ABSTRACT
The engineering of enzymes can be broadly classified into three different approaches. These are namely, rational design, random library creation, and semi-rational design (Quin and Schmidt-Dannert, 2011). In rational design, the protein structure and mechanism are analyzed through in silico modeling or protein crystallization techniques. The amino acid substitutions that are responsible for a target function are then identified and created by site-specific mutagenesis (Quin and Schmidt-Dannert, 2011). One major drawback of this approach is the need for detailed and accurate structural information for the active configuration of the protein, which is unavailable in many instances. Moreover, this method focuses more on mutations closer to the active site of an enzyme, which are effective at increasing enantioselectivity and creating new catalytic activity (catalytic promiscuity), but mutations more distant from the active site may also be needed to improve activity and stability of the enzyme (Morley and Kazlauskas, 2005).In the absence of detailed structural information, directed evolution strategies, encompassing both random library creation and
semi-rational approaches, have generated more promising results. The evolution and natural selection of enzymes had been occurring for millions of years without human intervention. However, this process occurs far too slowly in nature to meet the needs of industry. In directed evolution, different techniques are applied to evolve proteins and enzymes at a highly accelerated rate so as to generate desired mutants more quickly. In the random library creation approach, mutations are typically distributed randomly along the entire amino acid sequence. This approach has the great advantage of requiring no prior information on the protein being evolved. As such, it is highly suitable for the engineering of enzymes whose functions are not well understood (Chica et al., 2005). This method is also suitable for introduction of completely novel functions into existing enzymes. However, one major drawback to the random library creation approach is that it usually requires a large number of mutants to be created before the desired properties can be obtained. As such, the strategy depends greatly on the availability of high throughput methods for the screening of large numbers of variants. In the semi-rational approach, some knowledge of the enzyme of interest is usually required. This can be in the form of either sequence or structural information. By considering this information, a much smaller library of mutants is created, usually with different assumptions about the active sites. With the smaller library, sometimes known as a “smart” library, the number of mutants to be screened is cut down greatly (Chica et al., 2005).In this chapter, the improvement of industrial enzyme candidates through both random library creation and semi-rational directed evolution strategies will be discussed. As this topic has also been addressed in several earlier reviews (De Carvalho, 2011; Du et al., 2011; Patel, 2011; Rubin-Pitel and Zhao, 2006; Tang and Zhao, 2009; Wang et al., 2012), we will place our emphasis on more recent examples. 3.2 Directed Evolution of Industrial Enzymes
3.2.1 Steps Involved in Directed EvolutionPioneering work in laboratory-based directed evolution began in the 1960s (Lerner et al., 1964). However, the directed evolution strategies that we are familiar with today only started in the
1990s (Chen and Arnold, 1993; Stemmer, 1994). The framework of directed evolution of proteins involves four main steps. The first step entails the identification of the protein of interest, e.g. an enzyme with potential industrial applications. The second step involves the generation of a mutant library with each member of the mutant library carrying a slightly different amino acid sequence from the native protein. In the pioneering studies, the mutant libraries were often generated with methods such as error-prone polymerase chain reaction (epPCR) or chemical mutagenesis (Chen and Arnold, 1993; Lerner et al., 1964). These methods introduce random changes to the amino acid sequence and do not require any structural or sequence information about the protein of interest. In the 1990s, DNA shuffling was also adopted to create mutant libraries (Stemmer, 1994). In DNA shuffling, homologous genes were digested and reassembled in a process that mirrors recombination in nature. Although epPCR and DNA shuffling have resulted in many positive outcomes and are still widely adopted, these methods are often associated with very large mutant libraries and a high proportion of deleterious mutations. To address these limitations, many studies have looked into the creation of “smart” libraries to reduce the library size and also to increase the fraction of beneficial mutations (summarized in Table 3.1). As these library creation methods have been reviewed elsewhere, they will not be covered in detail in this review (Lutz, 2010; Rubin-Pitel et al., 2007; Tang and Zhao, 2009). Table 3.1 Summary of library creation methods Approach Summary ReferencesError-prone PCR (epPCR) Random mutations are introduced during PCR amplification. This is done by increasing the error-rate of the PCR reaction.
