Science：细菌-噬菌体快速协同进化在短短三周内推动了复杂生态网络的形成

文献引用信息：Borin, J.M., Lee, J.J., Lucia-Sanz, A., Gerbino, K.R., Weitz, J.S., Meyer, J.R., 2023. Rapid bacteria-phage coevolution drives the emergence of multiscale networks. Science. 382(6671), 674-678.

编辑评语：

真实群落的特点是物种之间存在复杂的相互作用网络。如果有专门物种与子集物种相互作用，这种相互作用就被称为 "嵌套式（nested）"；如果物种在群体内相互作用，但不同群体之间没有相互作用，这种相互作用就被称为 "模块式（modular）"。混合模式被称为 "多尺度 multiscale"网络。Borin 等人研究了在一个由一种细菌和一种噬菌体组成的简单群落中，是否以及如何通过快速进化形成多尺度相互作用网络。作者利用受体基因敲除实验重现了宿主相互作用的范围。这意味着实验室环境中的噬菌体-宿主相互作用足以形成复杂的生态模式，可以成为为噬菌体疗法提供信息的宝贵模型系统。

摘要：

物种之间的相互作用催化了多尺度生态网络的演化，包括调节不同群落功能的嵌套和模块化元素。一种常见的假设是，这种复杂模式的形成需要空间隔离或较长演化时间尺度。我们表明，多尺度网络结构可以在没有空间结构的简单生态条件下快速演化。在短短 21 天的实验室协同进化中，大肠杆菌及其噬菌体 Φ21 共同进化并多样化，形成复杂的交叉感染网络。通过测量约 10,000 种噬菌体-细菌感染并测试相互作用的遗传基础，我们确定了创建多尺度模式的每个组成部分的机制。我们的结果展示了多尺度网络如何在寄生-宿主系统中进化，阐明了达尔文的观点，即简单的自适应过程可以产生纠缠不清的生态互动库。

研究内容：

在这项研究中，作者证明了协同进化过程足以在没有外部影响或空间结构的情况下快速生成复杂的多尺度生态网络。通过确定关键相互作用的机理基础，揭示了点状和渐进进化过程是如何在网络中形成三种主要模式。最初，当细菌和噬菌体通过多次循环的arms race积累了越来越强的抵抗力和感染力时，就会出现嵌套现象，细菌通过进化突变在功能上消除受体，而噬菌体则通过创新使用新的受体。当噬菌体与变异的受体变体发生专门的相互作用时，就会形成模块，使它们能够获得对某些宿主的感染力，同时失去对其他宿主的感染力。最后，当噬菌体和细菌通过lock-and-key arms动态积累突变来提高感染力和抵抗力时，模块内部就会出现嵌套。

图表：

图 1. 噬菌体-细菌共同进化在21天内生成多尺度网络模式。(A) Phage–bacteria interaction network (PBIN) comprised of 9472 pairwise infections between 128 E. coli and 74 Φ21 strains isolated from various days of coevolution. (B) Community detection of the PBIN using the LP-BRIM algorithm reveals three modules (1, 2, and 3 indicated in black, pink, and green, respectively: Qb = 0.2100***, n = 3). (C) Nestedness (NODF) and (D) Modularity (Qb) of interactions between isolates from early (module 1, days 3 to 9: NODF = 0.7655, Qb = 0.1935, n = 2) and late (modules 2 and 3, days 12 to 21: NODF = 0.6019, Qb = 0.3151, n = 2) in the coevolution (P < 0.0001***). N refers to number of modules. P-values were obtained by comparing metrics (NODF or Qb) against a null model where overall and marginal connectances were held fixed on average. (E) and (F) Infections between representative isolates from modules 2 and 3 were remeasured and recapitulated within-module nestedness. (E) and (F) serve as roadmaps for subsequent figures. Label aesthetics are applied consistently for continuity. PBINs that include isolate labels are available in the supplement (fig. S2, A and B).

图 2. 初始的嵌套性是由宿主受体丢失和噬菌体受体创新的多个循环驱动。(A) Frequency of phage receptor use (solid) and bacterial receptor mutations (dotted) in populations across coevolution days 0 to 12. Frequency of receptor use was calculated as the titer of phages on ΔOmpC ΔOmpF, ΔLamB ΔOmpF, or ΔLamB ΔOmpC K-12 hosts relative to K-12 WT. Frequency of receptor mutations was calculated as the frequency of whole-population sequencing reads with mutations affecting each receptor. (B) Ability of 41 phage isolates from coevolution days 0 to 12 to use host receptors LamB, OmpC, or OmpF, determined by spotting phages on agar infused with different dual-receptor knockout hosts. (C) Ability of phage isolates with expanding receptor use to infect naturally coevolved and genetically engineered bacteria with mutations successively eliminating receptors.

图 3. 模块形成是由不同噬菌体和OmpF变体之间的专门相互作用导致的。(A) Infection assay of Φ21 and module entry phages spotted onto lawns of K-12 WT, module entry bacteria, and ΔLamB ΔOmpC hosts engineered with ompF mutations from corresponding entry bacteria. (B) AlphaFold predicted structures of the phage host-recognition central tail fiber protein J (top) and solved structure of host OmpF receptor (bottom). Mutations in module 1 and 2 isolates are annotated in pink and green, respectively. Phage T21_L-F-_1 has 7 mutations (green) in addition to the 5 mutations present in phage T12_WT_1 (pink). OmpF mutants do not share any mutations.

图 4. 锁与钥匙的竞争动态在模块内部产生嵌套。(A) Phylogenies of J and ompF mapped onto a PBIN of representative phage and bacterial isolates from module 2 (see Fig. 1E). Mutations are indicated on the phylogeny in red. OmpF mutations are labeled in pink, blue, and orange. Synonymous mutations and mutations in other genes are shown in figs. S5 and S6. (B) Infection assay of phages spotted onto lawns of coevolved bacteria or ΔLamB ΔOmpC hosts engineered with ompF mutations from corresponding bacteria.