CLIPR-59-deficient mice die at birth from respiratory failure. (A) Schematic representation of the wild-type and recombined CLIPR-59 loci after insertion of loxP sites and excision of exons 2-5 by Cre recombinase. (B) CLIPR-59 protein is undetectable by western blot in spinal cord lysates from E18.5 mutants, whereas the levels of CLIP-115 and CLIP-170 are similar to those of controls. (C) The genotype ratio in litters from heterozygous mating is Mendelian at different embryonic stages. (D) Following Cesarean section of pregnant mice and stimulation of the E18.5 embryos by tail pinching, mutants are never able to breathe, become cyanosed and die. By contrast, control embryos are able to breathe and their open lungs are evident by thoracic transparency (arrow).

Innervation of the diaphragm is impaired in Clipr-59^-/- embryos from E15.5 to E18.5. Anti-peripherin and α-bungarotoxin were used to label nerve intermediate filaments and nAChRs, respectively before macroscopic imaging of diaphragm whole mounts (orientation of the diaphragm: D, dorsal; V, ventral; R, right; L, left). (A) The innervation patterns of the diaphragms are similar for control and mutant embryos until E15.5, as the dorsal and ventral primary branches of the phrenic nerve reach both muscle extremities. (B) In the absence of CLIPR-59, the innervation pattern becomes incomplete from E15.5 until E18.5, in particular in the ventral region of the diaphragm muscle. (C) At E18.5, both ventral and dorsal branches of phrenic nerves do not reach the most distal part of the mutant diaphragm where clusters of nAChRs are normally distributed (arrows). (D) At E15.5, the dorsal and ventral nerve branches in embryos of all three genotypes exhibit no significant differences in length. (E) At E18.5 the ventral and dorsal primary branches of the phrenic nerve are both significantly shorter in the mutant. (F) In the right hemi-diaphragm, the lengths of the major secondary nerve branches are longer in Clipr-59^-/- mice. (G) In the ventral part of the left hemi-diaphragm, the band of nAChR clusters is slightly but not significantly larger in mutants. Data are means±s.e.m. ^*P<0.05; ^**P<0.001; Mann-Whitney U test. NS, non significant. Scale bars: 500 μm.

The isometric contraction force of the diaphragm muscle is impaired in CLIPR-59-deficient embryos at E18.5. (A) In the mutant, the isometric twitch tension elicited by single nerve stimulation is markedly reduced compared with that obtained by single direct muscle stimulation. (B) The percentage of maximum contraction as a function of the stimulation frequency are similar in nerve-stimulated hemi-diaphragms from both control and mutant embryos. (C) Example of sustained tetanic contraction in diaphragms obtained by nerve stimulation at 40 Hz for 6 seconds show that the tension dropped faster in the absence of CLIPR-59 (black) than in the control (gray). (D) The percentage of maximal contraction is more substantially reduced for Clipr-59^-/- embryos than in Clipr-59^+/- embryos after 1.5 seconds of nerve stimulation (solid bars), and to an even greater extent after 6 seconds of nerve stimulation (hatched bars). (E) An example of the tension recording graphs obtained by a direct tetanic stimulation at 40 Hz for 6 seconds of the muscle in presence of d-tubocurarine shows that the contraction patterns for mutant (black) and control (gray) diaphragms are very similar. (F) The percentage of maximal contraction upon direct stimulation remains high after 1.5 and 6 seconds for diaphragms from either Clipr-59^-/- or Clipr-59^+/- embryos. Data are mean±s.e.m.

Selective and premature denervation of ear muscles in Clipr-59^-/- mice. Two muscle layers located between the ears of the mouse and innervated by a branch of the facial nerve were studied for innervation defects. (A) The properties of these ear muscles are summarized using information taken from Murray et al. [^*(Murray et al., 2008); ^**(Murray et al., 2010b)]. (B) Innervation was revealed with anti-peripherin labeling, and the nerve trunk was highlighted by a color-coded broken line. The orientation of the muscles on the mouse head is indicated in the center of the figure (F, front; L, left ear). The innervation pattern of the superficial layer is severely impaired in E18.5 mutants. In particular the LALr (orange) is denervated in distal regions of the muscle, with fewer nerve branches in the mutant compared with the control, whereas the innervation of the LALc (red) is relatively normal in mutant mice. (C) In the muscle deep layer of E18.5 mutant, innervation is more affected in the AAL (a fast-twitch muscle, purple) with a severe axon loss and thinner axon branches than in the AS (slow-twitch muscle, blue). (D) Innervation defects in mutant ear muscles are already noticeable as early as E15.5, with thinner and less ramified nerve branches. Scale bars: 100 μm.

NMJs appear to have formed normally but are not maintained in distal regions of the diaphragm at E18.5. (A) Nerve terminals and nAChRs were labeled with anti-peripherin (red) and α-bungarotoxin (green), respectively, in diaphragms and imaged by confocal microscopy. (B) Terminal Schwann cells were labeled with anti-S-100 (red) and neuronal tubulin was labeled with a Tuj1 antibody (blue). This series of images shows the ventral region of the left hemi-diaphragm of E18.5 embryos at the core of the nerve branch and at its end. (A) In the core of the nerve branch, normal NMJs are evident and the presynaptic nerve terminals are apposed to post-synaptic nAChRs in control (a) and mutant (b) diaphragms. At the tip of the nerve branch (c,d), the thinning of the nerve is much more pronounced in the absence of CLIPR-59 than in the control, with visible retraction bulbs (arrowhead) and uninnervated nAChRs (star). (B) S-100-positive terminal Schwann cells capping the NMJs both in control and in CLIPR-59-deficient mice. In the core of the nerve branch upstream of the nerve tip, Schwann cell morphologies are very similar in control (a) and CLIPR-59-deficient (b) diaphragms. However, at the tip of the nerve branch in CLIPR-59-deficient embryos (d), abnormally shaped Schwann cells can be observed (thick arrows) next to shedding segments of disconnected axons (arrowheads). By comparison, normally shaped Schwann cells cap healthy NMJs in control diaphragms (c). Scale bars: 10 μm.

The ultrastructure is modified in some NMJs of E18.5 mutant diaphragms. Presynaptic nerve terminals and terminal Schwann cells were artificially colored in light red and green, respectively. (A) Bundle of nerve terminals capped by a Schwann cell process in a control NMJ. Numerous pre-synaptic synaptic vesicles and some mitochondria (asterisks) are present in nerve terminals. (B,C) Clipr-59 mutant NMJs showing either the chain-like arrangement of nerve terminals (B), or the single arrangement of the terminal (C) covered by thin Schwann cell processes. Both types of nerve terminals exhibit small clear synaptic vesicle accumulation and mitochondria. (D) Clipr-59 mutant NMJ exhibiting abnormal nerve terminals (stars), totally wrapped up in Schwann cell processes containing few synaptic vesicles. Scale bars: 1 μm.