InteractiveFly: GeneBrief

Zinc-finger protein: Biological Overview | References

How a cell decides to self-renew or differentiate is a critical issue in stem cell and cancer biology. Atypical protein kinase C (aPKC) promotes self-renewal of Drosophila larval brain neural stem cells, neuroblasts. However, it is unclear how aPKC cortical polarity and protein levels are regulated. This study identified a zinc-finger protein, Zif, which is required for the expression and asymmetric localization of aPKC. aPKC displays ectopic cortical localization with upregulated protein levels in dividing zif mutant neuroblasts, leading to neuroblast overproliferation. Zif was shown to be a transcription factor that directly represses aPKC transcription. It was further shown that Zif is phosphorylated by aPKC both in vitro and in vivo. Phosphorylation of Zif by aPKC excludes it from the nucleus, leading to Zif inactivation in neuroblasts. Thus, reciprocal repression between Zif and aPKC act as a critical regulatory mechanism for establishing cell polarity and controlling neuroblast self-renewal (Chang, 2010).

Drosophila larval brain neural stem cells, neuroblasts (NBs), divide asymmetrically to give rise to a self-renewing daughter and a ganglion mother cell (GMC) that is committed to differentiation. The mechanisms of NB asymmetric division have been studied primarily in embryonic NBs, and are conserved in larval NBs. Asymmetric division of NBs depends on asymmetric protein localization during mitosis. Apically localized proteins include the Bazooka (Drosophila Par3)/Par6/aPKC complex, Partner of Insc (Pins), Locomotion defects (Loco), and mushroom body defect (Mud). Gβγ and Ric-8 regulate Gαi localization. Apical proteins allow asymmetric localization of basal proteins Numb, Prospero (Pros), Brat, Partner of Numb (Pon), and Miranda (Mira), which are segregated into GMCs. Basal protein localization is also mediated by Discs large, Lethal (2) giant larvae (Lgl), Myosins II and VI, and Protein Phosphatases 2 and 4. aPKC can directly phosphorylate Numb and Mira to regulate their asymmetric localization (Chang, 2010).

Drosophila larval brain NBs utilize the asymmetric division machinery to distribute 'proliferation factors' and 'differentiation factors' to different daughters. Failure in asymmetric division can result in NB overproliferation. Larval brain tissues from mutants of pins, mira, and numb when transplanted into wild-type adults, can form malignant tumors. Two distinct populations of NBs have recently been identified in larval brains. Unlike type-I NBs, type-II NBs are Asense (Ase) negative and divide to produce a NB and an intermediate neural progenitor cell (INP) which produces multiple GMCs. aPKC is a NB proliferation factor in both type-I and -II NBs, as its ectopic expression throughout the entire cell cortex by CAAX motif leads to overgrowth in larval brains. Aurora-A and Polo kinases inhibit NB overgrowth primarily by regulating Numb asymmetry, while all basal proteins have brain tumor suppressor functions that prevent NB overgrowth. Aurora-A acts in both type-I and -II NBs; it also directly phosphorylates Par-6. Dap160/intersectin interacts with and stimulates aPKC activity, which is required for maintaining NB proliferation (Chang, 2010).

This study identifies a C2H2-type zinc-finger transcription factor that has been named 'Zif,' which is required for regulating both aPKC expression and cortical polarity. In addition, a reciprocal activity of aPKC on Zif—that aPKC directly phosphorylates Zif, regulating its activity via control of its subcellular localization (Chang, 2010).

A clonal screen was performed using mosaic analysis with a repressible cell marker (MARCM) and three ethlymethane sulfonate (EMS)-induced mutants (1L15, 2L745, and 2L497) belonging to a single complementation group were isolated. Compared with wild-type clones which mostly possess only one NB, supernumerary NB-like cells labeled by Mira were observed in these mutant clones. Genetic mapping and molecular analysis indicated that 1L15, 2L745, and 2L497 each carried a nonsense mutation in CG10267, a C2H2-type Zinc-finger protein (Zif), at Q110, Q192, and K306, respectively. Zif contains five zinc fingers at the C terminus which confers DNA-binding capability. All three mutants are embryonic lethal and are likely to be loss-of-function alleles, as Zif protein was undetectable in them. 1L15 (referred to as zif mutant hereafter) was used for most experiments. The NB overgrowth defects observed in all zif mutants could be completely rescued by expressing full-length Zif (Chang, 2010).

Examined next was whether Zif functions similarly to inhibit NB overgrowth within type-I and -II larval NB lineages. zif MARCM clones were generated by labeling NBs with Deadpan (Dpn)/Ase and differentiated neurons with Embryonic Lethal Abnormal Vision (Elav). Wild-type type-I NB clones contain one Dpn-positive/Ase-positive (Dpn⁺Ase⁺) NB with a large number of ELAV⁺ neurons. However, in zif mutant type-I NB clones, ectopic NBs were evident by the presence of multiple Dpn⁺Ase⁺ cells. This NB overgrowth occurs at the expense of neuron formation, as very few Elav⁺ neurons were observed in these clones. In type-II NB clones, zif mutants also had an increased number of Dpn⁺Ase^- NBs, compared to one Dpn⁺Ase^- NB in a wild-type type-II clone. zif mutant type-II clones contained 24.1 ± 8.1 INPs, similar to wild-type clones, indicating that Zif is not obviously required for INP self-renewal. Therefore, Zif functions in both type-I and -II NB lineages to inhibit NB overgrowth (Chang, 2010).

Zif protein, as detected by antibodies, is nuclear in interphase NBs and cytoplasmic throughout mitosis. This localization pattern of endogenous Zif is recapitulated by an inducible construct of Venus-tagged full-length Zif (UAS-Venus::zifWT-FL) expressed using Insc-Gal4. Zif is expressed in NBs, GMCs, and neurons, suggesting a ubiquitous expression of Zif in most cell types in larval brains (Chang, 2010).

Whether Zif regulates asymmetric protein localization in NBs was investigated. In contrast to wild-type prometa/metaphase NBs in which 84% had a robust localization of aPKC to the apical cortex (12% with weaker crescent of aPKC), 78% of the zif mutant NBs exhibited aPKC delocalization throughout the cortex. The remaining 22% of the zif prometa/metaphase NBs analyzed also exhibited disrupted aPKC localization, either appearing as punctate structures on the cortex or diffused crescents. Consistent with the observed aPKC delocalization, Baz is also delocalized or absent in 56% of prometa/metaphase zif mutant NBs. Zif is also required for asymmetric localization of Numb, Mira, Pros, and Pon. Numb is cortically localized or absent in zif mutant prometa/metaphase NBs. Pon, Mira, and Pros are also delocalized in zif prometa/metaphase NBs. Thus, Zif regulates asymmetric localization of apical and basal complex proteins during NB asymmetric divisions (Chang, 2010).

The NB overgrowth phenotype in zif mutant clones was significantly suppressed in aPKC^06403/+ heterozygous background. aPKC asymmetric localization during prophase/metaphase is also mostly restored. zif mutant NB overproliferation was also suppressed by overexpression of Lgl3A. These data indicate that Zif inhibits NB overgrowth primarily by suppressing aPKC function (Chang, 2010).

As a putative transcription factor, it was asked if Zif regulates the expression of asymmetric cell division genes. Examination of the mRNA transcript levels of most known asymmetric cell division genes by reverse transcriptase (RT)-PCR in zif/Df(3R)WIN11 hemizygous embryos versus wild-type showed that aPKC transcripts were dramatically increased in zif/Df(3R)WIN11 mutant embryos compared to wild-type. The levels of transcript of all other genes tested remained similar between zif/Df(3R)WIN11 and wild-type samples. Quantitative real-time PCR was performed using the same samples, and it was found that normalized transcript levels for aPKC in zif mutant were upregulated to 363% compared to wild-type (100%), while transcript levels of Pros and Dap160 in zif mutant remained comparable to wild-type. Consistently, aPKC protein levels were dramatically increased in zif^RNAi larval brains in which zif levels were further reduced by zif+/−. Knocking down zif levels in S2 cells also leads to an increase in aPKC levels; further supporting that Zif is able to repress aPKC levels in dividing cells. Knocking down of Zif in S2 cells did not change total Numb protein level but observed a dramatic increase was seen in phosphorylation of Numb by aPKC recognized by an anti-ps7Numb antibody, indicating that the excess aPKC protein in zif mutants was functionally active (Chang, 2010).

Chromatin immunoprecipitation (ChIP) was performed in S2 cells to investigate whether Zif can directly bind to aPKC promoter. At the aPKC regulatory region, a 500 bp DNA region was identified where Zif directly binds in the ChIP assay. The Zif-binding region was further narrowed down to a 200 bp DNA fragment (−652 to −424 bp upstream of the transcriptional start site, which is referred to as the mini-aPKC proximal promoter (aPKC pro). Next, a luciferase reporter assay was carried out to test the effect of Zif on aPKC pro. Zif was capable of suppressing the luciferase reporter activity under the control of aPKCpro in a dose-dependent manner, while no consistent effect was observed on the actin promoter. Taken together, Zif directly represses aPKC transcription by binding to an upstream proximal promoter region of aPKC (Chang, 2010).

Endogenous Zif was found to exist in two states differing in their net charges in two-dimensional polyacrylamide gel electrophoresis (2-D PAGE). The proportion of Zif that has more acidic residues (more negatively charged) was abolished by knockdown of aPKC in S2 cells, indicating that the putative posttranslational modification of Zif is aPKC dependent. Using group-based prediction system, two putative consensus phosphorylation sites for aPKC [S-X-(R/L)_1-3] were identified at serine residues 197 and 292 of Zif. Zif is a good substrate of aPKC as shown by an in vitro kinase assay using both radiometric and nonradioactive approaches. Mutating either S197 or S292 to alanine alone does not consistently abrogate aPKC phosphorylation, whereas mutations of both completely abolished phosphorylation by aPKC. An antibody was generated that specifically recognizes phosphorylation of S292 in Zif (pS292Zif) by aPKC in the kinase assay. The phosphorylated Zif is completely absent upon aPKC knockdown in S2 cells. In addition, phosphorylated Zif is also barely detectable in homozygous aPKC^k06403 larval brain extracts. In contrast, phosphorylation of Zif at S292 was unaffected in aur⁸⁸³⁹ larval brains. Taken together, these data strongly suggest that aPKC phosphorylates Zif both in vitro and in vivo (Chang, 2010).

To analyze the functional role of aPKC-directed phosphorylation of Zif, transgenic flies were generated expressing mutated Zif constructs fused to Venus reporter-phosphomimetic form of Zif by mutating S197 and S292 to aspartic acid residues (S197,292D) or a nonphosphorylatable form of Zif by mutating both serine residues to alanine residues (S197,292A). The strong nuclear localization of Venus-Zif^{S197, 292A} is reminiscent of wild-type Venus-Zif in NBs. In contrast, Venus-Zif^S197,292D which still possesses the putative nuclear localization sequence was completely excluded from the nucleus of interphase NBs. Localization of overexpressed Venus-Zif mutant proteins was verified by anti-Zif antibody. Therefore, phosphorylation of Zif by aPKC results in Zif exclusion from the nucleus in the NBs. Furthermore, the nonphosphorylatable form of Zif (Zif^{S197, 292A}) was capable of almost complete rescue of the NB overgrowth or Mira delocalization phenotypes (76% of metaphase NBs had asymmetric localized Mira) in zif mutant MARCM clones. In contrast, expressing the phosphomimetic form of Zif^{S197, 292D} could not significantly rescue the zif mutant overgrowth (71% of clones contained multiple NBs) or Mira mislocalization phenotypes. These data suggest that the nonphosphorylated form of Zif is active and can inhibit excess NB self-renewal (Chang, 2010).

This study has shown that Drosophila Zif is a new player regulating asymmetric cell division and the balance between self-renewal and differentiation. Several Zif-related proteins in mammals, including mouse zinc finger protein 160 and human zinc finger protein 287, both of which share 19% amino acid identities with Drosophila Zif and contain a serine that is conserved with S292 of Zif. These mammalian Zif-related proteins remain uncharacterized and their function during stem cell homeostasis will be of great interest (Chang, 2010).

Zif was shown to have a critical role in directly repressing aPKC expression to inhibit NB overgrowth. In zif embryos, aPKC expression in epithelium appears to be elevated compared to wild-type, indicating that the regulation of aPKC by Zif may occur in other tissues besides NBs. Overexpression of aPKC alone in NBs is insufficient to cause asymmetric division defect. It is currently unclear how Zif controls asymmetric localization of aPKC, which is likely to be independent of its regulation of aPKC transcription. It is possible that Zif regulates unidentified proteins that in turn control aPKC cortical polarity. It will be of great interest to identify such proteins and elucidate their functions in asymmetric division of NBs. In zif mutants, the delocalization of both aPKC and Numb to the entire cortex in metaphase NBs is similar to that caused by ectopic expression of aPKC-CAAX. This is inconsistent with a previous model where aPKC excludes Numb from the cortex in sensory organ precursor cells. The results may suggest that cortical aPKC does not always exclude Numb from the cortex in vivo (Chang, 2010).

Zinc-finger proteins that contain the C2H2 motif, a DNA-binding motif, are putative transcription factors. Mammalian C2H2-type zinc-finger transcription factors are abundant. However, very little is known about posttranslational modification of zinc-finger proteins, which is essential for regulating the functions of these transcription factors during development. Phosphorylation of a threonine residue in the linker region (interfinger spacer) of C2H2 zinc-finger proteins can inactivate their function during mitosis. The current data suggest that aPKC phosphorylates Zif on S197 and S292, resulting in Zif exclusion from the nucleus. It will be interesting to assess whether these findings represent a conserved common mechanism by which C2H2 zinc-finger transcription factors are inactivated (Chang, 2010).

Given that Zif is a nuclear transcription factor and aPKC is predominantly localized to the cortex and cytoplasm, it is conceivable that aPKC could have access to cytoplasmic Zif during mitosis when Zif is localized in the cytoplasm after nuclear membrane breakdown. Upon phosphorylation, Zif would be retained in the cytoplasm at the interphase of the following cell cycle. Consistent with this hypothesis, an increase was detected in aPKC-phosphorylated Zif in Nocodazole-treated S2 cells that are synchronously released into mitosis where Zif is localized to the cytoplasm. In addition, it was observed that by mutating the nuclear localization signal (NLS) in Zif, there was an increase in the cytoplasmic localization of Zif in interphase S2 cells. This suggests that nuclear-cytosol shuttling of Zif may allow phosphorylation of Zif by aPKC to still take place during interphase. In summary, these findings suggest that mutual inhibition between Zif and aPKC is critical for ensuring attainment of the appropriate levels and polarity of aPKC optimal for the proper control of asymmetric division and self-renewal of NBs (Chang, 2010).

Search PubMed for articles about Drosophila Zif

Chang, K. C., et al. (2010). Interplay between the transcription factor Zif and aPKC regulates neuroblast polarity and self-renewal. Dev. Cell 19(5): 778-85. PubMed ID: 21074726

date revised: 6 May 2012

Home page: The Interactive Fly © 2011 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.