人類二氫嘧啶水解酵素在人胚腎細胞293FT之功能性表達

二氫嘧啶水解酵素 (dihydropyrimidinase; DHPase)，在嘧啶代謝途徑中可將二氫尿嘧啶 (dihydrouracil; DHU) 與二氫胸腺嘧啶 (dihydrothymine; DHT) 進行催化開環反應。儘管人類的DHPase (human DHPase; huDHPase) 活性已被報導，但其功能表達的重組蛋白質尚未全面的分析，包括受質特異性、金屬需求以及結構分析。首先，我們將huDHPase轉殖並成功於大腸桿菌表現，並同時獲得蛋白質晶體以嘗試得到其結構；然而，此蛋白質幾乎無法測得其活性，包括DHU、DHT、phthalimide等可能的受質。因此我們改使用人胚腎細胞293FT (HEK293FT) 表現系統表現huDHPase，並成功純化獲得具催化活性的酵素。SDS-PAGE結果則顯示其分子量有差異，表明具較高分子量的HEK293FT細胞表達的huDHPase可能有未明的後修飾。光譜法與呈色法分析huDHPase對不同受質之催化效率與選擇性的結果顯示，huDHPase對DHU與DHT具有明顯催化能力: 濃度2 mM DHU作為受質時，無外加金屬的比活性為2.98 µmol/mg/min，而外加鈷離子或鋅離子後，其比活性分別提升至5.09與6.71 µmol/mg/min，而外加鋅離子的huDHPase其動力學參數為: Km = 0.40 ± 0.03 mM與Vmax = 8.29 ± 0.25 µmol/mg/min。在濃度2 mM DHT作為受質時，無外加金屬的比活性為1.83 µmol/mg/min，外加鈷離子與鋅離子則分別提升其比活性至3.31與4.04 µmol/mg/min，而外加鋅離子的huDHPase其動力學參數為: Km = 0.62 ± 0.04 mM, Vmax = 5.38 ± 0.16 µmol/mg/min。廣泛可被細菌DHPase催化的受質5端取代的乙內醯脲衍生物 (5-R-hydantoin) 在huDHPase中呈現弱催化反應，表示其反應機制可能具有物種差異性。由於huDHPase在C端多出一段較細菌DHPase長的29個胺基酸，因此我們亦以大腸桿菌表達了C端切除的1-490 huDHPase片段 (huDHPase1-490), 其分析結果顯示相較於HEK293FT細胞表現並外加鋅離子之huDHPase對DHU的比活性為5.97 µmol/mg/min，大腸桿菌表現之全長huDHPase與huDHPase1-490 的比活性為0.03與0.09 µmol/mg/min。在DHT反應中也呈現相似趨勢，即切除C端反而較有活性，約三倍左右。AlphaFold 3.0的結構預測顯示，比較huDHPase與綠膿桿菌DHPase的整體結構特徵，發現兩者在動態迴圈區域上都有受質結合位中的tyrosine殘基，然而其整體摺疊結構不完全相同。抑制劑分析與分子對接的結果顯示，傳統細菌DHPase類黃酮抑制劑，如quercetin、myricetin以及kaempferol，抑制僅7.8%、6.7% 及2.5%，對huDHPase均無法有效抑制。總的來說，此研究直接的比較了表達系統的不同，將明顯的造成huDHPase的活性迥異，後續研究可進一步直接探討，哪些差異因素影響huDHPase的催化反應性。

Dihydropyrimidinase (DHPase) catalyzes the ring-opening reactions of dihydrouracil (DHU) and dihydrothymine (DHT) in the pyrimidine metabolic pathway. Although the enzymatic activity of human DHPase (huDHPase) has been reported, the biochemical properties of functionally expressed recombinant huDHPase, including acceptor specificity, metal ion dependence, and structural features, have not been comprehensively characterized. In this study, huDHPase was first expressed in Escherichia coli (E. coli), and protein crystals were obtained for structural analysis; however, the recombinant protein exhibited almost no detectable activity toward DHU, DHT, or phthalimide. Consequently, huDHPase was expressed in a human embryonic kidney 293FT (HEK293FT) cell system, from which a catalytically active enzyme was successfully purified. SDS-PAGE analysis revealed an increased apparent molecular weight of huDHPase expressed in HEK293FT cells, suggesting the presence of post-translational modification(s). Spectroscopic and colorimetric assays demonstrated that huDHPase displayed substantial catalytic activity toward both DHU and DHT. Using 2 mM DHU as the acceptor, the specific activity without added metal ions was 2.98 µmol/mg/min, which increased to 5.09 and 6.71 µmol/mg/min upon the addition of cobalt and zinc ions, respectively. Under zinc-supplemented conditions, the kinetic parameters were Km = 0.40 ± 0.03 mM and Vmax = 8.29 ± 0.25 µmol/mg/min. When 2 mM DHT was used as the acceptor, the specific activity increased from 1.83 µmol/mg/min (metal-free) to 3.31 and 4.04 µmol/mg/min in the presence of cobalt and zinc ions, respectively, with corresponding kinetic parameters of Km = 0.62 ± 0.04 mM and Vmax = 5.38 ± 0.16 µmol/mg/min under zinc-supplemented conditions. In contrast, 5-R-hydantoin, a substrate efficiently catalyzed by bacterial DHPases, showed weak detectable activity with huDHPase, indicating species-specific differences in substrate recognition. Because huDHPase contains an additional 29 amino acids at the C-terminus compared with bacterial homologs, a C-terminally truncated variant (huDHPase1–490) was expressed in E. coli. Both full-length and truncated huDHPase expressed in E. coli exhibited markedly lower activities (0.03 and 0.09 µmol/mg/min, respectively) than huDHPase expressed in HEK293FT cells under zinc-supplemented conditions (5.97 µmol/mg/min). Notably, the truncated variant displayed approximately threefold higher activity than the full-length enzyme in the bacterial expression system. AlphaFold 3.0 structural predictions revealed that although huDHPase and Pseudomonas aeruginosa DHPase share conserved tyrosine residues within the metal-binding region of the dynamic loop, their overall folding architectures differ. Furthermore, flavonoid inhibitors effective against bacterial DHPases exhibited only weak inhibition of huDHPase. In summary, this study demonstrates that the expression system profoundly influences huDHPase activity and highlights the importance of expression-dependent factors in regulating its catalytic reactivity.