A hybrid strategy for massive acceleration of directed evolution: meeting the need for high-turnover enzymes in industrial biotechnology.

酶是一种微小的机器，可以加速生物体中的化学过程；没有酶，生命就不会存在，因为基本的化学反应不会发生得足够快。酶发挥这种神奇的功能，首先附着在化学物质上，暂时改变它们的形状，并利用这一点引发化学反应，然后释放产物。一些酶将大的化学物质分解成更简单的部分，而另一些则将较小的化学物质分解成更复杂的化学物质。所有这些都是在常压、室温、中性酸度下发生的，所有成分都是可生物降解的！相比之下，目前的工业化学需要高温和高压，并产生有机废物和污染物。因此，工业反应通常效率不是很高，并经常导致不想要的副产品的形成。有了酶，能量需求相对较少，副产品可以被根除。因此，酶代表着21世纪的一个巨大机遇，它将彻底改变工业化学反应，使其更具成本效益和环境友好性。不幸的是，有一个问题：自然产生的酶可能并不适合工业过程，这或许并不令人惊讶。例如，一种酶可能不能自然地用于手头的化学反应，它们在工业环境中可能不稳定，或者它们可能太慢而不符合成本效益(代谢反应不需要这样的速度)。然而，生物学确实为这些缺点提供了一个潜在的解决方案。每种酶都是由一串被称为氨基酸的100个较小的构件组成的。此外，一种酶可以通过改变(或突变)其氨基酸来转化为另一种酶。有20种可能的氨基酸，在一串100种氨基酸中，有比宇宙中的恒星多得多的组合，很明显，自然界只使用了可能色域中的极小部分。真正的机会是突变天然酶，使它们更稳定、更快，并量身定做它们的特异性，使它们可以用于工业，使化学过程更具成本效益和环境友好，这是这项研究计划的核心。酶的活性部位是一个只有合适的化学物质才能容易和完美地进入的空间，就像锁上的钥匙，正是这个空间的形状和内部运动决定了酶的速度和特异性；突变通过影响活性部位来改变速度和特异性。科学家们天真地认为，在活性部位周围随机突变几个氨基酸就足以实现他们的目标。在这种情况下，比如说，只需要生产和测试100万种组合来确定合适的一种。然而，现在已经发现，酶的氨基酸序列中任何地方的突变都可能影响活性部位，基于随机突变，组合的数量确实是天文数字。在被称为定向进化的过程中已经取得了一些进展，其中随机突变在迭代周期中进行，但远远不足以满足工业，因为对于某些酶来说，这个过程预计需要数千年。我们的愿景是从根本上改变定向进化的执行方式，并减少时间尺度，不是递增的，而是潜在的数百万倍，以促进具有工业相关性质的酶的快速生产。我们计划使用一种酶，单胺氧化酶，这种酶可以用于制造几乎一半的当前开发药物，以显示我们新方法的有效性。这个想法是利用为渲染电脑游戏而开发的硬件，对酶进行非常快速但准确的计算机模拟，以了解整个酶的突变是如何影响活性部位的，并利用这一点来预测定向进化的最佳突变，从而使这个过程在几周内发生，而不是数千年。

Enzymes are tiny machines that speed up chemical processes in living organisms; life would not exist without enzymes, because essential chemical reactions would not happen fast enough. Enzymes perform this miraculous function by first attaching to chemicals, changing their shapes temporarily and using this to elicit a chemical reaction, and then releasing the products. Some enzymes break down large chemicals into simpler parts, while some build smaller chemicals into more complex ones. All this happens at atmospheric pressure, room temperature, neutral acidity and all components are biodegradable! In contrast, current industrial chemistry needs high temperatures and pressures, and creates organic waste and pollutants. Consequently, industrial reactions are typically not very efficient and often lead to the formation of unwanted side products. With enzymes there is relatively little energy demand and side products can be eradicated. Enzymes, therefore, represent a huge opportunity in the 21st Century to revolutionise industrial chemical reactions and make them more cost-effective and environmentally friendly.Unfortunately, there is a catch: naturally occurring enzymes, perhaps unsurprisingly, are not suited to industrial processes. For example, an enzyme may not be naturally available for the chemical reaction at hand, they may be unstable in an industrial setting, or they may be too slow to be cost-effective (metabolic reactions don't demand such speed). Biology does, however, provide a potential solution to these shortcomings. Each enzyme is constructed as a string of 100s of smaller building blocks called amino acids. Furthermore, one enzyme can be converted to another by altering (or mutating) its amino acids. There are 20 possible amino acids and within a string of 100 there are vastly more combinations than stars in the universe and it is very clear that the natural world only uses the tiniest fraction of the possible gamut. The real opportunity is to mutate natural enzymes to make them more stable, faster and tailor their specificity so that they can be used in industry and make chemical processes much more cost-effective and environmentally friendly, which is at the heart of this research proposal.The active site of an enzyme is a space that only the right chemicals can slot into easily and perfectly, like a key in its lock, and it is the shape of this space and motions within that determine an enzyme's speed and specificity; mutations alter speed and specificity by affecting the active site. Scientists thought, naively, that randomly mutating a few amino acids around the active site would be sufficient to achieve their goals. In such a case there would be, say, only a million combinations to produce and test to identify a suitable one. However, it has now been found that mutations anywhere in an enzyme's string of amino acids may affect the active site and based on random mutations the number of combinations are truly astronomical. Some progress has been made in a process called directed evolution, where random mutations are made in an iterative cycle, but not nearly enough to satisfy industry since for some enzymes the process is predicted to take millennia.Our vision is to fundamentally change the way that directed evolution is performed, and reduce the timescales, not incrementally, but by potentially millions of times to facilitate rapid production of enzymes with industrially-relevant properties. We plan to use an enzyme, monoamine oxidase, which could be used in the manufacture of almost half of current developmental drugs, to show the validity of our new approach. The idea is to use very fast, but accurate, computer simulations of enzymes, leveraging hardware developed for rendering computer games, to understand how mutations throughout an enzyme affect the active site and use this to predict the optimal mutations for directed evolution, allowing the process to occur in weeks rather than millennia.