The mutagenic effects of 2-aminopurine (2AP) have been compared in the adenine-3 (ad-3) region of two-component heterokaryons of Neurospora crassa: nucleotide excision repair-proficient ( uvs-2 +/ uvs-2 +) heterokaryon 12 (H-12) and nucleotide excision repair-deficient ( uvs-2/ uvs-2) heterokaryon 59 (H-59). This forward-mutation, morphological and biochemical, specific-locus assay system permits the recovery of ad-3 A and / or ad-3 B mutants in 3 major classes: gene/point mutations, multilocus deletion mutations, and unknowns, and 3 different subclasses of multiple-locus mutations. Previous studies (Brockman et al., Mutation Res., 218 (1989) 1–11) showed that 2AP treatment of growing cultures of H-12 and H-59 gave no difference between ad-3 forward-mutation frequencies over a wide range of 2AP concentrations in each strain. In the present experiments, genetic analyses of ad-3 mutants recovered from these experiments has demonstrated qualitative differences between the spectra of the 3 main classes of ad-3 mutations. In H-12, 84.2% (203/241) resulted from gene/point mutation, 11.6% (28/241) from multilocus deletion mutation, and 4.1% (10/241) were unknowns. In contrast, in H-59 43.0% (99/230) resulted from gene/point mutation, 55.7% (128/230) from multilocus deletion mutation, and 1.3% (3/230) were unknowns. In additions, quantitative differences were also found between the spectra of ad-3 mutations in 1 subclass of multiple-locus mutations, but not 2 additional subclasses. The first subclass consisted of 1.7% (4/241) and 9.6% (22/230) gene/point mutations with a closely linked recessive lethal mutation, in H-12 and H-59, respectively. The second two subclasses consisted of (a) 0.4% (1/241) and 0.4% (1/230) multilocus deletion mutations with a closely linked recessive lethal mutation, and (b) 13.3% (32/241) and 15.2% (35/230) gene/point mutations with a separate recessive lethal mutation elsewhere in the genome, in H-12 and H-59, respectively. Data from studies by others have shown that 2AP inhibits adenosine deaminase, resulting in nucleotide precursor pool inbalance, and that 2AP can saturate the mismatch repair system. As a consequence of either effect of 2AP, the spectrum of 2AP-induced mutation could include frameshift mutations and chromosome aberrations such as multilocus deletions in addition to base-pair substitutions. The defect in DNA repair due to the uvs-2 allele, which has been shown to be a deficiency in pyrimidine dimer excision (Worthy and Epler, 1974), most probably has some other excision-repair deficiency (Macleod and Stadler, 1986; Baker et al., 1991). Uvs-2, which is homokaryotic in H-59, results in about 5-fold higher frequencies of 2AP-induced multilocus deletion mutations and multiple-locus mutations (gene/point mutations with a closely linked recessive lethal mutation), as compared to H-12, suggesting that H-59 lacks an important system that normally repairs or prevents these types of chromosomal damage. We conclude that 2AP induces ad-3 mutants by a wide variety of mechanisms including: (1) AT to GC and GC to AT base-pair transitions, (2) other, as yet unidentified, intragenic alterations, (3) small multilocus deletion mutations, and (4) multiple-locus mutations (gene/point mutations at the d-3 A locus or at the ad-3 B locus associated with separate, but closely linked, recessive lethal mutations, or multilocus deletion mutations with separate, but closely linked, recessive lethal mutations).