The unfolded protein response is required for dendrite morphogenesis

1) One weakness of the study concerns the sufficiency of dma-1 triggered UPR in PDE morphogenesis (Figure 7). The authors show that overexpression of dma-1 and hsp-4 can cause abnormal shape of PDE. However, the data do not convincingly or directly show that this is due to changes in UPR or ER folding capacity. Does this ectopic expression induce changes in the XBP-1 reporter? Is the degree of ectopic branching correlating with dma-1 protein folding or protein localization? Simply based on the data, the PDE morphological changes can also be due to signaling induced by ectopic dma-1 expression. Since the authors find kpc-1 and hpo-30 are not responsible for triggering UPR in PVD, they should include ectopic expression of either gene as control for the specificity of dma-1.

To answer if dma-1 expression in PDE turns on UPR, we tested the UPR activity with the Phsp-4::HIS-24::GFP reporter, which serves as a readout for the transcriptional activation of the HSP-4 chaperone. We found that HIS-24::GFP level is dramatically increased in PDE neurons expressing dma-1, suggesting that dma-1 expression indeed activates UPR (Figure 7–figure supplement 2). In the same experiments, we also attempted to visualize DMA-1 localization with a DMA-1::tagBFP fusion. However, unfortunately, the fluorescence of BFP protein fusion is too dim to clearly delineate the subcellular localization of DMA-1. We further tested the subcellular localization of DMA-1 and UPR activity in an attempt to address the reviewers’ next question.

As the reviewers requested, we also tested if overexpressing hpo-30 with hsp-4 in PDE caused ectopic branching of PVD and found no ectopic branches (Figure 7 panel F). So far, DMA-1 is the only cell surface molecule, which can induce ectopic branches.

2) DMA overexpression in IRE mutants sometimes enhances IRE phenotypes, sometimes suppresses. Can the authors provide any additional insight as to what might determine this outcome?

When we overexpressed dma-1 in the ire-1 mutants, we found that about 60% of the transgenic animals showed a “wt-like” dendrites, while about 25% of the transgenic animals showed enhanced “dma-1-like” dendrites. We hypothesized that the two types of phenotypes are due to variability of animals’ ability to fold overexpressed DMA-1. To directly test this hypothesis, we compared the subcellular localization of DMA-1::GFP between the “wt-like” animals and “dma-1-like” animals separated based on their dendrite morphology. We found that the “dma-1-like” animals showed 15 fold higher level of cell body DMA-1 compared with the “wt-like” animals. The correlation between intense cell body DMA-1::GFP and the enhanced dendrite phenotype is very tightly correlated. These data suggest that animals with slightly higher folding capacity can produce functional DMA-1 and rescue the dendrite morphology phenotypes. In animals with less folding capacity, this folding problem is exacerbated by overexpression of DMA-1, which leads to DMA-1 misfolding and accumulation in the ER.

3) Figure 5: the induction of XBP-1::GFP based on fluorescent intensity seems subtle. The image in L4 (panel E) is hardly different from that of L3. Is this due to poor folding of GFP? The data could be strengthened by including analyses of mRNA splicing of their xbp-1 reporter.

We agree with the reviewers that the XBP-1::GFP only yielded weak fluorescence. In the revised manuscript, we verified this result by using another UPR activity reporter (Phsp-4::HIS-24::GFP). As an ER chaperone, HSP-4 is a transcriptional target of activated XBP-1. We reasoned that this reporter should be more sensitive due to the transcriptional amplification. Indeed, we found that the GFP signal is significant higher but also more variable between animals. Using this reporter, we repeated developmental regulation of UPR. We found that URP was activated in L3. The UPR activity was further increased in L4 animals during which the menorahs form. The UPR activity was dramatically downregulated in adult animals (Figure 5–figure supplement 1). These results are consistent with the original Figure 5.

We did consider the mRNA analysis suggested by the reviewers. This experiment will require us to isolate mRNA specifically from PVD neurons. There are only 2 PVD neurons in each worm. We would need to develop an entirely new method to collect these cells before we can attempt the experiment. Due to the time constraint of resubmission, we decided not to pursue it at this moment. However, as part of our effort to understand the transcriptome of PVD, we are in the process of adopting cell isolation method in C. elegans. We will definitely perform this experiment once the cell isolation system works reliably.

4) It would be appropriate and straightforward for the authors to include additional genetic and transgenic controls for the rescue experiments. Specifically, please include transgene expression alone.

As the reviewers requested, we now provide images for the rescuing transgenes(PVD::ire-1 and PVD::xbp-1 cDNA) in the wild type background (Figure 1–figure supplement 3). We observed that neither of these transgenes altered the dendritic morphology of wt PVD. Overexpressing DMA-1 in the wt background causes overbranching of PVD dendrites and this effect was reported in Liu and Shen, 2012.

5) A tighter connection to cell size seems to be necessary given that the major assertion of the paper is that the UPR is particularly required for neurons with a large dendritic tree. For example, are there genetic manipulations of cell size that are independent of UPR that might be used to suppress the sensitivity to UPR stress in these large neurons?

We thank the reviewers for this insightful suggestion. We tested the reviewers’ hypothesis with the dpy-5 mutants, which have significantly reduced body size compared with wild type controls. As a result of the reduced body size, the PVD dendritic arbor is also significantly smaller in the dpy-5 mutant compared with the wild type. We found that dpy-5 ire-1 double mutants showed a partially suppressed the ire-1 phenotype (Figure 7–figure supplement 1) while dpy-5 single mutant showed normal dendrite patterns. Together, these results support the hypothesis that UPR is particularly required for neurons with large dendritic tree. By reducing the cell size, the burden of UPR is reduced accordingly.