Molecular Biotechnology: Principles and Applications - Glick B., Pasternak J. 2002

Fundamentals of Molecular Biotechnology
Site-Directed Mutagenesis and Protein Engineering
Protein Engineering

Out of the many thousands of studied and characterized Enzymes, just 20 account for over 90% of all enzymes currently used in industry. Table 8.1 lists some of the most important ones and their Applications. The remaining enzymes are not used because their inherent activity fails to meet the demands of highly specialized in vitro processes. Most enzymes rapidly denature at high temperatures and in the presence of organic Solvents—precisely the conditions under which many Industrial processes take place. Of course, thermostable Enzymes can be isolated from thermophilic microorganisms; however, these organisms do not always synthesize the specific enzymes required. Fortunately, these difficulties can be overcome using Site-Directed Mutagenesis and the cloning of target genes.

Class="center">Table 8.1. Selected enzymes and their applications

Formation of additional Disulfide Bonds

The thermostability of protein molecules can be enhanced by introducing modifications that prevent them from unfolding for longer periods at elevated temperatures. Furthermore, such thermostable Proteins often resist Denaturation in organic solvents and under non-physiological conditions (such as extreme pH). A significant increase in the stability of a protein molecule can be achieved by introducing additional disulfide bonds. The main challenge here is to ensure that these bonds do not interfere with the normal functioning of the protein. In one experiment, oligonucleotide-directed mutagenesis was used to generate six variants of phage T4 Lysozyme containing novel intrachain disulfide bonds. To accomplish this, two, four, or six specific amino acid residues in the polynucleotide chain were substituted with Cysteine residues, resulting in The formation of one, two, and three disulfide bonds, respectively (Table 8.2).

The amino acid residues substituted with cysteine were located in close proximity to one another in the active enzyme, so that the overall conformation of the molecule was not significantly altered upon the Formation of the new disulfide bonds. Moreover, they were situated outside the Active Site—the region most sensitive to the slightest conformational changes. Bonds were formed between residues 3 and 97, 9 and 164, and 21 and 142 (with numbering starting from the N-terminus). Following mutagenesis, the mutant genes were identified and expressed in E. coli; the recombinant proteins were then purified, and their enzymatic activity and thermostability were determined (Table 8.2). Thermostability is typically characterized by the Temperature at which 50% of the molecule denatures, with the degree of denaturation determined by monitoring Changes in the circular dichroism of the protein solution. The original (native) form of T4 lysozyme contains two free cysteine residues that do not participate in disulfide bond formation. In the so-called "pseudo-wild-type" enzyme, these are replaced by Thr and Ala residues, while the enzyme's activity and thermostability remain unchanged. The pseudo-wild-type sequence served as a standard for comparison with variants containing potentially thermostabilizing disulfide bonds and also prevented the random formation of disulfide bonds between newly introduced cysteine residues and those already present in the native protein.

Table 8.2. Properties of T4 lysozyme and its six engineered variants¹⁾

¹⁾ Based on data from Matsumura et al., Nature 342:291–293, 1989.

The results of this experiment demonstrated that the thermostability of the enzyme increases with the formation of new disulfide bonds, with the protein possessing the maximum number of such bonds exhibiting the highest thermostability. However, some variants (C, E, and F), while more thermostable than the native or pseudo-wild-type enzyme, lacked enzymatic activity. This is likely due to conformational distortion of the protein molecule caused by the disulfide bond formed between residues 21 and 142. Although engineering novel proteins through genetic techniques is frequently an empirical process (i.e., it is far from always obvious which specific Amino Acid Substitutions will yield the "best" variant), the described experiment clearly shows that producing thermostable proteins with additional disulfide bonds is entirely feasible.

Attempts have also been made to produce a thermostable mutant xylanase from Bacillus circulans, an enzyme that can be utilized in paper manufacturing. One of the steps in this process involves removing hemicellulose from pulp to bleach it, a Procedure that generates large quantities of toxic waste. Treating wood pulp with xylanase reduces The amount of bleaching chemicals required. Unfortunately, prior to enzyme addition, the pulp is treated with hot alkali; furthermore, as modern technologies tend to minimize the amount of Water consumed for pulp cooling, xylanase must remain active at relatively high temperatures.

To determine into which Regions of the polypeptide chain one, two, or three disulfide bonds could be introduced to stabilize the enzyme without disrupting its catalytic activity, computer modeling of the xylanase Spatial Structure was employed. Eight derivatives of B. circulans xylanase were obtained. All of them exhibited higher thermostability than the native enzyme; additionally, three were just as active at 60 °C as the native protein, and one—containing a disulfide bond between the N- and C-termini—was twice as active and retained over 85% of its activity after a 2-hour incubation at 60 °C, whereas the native enzyme completely lost its activity under these conditions within just 30 minutes. The success of these experiments indicates that this strategy can be applied to enhance the thermostability of various enzymes, provided that sufficiently detailed X-ray crystallographic data are available. Nevertheless, it remains uncertain whether thermostable xylanase will see widespread industrial application in paper manufacturing.

Substitution of asparagine with Other Amino Acids

At high temperatures, asparagine and glutamine residues can undergo deamidation, releasing ammonia. By losing their amide group, they are converted into aspartic and glutamic acids, respectively. This results in local Conformational Changes in the polypeptide chain and, consequently, the loss of activity in the proteins containing them.

To determine The Effect of substituting certain asparagine residues in the Triosephosphate isomerase molecule of Saccharomyces cerevisiae on The properties of the enzyme, specific experiments were conducted. Triosephosphate isomerase consists of two identical subunits; each contains two asparagine residues whose substitution can alter the thermal sensitivity of the protein because they are located at the subunit interface. Oligonucleotide-directed mutagenesis was used to replace asparagine residues at positions 14 and 78 (Table 8.3). Substituting one of these with a Threonine or isoleucine residue led to an increase in the enzyme's thermostability, whereas substitution with aspartic acid resulted in a decrease. The enzyme generated by replacing both asparagine residues with aspartic acid residues proved unstable even at normal temperatures and exhibited low enzymatic activity (not shown in Table 8.3).

Table 8.3. Stability at 100 °C of Yeast triosephosphate isomerase and its engineered variants¹⁾

1) Based on data from Ahern et al., Proc. Natl. Acad. Sci. USA 84:675–679, 1987.

The thermostability of the enzyme is evaluated as its half-life (rate of Enzyme inactivation) at 100 °C. The higher this value, the more stable the enzyme.

Assessing the susceptibility of recombinant proteins to proteolytic Cleavage revealed a positive correlation between protein thermostability and resistance to proteolysis. These findings suggest the possibility of engineering thermostable forms of Other Enzymes by substituting non-essential asparagine residues.

Reduction in the number of free sulfhydryl groups

A foreign protein synthesized in a host Organism sometimes proves less active than expected, and Introduction/32.html">Genetic Engineering techniques can be employed to enhance its activity. For instance, when cloned complementary DNA (cDNA) encoding human $\beta$-interferon (IFN-$\beta$) was expressed in E. coli, the protein product exhibited tenfold lower antiviral activity than the native glycosylated form. Although IFN-$\beta$ was synthesized in fairly large quantities, nearly all of its molecules formed dimers and higher-molecular-weight inactive complexes.

Nucleotide sequence Analysis of the IFN-$\beta$ Gene revealed the presence of three cysteine residues, one or more of which likely participate in disulfide bond formation, leading to the aggregation of dimers and oligomers in E. coli Cells, but not in human cells. It was hypothesized that replacing one or more cysteine codons with Serine codons would yield an interferon that does not form oligomers. Serine was chosen because its structure is similar to that of cysteine, except that it contains an oxygen atom instead of sulfur and therefore cannot form disulfide bonds.

At the outset of these experiments, the researchers lacked detailed information regarding the molecular structure of $\beta$-interferon and had to rely on data from related proteins. In other words, they did not know which of the three cysteine residues was responsible for forming intermolecular disulfide bonds. Fortunately, the locations of the cysteine residues involved in intramolecular disulfide bond formation in the structurally analogous IFN-$\alpha$ molecule were known, enabling a sequence comparison between the two molecules (Fig. 8.8). The analysis showed that residues Cys-31 and Cys-141 in IFN-$\beta$ occupy the exact same positions as residues Cys-29 and Cys-138 in IFN-$\alpha$. Since the latter participate in intramolecular disulfide bonding in IFN-$\alpha$, it was reasonable to assume that Cys-17 in IFN-$\beta$ is not involved in such bonding and can be substituted.

This assumption proved correct: when synthesized in E. coli cells, Ser-17-IFN-$\beta$ did not form multimeric complexes. Furthermore, this interferon exhibited the same specific activity as authentic native IFN-$\beta$ and demonstrated greater stability during long-term storage than the native form.

Enhancing Enzymatic Activity

Site-directed mutagenesis allows not only for improving enzyme stability but also for altering catalytic activity. Currently, to substantially modify the enzymatic activity of any well-characterized enzyme, detailed information regarding the geometry of its active site is required. In this case, it is possible to predict which amino acid substitutions must be introduced to alter the enzyme's substrate Specificity.

The potential of this approach is illustrated by an experiment modifying the substrate-binding specificity of tyrosyl-tRNA synthetase from *B. stearothermophilus*. This enzyme catalyzes the aminoacylation of tRNA that specifically binds Tyrosine (tRNA^Tyr) via a two-step reaction:

Fig. 8.8. Localization of cysteine residues in IFN-α and IFN-β molecules involved in disulfide bond formation. Experimentally identified intramolecular disulfide bonds in IFN-α are indicated by solid lines, while the putative bond in IFN-β is shown by a dashed line.

In step 1, ATP activates tyrosine (Tyr), yielding enzyme-bound tyrosyl adenylate (Tyr-AMP) and pyrophosphate (PP_i). In step 2, tyrosyl adenylate is hydrolyzed by the free 3'-hydroxyl group of the tRNA molecule, transferring tyrosine to the tRNA with the release of AMP. Throughout both reactions, the substrates remain bound to tyrosyl-tRNA synthetase.

By the time this experiment was conducted, the three-dimensional structure of *B. stearothermophilus* tyrosyl-tRNA synthetase had been resolved and its active site mapped, enabling computer modeling to predict the effect of substituting one or more amino acid residues on enzyme-substrate interactions. To test these predictions, oligonucleotide-directed mutagenesis was employed to introduce specific Mutations into the tyrosyl-tRNA synthetase gene. The threonine residue at position 51 (Thr-51) was replaced by either Alanine or Proline. In the native enzyme, the hydroxyl group of Thr-51 forms a Hydrogen bond with the ribose ring oxygen of tyrosyl adenylate; it was hypothesized that breaking this weak bond would increase the enzyme's affinity for ATP.

To characterize the resulting enzymes, their kinetic constants were determined. In some instances, the changes were more pronounced than anticipated (Table 8.4). For example, while the binding constant (K_M) for ATP in the Ala-51 enzyme decreased approximately twofold without a significant change in the catalytic constant (k_cat), it dropped more than 100-fold for the Pro-51 enzyme. Concurrently, the catalytic efficiency (k_cat/K_M) of the aminoacylation reaction increased in both cases. The result obtained for the Pro-51 enzyme was unexpected, because replacing threonine with proline should have disrupted (at least locally) the α-Helix structure in this region, which was presumed to negatively impact substrate binding.

Table 8.4. Aminoacylation efficiency of native (Thr-51) and modified (Ala-51 and Pro-51) tyrosyl-tRNA synthetases¹⁾

¹⁾ Based on data from Wilkinson et al., Nature 307: 187–188, 1984.

These findings demonstrate that despite The complexity of predicting the outcomes of specific amino acid substitutions, the described approach successfully identifies side chains whose replacement improves the kinetic properties of an enzyme. Furthermore, it became evident that both substrate affinity and catalytic efficiency can be enhanced *in vitro* by introducing targeted modifications into the cloned gene.

Altering Metal Cofactor Requirements of Enzymes

Subtilisins—Serine proteinases secreted into the culture medium by Gram-positive Bacteria—are widely used as biodegradable detergent additives. They tightly bind one or more Calcium Ions, which enhance their stability. Unfortunately, subtilisins are often employed in industrial processes involving high concentrations of metal-chelating agents, including calcium, under which conditions they rapidly lose activity. To overcome this limitation, researchers first attempted to abolish calcium binding in subtilisin and subsequently increase the Stability of the modified enzyme.

Engineering the modified subtilisin began with identifying the BPN' gene from *Bacillus amyloliquefaciens*. High-resolution X-ray crystallography was first used to determine the Cell/13.html">Protein Structure, after which oligonucleotide-directed mutagenesis was applied to create a mutant gene with a deletion of the NUCLEOTIDES encoding amino acid residues 75 through 83, which comprise the calcium-binding site. The deletion protein did not bind calcium and, surprisingly, retained a conformation nearly identical to that of the wild-type enzyme.

Next, to enhance the stability of the calcium-binding-site-deficient subtilisin, target sites for modification and appropriate amino acid replacements were identified. It was hypothesized that the amino acids interacting with the calcium-binding domain in the native enzyme were suboptimal under the new conditions. Ten candidate amino acids were selected. Because the specific residue substitution yielding the highest stabilization could not be predicted *a priori*, random mutagenesis was utilized to introduce changes at each of the 10 sites.

The selected amino acids were located in four distinct regions of the protein molecule: the N-terminal segment (residues 2 through 5), the omega-loop (residues 36 through 44), the α-helical region (residues 63 through 85), and one of the β-sheets (residues 202 through 220). To identify the optimal amino acid for each site, mutant clones were cultured in microtiter plates, incubated at 65 °C for 1 h, cooled, and assayed for subtilisin activity. *Bacillus subtilis* had to be used to synthesize the active, non-calcium-binding subtilisin, as this protein proved toxic to *E. coli*.

Preliminary screening revealed stabilizing mutations in 7 of the 10 investigated sites (Table 8.5). The enzyme generated by combining these alterations into a single gene exhibited kinetic properties comparable to those of the native subtilisin. Moreover, in the absence of calcium, the mutant subtilisin was nearly 10 times more stable than the wild-type enzyme and, remarkably, about 50% more stable than the native enzyme in the presence of calcium. These results indicate that despite the labor-intensive nature of such experiments, genetic engineering can successfully modify enzyme properties that depend on A large number of amino acid residues.

Table 8.5. Effect of random amino acid substitutions at specific sites On the Stability of calcium-binding-site-deficient subtilisin BPN'¹⁾

¹⁾ Based on data from Strausberg et al., Bio/Technology 13: 669–673, 1995.

²⁾ Amino acid substitutions to Cys at positions 3 and 206 occurred within the same clone; disulfide bond formation between these residues rendered the protein exceptionally stable.

Altering Enzyme Specificity. Oligonucleotide-directed mutagenesis is primarily used to enhance pre-existing enzyme properties, but it can potentially be adapted to engineer enzymes with novel specificities. For instance, this approach has been used to generate new site-specific endonucleases based on the relatively nonspecific endonuclease FokI. To date, over 2,500 restriction-modification systems have been identified across A wide variety of organisms. Many of these recognize identical nucleotide sequences, yielding a total of approximately 200 distinct restriction sites, the majority of which are 4 to 6 bp in length. Restriction Endonucleases recognizing such sites cleave DNA at numerous locations and are less frequently employed for generating large DNA fragments compared to restriction enzymes that recognize nucleotide sequences of 8 bp or longer. Screening for novel restriction endonucleases is a challenging and time-consuming task. It is unlikely that a sufficient number of enzymes recognizing sites of 8 bp or more can be discovered naturally, necessitating alternative genetic engineering strategies to obtain novel restriction enzymes.

Fig. 8.9. Genetically engineered construct encoding the recombinant "zinc finger - FokI restriction endonuclease" fusion protein.

There is a fascinating class of proteins whose molecules contain unique Structural domains that bind Zn²⁺ atoms—the so-called zinc fingers. These proteins bind to specific nucleotide sequences by inserting their a-helical region into the major groove of the DNA double helix. For instance, the Zif268 protein from mouse cells contains three zinc fingers, each interacting with a distinct DNA codon. Because these fingers bind to DNA independently of one another, they can be combined within a single peptide to target a specific site. This approach allows the creation of Nucleases that cleave DNA at unique sites by fusing The nucleotide sequences encoding zinc fingers with a portion of the gene for the non-specific FokI nuclease from the bacterium Flavobacterium okeanokoites. To test this hypothesis, a chimeric gene was constructed encoding a six-Histidine tag at the N-terminus of the protein molecule to facilitate purification of the recombinant protein, three zinc fingers, a (Gly₄Ser)₃ linker to impart flexibility to the recombinant molecule, and a segment containing the FokI nuclease gene portion (Fig. 8.9). Following purification of the recombinant protein, the N-terminal histidine residues were removed by Thrombin Treatment.

Bacteria that synthesize restriction endonucleases protect their own DNA from cleavage by using enzymes that methylate the specific recognition sites of the restriction endonuclease. However, the host cell genome is not protected against the recombinant FokI restrictase. To prevent the death of growing cells, the Synthesis of the hybrid enzyme was suppressed by placing its gene under the control of the bacteriophage T7 expression system.

As a result of these experiments, two recombinant FokI restriction endonucleases were obtained. One of them cleaved λ phage DNA at the expected site, whereas the other cleaved it at the expected site and—to a lesser extent—at two other sites. This is not surprising, given that zinc fingers primarily recognize two out of the three bases in a triplet. Although these recombinant enzymes cannot yet be used in laboratory practice, the described approach for generating custom restriction endonucleases holds significant promise.

Enhancing Enzyme Stability and Specificity

Tissue plasminogen activator (tPA) is a multi-domain serine proteinase used clinically to dissolve Blood clots. Unfortunately, tPA is rapidly cleared from the Circulatory system, necessitating continuous infusion. To achieve the desired therapeutic effect, high enzyme concentrations must be used, which can lead to nonspecific internal bleeding. Therefore, it would be highly desirable to develop a long-lived tPA variant that exhibits high affinity for fibrin in thrombi without causing hemorrhaging. Proteins with such properties can be engineered by introducing specific mutations into the native tPA gene. Substituting Asn for Thr-103 yielded an enzyme that persists in rabbit plasma approximately 10 times longer than the native variant. Replacing amino acids 296–299 (Lys-His-Arg-Arg) with Ala-Ala-Ala-Ala resulted in a substantial increase in the enzyme's affinity for fibrin. Substituting Gln for Asn-117 produced an enzyme with a fibrinolytic activity comparable to that of the wild-type enzyme. Combining all three mutations into a single protein yielded an enzyme possessing all three properties (Table 8.6). Further studies are required to determine whether it can effectively replace native tPA.

Table 8.6. Stability and activity of various mutant tPA enzyme variants^1),2)

¹⁾ Adapted from Pena et al., Proc. Natl. Acad. Sсi. USA 91: 3670–3674, 1994.

²⁾ All values are normalized relative to those of the native enzyme. Plasma stability is inversely proportional to the time required for plasma clarification; higher values indicate greater stability. Fibrin affinity correlates with clot dissolution efficiency. Plasma activity shows a negative correlation.