Phage (endo)lysins are nowadays one of the most promising way out of the current antibiotic resistance crisis. Either as sole therapeutics or as a complement to common antibiotic chemotherapy, lysins are already entering late clinical phases to get regulatory agencies authorization. Even the old paradigm of the inability of lysins to attack Gram-negative bacteria from without has already been overcome in a variety of ways: either by engineering approaches or investigating the natural mechanisms by which some wild type lysins are able to interact with the bacterial surface. Such inherent ability of some lysins has been linked to antimicrobial peptide (AMP)-like regions, which are, on their own, a significant source for novel antimicrobials. Currently, though, many of the efforts for searching novel lysin-based antimicrobial candidates rely on experimental screenings. In this work, we have bioinformatically analysed the C-terminal end of a collection of lysins from phages infecting the Gram-negative genus Pseudomonas. Through the computation of physicochemical properties, the probability of such regions to be an AMP was estimated by means of a predictive k-Nearest Neighbours (kNN) model. This way, a subset of putatively membrane-interacting lysins was obtained from the original database. Two of such candidates (named Pae87 and Ppl65) were prospectively tested in terms of muralytic, bacteriolytic and bactericidal activity. Both of them were found to possess an activity against Pseudomonas aeruginosa and other Gram-negative bacterial pathogens, implying that the prediction of AMP-like regions could be a useful approach towards the mining of phage lysins to design and develop antimicrobials or antimicrobial parts for further engineering.

中文翻译：

---

通过基于序列的理化性质计算来挖掘革兰氏阴性的表面活性酶生物候选物

噬菌体（内源）溶素是当今摆脱当前抗生素耐药性危机最有前途的方法之一。作为唯一的疗法或作为普通抗生素化学疗法的补充，溶素已经进入临床后期，以获得监管机构的批准。甚至赖氨酸不能从无赖细胞中攻击革兰氏阴性细菌的古老范例也已经通过多种方式得到了克服：通过工程方法或研究某些野生型赖氨酸能够与细菌表面相互作用的天然机制。 。一些溶素的这种固有能力已与抗微生物肽（AMP）样区域相关联，而AMP样区域本身就是新型抗微生物剂的重要来源。不过，目前，寻找新型基于溶素的抗微生物候选药物的许多努力都依赖于实验筛选。在这项工作中，我们已经从生物学上分析了感染革兰氏阴性假单胞菌属的噬菌体的溶菌素集合的C末端。通过理化性质的计算，借助于预测性k-最近邻居（kNN）模型估计了此类区域成为AMP的可能性。这样，从原始数据库中获得了假定的膜相互作用溶素的子集。前瞻性地测试了其中两个这样的候选物（分别命名为Pae87和Ppl65）的杀真菌活性，杀菌活性和杀菌活性。发现它们都具有抗铜绿假单胞菌和其他革兰氏阴性细菌病原体的活性，