LytA, the main autolysin of Streptococcus pneumoniae, was the first member of the bacterial N-acetylmuramoyl-l-alanine amidase (NAM-amidase) family of proteins to be well characterized. This autolysin degrades the peptidoglycan bonds of pneumococcal cell walls after anchoring to the choline residues of the cell wall teichoic acids via its choline-binding module (ChBM). The latter is composed of seven repeats (ChBRs) of approximately 20 amino acid residues. The translation product of the lytA gene is the low-activity E-form of LytA (a monomer), which can be "converted" (activated) in vitro by choline into the fully active C-form at low temperature. The C-form is a homodimer with a boomerang-like shape. To study the structural requirements for the monomer-to-dimer modification and to clarify whether "conversion" is synonymous with dimerization, the biochemical consequences of replacing four key amino acid residues of ChBR6 and ChBR7 (the repeats involved in dimer formation) were determined. The results obtained with a collection of 21 mutated NAM-amidases indicate that Ile-315 is a key amino acid residue in both LytA activity and folding. Amino acids with a marginal position in the solenoid structure of the ChBM were of minor influence in dimer stability; neither the size, polarity, nor aromatic nature of the replacement amino acids affected LytA activity. In contrast, truncated proteins were drastically impaired in their activity and conversion capacity. The results indicate that dimerization and conversion are different processes, but they do not answer the questions of whether conversion can only be achieved after a dimer formation step.