Many bacterial species have biotechnological applications, for example in the production of compounds of interest by fermentation or biotransformation, the handling of waste and the decontamination of polluted sites. The genetic manipulation of such bacteria may improve their suitability for these processes, and even allow them to synthesize or degrade new compounds. Obtaining good results usually requires that engineered microorganisms dedicate a significant amount of their resources (energy, reducing power, etc.) to the desired biocatalytic activity, but this can leave their general physiology under-optimized. When resources are compromised, a number of responses aimed at restoring proper equilibrium are triggered, likely resulting in the yield of the desired catalytic activities being reduced. This is an old problem for biotechnologists that has traditionally been tackled by optimizing the growth conditions, but it might also be confronted via a deeper understanding of the regulatory mechanisms that govern metabolite fluxes and bacterial physiology. In fact, for the optimal engineering of biocatalysts via the rational design and assembly of biological modules that work in a coordinated fashion, such knowledge is essential.

Bacteria have different global regulators that help adapt their metabolism to fluctuating conditions, such as when the supply of nutrients changes or when oxygen availability becomes limiting. But even when nutrients and oxygen are plentiful, other regulatory mechanisms may compromise biotechnological processes. For example, when different types of carbon sources are present, and in varying abundance, regulatory responses may prioritize one compound over the rest, resulting in a distribution of metabolite fluxes that may—or may not—be compatible with the microorganism's biotechnological use. Such regulation, generically known as carbon catabolite repression (CCR), has been studied for decades in a few model bacteria and has been shown not only complex but to be driven by distinct molecular mechanisms in each bacterial group (Görke & Stülke, 2008; Rojo, 2010). The preferred substrates also differ. Many bacteria prioritize glucose over other compounds (e.g. Escherichia coli or Bacillus sp.), while some prefer certain organic acids or amino acids over this sugar (e.g. Pseudomonads). These two strategies have been termed classical CCR and reverse CCR respectively (Park et al., 2020), and it is proposed that their complementarity helps bacteria specialize with respect to their carbon sources, thereby reducing competition for them in their natural environments (Park et al., 2020).

Optimizing a bacterium for a particular use, or engineering it to obtain a new metabolic pathway, often requires its substantial genetic modification. Powerful techniques are now available for manipulating bacterial genomes, although not all strains are equally tractable. Genome-editing techniques even allow large DNA segments to be deleted, reducing the size of the genome (Martínez-García et al., 2014; Martínez-García et al., 2015). However, it is difficult to predict the full consequences of directed genome edition on bacterial physiology. Clearly, the rational modification of any strain for a specific purpose requires prior in-depth knowledge of its metabolism, and of the regulatory networks that coordinate and optimize the expression of its genome. However, we are far from having such knowledge at our disposal despite the impressive progress made in recent years. Indeed, with the number of known regulatory elements and interconnections increasing every year, plus the realization of the importance of regulatory small RNAs (Aoyama et al., 2022; Bobrovskyy & Vanderpool, 2013), the puzzle is becoming ever more complex.

The control of metabolite fluxes is influenced by many factors, complicating their study (Chubukov et al., 2014). Certainly, the use of [¹³C]-labelled substrates to analyse these fluxes, plus the development of strain-specific genome-scale metabolic models that help in interpreting transcriptomic and metabolomic datasets, have greatly increased our understanding of which pathways metabolites flow through under given growth conditions, and of how these fluxes are controlled (Hyduke et al., 2013; Schwechheimer et al., 2018). However, unless growth occurs in a chemostat in which steady-state conditions can be maintained, the concentrations of nutrients must fall as they are consumed while those of by-products, and waste products, must increase. This necessarily influences the metabolite fluxes at work, as recently illustrated in the model bacterium Pseudomonas putida KT2440 when batch-cultured in a complete medium. The configuration of its metabolism changed substantially over the exponential growth phase as cells sequentially exhausted the different nutrients present and started to use those less preferred [Figure 1(A,B)]. In particular, the configuration of the tricarboxylic acid cycle changed over growth, providing no energy in the early phase (P. putida can use glucose as a source of energy only, or as an energy and carbon source), then switching to a reductive mode at mid-growth, and to an oxidative configuration in late exponential phase (Molina et al., 2019b). Furthermore, the inactivation of the main CCR regulatory network led to a metabolic imbalance in which the uptake and assimilation of substrates did not match cellular needs, leading to the overflow of some pathways and the leakage of pyruvate and acetate—something not seen in the wild-type strain (Molina et al., 2019a). The CCR-deficient strain also assimilated many of the available nutrients significantly more quickly than did the wild-type strain, but rather than increasing the growth rate, it actually reduced it, revealing the importance of CCR in coordinating metabolism and optimizing growth.

An important question is whether the addition of new genes and pathways to a bacterium might alter its metabolite fluxes and overall performance, and whether or not it is desirable to keep the new genes beyond the influence of the cell's regulatory networks. In connection with this, it is interesting to know whether native genes that provide bacteria with metabolic versatility, but which are present only in certain strains of a given species (and are therefore components of the so-called ‘accessory genome’), fall under the influence of these regulatory networks. Several Pseudomonas strains with accessory genes that allow them to assimilate hydrocarbons and aromatic compounds have been studied. In most cases, these genes were found to be under the strict influence of global regulatory networks, including CCR (Hernández-Arranz et al., 2013; Madhushani et al., 2015; Moreno et al., 2010). It would therefore seem to be advantageous for cells to keep newly acquired genes under the control of their global regulatory networks. But this might not always be desirable for genes involved in biotechnological applications. In fact, there are several examples in which global regulatory networks, and CCR in particular, have hampered such applications (Vinuselvi et al., 2012). In these cases, uncoupling the expression of genes of interest from global regulatory networks would therefore seem advisable [Figure 1(C)].

The question thus arises as to what might be the best strategy for separating a metabolic pathway of interest from CCR control. One solution might be to reduce as far as possible the interaction between pathways that generate biomass (growth) and those that produce compounds of interest (Pandit et al., 2017). This might be useful in some settings, but not in all, since many pathways are strongly interconnected. An alternative is to use a carefully selected mixture of carbon sources that optimizes both growth and biocatalyst performance (Liu et al., 2020), but again, no size fits all. A third possibility is to eliminate the influence of CCR on the biosynthetic pathway of interest by deactivating the complete CCR network, for example by deleting some key genes associated with it. Some examples are known in which inactivating the genes responsible for CCR improved the simultaneous utilization of several compounds that would otherwise be used sequentially (Elmore et al., 2020; Johnson et al., 2017). However, once again, this solution was found not to be helpful in other situations (Lu et al., 2021), probably because eliminating the global regulators responsible for CCR causes imbalanced metabolism. A further possibility might be to conserve the general CCR network but detach the genes of interest from its influence. This implies engineering these genes to eliminate the elements recognized by global CCR regulators, and indeed this strategy has proved helpful in optimizing a P. putida strain engineered to generate medium-chain-length α,ω-diols (building blocks for polymer production) (Lu et al., 2021). The drawback here is that such an approach requires detailed prior knowledge of the molecular mechanisms that drive CCR in the strain of interest, including the identification of the targets recognized by the global regulators of the genes to be manipulated. In addition, the regulatory mechanisms responsible for CCR differ substantially among bacterial groups. The molecular details are relatively well known for model bacteria such as E. coli, B. subtilis or Pseudomonads (Görke & Stülke, 2008; Moreno et al., 2015; Pei et al., 2019; Rojo, 2010; Sonnleitner et al., 2018), but certainly not for other bacteria of biotechnological importance (Ruiz-Villafán et al., 2021). Moreover, even if the problems related to CCR could be solved, other regulatory elements might also interfere transcriptionally or post-transcriptionally affecting the expression of the genes of interest. Particular metabolites might even allosterically modulate the activity of key enzymes. In other words, eliminating one regulatory element might not avoid the effects of others.

A final problem is that the signals that trigger CCR are not well understood. Cells probably sense the concentrations of key metabolites that act as flux sensors, although other elements might be monitored as well. A recent report indicated that the hierarchy of utilization of carbon sources in E. coli was ordered by the total carbon-uptake flux rather than by the precise compounds being used (Okano et al., 2020; Okano et al., 2021). Below a given threshold of carbon-uptake flux, CCR faded away and the compounds present were utilized simultaneously. In addition, certain metabolites such as fructose-1,6-bisphosphate (an intermediate of the glycolytic pathway), α-ketoglutarate (a component of the tricarboxylic acids cycle) and others may also act as flux sensors regulating the activity of particular enzymes (Chubukov et al., 2014; Okano et al., 2020). The final picture is that of finely tuned, highly intertwined and regulated metabolic pathways that are complicated to study, but much deserving of attention.

In summary, learning to control bacterial metabolism, or at least making good predictions regarding its behaviour, will require considerable effort. It would be useful to investigate this from complementary perspectives, for example examining how the expression of genes making up complete metabolic pathways is regulated, how global regulators change transcription programs, how the activity of particular enzymes is regulated by key signalling metabolites, which metabolites flow through which pathways under given conditions, how these flows change when conditions change, how pathways are interconnected, and so on. Obtaining this knowledge requires bringing together traditional molecular biology techniques, analysis of the metabolite fluxes under different conditions, the use of in silico genome-scale metabolic models of increasing complexity that will allow predictions to be experimentally verified, and so on. The more complete information we obtain, the better the chance of succeeding when engineering bacterial biocatalysts.

A detailed knowledge of how cells regulate and coordinate their metabolism is important in other fields too (Figure 2). For example, the nutrients used by pathogenic bacteria can influence the expression of virulence genes (Eisenreich et al., 2010), the adaptation of the pathogen to a specific host (La Rosa et al., 2018), or its susceptibility to different antibiotics, for example those that enter the cell using cell envelope proteins involved in the transport of nutrients (Martínez & Rojo, 2011). Metabolic preferences can also have a huge impact on the assembly and behaviour of bacterial communities, the members of which have to share—or compete for—nutrients present in limited amounts (Bajic & Sánchez, 2020; Estrela et al., 2021). CCR also affects the expression of genes involved in plant–microbe interactions (Franzino et al., 2022) with its many implications. Clearly, reaching an in-depth understanding of the regulation of metabolism will require much effort over the coming years, but will bring many benefits as well.

许多细菌物种具有生物技术应用，例如通过发酵或生物转化生产感兴趣的化合物、废物处理和污染场地的净化。对这些细菌进行基因操作可以提高它们对这些过程的适应性，甚至可以让它们合成或降解新化合物。获得良好的结果通常需要工程微生物将大量资源（能量、还原力等）用于所需的生物催化活性，但这可能会使它们的一般生理机能得不到优化。当资源受到损害时，会触发一些旨在恢复适当平衡的反应，可能导致所需催化活性的产量降低。对于生物技术学家来说，这是一个传统上通过优化生长条件来解决的老问题，但也可能通过更深入地了解控制代谢物通量和细菌生理学的调节机制来解决。事实上，为了通过合理设计和组装以协调方式工作的生物模块来优化生物催化剂工程，这些知识是必不可少的。

细菌具有不同的全局调节器，可帮助它们的新陈代谢适应波动的条件，例如当营养供应发生变化或氧气供应变得有限时。但即使营养物和氧气充足，其他调节机制也可能损害生物技术过程。例如，当存在不同类型的碳源且丰度不同时，监管反应可能会优先考虑一种化合物而不是其他化合物，从而导致代谢物通量的分布可能与微生物的生物技术用途相容，也可能不相容。这种调节，通常称为碳分解代谢物抑制 (CCR)，已经在一些模型细菌中研究了数十年，并且已被证明不仅复杂，而且在每个细菌群中由不同的分子机制驱动（Görke & Stülke， 2008年；罗霍，  2010）。优选的底物也不同。许多细菌优先选择葡萄糖而不是其他化合物（例如大肠杆菌或芽孢杆菌），而有些细菌更喜欢某些有机酸或氨基酸而不是这种糖（例如假单胞菌）。这两种策略分别被称为经典CCR 和反向CCR（Park 等人，  2020 年），有人提出它们的互补性有助于细菌在碳源方面进行专业化，从而减少它们在自然环境中的竞争（Park 等人，2020 年）。等人，  2020 年）。

针对特定用途优化细菌，或对其进行改造以获得新的代谢途径，通常需要对其进行大量的基因改造。现在可以使用强大的技术来操纵细菌基因组，但并非所有菌株都同样容易处理。基因组编辑技术甚至允许删除大的 DNA 片段，从而减小基因组的大小（Martínez-García 等人，  2014 年；Martínez-García 等人，  2015 年）). 然而，很难预测定向基因组编辑对细菌生理学的全部影响。显然，为特定目的对任何菌株进行合理修饰需要事先深入了解其新陈代谢，以及协调和优化其基因组表达的调控网络。然而，尽管近年来取得了令人瞩目的进展，但我们还远未掌握这些知识。事实上，随着已知调控元件和相互关联的数量逐年增加，再加上调控小 RNA 重要性的认识（Aoyama 等人，  2022 年；Bobrovskyy 和 Vanderpool，  2013 年），这个难题变得越来越复杂。

代谢物通量的控制受到许多因素的影响，使他们的研究复杂化（Chubukov 等人，  2014 年）。当然，使用 [ ¹³ C] 标记的底物来分析这些通量，加上有助于解释转录组学和代谢组学数据集的菌株特异性基因组规模代谢模型的开发，极大地增加了我们对代谢物在哪些途径下流动的理解给定生长条件，以及如何控制这些通量（Hyduke 等人，  2013 年；Schwechheimer 等人，  2018 年). 然而，除非生长发生在可以维持稳态条件的恒化器中，否则营养物的浓度必须在消耗时下降，而副产品和废物的浓度必须增加。这必然会影响工作中的代谢物通量，正如最近在完全培养基中分批培养的模型细菌恶臭假单胞菌 KT2440 中所说明的那样。随着细胞依次耗尽存在的不同营养物质并开始使用那些不太受欢迎的物质，其新陈代谢的结构在指数生长期发生了显着变化 [图 1(A,B)]。特别是，三羧酸循环的结构随着生长而改变，在早期阶段不提供能量（恶臭假单胞菌可以仅将葡萄糖用作能源，或作为能源和碳源），然后在生长中期切换到还原模式，并在指数后期切换到氧化配置（Molina 等人，2019b  ）。此外，主要 CCR 调节网络的失活导致代谢失衡，其中底物的摄取和同化与细胞需求不匹配，导致某些途径溢出以及丙酮酸和乙酸盐的泄漏——这在野外是看不到的型菌株 (Molina et al.,  2019a). 与野生型菌株相比，CCR 缺陷型菌株吸收许多可用营养素的速度也明显更快，但它并没有提高生长速度，反而降低了生长速度，这揭示了 CCR 在协调新陈代谢和优化生长方面的重要性。

一个重要的问题是，向细菌添加新基因和途径是否会改变其代谢物通量和整体性能，以及是否需要让新基因不受细胞调节网络的影响。与此相关，了解为细菌提供代谢多样性但仅存在于给定物种的某些菌株（因此是所谓的“辅助基因组”的组成部分）的天然基因是否属于这些监管网络的影响。几种假单胞菌已经研究了具有允许它们同化碳氢化合物和芳香族化合物的辅助基因的菌株。在大多数情况下，这些基因被发现受到包括 CCR 在内的全球监管网络的严格影响（Hernández-Arranz 等人，  2013 年；Madhushani 等人，  2015 年；Moreno 等人，  2010 年）。因此，细胞将新获得的基因置于其全球监管网络的控制之下似乎是有利的。但这对于涉及生物技术应用的基因来说并不总是可取的。事实上，有几个例子表明全球监管网络，尤其是 CCR，阻碍了此类应用（Vinuselvi 等人，2012 年）). 因此，在这些情况下，从全球监管网络中解耦感兴趣基因的表达似乎是可取的 [图 1(C)]。

因此，问题是什么可能是将感兴趣的代谢途径与 CCR 控制分开的最佳策略。一种解决方案可能是尽可能减少产生生物量（生长）的途径与产生感兴趣化合物的途径之间的相互作用（Pandit 等人，2017 年 ）。这在某些情况下可能有用，但并非在所有情况下都有用，因为许多路径都是紧密相连的。另一种方法是使用精心挑选的碳源混合物，以优化生长和生物催化剂性能（Liu 等人，  2020 年）)，但同样，没有适合所有人的尺码。第三种可能性是通过停用完整的 CCR 网络来消除 CCR 对感兴趣的生物合成途径的影响，例如通过删除与其相关的一些关键基因。一些已知的例子表明，使负责 CCR 的基因失活可以提高几种化合物的同时利用，否则这些化合物将按顺序使用（Elmore 等人，  2020 年；Johnson 等人，  2017 年）。然而，再次发现该解决方案在其他情况下没有帮助（Lu et al.,  2021), 可能是因为消除负责 CCR 的全球监管机构会导致新陈代谢失衡。另一种可能性可能是保护一般的 CCR 网络，但将感兴趣的基因从其影响中分离出来。这意味着对这些基因进行工程改造以消除全球 CCR 监管机构认可的元素，事实上，这一策略已被证明有助于优化P。恶臭菌株经过改造可产生中等链长的 α,ω-二醇（聚合物生产的基础材料）（Lu 等人，  2021 年）). 这里的缺点是，这种方法需要详细了解在感兴趣的菌株中驱动 CCR 的分子机制，包括识别被操纵基因的全球监管机构识别的目标。此外，负责 CCR 的调节机制在细菌群之间有很大差异。大肠杆菌等模型细菌的分子细节相对广为人知。大肠杆菌，B . subtilis或Pseudomonads（Görke & Stülke，  2008 年；Moreno 等人，  2015 年；Pei 等人，  2019 年；Rojo，  2010 年；Sonnleitner 等人，  2018 年)，但肯定不适用于其他具有生物技术重要性的细菌（Ruiz-Villafán 等人，  2021 年）。此外，即使与 CCR 相关的问题可以得到解决，其他调控元件也可能在转录或转录后干扰感兴趣基因的表达。特定的代谢物甚至可能变构调节关键酶的活性。换句话说，消除一个监管要素可能无法避免其他要素的影响。

最后一个问题是触发 CCR 的信号没有得到很好的理解。细胞可能会感知充当通量传感器的关键代谢物的浓度，尽管也可能会监测其他元素。最近的一份报告表明，大肠杆菌中碳源利用的等级。大肠杆菌按总碳吸收通量排序，而不是按使用的精确化合物排序（Okano 等人，  2020 年；Okano 等人，  2021 年）). 低于给定的碳吸收通量阈值，CCR 逐渐消失，同时使用存在的化合物。此外，某些代谢物如 1,6-二磷酸果糖（糖酵解途径的中间体）、α-酮戊二酸（三羧酸循环的一个组成部分）和其他代谢物也可以作为调节特定酶活性的通量传感器（ Chubukov 等人，  2014 年；Okano 等人，  2020 年）。最后的画面是精细调整、高度交织和调节的代谢途径，这些途径研究起来很复杂，但非常值得关注。

总之，学习控制细菌的新陈代谢，或者至少对其行为做出良好的预测，需要付出相当大的努力。从互补的角度对此进行研究将很有用，例如检查构成完整代谢途径的基因表达如何受到调节，全球监管机构如何改变转录程序，特定酶的活性如何受到关键信号代谢物的调节，哪些代谢物流动在给定条件下通过哪些路径，这些流量在条件变化时如何变化，路径如何相互关联等等。获得这些知识需要结合传统的分子生物学技术、不同条件下代谢物通量的分析、计算机模拟的使用复杂性不断增加的基因组规模代谢模型将使预测得到实验验证，等等。我们获得的信息越完整，设计细菌生物催化剂的成功机会就越大。

详细了解细胞如何调节和协调其新陈代谢在其他领域也很重要（图 2）。例如，病原菌使用的营养物质会影响毒力基因的表达（Eisenreich 等人，  2010 年）、病原体对特定宿主的适应性（La Rosa 等人，  2018 年），或者它对不同抗生素的敏感性，例如那些使用参与营养物质运输的细胞包膜蛋白进入细胞的细胞 (Martínez & Rojo,  2011 )。代谢偏好也会对细菌群落的组装和行为产生巨大影响，细菌群落的成员必须共享或竞争有限数量的营养素（Bajic & Sánchez，2020 年；Estrela 等 人， 2021 年）。CCR 还影响植物与微生物相互作用中涉及的基因的表达（Franzino 等人，  2022 年），并具有许多意义。显然，深入了解新陈代谢的调节需要在未来几年付出很多努力，但也会带来很多好处。