To experimentally determine the functional relevance of the Sec4p phosphorylation sites, we replaced these serine residues with the phosphomimetic residues glutamic acid or aspartic acid, or with nonphosphorylatable alanines. To assay the functionality of these constructs, we asked if they could replace the endogenous copy of SEC4. As SEC4 is an essential gene we made use of a plasmid shuffle system with the counter-selectable URA3 marker, to generate cells whose only source of the essential gene SEC4 was the novel construct provided. Briefly, the genomic copy of SEC4 is deleted and the cell viability maintained with an episomal copy of wild type SEC4 in a plasmid containing URA3. After transformation with an alternative plasmid containing the novel sec4 allele, the cells are plated on 5FOA to select against the URA3-containing plasmid, leaving cells with the novel sec4 allele as the only cellular source of Sec4p function. In our analysis, we also included the serine at position 10 reasoning that its proximity to the phosphorylated residues of S8 and S11 might allow it to be utilized as a bypass phosphorylation site in the case of relaxed positional preference. In addition, the inclusion of S10 may compensate for amino acid substitutions that are poorly phosphomimetic. Combined mutagenesis of the phosphorylated serine residues at all five positions revealed that an alanine replacement retained in vivo functionality. However, replacement of the serine residues with either of the phosphomimetic amino acids aspartic acid or glutamic acid resulted in an allele that was unable to functionally complement the wild type gene. Substitution with glutamine, a neutral polar amino acid of the same size as the phosphomimetic amino acids, had no effect on functionality indicative of the suggestion that aspartic and glutamic acid residues are serving as phosphomimetics. These data indicate that the phosphorylation of serines at position S8, S11, S201, S204 is not required for Sec4p functionality. Rather, they suggest that Sec4p is phosphorylated in vivo as part of a cycle where the phosphorylated state is inhibitory and the impact of phosphomimetic substitutions are that they lock Sec4p in an inhibitory state. These experiments used a GFP-tagged version of Sec4p, however the effect of the different mutants was unaffected by the nature of the GFP moiety as an alternative tag, maltose binding protein, had the same effect. We also used an alternative method to check the functionality of untagged SEC4 constructs in vivo. This method exploits the fact that a duplication of the SEC4 gene can suppress the exocytic mutant sec15-1, a component of the exocyst that is an effector for the Sec4p GTPase. In this experiment, shown in Fig. 1C, we used untagged versions of the wild type and Sec4p mutant constructs with a similar result, that alanine substitutions gave robust function and the aspartic acid substitutions abrogated the function of the protein. The alanine mutant showed a slight advantage over wild type in that it was able to weakly suppress at 40uC. We also examined untagged versions of these sec4 alleles incorporated into the GTP-hydrolysis deficient Q79L mutant of Sec4p, which yielded the same result, namely that a construct where the phosphorylated serine residues were substituted with phosphomimetics was non-functional, relative to the substitutions with alanine. Thus a situation that compromises Sec4p function, whether artificially with NH2-terminal tagging, or in vivo using a sec15 mutant or a GTP-hydrolysis deficient form of Sec4p, allows us to discriminate between levels of Sec4p action.