The global sugar consumption in 2022 has reached 176 million tonnes, equivalent to an average intake of 0.068 kg of sugar per person per day [1]. Excessive sugar intake has been linked to many diseases, such as type 2 diabetes and obesity [2], posing serious challenges to human health and social medical resources. To satisfy the demand for sweetness without causing excessive energy intake, artificial sweeteners have been used as sucrose alternatives in the past 30 years, including aspartame, acesulfame-K, and sucralose [3]. However, recent prospective cohort studies have shown that these artificial sweeteners may increase the risk of cancer and cardiovascular diseases [4,5], prompting a reassessment of their safety. Alternatively, a new generation of safer sugar substitutes needs to be developed.

D-allulose, also known as D-ribo-2-hexylose or D-psicose (C6H12O6), is a low-calorie rare sugar, with only 0.4 kcal/g and a sweetness level of approximately 70 % relative to sucrose [2]. D-allulose has excellent physical characteristics and beneficial physiological functions including reducing fat accumulation [6], anti-hyperglycemia [7], and protection against diabetic nephropathy caused by type 2 diabetes [8]. Since 2000, D-allulose has received approval for food use as a desirable sucrose substitute in the US, Mexico, Japan, Singapore, and South Korea [2], leading to the rapid growth of its global market, which is projected to reach US $59 million by 2027 [9].

The primary method for producing D-allulose is enzymatic synthesis using two types of ketose 3-epimerase (KEase), i.e. D-allulose/D-psicose 3-epimerase (DAE/DPE, EC 5.1.3.30) and D-tagatose 3-epimerase (DTE, EC 5.1.3.31) [10], of which, DAE usually exhibits higher catalytic activity [11]. So far, however, the number of available KEase enzymes is limited, and most of them are derived from bacteria, e.g., Agrobacterium tumefaciens [12], Clostridium cellulolyticum [13], Ruminococcus sp. 5_1_39BFAA [14]. Moreover, the catalytic equilibrium conversion rate of these enzymes ranges in 23–32 %, and their thermostability is poor, depending on Co2+ or Mn2+ as co-factor (Table S1), which limits their further applications [15]. Thus, there is a critical need in exploring new DAE variants or modifying existing ones [11].

As a tetramer, the thermostability of DAE can be improved by addressing the loss of integrity of its tertiary (intra-subunit) or quaternary structure (inter-subunit) [16,17]. DAE enzyme engineering primarily focuses on the intra-subunit level, including directed evolution [18], rational design [19], semi-rational design [20], and SUMO fused expression [21]. More recently, semi-rational approaches have been increasingly used in enzyme engineering to streamline the laborious process of developing high-throughput assay methods [16]. These approaches typically involve the prediction of hotspots, which are amino acids believed to be highly flexible, through sequence alignments or by referring to crystal or simulated structures [22]. In contrast, inter-subunit engineering involves the placement of interface residues [23] or disulfide bridges [24] to prevent subunit dissociation and increase thermostability, but successful cases are rare.

In this study, we show the discovery and engineering of a novel DAE enzyme for the production of D-allulose. First, we identified a DAE candidate gene, Rum55, from a human gut bacterium Ruminococcus sp. CAG55. To enhance the thermostability for Rum55, we employed a semi-rational design approach to identify and modify its interface interaction to stabilize the tetramer structure. Afterward, we attached a self-assembling peptide (SAP) to its C-terminus to immobilize the enzyme into active aggregates for further increase stability and operation robustness for Rum55. This study provides valuable insights into the impact of interface interactions and metal ion on the stability of DAEs, along with presenting a useful strategy for engineering DAEs through the stabilization of their tetramer structures.