Functional and bioinformatical studies of the TGFb superfamily : Identification and characterization of ALK7

The transforming growth factor beta (TGFbeta) superfamily of polypeptide growth factors encompasses more than 30 diverse mammalian members. They are involved in most cellular processes, ranging from cell proliferation and differentiation, to survival and death. Loss of TGFbeta activity is coupled to tumor development and many other disorders. Most TGFbeta superfamily ligands exert their actions through two transmembrane receptor serine/threonine kinases (RSTKs), type I and type II. Following ligand binding, the type I and type II receptors form a heterotetrameric complex that activates intracellular second messengers of the Smad family. Upon activation, the Smads translocate to the nucleus and initiate gene transcription. This thesis focuses on activin receptor-like kinase 7 (ALK7), a type 1 RSTK. We have cloned and sequenced the ALK7 gene and studied its expression and function in vitro as well as in vivo. We find ALK7 to be highly expressed in adult cerebellum and during mouse embryogenesis. The absence of a known ligand and a partner type II receptor initially constrained the functional analysis of this receptor. We therefore constructed a cell line derived from rat pheochromocytoma PC12 cells in which a tetracycline-inducible promoter controlled the expression of a constitutively active mutant of ALK7. This cell line was then used to study the biological relevance of ALK7 activation. Briefly, ALK7 signaling resembles that of its closest relatives, TbetaRI and ActRIB. However, only ALK7 has a profound effect on cell morphology. Next, we established nodal as a candidate ligand for ALK7. Indeed, constitutively active ALK7 mimics the effects of Xnr1, the Xenopus laevis nodal homologue, while kinase dead ALK7 specifically blocks those effects. We also show that ALK7 cooperates with a type II receptor for activin (ActRIIB), and the complex is able to mediate nodal signals. In parallel with these studies, successive efforts were made to generate a knockout mouse deficient in ALK7 activity. After a total of ten electroporations and 3,456 picked clones, nine clones that correctly targeted the ALK7 gene by homologous recombination were isolated. Fifteen ALK7 +/- pups were born after blastocyst injections, and they are now bred to generate ALK7 +/- animals. At this stage, we can only speculate on the outcome, but many related gene deletion studies result in mice that succumb either during early embryonic stages because of severe defects of gastrulation, or later during development because of impaired left-right patterning. The second part of this thesis involves bioinformatics and the computational analysis of the increasing amount of sequence information available. The first release in 1982 of the EMBL Nucleotide Sequence Database contained 568 sequences with 585,433 nucleotides. Today, nineteen years later, this database has reached its 68:th release with 12,964,797 sequences and 13,727,100,206 nucleotides - the size of the database has at least tripled every three years. When searching for distant homologies, where only a non-contiguous amino acid pattern is conserved, conventional search engines are not always capable of matching relevant sequences. I therefore developed Motifer, a software tool aimed at finding very remote protein homologues. A number of features enable Motifer to find sequence homologues where FASTA, BLAST, and other search algorithms would deem them unrelated. Using Motifer to search the human genome for sequences corresponding to the conserved cysteine pattern of the TGFbeta superfamily and that of the glial cell line-derived neurotrophic factor (GDNF) family receptor a (GFRalpha), two new TGFbeta superfamily ligands and one new GFRalpha were identified. The new sequences were provisionally named MDF45 1, NMF628 and MDRalpha (Motifer derived factor/receptor; numbers originating from sequence accession numbers). BLAST and FASTA failed to find any of these sequences using known TGFbeta ligands or GFRalpha receptors as query. Interestingly, the MDFs appear to be primate specific, while MDRalpha can also be found in mouse. Expression studies in human tissues have shown the MDFs to be mainly expressed in the adult brain, adult cerebellum, and fetal brain. MDRalpha mRNA was not detected in any of the mouse tissues investigated. Present work is now aimed at detecting a functional translational product and at cloning the full- length cDNA of all three sequences identified