Terrell Evan, Bruni Gillian O, Qi Yunci
US Department of Agriculture, Agricultural Research Service, New Orleans, LA, USA.
US Department of Agriculture, Agricultural Research Service, Peoria, IL, USA.
J Sci Food Agric. 2025 Aug 15. doi: 10.1002/jsfa.70125.
The presence of sucrose-consuming microorganisms and their associated exopolysaccharides (EPS) causes operational challenges for sugar crop (e.g. sugarcane, sugar beet) processing facilities during raw sugar manufacturing. To better prevent factory losses from issues related to microbially-derived EPS, it is necessary to develop easy-to-implement characterization methods for EPS. The two primary types of EPS associated with sugar crop processing are dextrans and levan fructans, which occur to varying degrees during raw sugar production. Chemometric techniques applied to characterization of these EPS can efficiently distinguish between dextran and levan based on their structural differences.
In this study, EPS was produced from several bacterial strains isolated from sugarcane and sugar beet factories. These EPS, along with dextran and levan standards, were analyzed with Fourier transform infrared (FTIR) spectroscopy and subjected to principal component analysis (PCA). Results from PCA applied to FTIR allow for the ability to distinguish dextran from levan EPS using machine learning techniques (linear discriminant analysis, support vector machine and k-nearest neighbors). These results are also consistent with dextran- and levan-producing genes identified through whole genome sequencing for the bacterial isolates.
The analyses presented in this study demonstrate a low cost, relatively non-labor-intensive analytical classification method for possible identification/characterization of EPS in raw sugar production. Implementation of this type of analytical approach at sugar crop processing facilities and/or laboratories has potential to improve engineering practices for EPS management. Future work applied to mixed samples, polycultures, or sugar crop juices could be explored to further develop chemometric applications. Published 2025. This article is a U.S. Government work and is in the public domain in the USA.
消耗蔗糖的微生物及其相关胞外多糖（EPS）的存在，给原糖生产过程中的糖料作物（如甘蔗、甜菜）加工设施带来了操作挑战。为了更好地防止工厂因微生物来源的EPS相关问题而造成损失，有必要开发易于实施的EPS表征方法。与糖料作物加工相关的EPS主要有两种类型，即葡聚糖和果聚糖，它们在原糖生产过程中出现的程度各不相同。应用化学计量技术对这些EPS进行表征，可以根据其结构差异有效地区分葡聚糖和果聚糖。
在本研究中，从甘蔗和甜菜工厂分离出的几种细菌菌株产生了EPS。这些EPS与葡聚糖和果聚糖标准品一起，用傅里叶变换红外（FTIR）光谱进行了分析，并进行了主成分分析（PCA）。应用于FTIR的PCA结果表明，使用机器学习技术（线性判别分析、支持向量机和k近邻）能够区分葡聚糖和果聚糖EPS。这些结果也与通过对细菌分离株进行全基因组测序鉴定出的葡聚糖和果聚糖产生基因一致。
本研究中的分析表明，一种低成本、相对不耗费人力的分析分类方法可用于原糖生产中EPS的可能鉴定/表征。在糖料作物加工设施和/或实验室实施这种分析方法，有可能改善EPS管理的工程实践。未来可以探索将其应用于混合样品、混合培养物或糖料作物汁液，以进一步开发化学计量学应用。2025年发表。本文是美国政府工作成果，在美国属于公共领域。
