Using a strategy of large-scale whole mount in situ hybridization, three genes were identified from a cDNA library constructed from endoderm-like tissue induced from activin treated animal caps. One gene, XODC2, encodes a paralogue to ubiquitous ODC (Genbank accession number: AF217544); another, XCL-2, encodes a tissue-specific m-type Calpain (Genbank accession number: AF212199); while the third one, XETOR, encodes a novel member of the ETO/MTG8 oncoprotein family (Genbank accession number: AF212198). Spatial and temporal expressions of these genes were examined by ways of whole mount in situ hybridization and RT-PCR. Functional analyses were focused on XCL-2 and XETOR. Overexpression of wild-type XCL-2 suggests that this gene is involved in gastrulation movement and convergent extension during gastrulation and neurulation. Overexpression of a dominant-negative mutant caused a phenotype morphologically similar to, but histologically different from, that caused by overexpression of wild-type XCL-2. The mutant phenotype can be rescued by injection of wild-type XCL-2. These data suggest that XCL-2 plays an important role in convergent extension movements during embryogenesis in Xenopus laevis. XETOR is expressed during neurula stage in three bilaterally symmetrical stripes at each side of dorsal midline, a pattern similar to that of the genes involved in primary neurogenesis. Indeed, overexpression of XETOR or a series of truncated mutants led to the inhibition of primary neuron formation without disruption of neural plates. Such an inhibitory effect is not mediated by lateral inhibition, but an independent action. Moreover, it was shown that XETOR and lateral inhibition antagonizes each other. Further evidence showed that expression of XETOR is activated or promoted by overexpressed proneural genes such as Xngnr-1, Xash-3, Xath3, and XNeuroD, and conversely, overexpressed XETOR inhibits the expression and function of Xath3 and XNeuroD. The neuron inducing activity but not expression of Xash-3 is inhibited in response to XETOR overexpression, while neither the expression nor the function of Xngnr-1 is inhibited. Thus a negative feedback loop is established between XETOR and proneural genes. Blocking of XETOR function in vivo resulted in a neurogenic phenotype of an enlarged neurogenic domain without alteration in neuron density. Nevertheless, double depletion of XETOR and lateral inhibition led to primary neuron formation with an increased density in enlarged domains. Based on these data, it is concluded that during primary neurogenesis, lateral inhibition and XETOR comprise a dual inhibitory mechanism to refine the number and localization of primary neurons via repression of the expression and function of proneural genes.