Eleven Lactobacillus strains with strong aggregation abilities were selected from a laboratory collection. In two of the strains, genes associated with aggregation capability were plasmid located and found to strongly correlate with collagen binding. The gene encoding the auto-aggregation-promoting protein (AggLb) of Lactobacillus paracasei subsp. paracasei BGNJ1-64 was cloned using a novel, wide-range-host shuttle cloning vector, pAZILSJ. The clone pALb35, containing a 11377-bp DNA fragment, was selected from the SacI plasmid library for its ability to provide carriers with the aggregation phenotype. The complete fragment was sequenced and four potential ORFs were detected, including the aggLb gene and three surrounding transposase genes. AggLb is the largest known cell-surface protein in lactobacilli, consisting of 2998 aa (318,611 Da). AggLb belongs to the collagen-binding superfamily and its C-terminal region contains 20 successive repeats that are identical even at the nucleotide level. Deletion of aggLb causes a loss of the capacity to form cell aggregates, whereas overexpression increases cellular aggregation, hydrophobicity and collagen-binding potential. PCR screening performed with three sets of primers based on the aggLb gene of BGNJ1-64 enabled detection of the same type of aggLb gene in five of eleven selected aggregation-positive Lactobacillus strains. Heterologous expression of aggLb confirmed the crucial role of the AggLb protein in cell aggregation and specific collagen binding, indicating that AggLb has a useful probiotic function in effective colonization of host tissue and prevention of pathogen colonization.