InteractiveFly: GeneBrief

Ets at 65A: Biological Overview | References

The remarkable diversity of neurons in the nervous system is generated during development, when properties such as cell morphology, receptor profiles and neurotransmitter identities are specified. In order to gain a greater understanding of neurotransmitter specification this study profiled the transcription state of cholinergic, GABAergic and glutamatergic neurons in vivo at three developmental time points. 86 differentially expressed transcription factors were identified that are uniquely enriched, or uniquely depleted, in a specific neurotransmitter type. Some transcription factors show a similar profile across development, others only show enrichment or depletion at specific developmental stages. Profiling of Acj6 (cholinergic enriched) and Ets65A (cholinergic depleted) binding sites in vivo reveals that they both directly bind the ChAT locus, in addition to a wide spectrum of other key neuronal differentiation genes. It was also shown that cholinergic enriched transcription factors are expressed in mostly non-overlapping populations in the adult brain, implying the absence of combinatorial regulation of neurotransmitter fate in this context. Furthermore, the data underlines that, similar to Caenorhabditis elegans, there are no simple transcription factor codes for neurotransmitter type specification (Estacio-Gómez, 2020).

Neurotransmitter identity is a key property of a neuron that needs to be tightly regulated in order to generate a properly functioning nervous system. This study has investigated the dynamics and extent of transcription factor specificity in fast-acting neurotransmitter neuronal types in Drosophila. The transcription state of cholinergic, GABAergic and glutamatergic neurons was profiled in the developing and adult brain of Drosophila. Enriched Pol II occupancy was observed at the relevant neurotransmitter synthesis genes and other genes associated with the activity of the specific types. The monoamine neurotransmitter related genes Vmat, DAT and Tdc2 are enriched in glutamatergic neurons, which is not unprecedented, as monoamine populations can also be glutamatergic (Aguilar, 2017; Trudeau, 2018). Cholinergic, GABAergic, serotonergic and dopaminergic receptors are enriched in embryonic GABAergic neurons relative to the other two fast-acting neurotransmitter types, which correlates with GABAergic interneurons acting as integrative components of neural circuits. The enrichment of MAP kinase pathway regulators in cholinergic neurons is intriguing, suggesting that this signalling pathway may have a specific role in these neurons. This is supported by a recent study showing that MAP kinase signalling acts downstream of Gq-Rho signalling in C. elegans cholinergic neurons to control neuron activity and locomotion (Estacio-Gómez, 2020).

Importantly, this study has uncovered and highlighted transcription factors and non-coding RNAs differentially expressed between these types. Some of these are expected based on previous studies in Drosophila, including acj6 (cholinergic) and Dbx (GABAergic) (Lacin, 2009). Also, studies in other model organisms fit with the current findings, for example, cholinergic enriched knot, whose orthologue, UNC-3 (C. elegans), is a terminal selector for cholinergic motor neuron differentiation. In addition, RFX, the vertebrate orthologue of Rfx, which was identified as glutamatergic enriched, can increase the expression of the neuronal glutamate transporter type 3. However, this study has identified many differentially expressed transcription factors that have not had their role studied with respect to neurotransmitter specification, or cases where there is supportive, but not direct, evidence for a role in neurotransmitter specification. For instance, vertebrate neuronal precursors expressing Nkx2.1 (HGTX orthologue) predominantly generate GABAergic interneurons, and a polyalanine expansion in ARX (homeobrain (hbn) orthologue) causes remodelling and increased activity of glutamatergic neurons in vertebrates. Acj6 is expressed in a subset of cholinergic neurons and Dbx in a subset of GABAergic neurons. None of the enriched transcription factors identified in this study are expressed in all of the neurons of a particular neurotransmitter type. This highlights that, similar to C. elegans, there are no simple transcription factor codes for neurotransmitter type specification in Drosophila (Estacio-Gómez, 2020).

Uniquely enriched factors are candidates for promoting a neurotransmitter fate, and a number of them were tested for their ability to reprogram neurons on a global scale in embryos. No obvious changes were observed, however, this is not particularly surprising considering the importance of cellular context for the reprogramming of neuronal properties. Successful reprograming may require intervention at a specific time point (e.g., at the progenitor stage), the co-expression of appropriate co-factors, and/or to exclusively target a neuronal subpopulation within each neurotransmitter type. Future work could investigate these factors in specific and relevant lineages, to shed light on important contextual information (Estacio-Gómez, 2020).

The majority of transcription factors identified as directly regulating neurotransmitter fate act in a positive manner, whereas only a handful of studies describe the role of repressors. Incoherent feedforward loops exist in C. elegans, where terminal selectors activate repressors, which feedback onto effector genes. In vertebrates, both Neurogenin 2 and Tlx3 are required for the specification of certain glutamatergic populations but also act to repress GABAergic fate. Whether this is direct repression of Glutamic acid decarboxylase (Gad) genes (required for the synthesis of GABA), or indirectly, through another transcription factor, is unclear. This study has identified several transcription factors that are expressed in two neurotransmitter types, but absent from the other. These include apterous (ap), Ets65A (long transcripts) and orthopedia (otp), which this study hypothesizes to be candidate repressors, given their absence from cells with a specific neurotransmitter identity. Profiling of Ets65A-PA binding in vivo, reveals that it directly binds ChAT, and therefore has the potential to directly regulate cholinergic fate. Similar to the candidate activators, ectopic expression of these candidates did not show any obvious repression of the respective neurotransmitter genes, however, again, this might be because they can only act as a repressor in specific contexts (e.g., when a co-repressor is present), or that they regulate genes associated with specific types but do not directly regulate neurotransmitter identity (Estacio-Gómez, 2020).

The development of single cell RNA-seq (scRNA-seq) technology has led to the profiling of several Drosophila tissues, including the whole adult brain, the central adult brain and the adult optic lobes. This study has mined the whole adult brain data (u, 2018) to compare and investigate the cholinergic enriched factors that this study has identified in adult brains. The enrichment of these transcription factors (compared to GABAergic and glutamatergic neurons) is also observed in the scRNAseq data. Furthermore, it was discovered that the cholinergic cells that these factors are expressed in are almost non-overlapping. This is an intriguing finding, as it suggests that these factors, if they are indeed acting to promote/maintain cholinergic fate, they are not acting together in this context. This scenario maybe different during development, where specification is occurring, and it will be interesting to test this when high coverage scRNAseq data is available for the third instar larval brain. This study observed more differentially expressed transcription factors in the L3 larval stage (58) compared to the embryo (40) or adults (33). This may reflect the existence of both the functioning larval nervous system (built during embryogenesis) and the developing adult nervous system at this stage. While both the embryo and larval data are similar on a global scale, Pol II occupancy and chromatin accessibility in the adult brain is less correlated. It is currently unclear whether this is due to adult VNCs being absent from the profiling experiments, or differences between immature and fully mature neurons, such as overall lower transcriptional activity in adults. Previous work has shown that global chromatin accessibility distribution in adult neurons is distinct from larval neurons, which may account for some of these differences (Estacio-Gómez, 2020).

Apart from the neurotransmitter synthesis genes, the chromatin accessibility of the different neuronal types, at a given stage, is surprisingly similar, as demonstrated in embryos. The enriched accessibility is not just restricted to the gene bodies of the neurotransmitter genes, and peaks are present upstream (Gad1) and downstream (VGlut), which are likely enhancers. Accessibility at the ChAT gene is clearly higher in cholinergic neurons at the embryonic and adult stages, however, in third instar larvae, the difference is less pronounced. This could reflect increased plasticity at this stage, possibly linked to the dramatic remodelling of larval neurons during metamorphosis, or that this accessibility across the types is due to non-specific expression of the VAChT gene that overlaps with ChAT at its 5' end. While a subset of transcription factors display obvious contrasts in Pol II occupancy, the same transcription factors have no observable, or minor, differences in accessibility. This could be due to transcription factors being expressed at relatively lower levels and/or that they are only expressed in a subset of the cells, therefore the difference is less prominent (Estacio-Gómez, 2020).

Evidence is emerging for the roles of miRNAs in generating neuronal diversity, including the differentiation of taste receptor neurons in worms and dopaminergic neurons in vertebrates. This study found the enriched expression of mir-184 in GABAergic cells, which is intriguing, as mir-184 has been shown to downregulate GABRA3 (GABA-A receptor) mRNA (possibly indirectly) in vertebrate cell lines, and may be a mechanism to help prevent GABAergic neurons self-inhibiting. Furthermore, mir-87 has enriched RNA polymerase II occupancy in cholinergic neurons, and when mutated causes larval locomotion defects in Drosophila (Estacio-Gómez, 2020).

Acj6 is expressed in adult cholinergic neurons, whilst Ets65A-PA is expressed in non-cholinergic adult neurons. However, despite this, they bind a large number of common target genes. This includes 20% (101/493) of all genes annotated for a role in 'neuron projection development' (GO:0031175). This is quite striking, especially as this is in the adult, where there is virtually no neurogenesis or axonogenesis. However, this may reflect dendritic re-modelling processes, or a requirement of neurons to continuously express transcription factors, even after development, to maintain their fate. The acj6 orthologues, unc-86 and Brn3a are both required to maintain the fate of specific cholinergic populations, and transcriptional networks that specific Tv1/Tv4 neurons in Drosophila are also required to maintain them in the adult. Therefore, the binding of Acj6 and Ets65A-PA to developmental genes and ChAT in adult neurons could be required for the continued activation (and repression) of genes governing neuronal identity. MAP kinase signalling genes are enriched in cholinergic neurons and Ets65A-PA specifically binds MAP kinase signalling genes, making it tempting to speculate that Ets65A-PA acts to repress cholinergic specific genes such as ChAT and MAP kinase genes. These Acj6 and Ets65A-PA data also emphasize the diverse set of neuronal differentiation genes a single transcription factor could regulate (Estacio-Gómez, 2020).

The precise synthesis and utilisation of neurotransmitters ensures proper information flow and circuit function in the nervous system. The mechanisms of specification are lineage specific, predominantly through the action of transcription factors. This study has provided further insights into the complement of different transcription factors that regulate neurotransmitter identity throughout development. Furthermore, this study identified the genomic binding of a known activator, and a candidate repressor, of cholinergic fate in the adult, emphasizing the broad spectrum of neural identity genes that they could be regulating outside of neurotransmitter use. Given the strong evidence for conserved mechanisms controlling neurotransmitter specification, these data will be a useful resource for not just researchers using Drosophila but other model systems too. Continued work to elucidate the mechanisms, co-factors and temporal windows in which these factors are acting will be fundamental in gaining a comprehensive understanding of neurotransmitter specification (Estacio-Gómez, 2020).

Temporal progression of Drosophila medulla neuroblasts generates the transcription factor combination to control T1 neuron morphogenesis

The Drosophila medulla, part of the visual processing center of the brain, contains more than 70 neural types generated by medulla neuroblasts which sequentially express several temporal transcription factors (TTFs), including Homothorax (Hth), eyeless (Ey), Sloppy paired 1 and 2 (Slp), Dichaete (D) and Tailless (Tll). However, it is not clear how such a small number of TTFs could give rise to diverse combinations of neuronal transcription factors that specify a large number of medulla neuron types. This study reports how temporal patterning specifies one neural type, the T1 neuron. The T1 neuron is the only medulla neuron type that expresses the combination of three transcription factors Ocelliless (Oc or Otd), Sox102F and Ets65A. Using CRISPR-Cas9 system, this study shows that each transcription factor is required for the correct morphogenesis of T1 neurons. Interestingly, Oc, Sox102F and Ets65A initiate expression in neurons beginning at different temporal stages and last in a few subsequent temporal stages. Oc expressing neurons are generated in the Ey, Slp and D stages; Sox102F expressing neurons are produced in the Slp and D stages; while Ets65A is expressed in subsets of medulla neurons born in the D and later stages. The TTF Ey, Slp or D is required to initiate the expression of Oc, Sox102F or Ets65A in neurons, respectively. Thus, the neurons expressing all three transcription factors are born in the D stage and become T1 neurons. In neurons where the three transcription factors do not overlap, each of the three transcription factors can act in combination with other neuronal transcription factors to specify different neural fates. This study shows that this way of expression regulation of neuronal transcription factors by temporal patterning can generate more possible combinations of transcription factors in neural progeny to diversify neural fates (Naidu, 2020).

T1 neurons are a class of mysterious neurons that connect the lamina and the medulla part of the optic lobe. They are uni-columnar neurons with one in each of the 800 columns of the medulla. The cell body of the T1 neuron is found in the medulla cortex, and its axon branches in a characteristic 'T' shape on the distal surface of the medulla. One branch projects through the outer optic chiasm back to the lamina and then forms a basket like structure of processes surrounding each lamina cartridge. The other branch arborizes in the M2 layer of the medulla with a dense bush like structure. T1 neuron is post-synaptic to amc (lamina amacrine cells), and the amc/T1 pathway was shown to enhance the lamina neuron L1 motion detection pathway at intermediate contrast (Rister, 2007). Depolarizing T1 neurons affected the flight steering responses to visual stimuli (Tuthill, 2013; Naidu, 2020).

Through screening antibodies and GFP fusion lines, this study found that Ocelliless (oc), Sox102F and Ets65A are expressed in T1 neurons, and the combination of these three TFs can distinguish T1 neurons from all other medulla neurons. Using CRISPR-Cas9 system, bi-allelic somatic mutations of each of the three TF genes were generated in T1 neurons; knock-down of each one affected different aspects of the T1 neuron morphology. Next, how the expression of each TF is controlled by temporal patterning to generate the combination code was examined. Oc expression in neurons starts in the Ey temporal stage, and continues in the Slp and D temporal stages, and Ey is required for the initiation of Oc expression in neurons; while Sox102F expression in neurons starts in the Slp temporal stage, and continues in the D temporal stage, and Slp is required for initiating the expression of Sox102F in neurons; finally, Ets65A is expressed in subsets of medulla neurons born in the D and later temporal stages, and D is required for the expression of Ets65A. Thus, the three TFs that control T1 neuron morphology initiate their expression in neurons beginning at different temporal stages controlled by different TTFs, but each of them spans a few temporal stages, and the neurons expressing all three TFs are born in the D stage and become T1 neurons. In neurons where the three transcription factors do not overlap, each of the three TFs could also act with other neuronal TFs to specify different neural fates. In this way, more combinations of TFs can be generated through temporal patterning (Naidu, 2020).

This study identified a combination of three transcription factors that control T1 neuron morphology, and examined how the expression of these three transcription factors are controlled by temporal patterning of medulla neuroblasts. Oc is turned on in neurons starting in the late Ey stage, and Oc expressing neurons continue to be generated in the Slp and D stages, although the fates will be different, possibly dependent on the co-expression with other neuronal TFs. Sox102F expressing neurons start to be generated in the Slp stage and continue in the D stage. Ets65A expressing neurons are generated in the D and later temporal stages. Thus, the three TFs that control T1 neuron morphology start their expression in neurons born at different temporal stages, and require the corresponding TTF for initiation of their expression, and each neuronal TF is expressed in neurons spanning a few temporal stages. One advantage for such temporal control of neuronal TFs is that more combinations of TFs can be generated to specify different fates. For example, Toy is expressed in the N-on neuronal progeny born from the Slp and D stages, and also in some N-off progeny born from the late Ey stage neuroblasts in some regions of the medulla. Results from this study and others suggest that the subset of Sox102F neurons that do not express Oc, express Toy and Ap instead, and they are specified as Tm5 neurons. In addition, the neurons that express both Toy and Oc in the N-off progeny of some late Ey stage neuroblasts could determine another unknown neural type. Although it remains to be determined whether these TF combinations are indeed required for the corresponding neural fates, these examples do suggest that different combinations of neuronal TFs can be created that might determine different fates (Naidu, 2020).

Mutation of each of the three TFs expressed in T1 caused a certain morphological defect, similar to the morphology TFs that act in combinations to determine motor neuron morphology. For oc and Ets65A mutant neurons, it appeared that they still maintained the T1 fate, but the morphology was abnormal. Some Sox102F mutant neurons resembled medulla intrinsic neurons, but without functional assay, it was not clear whether they were fully transformed to a normal Mi neuron fate, or they still maintained some T1 neuron charateristics but underwent dramatic morphological changes. One question is whether the combination of TFs regulate neuron morphology by simple addition (each TF determines one feature, and the simple addition of these features determines one neural type), or in a synergistic way (three TFs together can determine features not determined by either TF alone). In the case of T1, when Sox102F was removed from T1 neurons, the driver used (T1-LexA) was still expressed in the mutant neurons, but the neurons became more like medulla intrinsic neurons, and some neurons lost the projection back to the lamina. However, Sox102F is not expressed in other neurons that project back to the lamina like lamina wide field neurons (lawf 1/2) which express Hth and Eya. Instead, Sox102F is also expressed in a Transmedulla neural type (Tm5) which do not resemble T1 neurons. Thus, these results favor the synergistic action model of neuronal TFs to control neuron morphology (Naidu, 2020).

The results are consistent with the principle that integration between temporal/spatial patterning of neuroblasts and the Notch-dependent binary neuron fate choice further diversifies neural fates. This study found that T1 neurons are derived from the Notch-off hemilineage of D stage neuroblasts. In addition, although T1 neurons are uni-columnar neurons that are generated throughout the main medulla region, there is a spatial component that regulates Oc expression and neural fate specification. Neurons that co-express Oc and Forkhead are only localized in the Dpp domains. Through analyzing the sequencing data published for all medulla neurons, the neurons expressing both Oc and Fkh should become the Dm12 neuron, a multi-columnar neuron with arborizations spanning several columns. Thus, these results support the conclusion that uni-columnar neurons are generated throughout the medulla main region, while multi-columnar neurons are generated in special spatial domains determined by spatial patterning (Erclik, 2017; Naidu, 2020).

In summary, this study of T1 neuron specification illustrated an example how temporal patterning of neuroblasts sequentially turns on the expression of three TFs in neuronal progeny, and generates different combinational codes to determine neural fates. In the future it will be interesting to examine how TTFs in neuroblasts regulate the expression of neuronal TFs in neurons that often span a few temporal stages. Only a subset of neurons maintain the expression of TTFs, while other neurons do not. Thus the TTFs should determine the expression of neuronal TFs already in neuroblasts. It is possible that the TTF promotes epigenetic modifications in the neuronal TF gene locus, so that the TF will be turned on in its progeny as well as in neurons born in subsequent temporal stages. It is also possible that the expression of the same neuronal TF in two subsequent temporal stages are controlled by two separate enhancers that respond to different TTFs. Addressing these questions will further advance understanding of the link between neuroblast temporal patterning and neural fate specification (Naidu, 2020).

Search PubMed for articles about Drosophila Ets65A

Aguilar, J. I., Dunn, M., Mingote, S., Karam, C. S., Farino, Z. J., Sonders, M. S., Choi, S. J., Grygoruk, A., Zhang, Y., Cela, C., Choi, B. J., Flores, J., Freyberg, R. J., McCabe, B. D., Mosharov, E. V., Krantz, D. E., Javitch, J. A., Sulzer, D., Sames, D., Rayport, S. and Freyberg, Z. (2017). Neuronal depolarization drives increased dopamine synaptic vesicle loading via VGLUT. Neuron 95(5): 1074-1088 e1077. PubMed ID: 28823729

Davie, K., Janssens, J., Koldere, D., De Waegeneer, M., Pech, U., Kreft, L., Aibar, S., Makhzami, S., Christiaens, V., Bravo Gonzalez-Blas, C., Poovathingal, S., Hulselmans, G., Spanier, K. I., Moerman, T., Vanspauwen, B., Geurs, S., Voet, T., Lammertyn, J., Thienpont, B., Liu, S., Konstantinides, N., Fiers, M., Verstreken, P. and Aerts, S. (2018). A single-cell transcriptome atlas of the aging Drosophila brain. Cell 174(4): 982-998 e920. PubMed ID: 29909982

Erclik, T., Li, X., Courgeon, M., Bertet, C., Chen, Z., Baumert, R., Ng, J., Koo, C., Arain, U., Behnia, R., del Valle Rodriguez, A., Senderowicz, L., Negre, N., White, K. P. and Desplan, C. (2017). Integration of temporal and spatial patterning generates neural diversity. Nature 541(7637): 365-370. PubMed ID: 28077877

Estacio-Gomez, A., Hassan, A., Walmsley, E., Le, L. W. and Southall, T. D. (2020). Dynamic neurotransmitter specific transcription factor expression profiles during Drosophila development. Biol Open 9(5). PubMed ID: 32493733

Lacin, H., Chen, H. M., Long, X., Singer, R. H., Lee, T. and Truman, J. W. (2019). Neurotransmitter identity is acquired in a lineage-restricted manner in the Drosophila CNS. Elife 8. PubMed ID: 30912745

Lee, M. H. and Salvaterra, P. M. (2002). Abnormal chemosensory jump 6 is a positive transcriptional regulator of the cholinergic gene locus in Drosophila olfactory neurons. J Neurosci 22(13): 5291-5299. PubMed ID: 12097480

Naidu, V. G., Zhang, Y., Lowe, S., Ray, A., Zhu, H. and Li, X. (2020). Temporal progression of Drosophila medulla neuroblasts generates the transcription factor combination to control T1 neuron morphogenesis Dev Biol. PubMed ID: 32442418

Rister, J., Pauls, D., Schnell, B., Ting, C. Y., Lee, C. H., Sinakevitch, I., Morante, J., Strausfeld, N. J., Ito, K. and Heisenberg, M. (2007). Dissection of the peripheral motion channel in the visual system of Drosophila melanogaster. Neuron 56(1): 155-170. PubMed ID: 17920022

Trudeau, L. E. and El Mestikawy, S. (2018). Glutamate cotransmission in cholinergic, GABAergic and monoamine systems: contrasts and commonalities. Front Neural Circuits 12: 113. PubMed ID: 30618649

Tuthill, J. C., Nern, A., Holtz, S. L., Rubin, G. M. and Reiser, M. B. (2013). Contributions of the 12 neuron classes in the fly lamina to motion vision. Neuron 79(1): 128-140. PubMed ID: 23849200

date revised: 26 September 2020

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