Doublecortin-domain containing (DCDC) genes play key roles in the normal and pathological development of the human brain cortex. The origin of the cellular specialisation and the functional redundancy of these microtubule (MT)-associated proteins (MAPs), especially those of Doublecortin (DCX) and Doublecortin-like kinase (DCLKs) genes, is still unclear. The DCX domain has the ability to control MT architecture and bundling. However, the physiological significance of such properties is not fully understood. To address these issues, we sought post-mitotic roles for zyg-8, the sole representative of the DCX–DCLK subfamily of genes in C. elegans. Previously, zyg-8 has been shown to control anaphase-spindle positioning in one-cell stage embryos, but functions of the gene later in development have not been investigated.

Here we show that wild-type zyg-8 is required beyond early embryonic divisions for proper development, spontaneous locomotion and touch sensitivity of adult worms. Consistently, we find zyg-8 expression in the six touch receptor neurons (TRNs), as well as in a subset of other neuronal and non-neuronal cells. In TRNs and motoneurons, zyg-8 controls cell body shape/polarity and process outgrowth and morphology. Ultrastructural analysis of mutant animals reveals that zyg-8 promotes structural integrity, length and number of individual MTs, as well as their bundled organisation in TRNs, with no impact on MT architecture.

The Doublecortin (DCX) family of Microtubule-Associated Proteins (MAPs) includes the X-linked lissencephaly and doublecortex syndrome DCX gene, the Doublecortin-like kinase DCLK genes, the retinitis pigmentosa RP1 gene as well as the dyslexia-associated DCDC2 gene (for a review, see Reiner et al., 2006). This family of genes is conserved in animals, including mammals, fish, insects and nematodes. The mouse genome harbors 11 such genes, complicating functional analysis of these proteins in vivo. Nevertheless, considerable effort has been focused on understanding the physiopathology of DCX domain-containing proteins (DCDCs) in the mammalian nervous system, especially that of the Doublecortin and DCLKs subgroup. Thus, in mice, DCX and DCLK-1 function redundantly in several types of cortical neurons to ensure their migration as well as the control of axon growth and wiring during embryonic life (Deuel et al., 2006; Koizumi et al., 2006). In addition, DCX and DCLK-1 display earlier functions in neurogenesis (Pramparo et al., 2010; Shu et al., 2006) and DCLK-2 expression persists in mature neurons, in which it may play roles at the growth cone (Edelman et al., 2005).

Doublecortin promotes selective assembly and stability of 13-protofilament (13-pf) MTs in vitro, through the maintenance of lateral and longitudinal contacts between adjacent protofilaments (Fourniol et al., 2010; Moores et al., 2004). Many DCDCs also induce microtubule bundling when overexpressed in mammalian cells, suggesting that this family of proteins may also mediate intermolecular interactions between adjacent MTs (Bielas et al., 2007; Coquelle et al., 2006). To date, however, the contribution of DCDC proteins to MT organisation in post-mitotic cells such as neurons is poorly documented.

We address these issues by analysing zyg-8 in the nematode Caenorhabditis. elegans. The rationale for this choice is twofold. First, C. elegans has a simple and well-characterised neuroanatomy and powerful genetic tools allow for straightforward analyses of gene function at a whole organism scale. Second, zyg-8 is the unique representative of the DCX–DCLK subfamily in C. elegans, allowing for the analysis of evolutionary conserved functions of this group of genes. Previous studies established the essential role of zyg-8 in mitotic spindle positioning of one-cell stage C. elegans embryos, a function highly reminiscent of that of DCLK-1 in dividing murine neuronal precursor cells (Gönczy et al., 2001; Shu et al., 2006). Part of the DCLK properties may thus have been retained throughout metazoan evolution, although post-embryonic functions of zyg-8 have not been investigated to date, due to early lethality of homozygous mutant embryos.

Here we show that zyg-8 is required throughout gastrulation for proper development and survival of embryos, and later on for spontaneous locomotion and touch sensitivity of adult animals. We find that zyg-8 is expressed in several classes of neurons, including motoneurons and touch receptor neurons (TRNs), as well as in some non-neuronal tissues. We show that zyg-8 is required to maintain neuronal cell shape/polarity and process outgrowth/wiring in developing worms. In addition, we demonstrate that zyg-8 promotes bundling, length and structural integrity of the atypical 15-pf MTs that fill TRN processes, without affecting MT architecture. Therefore, in C. elegans neurons, zyg-8 recapitulates most functions attributed to DCX–DCLK genes in the mammalian brain, strongly suggesting an ancient origin for these properties.

zyg-8 is required beyond the first division for full embryonic survival

zyg-8 null mutant animals display a penetrant, maternal-effect embryonic lethality, likely resulting from the dramatic anaphase spindle positioning defects in one-cell stage C. elegans embryos (Gönczy et al., 2001; Wood et al., 1980). But, it is not known what functions zyg-8 has in development and physiology beyond the first mitotic division. To address this question, we took advantage of two conditional, temperature-sensitive alleles, zyg-8(or484) and zyg-8(or490), which encode proteins mutated in the DCX (G219E) and kinase (G491E) domains, respectively (Gönczy et al., 2001). Although we cannot formally preclude neomorphic actions of these mutations, we believe this is unlikely, i.e. the two alleles rather reflect loss of function of zyg-8 because (i) in early embryos, zyg-8(or484) and zyg-8(or490) display phenotypes indistinguishable from those of genetically null alleles (Gönczy et al., 2001), and (ii) in fully developped animals, the two mutations result in similar defects (this study).

We used temperature-shift experiments to investigate the contribution of zyg-8 to embryonic survival (Materials and Methods; supplementary material Fig. S1). This analysis revealed that zyg-8 function is required after the first division for full embryonic viability but not beyond gastrulation (supplementary material Fig. S1C). Thus, zyg-8 may control most, if not all of the mitotic divisions during the first hours of development, but appears to be dispensable for subsequent mitoses. Mitosis may not be the only function for zyg-8 since it is expressed throughout development and persists in several post-mitotic tissues in adult animals (see below). To study these functions, we prepared zyg-8 mutant adults derived from the eggs of zyg-8(or484) or zyg-8(or490) parents shifted to 25°C after the completion of gastrulation at the permissive temperature.

zyg-8 is expressed by neuronal and non-neuronal cells

To determine where zyg-8 functions in adult animals, we analysed the expression pattern of a Pzyg-8::YFP transcriptional fusion that encompasses 2 kb of the 5′ regulatory region of zyg-8 (Materials and Methods). The progeny of injected animals displayed YFP expression in both neuronal and non-neuronal structures, from late embryonic stages to adulthood. Post-embryonic neurons include a few unidentified cells clustered near the nerve ring, some of them projecting dendrites anteriorly like amphid neurons (Fig. 1A1, star and red arrowhead, respectively), numerous motoneurons in the ventral nerve cord (Fig. 1A1, white arrowheads) and, most obviously the six TRNs: ALML/R, PLML/R, AVM and PVM (Fig. 1A1,2, yellow arrowheads). The TRNs sense slight mechanical stimuli applied to the cuticle and transduce them into electrical currents for proper locomotory responses (Chalfie, 2009). In addition, the biolistic integration of our reporter construct led to the expression of YFP in the germline (Fig. 1A3), which was expected, from both genetic and molecular data (Gönczy et al., 2001), as well as in hypodermal cells (Fig. 1A4). The latter expression pattern was unexpected considering that other known DCDC genes act in neurons, but two distinct Pzyg-8-ZYG-8::YFP translational fusion reporters (Materials and Methods) displayed the same profile (data not shown).

Fig. 1.

zyg-8 is expressed in C. elegans neurons. (A) Confocal microscopy imaging of Pzyg-8::YFP transgenic animals generated by classical microinjection (1 and 2) or biolistic bombardment (3 and 4). Yellow arrowheads point to touch receptor neurons, white arrowheads to motoneurons along the ventral nerve cord, red arrowhead to anterior projections of unidentified head neurons marked by a white star. (B) Confocal microscopy images of three fixed N2 adult animals immunostained with affinity-purified anti-ZYG-8 rabbit antibodies (upper panels), monoclonal 6–11B-1 antibodies that specifically recognise acetylated tubulin (AcTub, middle panels). Lower panels are merged images (ZYG-8 in red, AcTub in green, colocalization in yellow).

Fig. 1.

zyg-8 is expressed in C. elegans neurons. (A) Confocal microscopy imaging of Pzyg-8::YFP transgenic animals generated by classical microinjection (1 and 2) or biolistic bombardment (3 and 4). Yellow arrowheads point to touch receptor neurons, white arrowheads to motoneurons along the ventral nerve cord, red arrowhead to anterior projections of unidentified head neurons marked by a white star. (B) Confocal microscopy images of three fixed N2 adult animals immunostained with affinity-purified anti-ZYG-8 rabbit antibodies (upper panels), monoclonal 6–11B-1 antibodies that specifically recognise acetylated tubulin (AcTub, middle panels). Lower panels are merged images (ZYG-8 in red, AcTub in green, colocalization in yellow).

To determine where ZYG-8 accumulates in the animal and confirm the TRN expression, we performed immunolocalization studies, using affinity-purified ZYG-8 antibodies, together with anti-acetylated tubulin (AcTub) monoclonal antibodies that preferentially decorate TRN processes (Fukushige et al., 1999). As shown in Fig. 1B, ZYG-8 colocalized with AcTub, indicating that ZYG-8 is present in touch cells and distributed along the length of their neurites. Anti-ZYG-8 antibodies also marked the gonad and other structures that include ventral nerve cords (data not shown).

Thus, in addition to its expected germline expression, zyg-8 is transcribed in hypodermal and neural tissues. Consistent with a role for zyg-8 in hypodermis function, mutant animals were slightly shorter than wild-type adults (the Dumpy or Dpy phenotype) and some animals displayed a roller (Rol) phenotype in which animals twist around their antero-posterior axis while crawling (data not shown). Below, we focus on characterising the neuronal functions of zyg-8.

zyg-8 function is required in adults for proper locomotion and touch sensitivity

We observed that zyg-8 mutant animals were lethargic (Fig. 2A), a phenotype that has been linked to synaptic defects and is also evident in animals with defects in TRN function (Syntichaki and Tavernarakis, 2004). We used video recordings of wild-type and mutant adults to evaluate the coordination and speed of locomotion. zyg-8 mutants left sinusoidal tracks in the bacterial lawn indistinguishable from that of wild-type animals and retained their ability to respond to harsh touch. However, they were immobile for long periods of time, displayed reduced speed when moving and, as a consequence, reduced net displacement when compared to wild type (Fig. 2A). Thus, wild-type zyg-8 is required for spontaneous motion of adults.

Fig. 2.

zyg-8ts mutant animals are lethargic and touch insensitive. (A) Still images from representative movies (3 min long, 1 image/sec) of adult animals with the indicated genotypes (N2: wild type). The position of the tip of each animal was recorded over 180 frames and is overlaid in blue on each panel. Peak velocity (V), average velocity (v) and lethargy ratio (L, corresponding to the fraction of time of total immobility during the 3 min recording) are indicated for each animal. Note the sinusoidal traces left in the bacterial lawn by mutant animals, which reflect their normal coordination. Similar results were obtained for at least 35 animals per genotype. (B) Touch sensitivity of wild-type and zyg-8 mutant animals shifted to 25°C as embryos or adults. At least 20 animals were tested for each condition, except for adults shifted for 44h (bar 6) for which 13 animals were tested. Note that mutants were less touch sensitive at 15°C and that a 20-h-long temperature shift has no effect (bars 4 and 5).

Fig. 2.

zyg-8ts mutant animals are lethargic and touch insensitive. (A) Still images from representative movies (3 min long, 1 image/sec) of adult animals with the indicated genotypes (N2: wild type). The position of the tip of each animal was recorded over 180 frames and is overlaid in blue on each panel. Peak velocity (V), average velocity (v) and lethargy ratio (L, corresponding to the fraction of time of total immobility during the 3 min recording) are indicated for each animal. Note the sinusoidal traces left in the bacterial lawn by mutant animals, which reflect their normal coordination. Similar results were obtained for at least 35 animals per genotype. (B) Touch sensitivity of wild-type and zyg-8 mutant animals shifted to 25°C as embryos or adults. At least 20 animals were tested for each condition, except for adults shifted for 44h (bar 6) for which 13 animals were tested. Note that mutants were less touch sensitive at 15°C and that a 20-h-long temperature shift has no effect (bars 4 and 5).

Next, we asked whether zyg-8 mutant adults also exhibit a touch-insensitive phenotype, using a classical gentle body-touch assay (Materials and Methods). When assayed in parallel, zyg-8 mutant animals were less touch sensitive than control animals, but more sensitive than mec-7 beta tubulin mutants (Fig. 2B, compare lanes 1 and 3). Thus, an intact zyg-8 gene is needed for full touch sensitivity.

We next tested whether this touch insensitivity phenotype reflects a requirement for zyg-8 during developing or in mature TRNs. As shown in Fig. 2B (lanes 4–6), shifting young zyg-8(or484) adults from 15°C to 25°C led to a severe reduction of their touch response, providing the shift to 25°C was long enough. Similar results were obtained with zyg-8(or490) animals and mutant larvae (data not shown). Since TRNs are fully specified and drive a functional mechanotransducing network by the end of the L4 larval stage, this result suggests that zyg-8 is required for the physiology of fully mature neurons.

zyg-8 controls morphology and growth of C. elegans neurons

MTs are essential for diverse aspects of neuronal biology such as cell division, neurogenesis, migration, vesicular transport, cell shape, polarity and process outgrowth, and guidance (Mollinedo and Gajate, 2003). We thus probed zyg-8 function further by analysing the position and morphology of TRNs as well as that of the well-characterised D-type motoneurons. These 19 ventral cord neurons (13 VDs and 6 DDs) express the inhibitory neurotransmitter GABA and extend processes within both the ventral and and dorsal nerve bundles of C. elegans to perform their function as cross-inhibitors of body-wall muscle contraction (McIntire et al., 1993). Using appropriate GFP markers, we observed significant morphological defects in mutants compared to control animals (Fig. 3; Tables 1, 3).

Fig. 3.

zyg-8 controls various aspects of neuronal morphology in C. elegans. (A) TRNs visualised in live animals expressing Pmec-4::GFP. (1) Combined DIC and GFP fluorescene image of an adult animal, highlighting positions and projections of the six touch cells. (2–10) Confocal microscopy imaging of representative TRNs from wild-type (2–4, ctr) and zyg-8(or484) animals (5–10, zyg-8). Panels 2, 5 and 8 illustrate ALM (plus one AVM in panel 2) neurons, panels 3, 6 and 9 PVM neurons, panels 4, 7 and 10 PLM neurons. Yellow arrowheads point to varicosities, yellow arrows to ectopic process outgrowth, red arrows to misshaped cell bodies, and the horizontal bracket marks a trajectory defect. A quantification of these defects is provided in Table 1. (B) D-type motoneuron morphology visualised with Punc-25::GFP or Pflp13::GFP transgenes and D neuron synapses visualised with the presynaptic vesicle marker Punc-25-snb-1::GFP. (1) GFP image showing positions and projections of the 13 VD and 6 DD motoneurons in young adult animals. (2) Diagram of the morphology of a single DD motoneuron with synapses along dorsal muscles. (3–12) Representative D neurons from wild-type (3–5, ctr) and zyg-8 mutant animals (6–12, zyg-8). Projections of wild-type D-type motoneurons in the dorsal nerve cord (3) and D-type neuron cell bodies and processes along the ventral nerve cord (4) are shown. Varicosities (yellow arrows, see text for details) and thinning (yellow bracket) along the dorsal cord of zyg-8(or484) animals are depicted in panel 6, and axon fasciculation and morphology defects (red arrows) in the ventral nerve cord of zyg-8(or490) worms are illustrated in panel 7. Dorsal synapses of the DD neurons are visualised by SNB-1::GFP expression in wild-type (5) and zyg-8(or490) animals (8 and 9). Additional defects observed in zyg-8(or484) mutants include DD migration defects (10) and ectopic synapses formed by a misplaced VD axon in the nose/head (12). Panel 11 depicts a newly hatched zyg-8(or484) L1 larvae in which the DD neurons extend normally along the dorsal cord, but each cell lacks a posterior dendrite along the ventral cord. A quantification of these defects is provided in Table 3. In this figure and the following ones, anterior is on the left side and the position of the vulva is marked by a white asterisk.

Fig. 3.

zyg-8 controls various aspects of neuronal morphology in C. elegans. (A) TRNs visualised in live animals expressing Pmec-4::GFP. (1) Combined DIC and GFP fluorescene image of an adult animal, highlighting positions and projections of the six touch cells. (2–10) Confocal microscopy imaging of representative TRNs from wild-type (2–4, ctr) and zyg-8(or484) animals (5–10, zyg-8). Panels 2, 5 and 8 illustrate ALM (plus one AVM in panel 2) neurons, panels 3, 6 and 9 PVM neurons, panels 4, 7 and 10 PLM neurons. Yellow arrowheads point to varicosities, yellow arrows to ectopic process outgrowth, red arrows to misshaped cell bodies, and the horizontal bracket marks a trajectory defect. A quantification of these defects is provided in Table 1. (B) D-type motoneuron morphology visualised with Punc-25::GFP or Pflp13::GFP transgenes and D neuron synapses visualised with the presynaptic vesicle marker Punc-25-snb-1::GFP. (1) GFP image showing positions and projections of the 13 VD and 6 DD motoneurons in young adult animals. (2) Diagram of the morphology of a single DD motoneuron with synapses along dorsal muscles. (3–12) Representative D neurons from wild-type (3–5, ctr) and zyg-8 mutant animals (6–12, zyg-8). Projections of wild-type D-type motoneurons in the dorsal nerve cord (3) and D-type neuron cell bodies and processes along the ventral nerve cord (4) are shown. Varicosities (yellow arrows, see text for details) and thinning (yellow bracket) along the dorsal cord of zyg-8(or484) animals are depicted in panel 6, and axon fasciculation and morphology defects (red arrows) in the ventral nerve cord of zyg-8(or490) worms are illustrated in panel 7. Dorsal synapses of the DD neurons are visualised by SNB-1::GFP expression in wild-type (5) and zyg-8(or490) animals (8 and 9). Additional defects observed in zyg-8(or484) mutants include DD migration defects (10) and ectopic synapses formed by a misplaced VD axon in the nose/head (12). Panel 11 depicts a newly hatched zyg-8(or484) L1 larvae in which the DD neurons extend normally along the dorsal cord, but each cell lacks a posterior dendrite along the ventral cord. A quantification of these defects is provided in Table 3. In this figure and the following ones, anterior is on the left side and the position of the vulva is marked by a white asterisk.

Table 1.
Touch receptor neuron morphology and outgrowth
Genotype (25°C) Morphology defects (% animals, n = 25) 
Cell body Process 
Nb, position Shapea Polarityb (ALM) Varicositiesc (ALM; PLM) Trajectoryd (ALM; PVM; PLM) Ectopic branch Lateral growthe (ALM; PLM) 
N2; zdIs5 16 24 8; 0 8; 0; 0 16 0; 0 
zyg-8(or484); zdIs5 40 88 20; 32 8; 36; 20 32 56; 44 
zyg-8(or490); zdIs5 40 92 72; 100 4; 36; 0 12 48; 60 
Genotype (25°C) Morphology defects (% animals, n = 25) 
Cell body Process 
Nb, position Shapea Polarityb (ALM) Varicositiesc (ALM; PLM) Trajectoryd (ALM; PVM; PLM) Ectopic branch Lateral growthe (ALM; PLM) 
N2; zdIs5 16 24 8; 0 8; 0; 0 16 0; 0 
zyg-8(or484); zdIs5 40 88 20; 32 8; 36; 20 32 56; 44 
zyg-8(or490); zdIs5 40 92 72; 100 4; 36; 0 12 48; 60 

N2 (wild type) and zyg-8 alleles were crossed to zdIs5 [Pmec-4::GFP] worms to visualize touch receptor neuron morphology. Wild-type and mutant worms were cultured at 15°C as described in Materials and Methods. Sets of N2, or484 and or490 late embryos were shifted to 25°C and 25 adult animals of each genotype were analyzed.

a

Most of the cell bodies in control animals were round and regular in shape, as illustrated in Fig. 3A2–4. Defects scored here include irregular, elongated to spiky shapes.

b

In our hands, control animals typically had bipolar ALM neurons due to a posterior extension longer than the cell body. zyg-8 mutations reduced this posterior process. Occasionally, the polarity of other TRNs was altered in zyg-8 mutants, as illustrated in Fig. 3A9 (PVM).

c

Most of the TRN processes in control animals were straight and regular, as illustrated in Fig. 3A2–4. In contrast, zyg-8 mutant animals exhibited repeated inflated areas along the processes of mostly PLM and AML, as illustrated in Fig. 3A5 and 3A7 (yellow arrowheads).

d

Although TRN processes followed a straight ventral or anterior path in control animals, their trajectory in zyg-8 mutants was altered, as exemplified by panel 6 of Fig. 3A. It is worth noting, however, that we could not find any case of missing connections.

e

See Fig. 3A5, yellow arrow for an ALM cell.

Table 3.
D neuron morphology and outgrowth
Genotype °C Morphology defects (% animals) 
Cell bodies Commissures Fasciculation 
N2; juIs76 15 0/30 (0) 0/30 (0) 0/30 (0) 
N2; juIs76 25 0/50 (0) 0/50 (0) 1/50 (2) 
zyg-8(or484); juIs76 15 0/22 (0) 2/22 (9) 4/22 (18) 
zyg-8(or484); juIs76 25 2/25 (8) 3/25 (12) 6/25 (24) 
zyg-8(or490); juIs76 15 0/28 (0) 2/28 (7) 2/28 (7) 
zyg-8(or490); juIs76 25 6/31 (19) 5/31 (16) 23/31 (74) 
Genotype °C Morphology defects (% animals) 
Cell bodies Commissures Fasciculation 
N2; juIs76 15 0/30 (0) 0/30 (0) 0/30 (0) 
N2; juIs76 25 0/50 (0) 0/50 (0) 1/50 (2) 
zyg-8(or484); juIs76 15 0/22 (0) 2/22 (9) 4/22 (18) 
zyg-8(or484); juIs76 25 2/25 (8) 3/25 (12) 6/25 (24) 
zyg-8(or490); juIs76 15 0/28 (0) 2/28 (7) 2/28 (7) 
zyg-8(or490); juIs76 25 6/31 (19) 5/31 (16) 23/31 (74) 

N2 (wild type) and zyg-8 alleles were crossed to juIs76 [Punc25::GFP; lin-15] worms to visualize GABAergic DD and VD motor neuron morphology. Wild-type and mutant worms were cultured at 15°C as described in Materials and Methods. Sets of N2, or484 and or490 late embryos were shifted to 25°C while corresponding controls remained at 15°C. Six DD neurons extend axons during embryogenesis, and 13 VD neurons are born in L1 larvae and extend processes along the ventral and dorsal cord after the animals are shifted to 25°C. Adult worms were scored for alterations in cell body size and shape, commissures that were misplaced or extended along the wrong side of the body, and defasciculation of axons along the ventral and dorsal nerve cords.

First, although TRN neurons were properly positioned in the vast majority of cases, cell bodies were often misshapen (red arrows in Fig. 3A) and displayed polarity defects, as assessed by the higher rate of monopolar ALM neurons in mutant worms compared to control (∼90% versus 24%, Table 1). Misshapen cell bodies were also noted in mutant D neurons at lower frequency (8% of or484 and 19% of or490 mutants). PVM displayed defective ventral projections in 36% of cases in both zyg-8(or484) and zyg-8(or490) strains (yellow bracket in Fig. 3A), and about 50% of the mutant animals exhibited shortened lateral branches along ALM or PLM neurons (yellow arrows in Fig. 3A), suggesting that in wild-type touch cells, zyg-8 restricts ectopic process outgrowth.

Second, we observed with high penetrance (up to 100% for PLMs in zyg-8(or490) animals) the formation of numerous varicosities along some D-neuron, ALM and PLM processes (yellow arrowheads in Fig. 3, panels B6, A5 and A7). These varicosities are potential indications of transport issues. However, our analysis of the distribution of the mechanotransduction channel subunits MEC-4 and MEC-2, as well as that of the synaptic vesicle marker GFP-RAB-3, does not support the hypothesis of a global disruption of axonal transport in zyg-8 mutant TRNs (Fig. 4, Table 2, respectively).

Fig. 4.

zyg-8 does not regulate protein levels and transport in TRNs. (A) Western blot analysis of MEC-18, MEC-4, MEC-2 protein content in extracts prepared from wild-type and zyg-8(or484) mutant adult worms. α-Tubulin was used as a loading control and MEC-18 is recognised as a doublet. (B) Confocal microscopy imaging of TRNs in fixed representative wild-type (left panels) and zyg-8(or484) mutant (right panels) adult animals immunostained with affinity-purified anti-MEC-2 (top panels) or anti-MEC-4 (middle panels) polyclonal antibodies. Note on the merge image (lower panels) that MEC-2 (green) and MEC-4 (red) colocalize both in wild-type and mutant worms. Scale bars: 10 µm.

Fig. 4.

zyg-8 does not regulate protein levels and transport in TRNs. (A) Western blot analysis of MEC-18, MEC-4, MEC-2 protein content in extracts prepared from wild-type and zyg-8(or484) mutant adult worms. α-Tubulin was used as a loading control and MEC-18 is recognised as a doublet. (B) Confocal microscopy imaging of TRNs in fixed representative wild-type (left panels) and zyg-8(or484) mutant (right panels) adult animals immunostained with affinity-purified anti-MEC-2 (top panels) or anti-MEC-4 (middle panels) polyclonal antibodies. Note on the merge image (lower panels) that MEC-2 (green) and MEC-4 (red) colocalize both in wild-type and mutant worms. Scale bars: 10 µm.

Table 2.
Touch receptor neuron synaptic vesicles
Genotype (L4s)a °C Synaptic vesicle defects (% animals, n = 35) 
Nerve ring puncta PLM puncta 
Nb Diffuse/localisation Size/spacing Diffuse/localisationa 
N2; jsIs821 15 14 14 
N2; jsIs821 25 17 29 
zyg-8(or484); jsIs821 15 32 40 
zyg-8(or484); jsIs821 25 37 43 29 
zyg-8(or490); jsIs821 15 26 20 
zyg-8(or490); jsIs821 25 43 46 23 
Genotype (L4s)a °C Synaptic vesicle defects (% animals, n = 35) 
Nerve ring puncta PLM puncta 
Nb Diffuse/localisation Size/spacing Diffuse/localisationa 
N2; jsIs821 15 14 14 
N2; jsIs821 25 17 29 
zyg-8(or484); jsIs821 15 32 40 
zyg-8(or484); jsIs821 25 37 43 29 
zyg-8(or490); jsIs821 15 26 20 
zyg-8(or490); jsIs821 25 43 46 23 

N2 (wild type) and zyg-8 alleles were crossed to [Pmec-7::GFP-RAB-3] worms to visualize touch receptor neuron synaptic vesicles. Wild-type and mutant worms were cultured at 15°C as described in Materials and Methods. Sets of N2, or484 and or490 late embryos were shifted to 25°C. Thirty-five worms of each genotype were analyzed as L4s, a larval stage at which the marker is more consistent and reliable.

a

In wild-type L4 animals, GFP–RAB-3 appears as two strong puncta in the VNC, posterior to the vulva. Animals were scored for the presence of diffuse GFP or additional very small puncta mislocalized along the VNC or in the PLM dendrites.

Third, we documented frequent alterations of process growth and wiring in the D neurons, consistent with disorganization of the ventral nerve cord. Seventy-four percent of zyg-8(or490) animals displayed defasciculation of D-neurons along the nerve cords, and commissures were missing or extended along the wrong side of the body in 19% of these animals (Table 3). In several zyg-8 mutant animals, a D axon with SNB-1::GFP clusters that is normally restricted to body wall muscles in wild-type worms, extended into the head region (Fig. 3, panel B12, arrows). Many zyg-8(or484) and zyg-8(or490) mutants exhibited small gaps or thinning in GFP expression along ventral or dorsal nerve cords indicating that some axons failed to extend their full length (Fig. 3, panels B7, B10 and B11, compare to B4). Regions of altered expression of the presynaptic vesicle marker Punc-25–snb-1::GFP (Hallam and Jin, 1998) were also observed along the dorsal cord of zyg-8 mutants at 25°C (Fig. 3, panels B8 and B9, compare to panel B5). Although outgrowth of the six DD neurons occurs during embryogenesis, DD synapses form en passant along the dorsal muscles only at the end of the L1 larval stage. These synapses increase in size by about 30% as the animals grow to their full adult size, and new membrane and synapses are added along the axon (Yeh et al., 2005). Variations in the snb-1::GFP expression pattern could be due to abnormal axon morphology, but may also reflect defective transport during axon and synapse growth.

Collectively, these data indicate that an intact zyg-8 gene is required for neuronal morphology and polarity, restricts ectopic process outgrowth and may also promote axon and synapse growth. Because these aspects of neuronal development are known to depend on MTs, our findings suggest that zyg-8 has a role in maintaining neuronal MT integrity.

zyg-8 does not affect global protein levels and transport in touch receptor neurons

Touch receptor neuron MTs are unique within C. elegans in at least two respects: they display a larger diameter due to their 15-protofilament architecture (versus 11 in the rest of the animal) and they are organised into dense bundles that fill touch cell processes (Chalfie and Thomson, 1979). Since 15-pf MT bundles are required for efficient mechanotransduction (Chalfie, 2009), TRNs constitute an exquisite model system to investigate the structural and physiological role of a MAP on MT dynamics and organisation. Fifteen-protofilament MTs affect touch sensitivity in several ways that include the control of protein levels and the transport of important components of the mechanotransduction machinery (Bounoutas et al., 2009a). We thus investigated the impact of zyg-8 mutations on protein levels and the distribution of three TRN-specific components: MEC-18, MEC-2 and MEC-4. Antibodies to all three proteins recognised similar amonts of protein in extracts prepared from wild-type and zyg-8(or484) animals (Fig. 4A), suggesting that zyg-8 does not affect protein levels. The cytoplasmic distribution of MEC-18 in mutant TRNs was indistinguishable from that in wild-type cells (data not shown) and MEC-2 and MEC-4 displayed overlapping punctate patterns along TRN processes, with no gross reduction in the size or number of individual puncta (Fig. 4B). Together, these findings indicate that the loss of touch sensitivity in zyg-8 mutant animals is unlikely to result from altered protein levels or from mislocalization of mechanotransduction channels.

To further probe putative roles of zyg-8 in axonal transport, we examined the distribution of the synaptic vesicle marker GFP-RAB-3 (Bounoutas et al., 2009b) in mutant TRNs. Our analysis revealed a small increase in defective shape and position of GFP-RAB-3 puncta in the nerve ring compared to controls, and more diffuse GFP and ectopic puncta were observed in both or484 and or490 genotypes (Table 2). A fourfold increase in the number of animals with ectopic puncta in PLM processes was also noted for both zyg-8 alleles (Table 2). These defects are compatible with weak transport and polarity defects, so that not all synaptic vesicles are tightly clustered at synapses. Although they could contribute to the Mec phenotype of zyg-8 mutant animals, the primary synaptic output of the mechanosensory neurons is via gap junctions, and mutants that exhibit significant defects in differentiation of chemical synapses are not necessarily Mec (Zheng et al., 2011). Moreover, these defects in synaptic vesicle localisation do not reflect global transport failure, as assessed by the proper local delivery of MEC-2 and MEC-4 (Fig. 4B).

zyg-8 maintains the integrity, number, length and bundling of 15-pf MTs

To test whether zyg-8 promotes MT stabilisation and to gain insight in the corresponding mechanism, we compared the phenotypes induced by zyg-8 loss of function and treatment with the MT-depolymerizing drug colchicine. As shown in Fig. 5A, zyg-8 loss of function and colchicine caused distinct morphological phenotypes. Indeed, process varicosities observed on TRN processes of zyg-8 mutants were never found upon colchicine treatment (compare Fig. 3A7 and Fig. 5A1–4 for instance). Instead, we noted a colchicine dose-dependent decrease in the Pmec-4::GFP signal, a phenotype never observed in zyg-8 mutants. In addition, we observed strong synthetic effects of zyg-8 loss of function and colchicine, both in the morphology (Fig. 5A5–8) and function (Fig. 5B) of TRNs. TRN processes of zyg-8 mutant animals treated with low doses of colchicine consistently developed numerous ectopic secondary branches (yellow arrows) that could further branch and cell bodies were often jagged, stellate and disconnected from their physiological targets, as illustrated in panel A6 for instance. The functional consequences of those synthetic effects are best evidenced at 0.1 mM colchicine, a drug concentration that reduces body-touch response of zyg-8 mutant animals by 10-fold versus only 2-fold in control animals (Fig. 5B).

Fig. 5.

zyg-8 loss of function synergizes with colchicine-elicited MT depolymerization in TRNs. (A) Confocal microscopy imaging of representative GFP-expressing TRNs from wild-type (1–4, ctr) or zyg-8(or484) (5–8) adult animals grown on 0.1 mM colchicine. 1 and 6: AVM; 3 and 5: ALM; 2 and 7: PVM; 4 and 8: PLMs. Yellow arrows point to ectopic process outgrowth/branching, red arrows to misshaped cell bodies, red arrowheads to trajectory defects and red stars mark process arborizations. Compare these images with Fig. 3A5–10 to appreciate the synthetic effects of colchicine and zyg-8 loss of function on TRN morphology. (B) Gentle body-touch assay. At least 20 animals were tested under each condition.

Fig. 5.

zyg-8 loss of function synergizes with colchicine-elicited MT depolymerization in TRNs. (A) Confocal microscopy imaging of representative GFP-expressing TRNs from wild-type (1–4, ctr) or zyg-8(or484) (5–8) adult animals grown on 0.1 mM colchicine. 1 and 6: AVM; 3 and 5: ALM; 2 and 7: PVM; 4 and 8: PLMs. Yellow arrows point to ectopic process outgrowth/branching, red arrows to misshaped cell bodies, red arrowheads to trajectory defects and red stars mark process arborizations. Compare these images with Fig. 3A5–10 to appreciate the synthetic effects of colchicine and zyg-8 loss of function on TRN morphology. (B) Gentle body-touch assay. At least 20 animals were tested under each condition.

Because colchicine affects MT stability by preventing polymerization, the synergistic effects of zyg-8 loss of function and colchicine suggest that zyg-8 promotes MT stability through mechanisms at least partially distinct from the sole control of polymer dynamics. To gain further insights in such mechanism in TRNs, we proceeded to the ultrastructural analysis of these neurons in zyg-8 mutant adults. High-pressure frozen (HPF) animals prepared for transmission electron microscopy (Materials and Methods) revealed several microtubule defects in mutant TRNs (Fig. 6). First and most obviously, the average number of MTs that normally fill TRN processes was reduced. On average, there were 12.6±3.5 (N = 14 touch receptor neurons) microtubules per cross-section in zyg-8 mutants compared to 25.3±3.0 (N = 25 touch receptor neurons) in wild-type animals raised at 25°C. Thus, wild-type zyg-8 is required for the assembly or maintenance of the large MT array that are hallmarks of the TRNs.

Fig. 6.

zyg-8 controls MT stability and organisation in TRNs. (A) (1–3) Transmission electron micrographs of cross sections of a TRN (PLM) with the indicated genotypes. Arrowheads mark the TRN-associated extracellular matrix and arrows mark ordinary (11-pf) MTs found in hypodermal cells. Scale bar: 100 nm. (4–6) High magnification views of four selected MTs for each genotype showing protofilament organisation. Note the unaltered15-pf content of TRN MTs in zyg-8 mutants compared to those of wild-type animals. Scale bar: 25 nm. (B) Bar graph representing the average number of MTs per TRN in wild-type and zyg-8(or484) animals. (C) Probability distribution of nearest neighbour inter-MT distances in TRNs of wild-type and zyg-8(or484) animals. The median values are indicated by black and blue arrowheads and are 39.0 and 34.6 nm, respectively. The distributions are significantly different (P<0.01, KS test). (D) Bar graph representing the average microtubule length per TRNs of wild-type and zyg-8(or484) animals, as determined from serial section datasets and digital alignment (Materials and Methods). (E) Bar graph representing the average diameter of MTs per TRN in wild-type and zyg-8(or484) animals. *Statistically significant difference from wild type (P<0.01, Student’s t-test). N = Number of TRNs scored, n = number of individual MTs (E) or MT-MT distances (C) scored.

Fig. 6.

zyg-8 controls MT stability and organisation in TRNs. (A) (1–3) Transmission electron micrographs of cross sections of a TRN (PLM) with the indicated genotypes. Arrowheads mark the TRN-associated extracellular matrix and arrows mark ordinary (11-pf) MTs found in hypodermal cells. Scale bar: 100 nm. (4–6) High magnification views of four selected MTs for each genotype showing protofilament organisation. Note the unaltered15-pf content of TRN MTs in zyg-8 mutants compared to those of wild-type animals. Scale bar: 25 nm. (B) Bar graph representing the average number of MTs per TRN in wild-type and zyg-8(or484) animals. (C) Probability distribution of nearest neighbour inter-MT distances in TRNs of wild-type and zyg-8(or484) animals. The median values are indicated by black and blue arrowheads and are 39.0 and 34.6 nm, respectively. The distributions are significantly different (P<0.01, KS test). (D) Bar graph representing the average microtubule length per TRNs of wild-type and zyg-8(or484) animals, as determined from serial section datasets and digital alignment (Materials and Methods). (E) Bar graph representing the average diameter of MTs per TRN in wild-type and zyg-8(or484) animals. *Statistically significant difference from wild type (P<0.01, Student’s t-test). N = Number of TRNs scored, n = number of individual MTs (E) or MT-MT distances (C) scored.

Second, we observed in zyg-8 mutant TRNs, a notable disorganization of MTs that failed to cluster in the regular bundle-like arrays found in wild-type neurons (Fig. 6A1–3). In a geometrical point of view, the level of organisation of these structures can be accounted for by the variability of the distance separating adjacent MTs within the array: in principle, a regular bundle has normal (Gaussian) distribution of nearest-neighbour inter-MT distances, while a disorganised bundle will exhibit a non-normal distribution. Consistent with the idea that zyg-8 is required to properly organise the MT bundle, the distribution of inter-MT distances differ in wild-type and zyg-8(or484) mutant TRNs (Fig. 6C). In addition, MTs are closer to each other in zyg-8 mutant TRNs, as revealed by the median of the distributions: 34.6 nm versus 39.0 nm, respectively. We propose from this analysis that zyg-8 plays key roles in the establishment or maintenance of the regular, bundled organisation of TRN MTs.

Third, the analysis and digital alignment of serial sections encompassing up to 2.4 µm of body length (Materials and Methods) revealed a 5-fold decrease in MT length in TRNs of mutant animals (Fig. 6D): on average, wild-type TRN MTs were 14.8±1.4 µm long whereas in zyg-8(or484) animals, this length was reduced to 3.0±0.4 µm. We conclude that, as a MAP, wild-type zyg-8 is needed to maintain long MTs in TRNs.

Fourth, when looking at high magnifications of individual MT sections, we observed no difference in protofilament content between zyg-8 mutant and wild-type animals (Fig. 6A4–6), indicating that, in the context of TRNs, the DCDC gene zyg-8 is dispensable for the assembly or maintenance of 15-pf MTs. However, we also noticed on the same images, a slight but significant reduction of MT caliber in mutant TRNs, as well as a lower signal-to-noise ratio that hinders the sharp resolution of individual protofilaments displayed on wild-type sections (compare Fig. 6A4 to Fig. 6A5,6 and Fig. 6E). The gene may also be dispensable for regulating the protofilament content of the 11-pf MTs found in other neurons, including those in the ventral nerve cord (supplementary material Fig. S2).

Altogether, our ultrastructural analysis establishes that, in TRNs, wild-type ZYG-8 is needed to promote the integrity of the 15-pf MT array through the maintenance of MT number, MT length and MT bundling. The gene is dispensable for the control of pf content in vivo although it may still contribute to the structural integrity of individual MTs.

zyg-8 physiology at a whole organism scale

We show here that zyg-8 is expressed in C. elegans, in post-mitotic neurons where it contributes to most of the functions attributed to DCDC genes in mammals, including cellular morphology, axon outgrowth and guidance as well as synaptic vesicle localisation. These functions may reflect underlying roles of the gene in the establishment or maintenance of neuronal polarity. Additional phenotypic and genetic analyses will be necessary to determine if there are signalling pathways connecting zyg-8 directly to the polarity machinery, or if these phenotypes are secondary to functions of zyg-8 in MT dynamics, stabilisation or bundling. However, our findings indicate that the pleiotropic functions of DCDC genes in the cortex of mammals are likely to be of ancient origin rather than resulting from recent sub-specialisation of these MAPs upon cerebrocortical tissue compartmentalization.

In addition to these functions that may rely on intrinsic properties of the DCX domain, our study reveals an unanticipated role for zyg-8 in mechanosensation. A recent study in fly sensory cilia showed that the Drosophila DCDC protein DmELP-1 is implicated in mechanotransduction (Bechstedt et al., 2010). DmELP-1 is a multidomain protein containing both a DCX domain and a domain related to the Echinoderm Microtubule-Associated Protein (EMAP) family of proteins. It is not known whether the contribution of DmELP-1 to mechanotransduction is conferred by its DCX or EMAP-like domains. Because the C. elegans EMAP, CeELP-1, is expressed in TRNs and is required for touch sensation (Hueston et al., 2008), it may be concluded from these two studies that EMAP-like proteins display conserved function in mechanosensation in a way that might not involve the DCX domain. Our work suggests that the DCX domain displays the ability to stabilise MT-based mechanosensors. In this context, it would be interesting to test whether ZYG-8 and CeELP-1 functionally cooperate in touch neurons and recapitulate the activity of DmELP-1 in mechanotransduction. The stabilising function of DCDC proteins on large MT-based structures may not be restricted to invertebrate cells, as a recent report established the functional association of DCDC2 with primary cilia in rat hippocampal neurons (Massinen et al., 2011). The presence in the inner ear of mammals of specialised cells filled with complex MT structures that display both unusual architecture and extensive bundling and that are essential to the conversion of sounds into electrical currents, should stimulate closer inspection of the expression of DCDC genes in this tissue (Tolomeo and Holley, 1997).

Most eukaryotic mechanotransducing systems involve large cytoskeletal structures that are thought to constitute a rigid support towards mechanical deformations and contribute to gating transducing ion channels (Syntichaki and Tavernarakis, 2004). Both microtubule internal architecture and intermolecular organisation such as their bundling may affect the mechanical properties of these structures, by a direct effect on the intrinsic rigidity of the polymer or by an increase of the overall stiffness of the system (Tolomeo and Holley, 1997). However, MT architecture and MT bundling are often intimately linked, so that defining the relative contribution of these two levels of MT organisation in mechanosensitive responses is not an easy task. Caenorhabditis elegans TRNs represent a useful model system to tackle this issue because the 15-pf MT bundles that fill their processes are unique within the animal and their physiological requirement is genetically tractable. (In nematodes, ordinary MTs are built from 11 pfs and are not clustered in bundles). Chalfie and colleagues demonstrated that 15-pf MT bundles play specific role in mechanosensation, in part by insuring normal protein levels of components of the transduction machinery and transport of mechanosensory channel subunits along the TRN processes (Bounoutas et al., 2009a). However, because 11-pf MTs do not form bundles, no conclusion could be drawn regarding the physiological relevance of MT architecture and MT bundling.

Here we show that zyg-8 is needed for full mechanosensation, with no impact on protein levels or axonal transport in TRNs, suggesting that the gene plays rather direct roles in the process. Our findings also show that MT architecture and MT bundling are submitted to separate regulatory controls. Indeed, zyg-8 mutant animals display disorganised TRN MT bundles, with fewer and shorter MTs, but no disruption of the 15-pf MT architecture. Therefore, it could be concluded that maintaining 15-pf MTs in C. elegans TRNs is not sufficient, per se, to ensure proper mechanosensation and that MT bundling plays crucial role in that process. Although we do not have the ability to measure the stiffness of TRN MT bundles, our findings are compatible with a scenario in which both the 15-pf MT architecture and MT bundling contribute to the rigidity of the array, which, in turn, defines the level of responsiveness of the system.

Our reporter gene analysis suggests that zyg-8 is somatically transcribed in non-neuronal tissues such as the hypodermis. Although the presence of the ZYG-8 protein remains to be demonstrated in these cells, the Rol and Dpy phenotypes of zyg-8 mutant animals are compatible with a role of the gene in that tissue. The C. elegans hypodermis performs many functions during early development, including establishing the basic body plan, depositing basement membrane components, regulating cell fate specification of neighbouring cells, guiding cell and axon migrations, and taking up apoptotic cell bodies by phagocytosis (http://www.wormatlas.org/hermaphrodite/hypodermis/Hypframeset.html). Microtubules are likely to participate to most of these developmental processes but to date, their physiological control has not been thoroughly investigated in this tissue. Interestingly, in mammals, individual cells of the olfactory epithelium also express DCX and DCLKs (Reiner et al., 2006), suggesting a possible ancestral function of these genes in epithelia formation. More work will be necessary to shed light on putative non-neuronal functions of DCDC genes, but C. elegans and zyg-8 will undoubtedly be instrumental in such a quest.

Neuronal zyg-8 functions as a MT stabiliser and organiser

The ultrastructural study that we performed allowed us to precisely and quantitatively describe the MT stabilising role of zyg-8, in the physiological context of a developing organism. We show that the gene controls MT number, MT bundling, MT length, but with no associated modification of pf number. One may ask whether these activities are interdependent or rather involve distinct properties of ZYG-8. MT bundling and number may be governed by independent processes, as we consistently find organised MTs in wild-type TRNs, even in the ones that contain less than 10 MTs (data not shown). Conversely, we hardly find well-organised MT bundles in the TRNs of zyg-8 mutant animals, even in those displaying 20–40 MTs (data not shown).

In zyg-8(or484) mutant animals, the average MT content of TRNs is reduced by nearly twofold and MTs are, on average, fivefold shorter than their wild-type counterparts. This suggests that zyg-8 displays important stabilising and/or nucleating functions in TRNs. In one-cell stage C. elegans embryos, zyg-8 also affects spindle MT growth and nucleation and loss of its function can be phenocopied by the MT-depolymerizing drug nocodazole (Gönczy et al., 2001; Srayko et al., 2005). Therefore, zyg-8 may display similar regulatory activities towards MTs in vivo, both in mitotic and post-mitotic contexts.

One of the most striking defects of zyg-8 mutant TRNs is the disorganization of MTs that fail to cluster as a regular bundle. Consistent with this idea, inter-MT distance varies much more in mutant TRNs than wild-type neurons. In addition, MTs are closer to each other in mutant cells, indicating that ZYG-8 is unlikely to hold bundled MTs together. Our data are more compatible with a model in which ZYG-8 controls the regular spacing between adjacent MTs. Doublecortin-domain containing proteins, including ZYG-8, are usually considered as MT bundling factors, based on their ability to elicit these large MT structures when expressed in cultured cells, but our analysis suggests that more subtle functions of these proteins may be unravelled in systems where MT bundles constitute the physiological state, such as C. elegans TRNs (Coquelle et al., 2006; Gönczy et al., 2001).

Moores and colleagues have dissected the mechanism of MT stabilisation by human DCX (Moores et al., 2004; Moores et al., 2006). Notably, they established that DCX selectively promotes the assembly of 13-pf MTs (the major MT architecture in mammals) and stabilises this architecture in vitro. This would result from the unique binding site of DCX in between pfs, where it would ideally control inter-pf lateral contacts and constrain pf number within MTs. DCX may also stabilise longitudinal interactions along pfs in addition to stapling them laterally (Fourniol et al., 2010). This would account for the potent anti-catastrophe and nucleation activities of DCX towards mammalian MTs in vitro, and, by extension, predict similar behaviours of DCDC proteins in various systems. However, the model does not predict the architectural properties of endogenous DCDCs in physiological systems deprived of 13-pf MTs. Do they retain and adapt their stapling function to MT architectures with less or more than 13 pfs or is this activity governed by the geometry of 13-pf MTs? In this context, C. elegans constitutes a model of choice because its cells are filled with either 11 pf-MTs or 15-pf MTs. Our findings that a DCDC protein is expressed in C. elegans TRNs and other neurons indicates that the DCX domain per se is not sufficient to constrain any type of MT to adopt the 13-pf architecture in vivo. Because zyg-8 function is dispensable to the assembly or maintenance of both types of MT architectures in worms, it may be concluded that the stapling model proposed by Moores and colleagues may not apply to all DCDC proteins or is restricted to 13-pf MTs. While it has been suggested in the literature that 15-pf MTs are the result of specific alpha and beta tubulins (Chalfie and Thomson, 1982), respectively MEC-12 and MEC-7, these isotypes are not exclusive to the TRNs. Potential alternative mechanisms include both other MAPs and post-translational modifications like α-tubulin acetylation, recently shown to be needed for the formation of uniform bundles of 15-pf microtubules in C. elegans touch receptor neurons (Cueva et al., 2012; Topalidou et al., 2012). zyg-8 has no paralog in C. elegans genome but two unrelated MAPs, PTL-1 and ELP-1, have been shown to be expressed in TRNs and required for mechanosensation (Gordon et al., 2008; Hueston et al., 2008). Although the role of these proteins on MT architecture remains to be demonstrated, they represent interesting candidates to supply zyg-8 loss of function.

Despite the lack of a strict requirement for zyg-8 in the maintenance of TRN MT architecture, we measure a slight but significant decrease of MT diameter in mutant animals, associated to less defined protofilaments when compared to wild type. Because the animals have been grown, fixed and processed for TEM analysis the same way in the same conditions, we believe these subtle defects may reflect cryptic alterations of the structure. Taken together, our ultrastructural findings demonstrate that an intact zyg-8 gene acts to maintain MT length, number and organisation in TRNs.

Nematode strains and transgenesis

C. elegans (N2) culture was performed according to standard methods (Brenner, 1974). zyg-8 mutant alleles were described previously (Gönczy et al., 2001). In the course of this study, we outcrossed the two zyg-8ts alleles zyg-8(or484) and zyg-8(490) from their original lin-2 background, through conventional genetic crosses. The resulting strains were subsequently crossed with SK4005 zdIs5 (Pmec-4::GFP), NM2689 jsIs821 (Pmec-7::GFP-RAB-3) (Bounoutas et al., 2009b), CZ333 juIs1 (Punc-25::GFP-SNB-1) (Zhen and Jin, 1999), juIs145 (Pflp-13::GFP) (Sakaguchi-Nakashima et al., 2007), and CZ1200 juIs76 (Punc-25::GFP) (Huang et al., 2003) integrated transgenic lines to examine, respectively, touch receptor neurons or D-type motoneuron morphology upon zyg-8 loss of function. The CB1477 mec-7(e1343)X strain (Savage et al., 1989) was used as positive control in gentle body-touch assays.

Transgenesis was performed by conventional microinjection methods (Mello et al., 1991) and by biolistic bombardment on a Bio-RadPDS-1000/He Hepta System, following previously reported procedures (Praitis et al., 2001) optimized by C. Canard (P. Gönczy lab).

Constructs and zyg-8 promoter validation

The YFP reporter construct Pzyg-8::YFP was engineered by subcloning 2 kb of genomic DNA sequence located immediately upstream of the predicted +1 codon of zyg-8 in the BamHI/NheI sites of the Addgene plasmid 1662. This construct was subsequently used as a backbone to insert the zyg-8 full-length cDNA sequence at the NheI/KpnI sites and thus generate Pzyg-8-zyg-8c::YFP. We also created the intron-1-containing Pzyg-8-zyg-8g::YFP construct by replacing the 1.65 kb BglII/PasI fragment from Pzyg-8-zyg-8c::YFP by the corresponding zyg-8 genomic sequence. All these constructs have been verified by sequencing. In addition, we made use of the previously described Ppie1::GFP-zyg-8 plasmid, that has been shown to partially rescue the Maternal-effect Embryonic Lethality (MEL) of non-conditional, balanced, unc-32(e189) zyg-8(t1650) homozygous animals (Gönczy et al., 2001).

The four constructs, upon biolistic bombardment, line selection and western blot extract analyses, encoded proteins with the expected sizes that displayed epitopes for anti-GFP or anti-ZYG-8 antibodies (data not shown). Furthermore, when crossed with the unc-32(e189) zyg-8(t1650) strain, Pzyg-8-zyg-8g::YFP (as well as Ppie1::GFP-zyg-8) resulted in the segregation of living (Unc) progeny from zyg-8 homozygous Unc F2 animals (data not shown). This partial rescue of the MEL phenotype of zyg-8(t1650)−/− animals biologically validated our 2 kb regulatory sequence as a bona fide zyg-8 promoter sequence.

Anti-ZYG-8 antibodies

For this work, we generated new affinity-purified anti-ZYG-8 antibodies, essentially as described before, with few modifications (Bellanger and Gönczy, 2003). Briefly, the zyg-8 full-length cDNA was subcloned into a modified pMalC2X prokaryote expression vector and recombinant MBP-ZYG-8 was produced in E. coli BL21 cells, as soluble and insoluble (inclusion bodies) fractions. A rabbit was immunized with inclusion bodies while the soluble fraction was cleaved with Factor Xa. Cleaved products (250 µg) were separated by SDS-PAGE, transferred onto nitrocellulose and stained with Ponceau Red to cut a strip that encompassed the ∼100 kDa ZYG-8 band. We proceeded to affinity purification by incubating the strip with sera for 2 h at 4°C, then eluting specific antibodies with 0.1 M glycine pH 2.0 before neutralization with 1 M Tris pH 8.0.

Temperature-shift experiments

Temperature-shift experiments and progeny-tests were conducted as described previously (Bellanger et al., 2007). The end of the zyg-8 Temperature-Sensitive Period (TSP) was determined by collecting embryos at different developmental stages, and measuring their survival rate upon temperature shift. ‘Early’ embryos were collected by dissecting gravid hermaphrodites and thus correspond to 1- to 30-cell stage embryos approximately. Older embryos were collected as eggs laid over a 10 h period of time (at 15°C). They thus correspond to ∼30–400 cell stage embryos. Since (1) only ‘early’ embryos displayed embryonic lethality (supplementary material Fig. S1C), (2) gastrulation ends 3 h (at 25°C) after eggs have been laid and (3) we previously determined that the maternal pool of zyg-8(or484) inactivates within 12 h following their shift to 25°C, the zyg-8 TSP likely encompasses most of the gastrulation process.

Larvae and adult animals that hatched from eggs shifted to 25°C after the completion of gastrulation, displayed no or minor developmental defects. We noted slight Dumpy phenotypes and occasional rollers, suggesting a role for zyg-8 in the metabolism of collagen, beyond the scope of this study.

Gentle body-touch assay

Body touch sensitivity of worms was assayed as follows: 20–30 animals/condition were gently touched with an eyebrow hair glued to a toothpick, five times to the anterior body, alternately with five times to the posterior body. Control animals move backward when touched anteriorly and forward when stroked posteriorly. We scored as positive the animals that responded four or five times out of the five touches. Since we noticed higher variability in the anterior response, we chose not to pool both responses and to present only data corresponding to the posterior body-touch sensitivity. Fig. 2B and Fig. 5B illustrate the response of adult animals that have been shifted to 25°C as post-gastrulating embryos or adults, as indicated. CB1477 mec-7 animals and the MT-depolymerizing drug colchicine, were used as genetic and chemical positive controls in Fig. 2B and Fig. 5B, respectively (Chalfie and Thomson, 1982).

Microscopy

Locomotion and morphological defects were analysed on a Leica Z16 APO stereomicroscope. Lethargy was monitored at 25°C by recording adult mutant animals transferred onto fresh plates 15 min prior to starting the movies. Movies were 3 min long, 1 image per second. The position of the worm was marked on each frame to reconstruct its track and to extract three parameters: peak velocity ‘V’, averaged velocity ‘v’ and lethargy ratio ‘L’ (defined as the fraction of time of total immobility).

ZYG-8 immunofluorescence and YFP reporter studies were exclusively performed on a Zeiss Axioplan2/LSM510 Meta confocal microscope, at the Montpellier RIO Imaging facility. Optical slices (0.7–0.9 µm thick) were collected and used as such for colocalization studies, or z-stacked to reconstruct the whole animal thickness. Motoneuron GFP reporters were imaged using a Nikon E800 microscope equipped with Photometrics CoolsnapES CCD camera and Metamorph software.

Serial-section electron microscopy

Nematodes were raised at 25°C from hatching and prepared for electron microscopy by high-pressure freezing (HPF) followed by freeze substitution (in 2% glutaraldehyde, 1% osmium tetroxide), as described (Cueva et al., 2007). Following embedding in Eponate 12/Araldite 502, serial, 40-nm sections were cut and collected on Formvar-coated, copper-slot grids. A novel post-staining method, which will be described in detail elsewhere (J.G.C. and M.B.G., unpublished data), was used to enhance contrast and to visualise microtubule protofilaments. Sections were tilted 20°C, imaged on a TEM (JEOL TEM 1230, Tokyo, Japan) and digital images were acquired with an 11-megapixel bottom-mounted cooled CCD camera (Orius SC1000, Gatan, Pleasanton, California). Images from serial sections were manually aligned using Reconstruct (Fiala, 2005). Microtubule number was counted by hand and used to calculate MT length, as described (Chalfie and Thomson, 1979). The bundled organisation of TRN microtubules was estimated via the distribution of inter-MT distances within a given cell. Indeed, a perfect bundle displays regularly spaced MTs, associated with inter-MT distances that peak around the average value. In contrast, the variable MT-MT spacing that features disorganised MT geometries is predicted to yield a distribution of inter-MT distances more spread along the length axis. In practice and to compensate for the intracellular variability of MT organisation along TRN processes, we first selected, in each serial section, five or six images distant from a 300–500 nm step. On each image, we drew lines connecting the centre of every MT to the centre of its immediate neighbours. The length of each line was then determined, collected and plotted as a function of its relative abundance within the whole collection.

Indirect immunofluorescence

Fixation of nematodes for indirect immunofluorescence was performed according to established protocols (Gendrel et al., 2009), with few modifications introduced by M. Gendrel (J.-L. Bessereau lab). Briefly, worms were sandwiched between two SuperFrost Plus glass slides (ThermoScientific) that do not require any coating. The sandwich was then warmed for 5 sec on a hot aluminium block and immediately placed on another block that was sitting on dry ice for 20 min or more. Cuticles were cracked by splitting the two frozen slides apart with a razor blade, and open worms were immediately fixed and permeabilized in 3.2% paraformaldehyde and 0.1% Triton X-100 (ZYG-8 staining) or in cold acetone and methanol (other stainings). After washing, worms were split in as many tubes as intended conditions and further processed to immunostaining. The following primary antibodies were utilised: 1∶50 rabbit affinity-purified anti-ZYG-8, 1∶500 mouse anti-acetylated tubulin (6–11B-1, Sigma, kindly provided by C. Janke), 1∶100 rabbit anti-MEC-2 (kindly provided by A. Bounoutas, M. Chalfie lab), 1∶100 mouse anti-MEC-4 (ab22184, Abcam) and 1∶1000 rabbit anti-MEC-18 (kindly provided by A. Bounoutas, M. Chalfie lab). Secondary antibodies were 1∶1000 goat anti-mouse conjugated to Alexa Fluor 488 and 1∶1000 goat anti-rabbit conjugated to Alexa Fluor 546 (Molecular Probes).

Western blotting

Western blot analysis was performed as described (Bellanger and Gönczy, 2003) using nitrocellulose membranes (Protran, Schleicher and Schuell). Primary antibodies used were: 1∶1000 rabbit anti-GFP (Torrey Pines), 1∶2000 mouse anti-α-tubulin (DM1α, Sigma), 1∶100 rabbit anti-ZYG-8, 1∶500 rabbit anti-MEC-2 (kindly provided by A. Bounoutas, M. Chalfie lab), 1∶500 mouse anti-MEC-4 (ab22184, Abcam) and 1∶1000 rabbit anti-MEC-18 (kindly provided by A. Bounoutas, M. Chalfie lab). The secondary HRP-coupled anti-mouse and anti-rabbit antibodies were diluted in PBS and utilised at 1∶10,000. The ECL reaction was performed as recommended by the manufacturer (Amersham Pharmacia).

We thank Jean-Louis Bessereau, Jérôme Boudeau, Pierre Gönczy, Anne Houdusse, Lionel Pintard and Susanne Schmidt for critical reading of the manuscript, Sylvie Fromont for technical help in molecular biology, Riva DiGiacomo for help in strain construction, Montpellier RIO Imaging for confocal microscopy analyses, the antibody production facility of the CRBM and the Caenorhabditis Genetic Center for strains.

Funding

This work was supported by Centre National de la Recherche Scientifique institutional grants (to A.D.); National Institutes of Health [grant numbers NS047715 and EB 006745 to M.B.G., NSF 064221 to R.B.]. Deposited in PMC for release after 12 months.

Bechstedt
S.
,
Albert
J. T.
,
Kreil
D. P.
,
Müller–Reichert
T.
,
Göpfert
M. C.
,
Howard
J.
(
2010
).
A doublecortin containing microtubule-associated protein is implicated in mechanotransduction in Drosophila sensory cilia.
Nat. Commun.
1
,
11
.
Bellanger
J. M.
,
Gönczy
P.
(
2003
).
TAC-1 and ZYG-9 form a complex that promotes microtubule assembly in C. elegans embryos.
Curr. Biol.
13
,
1488
1498
.
Bellanger
J. M.
,
Carter
J. C.
,
Phillips
J. B.
,
Canard
C.
,
Bowerman
B.
,
Gönczy
P.
(
2007
).
ZYG-9, TAC-1 and ZYG-8 together ensure correct microtubule function throughout the cell cycle of C. elegans embryos.
J. Cell Sci.
120
,
2963
2973
.
Bielas
S. L.
,
Serneo
F. F.
,
Chechlacz
M.
,
Deerinck
T. J.
,
Perkins
G. A.
,
Allen
P. B.
,
Ellisman
M. H.
,
Gleeson
J. G.
(
2007
).
Spinophilin facilitates dephosphorylation of doublecortin by PP1 to mediate microtubule bundling at the axonal wrist.
Cell
129
,
579
591
.
Bounoutas
A.
,
O’Hagan
R.
,
Chalfie
M.
(
2009a
).
The multipurpose 15-protofilament microtubules in C. elegans have specific roles in mechanosensation.
Curr. Biol.
19
,
1362
1367
.
Bounoutas
A.
,
Zheng
Q.
,
Nonet
M. L.
,
Chalfie
M.
(
2009b
).
mec-15 encodes an F-box protein required for touch receptor neuron mechanosensation, synapse formation and development.
Genetics
183
,
607
617
.
Brenner
S.
(
1974
).
The genetics of Caenorhabditis elegans.
Genetics
77
,
71
94
.
Chalfie
M.
(
2009
).
Neurosensory mechanotransduction.
Nat. Rev. Mol. Cell Biol.
10
,
44
52
.
Chalfie
M.
,
Thomson
J. N.
(
1979
).
Organization of neuronal microtubules in the nematode Caenorhabditis elegans.
J. Cell Biol.
82
,
278
289
.
Chalfie
M.
,
Thomson
J. N.
(
1982
).
Structural and functional diversity in the neuronal microtubules of Caenorhabditis elegans.
J. Cell Biol.
93
,
15
23
.
Coquelle
F. M.
,
Levy
T.
,
Bergmann
S.
,
Wolf
S. G.
,
Bar–El
D.
,
Sapir
T.
,
Brody
Y.
,
Orr
I.
,
Barkai
N.
,
Eichele
G.
et al.  (
2006
).
Common and divergent roles for members of the mouse DCX superfamily.
Cell Cycle
5
,
976
983
.
Cueva
J. G.
,
Mulholland
A.
,
Goodman
M. B.
(
2007
).
Nanoscale organization of the MEC-4 DEG/ENaC sensory mechanotransduction channel in Caenorhabditis elegans touch receptor neurons.
J. Neurosci.
27
,
14089
14098
.
Cueva
J. G.
,
Hsin
J.
,
Huang
K. C.
,
Goodman
M. B.
(
2012
).
Posttranslational acetylation of α-tubulin constrains protofilament number in native microtubules.
Curr. Biol.
22
,
1066
1074
.
Deuel
T. A.
,
Liu
J. S.
,
Corbo
J. C.
,
Yoo
S. Y.
,
Rorke–Adams
L. B.
,
Walsh
C. A.
(
2006
).
Genetic interactions between doublecortin and doublecortin-like kinase in neuronal migration and axon outgrowth.
Neuron
49
,
41
53
.
Edelman
A. M.
,
Kim
W. Y.
,
Higgins
D.
,
Goldstein
E. G.
,
Oberdoerster
M.
,
Sigurdson
W.
(
2005
).
Doublecortin kinase-2, a novel doublecortin-related protein kinase associated with terminal segments of axons and dendrites.
J. Biol. Chem.
280
,
8531
8543
.
Fiala
J. C.
(
2005
).
Reconstruct: a free editor for serial section microscopy.
J. Microsc.
218
,
52
61
.
Fourniol
F. J.
,
Sindelar
C. V.
,
Amigues
B.
,
Clare
D. K.
,
Thomas
G.
,
Perderiset
M.
,
Francis
F.
,
Houdusse
A.
,
Moores
C. A.
(
2010
).
Template-free 13-protofilament microtubule-MAP assembly visualized at 8 A resolution.
J. Cell Biol.
191
,
463
470
.
Fukushige
T.
,
Siddiqui
Z. K.
,
Chou
M.
,
Culotti
J. G.
,
Gogonea
C. B.
,
Siddiqui
S. S.
,
Hamelin
M.
(
1999
).
MEC-12, an alpha-tubulin required for touch sensitivity in C. elegans.
J. Cell Sci.
112
,
395
403
.
Gendrel
M.
,
Rapti
G.
,
Richmond
J. E.
,
Bessereau
J. L.
(
2009
).
A secreted complement-control-related protein ensures acetylcholine receptor clustering.
Nature
461
,
992
996
.
Gönczy
P.
,
Bellanger
J. M.
,
Kirkham
M.
,
Pozniakowski
A.
,
Baumer
K.
,
Phillips
J. B.
,
Hyman
A. A.
(
2001
).
zyg-8, a gene required for spindle positioning in C. elegans, encodes a doublecortin-related kinase that promotes microtubule assembly.
Dev. Cell
1
,
363
375
.
Gordon
P.
,
Hingula
L.
,
Krasny
M. L.
,
Swienckowski
J. L.
,
Pokrywka
N. J.
,
Raley–Susman
K. M.
(
2008
).
The invertebrate microtubule-associated protein PTL-1 functions in mechanosensation and development in Caenorhabditis elegans.
Dev. Genes Evol.
218
,
541
551
.
Hallam
S. J.
,
Jin
Y.
(
1998
).
lin-14 regulates the timing of synaptic remodelling in Caenorhabditis elegans.
Nature
395
,
78
82
.
Huang
X.
,
Huang
P.
,
Robinson
M. K.
,
Stern
M. J.
,
Jin
Y.
(
2003
).
UNC-71, a disintegrin and metalloprotease (ADAM) protein, regulates motor axon guidance and sex myoblast migration in C. elegans.
Development
130
,
3147
3161
.
Hueston
J. L.
,
Herren
G. P.
,
Cueva
J. G.
,
Buechner
M.
,
Lundquist
E. A.
,
Goodman
M. B.
,
Suprenant
K. A.
(
2008
).
The C. elegans EMAP-like protein, ELP-1 is required for touch sensation and associates with microtubules and adhesion complexes.
BMC Dev. Biol.
8
,
110
.
Koizumi
H.
,
Tanaka
T.
,
Gleeson
J. G.
(
2006
).
Doublecortin-like kinase functions with doublecortin to mediate fiber tract decussation and neuronal migration.
Neuron
49
,
55
66
.
Massinen
S.
,
Hokkanen
M. E.
,
Matsson
H.
,
Tammimies
K.
,
Tapia–Páez
I.
,
Dahlström–Heuser
V.
,
Kuja–Panula
J.
,
Burghoorn
J.
,
Jeppsson
K. E.
,
Swoboda
P.
et al.  (
2011
).
Increased expression of the dyslexia candidate gene DCDC2 affects length and signaling of primary cilia in neurons.
PLoS ONE
6
,
e20580
.
McIntire
S. L.
,
Jorgensen
E.
,
Horvitz
H. R.
(
1993
).
Genes required for GABA function in Caenorhabditis elegans.
Nature
364
,
334
337
.
Mello
C. C.
,
Kramer
J. M.
,
Stinchcomb
D.
,
Ambros
V.
(
1991
).
Efficient gene transfer in C.elegans: extrachromosomal maintenance and integration of transforming sequences.
EMBO J.
10
,
3959
3970
.
Mollinedo
F.
,
Gajate
C.
(
2003
).
Microtubules, microtubule-interfering agents and apoptosis.
Apoptosis
8
,
413
450
.
Moores
C. A.
,
Perderiset
M.
,
Francis
F.
,
Chelly
J.
,
Houdusse
A.
,
Milligan
R. A.
(
2004
).
Mechanism of microtubule stabilization by doublecortin.
Mol. Cell
14
,
833
839
.
Moores
C. A.
,
Perderiset
M.
,
Kappeler
C.
,
Kain
S.
,
Drummond
D.
,
Perkins
S. J.
,
Chelly
J.
,
Cross
R.
,
Houdusse
A.
,
Francis
F.
(
2006
).
Distinct roles of doublecortin modulating the microtubule cytoskeleton.
EMBO J.
25
,
4448
4457
.
Praitis
V.
,
Casey
E.
,
Collar
D.
,
Austin
J.
(
2001
).
Creation of low-copy integrated transgenic lines in Caenorhabditis elegans.
Genetics
157
,
1217
1226
.
Pramparo
T.
,
Youn
Y. H.
,
Yingling
J.
,
Hirotsune
S.
,
Wynshaw–Boris
A.
(
2010
).
Novel embryonic neuronal migration and proliferation defects in Dcx mutant mice are exacerbated by Lis1 reduction.
J. Neurosci.
30
,
3002
3012
.
Reiner
O.
,
Coquelle
F. M.
,
Peter
B.
,
Levy
T.
,
Kaplan
A.
,
Sapir
T.
,
Orr
I.
,
Barkai
N.
,
Eichele
G.
,
Bergmann
S.
(
2006
).
The evolving doublecortin (DCX) superfamily.
BMC Genomics
7
,
188
.
Sakaguchi–Nakashima
A.
,
Meir
J. Y.
,
Jin
Y.
,
Matsumoto
K.
,
Hisamoto
N.
(
2007
).
LRK-1, a C. elegans PARK8-related kinase, regulates axonal-dendritic polarity of SV proteins.
Curr. Biol.
17
,
592
598
.
Savage
C.
,
Hamelin
M.
,
Culotti
J. G.
,
Coulson
A.
,
Albertson
D. G.
,
Chalfie
M.
(
1989
).
mec-7 is a beta-tubulin gene required for the production of 15-protofilament microtubules in Caenorhabditis elegans.
Genes Dev.
3
,
870
881
.
Shu
T.
,
Tseng
H. C.
,
Sapir
T.
,
Stern
P.
,
Zhou
Y.
,
Sanada
K.
,
Fischer
A.
,
Coquelle
F. M.
,
Reiner
O.
,
Tsai
L. H.
(
2006
).
Doublecortin-like kinase controls neurogenesis by regulating mitotic spindles and M phase progression.
Neuron
49
,
25
39
.
Srayko
M.
,
Kaya
A.
,
Stamford
J.
,
Hyman
A. A.
(
2005
).
Identification and characterization of factors required for microtubule growth and nucleation in the early C. elegans embryo.
Dev. Cell
9
,
223
236
.
Syntichaki
P.
,
Tavernarakis
N.
(
2004
).
Genetic models of mechanotransduction: the nematode Caenorhabditis elegans.
Physiol. Rev.
84
,
1097
1153
.
Tolomeo
J. A.
,
Holley
M. C.
(
1997
).
Mechanics of microtubule bundles in pillar cells from the inner ear.
Biophys. J.
73
,
2241
2247
.
Topalidou
I.
,
Keller
C.
,
Kalebic
N.
,
Nguyen
K. C.
,
Somhegyi
H.
,
Politi
K. A.
,
Heppenstall
P.
,
Hall
D. H.
,
Chalfie
M.
(
2012
).
Genetically separable functions of the MEC-17 tubulin acetyltransferase affect microtubule organization.
Curr. Biol.
22
,
1057
1065
.
Wood
W. B.
,
Hecht
R.
,
Carr
S.
,
Vanderslice
R.
,
Wolf
N.
,
Hirsh
D.
(
1980
).
Parental effects and phenotypic characterization of mutations that affect early development in Caenorhabditis elegans.
Dev. Biol.
74
,
446
469
.
Yeh
E.
,
Kawano
T.
,
Weimer
R. M.
,
Bessereau
J. L.
,
Zhen
M.
(
2005
).
Identification of genes involved in synaptogenesis using a fluorescent active zone marker in Caenorhabditis elegans.
J. Neurosci.
25
,
3833
3841
.
Zhen
M.
,
Jin
Y.
(
1999
).
The liprin protein SYD-2 regulates the differentiation of presynaptic termini in C. elegans.
Nature
401
,
371
375
.
Zheng
Q.
,
Schaefer
A. M.
,
Nonet
M. L.
(
2011
).
Regulation of C. elegans presynaptic differentiation and neurite branching via a novel signaling pathway initiated by SAM-10.
Development
138
,
87
96
.

Supplementary information