1. Cell Biology
  2. Immunology and Inflammation

Microtubule plus-end dynamics link wound repair to the innate immune response

  1. Clara Taffoni
  2. Shizue Omi
  3. Caroline Huber
  4. Sébastien Mailfert
  5. Mathieu Fallet
  6. Jean-François Rupprecht
  7. Jonathan J Ewbank
  8. Nathalie Pujol  Is a corresponding author
  1. Aix Marseille Univ, INSERM, CNRS, France
  2. Aix Marseille Univ, CNRS, France
https://doi.org/10.7554/eLife.45047
Share this article
Cite this article
  1. Clara Taffoni
  2. Shizue Omi
  3. Caroline Huber
  4. Sébastien Mailfert
  5. Mathieu Fallet
  6. Jean-François Rupprecht
  7. Jonathan J Ewbank
  8. Nathalie Pujol
(2020)
Microtubule plus-end dynamics link wound repair to the innate immune response
eLife 9:e45047.
https://doi.org/10.7554/eLife.45047
  2. Download BibTeX
  3. Download .RIS

Abstract

The skin protects animals from infection and physical damage. In Caenorhabditis elegans, wounding the epidermis triggers an immune reaction and a repair response, but it is not clear how these are coordinated. Previous work implicated the microtubule cytoskeleton in the maintenance of epidermal integrity (Chuang et al., 2016). Here, by establishing a simple wounding system, we show that wounding provokes a reorganisation of plasma membrane subdomains. This is followed by recruitment of the microtubule plus end-binding protein EB1/EBP-2 around the wound and actin ring formation, dependent on ARP2/3 branched actin polymerisation. We show that microtubule dynamics are required for the recruitment and closure of the actin ring, and for the trafficking of the key signalling protein SLC6/SNF-12 toward the injury site. Without SNF-12 recruitment, there is an abrogation of the immune response. Our results suggest that microtubule dynamics coordinate the cytoskeletal changes required for wound repair and the concomitant activation of innate immunity.

Introduction

Inducible immune responses are ubiquitous features of animal defences against infection. In one well-studied example, when the fungus Drechmeria coniospora infects Caenorhabditis elegans, by piercing the cuticle and growing through the underlying epidermis, it triggers an innate immune response, characterized by the induction of expression of a battery of antimicrobial peptide (AMP) genes (Pujol et al., 2008b; Taffoni and Pujol, 2015). These include the genes of the nlp-29 cluster (Pujol et al., 2012), which are principally controlled in a cell autonomous manner (Lee et al., 2018). Sterile wounding also upregulates the expression of the nlp-29 cluster (Pujol et al., 2008a). In both cases, these changes in gene expression result from the activation of the GPCR DCAR-1 that acts upstream of a conserved p38 MAPK signalling cassette (Zugasti et al., 2014). The p38 MAPK PMK-1 itself acts upstream of the STAT transcription factor-like protein STA-2. Although the precise details are as yet unclear, activation of STA-2 is believed to require the SLC6 family protein SNF-12 (Dierking et al., 2011).

Triggering the immune response upon wounding has a prophylactic role and is a safeguard against the potential entry of pathogenic microbes. To ensure longer term protection, a wounded tissue must also be repaired to maintain organismal integrity. Thus, in addition to AMP gene induction, skin wounding in C. elegans also triggers a Ca2+-dependent signalling cascade that promotes wound repair through actin ring closure (Xu and Chisholm, 2011). The nematode skin is mainly a syncytium. Except at the head and tail, a single multinucleate epidermal cell, hyp7, covers the whole animal (Altun and Hall, 2014). Its repair processes exhibit some of the characteristics previously described for single cell wound repair in other organisms (Sonnemann and Bement, 2011). On the other hand, actin ring closure in C. elegans does not rely on myosin-II contractility and the ‘purse string’ mechanism described in other organisms (Begnaud et al., 2016), but instead requires Arp2/3-dependent actin polymerization (Xu and Chisholm, 2011).

A small number of reports have indicated that microtubules (MTs) cooperate with the actomyosin cytoskeleton during wound closure. In a single-cell wound-healing model using Xenopus laevis oocytes, MTs were shown to be required to restrict the assembly zone of actin and myosin around the wound edge (Bement et al., 1999; Mandato and Bement, 2003). A study in D. melanogaster embryos also revealed that perturbation of MT dynamics, in a + end-binding protein 1 (EB1) mutant, resulted in a delay of actomyosin assembly at the wound edge in multi-cellular wounds (Abreu-Blanco et al., 2012). The direct role of MTs in the C. elegans wound response has not been previously addressed.

The tissue repair processes in C. elegans have been described to act in parallel to the innate immune response that accompanies wounding (Xu and Chisholm, 2011). Several pieces of evidence suggest, however, a coordination of the two responses. Earlier work had shown that loss of DAPK-1, homolog of death-associated protein kinase, causes both undue tissue repair and an inappropriate activation of an immune response (Tong et al., 2009). A more recent study showed that loss of dapk-1 function causes excessive MT stabilization (Chuang et al., 2016). These results suggest that MT stability could directly influence immune signalling.

Here, we develop the use of a 405 nm laser to wound the C. elegans epidermis and characterise the subsequent subcellular events in vivo. We show for the first time a rapid membrane reorganisation of phosphatidylinositol 4,5-bisphosphate (PIP2) domains, as well as a recruitment of EB1 and reorganisation of MTs around the wound. Rings of EB1 and actin close at the same pace, with EB1 at the leading edge. We demonstrate that specific inactivation in the adult epidermis of several alpha and beta tubulin isotypes leads to a decrease in MT dynamics, revealed by an absence of EB1 comets. This inactivation also blocks EB1 accumulation and decreases the recruitment of actin at the wound site. Further, it also abrogates the directed trafficking to the wound site of the key signalling protein SNF-12 and causes a block of the subsequent immune response. These results suggest that the recruitment of MT + end and EB1 to wounds coordinates the activation of the innate immune response and wound repair.

Results

The C. elegans epidermis can be wounded with a 405 nm laser

The adult epidermis in C. elegans is a single cell layer, covered on its apical surface by an impermeable cuticle that serves as an exoskeleton. The epidermal syncytium hyp7 can be divided into two distinct regions, one lateral, where the nuclei and most of the cytoplasm are, and the dorso-ventral part above the muscles where the apical and basolateral plasma membranes of hyp7 are juxtaposed and crossed by hemidesmosomes that anchor the muscles to the cuticle exoskeleton (Figure 1A; Altun and Hall, 2014). The large size of hyp7 makes it an attractive model for the study of cellular wound repair mechanisms (Pujol et al., 2008a; Xu and Chisholm, 2011). Previous studies on Xenopus oocytes have demonstrated that the plasma membrane can be perforated in an extremely targeted manner using standard 405 nm microscopy imaging lasers (Burkel et al., 2012; Mandato and Bement, 2001). We found that we could efficiently wound the hyp7 lateral syncytial epidermis of C. elegans (Figure 1A) with a short pulse of 405 nm laser light (maximum power for 1–3 s) in Fluorescence Recovery After Photobleaching (FRAP) mode. Coupled with spinning disk or confocal microscopy, it provides a reproducible way to monitor in vivo the immediate consequences of a carefully controlled injury and the subsequent steps in the wound healing process. Using this system, we first confirmed all the previously described wound hallmarks (Pujol et al., 2008a; Xu and Chisholm, 2011), such as an immediate autofluorescent scar, a Ca2+ burst, the formation of an actin ring and later the induction of AMP reporter gene expression (Figure 1—figure supplement 1).

Figure 1 with 3 supplements see all
Download asset Open asset
Using a 405 nm laser to wound the worm epidermis reveals rapid membrane reorganisation.

(A) Schematic view of a section through an adult C. elegans worm near the mid-body. The internal organs are omitted for the sake of clarity. The main syncytial epidermis, hyp7 (buff), with its adjacent cells (seam cells in pink, nerve cord in red and muscles in green) can be divided into two main regions, the lateral epidermis (delimited by the dashed black line) and the thin dorso-ventral epidermis where muscles are anchored to the cuticle. Microtubules (thin taupe lines) are disordered on the apical surface of the lateral epidermis. The typical position of a wound is indicated by the yellow circle. Figure adapted from one kindly provided by Christopher Crocker, WormAtlas (Altun and Hall, 2014). (B) Wounding the lateral epidermis with a 405 nm laser causes a rapid increase in intracellular Ca2+, measured using GCamp3. Representative spinning disk images from a worm carrying a col-19p::GCamp3 reporter transgene. The dashed circle is centred on the wound site; bw, before wound; pw, post-wound; sc, seam cells. Scale bar 10 µm. (C–D) Upon laser wounding, membrane and lipids (red arrows) are rapidly recruited to the wound site (asterisk). Representative spinning disk images of worms carrying dpy-7p::GFP::CAAX (C) and wrt-2p::GFP::PH-PLC1δ transgenes (D). bw, before wound; time post-wound [min:s]; sc, seam cells. Scale bar 5 µm. The last panel on the right in both C and D is a focal plane two microns deeper than the previous images, showing recruitment of cytoplasmic vesicles to the wound.

Membrane is rapidly reorganised at the wound site

In Xenopus oocytes, a wound is rapidly patched through mobilisation and recruitment of membrane from a local pool of vesicles (Davenport et al., 2016). We could similarly observe rapid membrane recruitment at the wound site in the epidermal syncytium using a strain expressing a prenylated form of GFP (dpy-7p::GFP::CAAX). Within 15 s after wounding, large heterogeneous membrane domains that are normally stable regrouped and extended towards the wound (Figure 1C, Figure 1—video 1). PIP2 has been found segregated into distinct membrane pools in the plasma membrane in other species, consistent with its wide range of cellular functions including regulating the adhesion between the actin-based cortical cytoskeleton and the plasma membrane (Raucher et al., 2000). We found that the plasma membrane of the hyp7 syncytium was also heterogeneously labelled using a strain in which PIP2 was specifically labelled (wrt-2p::GFP::PH-PLC1δ Wildwater et al., 2011); these membrane domains were relatively static. Upon wounding, we observed a rapid but transient disappearance of the signal in a larger area than the wound. There was then a reorganization of these membrane domains around the wound after 15 to 30 s, followed by appearance of cytoplasmic vesicles below the plasma membrane (Figure 1D, Figure 1—video 2). These processes are likely to provide an immediate barrier preventing the potential leakage of cellular components, followed by a restoration of membrane integrity as previously proposed in other systems (Davenport et al., 2016; Nakamura et al., 2018; Vaughan et al., 2014).

Actin is recruited at the wound site forming concentric rings

Wounding the C. elegans epidermis causes cytoskeleton rearrangements, with an actin ring forming at the wound site (Xu and Chisholm, 2011). We wanted to monitor the dynamics of actin recruitment and so constructed a strain in which the filamentous actin binding peptide Lifeact (Riedl et al., 2008), labelled with mKate2, was specifically expressed in the adult epidermis (col-62p::Lifeact::mKate2). Under resting condition in the young adult, in the lateral epidermis, we observed a disorganised pattern of foci linked with fine filaments (Figure 2A, bw (before wound)). This is in contrast to the highly organised circumferential bundles of actin found above the muscles in the same hyp7 cell, as previously described (Costa et al., 1997; Lažetić et al., 2018; Figure 2—figure supplement 1A). Upon wounding, while the pre-existing actin filaments appeared stable, actin was recruited as a wave to the wound site within 2–4 min and formed a well-defined actin ring after 5 min (n > 100). This ring subsequently constricted as the wound closed (Figure 2A and Figure 2—video 1), in line with previous observations (Xu and Chisholm, 2011).

Figure 2 with 8 supplements see all
Download asset Open asset
EB1 associates with actin at the wound site in the lateral epidermis.

(A) In unwounded worms, in the lateral epidermis actin is sparsely structured. Laser wounding causes actin recruitment (as early as 2.5 min) and actin ring formation (clearly visible by 5 min). These rings close with time. Representative spinning disk images of a strain carrying a col-62p::Lifeact::mKate reporter. Laser wounding also causes ARP2/ARX-2 and EB1 recruitment at the wound site as visualized in strains ARX-2::GFP KI (B) and col-19p::EBP-2::GFP (C), respectively. Simultaneous visualization of MTs (green) and actin (red) using strains carrying EB1::GFP (D) or DNC-2::GFP (E) together with col-62p::Lifeact::mKate2. The dashed circle is centred on the wound site; bw, before wound; time post-wound [min:s]; scale bar 10 µm. (F) Kymograph of EB1 signal along the anterio-posterior (AP) direction normalised according to maximal intensity, see Materials and methods, reveals two phases of fast (P1) and slow (P2) contraction. (G) Kymograph of EB1 (green) and actin (red) during wound closure, EB1 ring contraction precedes that of actin. (H) Velocities of contraction in the fast (P1) and slow phases for EB1 and actin (n = 4).

Figure 2—source data 1

Reporter protein dynamics in the epidermis.

https://cdn.elifesciences.org/articles/45047/elife-45047-fig2-data1-v2.docx

In the adult epidermis, wounds do not close by a purse-string mechanism but by local actin polymerization involving the Arp2/3 complex (Xu and Chisholm, 2011). The Arp2/3 complex produces branched actin networks, and it contains in C. elegans seven subunits, encoded by arx-1 through arx-7 (Sawa et al., 2003). To confirm its role in the formation of the actin ring, we examined ARX-2/ARP2 during wounding, using an ARX-2::GFP knock-in strain (Wu et al., 2017). We could observe a recruitment of ARX-2/ARP2 to the wound with a pattern similar to actin (Figure 2B, Figure 2—video 2). These results reinforce the hypothesis that the actin ring closes due to polymerisation of branched actin network.

EB1 accumulates at the wound site with actin

Microtubules (MTs) are another essential component of the cytoskeleton, but their role in wound healing has not previously been addressed directly in C. elegans. To look at MT dynamics, we imaged worms expressing a tagged form of EB1/EBP-2, a protein that binds to the + end of growing MTs (Akhmanova and Hoogenraad, 2005; Srayko et al., 2005), specifically in the adult epidermis (col-19p::EBP-2::GFP; EB1::GFP for brevity) (Chuang et al., 2016). Under resting conditions EB1::GFP exhibited rapid, comet-like movement throughout hyp7, with a speed (0.26 ± 0.10 µm/s, Figure 2—source data 1) similar to the speed previously determined (0.3 µm/s) (Chuang et al., 2016). Upon wounding, EB1::GFP comets accumulated within 2 min around the wound site (n > 50, Figure 2C and Figure 2—video 3). This was concomitant with a small reduction of their average speed in the vicinity of the wound, to 0.16 ± 0.06 µm/s (Figure 2—source data 1). When actin and EB1 were visualised simultaneously, EB1 was found associated with the actin ring as it closes, with a preferential localisation at the leading edge (Figure 2D, Figure 2—videos 4 and 5). Using a dynactin reporter previously shown to be associated with MT + tips in C. elegans, DNC-2::GFP (Barbosa et al., 2017), we observed that there was a similar recruitment of DNC-2 at the inner edge of the actin ring (Figure 2E, Figure 4—video 1).

To address further the nature of the association of EB1 with actin, we quantified the respective speeds of their accumulation during ring closure. Kymograph analyses of the EB1 signal revealed two phases, the first one almost 10 times more rapid that the subsequent one (Figure 2F). Simultaneous analyses of both EB1 and actin signals revealed the existence of two regimes for both (Figure 2G). Notably, the speed of EB1 ring closure (0.008 µm/s) was almost 2 orders of magnitude slower than the speed of the EB1 comets (0.26 µm/s), and was similar to that of actin ring closure (Figure 2H, n = 4). Moreover, the specific enrichment of EB1 at the leading edge and EB1 ring closure was ahead of that of actin (Figure 2D and G, Figure 2—video 6, Figure 2—figure supplement 2). These results show that EB1 associated with actin at the time of the formation of the ring up to its closure.

MT dynamics at the wound

As EB1 comets reflect MT + tip dynamics, we then directly looked at the microtubule network after wounding. We used strains where one of the 16 C. elegans tubulin proteins, TBB-2, was labelled with GFP (Si [col-19p::GFP::TBB-2] and KI [tbb-2p::TBB-2::GFP]; both here termed TBB-2::GFP). We observed that MTs were arranged predominantly in longitudinal but haphazard bundles in the lateral epidermis, in marked contrast to their ordered, parallel and circumferential organisation underneath the muscle quadrants in the dorsal and ventral epidermis as well as during molting (Figure 2—figure supplement 1B–D), as previously described (Costa et al., 1997; Wang et al., 2015). Consistent with previous reports, EB1 comets were observed to run along MT bundles, with a preferred directionality in the A/P axis as observed for MT bundles in the main epidermis (Figure 2—figure supplement 1E–G). Tracking of EB1 comets confirmed this observation (Figure 3A and B). After wounding, no change in the direction of EB1 comets was observed (Figure 3B). Detailed observations with confocal microscopy revealed MT tips growing and retracting within second at the site of injury (Figure 3C and Figure 3—video 1). The dynamics of the MT tips paralleled the arrival of EB1 at the wound. At later time points, a global reorganisation of the MT network was observed as a knitting web with orthoradial MTs around the wound (Figure 3D, Figure 3—videos 2 and 3). The orientation of the MTs was analysed using a MAP1/MAPH-1.1::GFP reporter strain by quantifying the direction of the largest eigenvector of a local intensity structure tensor. Maps of their local orientation were defined within three concentric regions (R1-R3). The angle θ between the local direction of microtubule bundles and the direction to the centre of the wound region was then estimated (Figure 3E). After wounding, microtubules in the immediate wound periphery (R1) were perpendicular to the wound centre direction (θ ≈ 80). Such preferential orientation didnot exist in controls, for example in the distant sector R3 to the wound centre nor in any sectors R1-3 before laser ablation (Figure 3F). Thus, in concordance with the apparent random orientation of EB1 comets, these results show that MT were not arranged radially towards the wound, but displayed an increased orthoradial orientation proximal to the wound after 10 min. Together, these results lead to a working model in which MTs rapidly seed the accumulation of EB1, so facilitating the recruitment of actin, as well as knitting a webbed network around the site of injury.

Figure 3 with 3 supplements see all
Download asset Open asset
Microtubules are dynamic and orthoradially organised at the wound site.

(A) EB1 comets are tracked and mapped in a region of 150 × 150 pixels centred on the wound (black dot) for 5 min. (B) The angles of the EB1 comet vectors are calculated either to the A/P axis of the worm (left) or to the wound or the centre of a control region (right); their relative frequency are compared between control (n = 5) and wounded regions (n = 9). (C) Representative confocal images of a strain carrying a col-62p::Lifeact::mKate2 and TBB-2::GFP reporter; time post-wound [min:sec]; scale bar 5 µm. An asterisk marks the position of the tips of MTs extending into the site of the wound. The final panel is a temporal projection (Figure 3—video 1). (D) Representative spinning disk images of a strain carrying a TBB-2::GFP reporter. The dashed circle is centred on the wound site; time post-wound [min:sec]; scale bar 5 µm. (E) Map of the orientation of MTs 10 min post wounding, coloured according to their distance to the wound centre, defining sectors R1 (blue), R2 (red), R3 (green) in a MAPH-1::GFP reporter strain. Definition of the angle θ between the local direction of MT and the direction to the wound centre. (F) Statistics of the orientation θ for each sector just before (t = 0 min) and after (10 min) wounding. The increase in the mean value of θ in R1 corresponds to orthoradial MTs; the decrease in angle in R2 corresponds to the appearance of radially oriented structures; no global structures were observed at longer distances R3 (n = 3).

tba-2 or tbb-2 inactivation results in abrogation of EB1 dynamics in the epidermis

To test whether MT growth at the wound site indeed influences actin ring formation, we sought a way to block MT dynamics prior to wounding the epidermis. Due presumably to their impermeable cuticle, treatment of adult worms with commonly used drugs, such as colchicine, was either ineffective or had irreproducible effects in our hands. Although inactivating various genes that affect cuticle formation can increase permeability and so render C. elegans more susceptible to drugs (Loer et al., 2015; Partridge et al., 2008), this uniformly affects the epidermal innate immune response (Dodd et al., 2018; Zugasti et al., 2016). We therefore used RNAi to interfere directly with tubulin gene expression. Since eliminating MTs provokes developmental delay and lethality, worms were subjected to a relatively brief (24 hr) RNAi treatment from the L4 stage. Under these conditions, worms completed their development normally and exhibited no overt morphological defects. We chose one tubulin alpha tba-2 and one tubulin beta tbb-2, as these genes have been reported to be highly expressed in the epidermis (Cao et al., 2017; Hutter and Suh, 2016). In young adult tba-2(RNAi) or tbb-2(RNAi) worms, EB1::GFP comets were no longer visible suggesting that MT dynamics had been severely compromised (Figure 4A and B and Figure 4—figure supplement 1A). Interestingly, this effect was specific to the epidermis (hyp7) since highly motile comets were still visible in the seam cells (Figure 4A and B). Upon wounding, EB1::GFP was not recruited to the wound site (Figure 4A and C). This suggests that constant expression of tba-2 and tbb-2 genes is required for the normal dynamic behaviour of the non-centrosomal MTs in the epidermal syncytium, as well as for the redistribution of EB1 caused by injury.

Figure 4 with 2 supplements see all
Download asset Open asset
Non-centrosomal microtubule dynamics are required for actin ring closure upon wounding.

(A) Tubulin α TBA-2 is required for the presence of EB1 comets and EB1 recruitment at the wound site. Representative spinning disk images of col-19p::EBP-2::GFP in control (ctrl) and tba-2 RNAi-treated worms; time post wound [min:s]; scale bar 5 µm. (B) Quantification of EB1 comet number in either lateral epidermis or seam cells (n = 5) before wounding. (C) Quantification of EB1 recruitment at the wound site, at 10 min post-wounding (n = 10). Representative spinning disk images of Lifeact::mKate (D) kymograph analysis (E), intensity across the wound at 10 and 15 min (F) and quantification of the intensity of actin (G) and the area inside the actin ring (H) 15 min after wounding, in control and tubulin β encoding gene tbb-2 RNAi treated worms; ns p>0,05, *p<0.05, **p<0.01 and ****p<0.0001 non-parametric Mann-Whitney test.

Abrogation of MT dynamics decreases actin recruitment at the wound site

To address the question of whether MTs are required for the formation of the actin ring, we used the tbb-2(RNAi) conditions defined above. In resting conditions, short abrogation of tbb-2 from the L4 stage, altered the MT network in young adult worms, as visualised with TBB-2::GFP; it became less dense and less longitudinally structured in the lateral epidermis (Figure 4—figure supplement 1B). The actin network was also less dense in these worms (Figure 4—figure supplement 1B). Upon wounding, less actin was recruited around the site of injury following tbb-2 RNAi (Figure 4D and G). Further, analyses of the dynamics of radial intensities across the wound revealed that as well as accumulating to lower levels, actin was recruited later in tbb-2 RNAi treated than in control worms (Figure 4E–F). Moreover, the actin ring did not constrict as quickly as in control worms, as measured by the area inside the ring after 15 min (Figure 4H). Notably, the relationship between MTs and actin was not reciprocal. Thus, reduction of the level of actin via inactivation of act-2 did not change the overall pattern of MTs nor EB1 dynamics (Figure 4—figure supplement 1A and B). This suggests that MT dynamics do indeed facilitate actin ring closure.

Microtubule dynamics is required for the immune response

We had previously demonstrated in a genome-wide RNAi screen that knocking down certain genes associated with MT function, including tba-2 and tba-4, leads to an abrogation of the induction of the AMP gene reporter nlp-29p::GFP usually caused by Drechmeria coniospora infection (Supplementary file 1-Table S1, Zugasti et al., 2016). Using the same protocol of short inactivation from the L4 stage presented above, we confirmed that inactivation of these two tubulin α genes led to a block of AMP induction after infection in the adult (Figure 5A). In addition, we showed that tba-2 and tba-4 are also required for AMP reporter gene induction after wounding (Figure 5A). MTs are formed by dimers of tubulin, each consisting of two subunits, tubulin α and tubulin β. In the C. elegans genome, there are 9 α (TBA) and 6 β (TBB) genes. Some have been shown to act in specific cells or tissue, such as mec-7 (β) and mec-12 (α) in the mechanosensory cells (Savage et al., 1989), other are predicted to be ubiquitously expressed, and several to be expressed in the epidermis (Harris et al., 2010). We extended our investigation to all tubulin genes by assaying the effect of individually knocking down their expression by RNAi specifically in the adult epidermis, using an rde-1 RNAi resistant mutant rescued in the adult epidermis (Xu and Chisholm, 2011; Zugasti et al., 2014). In addition to tba-2 and tba-4, we found that inactivation of tubulin β genes tbb-1 and tbb-2 was also associated with a block of AMP reporter gene induction upon infection (Figure 5B and C). We found that this was not a consequence of reduced spore binding to the worm cuticle (Figure 5—figure supplement 1A). Using the TBB-2::GFP strain, we also showed that inactivation of the genes that blocked the immune response altered the pattern of MTs, confirming the expression and functional importance of these tubulin isotypes in the adult epidermis (Figure 5—figure supplement 2).

Figure 5 with 2 supplements see all
Download asset Open asset
Specific tubulin isoforms are required for the activation of the immune response upon wounding and fungal infection.

Quantification of green fluorescence in a strain carrying the nlp-29p::GFP transcriptional reporter after RNAi against different tubulin α, β and γ genes in non-infected worms or after infection with D. coniospora or after wounding (blue, orange and green symbols, respectively). The ratio between GFP intensity and size (time of flight; TOF) is represented in arbitrary units. (A) Worms were fed on RNAi bacteria from the L4 stage and after 24 hr infected or wounded; sta-2(RNAi) is known to block the immune response (Dierking et al., 2011). (B, C) Worms sensitive to RNAi primarily in the adult epidermis were fed on RNAi clones from the L1 stage and infected with D. coniospora at the young adult stage. Mean are represented in black, numbers of worms in 5A: 47, 87, 25, 48, 46, 26, 66, 107, 21, 33, 52, 28; in 5B: 106, 82, 67, 59, 63, 57, 112, 51, 95, 65, 88, 76, 79, 65, 128, 58, 98, 121, 69, 138; in 5C: 228, 180, 199, 210, 201, 162, 145, 232, 220, 224, 147, 207, 196, 188, 133, 209. Only ****p<0.0001 is presented; ANOVA Bonferoni’s test. Graphs are representative of the results obtained from at least three independent replicates.

On the other hand, inactivation of tubulin-γ (TBG-1) expression in the adult epidermis a few hours before infection did not affect AMP reporter gene induction nor the overall MTs pattern (Figure 5C and Figure 5—figure supplement 2). Since tubulin-γ nucleates MTs at their minus end (Quintin et al., 2016; Wang et al., 2015), this could suggest that unlike MT + ends, – ends are less dynamic in the adult epidermis and are not required for the immune response upon wounding. As previously shown in the C. elegans epidermis (Wang et al., 2015), however, TBG-1 could act redundantly with the minus-end stabiliser PTRN-1/patronin, which has a known role in epidermal wound repair (Chuang et al., 2016). Thus, independent of their role in the maintenance of epidermal structure, and in addition to a function in mediating wound closure upon wounding, MT + end dynamics appear to be required for the regulation of the transcriptional response to injury and fungal infection in the epidermis.

SNF-12, RAB-5 and RAB-11 get locally recruited upon wounding

We hypothesised that microtubules could be required for controlling the subcellular localization of key signalling molecules that are necessary for the induction of the epidermal immune response. The SLC6 family member SNF-12 was a prime candidate, as we have previously shown that it is localized in clusters at the apical surface of the hyp7 syncytium (Dierking et al., 2011). We therefore characterised further the localization of SNF-12 and its behaviour following wounding. We performed colocalisation analyses between SNF-12::GFP and available vesicular and plasma membrane markers (early, late and recycling endosomes, lysosomes, plasma membrane lipids). We did not detect any colocalisation, but we observed that SNF-12 clusters were in the same focal plane as PIP2 and the most apical RAB-11 recycling endosomes, and apical to LAAT-1-positive lysosomes (Figure 6A and B, Figure 6—figure supplements 1 and 2). This indicates that SNF-12 is in a yet-to-be defined apical membrane compartment. Interestingly, in the dorsal and ventral epidermis, SNF-12 was found in a banded pattern, like the highly organised circumferential cytoskeleton (Figure 6C, Figure 6—video 1). Compared to EB1 and RAB-11 (see below), the majority of SNF-12 clusters were static or just vibrating, with only very few moving longer straight distances, presumably along MT tracks, with a low speed of 0.017 ± 0.005 µm/s (Figure 6D, Figure 6—video 1, Figure 2—source data 1).

Figure 6 with 10 supplements see all
Download asset Open asset
SNF-12 localizes to apical clusters that are recruited at the wound site in a MT-dependent way.

Representative spinning disk images of worms carrying col-12p::SNF-12::GFP as well as a red marker of early endosome (snx-1p::mRFP::SNX-1; A), membrane lipids (PIP2;ced-1p::mCherry::PH-PLC1δ; B) or actin (col-62p::Lifeact::mKate; C); scale bar 10 µm. (D) The different dynamics of SNF-12, EB1 and RAB-11 are represented with a temporal color-coded projection of 120 frames over 2 min (one fps) before wounding. (E) After wounding, SNF-12::GFP clusters move towards the wound as represented with a color coded temporal projection of 32 frames over 10 min before wounding and 40 after, 6.30 min after wounding. (F–G) The SNF-12 recruitment to the wound site seen in control animals (F) is abrogated upon tba-2 RNAi (G). The dashed circle is centred on the wound site; time post wound [min:s]; scale bar 5 µm. (H) Quantification of SNF-12 cluster number in the lateral epidermis (worm n = 6). (I) Quantification of SNF-12 recruitment at the wound site at different time points post wounding (wound n = 12 for control worms and 10 for tba-2 RNAi). **p<0.01 and ***p<0.001; non-parametric Mann-Whitney test.

When we wounded worms carrying the SNF-12::GFP reporter, we observed a progressive recruitment of SNF-12 to the wound site (Figure 6E and F, Figure 6—video 2), with the majority of the clusters around the wound moving in a directed manner at a low speed of 0.007 ± 0.004 µm/s (Figure 2—source data 1). A similar recruitment was also seen in several other strains containing different tagged forms of SNF-12 (Figure 6—figure supplement 3) suggesting that it reflects the behaviour of the endogenous protein. When we disrupted MTs, via RNAi of tba-2, SNF-12 patterning and recruitment were severely compromised (Figure 6G–I). Similarly, both EB1 and SNF-12 patterning and dynamics upon wounding were also affected when the MT severing protein SPAS-1 was overexpressed in the worm epidermis (Figure 6—figure supplement 4). Together, these results suggest that MTs play an important role in SNF-12 localisation and dynamics and thereby in the induction of AMP gene expression.

The behaviour of SNF-12 was unusual as in almost all cases when we wounded strains in which vesicles markers were fluorescently tagged, there was no such recruitment (Figure 6—figure supplement 2). Thus, the movement of SNF-12 towards the wound site was not simply a reflection of cytoplasmic flow. Two exceptions were RAB-5, a marker of early endosomes, and RAB-11. RAB-5::GFP appeared localised in large static donut-shaped structures (0.3–0.8 µm) arranged in a reticular pattern, exchanging with smaller vesicles (0.2–0.4 µm), as described previously (Chuang et al., 2016; Figure 7A, Figure 6—video 3). Upon wounding these small vesicles accumulated around the wound site within 10–15 min (Figure 7A, Figure 6—video 4). RAB-11::GFP was found in extremely motile (1.31 ± 0.33 µm/s) vesicles (0.2–0.4 µm) (Figure 2—source data 1). They moved long distances in one direction, presumably on microtubule tracks (Figure 6D, Figure 6—video 5). Upon wounding they started to accumulate around the wound site within 10–15 min (Figure 7B, Figure 6—video 6) and reached a maximum density between 30 and 60 min post wounding. Notably, we previously demonstrated that both rab-5 and rab-11 are required non-redundantly for the epidermal innate immune response to infection (Dierking et al., 2011; Zugasti et al., 2016). Together, these results lead us to propose a timeline of events after wounding (Figure 8, Figure 8—figure supplement 1). In between the reorganisation of PIP2 membrane subdomains and the restoration of membrane integrity involving the endocytic pathway, an actin ring forms in a myosin-independent, ARP2/3-dependent manner, tightly linked to the recruitment of MT + tip binding-protein EB1. The coordinated recruitment of endosomal vesicles and the uncharacterised SNF-12-associated clusters, which are both dependent upon MT reorganisation, tie together wound repair and the innate immune response.

Figure 7
Download asset Open asset
Early and recycling endosomes are recruited to the wound site.

Representative spinning disk images of worms carrying dpy-7p::GFP::RAB-5 (A) and dpy-7p::GFP::RAB-11 (B). A fraction of the early endosome marker RAB-5 localizes to donut-shaped structures (red arrows). Upon laser wounding these RAB-5-positive structures (A) as well as RAB-11-positive recycling endosomes (B) are recruited to the wound site within 10 min. The dashed circle is centred on the wound site; time post wound [min:s]; scale bar 5 µm.

Figure 8 with 1 supplement see all
Download asset Open asset
Time line of events.

Discussion

Previous studies of wound healing in C. elegans used either imprecise and error prone mechanical injury with a needle, or sophisticated laser rigs not accessible to many laboratories. We demonstrated that a simple FRAP laser can be used to induce wounds with high precision, and can be combined easily with fast image acquisition. Using this system, we observed in the C. elegans epidermis the phenomenon of rapid membrane reorganisation previously described in Xenopus (Davenport et al., 2016), suggesting that these initial steps of plasma membrane repair, as well as subsequent steps in wound healing (Nakamura et al., 2018), may be conserved throughout Metazoans.

In addition to the acute patching of the membrane, wounding was followed by reorganisation of PIP2 domains. PIP2 is known to influence the adhesion between the actin-based cortical cytoskeleton and the plasma membrane (Raucher et al., 2000). Thus, its redistribution could be linked to the dramatic changes in the actin cytoskeleton that occurred in the minutes after wounding. The C. elegans epidermis is bounded on its apical side by a flexible collagen-rich cuticle that serves as an exoskeleton. As in plant cells, this is likely to give the structural support that is otherwise provided by a highly organised cortical actin network in most animal cells (Salbreux et al., 2012). This may explain why cortical actin is not organized in the adult lateral epidermis. During moulting, however, when the cuticle is remodelled, a highly organized cytoskeleton comprised of circumferential actin and microtubule bundles forms transiently, disappearing when the new cuticle is formed. This parallels the situation in the epidermis upon injury. Actin is rapidly assembled around the wound, and disassembled once it is repaired. Thus changes in membrane tension, either linked to alterations of the cuticle during development, or caused by a breach in its integrity are closely linked to cytoskeleton dynamics.

Wound repair has been previously suggested to involve a recapitulation of developmental programs (Martin and Parkhurst, 2004; Razzell et al., 2014). Thus for example, in the syncytial Drosophila early embryo, wound closure occurs through a myosin-II-dependent purse-string mechanism, driven mainly by the contractile force exerted by myosin (Abreu-Blanco et al., 2011; Bement et al., 1999). This parallels the tight control exerted by myosin on the actin cytoskeleton during morphogenetic cell shape changes in Drosophila embryogenesis (Lecuit and Lenne, 2007). In Xenopus oocytes, if myosin motor activity is inhibited, wounds still close, though at a much slower rate. The residual force driving wound closure is thought to be given by the treadmilling of actin, with a preferential recruitment of Rho GTPase at the wound (Burkel et al., 2012). The recruitment we observed of ARX-2/ARP2 to wounds reinforces the hypothesis that the actin ring in the C. elegans epidermis closes due to polymerisation of branched actin networks.

Actin-dependent processes, including cell migration and cell protrusion, can be influenced by MT dynamics. The mechanisms that underlie their functional crosstalk are still being characterised (Dogterom and Koenderink, 2019). MT + tip proteins like EB1 have been suggested to be involved in microtubule guidance, attachment to cellular structure or acting as concentration devices for signalling events (Akhmanova and Steinmetz, 2015). For example, a recent in vitro study showed that the MT + end-associated protein CLIP-170, together with the formin mDia1, can accelerate the actin filament polymerization which occurs at the tip of growing MTs (Henty-Ridilla et al., 2016). We showed here that in C. elegans the recruitment of MT + end binding protein EB1 to the wound is concomitant with that of actin. The two proteins then converge during wound closure at the same rate, with EB1 being at the leading edge. Both in vitro and in vivo studies have indicated that EB1 can interact directly with actin (Alberico et al., 2013). We demonstrated that perturbation of MT dynamics in vivo slows actin ring closure in C. elegans epidermis. Together, this suggests that MT dynamics could facilitate actin ring closure by bringing EB1 to the wound site. This MT-dependent mechanism, which is independent of myosin-based contractility (Xu and Chisholm, 2011), could arise because the epidermis is bounded by a rigid apical extracellular matrix, the tensile cuticle exoskeleton.

Injuring the epidermis not only leads to wound healing, but also provokes an innate immune response. Previous results in C. elegans suggested that the pathways involved in these two processes were distinct. Thus, although mutants for the p38 MAPK PMK-1 and for the SARM orthologue TIR-1, that acts upstream of the p38 cascade, are defective for the induction of AMP gene expression after wounding (Pujol et al., 2008a), they exhibit normal actin ring closure (Xu and Chisholm, 2011). If innate immune signal transduction pathway components are not required for wound repair, our results suggest, on the other hand, that cytoskeleton components can play a role in regulating AMP production. Indeed, several tubulin genes are required for AMP expression upon infection and wounding, potentially through their effect on the localization and dynamics of innate immune signalling proteins including SNF-12. The fact that overexpression in the epidermis of the MT severing protein Spastin caused a similar effect on SNF-12 dynamics as knocking down tubulin expression, supports a role for MTs in positioning SNF-12-containing vesicles. Determining the precise nature of these vesicles, their connection with MTs and the mode of SNF-12 activation remain challenges for the future. We speculate that this spatial regulation of signalling may contribute to restricting defence gene activation to the immediate vicinity of a wound or site of infection within the major epidermal syncytium.

A further link between cytoskeleton reorganization and AMP production comes from the study of the protein DAPK-1, a negative regulator of both wound closure and innate immune gene expression (Tong et al., 2009). Interestingly, dapk-1 mutants have defects in the epidermis that were shown to result from the hyperstabilisation of MTs. DAPK-1 inhibits the function of PTRN-1, a MT - end binding protein, which acts as a MT nucleator and promotes MT stabilization. Loss of ptrn-1 function suppresses the epidermal defects seen in dapk-1 mutants as well as their constitutive AMP gene expression and accelerated wound closure. Loss of dapk-1 function is associated with an increase in EB1 comets. This too is suppressed in a dapk-1;ptrn-1 double mutant (Chuang et al., 2016). Taken together with our results, this suggests that the accelerated wound healing seen in dapk-1 mutant worms may be due to an increased recruitment of EB1 protein thus speeding up MT-dependent actin ring closure.

The cytoskeleton is an assembly of co-regulating components that function together in a precise and dynamic way. Our results suggest that this co-regulation extends to the coordination of physical changes required for membrane wound healing and immune signal transduction driving transcriptional responses to injury.

Materials and methods

Nematode strains

Request a detailed protocol

All C. elegans strains were maintained on nematode growth medium (NGM) and fed with E. coli OP50, as described (Stiernagle, 2006): the wild-type N2, IG274 frIs7[col-19p::DsRed, nlp-29p::GFP] IV (Pujol et al., 2008a), IG823 frIs43[col-12p::SNF-12::GFP, ttx-3p::DsRed2] and IG1235 cdIs73[RME-8::mRFP, ttx-3p::GFP, unc-119(+)]; frIs43[col-12p::SNF-12::GFP, ttx-3p::DsRed2] (Dierking et al., 2011), IG1327 rde-1(ne219) V; juIs346[col-19p::RDE-1, ttx-3p::GFP] III; frIs7[nlp-29p::GFP, col-12p::DsRed] IV (Zugasti et al., 2014), CZ13896 juIs319[col-19p::GCaMP3, col-19p::tdTomato] (Xu and Chisholm, 2011), CZ14453 juEx3762[col-19p::EBP-2::GFP, ttx-3p::RFP], JLF302 ebp-2(wow47[EBP-2::GFP]) II; zif-1(gk117) III (Sallee et al., 2018), CZ14748 juls352[GFP::moesin, ttx-3p::RFP] I, CZ21789 juSi239[col-19p::GFP::TBB-2] I, CZ9334 juEx1919[dpy-7p::GFP::RAB-5, ttx-3p::RFP] (Chuang et al., 2016), SA854 tbb-2(tj26[GFP::TBB-2]) III (Honda et al., 2017), SV1009 Is[wrt-2p::GFP::PH-PLC1δ, wrt-2p::GFP::H2B, lin-48p::mCherry] (Wildwater et al., 2011), BOX188 maph-1.1(mib12[GFP::maph-1.1]) I (Waaijers et al., 2016), GCP417 dnc-2[prt42(3xflag::GFP::DHC-2)] III (Barbosa et al., 2017), GN675 tba-1(pg77[TagRFP-T::TBA-1]) I (Lockhead et al., 2016), GOU2047 arx-2(cas607[GFP::ARX-2]) (Wu et al., 2017), MBA365 Ex[dpy-7p::GFP::CAAX, myo-2p::GFP] kindly provided by M. Barkoulas (UCL), RT343 pwIs82[snx-1p::mRFP::SNX-1, unc-119(+)] (Sato et al., 2014, NP878 cdIs73[RME-8::mRFP, ttx-3p::GFP, unc-119(+)] (Shi et al., 2009), XW10992 qxIs513[ced-1p::mCherry::PLC-1-PH] and XW9653 qxIs68[ced-1p::mCherry::RAB-7] (Liu et al., 2012), ML1896 mcIs35[lin-26p::GFP::TBA-2, pat-4p::CFP, rol-6(su1006)]; mcIs54[dpy-7p::SPAS-1_IRES_NLSmCherry, unc-119(+)] X (Wang et al., 2015). All the multiple reporter strains generated in this study were obtained by conventional crosses (see Supplementary file 2-Key resource table for a list of all strains).

Constructs and transgenic lines

Request a detailed protocol

frSi9 is a single copy insertion on chromosome II at the location of the Nemagenetag Mos1 insertion (Vallin et al., 2012) ttTi5605 of pNP151 (col-62p::Lifeact::mKate2_c-nmy3’utr). pNP151 was obtained by insertion of the col-62 promoter fused to Lifeact::mKate2 (Reymann et al., 2016) into the MosSCI vector pCFJ151 (Frøkjaer-Jensen et al., 2008). It was injected into the EG6699 strain at 20 ng/µl together with pCFJ90 (myo-2p::mCherry) at 1.25 ng/µl, pCFJ104 (myo-3p::mCherry) at 5 ng/µl, pMA122 (hsp16.41p::PEEL-1) at 10 ng/µl, pCFJ601 (eft-3p::Mos1 transposase) at 20 ng/µl and pNP21 (unc-53pB::GFP Stringham et al., 2002) at 40 ng/µl. A strain containing the insertion was obtained following standard selection and PCR confirmation (Frøkjaer-Jensen et al., 2008).

frSi13 is a single copy insertion on chromosome II (ttTi5605 location) of pNP159 (dpy-7p::GFP::RAB-11) by CRISPR using a self-excising cassette (SEC) (Dickinson et al., 2015). pNP159 was obtained by insertion of dpy-7p::GFP::RAB-11 (kindly provided by Grégoire Michaux) into the pNP154 vector. pNP154 was made from a vector containing the SEC for single insertion on Chromosome II at the position of ttTi5605 (pAP087, kindly provided by Ari Pani). pNP159 was injected in N2 at 10 ng/µl together with pDD122 (eft-3p::Cas9) at 40 ng/µl, pCFJ90 (myo-2p::mCherry) at 2.5 ng/µl, pCFJ104 (myo-3p::mCherry) at 5 ng/µl, and #46168 (eef-1A.1p::CAS9-SV40_NLS::3’tbb-2 Friedland et al., 2013) at 30 ng/µl. Non-fluorescent roller worms were selected then heat shocked to remove the SEC by FloxP as described in Dickinson et al. (2015). All constructs were made using Gibson Assembly (NEB Inc, MA) and confirmed by PCR or sequencing. Plasmid sequences are available upon request.

Different transgenic strains containing several tagged version of SNF-12 were generated including IG1784 frEx597[pSO16(col-12p::SNF-12::mKate_3’unc-54), pCFJ90(myo-2p::mCherry], IG1270 frEx453[pMS8(col-12p::GFP::STA-2), pMS9(col-12p::mCherry::SNF-12)] and IG1663 frEx577[pNP158(snf-12p::SNF-12::GFP_3'snf-12, pCFJ90(myo-2p::mCherry), pCFJ104(myo-3p::mCherry)].

RNAi

Request a detailed protocol

The RNAi bacterial clones were obtained from the Ahringer library (Kamath et al., 2003) or Vidal library (Rual et al., 2004) and sequenced to confirm their identity using CloneMapper (Thakur et al., 2014). RNAi feeding experiments were performed at 25°C and the RNAi clone sta-1 was used as a negative control. To avoid developmental delay or lethality, RNAi experiments of tubulin or actin genes were either performed from the L4 stage or in a strain allowing gene silencing primarily in the epidermis from the L4 stage (IG1327 (rde-1(ne219) V; juIs346[col-19p::RDE-1, ttx-3p::GFP] III; frIs7[nlp-29p::GFP, col-12p::DsRed] IV) Zugasti et al., 2016), by rescuing the rde-1 RNAi resistant mutant in the adult epidermis as previously described (Xu and Chisholm, 2011).

Infection and needle wound

Request a detailed protocol

Eggs prepared by the standard bleach method were allowed to hatch in 50 mM NaCl in the absence of food at 25°C over night. Synchronized L1 larvae were transferred to NGM agar plates spread with E. coli OP50 and cultured at 25°C until the L4 stage (40 hr) before being exposed to fungal spores as previously described (Pujol et al., 2001). To count the number of spores, young adult worms were infected for 8 hr at 25°C then observed on a slide under the microscope. Statistical significance was determined using ANOVA Bonferoni’s test (Graphpad Prism). Needle wounding was performed as previously described (Pujol et al., 2008a) with a standard microinjection needle under a dissecting microscope by pricking the worm’s posterior body or tail on agar plates; worms were analysed after 6 hr.

Fluorescent reporter analyses

Request a detailed protocol

Analysis of nlp-29p::GFP expression was quantified with the COPAS Biosort (Union Biometrica; Holliston, MA) as described in Pujol et al. (2008b). In each case, the results are representative of at least three independent experiments with more than 70 worms analysed. The ratio between GFP intensity and size (time of flight; TOF) is represented in arbitrary units. Statistical significance was determined using a non-parametric analysis of variance with a Dunn’s test (Graphpad Prism). Fluorescent images were taken of transgenic worms mounted on a 2% agarose pad on a glass slide anesthetized with 0.01% levamisole, using the Zeiss AxioCam HR digital colour camera and AxioVision Rel. 4.6 software (Carl Zeiss AG).

Image acquisition and laser wound

Request a detailed protocol

Since in preliminary experiments we observed that immobilization using latex beads impacted vesicle and protein dynamics in the epidermis, young adult worms (containing less than five eggs) were immobilized on a slide using 0.01% levamisole. Time lapse and/or Z stack were acquired using two different spinning disk microscopes: an inverted Visitron Systems GmbH spinning disk with a Nikon 40X oil 1.3 NA objective and 1.5X lens, and a Roper inverted spinning disk with a Nikon 100X oil 1.4 NA objective, were used with their respective acquisition software (VisiView or Metamorph). For spatially resolved time-lapse imaging, a Zeiss LSM780 confocal microscope whose GaAsP detector was chosen for its high sensitivity and controlled through the Zen acquisition software.

Laser wounding was performed by using a 405 nm laser at its maximum power using either Visitron, iLas FRAP or Zeiss FRAP systems. The respective laser power before objectives was: 7.3 mW, 1–2 mW and 3 mW. 1–3 s of laser pulses were applied on a circular region of interest (ROI) of 8 × 8 pixels with the Visitron (~2.13 µm in diameter), 21 × 21 pixels with the Nikon (~1.68 µm in diameter) or with the confocal (~1.8 µm in diameter) on the apical plane of the lateral epidermis either left or right of the seam cells. Each worm was wounded from 1 to 3 times.

Time lapse acquisitions were 10 to 20 min long. Z stack acquisitions were performed on 4–6 planes spaced by 0.3 or 0.5 µm. High temporal resolution was obtained on one single plane with a time interval of 0.2 to 0.5 ms. Laser wound was performed between the first and the second Z stack. For colocalisation analysis, we initially controlled that there was no signal cross talk using single labelled strains. We then performed sequential acquisitions on the strains expressing two fluorescent proteins. Experiments requiring comparative analysis were performed on the same day.

Image analysis

Request a detailed protocol

All image processings were done using Fiji software. A custom Image J macro was used to correct for Z-drift, while XY drift was corrected either with the Image Stabilizer or the Correct 3D drift plugin from ImageJ or the online available macro ‘NMSchneider/FixTranslation’.

To analyse EB1 density before wounding, we manually counted fluorescent dots as described in Chuang et al. (2016), in a 300 µm2 ROI in the lateral hyp7 or in an 80 µm2 ROI in the seam cell in five worms per RNAi condition. Similarly, we measured SNF-12 cluster density in the lateral hyp7, analysing six worms per RNAi condition. Quantification of EB1 recruitment at the wound site was performed by measuring the RawIntDen (sum of the pixel intensities) of a selected ROI of 256 µm2 for EB1 centred at the wound site. Since EB1 brightness differed before wounding between worms, for each single wound, only pixels with values above a defined brightness threshold were considered, the threshold being set to the [mean + 3 x standard deviation] in the ROI before wounding. Quantification of SNF-12 recruitment was performed by measuring the mean intensity of a selected ROI of 18 µm2 centred at the wound site at specific time points. Statistical significance was determined using a non-parametric Mann-Whitney test (Graphpad Prism).

To visualise the speed and directionality of EB1, RAB-11 and SNF-12, a temporal projection was done using a temporal-Color code (plugin from Kota Miura) on 120 frames over 2 min (one frame per second) before wounding for all tree makers and on 32 frames over 10 min after wounding for SNF-12.

To analyse particle motility before and after wounding, we used the macro KymographClear described in Mangeol et al. (2016). For EB1, RAB-11 and SNF-12, we acquired movies with time interval of 300 ms, 400 ms and 200 ms, respectively. Since SNF-12 was observed to move much slower than the other dynamic proteins studied, a substack with a time interval of 20 s was generated. SNF-12 quantifications were done for particles with a minimum displacement of 5 pixels within 120 s for before wounding and within either 210 or 420 s for after wounding. To analyse the directionality of EB1 movement, we developed a workflow using first available Image J/Fuji plugins to correct image drift (‘Image stabilizer’), crop movies to a standard size and number of frames (100*100 pixels *1000;”crop’ and ‘make substack’), subtract the stable background with a median temporal filter (‘median 3D’, treating the time series as a Z-stack, setting the filter at 31 (temporal), and then subtracting from each original image). We then defined particles using the Trackmate plugin (Tinevez et al., 2017), with estimated blob diameter of 3 pixels, and a threshold at 900, linking distance and gap-closing max distance two pixels, gap-closing max frame gap of 1. From the resulting file, the track ID and x,y positions were analysed using a custom R script that calculated vectors relative to the wound site for each moving particle.

To compare the actin ring closure upon different RNAi condition, a ROI of 80 pixels2 located on the centre of the wound was selected 15 min post-wounding. The area inside the actin ring was detected using the Huang automatic threshold method from Image J. To first compare the recruitment of EB1 versus actin, the radial profile (plugin from Paul Baggethun) was calculated, after determination of the centroid of the wound for each time point by manually selecting the wound area. For each radial profile, we extracted the radius corresponding to the maximum intensity and plotted this radius versus time. To further quantify, we build kymographs by considering masks that corresponds to stripes oriented in the anterior-posterior (AP) direction and including the ablation region. We obtained one-dimensional profiles by averaging the signal intensity over the circumferential direction; to account for variations in signal intensity during time, we normalised each profile by its maximal value. These colour-coded profiles were concatenated to produce a kymograph.

To analyse the microtubule orientation, we obtained maps of the local orientation of microtubules of an ROI of 20 × 20 µm of confocal images of MAPH-1::GFP reporter strain using the ImageJ plugin OrientationJ and custom Matlab scripts (Source file plotorientation), which provide the direction of the largest eigenvector of a local intensity structure tensor (Rezakhaniha et al., 2012). We then estimated the angle θ between the local direction of microtubule bundle and the direction to a point of reference (the centre of the wound region); θ lies in the 0 to +90 sector since it measures the discrepancy between two directions. We then measured the averaged angle θ over three annular sectors labelled according to their distance to the centre of the wound: R1, R2, R3. Statistics were performed in Matlab(R) using the circular Statistic toolbox (Berens, 2009). Custom R and Matlab scripts are either available as source file or upon request.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for all quantitative figures.

References

    1. Altun ZF
    2. Hall DH
    (2014)
    WormAtlas
    Epithelial system, hypodermis, WormAtlas, Albert Einstein College of Medicine, 10.3908/wormatlas.1.13.
    1. Lockhead D
    2. Schwarz EM
    3. O'Hagan R
    4. Bellotti S
    5. Krieg M
    6. Barr MM
    7. Dunn AR
    8. Sternberg PW
    9. Goodman MB
    (2016)
    The tubulin repertoire of C. elegans sensory neurons and its context-dependent role in process outgrowth
    Molecular Biology of the Cell 27:3717–3728.
    1. Stiernagle T
    (2006)
    WormBook
    Maintenance of C. elegans, WormBook, Albert Einstein College of Medicine, 10.1895/wormbook.1.101.1, 18050451.
    1. Stringham E
    2. Pujol N
    3. Vandekerckhove J
    4. Bogaert T
    (2002)
    unc-53 controls longitudinal migration in C. elegans
    Development 129:3367–3379.

Decision letter

  1. Terry Lechler
    Reviewing Editor; Duke University, United States
  2. Didier YR Stainier
    Senior Editor; Max Planck Institute for Heart and Lung Research, Germany
  3. William M Bement
    Reviewer; University of Wisconsin-Madison, United States
  4. Andrew D Chisholm
    Reviewer; University of California, San Diego, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Thank you for submitting your article "Microtubule plus-end dynamics link wound repair to the innate immune response" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen Didier Stainier as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: William Bement (Reviewer #2); Andrew D Chisholm (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This manuscript establishes a new model for cellular wound healing, using laser induced cutting of the C. elegans epidermis. The authors characterize the cellular responses to wound healing and note a remarkable accumulation of the microtubule plus end binding protein Eb1 at the wound site. They further suggest that this regulates both subsequent actin accumulation and the enrichment of a transcription factor that mediates the immune response to wounding. These interesting observations link a novel microtubule-dependent wound response to a local immune response.

Essential revisions:

1) Organization of microtubules. The data on accumulation of Eb1 puncta at the wound site is very clear. However, neither the organization of their paths, nor a clear image of microtubule organization is presented. Further investigation of this will allow increased clarity of the model for microtubule function in wound healing.

2) Quantitation of data. This manuscript characterizes a new method for wounding and describes the dynamics of various cellular components during the healing process. To make it a more valuable resource for future research, increased quantitation of the data, showing both reproducibility over a number of cells, as well as showing the kinetics of various components relative to each other should be provided.

There were a number of points that were not entirely clear in the manuscript, such as how microtubules control SNF-12 localization (and further discussed in detail below). Changes to the text to clarify many of these confusions, listed in the reviews below, would strengthen the manuscript.

Reviewer #1:

This manuscript establishes a robust assay for wounding the C. elegans lateral epidermis. Using this assay, the authors characterize the membrane and cytoskeletal responses to wounding, which are similar to previously studies Drosophila and Xenopus models. They also show a remarkable increase in the microtubule end-binding protein Eb1, at the wound site. Functionally, they find that disrupting microtubule dynamics inhibits actin accumulation and the normal immune response. There are several really interesting observations in this paper, however, the conclusions at present could be strengthened by increased rigor throughout as different interpretations of the data exist.

1) As they are establishing a novel wounding methodology and characterizing the response, increased quantitation of the dynamics and extent of responses in Figures 1-3 are required. For all of these, it would be ideal to have (in addition to the videos shown) quantitations over several videos to see the reproducibility of the responses. This will also be helpful in laying out the temporal order of changes occurring. I find this to be especially important as the authors suggest that the microtubule response occurs before the actin accumulation, though this is not clear from the video presented.

2) Constriction of the F-actin at the wound. This is documented for a single wound in Figure 3H, but is not obvious in the videos provided. It is important to document whether this is a constriction or just a loss of F-actin over time. This is especially confusing as the Introduction states that healing hear is not driven by acto-myosin contraction, but rather by Arp2/3 mediate F-actin assembly.

3) The accumulation of tubulin at wound sites was dramatically less than Eb1. This is perhaps surprising and should be discussed. Along these lines, it was not clear whether the authors could track Eb1 puncta within the wound site due to density of puncta. Are all puncta motile and analyzable?

4) Epistasis of actin and microtubule accumulation. The authors state in the Discussion that Eb1 accumulates before actin. However, this is not clear from the videos provided. Functionally, there is a decrease in F-actin recruitment when tubulin genes are knocked down or microtubules disrupted. However, the timepoints here were late and don't address whether initial accumulation occurs but does not increase or whether it peaks and goes down, or never increases in the first place. An additional concern along these lines is that the F-actin as reported by LifeAct appears to be significantly disrupted in the unwounded tubulin mutant epidermis (S2B – to a similar extend as the Act2 mutant). This raises the caveat that it may not be a wound-induced effect, but a general effect on cytoskeletal organization.

5) Connection to the immune response. I felt that I needed more discussion of how the SNF-12 localization connected to the phenotype. Is the localization of the SNF-12 transcription factor at the wound site necessary for its transcriptional activity? How do you imagine this occurs. Are microtubules required for MAPK kinase activity or DCAR-1 signaling in any way? Authors should rule out the possibility that microtubules affect this signaling axis and that they work through localization of SNF-12. Along these lines, in the induction of AMP's in response to cuticle defects similarly dependent on tubulin genes?

6) SNF-12 clusters were significantly lower in unwounded epidermis of tubulin mutants compared to controls. One possibility is that microtubules are required for cluster formation and not localization of SNF-12.

Reviewer #2:

While not widely recognized, the response of single cells to damage is just as important as the response of tissues to damage. Not surprisingly then, a growing body of work suggests that cell damage elicits a conserved response that involves the mobilization of a variety of signaling, lipid, and cytoskeletal components. There are at least two biological consequences of this response: First, the hole in the cell is patched or otherwise closed, preventing the cell from simply lysing. Second, the damage material is cleaned up and removed, such that the wound site is eventually restored to its original state (Sonnemann and Bement, 2011). To a first approximation, these biological imperatives parallel those of damaged tissue, with clotting being responsible for plugging the hole and the action of immune system being responsible for the clean-up.

In the current paper, Taffoni et al. ask whether the single cell wound response is somehow linked to and coordinated with the immune response in the C. elegans embryo. To accomplish this, they develop a laser wounding system that permits them to make relatively precise wounds in the embryonic syncytial epiderms. They first show that repair of these wounds has several characteristics in common with the repair response in other damaged cell types. They then identify a striking accumulation of EB1, a protein that binds to the plus ends of growing microtubules, around wounds, leading them to conclude that wounding elicits local microtubule reorganization around wounds. They report that they can block this response by depleting microtubule subunit proteins (i.e. alpha and beta tubulin) and that this manipulation not only impairs the immediate wound response, but also impairs the innate immune response as monitored by expression of an AMP reporter gene. They further link microtubules to recruitment of SNF-12-a protein involved in the innate immune response-to wound sites. Based on these results the authors conclude that the microtubule response is an important feature of both the basic (cell healing) wound response and the innate immune response.

Several features of this paper are very appealing: the question is of fundamental importance for those who seek to understand wound repair, the authors nicely make the link between the basics of cell damage repair as it occurs in their system and in other systems, and much of the imaging clearly illustrates the points the authors are making. Perhaps most importantly, the link between the microtubule response in a single cell (albeit a very big cell) and the mobilization of the innate immune response is an important one.

However, several points need to be addressed:

1) It is very difficult to tell what, exactly, is happening with the microtubules at the wound site because the different approaches seem to be suggesting different things. That is, in some cell types microtubules accumulate around wounds in a radial array such that many microtubules are focussed toward the wounds (e.g. Mandato and Bement, 2003). Such an arrangement would be consistent with the accumulation of the EB1-GFP fluorescence around wounds (but see below). However, it is not consistent with the movement of the EB1-GFP comets, which appear to move in all directions and in the absence of any obvious bias toward the wounds. The live imaging of the microtubules themselves (Figure 3C) is also inconsistent with a radial array. As with the movement of the comets, it seems to indicate that the microtubules are randomly oriented with respect to the wound. Now, one might be tempted to imagine that the intense EB1-GFP signal that accumulates around the edge of the wound is sufficient evidence of some kind of wound-localized array of microtubules but there are two problems with this notion. First, that signal is not obviously in comet form, rather, it just appears to be a uniform slathering of GFP on the plasma membrane. Second, EB1 only associates with growing microtubules plus ends; it does not associate with stable microtubule plus ends (e.g. Akhmanova and Steinmetz, 2015). Thus, the apparently stable accumulation of EB1-GFP at the wound edge is rather mysterious. To settle this point the authors need to fix wounded embryos and stain for microtubules or find an alternative way to characterize the organization of microtubules around wounds.

2) This may reflect my ignorance, but it is not at all clear to me what links the transport of SNF-12 to wounds and the triggering of the innate wound response. Does something special happen to the SNF-12 at wound edges? I was under the impression that SNF-12 exerts its effects on transcriptional control of the response after entering the nucleus as previously reported by the authors (Dierking et al., 2011). While it is not incumbent on the authors to trace out all of the steps in the pathway that leads to an upregulation of the immune transcriptional response, it would be helpful to have a least a little bit of information on why recruitment to the wound is important.

Reviewer #3:

Taffoni et al. examine the role of microtubule dynamics in epithelial wound responses in the epidermis of the nematode C. elegans. Microtubule dynamics have been implicated in several cell repair processes, and were implicated in C. elegans skin wound repair, but had not previously been examined directly. The authors use extensive live imaging of fluorescently tagged markers to show that MT dynamics are affected by wounding and link physical wound repair to innate immune response to damage via transport of the signalling molecule SNF-12. The authors provide extensive descriptions of the dynamics of various membrane and organelle markers after injury. Overall this is a thorough and careful study that advances our understanding of the cell biology of wound repair in vivo.

https://doi.org/10.7554/eLife.45047.sa1

Author response

Essential revisions:

1) Organization of microtubules. The data on accumulation of Eb1 puncta at the wound site is very clear. However, neither the organization of their paths, nor a clear image of microtubule organization is presented. Further investigation of this will allow increased clarity of the model for microtubule function in wound healing.

We thank the reviewers for their insightful comments that led us to define a more precise model. We have now better characterised EB1 and microtubule dynamics upon wounding. We have generated several new strains and combined imaging with advanced analysis techniques.

Detailed analysis of EB1 revealed 2 regimes. The first was analysed by tracking EB1 comets after wounding. It matches the typical speed of microtubule + tips. This tracking also revealed that EB1 comets were not associated with any directional bias toward the wound, but seem to stop at the wound en-passant. The second regime of EB1 starts when it accumulates at the wound in the form of a ring. Kymograph analyses of both EB1 and actin signals reveal that the EB1 ring closes at the same speed as the actin ring, while also confirming our previous observation that EB1 precedes actin, EB1 being found at the leading edge of actin while the ring closes.

We have complemented these observations with more detailed examination of MT by analysing several MT-binding protein reporters, like the dynactin protein DNC-2, known to bind MT + tips in the C. elegans embryo (Barbosa et al., 2017), and MAPH-1, a MT associated protein MAP1, as well as more resolutive confocal analyses of the tubulin TBB-2 reporter. The observations with DNC-2 were fully consistent with what we had seen with EB1 and confirmed an accumulation of MT + tips at the wound at the same time as actin. Second, confocal analysis revealed MT growing and retracting at the wound just before and during actin accumulation. Third, quantification of MT orientation and tracking of EB1 comets, allow us to rule out a radial organisation of MT at the wound, consistent with the non-directional dynamics of EB1 comets. These data support a dynamic orthoradial organisation of the MT at the wound, and suggest that the microtubules reorganised to knit a web around the wound as the actin ring closes.

Reviewer #2 pointed out (see below), generally, “EB1 only associates with growing microtubules plus ends; it does not associate with stable microtubule plus ends (e.g. Akhmanova and Steinmetz, 2015)”. To reconcile this model with our observations that EB1 stops at the wound site and then migrates inward with the actin as the wound closes, we propose as a working model that MT-associated EB1 binds to actin at the wound and then dissociates from MT during closure. The actin and MT cytoskeleton have long been known to be intimately linked, for example during the directional migration of axonal growth cones. Indeed the main molecular linkers connecting actin and microtubules belong to the microtubule plus-end tracking proteins (+TIPs), including EB1 (Coles C. H. and Bradke F. 2015. Coordinating neuronal actin-microtubule dynamics. Curr. Biol.), and a direct association of EB1 with actin has been demonstrated in vitro (Alberico et al. 2016. Interactions between the Microtubule Binding Protein EB1 and F-Actin. J Mol Biol.). Recent studies also proposed a role of actin in the MT spindle formation during mitosis (Farina F. el al 2019. Local actin nucleation tunes centrosomal microtubule nucleation during passage through mitosis. EMBO J). Our results suggest a model where MT dynamics at the wound would permit the association of the +TIPs EB1 with actin that would then facilitate ring closure.

2) Quantitation of data. This manuscript characterizes a new method for wounding and describes the dynamics of various cellular components during the healing process. To make it a more valuable resource for future research, increased quantitation of the data, showing both reproducibility over a number of cells, as well as showing the kinetics of various components relative to each other should be provided.

The formation of an actin ring is a highly reproducible phenotype in our hands. It has become our general control (after the calcium wave) that we use to adjust the 405 laser settings to obtain a normal wound response in control worms. And it has allow us to implement a new wounding platform on a confocal microscope, still using a 405 laser from a FRAP module, in addition to our previous setting on two different spinning disk microscopes. We have now generated several new strains combining the actin reporter with MT binding proteins, and added quantification for actin and EB1 recruitment as mentioned above. In addition to new videos, we now provide kymograph analysis. We also have added the number of worm analysed in the different figures and a detailed source data file reporting for all quantification we have made (Figure 2H, 3B, 3F, 4B, 4C, 4F, 4G, 4H, 5A, 5B, 5C, 6H, 6I, Figure 8—figure supplement 1). Finally, we have added a supplementary figure (Figure 8—figure supplement 1) with quantification of the onset of recruitment and a proposed timeline of events.

There were a number of points that were not entirely clear in the manuscript, such as how microtubules control SNF-12 localization (and further discussed in detail below). Changes to the text to clarify many of these confusions, listed in the reviews below, would strengthen the manuscript.

We have changed the text as explained in more detail below.

Reviewer #1:

This manuscript establishes a robust assay for wounding the C. elegans lateral epidermis. Using this assay, the authors characterize the membrane and cytoskeletal responses to wounding, which are similar to previously studies Drosophila and Xenopus models. They also show a remarkable increase in the microtubule end-binding protein Eb1, at the wound site. Functionally, they find that disrupting microtubule dynamics inhibits actin accumulation and the normal immune response. There are several really interesting observations in this paper, however, the conclusions at present could be strengthened by increased rigor throughout as different interpretations of the data exist.

1) As they are establishing a novel wounding methodology and characterizing the response, increased quantitation of the dynamics and extent of responses in Figures 1-3 are required. For all of these, it would be ideal to have (in addition to the videos shown) quantitations over several videos to see the reproducibility of the responses. This will also be helpful in laying out the temporal order of changes occurring. I find this to be especially important as the authors suggest that the microtubule response occurs before the actin accumulation, though this is not clear from the video presented.

We have added quantifications as mentioned above.

2) Constriction of the F-actin at the wound. This is documented for a single wound in Figure 3H, but is not obvious in the videos provided. It is important to document whether this is a constriction or just a loss of F-actin over time. This is especially confusing as the Introduction states that healing hear is not driven by acto-myosin contraction, but rather by Arp2/3 mediate F-actin assembly.

We provide clearer evidence for the closure of the actin ring in new figures and videos. We indeed believe that the actin ring is not closing by contraction but by inward directed polymerisation of actin. We included the recruitment of Arp2/3 complex protein ARX-2 (Figure 2), confirming the role of branched actin polymerisation in actin closure demonstrated in (Xu and Chisholm, 2011).

3) The accumulation of tubulin at wound sites was dramatically less than Eb1. This is perhaps surprising and should be discussed.

Please see our answer above and to point 1 of Reviewer #2.

Along these lines, it was not clear whether the authors could track Eb1 puncta within the wound site due to density of puncta. Are all puncta motile and analyzable?

It was indeed not possible to track EB1 comets after they reach the wound due to their dense accumulation. But we were able to track the peripheral comets in 9 wounds and 5 control regions. These analyses reveal that their trajectories were not directed toward the wound. We believe that EB1 signal do reflect the dynamics of the + tips MT until they reach the wound (see Figure 2—figure supplement 1). As mentioned above, we have now visualised the recruitment at the wound of another MT + tip protein dynactin 2.

When at the wound, EB1 comets stop, accumulate as a ring that migrates at the same speed as the actin ring to close the wound (see kymographs Figure 2). As we now discuss, these observations are compatible with EB1 binding to actin at the wound.

4) Epistasis of actin and microtubule accumulation. The authors state in the Discussion that Eb1 accumulates before actin. However, this is not clear from the videos provided.

As mentioned above, in addition to new videos, we now provide kymograph analysis. Both show that EB1 associated at the leading edge of the actin while the ring forms and as the ring closes.

Functionally, there is a decrease in F-actin recruitment when tubulin genes are knocked down or microtubules disrupted. However, the timepoints here were late and don't address whether initial accumulation occurs but does not increase or whether it peaks and goes down, or never increases in the first place.

We have added a quantification of the phenotype and show that actin is recruited later and to a lesser extent in tbb-2 RNAi treated worms (Figure 4).

An additional concern along these lines is that the F-actin as reported by LifeAct appears to be significantly disrupted in the unwounded tubulin mutant epidermis (S2B – to a similar extend as the Act2 mutant). This raises the caveat that it may not be a wound-induced effect, but a general effect on cytoskeletal organization.

We thank the reviewer for highlighting this point that we had reported but not discussed in the original submission. The actin pattern is somewhat affected in young adults after tbb-2 abrogation during the L4 larval stage. Between L4 and adult, worms molt. There is a major reorganisation of both actin and MT networks during molting (Costa et al., 1997; Lazetic et al., 2018); and the reorganisation of both networks seems to highly coordinated (NP unpublished). Altering MTs could therefore impact the organisation of the actin cytoskeleton. We now mention this briefly in the Discussion. We expect that the coordination of MT and actin networks upon molting or after wounding will use the same mechanisms, and this will be the subject of future studies.

Returning to the reviewer’s concern, within a population of tbb-2(RNAi)-treated worms, some worms have an almost normal actin network. If we analyse these worms, we can still observe the concentric wave of actin as in the control but less actin accumulates at the wound site and wound closure is less efficient (Figure 4). High resolution live imaging with a confocal microscope revealed that some MT at the wound site do pattern the growing actin patch (Figure 3C, Video 9). Together with the fact that EB1 is found at the leading edge of the actin ring during its formation and closure, this suggests that MT dynamics facilitate actin accumulation at the wound. We have modified the text accordingly.

5) Connection to the immune response. I felt that I needed more discussion of how the SNF-12 localization connected to the phenotype. Is the localization of the SNF-12 transcription factor at the wound site necessary for its transcriptional activity? How do you imagine this occurs.

Please see our answer to point 2 of Reviewer #2. We regret that we didn’t make clear that SNF-12 is an amino acid transporter that we believe is acting as a signalling protein. We have added to the Discussion “We speculate that this spatial regulation of signalling may contribute to restricting defence gene activation to the immediate vicinity of a wound or site of infection within the major epidermal syncytium.”

Are microtubules required for MAPK kinase activity or DCAR-1 signaling in any way? Authors should rule out the possibility that microtubules affect this signaling axis and that they work through localization of SNF-12.

We apologize if we were not clear enough in the text. DCAR-1 signalling is entirely dependent upon SNF-12, so yes, microtubules affect this signalling axis and they work through localization of SNF-12. We have changed the text in the Discussion to make it clearer. We know that DCAR-1 (the most upstream element of the MAPK pathway described so far) is not recruited to the wound site before SNF-12 but much later, one hour after wounding with a general increased expression. We believe that these observations go beyond the scope of the current paper.

Along these lines, in the induction of AMP's in response to cuticle defects similarly dependent on tubulin genes?

We previously reported that in furrow-less Dpy mutants there is a constitutive high expression of the nlp AMP genes (Dodd et al., 2018). We have found that blocking MT dynamics suppresses this high expression (N.P. unpublished). We could add these results if absolutely required but believe they are beyond the scope of the current paper.

6) SNF-12 clusters were significantly lower in unwounded epidermis of tubulin mutants compared to controls. One possibility is that microtubules are required for cluster formation and not localization of SNF-12.

We agree and we now write in the text, “When we disrupted MTs, via RNAi of tba-2, SNF-12 patterning and recruitment were severely compromised (Figure 6G-I).” “Together, these results suggest that MTs play an important role in SNF-12 localization and dynamics and thereby in the induction of AMP gene expression.”

Reviewer #2:

[…] Several points need to be addressed:

1) It is very difficult to tell what, exactly, is happening with the microtubules at the wound site because the different approaches seem to be suggesting different things. That is, in some cell types microtubules accumulate around wounds in a radial array such that many microtubules are focussed toward the wounds (e.g. Mandato and Bement, 2003).. Such an arrangement would be consistent with the accumulation of the EB1-GFP fluorescence around wounds (but see below). However, it is not consistent with the movement of the EB1-GFP comets, which appear to move in all directions and in the absence of any obvious bias toward the wounds. The live imaging of the microtubules themselves (Figure 3C) is also inconsistent with a radial array. As with the movement of the comets, it seems to indicate that the microtubules are randomly oriented with respect to the wound. Now, one might be tempted to imagine that the intense EB1-GFP signal that accumulates around the edge of the wound is sufficient evidence of some kind of wound-localized array of microtubules but there are two problems with this notion. First, that signal is not obviously in comet form, rather, it just appears to be a uniform slathering of GFP on the plasma membrane. Second, EB1 only associates with growing microtubules plus ends; it does not associate with stable microtubule plus ends (e.g. Akhmanova and Steinmetz, 2015). Thus, the apparently stable accumulation of EB1-GFP at the wound edge is rather mysterious. To settle this point the authors need to fix wounded embryos and stain for microtubules or find an alternative way to characterize the organization of microtubules around wounds.

We agree with the reviewer that our observation in the nematode epidermis are not wholly consistent with what has been reported in other systems. To address the reviewer’s point, we chose not to fix the worm as explained above, but have now analysed more MT markers like DNC-2, MAPH-1 and another tubulin isotype TBA-1, in each case generating strains that allow us to visualise at the same time EB1 and/or actin recruitment. Further, we have increased the spatial resolution of our study by implementing the FRAP module for wounding on a confocal microscope, tracked EB1 comets and analysed EB1 and actin ring closure. Together these allow us to have a more complete description of the reorganisation of MTs that accompanies wounding, as explained above.

2) This may reflect my ignorance, but it is not at all clear to me what links the transport of SNF-12 to wounds and the triggering of the innate wound response. Does something special happen to the SNF-12 at wound edges? I was under the impression that SNF-12 exerts its effects on transcriptional control of the response after entering the nucleus as previously reported by the authors (Dierking et al., 2011). While it is not incumbent on the authors to trace out all of the steps in the pathway that leads to an upregulation of the immune transcriptional response, it would be helpful to have a least a little bit of information on why recruitment to the wound is important.

This is a very good question. We do not know how or why the localisation of the amine transporter SNF-12 would affects its signalling. It is known that, to be functional, some signalling complexes need to be assembled on particular vesicular compartments (e.g. TLR7). In our system, SNF-12 localisation could potentially affect its interaction with the Stat transcription factor STA-2.

Understanding this will be a challenge for the future. We have altered the Discussion to address the point explicitly: “The fact that overexpression in the epidermis of the MT severing protein Spastin caused a similar effect on SNF-12 dynamics as knocking down tubulin expression, supports a role for MT in positioning SNF-12-containing vesicles. […] We speculate that this spatial regulation of signalling may contribute to restricting defence gene activation to the immediate vicinity of a wound or site of infection within the major epidermal syncytium.”

https://doi.org/10.7554/eLife.45047.sa2

Article and author information

Author details

  1. Clara Taffoni

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Present address
    Institute of Human Genetics, University of Montpellier, CNRS, Montpellier, France
    Contribution
    Formal analysis, Investigation, Visualization, Methodology, Writing - original draft
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9265-5909

  2. Shizue Omi

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2711-2016

  3. Caroline Huber

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Contribution
    Investigation, Visualization
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3767-2719

  4. Sébastien Mailfert

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Contribution
    Visualization, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0409-0432

  5. Mathieu Fallet

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Contribution
    Visualization, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8144-6159

  6. Jean-François Rupprecht

    CPT, Turing Centre for Living Systems, Aix Marseille Univ, CNRS, Marseille, France
    Contribution
    Software, Validation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8904-5878

  7. Jonathan J Ewbank

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Contribution
    Funding acquisition, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1257-6862

  8. Nathalie Pujol

    CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Aix Marseille Univ, INSERM, CNRS, Marseille, France
    Contribution
    Conceptualization, Formal analysis, Supervision, Funding acquisition, Validation, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    pujol@ciml.univ-mrs.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8889-3197

Funding

Agence Nationale de la Recherche (ANR-16-CE15-0001-01)

  • Clara Taffoni
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jonathan J Ewbank
  • Nathalie Pujol

Institut National de la Santé et de la Recherche Médicale

  • Jonathan J Ewbank

Centre National de la Recherche Scientifique

  • Shizue Omi
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jean-François Rupprecht
  • Nathalie Pujol

Aix-Marseille Université

  • Clara Taffoni
  • Caroline Huber

Agence Nationale de la Recherche (ANR-12-BSV3-0001-01)

  • Clara Taffoni
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jonathan J Ewbank
  • Nathalie Pujol

Agence Nationale de la Recherche (ANR-11-LABX-0054)

  • Clara Taffoni
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jonathan J Ewbank
  • Nathalie Pujol

Agence Nationale de la Recherche (ANR-11-IDEX-0001-02)

  • Clara Taffoni
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jonathan J Ewbank
  • Nathalie Pujol

Agence Nationale de la Recherche (ANR-10-INBS-04-01)

  • Clara Taffoni
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jonathan J Ewbank
  • Nathalie Pujol

Agence Nationale de la Recherche (ANR-16-CONV-0001)

  • Shizue Omi
  • Sébastien Mailfert
  • Mathieu Fallet
  • Jean-François Rupprecht
  • Jonathan J Ewbank
  • Nathalie Pujol

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Annie Bonnet, Vincent Rouger, Guillaume Bordet, Sham Tlili, Claudio Collinet and Pierre Mangeol for their contributions, Asako Sugimoto, Andrew Chisholm, Michel Labouesse, Sander van den Heuvel, Michalis Barkoulas, Bart Grant, Xiaochen Wang, Jessica Feldman, Reto Gassmann and the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MN) supported by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440) for strains, Ari Pani, Anne Cecile Rayman, Gregoire Michaux and Michael Sixt for plasmids, Chris Crocker at Wormatlas (supported by NIH OD010943 to David Hall) for diagrams, and Didier Marguet, Ariane Abrieu, Alphée Michelot, Thomas Lecuit and members of our lab for discussions and critical reading of the manuscript. Worm sorting was performed by Jerome Belougne using the facilities of the French National Functional Genomics platform, supported by the GIS IBiSA and Labex INFORM. We thank the imaging core facility (ImagImm) of the Centre d'Immunologie de Marseille-Luminy (CIML) supported by the French National Research Agency program (France-BioImaging ANR-10-INBS-04), and Centuri funded by the « Investissements d'Avenir » French Government program managed by the French National Research Agency (ANR-16-CONV-0001) and from Excellence Initiative of Aix-Marseille University - A*MIDEX.

Senior Editor

  1. Didier YR Stainier, Max Planck Institute for Heart and Lung Research, Germany

Reviewing Editor

  1. Terry Lechler, Duke University, United States

Reviewers

  1. William M Bement, University of Wisconsin-Madison, United States
  2. Andrew D Chisholm, University of California, San Diego, United States

Publication history

  1. Received: January 11, 2019
  2. Accepted: January 27, 2020
  3. Accepted Manuscript published: January 29, 2020 (version 1)
  4. Version of Record published: February 26, 2020 (version 2)

Copyright

© 2020, Taffoni et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 2,588
    Page views
  • 407
    Downloads
  • 16
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Clara Taffoni
  2. Shizue Omi
  3. Caroline Huber
  4. Sébastien Mailfert
  5. Mathieu Fallet
  6. Jean-François Rupprecht
  7. Jonathan J Ewbank
  8. Nathalie Pujol
(2020)
Microtubule plus-end dynamics link wound repair to the innate immune response
eLife 9:e45047.
https://doi.org/10.7554/eLife.45047

Categories and tags

Research organism

Further reading

    1. Cell Biology
    2. Developmental Biology

    DAPK interacts with Patronin and the microtubule cytoskeleton in epidermal development and wound repair

    Marian Chuang, Tiffany I Hsiao ... Andrew D Chisholm
    Research Article Updated

    Epidermal barrier epithelia form a first line of defense against the environment, protecting animals against infection and repairing physical damage. In C. elegans, death-associated protein kinase (DAPK-1) regulates epidermal morphogenesis, innate immunity and wound repair. Combining genetic suppressor screens and pharmacological tests, we find that DAPK-1 maintains epidermal tissue integrity through regulation of the microtubule (MT) cytoskeleton. dapk-1 epidermal phenotypes are suppressed by treatment with microtubule-destabilizing drugs and mimicked or enhanced by microtubule-stabilizing drugs. Loss of function in ptrn-1, the C. elegans member of the Patronin/Nezha/CAMSAP family of MT minus-end binding proteins, suppresses dapk-1 epidermal and innate immunity phenotypes. Over-expression of the MT-binding CKK domain of PTRN-1 triggers epidermal and immunity defects resembling those of dapk-1 mutants, and PTRN-1 localization is regulated by DAPK-1. DAPK-1 and PTRN-1 physically interact in co-immunoprecipitation experiments, and DAPK-1 itself undergoes MT-dependent transport. Our results uncover an unexpected interdependence of DAPK-1 and the microtubule cytoskeleton in maintenance of epidermal integrity.

    1. Cell Biology

    Global analysis of contact-dependent human-to-mouse intercellular mRNA and lncRNA transfer in cell culture

    Sandipan Dasgupta, Daniella Y Dayagi ... Jeffrey E Gerst
    Research Article Updated

    Full-length mRNAs transfer between adjacent mammalian cells via direct cell-to-cell connections called tunneling nanotubes (TNTs). However, the extent of mRNA transfer at the transcriptome-wide level (the ‘transferome’) is unknown. Here, we analyzed the transferome in an in vitro human-mouse cell co-culture model using RNA-sequencing. We found that mRNA transfer is non-selective, prevalent across the human transcriptome, and that the amount of transfer to mouse embryonic fibroblasts (MEFs) strongly correlates with the endogenous level of gene expression in donor human breast cancer cells. Typically,<1% of endogenous mRNAs undergo transfer. Non-selective, expression-dependent RNA transfer was further validated using synthetic reporters. RNA transfer appears contact-dependent via TNTs, as exemplified for several mRNAs. Notably, significant differential changes in the native MEF transcriptome were observed in response to co-culture, including the upregulation of multiple cancer and cancer-associated fibroblast-related genes and pathways. Together, these results lead us to suggest that TNT-mediated RNA transfer could be a phenomenon of physiological importance under both normal and pathogenic conditions.

    1. Cell Biology
    2. Neuroscience

    Repair of noise-induced damage to stereocilia F-actin cores is facilitated by XIRP2 and its novel mechanosensor domain

    Elizabeth L Wagner, Jun-Sub Im ... Jung-Bum Shin
    Research Article

    Prolonged exposure to loud noise has been shown to affect inner ear sensory hair cells in a variety of deleterious manners, including damaging the stereocilia core. The damaged sites can be visualized as ‘gaps’ in phalloidin staining of F-actin, and the enrichment of monomeric actin at these sites, along with an actin nucleator and crosslinker, suggests that localized remodeling occurs to repair the broken filaments. Herein, we show that gaps in mouse auditory hair cells are largely repaired within 1 week of traumatic noise exposure through the incorporation of newly synthesized actin. We provide evidence that Xin actin binding repeat containing 2 (XIRP2) is required for the repair process and facilitates the enrichment of monomeric γ-actin at gaps. Recruitment of XIRP2 to stereocilia gaps and stress fiber strain sites in fibroblasts is force-dependent, mediated by a novel mechanosensor domain located in the C-terminus of XIRP2. Our study describes a novel process by which hair cells can recover from sublethal hair bundle damage and which may contribute to recovery from temporary hearing threshold shifts and the prevention of age-related hearing loss.