N-terminal acetylation is a highly abundant protein modification catalyzed by N-terminal acetyltransferases (NATs) NatA-NatG. The Saccharomyces cerevisiae protein Arl3 depends on interaction with Sys1 for its localization to the Golgi and this targeting strictly requires NatC-mediated N-terminal acetylation of Arl3. We utilized the Arl3 acetylation-dependent localization phenotype as a model system for assessing the functional conservation and in vivo redundancy of several human NATs. The catalytic subunit of human NatC, hNaa30 (Mak3), restored Arl3 localization in the absence of yNaa30, but only in the presence of either yeast or human Naa35 subunit (Mak10). In contrast, hNaa35 was not able to replace its yeast orthologue without the co-expression of hNaa30, suggesting co-evolution of the two NatC subunits. The most recently discovered and organellar human NAT, NatF/Naa60, restored the Golgi localization of Arl3 in the absence of yNaa30. Interestingly, this was also true for hNaa60 lacking its membrane-binding domain whereas hNaa50 did not complement NatC function. This in vivo redundancy reflects NatC and NatF´s overlapping in vitro substrate specificities. The yeast model presented here provides a robust and rapid readout of NatC and NatF activity in vivo, and revealed evolutionary conservation of the NatC complex and redundancy between NatC and NatF.