Subsequently, the PCR fragments were purified and subcloned into the SalI and BamHI sites of the expression vector pEGFP-N1. The mammalian expression plasmids pTL-Flag-TPI-WT and pTL-Flag-TPI-Glu104Asp were generated by subcloning the corresponding DNA fragments from pACT-TPI-WT or pACTTPI-Glu104Asp into the XhoI/NotI sites of the vector pTL-FlagC, respectively. PCR was performed under conditions described earlier and all resultant products were verified by sequencing. As aforementioned, the frequency of TPI null alleles is much higher as the rare incidence of TPI deficiency, and in vitro measurements of a mutant TPI variant purified from E. coli demonstrated that the purified TPI protein carrying the Phe240Leu mutation exhibited a 6-fold higher activity than expected from the measurements performed with the patient erythrocyte extract. Moreover, theoretical calculations demonstrated that the extremely high DHAP level observed in patients’erythrocytes does not correlate with the measured residual TPI activity. Since patients’metabolites highly depend on the genetic background and environmental factors, we have generated an in vivo system Folinic acid calcium salt pentahydrate allowing analysis of the enzymatic activity of wild-type or pathogenic TPI variants inside a cell without additional side effects. At least three different isoelectric TPI variants encoded by the single TPI gene have been observed in human red blood cells or mouse-brain capillary endothelial cells. Furthermore, the Alternative Splicing Database predicts three alternative splice variants for human TPI, but none of them has been identified experimentally to date. Along these lines, protein diversity is often generated by utilizing alternative translation Chlorhexidine hydrochloride initiation sites, mainly if a second start site is located within the first 20 codons. The usage of alternative translation initiation sites have been discussed for disease-causing proteins such as parkin, the prion protein and the breast cancer antigen BRCA1. Here, we provided evidence that the second in-frame ATG codon in the TPI gene encoding MET14 can be used as an alternative translation initiation site in yeast as well as in mammalian cells. However, the resulting protein did not suppress the growth defect of Dtpi1 yeast on glucose medium indicating that this variant lacks catalytic activity. This result is not unexpected, since this TPI variant lacks the catalytic lysine and at least three residues of the intersubunit interface. Interestingly, a TPI deficiency patient which has inherited a start codon mutation in combination with the mutation at position 104, has a stronger pathology in comparison to Glu104Asp homozygotes. Structural alterations of the pathogenic TPI variants have often been speculated to contribute to TPI deficiency as a number of mutations seem to affect the dimerization interface of TPI which forms a stable dimer in most investigated organisms. To address this issue, we have analyzed the dimerization properties between wild-type and the pathogenic TPI variants by a quantitative assay that is based on the lacZ gene as reporter. This allows determination of the relative strength of protein-protein interactions in yeast, however, these values can not be directly correlated with binding constants determined by in vitro measurements. We discovered that the dimerization behavior between wild-type TPI and the two pathogenic variants Cys41Tyr and Glu104Asp is strongly altered compared to the dimerization behavior between wild-type proteins. We further observed that the glycolytically inactive TPI2ndATG variant dimerizes with wild-type TPI, although at reduced levels in comparison to the dimerization of wild-type TPI proteins.