Arrest defective 1 (ARD1), also known as N(alpha)-acetyltransferase 10 (NAA10) was originally identified as an N-terminal acetyltransferase (NAT) that catalyzes the acetylation of N-termini of newly synthesized peptides. After that, mammalian ARD1/NAA10 expanded its’ role to lysine acetyltransferase (KAT) that post-translationally acetylates internal lysine residues of proteins. ARD1/NAA10 is the only enzyme with both NAT and KAT activities. However, recent studies on the role of human ARD1/NAA10 (hARD1/NAA10) in lysine acetylation are contradictory, as crystal structure and in vitro acetylation assay results revealed the lack of KAT activity. Thus, the role of hARD1/NAA10 in lysine acetylation is still debating. Here, we found a clue that possibly explains these complicated and controversial results on KAT activity of hARD1/NAA10. Recombinant hARD1/NAA10 exhibited KAT activity, which disappeared soon in vitro. Size-exclusion analysis revealed that most recombinant hARD1/NAA10 formed oligomers over time, resulting in the loss of KAT activity. While oligomeric recombinant hARD1/NAA10 lost its ability for lysine acetylation, its monomeric form clearly exhibited lysine acetylation activity in vitro. We also characterized the KAT activity of hARD1/NAA10 that was influenced by several experimental conditions, including concentration of reactants and reaction time. Taken together, our study proves that recombinant hARD1/NAA10 exhibits KAT activity in vitro but only under accurate conditions, including reactant concentrations and reaction duration.