The coding region for a Leuconostoc mesenteroides NRRL B-1299 dextransucrase gene (dsrA) was isolated and sequenced. Using a pair of primers designed on the basis of two highly conserved amino-acid (aa) sequences in L. mesenteroides NRRL B-512F dextransucrase and streptococcal glucosyltransferases (GTFs), a fragment of dsrA was amplified by the polymerase chain reaction (PCR). This PCR product was used as an hybridization probe to isolate a 1.8-kb fragment identified as the central region of dsrA. Cleavage by Sac I of this fragment allowed two probes to be obtained to isolate the 5' and the 3' ends of dsrA. The nucleotide sequence of the dsrA gene was determined and found to consist of an open reading frame (ORF) of 4870 base pairs (bp) coding for a 1290-aa protein with an M(r) of 145590. The aa sequence exhibited a high similarity with other GTFs. The two domains previously described in GTFs are conserved in DSRA: an N-terminal conserved domain and a C-terminal domain composed of a series of repeats. Surprisingly, the expected signal peptide was not detected. The entire gene was reconstructed and the activity of DSRA was investigated. The dextran produced appeared to be composed of 85% alpha (1-6) and 15% alpha (1-3) linkages and the oligosaccharides synthesized in the presence of maltose were mainly composed of alpha (1-6) linkages. This enzyme is a novel dextransucrase from L. mesenteroides NRRL B-1299 producing no alpha (1-2) linkages and is the first glucosyltransferase having no signal peptide described.