Post-transcriptional regulation of transcription factor codes in immature neurons drives neuronal diversity

Abstract

How the vast array of neuronal diversity is generated remains an unsolved problem. Here, we investigate how 29 morphologically distinct leg motoneurons are generated from a single stem cell in Drosophila. We identify 19 transcription factor (TF) codes expressed in immature motoneurons just before their morphological differentiation. Using genetic manipulations and a computational tool, we demonstrate that the TF codes are progressively established in immature motoneurons according to their birth order. Comparing RNA and protein expression patterns of multiple TFs reveals that post-transcriptional regulation plays an essential role in shaping these TF codes. Two RNA-binding proteins, Imp and Syp, expressed in opposing gradients in immature motoneurons, control the translation of multiple TFs. The varying sensitivity of TF mRNAs to the opposing gradients of Imp and Syp in immature motoneurons decrypts these gradients into distinct TF codes, establishing the connectome between motoneuron axons and their target muscles.

Keywords: CP: Developmental biology; CP: Neuroscience; Drosophila; motoneurons; neurodevelopment; post-transcriptional regulation; transcription factor.

Figure 1.. Lin A, a model to study how the MN-muscle connectome is built

(A) Drawing of a fly showing the CNS and Lin A/15 MNs. (B) VNC with six Lin A/15 labeled with myr:GFP (green) and immunostained with anti-BRP (blue). Arrowheads indicate MN cell bodies. (C) T1 leg with Lin A/15 axons labeled with myr:GFP (green). (D) T1 leg with a Lin A/15 MARCM clone expressing mCD8:GFP (green) and all muscles were labeled with Mhc -RFP (red). The number of MNs innervating the corresponding leg segment are indicated in (C). Leg muscles innervated by Lin A/15 are indicated in (D2). (E) Schematic showing the link between birth order and muscle targeting, modified from Baek and Mann (2009) (see link between birth order and muscle targets in STAR Methods). Top: schematic of the cell body of LinA/15 iMNs is shown. The numbers inside indicate their birth order, and the abbreviations below indicate the name of the MNs based on the nomenclature from Baek and Mann (2009). Bottom: schematic of a T1 leg innervated by Lin A/15 is shown; the muscles innervated by Lin A/15 are color coded based on their innervation. The line between the cell body and leg muscles indicates the relationships between MN birth order and muscle innervation. Muscle nomenclature is based on Soler et al. (2004): d, depressor; fe, femur; l, levator; m, muscle; r, reductor; t, tendon; ta, tarsus; ti, tibia; tr, trochanter. (F) Drawing of the anterior region of an L3 larva showing the CNS and Lin A/15 iMNs. (G and H) 3D reconstruction of six (G) and one (H1) right thoracic hemisegment 2 (T2R) Lin A/15 in an L3 larva labeled with myr:GFP (green). Ventral (G) and lateral (H1) views are shown; axes: A, anterior; L, lateral; V, ventral. (H2) Plot of the relative position of each Lin A/15 cell in (H1) from a lateral perspective is shown. Lin A/15 proliferative glia (PG) are in white, iMNs are in blue, GMCs are in white, and the NB is in cyan. (H3–H6) Confocal sections of the Lin A/15 in (H1) and (H2) immunostained with anti-Elav (blue) and Dpn (cyan). (I–K) Plots of the relative position of each Lin A/15 cell from a lateral perspective in L3 larvae fed with EdU at indicated time points: EdU⁺ cells (red) and EdU⁻ cells (blue). The horizontal axis indicates the time point of EdU feeding. On the right of each graph, confocal sections of Lin A/15 are shown. (L) Schematic of Lin A/15 cell bodies in an L3 larva and of an adult leg that shows the correlation between the position of the iMNs and the muscle targeting of adult MNs.

Figure 2.. Correlation between birth order and TF codes

(A–G) Plot (A) of the relative position of each Lin A/15 cell from a lateral perspective and confocal sections (B1–G2) in an L3 larva immunostained with anti-Jim (purple), anti-Elav (blue), and anti-Dpn (cyan) and where Lin A/15 is labeled with myr:GFP (green). Black lines in (A) indicate the positions of the confocal section in (B1)–(G2) (see Figure S1 for the expression pattern of all TFs). (H) On the left (specimen 1), same graph as in (A) but where the spatial axis is represented (units are in micrometer). On the right, graph from another specimen (specimen 2) is shown. (I) MN ordering on an x’ axis according to their relative distance from the NB; Jim⁺ iMNs are in purple. (J and K) Frequency of Jim expression as a function of x’ among 15 specimens before (J) and after (K) applying a Savitzky-Golay filter. (K) The arrow indicates the peak of the Jim⁺ cell cluster. The horizontal bar indicates the Jim⁺ cell cluster detected with the PCCD method (see STAR Methods). (L) Same plot as in (K) for 15 TFs. (M) Schematic of the TF codes expressed in each iMN predicted by the PCCD method in an L3 larva. Bottom: schematic of the cell body of Lin A/15 iMNs is shown. The numbers inside indicate their relative distances from NB. Top: the horizontal bars indicate the TF⁺ cell clusters detected with the PCCD method. The dotted lines indicate the coverage index at the border. The numbers in parentheses indicate the length of the TF-positive cell clusters (see STAR Methods). (N) Three thoracic hemisegments (T1–T3) (N1) and two confocal (N2 and N3) sections of the dotted box in (N1) in an L3 CNS with a myr:GFP⁺ Lin A/15 (green) immunostained with anti-Jim (red) and anti-Chinmo (blue). The numbers in white and orange indicate the Jim⁺ iMNs and the Jim⁺ and Chinmo⁺ iMNs, respectively. The position of the confocal sections is indicated in (N4). (N4) Plots of the relative position of the Lin A/15 corresponding to the Lin A/15 boxed in (N1) are shown. One out of the four Jim⁺ Chinmo⁺ cells expressed a very low level of Chinmo (number 6 in N3) (see Figure S2 for all co-stainings). For each co-staining, a minimum of 10 Lin A/15 samples are analyzed. (O) Same schematic as in (M) after validation and small corrections of the PCCD method by performing co-staining (see STAR Methods). TF gradient expressions of Castor, Br, Runt, Pros, and Oli were not taken into consideration when assigning the codes. The weak expression of Chinmo, Nvy, and Mamo is indicated in lighter colors. The color of the Lin A/15 iMN cell bodies change from white to gray at each switch to a new TF code. The name of the categories to which a TF belongs is indicated on left (see Figure 3).

Figure 3.. TF expression is progressively shaped in iMNs during development

(A–C, E–G, I–K, and M–O) Plot of the relative position of each Lin A/15 cell from a lateral perspective (A1, B1, C1, E1, F1, G1, I1, J1, K1, M1, N1, and O1) and confocal sections (A2, A3, B2, B3, C2, C3, E2, E3, F2, F3, G2, G3, I2, I3, J2, J3, K2, K3, M2, M3, N2, N3, O2, and O3) showing the expression of Jim (A2, A3, B2, B3, C2, and C3), Oli (E2, E3, F2, F3, G2, and G3), Br (I2, I3, J2, J3, K2, and K3), and Chinmo (M2, M3, N2, N3, O2, and O3) (red) in GFP⁺ Lin A/15 (green) immunostained with anti-Dpn (cyan) and anti-Elav (blue) throughout development. Arrowheads indicate the NB. The developmental time points are indicated on top (EL3/ML3/LL3: early/mid/late L3). The expression patterns were evaluated at seven different time points with 2 to 3 CNS samples, and only representative time points were chosen for EL3/ML3/LL3. Boxed regions in (I3) and (K2) show the weak expression of Br in GFP-labeled cells. (A1, B1, C1, E1, F1, G1, I1, J1, K1, M1, N1, and O1) Left panel: Lin A/15 NB is in cyan, GMCs and proliferative glia (intermingled with MNs at early time points) are in white, iMNs are in blue, and the cells expressing a TF are in red. Right panel: only the TF-expressing cells are color coded in red. The black lines indicate the positions of the confocal sections of the lower panels. (D, H, L, and P) Schematic of the expression (red) of Jim (D), Oli (H), Br (L), and Chinmo (P) in Lin A/15 during larval stages. The TFs in the same categories are listed on top left of each schematic (see Figure S3 for all TF stainings).

Figure 4.. Pattern of transcripts of indicated TFs and Imp and Syp proteins in Lin A/15

(A) Lin A/15 labeled with myr:GFP (green). (B) Plots of the relative position of each Lin A/15 cell in (A) from lateral (top) and ventral (bottom) perspectives. Axes: M, medial. iMNs and GMCs are both in blue; NB is in cyan. (C) Confocal sections of the Lin A/15 in (A) labeled with DAPI (blue), jim mRNA (purple), and Syp mRNA (cyan). Note: Syp mRNAs are used to indicate the NB since it is highly abundant in NBs (boxed regions in C1). The boxed regions in (C1)–(C3) are enlarged at the top-right region of each panel, where the intensity of the GFP and smFISH signals was numerically enhanced in (C3) to highlight the dorsal weak signal due to the thickness of the tissue. (D) 3D segmented Lin A/15 cells of (A) (see STAR Methods). (Left) Lateral view is shown; (right) ventral view is shown. Each cell is color coded from low (blue) to high (red) based on the number of jim mRNA spots (spheres). (E) Plot of the expression level (E1) of the jim mRNA in Lin A/15 iMNs in (A) and plot of the average expression level (E2) of the jim mRNA in six Lin A/15 samples, as a function of their relative distance to the NB. Only the 29 iMNs most distant to NB are represented. See Table S1 for average number of jim mRNA spots from all samples analyzed. (F–I) (Top) Same graph as in (E2) for indicated TFs; n R ≥ 4 Lin A/15 samples (n, number of Lin A/15 samples analyzed) (see Figure S4 for oli mRNA expression). (Bottom left) 3D segmented Lin A/15 cells where chinmo (F), mamo (G), br (H), and nvy (I) mRNA are detected are shown. (Bottom right) One representative confocal section is shown. See Table S1 for average number of TF mRNA spots from all samples analyzed. (J) Confocal sections of a T2 Lin A/15 labeled with myr:GFP (green) and immunostained with anti-Jim (purple), anti-Syp (cyan), and anti-Imp (red). Boxed regions and arrowheads indicate the eight Jim⁺ iMNs. (K) Confocal section of a T2 Lin A/15 labeled with myr:GFP (green) and immunostained with anti-Dpn (purple), anti-Syp (cyan), and anti-Imp (red). The CNS was mounted laterally in (J1)–(K3). (L) Schematic of the expression of Imp, Syp, Chinmo, Mamo, Jim, Br, and Nvy proteins (horizontal bar, from Figure 2) and TF mRNAs (gradient) in LinA/15 iMNs (bottom).

Figure 5.. Function of Imp, Syp, and TF codes in building axonal-muscle connectome

(A–J) Axonal targeting phenotypes of WT tub > GFP⁺ P35 (A and E), WT VGlut > GFP (B–D), Imp^−/− tub > P35 (F), Syp^−/− (G), tub > Imp (H), nvy^−/− (I), and jim KD (J) Lin A/15. Axons are mCD8:GFP⁺ (green); muscles were labeled with Mhc-RFP (red) in (B), (D), (E), (G), (I), and (J). Insets show leg regions most affected. Arrowheads and arrows point to normal targeting (white), the absence or reduction of targeting (orange), and extra targeting (blue). Asterisks indicate an absence of trochanter targeting in Imp^−/− tub > P35 (F) that was not considered as part of the phenotypes because the WT tub-P35 showed similar defects (A). In WT tub > GFP⁺ P35 LinA/15, the supernumerary neurons targeted a new region in the coxa (n = 8/9). Note: for each indicated phenotype, a minimum of six legs are analyzed; see Table S2 for details. (K) Schematic of the axonal targeting phenotypes. See Figure S6 for Imp^−/−, br^−/−, chinmo^−/−, and Oli^−/− phenotypes and the number of Lin A/15 cells in each genetic background. See Figure 7 for UAS-Jim phenotype. See Table S2 for the penetrance.

Figure 6.. Opposite expression of Imp and Syp shapes the TF codes

(A–I) WT (A1–C4), Syp^−/− (D1–F4), and tub > Imp (G1–I4) Lin A/15 expressing mCD8:GFP (green) and immunolabeled with anti-Br (red) and anti-Chinmo (blue) (A1–A4, D1–D4, and G1–G4), anti-Nvy (red) and anti-Jim (blue) (B1–B4, E1–E4, and H1–H4), or anti-Mamo (red) (C1–C4, F1–F4, and I1–I4). (A1, B1, C1, D1, E1, F1, G1, H1, and I1) Plots of the relative position of each Lin A/15 cell from a lateral perspective are shown where cells expressing a given TF are color coded in red or blue according to the immunostaining in (A2)–(A4), (B2)–(B4), (C2)–(C4), (D2)–(D4), (E2)–(E4), (F2)–(F4), (G2)–(G4), (H2)–(H4), and (I2)–(I4), which are confocal sections from ventral to dorsal. Note: for each co-staining, a minimum of seven Lin A/15 samples are analyzed. (J–O) Graph of the number of VGlut⁺ Lin A/15 iMNs expressing Chinmo (J and M), Br (J and M), Nvy (K and N), Jim (K and N), and Mamo (L and O) in Syp^−/− VGlut > GFP versus WT VGlut > GFP (J–L) and in tub > Imp VGlut > GFP versus WT VGlut > GFP (M–O). The number of Lin A/15 VGlut⁺ cells sometimes varied between controls or other genetic backgrounds due to the difficulty of having L3 larvae perfectly staged. However, the variations in (J) and (O) could not explain the large variation of the number of iMNs expressing Chinmo, Br, and Mamo. ns, not significant; differences of p < 0.05 were considered significant. *0.01 < p < 0.05; **0.001 < p < 0.01; ***0.0001 < p < 0.001, **** p < 0.0001. (P) Schematic of the epistasis between the Lin A/15 TFs and Imp and Syp. (Q) Schematic of the expression pattern of Chinmo, Mamo, Jim, Br, Nvy, and Oli in WT, Syp^−/−, and Imp overexpression in an L3 Lin A/15. Note: only 33 iMNs are indicated because the readout was VGlut > GFP, which was expressed in around 30 iMNs at that stage (see Figure S7 for Oli expression).

Figure 7.. Post-transcriptional regulations of TFs in iMNs by Imp shape the axon-muscle connectome

(A–C) Axon targeting phenotypes of WT (A1 and A2), VGlut > Imp (B1 and B2), and VGlut > jim (C1 and C2) Lin A/15 expressing mCD8:GFP (green). Muscles were labeled with Mhc-RFP (red). Arrowheads and arrows point to normal targeting (white) and the absence or reduced targeting (orange). Note: for each indicated phenotype, a minimum of 11 legs are analyzed; see Table S2 for the penetrance. (D–I) WT (D1–F4) and VGlut > Imp (G1–I4) LinA/15 expressing VGlut > mCD8:GFP (green) and immunolabeled with anti-Br (red) and anti-Jim (blue) (D1–D4 and G1–G4), anti-Nvy (red) and anti-Chinmo (blue) (E1–E4 and H1–H4), or anti-Mamo (red) (F1–F4 and I1–I4). (D1, E1, F1, G1, H1, and I1) Plots of the relative position of each Lin A/15 cell from a lateral perspective, where cells expressing a given TF are color coded in red or blue according to the immunostaining in (D2)–(D4), (E2)–(E4), (F2)–(F4), (G2)–(G4), (H2)–(H4), and (I2)–(I4), which are confocal sections from ventral to dorsal; for each co-staining, a minimum of eight Lin A/15 samples are analyzed. (J) Graph of the number of VGlut⁺ Lin A/15 MNs in WT VGlut > GFP versus VGlut > Imp, VGlut > GFP expressing Chinmo, Br, Nvy, Jim, and Mamo. Note: The increase of the number of iMNs expressing Br was also observed when affecting the level of Imp by using tub > GAL4 driver, but the difference was not statistically significant (Figure 7J versus Figure 6M). ns, not significant; differences of p < 0.05 were considered significant. *0.01 < p < 0.05; **0.001 < p < 0.01; ***0.0001 < p < 0.001, **** p < 0.0001. (K) Schematic of a model explaining how the architecture of the axon-muscle connectome is shaped. The dotted and continuous lines indicate low and high expression levels of the TF proteins, respectively.