The POU transcription factor UNC-86 controls the timing and ventral guidance of Caenorhabditis elegans axon growth

Abstract

The in vivo mechanisms that coordinate the timing of axon growth and guidance are not well understood. In the Caenorhabditis elegans hermaphrodite specific neurons (HSNs), the lin-4 microRNA controls the stage of axon initiation independent of the UNC-40 and SAX-3 ventral guidance receptors. lin-4 loss-of-function mutants exhibit marked delays in axon outgrowth, while lin-4 overexpression leads to precocious growth in the L3 larval stage. Here, we show that loss of the POU transcription factor UNC-86 not only results in penetrant ventral axon growth defects in in the HSNs, but also causes processes to extend in the L1, three stages earlier than wild-type. This temporal shift is not dependent on UNC-40 or SAX-3, and does not require the presence of lin-4. We propose that unc-86(lf) HSN axons are misguided due to the temporal decoupling of axon initiation and ventral guidance responses.

Fig. 1. Wild-type HSNs initiate axon growth in the ventral direction

Top: Cartoon representation of an L4-stage animal, with the CAN neuron and excretory canal dorsal to the HSN. Image not drawn exactly to scale. Bottom: Images of a representative L4-stage wild-type HSN expressing lin-4∷GFP(zaIs1) that extends its axon ventrally and then anteriorly. Left: DIC merged with GFP. Right: GFP only. The arrow points to the excretory canal, arrowheads designate the HSN axon paths, the scale bars represent 5 μm, and anterior is to the left and ventral is down. The CAN neuron is in a distinct focal plane and is thus undetectable. The GFP images were pseudocolored green.

Fig. 2. UNC-86 is required for the ventral growth of HSN axons

A: The proportion of left HSNs with detectable axon outgrowth, visualized with the lin-4∷GFP(zaIs1) reporter strain. n>50 for each genotype. B: The percentage of HSNL axons scored in (A) extending in the direction indicated, and representing the first direction in which axons extended a distance of 1.5 anterior/posterior cell diameters. ***: p<0.0001 when comparing proportions between wild-type and the indicated mutants, calculated using the two-sample z-test. For (A) and (B), error bars represent standard error of proportion. C: Representative images of a single unc-86(n846) (upper two panels) or unc-86(e1416) (bottom panel) L4-stage HSN with ventral guidance defects. Cell somas were often directly adjacent to the excretory canal, a more dorsal position than in wild-type (top panels; data not shown). For unc-86(n846), images of the same animal were acquired using a GFP bandpass filter (upper image) or DIC (merged with GFP, lower image). Arrows point to the excretory canal, just ventral to the CAN neuron. Arrowheads designate the HSN axon paths, the scale bars represent 5 μm, and anterior is to the left and ventral is down. GFP images for unc-86(n846) were pseudocolored green.

Fig. 3. unc-86 controls the timing and guidance of HSN axon outgrowth

A-B: Representative wild-type HSN (A) or unc-86(n846) (B) expressing the unc-86∷myr-GFP reporter at the L1 stage. No axon has extended in wild-type, while unc-86(n846) exhibits precocious outgrowth. For (A) and (B), scale bars represent 5 μm, and anterior is to the left and ventral is down. C-D: The proportion of wild-type and unc-86 mutant animals with axon outgrowth in mixed L1 populations using unc-86∷myr-GFP (myr-GFP) (C) or lin-4∷GFP (GFP) (D) reporter strains. n>50 for all genotypes. E: The proportion of HSN axons scored in (C-D) with extension along the anterior/posterior (A/P) or dorsal/ventral (D/V) axes for the indicated genetic backgrounds. Assigned axes represent the first direction in which axons extended a distance of 1.5 anterior/posterior cell diameters. For C-E, error bars represent standard error of proportion, and ***: p<0.0001 for the difference between wild-type and unc-86(n846) (C-D) or A/P and D/V axon outgrowth (E), calculated using the two-sample z-test.

Fig. 4. The UNC-40/DCC receptor is expressed and membrane-targeted in unc-86(n846)

A-B: The proportion of wild-type or unc-86(n846) animals with detectable unc-40∷SL2∷YFP expression in mixed populations at each larval stage (A: Line 1c; B: Line 6). Error bars represent standard error of proportion, and for each condition, n≥50 (A) or ≥25 (B). ***: p<0.0001. **: p=0.007. *: p<0.05 (0.043 for L1, 0.035 for L2, and 0.033 for L4). p-values were calculated using the two-sample z-test. C: For Line 6, representative L4-stage HSNs from wild-type (upper row) and unc-86(n846) (lower row) animals. Images of the same animal were acquired using the Yellow GFP (YFP) filter or DIC, and YFP-expressing cells were pseudocolored green. Left: YFP; Right: YFP/DIC merge. Arrowheads indicate the CAN neuron, while arrows refer to the HSN. Note that in unc-86(n846), the HSN cell body is dorsally displaced relative to the developing vulva, reflecting a failure in ventral migration. D: A strain expressing the UNC-40∷GFP translational fusion construct was crossed into unc-86(n846) and scored for membrane localization. Wild-type and unc-86(n846) L1-stage animals display UNC-40 membrane targeting with no preferential ventral localization (arrowhead). In the L2, UNC-40 has accumulated primarily at the ventral surface (arrowhead) in wild-type animals. In L2-stage unc-86(n846) mutants, UNC-40 is distributed throughout the precociously extending axon (arrowheads), with preferential accumulation in the ventral/posterior cell soma and a posterior membrane protrusion (arrow). For C and D, scale bars represent 5 μm, and anterior is to the left and ventral is down.

Fig. 5. Precocious HSN axon initiation in unc-86(lf) is not dependent on signaling through the UNC-40/DCC or SAX-3/Robo guidance receptors

A: The proportion of L1-stage HSN axon outgrowth in unc-86(n846) animals that also contained a mutation in either unc-6/netrin or its receptor unc-40/DCC. n≥50 for all three genotypes. B: The proportion of axons in (A) with initial outgrowth along the A/P or D/V axes for the genotypes indicated. C: The proportion of L1-stage HSN axon outgrowth in unc-86(n846); sax-3(ky123) double mutant animals (n=30). D: The proportion of HSN axons in (C) with initial extension along the A/P versus D/V axes. For (A)-(D), axons were visualized using the unc-86∷myr-GFP reporter strain, and error bars represent standard error of proportion. For (B) and (D), assigned axes reflect the first direction in which axons extended a distance of 1.5 anterior/posterior cell diameters. ***: p<0.0001 for the difference in the percentage of HSNs extending axons along each axis, calculated using the two-sample z-test.

Fig. 6. unc-86(n846) suppresses the delay in HSN axon outgrowth observed in loss-of-function mutants for the lin-4 microRNA

A: The percentage of unc-86(n846); lin-4(e912) animals that extended axons during the L1 stage. Axons were visualized with the unc-86∷myr-GFP reporter (designated myr-GFP) and lin-4∷GFP (designated GFP). Error bars represent standard error of proportion. n>50 B: GFP expression in the HSNs normalized to L1-stage values for two unc-86 transcriptional fusion strains: unc-86∷myr-GFP containing the unc-54 3’UTR and unc-86∷GFP Line 2 containing the endogenous unc-86 3’UTR. No statistically significant differences in relative mean pixel intensity were observed between the two strains at the L2, L3, and L4 larval stages. C: Expression of the lin-14∷GFP(zaIs2) reporter strain in a wild-type or unc-86(n846) mutant background at the L1 and L2 larval stages. The mean pixel intensity values were not statistically different for wild-type and unc-86(n846) at both stages, but significant down-regulation between L1 and L2 was observed for each strain (**: p=0.002). D: Expression of lin-28∷GFP Line 10-2 in a wild-type or unc-86(n846) mutant background. While no difference was observed between wild-type and unc-86(n846) values at L1 or L3, each background exhibited significant down-regulation by the L3 (***: p<0.0001 for wild-type and p=0.0002 for unc-86(n846)). For A, n>50 for each genetic background. For B, C, and D, n≥10 for each time point and error bars represent standard error of the mean. p-values were calculated using the two-sample t-test.