A polarity pathway for exocyst-dependent intracellular tube extension

Lumen extension in intracellular tubes can occur when vesicles fuse with an invading apical membrane. Within the Caenorhabditis elegans excretory cell, which forms an intracellular tube, the exocyst vesicle-tethering complex is enriched at the lumenal membrane and is required for its outgrowth, suggesting that exocyst-targeted vesicles extend the lumen. Here, we identify a pathway that promotes intracellular tube extension by enriching the exocyst at the lumenal membrane. We show that PAR-6 and PKC-3/aPKC concentrate at the lumenal membrane and promote lumen extension. Using acute protein depletion, we find that PAR-6 is required for exocyst membrane recruitment, whereas PAR-3, which can recruit the exocyst in mammals, appears dispensable for exocyst localization and lumen extension. Finally, we show that CDC-42 and RhoGEF EXC-5/FGD regulate lumen extension by recruiting PAR-6 and PKC-3 to the lumenal membrane. Our findings reveal a pathway that connects CDC-42, PAR proteins, and the exocyst to extend intracellular tubes.


Introduction
Most organs contain tubes, which are used to transport gases and fluids from one site within the body to another. The circumference of larger tubes, such as the human intestine, is lined by many cells connected to one another with junctions. By contrast, the smallest tubes have intracellular lumens that are contained entirely within the cytoplasm of a cell. Although some intracellular tubes arise when a cell wraps circumferentially and recontacts itself to hollow out a lumen from the extracellular space (Rasmussen et al., 2008;Stone et al., 2009), many intracellular tubes are thought to form when an apical membrane domain invades into the cytoplasm to become the lumen (Lubarsky and Krasnow, 2003;Sundaram and Cohen, 2017). The Caenorhabditis elegans excretory cell provides a powerful model system for studying this mechanism of intracellular lumen extension. Born during the first half of embryogenesis, the H-shaped excretory cell contains four long canal arms that grow during larval stages to extend nearly the full length of the worm by the beginning of the L2 larval stage (Nelson et al., 1983;Sundaram and Buechner, 2016). An intracellular lumen initiates within the cell body and invades the length of each canal arm, functioning in osmoregulation (Buechner et al., 1999;Mancuso et al., 2012;Nelson and Riddle, 1984;Sundaram and Buechner, 2016). Vertebrate capillaries, as well as terminal and fusion cells of the Drosophila trachea and the Ciona notochord, are additional examples of cells containing intracellular tubes that are thought to form through an apical invasion mechanism (Denker et al., 2013;Gervais and Casanova, 2010;Herwig et al., 2011;Lenard et al., 2013).
Extension of an intracellular lumen by apical domain invasion requires the polarized delivery and fusion of vesicles, which supply the new membrane needed to expand the lumenal surface (Berry et al., 2003;Gervais and Casanova, 2010;Khan et al., 2013;Kolotuev et al., 2013;Schottenfeld-Roames and Ghabrial, 2012). The highly conserved, eight-protein exocyst complex and the small GTPase exocyst activator Ral are required for polarized membrane targeting of vesicles in Here, we utilize degron-tagged alleles of SEC-5, RAL-1, PAR-3, PAR-6, PKC-3, CDC-42, and EXC-5 to establish the roles of these proteins in extending the excretory cell intracellular lumen. We show that PAR-6 and PKC-3, but not PAR-3, are essential for lumen extension, and using acute protein depletion we demonstrate that PAR-6, but not PAR-3, is needed to recruit the exocyst to the lumenal membrane. Finally, we provide evidence that EXC-5 and CDC-42 function upstream of PAR-6 and PKC-3 as polarity cues, recruiting these proteins to the lumenal membrane. Our findings identify a pathway that connects Rho GTPase, cell polarity, and vesicle-tethering proteins to lumen extension during intracellular tubulogenesis.

SEC-5 and RAL-1 function within the excretory cell to promote lumen extension
The enrichment of the exocyst at the excretory cell lumenal membrane and its requirement for proper lumen extension suggest that exocyst-dependent vesicle delivery provides the new membrane needed for lumen expansion (Armenti et al., 2014a). If so, the exocyst, which is broadly expressed and needed for embryonic development (Armenti et al., 2014a;Frische et al., 2007), should be required autonomously within the excretory cell. To test this hypothesis, we designed a degron-based strategy to conditionally deplete exocyst component SEC-5 and exocyst activator RAL-1 (the sole C. elegans Ral GTPase homologue) specifically within the excretory cell ( Figure 1A); this approach removes zygotically expressed protein as well as inherited maternal protein, which can otherwise mask mutant phenotypes (Nance and Frøkjaer-Jensen, 2019). Proteins tagged with the ZF1 degron are rapidly degraded to undetectable levels by expressing the E3 ubiquitin ligase substrate-adapter protein ZIF-1 DeRenzo et al., 2003;Reese et al., 2000). In order to express ZIF-1 specifically within the excretory cell, we searched for an excretory cell-specific promoter. Existing transcriptional reporters for two promoters described to be active predominantly or exclusively in the excretory cell, pgp-12 (Zhao et al., 2005) and glt-3 (Mano et al., 2007), showed additional expression in other embryonic tissues. Using the WormBase (https://wormbase.org/) data-mining platform WormMine, we identified additional candidate promoters among a set of genes described to be expressed specifically within the excretory cell. Upstream sequences of one gene, T28H11.8, drove detectable mCherry expression specifically in the excretory cell from embryogenesis onward (Figure 1-figure supplement 1), and endogenous T28H11.8 mRNA is first detected by single-cell RNA sequencing in the excretory cell several hours after its birth (Packer et al., 2019). To determine if ZIF-1 expressed from the T28H11.8 promoter (hereafter excP) was sufficient to degrade ZF1-tagged proteins specifically within the excretory cell, we introduced a high-copy array containing excP::zif-1 into worms expressing a ZF1-tagged reporter protein, ZF1:: GFP::CDC-42. Control larvae, which did not inherit the excP::zif-1 array, robustly expressed ZF1:: GFP::CDC-42 in the excretory cell and other tissues ( Figure 1B). By contrast, ZF1::GFP::CDC-42 was depleted below detectable levels within the excretory cell in larvae that inherited the excP::zif-1 transgenic array ( Figure 1C), whereas expression of ZF1::GFP::CDC-42 persisted in other tissues. We conclude that excP::zif-1 can be used to deplete ZF1-tagged proteins from the excretory cell.
We measured canal length by examining where the posterior canal lumens ended relative to body length in L1 and L4 larvae, as these stages represent active outgrowth (L1) and maintenance (L4) of the canal lumen. Dividing the body into quartiles along its anterior-posterior axis, nearly all control larvae extended canals to the third quartile (51-75% of body length) at the L1 stage and the fourth quartile (76-100% of body length) by the L4 stage ( Figure 2). However, in both SEC-5 exc (-) and RAL-1 exc(-) larvae, canal lumen length was significantly reduced at both L1 and L4 stages, with nearly all larvae containing canal lumens that extended to less than 50% body length ( Figure 2). The Schematics of L4 larval stage worms depicting excretory cellspecific protein depletion using excP::zif-1. The H-shaped excretory canal is outlined and a hypothetical ubiquitous ZF1-tagged protein is depicted in green. The typical region of the canal examined by microscopy is enlarged to show cytoplasmic (yellow, excP::YFP) and lumenal membrane (cyan, IFB-1::CFP) markers used for analyzing excretory canal morphology. Anterior left, dorsal top. (B and C) L4 stage excretory canal in transgenic control (B) and excP::zif-1 (C) animals expressing ZF1::GFP::CDC-42. Outline of excretory canal cytoplasm is indicated by dotted line. ZF1::GFP::CDC-42 is degraded in the excretory cell, but not surrounding cells (arrowhead), in excP::zif-1 animals. (D) Endogenous expression of SEC-5::ZF1::YFP at the excretory canal lumenal membrane of L4 stage larva. (E-J'') Larval excretory canal phenotypes in control (E-F''), SEC-5 exc(-) (G-H''), and RAL-1 exc(-) (I-J''). Canal cytoplasm and lumenal membrane are marked by an extrachromosomal array expressing excretory cell-specific cytoplasmic and lumenal membrane markers (see panel A). Confocal images were acquired using Â20 (E, G, I) and Â63 objectives (F-F'', H-H'', J-J''). Excretory cell body indicated by asterisk. Posterior tip of excretory canal indicated by white arrow. Posterior excretory canal that has extended beyond the focal plane is indicated by dashed white arrow. Dashed box indicates approximate region represented in high magnification images. Outline of each animal is indicated by solid white line. Scale bars, 10 mm. The online version of this article includes the following figure supplement(s) for figure 1: Abrams and Nance. eLife 2021;10:e65169. DOI: https://doi.org/10.7554/eLife.65169 canal lumen length defect of SEC-5 exc(-) larvae did not become more severe when we replaced one sec-5(xn51: sec-5::zf1::yfp) allele with the sec-5(tm1443) predicted null allele (Frische et al., 2007; Figure 2-figure supplement 1), suggesting that SEC-5 exc(-) phenotypes result from nearly complete or complete loss of SEC-5 protein once the excP::zif-1 transgene is expressed. Together, these data indicate that exocyst activity within the excretory cell is needed for proper organization and extension of its intracellular lumen.
The localization of CDC-42 within the excretory cell has only been described using high-copy transgenes and heterologous promoters (Lant et al., 2015;Mattingly and Buechner, 2011), and the high-copy transgene expressing ZF1::GFP::CDC-42 that we used to test the efficacy of excP::zif-1 ( Figure 1B; Armenti et al., 2014b). We examined CDC-42 subcellular localization in the excretory cell using a functional endogenously tagged zf1::yfp::cdc-42 allele (Zilberman et al., 2017). ZF1:: YFP::CDC-42 protein was expressed in the excretory cell and showed a broader distribution than PAR-6::mKate ( Figure 3G-G''). ZF1::YFP::CDC-42 extended well into the excretory cell cytoplasm compared to endogenously expressed PAR-6::mKate present within the same animal ( Figure 3D,G, H), whereas endogenously tagged PAR-6::ZF1::YFP and PAR-3::mCherry showed a similar . Canal outgrowth phenotypes upon exocyst or PAR protein depletion. Schematics of the excretory cell are shown at the L1 stage, when the canal is extending, and the L4 larval stage, when the canal is fully extended. Canal outgrowth defects upon depleting the indicated proteins in the excretory cell are depicted as the percentage of animals in each of four phenotypic categories (quartiles) that measure posterior canal extension relative to body length. The relative intensity of green shading reflects the percentage of larvae observed in each phenotypic category. p values were calculated using Fisher's exact test after pooling quartiles and comparing each genotype to the control group (L1 stage:<50% versus>50% canal outgrowth; L4 stage:<75% versus>75% canal outgrowth). p value significance was adjusted using Bonferroni correction to account for multiple comparisons to a common control, such that p 0.008 is considered statistically significant.  enrichment to the lumenal membrane ( Figure 3E-F). While the peak localization intensities of ZF1:: YFP::CDC-42 and PAR-6::mKate in transects across the width of the excretory cell do not align, as they do with PAR-6::ZF1::YFP and PAR-3::mCherry, super-resolution imaging would be required to determine whether ZF1::YFP::CDC-42 is present at the lumenal domain. Therefore, consistent with previous findings made using immunostaining and transgenes (Armenti et al., 2014a), endogenously tagged PAR-3, PAR-6, and PKC-3 are each expressed within the excretory cell and are present at the lumenal membrane, and CDC-42 is expressed more broadly within the cytoplasm.
PAR-6, PKC-3, and CDC-42 are required in the excretory cell for lumen extension To determine if PAR proteins and CDC-42 are required within the excretory cell for lumen extension, we crossed excP::zif-1 with each par or cdc-42 knock-in allele and examined excretory canal morphology using cytoplasmic and lumenal membrane markers (see Figure 1A). PAR-6 exc(-) and PKC-3 exc(-) L4 stage larvae had severely truncated canals with dilated and cystic lumens ( Figure   and L4 stages ( Figure 2). Unexpectedly, PAR-3 exc(-) larvae had a distinct and comparatively mild phenotype. At the L1 stage, canal lumens in PAR-3 exc(-) larvae had an irregular diameter (Figure 4-figure supplement 2), and were significantly shorter than controls ( Figure 2). However, by the L4 stage, the canals of PAR-3 exc(-) larvae resembled those of controls ( Figure 4G-H'') and were not significantly shorter ( Figure 2). Although the phenotype of PAR-3 exc(-) larvae appears distinct, more subtle differences in excretory canal length following the depletion of specific proteins might reflect variation in degradation rates or efficiency (Nance and Frøkjaer-Jensen, 2019). All together, these findings suggest that PAR-6, PKC-3, and CDC-42 function within the excretory cell to promote extension of the lumen. PAR-3 is likely only important for lumen outgrowth during early stages, although we cannot exclude the possibility that an undescribed isoform of par-3 with a different 3' end, and thus lacking the ZF1 tag, is expressed within the excretory cell and buffers mutant phenotypes. Our findings also show that, in addition to promoting lumen extension, CDC-42 functions to prevent canal arms from bifurcating.

PAR-6, but not PAR-3, is required for exocyst lumenal membrane localization
The results above suggest that exocyst function or localization may require PAR-6, PKC-3, and CDC-42, but not PAR-3. To determine if PAR proteins regulate lumen extension by recruiting exocyst to the lumenal membrane, we acutely degraded PAR-6::ZF1::YFP and PAR-3::ZF1::YFP protein at the L4 larval stage, after canal growth was complete, by expressing ZIF-1 from a heat-shock promoter. This approach allowed us to analyze exocyst localization in anatomically normal canals, immediately after rapid PAR protein depletion ( Figure 5A). Following a 30 minute heat shock to induce ZIF-1 expression at the L4 stage, PAR-6::ZF1::YFP degraded rapidly within 1 hour ( Figure 5B Figure 5B',D). After PAR-6::ZF1::YFP degraded, mCherry::SEC-10 was no longer enriched at the lumenal membrane, but instead, appeared evenly distributed throughout the cytoplasm ( Figure 5C',E). We quantified these changes in localization by comparing mCherry::SEC-10 intensity along the lumenal membrane to that within the adjacent cytoplasm by generating a lumen/cytoplasm intensity ratio ( Figure 5A), which was significantly reduced in PAR-6depleted larvae ( Figure 5F). We performed analogous experiments to determine the role of PAR-3 in exocyst localization. In contrast to PAR-6::ZF1::YFP depletion, loss of PAR-3::ZF1::YFP did not decrease the enrichment of mCherry::SEC-10 at the lumenal membrane, despite a lack of visible PAR-3::ZF1::YFP protein following ZIF-1 induction ( Figure 5G-K). We conclude that PAR-6 is required to enrich the exocyst complex at the lumenal membrane, whereas PAR-3 is likely dispensable for exocyst lumenal membrane enrichment.

PAR-3 promotes PAR-6 lumenal membrane localization
In many polarized cell types, PAR-3 helps enrich PAR-6 at the membrane (Nance and Zallen, 2011;St Johnston and Ahringer, 2010). Therefore, the requirement for PAR-6, but not PAR-3, in mCherry::SEC-10 lumenal membrane enrichment was surprising. To investigate the epistatic relationship between PAR-3 and PAR-6 within the excretory cell, we first expressed ZIF-1 from a heat shock promoter and degraded PAR-3::ZF1::YFP after canal growth was complete ( Figure 6A-B). Surprisingly, endogenously tagged PAR-6::mKate (Dickinson et al., 2017) was significantly less enriched at the lumenal membrane and increased within the cytoplasm after depletion of PAR-3::ZF1::YFP when compared to control larvae ( Figure 6A'-E), although some puncta of PAR-6::mKate remained at the lumenal membrane ( Figure 6B', arrowheads). In reciprocal experiments, we degraded PAR-6::ZF1:: YFP by expressing ZIF-1 from a heat shock promoter and examined endogenously tagged PAR-3:: mCherry localization. PAR-3::mCherry remained enriched at the lumenal membrane in PAR-6depleted L4 worms, and unexpectedly, its lumen/cytoplasm ratio was significantly increased ( Figure 6F-J). We propose that PAR-3 is required to recruit most PAR-6 to the lumenal membrane, but that the PAR-6 puncta remaining after PAR-3::ZF1::YFP depletion are sufficient to recruit the exocyst to the lumenal membrane (see Discussion). In addition, these findings show that PAR-6 limits PAR-3 lumenal membrane enrichment. We next asked what other factors act upstream to regulate the lumenal membrane enrichment of PAR-6 and PKC-3 within the excretory cell. One candidate is CDC-42, which binds to the PAR-6 CRIB domain and can recruit PAR-6 to the membrane in parallel to PAR-3 in the one-cell C. elegans embryo (Aceto et al., 2006;Beers and Kemphues, 2006;Gotta et al., 2001;Joberty et al., 2000;Kay and Hunter, 2001;Rodriguez et al., 2017;Wang et al., 2017). CDC-42 exc(-) and PAR-6 exc(-) larvae displayed a similar canal outgrowth phenotype (Figure 2), consistent with these two proteins acting in the same lumen extension pathway within the excretory cell. To determine if CDC-42 is required for PAR-6 enrichment at the lumenal membrane, we acutely degraded ZF1::YFP::CDC-42 by heat shock expression of ZIF-1 in L4 larvae. PAR-6::mKate lumenal membrane enrichment was significantly decreased after loss of CDC-42 ( Figure 7A-E). Together, these results suggest that CDC-42 promotes lumen extension by helping to enrich PAR-6 at the lumenal membrane.

Discussion
An intracellular lumenogenesis pathway bridging Rho GTPase, cell polarization, and vesicle-tethering proteins During tubulogenesis within the C. elegans excretory cell, it has been proposed that the docking and subsequent fusion of large 'canalicular' vesicles at the lumenal membrane domain provides the membrane needed for tube extension (Khan et al., 2013;Kolotuev et al., 2013). We showed  previously that exocyst complex activity is required for canalicular vesicles to connect with the lumenal membrane domain and for normal lumen extension to occur (Armenti et al., 2014a). Here, based on cell-specific protein depletion experiments during lumen extension, and protein localization analysis following acute protein degradation in fully developed excretory cells, we propose a pathway for lumen extension ( Figure 7K). Most upstream, RhoGEF EXC-5 at the lumenal membrane activates the Rho GTPase CDC-42. Although EXC-5 has been proposed previously as an activator of CDC-42 at the lumenal membrane (Mattingly and Buechner, 2011;Shaye and Greenwald, 2016), our findings show for the first time that its depletion causes a similar molecular defect as depletion of CDC-42 (loss of PKC-3 or PAR-6 from the lumenal membrane). Downstream of EXC-5, we propose that active CDC-42 recruits PAR-6 and PKC-3 through interactions with the PAR-6 CRIB domain. In turn, PAR-6 and PKC-3 function to recruit the exocyst. RAL-1 has previously been shown to promote exocyst membrane localization, including in the early C. elegans embryo (Armenti et al., 2014a). The strong phenotypes we observe in RAL-1 exc(-) larvae suggest that RAL-1 has a similar function within the excretory cell. Although PAR-6 and PKC-3 bind one another and are typically thought to function as an obligate pair, we note that our experiments do not directly address whether they function together in lumen extension. In addition, further experiments will be required to determine whether EXC-5 activates CDC-42 specifically at the lumenal membrane, as our model predicts, and to identify the biochemical links between EXC-5, CDC-42, PAR-6, PKC-3, and the exocyst complex.
Even though lumen extension is severely compromised in SEC-5 exc(-) , RAL-1 exc(-) , PAR-6 exc(-) , PKC-3 exc(-) , and CDC-42 exc(-) larvae, the initial stages of lumenogenesis still occur. One possible explanation is that a distinct pathway directs the initial stages of lumen formation. Alternatively, since it is unclear whether the excP::zif-1 transgene is active at the very early stages of lumenogenesis (see Results), it is possible that complete loss of the targeted proteins immediately after excretory cell birth would block lumen formation entirely. Finally, it is possible that degradation of the targeted ZF1-tagged proteins, while visibly below our level of detection by fluorescence, is not complete and phenotypes are hypomorphic. Resolving these possibilities will require the use of earlier-acting zif-1 drivers and alternative genetic methods. (E) Quantification of lumenal membrane to cytoplasm intensity ratio of PAR-6::mKate expression in the excretory canal of control and hspP::zif-1 larvae expressing ZF1::YFP::CDC-42. Individual data points (small dots) are color-coded (orange, purple, and light blue) from three independent replicates. Large dots represent the mean of each replicate, horizontal bar is the mean of means, and error bars are the SEM. p values were calculated using a ratio paired t-test of the means. n = 8, 7, 7 for control; n = 9, 7, 8 for hspP::zif-1. (F-G) Distribution of EXC-5::ZF1::mScarlet in the larval excretory canal in control (F) and hspP::zif-1 (G) worms. (F'-G') Distribution of GFP:: PKC-3 in the larval excretory canal of control (F') and hspP::zif-1 (G') worms expressing EXC-5::ZF1::mScarlet. (H-I) Line trace of GFP::PKC-3 (green) and EXC-5::ZF1::mScarlet (magenta). Intensities were normalized to compare peak values of each channel. '0.0' on x-axis represents the center point of the canal lumen. n = 5 larvae. (J) Quantification of lumenal membrane to cytoplasm intensity ratio of GFP::PKC-3 expression in the excretory canal of control and hspP::zif-1 larvae expressing EXC-5::ZF1::mScarlet. Data are depicted as in panel E. p values were calculated using a ratio paired t-test of the means. n = 5, 6, 6 for control; n = 5, 5, 6 for hspP::zif-1. Although we found that in PAR-3-depleted larvae, most PAR-6 was lost from the excretory cell lumenal membrane -a phenotype that could be predicted based on previous studies of PAR-3 in other cell types -the relatively mild lumen extension phenotype of PAR-3 exc(-) larvae (shortened canals in the L1 stage that recovered to normal length by the L4 stage) and lack of requirement for PAR-3 in mCherry::SEC-10 localization were somewhat surprising. Recently, using auxin-inducible protein degradation, it was shown that PAR-3 is not essential for C. elegans larval development, in contrast to PAR-6 and PKC-3 (Castiglioni et al., 2020). Although further experiments will be needed to determine if an alternative form of PAR-3 protein lacking the ZF1 degron is produced, we consider this unlikely, as no such isoforms have been described, and the loss of PAR-6 at the lumenal membrane suggests that PAR-3 depletion was effective. Instead, we favor the hypothesis that PAR-3 makes lumen extension more efficient by augmenting PAR-6 lumenal enrichment, and that partial PAR-6 recruitment by CDC-42 is sufficient for lumen extension. Studies in the zygote have shown that in addition to localizing PAR-6 and PKC-3 to the membrane, CDC-42 also promotes PKC-3 activity (Rodriguez et al., 2017), raising the possibility that it plays a more consequential role during lumen extension than PAR-3 by both localizing and activating the PAR-6/PKC-3 complex. Such a relationship between PAR-3 and CDC-42 in recruiting PAR-6 likely occurs in additional cell types, as PAR-3 depletion in the epidermis causes PAR-6 mislocalization but not the junction defects that occur following PAR-6 depletion in the same cells (Achilleos et al., 2010). While it is not yet clear why PAR-3 appears to be more important for lumen extension at earlier larval stages, this is when active lumen outgrowth occurs. A reasonable hypothesis is that partially compromised PAR-6 function (because of reduced enrichment at the lumenal membrane) may be more consequential at this stage of lumenogenesis.
par-6, aPKC, and the exocyst are also required for proper intracellular lumen growth in Drosophila tracheal cells (Jones et al., 2014), suggesting that this pathway may function as a general mechanism promoting intracellular tube extension. Notably, and consistent with our findings in the C. elegans excretory cell, mutations in Drosophila baz (par-3) do not prevent tracheal lumen extension, suggesting that in both cell types PAR-6 and PKC-3/aPKC perform the major role in exocyst regulation. PAR proteins and the exocyst are also required for organized lumen expansion in mammalian cell cysts grown in 3D culture (Bryant et al., 2010). Thus PAR-mediated exocyst recruitment to sites of lumen expansion, where additional membrane is needed, appears to be a feature common to both intracellular and multicellular tubes despite their dramatically different organization.

Exocyst recruitment by PAR proteins
Together with previous studies, our findings suggest that PAR proteins and the exocyst may interface in multiple ways. In mammary epithelial cells, Par3 functions as an exocyst receptor, utilizing a lysine-rich domain to bind Exo70 and recruit the complex (Ahmed and Macara, 2017). However, in these cells, the exocyst also mediates membrane fusion at the basal membrane, where Par3 is not detected, suggesting that alternative exocyst receptors exist (Ahmed et al., 2018). Biochemical studies have also revealed interactions between the exocyst, PAR-6, and aPKC. For example, coimmunoprecipitation experiments in cultured rat kidney epithelial cells and in cortical neurons showed that aPKC immunoprecipitates with the exocyst proteins Sec8, Sec6, or Exo84 (Lalli, 2009;Rosse et al., 2009). Furthermore, Par6 can directly bind Exo84 in cultured mammalian neurons, and this interaction is promoted by the RAL-1 homologue RalA (Das et al., 2014). Finally, in rat kidney epithelial cells, aPKC helps recruit exocyst through the aPKC-interacting protein Kibra (Rosse et al., 2009). Together with these studies, our finding that PAR-6 but not PAR-3 is required to recruit SEC-10 to the lumenal membrane suggests that PAR-6 functions as an alternative means to recruit the exocyst complex to the membrane. Further studies will be needed to clarify whether it does so directly by functioning as an exocyst receptor, analogous to mammalian Par3 (Ahmed and Macara, 2017), or indirectly, for example through the kinase activity of aPKC. Because aPKC and Par6 localize interdependently in nearly all cell types examined, the fact that PKC-3 exc(-) and PAR-6 exc(-) larvae have similar lumen extension defects does not clarify how PKC-3 contributes to exocyst recruitment. Notably, C. elegans lacks a clear Kibra orthologue (Yoshihama et al., 2012), suggesting that if PKC-3 interfaces with the exocyst directly, it does so utilizing a distinct mechanism.   This study FT1692 Shown in Figure 1C, Figure 1- This study FT1834 FT1523 crossed to FT1837 Shown in Figure 1G  Continued on next page This study FT1866 Shown in Figure 1I-J'', Figure 2 Strain, strain background (C. elegans) pkc-3(xn84[zf1:: gfp::pkc-3]); xnIs547; xnEx466

Transgene integration
pJA027 (t28h11.8p::zif-1), which contains an unc-119(+) transformation marker, was injected into unc-119(ed3) worms to obtain a stably inherited, high-copy extrachromosomal array. The array was integrated using Trioxsalen (Sigma) and UV irradiation. A mixed population of washed transgenic worms was incubated in 600 ml of 33.3 ng/ml Trioxsalen in DMSO in the dark for 15 min. Worms were dripped onto an unseeded NGM agar plate and, after the solution soaked in, the agar plate was irradiated with 360 mJ of UV light in a Stratalinker. NA22 bacterial food was dripped onto the worms and, after 5 hr in darkness, 20 L4 stage transgenic worms were picked to each of 20 peptone plates (10 cm) seeded with NA22 bacteria. F1 adults were bleached to collect eggs, which were plated 200 per plate onto 70 NGM plates (6 cm). Nine hundred eighty-four transgenic F2s were picked into individual wells of 24-well plates, and those with an F3 brood containing only non-Unc progeny were saved. Transgenic insertion xnIs547 was isolated and outcrossed three times to unc-119(ed3).

Imaging
For all live-imaging experiments, larvae were mounted onto 5% agarose pads in a 2 mM Levamisole solution in M9 buffer to induce paralysis. Fluorescent images were acquired using an SP8 confocal microscope (Leica), 63 Â 1.4 NA oil-immersion objective, 458, 488-, 514-, 561 nm lasers, and 1-5x zoom. For intensity measurements, larvae were imaged using HyD detectors and the photon-counting mode. Images were analyzed and processed in ImageJ (NIH) with no g adjustments and level adjustments across pixels. For quantifications, the same laser power and exposure times were used within experiments and control and mutant images were processed similarly. After processing in ImageJ, images were rotated and cropped using Illustrator (CC2020, Adobe).

Heat-shock expression of ZIF-1
Plates containing late L4/young adult animals were placed in a water bath at 34˚C for 30 min and then transferred to 15˚C to recover. In each experiment, control and experimental animals were imaged 2-4 hr following heat shock.
Excretory canal outgrowth measurements  , RAL-1 exc(-) , PKC-3 exc(-) , PAR-6 exc(-) , CDC-42 exc(-) , and PAR-3 exc(-) strains were all homozygous viable when grown on NGM plates. Excretory canal length was scored visually using a canal-specific cytoplasmic marker (t28h11.8p::yfp) at L1 and L4 larval stages. Both posterior canal arms were scored in each animal. In cases where the canal arms differed in length, an approximate average of the two lengths was recorded for that animal.

Image analysis
All measurements were performed using ImageJ and raw SP8 confocal image files. For lumen/cytoplasm intensity measurements, a line four pixels in width was drawn along the lumenal membrane and a second line was drawn along an adjacent region within the canal cytoplasm, as shown in Figure 5A. Mean pixel intensity values along each line were calculated using the ImageJ measuring tool. Both faces of the lumenal membrane were measured in each image and two images were acquired of different regions of the posterior canal arms within each animal. Four such measurements were taken for each animal and an average 'lumen/cytoplasm intensity ratio' was calculated, which is represented by small colored dots in plots in Figures 5F, K, 6E, J, 7E and J.
For intensity profiles of the excretory canal, a line 30 pixels in width was drawn across a 3 mm region of the excretory canal cytoplasm, as shown in Figure 3D. Three measurements were acquired for each animal and averaged to generate a single intensity profile per animal. Measurements from five animals are shown in each graph. Values were copied into GraphPad Prism 8 to generate an XY line plot displaying the average and standard deviation.
To measure excretory canal fluorescence intensity after ZIF-1 degradation, the polygon tool in ImageJ was used to draw a region of interest (ROI) around the canal cytoplasm using the CFP cytoplasm marker. Mean pixel intensity values within each polygon were calculated using the ImageJ measuring tool. To measure degradation, fluorescent intensity of PAR-6::ZF1::YFP was calculated in control and hspP::zif-1 animals 2 hr after a 30 min heat shock at 34˚C. Two images were acquired of different regions of the posterior canal arms of each animal and averaged. Background YFP autofluorescence was calculated in wild type larvae carrying the pgp-12p::mCherry transgene to mark canal cytoplasm. Average background autofluorescence was subtracted from control and hspP::zif-1 animals prior to calculating percent of YFP depletion. Error bars represent standard deviation, and were calculated from the change in mean fluorescence intensity between control and experimental animals.
For plotting image quantification and statistical analysis, mean values for each animal and each biological replicate were copied to GraphPad Prism 8. SuperPlots were generated in GraphPad Prism 8 as previously described (Lord et al., 2020), with dots of the same color representing individual data points from the same experiment.

Statistics
Statistical analysis was performed in GraphPad Prism 8. Statistical tests, number of embryos, and number of experiments are indicated in the figure legends. No statistical tests were used to predetermine sample size. Animals were selected for measurements based on developmental stage, orientation on the slides, and health. No animals were excluded from analyses post-hoc. Investigators were not blinded to allocation during experiments and outcome assessment.
In Figure 2, data from quartiles was pooled into two categories and Fisher's exact test was then performed (see Figure Legend). Some categories (i.e. quartiles) contained small numbers (<10 larvae) which can cause the p value to be inaccurate for a test of independence and therefore pooling categories is appropriate in this instance (McDonald, 2014). Where multiple comparisons were made to a common control, p values were corrected using the Bonferroni method.