Genetic deficiency of the purine salvage enzyme adenosine deaminase (ADA) results in varying degrees of immunodeficiency, ranging from neonatal onset Severe Combined Immunodeficiency (SCID) to an adult onset immunodeficiency disorder. Multiple different mutations have now been identified in these immunodeficient patients. Additional mutations, initially identified in healthy individuals, abolish ADA in erythrocytes but retain 10-80% of activity in non-erythroid cells ('partial deficiency mutations'). In general, severity of disease correlates inversely with the amount of residual ADA expressed by the mutant enzymes and directly with the accumulation of the toxic metabolites deoxyATP and deoxyadenosine. We report two newly identified mutations (Y97C and L106V), both carried on the same allele of an immunodeficient patient who was diagnosed prenatally and successfully transplanted with haploidentical bone marrow. Based on the ability of mutant cDNAs to express ADA in vitro , the L106V mutation resulted in activity similar to 'partial' mutations (30% of normal) while the Y97C mutation resulted in detectable but markedly reduced activity (1.5% of normal). However, the presence of both mutations on the same allele virtually abolished detectable enzyme activity. Analysis of the crystallographic structure of ADA to understand the marked deleterious effect of the Y97C mutation suggested a previously unappreciated role of salt bridges in the catalytic mechanism of ADA. The patient was also heteroallelic for a previously described deletion of the promoter and exon 1. Testing of additional patients in whom we had not identified a mutation on the second allele revealed presence of this deletion in three of four patients tested. This deletion is therefore relatively common, accounting for 10% of almost 100 chromosomes studied by this and other laboratories, but is easily missed by currently used methods of mutation detection. Lastly, the finding of two mutations on the same allele that interact to reduce residual enzyme function emphasizes hazards in evaluating potential genotype-phenotype correlations in individuals analyzed only for the presence of single specific mutations.