TBA-7 is not a microtubule-destabilizing tubulin.

To the Editor:

Eukaryotic genomes contain multiple tubulin genes that code for different tubulin isotypes. Whether the incorporation of certain tubulin isotypes into the microtubules (MTs) can lead to the destabilization of MTs has been a long-standing question. In a previous paper that analyzed the effects of tubulin mutations on neurite growth in Caenorhabditis elegans touch receptor neurons (Zheng ), we reported that the α-tubulin TBA-7 may be such a MT-destabilizing tubulin because the loss of tba-7 resulted in increased MT stability and excessive neurite growth. In light of our more recent data, however, we realized that the above conclusion is not warranted.

Previously, in a genetic screen (Zheng ), we isolated a mutant allele, u1105, that induced the growth of an ectopic neurite in the ALM neurons (we called this neurite ALM-PN for the posterior neurite of ALM). Whole genome sequencing of strain TU4696, u1105; uIs134[mec-17p::TagRFP], found a missense mutation, G92D, in the α-tubulin gene tba-7. To confirm that tba-7(G92D) was the phenotype-causing mutation, we obtained a million-mutation project strain (VC40740) containing tba-7(gk787939; Q230*), a presumptive null allele of tba-7. We crossed VC40740 with uIs115[mec-17p::TagRFP] to create the TU5198 strain and found the same extended ALM-PN phenotype. (Both uIs115 and uIs134 transgenes specifically labeled the touch receptor neurons.) We also performed a complementation test and found that TU4696 failed to complement with VC40740. Finally, we overexpressed tba-7(+) in the touch receptor neurons using the cell-specific mec-17 promoter and observed the rescue of the ALM-PN phenotype. These results led us to conclude that u1105 was a loss-of-function mutation of tba-7 and that the inactivation of tba-7 resulted in the growth of ectopic ALM-PN.

In a continuation of this research on tba-7, we created knockout alleles of tba-7 through CRISPR/Cas9-mediated gene editing (see Materials and Methods) and found that the deletion of tba-7 did not create the ALM-PN phenotype observed in TU4696 and TU5198. Neither of the two independently generated tba-7 deletion alleles (unk70 and unk74) resulted in a prominent ALM-PN (Figure 1). To test whether u1105 and gk787939 were gain-of-function alleles, we deleted the tba-7 gene using CRISPR/Cas9 gene editing in the TU5198 strain. The resulted strain (CGZ549) still showed a long ALM-PN, suggesting that the ectopic neurite phenotype was not induced by the tba-7(gk787939) mutation. Finally, we used gene editing to introduce the G92D and Q230* mutations into the tba-7 locus in the wild-type background but failed to produce the ALM-PN phenotype (unk77 and unk78 in Figure 1).

FIGURE 1:

Deletion of tba-7 did not induce ALM-PN growth. (A) Gene structure of tba-7 and the position of various alleles. u1015 and gk7879393 were previously analyzed alleles. unk70 and unk74 are deletions alleles created in the wild-type background through CRISPR/Cas9 gene editing. unk77 and unk78 were missense mutations generated by CRISPR/Cas9 gene editing with repair donor ssDNA. Single asterisks indicate the position of the missense mutation in the gene; the lines indicate the deleted regions in the knockout aleles. (B) ALM-PN length in various strains. TU4686 and TU5198 carried closely linked tba-7 and klp-7 alleles. CGZ549 was generated by creating the tba-7(unk76) deletion allele in the TU5198 strain. Thus, the tba-7(gk787939) allele but not klp-7(gk787931) was removed in the resulting CGZ549 strain.

Being puzzled by the results, we decided to reanalyze the whole genome sequencing data of the TU4696 strain and found a klp-7(E95K) mutation closely linked to the tba-7(G92D) mutation. The genetic distance between klp-7 (III:5.41) and tba-7 (III:8.79) is 3.38 cM, making them difficult to separate during crossing. We also analyzed the sequence of VC40740 and found that it also contained a missense mutation in klp-7, named gk787931(E334K). We then found that both u1105 and the strain that contained gk787939 failed to complement with the klp-7(tm2143) deletion allele in inducing the growth of ALM-PN. These results indicate that u1105 should be mapped to klp-7 and not tba-7. Thus, we have no evidence that the loss of tba-7 causes any morphological defects in the touch receptor neurons; our interpretation of the previous tba-7 results published in Zheng is incorrect. All the phenotypes observed for tba-7 are likely caused by the loss-of-function allele klp-7(E95K).

Coincidently, this effect of klp-7 was also reported in Zheng , which examined the klp-7(tm2143) allele. klp-7 codes for a kinesin-13 protein that promotes the destabilization of MTs. So, the loss of klp-7 is expected to increase MT stability and cause excessive neurite growth in the ALM neurons. It is worth noting that Kim reported enhanced axonal regeneration in tba-7 mutants using the u1105 and gk787939 alleles, but we believe that this enhanced regeneration phenotype is likely caused by mutations in klp-7, because the loss of klp-7 is known to enhance axonal regeneration (Ghosh-Roy ).

We do not know why overexpression of tba-7(+) in klp-7(E95K) animals suppressed the ALM-PN phenotype but suspect that it may have triggered the down-regulation of other tubulins, such as mec-7 and mec-12, by tubulin autoregulation (Lin ). Such down-regulation could suppress the ALM-PN growth, because we had already shown that the loss-of-function mutations in mec-7 and mec-12 suppressed the ALM-PN growth in klp-7(tm2143) mutants (Zheng ).

In summary, a rare incidence of secondary mutations in the same gene in two independently isolated strains misled us in previous studies. Recent data with the tba-7 knockout alleles clearly indicate that the loss of tba-7 does not induce increased MT stability and excessive neurite growth. Whether the C. elegans genome contains MT-destabilizing tubulins still remains unclear.

MATERIALS AND METHODS

CRISPR/Cas9 gene editing was performed according to a published protocol (Dokshin ). Recombinant spCas9 (catalogue #M0646T) was purchased from New England Biolabs (NEB), and single guide RNA (sgRNA) was synthesized using the sgRNA synthesis kit (catalogue #E3322S) from NEB. To generate unk70, which deleted exon 1-3 of tba-7, 5′-aagtaatctcaatccacgtg-3′ in exon 1 and 5′-cgatttatgatatctgccgg-3′ in exon 3 were chosen as the targets. To generate unk74 and unk76, which deleted the whole tba-7 gene, 5′-aagtaatctcaatccacgtg-3′ in exon 1 and 5′-ttccgatgctaatgataatg-3′ in exon 5 were chosen as the targets. Cas9 assembled with two different sgRNAs were injected together, and progeny with large deletions were screened for by PCR. For precise editing to generate unk77(G92D) and unk78(Q230*) in the wild-type background, 5′-gcaaatgctcactggcaaag-3′ and 5′-attatttctcaggtttggag-3′ were used as targets, respectively. Synthetic single-stranded oligodeoxynucleotide (ssODN) donors containing the desired mutations and 35-base-pair flanking sequences were used as the repair templates and injected together with the Cas9/sgRNA complex. We screened for successfully edited progeny by worm PCR using primers that bound specifically to the edited site. For each edited line, at least 25 adults were imaged for their ALM-PN length, and double asterisks (Figure 1B) indicate a significant difference in a two-way analysis of variance with Tukey’s post hoc analysis.