The design-build-test-learn (DBTL) cycle has been implemented in metabolic engineering processes for optimizing the production of valuable compounds, including food ingredients. However, the use of recombinant microorganisms for producing food ingredients is associated with different challenges, e.g., in the EU, a content of more than 0.9% of such ingredients requires to be labeled. Therefore, we propose to expand the DBTL cycle and use the “learn” module to guide the development of non-engineered strains for clean label production. Here, we demonstrate how this approach can be used to generate engineered and natural cell factories able to produce the valuable food flavor compound - butanedione (diacetyl). Through comprehensive rerouting of the metabolism of Lactococcus lactis MG1363 and re-installment of the capacity to metabolize lactose and dairy protein, we managed to achieve a high titer of diacetyl (6.7 g/L) in pure dairy waste. Based on learnings from the engineering efforts, we successfully achieved the production of diacetyl without using recombinant DNA technology. We accomplish the latter by process optimization and by relying on high-throughput screening using a microfluidic system. Our results demonstrate the great potential that lies in combining metabolic engineering and natural approaches for achieving efficient production of food ingredients.

设计-构建-测试-学习 (DBTL) 循环已在代谢工程过程中实施，以优化包括食品成分在内的有价值化合物的生产。然而，使用重组微生物生产食品成分面临着不同的挑战，例如，在欧盟，超过 0.9% 的此类成分的含量需要进行标记。因此，我们建议扩展 DBTL 周期并使用“学习”模块来指导用于清洁标签生产的非工程菌株的开发。在这里，我们展示了如何使用这种方法来生成能够生产有价值的食用风味化合物 - 丁二酮（双乙酰）的工程和天然细胞工厂。通过乳酸乳球菌代谢的全面改道MG1363 和重新安装代谢乳糖和乳蛋白的能力，我们设法在纯乳制品废物中实现了高滴度的双乙酰 (6.7 g/L)。基于从工程工作中吸取的经验，我们在不使用重组 DNA 技术的情况下成功实现了双乙酰的生产。我们通过工艺优化和依赖使用微流体系统的高通量筛选来完成后者。我们的研究结果表明，将代谢工程与自然方法相结合以实现食品成分的高效生产具有巨大的潜力。