CED-4 CARD domain residues can modulate non-apoptotic neuronal regeneration functions independently from apoptosis

Abstract

A major challenge in regenerative medicine is the repair of injured neurons. Regeneration of laser-cut C. elegans neurons requires early action of core apoptosis activator CED-4/Apaf1 and CED-3/caspase. While testing models for CED-4 as a candidate calcium-sensitive activator of repair, we unexpectedly discovered that amino acid substitutions affecting alpha-helix-6 within the CED-4 caspase recruitment domain (CARD) confer a CED-4 gain-of-function (gf) activity that increases axonal regrowth without disrupting CED-4 apoptosis activity. The in vivo caspase reporter CA-GFP reveals a rapid localized increase in caspase activity upon axotomy, which is absent in ced-4 and ced-3 loss-of-function mutants but present in the ced-4(gf) mutant. The ced-3 loss-of-function mutation can significantly suppress the axonal regrowth of the ced-4(gf) mutant, indicating that CED-4(gf) regeneration depends on CED-3 caspase. Thus, we identified a subdomain within the CED-4 CARD that regulates the dynamic and controlled caspase activity required for efficient regeneration.

Figure 1

CED-4 EF-hand mutations do not disrupt CED-4 apoptosis functions. (a) Amino acid substitutions in the predicted CED-4 Ca²⁺-binding domains, EF-hand 1 (EFH1) and EF-hand 2 (EFH2) were introduced via CRISPR genome editing. Amino acids labeled green are potential Ca²⁺-binding amino acids; in each domain, we replaced two of these residues with alanine (red label). ced-4(∆) is a likely null allele that could potentially only express a partial CARD domain followed by a random short sequence (labeled purple). Actual allele names are: EFH1 ced-4(bz401); EFH2 ced-4(bz410); EFH(1 + 2) ced-4(bz406); ∆ced-4(bz404); CARD: caspase recruitment domain; HD: Helical domain; WHD: Winged-helix domain. (b) Representative pictures of the tail touch receptor neurons labeled by P_mec-4GFP in the background of various ced-4 alleles. In WT, two PLM tail neurons survive; in ced-4 null mutants, as many as two additional neurons survive and differentiate, so 3–4 cells may be visualized. (c) The PLM tail touch receptor neuron count for ced-4 mutants, n = 40 animals for each group in one trial. ****p < 0.0001 as compared to the wild type, Chi-square test.

Figure 2

The ced-4(EFH1) mutant exhibits increased axonal regrowth consequent to axotomy. (a) Depiction of measures of axonal regrowth from ALM touch receptor neurons. (b–d) The axonal regrowth measures for the ced-4 null mutant and the engineered EF-hand mutants. Total n numbers are given inside bars. All data come from >3 independent trials, *p ≤ 0.05, **p < 0.01, or ns (not significant) as compared to the ced-4(+) wild type, unpaired two-tail t-test (b,d) or one-way ANOVA with Tukey adjustment for multiple comparison (c); for (d), p = 0.0545.

Figure 3

A ced-3 mutant allele that disrupts the caspase active site suppresses the enhanced axonal regrowth of the ced-4(EFH1) mutant. (a) ced-3(n2433), lacking the caspase active site (indicated ced-3(lf)), exhibits reduced regeneration. Data come from >3 independent trials, *p < 0.05, unpaired t-test. (b) impact of ced-4(EFH1) on regeneration outcomes. Data come from >3 independent trials, **p < 0.01 or ****p < 0.0001, one-way ANOVA with Tukey adjustment for multiple comparison; numbers of animals scored indicated in bars. Separate graphs are given for time-separated trials.

Figure 4

In vivo caspase activity assay supports localized CED-3 activation consequent to axotomy, with CED-4(EFH1) capable of caspase activation. (a) Cartoon depiction of the CA-GFP reagent for in vivo measure of caspase activity. QP: quenching peptide, DEVD is the added consensus caspase cleavage site (purple). When the quenching peptide is removed, GFP fluorescence increases. (b) Average time-lapse of in-situ CA-GFP measurements for each region indicated. In our studies, we also included a co-expressed mCherry marker in the touch neurons to enable us to normalize GFP values at specific times after axotomy. Measurements are displayed as normalized fluorescence relative to the initial value. R stands for the ratio of CA-GFP/mCherry, and R_o is the R before laser severing, error bars represent standard error. (c) Images of in-situ CA-GFP measurements before laser axotomy at 0 min and after axotomy at 5 min and 30 min. The regions of axon near the point of axotomy (yellow boxes) are expanded for visual comparison (right). GFP signals are weak but not absent at baseline, especially in the soma. Image contrast has been enhanced to aid visualization. For accurate quantification, CA-GFP intensity measurements were normalized using simultaneous images of co-expressed mCherry (see Methods). (d) CA-GFP measurements of the severed axon near the point of axotomy, 30 min post axotomy. Measurements are displayed as normalized fluorescence relative to the initial value (R_o, before laser severing). *p < 0.05 or ns (not significant), one-way ANOVA with Tukey adjustment for multiple comparisons.

Figure 5

CED-4 S81 and E88 face away from the CED-3 or CED-9 interaction sites that regulate apoptosis, but might be available to interact with a factor that licenses transient CED-3 activation for localized repair. (a) Ribbon structure shows a CED-4 dimer (most residues in tan, but CARD domain in green with S81 and E88 highlighted in red), with two interacting CED-3 fragment residues (indicated in blue) as determined from a crystal structure. Note that the CARD domain is positioned apart from the CED-3 interaction domain and that S81 and E88 are oriented away from the main structure such that the subdomain that includes S81 and E88 might interact with another protein (or another region of a dynamic CED-4). (b) Ribbon structure shows a partial CED-4 dimer with one CARD domain depicted (CARD domain green, with S81 and E88 in red), binding to the CED-9 protein (magenta). Note that the CARD domain helix 6 containing S81 and E88 is positioned apart from the CED-9 interaction domain, with those AA positions facing outward from the protein. (c) A proposed model of CED-4 non-apoptotic function in neuronal repair shows (i) the two essential amino acids (S81 and E88 shown as yellow dots) in CED-4 helix 6 binding to a hypothetical inhibitor (red, indicated as a protein but could be an ion-binding site associated with conformational change); two green shades of CED-4 represent S and L isoforms. Upon injury (ii), the inhibition might be transiently relieved, enabling CED-4 to locally activate CED-3 caspase. Note that known apoptosis inhibitor CED-9 binds to the opposite side of the CARD helix 6 domain and is unlikely to be directly involved at the helix 6 sites. (iii) The substitution of alanine for S81 and E88 may reduce the affinity of the inhibitor to bind to the CED-4 dimer and subsequently increase the capacity to activate CED-3 in response to injury stress. When S81 and E88 in the CARD domain are absent, the interaction with the inhibitor is weakened and regeneration outgrowth is increased in a CED-3 caspase-dependent mechanism under injury conditions.