Worm Breeder's Gazette 12(1): 35 (September 1, 1991)

These abstracts should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with the consent of the author.

tra-3 encodes a nematode (cal)pain

Tom Barnes

MRC LMB, Hills Road, Cambridge, CB2 2QH, ENGLAND but now at (Dept. Mol. Biol., MGH + Dept. Genetics, HMS) Boston 02114

The sex determination pathway in the nematode Caenorhabditis elegans is a scheme summarizing genetic observations and predicting how various genes and gene products interact in order to implement the development of the correct sexual phenotype as dictated by the chromosomal sex. Analysis of tra-3 mutants singly and in combination with other sex determining genes suggests the role of the tra-3 gene product is to promote or prolong the action of tra-2 (Hodgkin, Genetics 1980; Doniach, Genetics 1986; Schedl and Kimble, Genetics 1988).

The tra-3 gene was delimited to a 13.5 kb region by RFLP analysis and transformation rescue (Barnes, 1991 Meeting), and 11.4 kb of this has been sequenced. Analysis of this sequence reveals a single likely gene, whose proposed product has extended similarity to calpain, a vertebrate calcium-regulated protease. In order to confirm this assignment, amber mutations were sought amongst the three amber-suppressible tra-3 mutants, e1107 , e1525 and e1903 .TGG(Trp) to TAG(amber) mutations were found in e1107 and e1903 at different sites, neither of which was altered in e2525 (which is therefore at a third site). Tra-3 protein ( Tra-3 p)has many Trps (12/510 so far), which suggests why ambers are so prevalent amongst tra-3 (0)alleles (3/4). Their clustering into two main regions made searching for the lesions convenient. e1903 occurs at the Trp immediately C-terminal to the Cys active site, and this Trp is found in all thiol proteases (e.g. the cathepsins). e1107 occurs about 45 residues further C-terminal. These mutations should lead to severely truncated proteins (83 and 125 residues respectively) lacking almost the entire proteolytic domain and the second active site (His), yet both are smg-suppressible (Hodgkin et al., Genetics 1989). This suggests that a low level of endogenous readthrough is occurring, which is proportionately boosted with message levels in a smg background (see Jonathan Hodgkin's article in this issue for a more detailed treatise).

The gene does not appear to be wholly contained within the rescuing region. 3' RACE analysis suggests that an additional 700 bp of exonic material (as yet unsequenced) lies outside the rescuing region. Interestingly, this region would correspond to the presumptive C-terminal EF hand (metal binding) domain. This observation is readily reconciled by the fact that this is used as a site of negative regulation by the calpain small subunit in vertebrates, which association is disrupted by the presence of calcium.

The 5' end of the gene has a structure typical of an outron, and PCR analysis reveals that SL1 (but not SL2 )is capable of splicing to the putative splice acceptor. The 5'-UTR (excluding SL1 )would therefore be quite short at 35 bases. The predicted intron/exon structure is supported by maintenance of similarity, computer prediction and sizes of various RT-PCR products. To date, the 1550 bases of the predicted mRNA are spread over 6.3 kb.

Calpains are ubiquitous (in vertebrates) cytosolic neutral thiol proteases which exist as a complex of three, if not four proteins: the large (catalytic) subunit (= Tra-3 p),the small (regulatory) subunit, the endogenous inhibitor calpastatin, and possibly the natural activator protein. Calpain-like activity has been detected in a number of invertebrate species, including Drosophila. The principal role of calpain appears to be as a processing protease in conjunction with signal transduction events. Proteolysis can either activate or inactivate the various substrates, and sites of cleavage correspond to PEST sequences, at least in some cases. In vertebrates, there appears to be single genes for the small subunit and calpastatin, but at least three different large subunits are known, with differing tissue distribution, calcium responsiveness and substrate specificity.

Considering the genetic data and applying Occam's razor, a model in which the tra-3 product activates the tra-2 product by cleavage is attractive. Many calpain substrates are integral membrane or membrane-associated proteins, such as Tra-2 pis proposed to be (Kuwabara et al., 1991 Meeting). If Tra-2 pinteracts with the fems through the cytoplasmic domain (as per Kuwabara et al., 1991 Meeting), Tra-3 pcould potentiate Tra-2 pby liberating the cytoplasmic domain from the membrane allowing more efficient sequestration. This model would eliminate the conundrum of the germ-line-specific, cytoplasmic-domain-only tra-2 transcript, as it would imply that cytoplasmic domain-only Tra-2 pwould exist in the soma as well. In this model, the membrane acts as a reservoir of tra-2 activity.

Other models are also possible, involving fem-1 cleavage (calpain cleaves ankyrin) or tra-2 3'-UTR repeat-binding-protein (Goodwin et al., 1991 Meeting) cleavage.