD25333) cause underproliferation in both type I and type II lineages (Figure 3B; Figure S2; online DB). This difference is not due to off-targets, as the overproliferation phenotype could be confirmed by another nonoverlapping RNAi line from the KK library (data not shown, see Experimental Procedures for details). Lola is a transcription factor involved in axon guidance during nervous system development (Goeke et al., 2003). The gene encodes at least 20 different isoforms that share a common N terminal BTB-domain, but differ in their C terminal Zn-finger region (Goeke et al., 2003). Both RNAi lines that cause overproliferation target the common N terminal region and are predicted to affect all of the 20 isoforms. The two lines causing Nb loss, however, specifically target the lola splicing isoforms B and N (Figure 3C). Thus, different splicing isoforms of lola seem to promote or inhibit Nb self-renewal.

To test whether lola isoforms are differentially expressed in Nbs versus neurons, we performed a microarray analysis of wild-type (WT) brains and brat-RNAi brains, which mostly consist of overproliferating type II Nbs. Interestingly, while isoforms D and H are downregulated in brat mutant brains, the expression of isoforms B, C, and S is significantly increased (see arrowheads in Figure 3C). Thus, alternative splicing of lola seems to be important for controlling proliferation in the developing Drosophila brain. Lola has been shown to not only antagonize the Notch pathway (Zheng and Carthew, 2008), but also to enhance Delta-induced tumor formation (Ferres-Marco et al., 2006). Notch-Delta signaling is important for balancing proliferation and differentiation in Drosophila Nbs. Thus, the different phenotypes of the various lola-RNAi lines could be explained if different splicing isoforms would have distinct effects on Notch-Delta signaling.

TRiC/CCT Complex

The TRiC/CCT complex is the major ATP-dependent chaperonin in the eukaryotic cytoplasm (Hartl and Hayer-Hartl, 2009). It contains eight proteins that copurify with the protein phosphatase PP4 (Gingras et al., 2005). In our screen, four of these (Cct5, Tcp-1eta, CG7033, and CG5525) result in underproliferation phenotypes, while the others are early lethal (Cct1γ) or cause lethality without an obvious Nb phenotype (Tcp-1like, CG8258, and Tcp-1ζ). The catalytic subunit of PP4 itself (Pp4-19C) also causes Nb underproliferation, although a subset of Nbs displays an overproliferation phenotype (Figures 3A and 3D). The overproliferation phenotype is not confined to type II lineages, as it is also seen with ase-Gal4, which drives RNAi in type I lineages only (Figure 3D). In addition, one of the two regulatory subunits (PPR2/PPp4R2r) also results in Nb underproliferation.

In Drosophila, a previous study has identified PP4 as a regulator of the asymmetric localization of Mira and its cargo proteins Pros and Brat in dividing Nbs (Sousa-Nunes et al., 2009). Whether PP4 acts in the TRiC/CCT complex to perform its role in Mira localization is currently unclear. PP4 has also been described to localize to centrosomes and act in centrosome maturation and spindle formation (Helps et al., 1998), and this might provide an alternative explanation for the Nb phenotypes.

Hierarchical Clustering of Phenotypes

Our analysis has assigned to each gene a string of numbers describing a putative loss-of-function phenotype in Nbs. This “phenotypic barcode” allows us to computationally analyze and group genes based on the similarity of their Nb phenotypes. We used a hierarchical clustering algorithm on a reduced set of phenotypic categories describing Nb and GMC size, an increase or decrease in Nb number, and an increase or decrease in the number of daughter cells generated by each Nb. In addition, we included the presence of GFP-aggregates within a cell, a phenotype that might indicate cell death or any other disturbance of internal cellular membranes. The analysis identified several clusters of similar phenotype combinations and allowed for a clearly arranged visualization of our screening results (Figure 4A).

Figure 4
Hierarchical Clustering of Phenotypes

To define groups of genes that might perform a similar function, we used specific combinatorial criteria. Genes required for restricting self-renewal and potential tumor suppressors are expected to cause an increase in Nb or total cell number. We therefore generated an “overproliferation” group of 29 genes where “Nb_more ≥ 2” or “GMC_more ≥ 2.” Analysis of genes promoting self-renewal is more complex. Such genes should cause a reduction of total cell number within Nb lineages. However, underproliferation is the most common phenotype in our analysis (538 of 620 genes) and can also arise from a number of unspecific biological defects, such as cell death or cell-cycle block. We therefore analyzed all genes where “Nb_less ≥ 2” and defined three mutually exclusive groups depending on whether Nb loss is associated with increased, decreased, or unchanged Nb size (Figure 4A). The Nb_loss_small, Nb_loss_large, and Nb_loss_normal groups were defined as “Nb_less ≥ 2 AND Nb_small ≥ 2,” “Nb_less ≥ 2 AND Nb_large ≥ 2,” and “Nb_less ≥ 2 AND Nb_small ≤ 1 AND Nb_large ≤ 1,” respectively. Genes in the Nb_loss_large and the Nb_loss_normal groups are almost always associated with the accumulation of GFP aggregates, suggesting that they act in basic cellular processes (Figure 4A). Such aggregates are almost never observed in the Nb_loss_small group, indicating that those genes might play a role in cell growth, but are not essential for cell survival. In addition, the clustering algorithm identified a group of genes causing an extreme increase in Nb size without loss, and we combined those genes in the Nb_huge group (Figure 4A).

A GO term analysis showed that the various phenotypic groups are highly enriched for genes regulating specific cellular functions (Figure 4B; Figure S3). This was further confirmed by analyzing the Nb_huge group in more detail. Interaction network analysis of this group identified a complex containing the known cytokinesis regulators Incenp, zipper, aurora B, and Deterin (Figures S4A and S4B), as well as the tubulin subunits betaTub60D and alphaTub67C and the kinase Pka-C2. Indeed, actin staining of these RNAi lines reveals massive enlargement of Nbs and increased cellular DNA content characteristic of a cytokinesis defect (Figure S4C).

Cell Growth and Nb Self-Renewal

Genes required for Nb self-renewal should cause a loss of Nbs. Together with genes required for Nb survival, they should be in the groups Nb_loss_small, Nb_loss_large, and Nb_loss_normal. As Caspase staining revealed that the number of apoptotic cells was increased in the Nb_loss_normal and Nb_loss_large groups, but not in the Nb_loss_small group (Figure 2A, data not shown), we focused on the Nb_loss_small group for further analysis. Network analysis of this group identified a cluster of genes regulating ribosome biogenesis and protein biogenesis (Figure 5A). Surprisingly, however, most other ribosomal subunits cause early lethality (data not shown). Thus, most ribosomal subunits are required in all Drosophila cells, while some have a more specific function. Among those are RpL10Aa and RpS10a (Figure 5B), but also RpS5b, a subunit that causes an Nb underproliferation phenotype below the cutoff threshold. These subunits are duplicated in the Drosophila genome (Marygold et al., 2007) and, in all cases, the alternative isoform (RpL10Ab, RpS10b, and RpS5a, respectively) causes early lethality. A previous gene expression analysis has revealed that RpS10a and RpS5b are significantly overrepresented in Drosophila ovarian stem cells (Kai et al., 2005). Our results indicate that certain ribosomal subunits are duplicated in the fly genome with one isoform being required in all cells and another isoform acting more specifically in stem cell lineages.

Figure 5
Cell Growth and Nb Self-Renewal

Functional diversification of duplicated ribosomal subunits has been demonstrated before in yeast. In S. cerevisiae, several ribosomal subunits exist as two isoforms that serve distinct functions and cause different phenotypes when deleted (Komili et al., 2007). In vertebrates, duplication of ribosomal protein genes is rare, although multiple splice variants exist that can be expressed in a tissue-specific manner (Nakao et al., 2004). In humans (but not mice), the Rps4 gene is duplicated. The two isoforms RPS4X and RPS4Y are located on the X and Y chromosomes, respectively (Fisher et al., 1990), and RPS4Y has been implicated in Turner syndrome (Fisher et al., 1990). Taken together, these data suggest that cell growth and ribosome biogenesis are rate-limiting for stem cell self-renewal in the Drosophila brain.

Differentiation and Tumor Suppression

Inhibitors of self-renewal and genes required for differentiation are expected to be in the overproliferation group. However, apparent overproliferation phenotypes could also be generated by longer GFP expression in Nb lineages or by an increase in proliferation. Since the CD8::GFP reporter did not allow us to distinguish those, we stained all 29 lines in the overproliferation group for the Nb marker Mira and for Pros, a marker for GMCs and neurons. This analysis identified 18 genes where RNAi results in the formation of extra Nbs (Figure 6; Figure S5).

Figure 6
Differentiation and Tumor Suppression

Interaction network analysis of the overproliferation genes revealed two protein complexes (Figure 6A). The first complex contains the segregating determinants Numb, Pros, Mira, and Brat and the phosphatase PP4 (see above, Sousa-Nunes et al., 2009). Numb connects to α-Adaptin and AP-2σ, two components of the AP2 complex that has been shown to bind to Numb (Berdnik et al., 2002). Knockdown of α-Adaptin or AP-2σ results in the formation of ectopic Nbs that coexpress Mira and Dpn, but are negative for the neuronal markers Pros and Elav (Figure 6B; Figure S5A; data not shown). A similar overproliferation phenotype is evident in clones of α-Adaptin mutants, confirming the specificity of the RNAi line (Figure S5D). To address where α-Adaptin and AP-2σ are required, we used ase-Gal4, which is specific to type I lineages, and wor-Gal4, ase-Gal80, which is specific to type II lineages (Figure S5E, see Supplemental Information). Using those lines, we found that the AP-2 complex is required in both type I and type II Nbs (Figures S5B and S5C). Thus, Numb might exert its tumor suppressor function by regulating endocytic trafficking via the AP-2 complex.

Chromatin Remodeling

The second complex contains the genes brahma (brm), moira (mor), and osa, which are part of the chromatin remodeling Brm complex (Papoulas et al., 1998). RNAi of either brm, mor, or osa results in the generation of extra Mira-positive Nbs at the expense of Pros-positive neurons (Figure 6C). Expressing RNAi with wor-Gal4, ase-Gal80 resulted in an overproliferation phenotype, while ase-Gal4 had no effect, indicating that type II Nbs are more sensitive to the loss of the Brm complex (Figures 6D and 6E).

Components of the mammalian SWI/SNF complexes have been implicated in tumor suppression (Reisman et al., 2009) and in controlling the balance between proliferation and differentiation in mammalian neural stem cells (Yoo and Crabtree, 2009). Consistent with this, mammalian homologs of Brm and Osa are upregulated in mouse neural stem cells (Figure 6A).

Transcriptional Elongation

In the overproliferation network, the Brm complex connects to the genes Ssrp and CG6049 (Figure 6A). RNAi targeting Ssrp or CG6049 results in an expansion of Nbs at the expense of neurons (Figure 6B; Figure 7B). Ssrp is a subunit of the so-called FACT complex that is required for transcriptional elongation on chromatin templates (Belotserkovskaya and Reinberg, 2004). The FACT complex acts by destabilizing nucleosomes and facilitates transcription by allowing PolII to pass. It is connected to the Brm complex because the yeast version of the Brm complex (Swi/Snf) is also implicated in histone disassembly and removal during transcriptional elongation (Schwabish and Struhl, 2007).

Figure 7
Barc Regulates Intermediate Neural Progenitors

CG6049 is the Drosophila homolog of human Tat-SF1 (Zhou and Sharp, 1996). Besides roles in HIV infection and RNA splicing, Tat-SF1 also has a prominent function in transcriptional elongation. It is an activator and binding partner of the Paf1 complex and the transcription elongation factor DSIF (Chen et al., 2009). Together with the negative elongation factor NELF, DSIF can cause stalling of PolII transcription, while Paf1 and Tat-SF1 are cofactors for activating transcriptional elongation. As the FACT complex is one of the most prominent genetic and physical interaction partners of Paf1 (Belotserkovskaya and Reinberg, 2004), all genes in the Brm/CG6049/Ssrp interaction network may regulate Nb self-renewal through a common mechanism.

Barc Regulates Intermediate Neural Progenitors

As CG6049 had not been characterized before, we chose this gene for in-depth analysis. We renamed CG6049 into barricade (barc) to indicate the block in Nb lineage progression we observed upon RNAi. Barc is conserved from yeast to humans. Like its vertebrate homolog Tat-SF1, it contains two RNA recognition modules (RRM), a nuclear localization signal, and a conserved region that contains two motifs that are known to bind to FF domains (Smith et al., 2004) and that we named the Barc-Tat-SF1 (BTS) motif (Figure 7A). To determine the specificity of the barc-RNAi phenotype, we generated an RNAi-resistant barc construct (see Experimental Procedures for details). When expressed together with barc-RNAi, this construct can rescue both lethality and the Nb phenotype (Figures 7B and 7E). In addition, the barc-RNAi phenotype could be confirmed by a second, nonoverlapping RNAi line (Figure S6A). Thus, barc is a regulator of lineage progression in Drosophila Nbs.

While barc-RNAi in type II lineages using wor-Gal4; ase-Gal80 causes overproliferation (Figure 7C), barc-RNAi induced by ase-Gal4 has no overproliferation phenotype (data not shown). The additional CD8::GFP-positive cells in the type II lineages express Cyclin E, indicating active proliferation (Figure 7B), and do not express the neuronal marker Elav (Figure 7E). We observe more cells positive for Mira and Dpn (Figures 7B and 7C; Figure S6B), which are expressed both in Nbs and in INPs. On average, the number of Mira-positive cells is approximately 4-fold increased (Figure 7D). As we only detect one large Ase-negative type II Nb, and the extra cells express the INP marker Ase (Figure 7F), we conclude that barc is required for INPs to generate differentiating neurons. Upon barc-RNAi, the daughter cells retain the INP fate, and this results in the overproliferation phenotype. Although barc-RNAi does not cause a similar overproliferation phenotype in type I lineages, we observe that the diameter of type I Nbs is increased from 15 ± 0.31 μm (SEM, n = 47) to 17.16 ± 0.27 μm (SEM, n = 43) (Figure 7G). This phenotype could either indicate an increase in growth rate or a decrease in cell cycle progression. Thus, barc is required for lineage progression in type II Nb lineages, but might also have a function in mitotic progression of type I Nbs.

To test Barc expression and subcellular localization, we generated a peptide antibody. The antibody detects a single band of approximately 75 kD on a western blot (Figure S6C), which can be blocked by the antigenic peptide (Figure S6C). The anti-Barc immunofluorescence signal is absent after barc-RNAi (Figure S6D) and increases upon Barc overexpression (data not shown). Barc antibody staining revealed that Barc is a nuclear protein that is predominantly expressed in both type I and type II Nbs and to a lesser extent in INPs, GMCs, and differentiated neurons (Figures S6D and S6E). Thus, we have identified a nuclear regulator of type II Nb lineages that allows INPs to generate daughter cells, which undergo terminal neural differentiation.


Our screen has identified a total of 620 genes that are potentially involved in controlling self-renewal in Drosophila neural stem cells. We demonstrate that precise quantification of phenotypic data allows for a computer analysis that can lead to biological insights that are not easily obtained through classic single-gene approaches. Through network analysis, we have identified splicing control as a key regulator of Nb self-renewal. Alternative splicing of lola might be one of the targets of this machinery as different isoforms of this transcription factor are differentially expressed and phenotypically distinct. We also show that duplicated forms of ribosomal subunits are functionally distinct, with one form being more specifically required in Nbs. Finally, we demonstrate that genes involved in transcriptional elongation and chromatin remodeling are important regulators of Nb self-renewal and differentiation. It is known that more than one third of all Drosophila genes are in a poised state where active RNA polymerase is stalled in a promoter proximal position. Release of stalled polymerases might contribute to the rapid activation of differentiation genes during Nb ACD. Transcriptional elongation is important for controlling vertebrate stem cell lineages as well (Bai et al., 2010), but how stalled promoters are released in a cell-type-specific manner is currently unknown. Analysis in the simple Drosophila Nb lineage could shed some light on this important question in stem cell biology.

Experimental Procedures

Genome-Wide RNAi Screen

Males from the Vienna Drosophila RNAi Center (VDRC) carrying an inducible UAS-RNAi construct were crossed to virgins of the driver line containing UAS-Dicer-2; insc-Gal4, UAS-CD8::GFP. Crosses were set up at 25°C and transferred to 29°C after 1 day. In case of homozygous RNAi lines, crosses were flipped after 4 days. Lethality was determined in the first cross by the presence of balancer chromosome flies only. In case of lethality, larvae of the right genotype were identified by insc-Gal4-driven CD8::GFP (brain and salivary gland) using a fluorescence microscope. Six larvae per genotype were dissected in PBS and fixed in 5% PFA in PBS for 20 min. After mounting the specimen in Vectashield, a confocal stack of the brain was recorded using a ZEISS LSM confocal microscope. All phenotypic abnormalities were recorded and stored in a database (http://neuroblasts.imba.oeaw.ac.at).

Phenotypic annotations were performed with the LSM image examiner software. Phenotypes affecting the whole brain were described using numbers ranging from 0 to 10 depending on their strength (0 indicates WT, 1 indicates a possible phenotype that is below a cutoff, and numbers from 2 to 10 represent definitive phenotypes with increasing strength). Phenotypes that are only detectable in a defined region of the brain, whereas other parts are not affected, were quantified using a similar scoring system in a “regional” field, but back-calculated into a global average for computational analysis. These quantified “phenotypes” were then used to create different “categories” for the bioinformatic analysis (see text and Supplemental Information for details).

Fly Strains and Clonal Analysis

Besides the driver line, the following fly strains were used: ase-Gal4 (Zhu et al., 2006) and wor-Gal4, ase-Gal80 always in combination with UAS-Dicer-2; UAS-barc, UAS-barc-resistant, UAS-barc-RNAi-2nd line (see Supplemental Information for details on cloning and constructs), MARCM stocks using elav-Gal4 (C155) (Lee and Luo, 1999), FRT40A, and alpha-Adaptin3 (Gonzalez-Gaitan and Jackle, 1997).

RNAi crosses were set up at 25°C and larvae were raised at 29°C. Brains of wandering third instar larvae were dissected and further processed for immunofluorescence. The KK line (KK101925) targeting lola was early lethal when crossed to insc-Gal4 and was therefore crossed to wor-Gal4, ase-Gal80 that has a more restricted expression area. For MARCM experiments, larvae were heat-shocked for 1 hr at 37°C and dissected 3 or 4 days later as wandering third instar larvae.


Antibodies used in this study are rabbit anti-Mira (1:100, Betschinger et al., 2006), mouse anti-Pros (1:10, Developmental Studies Hybridoma Bank, University of Iowa [DSHB]), guinea pig anti-Ase (1:100, Bhalerao et al., 2005), guinea pig anti-Dpn (1:1000, gift from J. Skeath), rabbit anti-Elav (1:300 [DSHB]), rat anti-Mira (1:100), rabbit anti-Caspase (1:200, Cell Signaling Technology), and mouse anti-PhosphoH3 (1:1000, Cell Signaling Technology). Barc-specific antisera were generated in rabbits against the C-terminal peptide: MKEEDVDSPENQLLPGDATP. Immunohistochemistry experiments were performed as previously described (Betschinger et al., 2006).

Gene Expression and Bioinformatics Analyses

Total RNA was isolated from third instar larval brains of either wild-type or brat-RNAi (GD31333 and KK105054) crossed to UAS-Dicer-2; insc-Gal4, UAS-CD8::GFP/CyO. The experiments were done in triplicates. For details on sample preparation, data processing—which was done at the Microarray DNA Facility of the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden—and data analysis, see Supplemental Information. The data is deposited at the ArrayExpress Archive (http://www.ebi.ac.uk/arrayexpress/) and has the ArrayExpress accession: E-MEXP-3112. Details on bioinformatics tools used for GO term, network, and cluster analyses are in Supplemental Information.


We wish to thank B. Dickson, K. Keleman, J. Skeath, H. Richardson, G. Dietzl, S. Bhalerao, the Vienna Drosophila RNAi Center (VDRC), the Developmental Studies Hybridoma Bank (DSHB), and the Bloomington Drosophila Stock Center for constructs, antibodies, and flystocks; E. Kleiner and L. Kirschner for technical assistance and C. Jüschke for help with data analysis; N. Corsini, E. Eroglu, and M. Lancaster for comments on the manuscript; M. Madalinski for peptide synthesis and affinity purification; and members of the Knoblich lab for discussions. Work in J.A.K.'s lab is supported by the Austrian Academy of Sciences, Wiener Wissenschafts-, Forschungs- und Technologiefonds (WWTF), the EU seventh framework program EuroSyStem, and the FWF.

Accession Numbers

The accession number for the new microarray data reported in this paper is E-MEXP-3112.

Supplemental Information

Document S1. Six figures and Supplemental Experimental Procedures:
Table S1:

List of genes with their respective neuroblast RNAi phenotypes and mammalian orthologs (s19filtered > 0.85).

Table S2:

List of genes with their respective neuroblast RNAi phenotypes and mammalian orthologs (s19filtered > 0.85) AND their expression in mammalian stem cells (MSigDBv2.5 data).


Adryan B., Teichmann S.A. FlyTF: a systematic review of site-specific transcription factors in the fruit fly Drosophila melanogaster. Bioinformatics. 2006;22:1532–1533. [PubMed]
Bai X., Kim J., Yang Z., Jurynec M.J., Akie T.E., Lee J., LeBlanc J., Sessa A., Jiang H., DiBiase A. TIF1gamma controls erythroid cell fate by regulating transcription elongation. Cell. 2010;142:133–143. [PMC free article] [PubMed]
Bello B., Reichert H., Hirth F. The brain tumor gene negatively regulates neural progenitor cell proliferation in the larval central brain of Drosophila. Development. 2006;133:2639–2648. [PubMed]
Bello B., Holbro N., Reichert H. Polycomb group genes are required for neural stem cell survival in postembryonic neurogenesis of Drosophila. Development. 2007;134:1091–1099. [PubMed]
Bello B.C., Izergina N., Caussinus E., Reichert H. Amplification of neural stem cell proliferation by intermediate progenitor cells in Drosophila brain development. Neural Develop. 2008;3:5. [PMC free article] [PubMed]
Belotserkovskaya R., Reinberg D. Facts about FACT and transcript elongation through chromatin. Curr. Opin. Genet. Dev. 2004;14:139–146. [PubMed]
Berdnik D., Torok T., Gonzalez-Gaitan M., Knoblich J.A. The endocytic protein alpha-Adaptin is required for numb-mediated asymmetric cell division in Drosophila. Dev. Cell. 2002;3:221–231. [PubMed]
Betschinger J., Mechtler K., Knoblich J.A. Asymmetric segregation of the tumor suppressor brat regulates self-renewal in Drosophila neural stem cells. Cell. 2006;124:1241–1253. [PubMed]
Bhalerao S., Berdnik D., Torok T., Knoblich J.A. Localization-dependent and -independent roles of numb contribute to cell-fate specification in Drosophila. Curr. Biol. 2005;15:1583–1590. [PubMed]
Boone J.Q., Doe C.Q. Identification of Drosophila type II neuroblast lineages containing transit amplifying ganglion mother cells. Dev. Neurobiol. 2008;68:1185–1195. [PMC free article] [PubMed]
Boulay J.L., Stiefel U., Taylor E., Dolder B., Merlo A., Hirth F. Loss of heterozygosity of TRIM3 in malignant gliomas. BMC Cancer. 2009;9:71. [PMC free article] [PubMed]
Bowman S.K., Rolland V., Betschinger J., Kinsey K.A., Emery G., Knoblich J.A. The tumor suppressors Brat and Numb regulate transit-amplifying neuroblast lineages in Drosophila. Dev. Cell. 2008;14:535–546. [PMC free article] [PubMed]
Caussinus E., Gonzalez C. Induction of tumor growth by altered stem-cell asymmetric division in Drosophila melanogaster. Nat. Genet. 2005;37:1125–1129. [PubMed]
Chen Y., Yamaguchi Y., Tsugeno Y., Yamamoto J., Yamada T., Nakamura M., Hisatake K., Handa H. DSIF, the Paf1 complex, and Tat-SF1 have nonredundant, cooperative roles in RNA polymerase II elongation. Genes Dev. 2009;23:2765–2777. [PMC free article] [PubMed]
Chintapalli V.R., Wang J., Dow J.A. Using FlyAtlas to identify better Drosophila melanogaster models of human disease. Nat. Genet. 2007;39:715–720. [PubMed]
Choksi S.P., Southall T.D., Bossing T., Edoff K., de Wit E., Fischer B.E., van Steensel B., Micklem G., Brand A.H. Prospero Acts as a Binary Switch between Self-Renewal and Differentiation in Drosophila Neural Stem Cells. Dev. Cell. 2006;11:775–789. [PubMed]
Cicalese A., Bonizzi G., Pasi C.E., Faretta M., Ronzoni S., Giulini B., Brisken C., Minucci S., Di Fiore P.P., Pelicci P.G. The tumor suppressor p53 regulates polarity of self-renewing divisions in mammary stem cells. Cell. 2009;138:1083–1095. [PubMed]
Copley R.R. The EH1 motif in metazoan transcription factors. BMC Genomics. 2005;6:169. [PMC free article] [PubMed]
Davis R.J., Tavsanli B.C., Dittrich C., Walldorf U., Mardon G. Drosophila retinal homeobox (drx) is not required for establishment of the visual system, but is required for brain and clypeus development. Dev. Biol. 2003;259:272–287. [PubMed]
Dietzl G., Chen D., Schnorrer F., Su K.C., Barinova Y., Fellner M., Gasser B., Kinsey K., Oppel S., Scheiblauer S. A genome-wide transgenic RNAi library for conditional gene inactivation in Drosophila. Nature. 2007;448:151–156. [PubMed]
Ding L., Paszkowski-Rogacz M., Nitzsche A., Slabicki M.M., Heninger A.K., de Vries I., Kittler R., Junqueira M., Shevchenko A., Schulz H. A genome-scale RNAi screen for Oct4 modulators defines a role of the Paf1 complex for embryonic stem cell identity. Cell Stem Cell. 2009;4:403–415. [PubMed]
Doe C.Q. Neural stem cells: balancing self-renewal with differentiation. Development. 2008;135:1575–1587. [PubMed]
Fasano C.A., Phoenix T.N., Kokovay E., Lowry N., Elkabetz Y., Dimos J.T., Lemischka I.R., Studer L., Temple S. Bmi-1 cooperates with Foxg1 to maintain neural stem cell self-renewal in the forebrain. Genes Dev. 2009;23:561–574. [PMC free article] [PubMed]
Ferres-Marco D., Gutierrez-Garcia I., Vallejo D.M., Bolivar J., Gutierrez-Avino F.J., Dominguez M. Epigenetic silencers and Notch collaborate to promote malignant tumours by Rb silencing. Nature. 2006;439:430–436. [PubMed]
Fisher E.M., Beer-Romero P., Brown L.G., Ridley A., McNeil J.A., Lawrence J.B., Willard H.F., Bieber F.R., Page D.C. Homologous ribosomal protein genes on the human X and Y chromosomes: escape from X inactivation and possible implications for Turner syndrome. Cell. 1990;63:1205–1218. [PubMed]
Gingras A.C., Caballero M., Zarske M., Sanchez A., Hazbun T.R., Fields S., Sonenberg N., Hafen E., Raught B., Aebersold R. A novel, evolutionarily conserved protein phosphatase complex involved in cisplatin sensitivity. Mol. Cell. Proteomics. 2005;4:1725–1740. [PubMed]
Giniger E., Tietje K., Jan L.Y., Jan Y.N. lola encodes a putative transcription factor required for axon growth and guidance in Drosophila. Development. 1994;120:1385–1398. [PubMed]
Goeke S., Greene E.A., Grant P.K., Gates M.A., Crowner D., Aigaki T., Giniger E. Alternative splicing of lola generates 19 transcription factors controlling axon guidance in Drosophila. Nat. Neurosci. 2003;6:917–924. [PubMed]
Gonzalez-Gaitan M., Jackle H. Role of Drosophila alpha-adaptin in presynaptic vesicle recycling. Cell. 1997;88:767–776. [PubMed]
Hartl F.U., Hayer-Hartl M. Converging concepts of protein folding in vitro and in vivo. Nat. Struct. Mol. Biol. 2009;16:574–581. [PubMed]
Helps N.R., Brewis N.D., Lineruth K., Davis T., Kaiser K., Cohen P.T. Protein phosphatase 4 is an essential enzyme required for organisation of microtubules at centrosomes in Drosophila embryos. J. Cell Sci. 1998;111:1331–1340. [PubMed]
Hu G., Kim J., Xu Q., Leng Y., Orkin S.H., Elledge S.J. A genome-wide RNAi screen identifies a new transcriptional module required for self-renewal. Genes Dev. 2009;23:837–848. [PMC free article] [PubMed]
Kai T., Williams D., Spradling A.C. The expression profile of purified Drosophila germline stem cells. Dev. Biol. 2005;283:486–502. [PubMed]
Komili S., Farny N.G., Roth F.P., Silver P.A. Functional specificity among ribosomal proteins regulates gene expression. Cell. 2007;131:557–571. [PMC free article] [PubMed]
Lee T., Luo L. Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron. 1999;22:451–461. [PubMed]
Lee C.Y., Wilkinson B.D., Siegrist S.E., Wharton R.P., Doe C.Q. Brat Is a Miranda Cargo Protein that Promotes Neuronal Differentiation and Inhibits Neuroblast Self-Renewal. Dev. Cell. 2006;10:441–449. [PubMed]
Loop T., Leemans R., Stiefel U., Hermida L., Egger B., Xie F., Primig M., Certa U., Fischbach K.F., Reichert H., Hirth F. Transcriptional signature of an adult brain tumor in Drosophila. BMC Genomics. 2004;5:24. [PMC free article] [PubMed]
Martin-Bermudo M.D., Martinez C., Rodriguez A., Jimenez F. Distribution and function of the lethal of scute gene product during early neurogenesis in Drosophila. Development. 1991;113:445–454. [PubMed]
Marygold S.J., Roote J., Reuter G., Lambertsson A., Ashburner M., Millburn G.H., Harrison P.M., Yu Z., Kenmochi N., Kaufman T.C. The ribosomal protein genes and Minute loci of Drosophila melanogaster. Genome Biol. 2007;8:R216. [PMC free article] [PubMed]
Mollereau B., Dominguez M., Webel R., Colley N.J., Keung B., de Celis J.F., Desplan C. Two-step process for photoreceptor formation in Drosophila. Nature. 2001;412:911–913. [PubMed]
Mummery-Widmer J.L., Yamazaki M., Stoeger T., Novatchkova M., Bhalerao S., Chen D., Dietzl G., Dickson B.J., Knoblich J.A. Genome-wide analysis of Notch signalling in Drosophila by transgenic RNAi. Nature. 2009;458:987–992. [PMC free article] [PubMed]
Nakao A., Yoshihama M., Kenmochi N. RPG: the Ribosomal Protein Gene database. Nucleic Acids Res. 2004;32:D168–D170. [PMC free article] [PubMed]
Neumüller R.A., Knoblich J.A. Dividing cellular asymmetry: asymmetric cell division and its implications for stem cells and cancer. Genes Dev. 2009;23:2675–2699. [PMC free article] [PubMed]
Neumüller R.A., Betschinger J., Fischer A., Bushati N., Poernbacher I., Mechtler K., Cohen S.M., Knoblich J.A. Mei-P26 regulates microRNAs and cell growth in the Drosophila ovarian stem cell lineage. Nature. 2008;454:241–245. [PMC free article] [PubMed]
Papoulas O., Beek S.J., Moseley S.L., McCallum C.M., Sarte M., Shearn A., Tamkun J.W. The Drosophila trithorax group proteins BRM, ASH1 and ASH2 are subunits of distinct protein complexes. Development. 1998;125:3955–3966. [PubMed]
Park J.W., Parisky K., Celotto A.M., Reenan R.A., Graveley B.R. Identification of alternative splicing regulators by RNA interference in Drosophila. Proc. Natl. Acad. Sci. USA. 2004;101:15974–15979. [PMC free article] [PubMed]
Pece S., Serresi M., Santolini E., Capra M., Hulleman E., Galimberti V., Zurrida S., Maisonneuve P., Viale G., Di Fiore P.P. Loss of negative regulation by Numb over Notch is relevant to human breast carcinogenesis. J. Cell Biol. 2004;167:215–221. [PMC free article] [PubMed]
Petrova T.V., Nykanen A., Norrmen C., Ivanov K.I., Andersson L.C., Haglund C., Puolakkainen P., Wempe F., von Melchner H., Gradwohl G. Transcription factor PROX1 induces colon cancer progression by promoting the transition from benign to highly dysplastic phenotype. Cancer Cell. 2008;13:407–419. [PubMed]
Pfreundt U., James D.P., Tweedie S., Wilson D., Teichmann S.A., Adryan B. FlyTF: improved annotation and enhanced functionality of the Drosophila transcription factor database. Nucleic Acids Res. 2010;38:D443–D447. [PMC free article] [PubMed]
Ramalho-Santos M., Yoon S., Matsuzaki Y., Mulligan R.C., Melton D.A. “Stemness”: transcriptional profiling of embryonic and adult stem cells. Science. 2002;298:597–600. [PubMed]
Reisman D., Glaros S., Thompson E.A. The SWI/SNF complex and cancer. Oncogene. 2009;28:1653–1668. [PubMed]
Schwabish M.A., Struhl K. The Swi/Snf complex is important for histone eviction during transcriptional activation and RNA polymerase II elongation in vivo. Mol. Cell. Biol. 2007;27:6987–6995. [PMC free article] [PubMed]
Schwamborn J.C., Berezikov E., Knoblich J.A. The TRIM-NHL protein TRIM32 activates microRNAs and prevents self-renewal in mouse neural progenitors. Cell. 2009;136:913–925. [PMC free article] [PubMed]
Shen Q., Wang Y., Dimos J.T., Fasano C.A., Phoenix T.N., Lemischka I.R., Ivanova N.B., Stifani S., Morrisey E.E., Temple S. The timing of cortical neurogenesis is encoded within lineages of individual progenitor cells. Nat. Neurosci. 2006;9:743–751. [PubMed]
Smith M.J., Kulkarni S., Pawson T. FF domains of CA150 bind transcription and splicing factors through multiple weak interactions. Mol. Cell. Biol. 2004;24:9274–9285. [PMC free article] [PubMed]
Sonoda J., Wharton R.P. Drosophila Brain Tumor is a translational repressor. Genes Dev. 2001;15:762–773. [PMC free article] [PubMed]
Sousa-Nunes R., Chia W., Somers W.G. Protein phosphatase 4 mediates localization of the Miranda complex during Drosophila neuroblast asymmetric divisions. Genes Dev. 2009;23:359–372. [PMC free article] [PubMed]
Southall T.D., Brand A.H. Neural stem cell transcriptional networks highlight genes essential for nervous system development. EMBO J. 2009;28:3799–3807. [PMC free article] [PubMed]
Subramanian A., Tamayo P., Mootha V.K., Mukherjee S., Ebert B.L., Gillette M.A., Paulovich A., Pomeroy S.L., Golub T.R., Lander E.S., Mesirov J.P. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA. 2005;102:15545–15550. [PMC free article] [PubMed]
Sur I. Kruppel-like factors 4 and 5: unity in diversity. Curr. Genomics. 2009;10:594–603. [PMC free article] [PubMed]
Xi R., Xie T. Stem cell self-renewal controlled by chromatin remodeling factors. Science. 2005;310:1487–1489. [PubMed]
Yoo A.S., Crabtree G.R. ATP-dependent chromatin remodeling in neural development. Curr. Opin. Neurobiol. 2009;19:120–126. [PMC free article] [PubMed]
Zheng L., Carthew R.W. Lola regulates cell fate by antagonizing Notch induction in the Drosophila eye. Mech. Dev. 2008;125:18–29. [PMC free article] [PubMed]
Zhou Q., Sharp P.A. Tat-SF1: cofactor for stimulation of transcriptional elongation by HIV-1 Tat. Science. 1996;274:605–610. [PubMed]
Zhu S., Lin S., Kao C.F., Awasaki T., Chiang A.S., Lee T. Gradients of the Drosophila Chinmo BTB-Zinc Finger Protein Govern Neuronal Temporal Identity. Cell. 2006;127:409–422. [PubMed]

Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Gene
    Gene records that cite the current articles. Citations in Gene are added manually by NCBI or imported from outside public resources.
  • GEO Profiles
    GEO Profiles
    Gene Expression Omnibus (GEO) Profiles of molecular abundance data. The current articles are references on the Gene record associated with the GEO profile.
  • MedGen
    Related information in MedGen
  • Nucleotide
    Primary database (GenBank) nucleotide records reported in the current articles as well as Reference Sequences (RefSeqs) that include the articles as references.
  • Pathways + GO
    Pathways + GO
    Pathways and biological systems (BioSystems) that cite the current articles. Citations are from the BioSystems source databases (KEGG and BioCyc).
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.
  • Taxonomy
    Taxonomy records associated with the current articles through taxonomic information on related molecular database records (Nucleotide, Protein, Gene, SNP, Structure).
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...