Antisense morpholino oligonucleotides (MO) have been used successfully in zebrafish and Xenopus to knock down gene function by gene-specific inhibition of mRNA translation (Ekker, 2000). In addition to their ability to block cytosolic processes, MO can enter the nucleus (Partridge et al., 1996) and have been shown to be effective inhibitors of pre-mRNA splicing in mammalian tissue-culture cell lines (Schmajuk et al., 1999). We show here that MO efficiently block pre-mRNA splicing in zebrafish embryos. Splice-blocking MO have the advantages that the efficacy of gene knockdown can be quantified without the use of antibodies, and that they specifically target zygotic, and not maternal, transcripts.We targeted the fgf8 gene (Furthauer et al., 1997;Reifers et al., 1998) with splice-blocking MO. An ENU induced mutation in fgf8, acerebellar (ace; referred to here as fgf8 it282 ), has previously been described (Reifers et al., 1998). fgf8 it282 is a splice donor mutation that results in the production of an aberrantly spliced mRNA (Reifers et al., 1998). Hence, blocking the same splicing event using MO should result in a phenotype similar to fgf8 it282 .We determined the fgf8 exon/intron structure and designed two 25-mer MO complementary to the exon 2 and exon 3 splice donor sites (designated E2I2 and E3I3, respectively; Fig. 1a). Both MO span the exon/intron junction, including the most conserved residues of the splice donor consensus sequence (Fig. 1 legend).We first asked whether splice site-targeted MO can alter splicing of fgf8 mRNA. Using reverse transcriptase polymerase chain reaction, we found that injection of the E3I3 MO into early zebrafish embryos results in the production of two aberrantly spliced messages we term variant 1 and 2 (Fig. 1a, b). Sequence analysis of the corresponding cDNAs revealed that variant 1 results from excluding exon 3, whereas variant 2 results from the use of a cryptic splice donor located 32 bases 5Ј of the normal exon 3 splice donor (Fig. 1c). Use of the cryptic splice donor creates a premature termination codon at the novel splice junction (Fig. 1c). Additionally, we found that injected splice site-targeted MO only alter the structure of zygotically expressed, but not maternally supplied, mRNA, as expected (Fig. 2d). Thus, E3I3 can