International Journal of Biological Macromolecules 214 (2022) 642–654

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules

journal homepage: [Link]/locate/ijbiomac

Review

Thermal stability enhancement: Fundamental concepts of protein

engineering strategies to manipulate the flexible structure
Mahdie Rahban a, Samaneh Zolghadri b, Najmeh Salehi c, Faizan Ahmad d, Thomas Haertlé e,
Nasrollah Rezaei-Ghaleh f, g, Lindsay Sawyer h, Ali Akbar Saboury a, *
a
Institute of Biochemistry and Biophysics, University of Tehran, Tehran 1417614335, Iran
b
Department of Biology, Jahrom Branch, Islamic Azad University, Jahrom 7414785318, Iran
c
School of Biological Science, Institute for Research in Fundamental Sciences (IPM), Tehran, Iran
d
Department of Biochemistry, Jamia Hamdard, New Delhi 110062, India
e
National Institute of Agronomic and Environmental Research, 44316 Nantes, France
f
Institute of Physical Biology, Heinrich Heine University Düsseldorf, 40225 Düsseldorf, Germany
g
Institute of Biological Information Processing, IBI-7: Structural Biochemistry, Forschungszentrum Jülich, 52428, Jülich, Germany
h
Institute of Quantitative Biology, Biochemistry and Biotechnology, The University of Edinburgh, King's Buildings, Edinburgh EH9 3JR, UK

A R T I C L E I N F O A B S T R A C T

Keywords: Increasing the temperature by just a few degrees may lead to structural perturbation or unfolding of the protein
Flexible loops and consequent loss of function. The concepts of flexibility and rigidity are fundamental for understanding the
Thermal stability relationships between function, structure and stability. Protein unfolding can often be triggered by thermal
Protein engineering
fluctuations with flexible residues usually on the protein surface. Therefore, identification and knowledge of the
effect of modification to flexible regions in protein structures are required for efficient protein engineering and
the rational design of thermally stable proteins. The most flexible regions in protein are loops, hence their
rigidification is one of the effective strategies for increasing thermal stability. Directed evolution or rational
design by computational prediction can also lead to the generation of thermally stable proteins. Computational
protein design has been improved significantly in recent years and has successfully produced de novo stable
backbone structures with optimized sequences and functions. This review discusses intramolecular and inter­
molecular interactions that determine the protein structure, and the strategies utilized in the mutagenesis of
mesophilic proteins to stabilize and improve the functional characteristics of biocatalysts by describing efficient
techniques and strategies to rigidify flexible loops at appropriate positions in the structure of the protein.

1. Introduction heat-induced reversible transition between N and U. In protein

biophysics, the standard conditions are set to a temperature of 298.15 K
Two protein states are involved in the heat-induced denaturation: (25 ◦ C), a pressure at 105 Pa (1 bar) and the solvent as water (or dilute
the natively folded state (N) and the unfolded (denatured) state (U). This buffer) containing no chemical denaturant. The Gibbs free energy
process can be described variously as “thermostability”, “thermal sta­ change is usually denoted by ΔG0U, where subscript ‘U' refers to the
bility” and “thermodynamic stability”. In the literature, these stability process N ↔ U, and superscript ‘0 (zero)’ refers to the standard state with
concepts are often used interchangeably. A few terminological com­ the solvent water (or dilute buffer) containing no chemical denaturant.
ments are therefore necessary: (i) Thermostability is defined as the (iii) Two different proteins having same Tm (thermostability) frequently
temperature at which protein is 50 % denatured (or in other words, in a have different ΔG0U (thermodynamic stability) values and vice versa
two-state model, the fraction of the native (fN) and that of the unfolded [1,2]. For example, Tm of cytochrome-c and lysozyme is, within exper­
(fU) protein molecules equal to 0.5). This temperature is conventionally imental error, the same (82 ◦ C), but the ΔG0U values are 9 and 15 kcal/
called Tm (midpoint of denaturation). (ii) Thermodynamic stability is mol, respectively. On the other hand, myoglobin and RNase have Tm
defined as the standard Gibbs free energy change associated with the values of 82 and 63 ◦ C, respectively, but their ΔG0U values are the same

* Corresponding author.
E-mail addresses: mrohban@[Link] (M. Rahban), [Link]@[Link] (S. Zolghadri), nsalehi@[Link] (N. Salehi), faizanahmad@[Link] (F. Ahmad),
haertle@[Link] (T. Haertlé), [Link]@[Link] (N. Rezaei-Ghaleh), [Link]@[Link] (L. Sawyer), saboury@[Link] (A.A. Saboury).

[Link]
Received 17 March 2022; Received in revised form 22 June 2022; Accepted 23 June 2022
Available online 27 June 2022
0141-8130/© 2022 Elsevier B.V. All rights reserved.
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

(12 kcal/mol). (iv) Thermal stability may mean thermostability and/or role in the function of proteins. Evolutionary studies show that sequence
thermodynamic stability. Often in the literature, the protein with the changes occur commonly in loop regions [22]. Residues in loop regions
higher Tm is described as being the more thermally stable. In this review have fewer contacts with other residues in the protein structure.
article, we shall use the description of thermostability in terms of Tm. Therefore, they are subject to more fluctuations and flexibility in the
The proteins and enzymes are classified into four categories based on protein conformation and may trigger protein thermal unfolding. Hence,
the optimum growth temperature (OGT) of the organisms from which a working strategy to improve thermal stability is to introduce mutations
they were extracted: psychrophile, mesophile, thermophile, and hy­ rigidifying these residues. Successful point mutations for designing
perthermophile [3]. Accordingly, Psychrophilic enzymes are active up thermally stable proteins include engineering surface salt bridges
to 25 ◦ C while mesophilic enzymes are optimally active between 25 and [23–27], introducing supplementary disulfide bridges [28–32], loop
50 ◦ C [4]. Thermophilic enzymes act at optimal temperatures from 60 to anchoring by H-bonding and hydrophobic interactions [33–36],
110 ◦ C. The optimum temperature of hyperthermophilic enzymes is decreasing the entropy of unfolding in order to rigidify these flexible
above 80 ◦ C and these enzymes show low activity at lower temperatures loops [18,37–40]. In addition, deletion or shortening of external loops or
[5]. Intermolecular interactions are correlated with the entropy- creating better packing on the surface of the protein may be tried
enthalpy balance and characterized by sequence-structure de­ [22,41,42].
terminants. The specific function requires a unique structure. Hence, the This review centers on a brief description of the fundamental con­
intra- and inter-molecular contacts which provide thermal stability at cepts and underlying mechanisms of protein thermal stability and the
the optimal growth temperature of the organisms are related to the factors involved in thermal stability. It also discusses the concepts of
amino acid compositions [6]. For example, among all possible combi­ flexibility and rigidity in thermophilic proteins. Besides, it describes
nations of amino acids, the fraction of Ile, Val, Tyr, Trp, Arg, Glu, and rigidifying flexible structures as an effective strategy for engineering
Leu (IVYWREL) in a proteome can be a predictor of thermal adaptation proteins to increase their thermal stability and reviews techniques for
of proteins which has been studied by Berezovsky and co-authors [7]. identifying and predicting flexible loops structures. Consequently, it
The thermostability of the majority of natural proteins is moderate. highlights the current state of knowledge by focusing on thermal sta­
Hence, they cannot be effective as biocatalysts, biosensors or therapeutic bility improvement by mutation of the flexible loops of proteins as well
agents in industry at higher temperatures. Elimination of microbial as discussing the opportunities and limitations of protein thermal en­
contamination, solubility of substrate, protein production yield, and gineering strategies. The readers who are interested in new approaches
reaction rate improve with an increase in the protein thermal stability in protein engineering are highly recommended to review the papers of
[8]. Increased stability can also influence local or global unfolding rates Narayanan et al. [43], Greenhalgh et al. [44], Pongsupasa, et al. [45],
of proteins. Therefore, the immunogenic response to aggregate forma­ Liu et al. [46] and Roda et al. [47].
tion is reduced in therapeutic applications [9]. In addition, due to
increased kinetic stability parameters (which is described in Section 2), 2. Stability definition in thermal denaturation
thermostable biocatalysts can be used frequently in harsh conditions
[10]. We have defined above the thermodynamic stability (ΔG0U) of pro­
Proteins as sustainable biomolecules or natural catalysts are widely teins. Estimates of ΔG0U generally come from the study of heat dena­
used in industry, medicine, scientific research and biotechnology. turation using differential scanning calorimetry (DSC) and equilibrium
Therefore, designing and constructing proteins with a desired function methods [48]. DSC provides a direct estimate of the enthalpy change of
has turned to protein engineering by gene modification or even the denaturation (ΔH0U) and the constant-pressure heat-capacity change
creation of a protein ab initio [11]. Various protein engineering strate­ (ΔC0p). In the second approach, the equilibrium constant is measured
gies have designed many proteins that demonstrate better catalytic ac­ from the heat-induced conformational-transition curve representing the
tivities and stabilities in severe conditions [12–14]. However, in some equilibrium between N and U [49–52]. The latter method is hence called
cases these strategies were unsuccessful largely because the relevant the equilibrium method. The globular mesophilic proteins typically have
information concerning the sequence-structure-function of proteins was ΔG0U between 5 and 15 kcal/mol [53]. In a review article, Vieille and
insufficient [15]. Therefore, rational protein engineering via sequence Zeikus [53] demonstrated a slight variation in ΔG0U values of thermo­
alteration aiming to increase protein thermostability is still a challenge. philic and mesophilic proteins, in the range of 5 to 20 kcal/mol. Thermal
As a result, deep sequencing, directed evolution, and different experi­ stability concerns with ΔG0U and Tm measurements for the equilibrium
mental and theoretical methods are commonly used to improve thermal process, N ↔ U. In contrast, kinetic stability involves a particular free
stability providing necessary clues to protein engineering techniques energy barrier that separates N and U. It is usually characterized by
[16]. The ability to predict reliably the alteration of protein stability parameters such as activation energy for unfolding (ΔG⧧U), optimum
would be valuable in the understanding of the relationship between reaction temperature (Topt), apparent inactivation constant (kobs), semi-
protein structure and function, and could assist in the evolution of inactivation temperature (T50) and it is generally expressed as half-life
protein sequences for predetermined structural properties. Furthermore, (t1/2) at a defined temperature [13,54]. In fact, kinetic stability is
unraveling the mechanisms behind the protein modifications required related to the time-dependent process of irreversible denaturation of
by the extreme environmental conditions of their host helps us to opti­ protein [55]. Stability in enzyme applications in industry refers to the
mize the rational design of proteins and paves the way toward efficient irreversible process of protein denaturation with time and is related to
de novo design [17]. activation energy barrier between the native state and the non-
Recently, a remarkably successful strategy in the engineering of functional forms, such as the unfolded state (s) and the irreversibly
protein thermal stability has been to rigidify flexible sites in protein denatured protein [56].
structures [18]. Thermophilic proteins are generally characterized by an
increase in global conformational rigidity relative to those from meso­ 2.1. Thermally stable proteins
philic organisms, while psychrophilic proteins exhibit a relatively high
degree of flexibility [19]. Therefore, identification of amino acids in the Thermophilic enzymes (thermozymes) have unique structural and
desired regions from structural data and mutation of them by random or functional features of stability, which help them to act at optimal tem­
site-directed mutagenesis, is applied to enhance the thermostability of peratures from 60 to 110 ◦ C [5]. Some theories and experimental results
proteins. In addition to helices and β-strands, folded proteins have loops. suggest that hyperthermophile bacterial and archaeal species were the
Loops generally belong to the most flexible parts of structures, which first life forms on Earth [53,57]. Proteins are highly dynamic structures
include turns and random coils [20,21]. Loops connect helices and/or with specific functions. The structural position of each residue, its
β-strands, are often exposed on the outer protein surfaces and play a vital pairwise interactions and conformational energy have evolved to allow

643
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

the whole protein to tolerate the thermal stress in which it is to be found general, the intramolecular interactions underlying protein thermal
[58]. As proposed by Berezovsky et al. [59] the thermostability of pro­ stability involve a wide range of mechanisms such as higher packing
teins is governed by “structure-based” and “sequence-based” physical efficiency through van der Waals interactions, disulfide bridges, elec­
mechanisms depending on the evolutionary history of organisms. The trostatic interactions, hydrogen bonds, hydrophobic interactions,
structure-based mechanisms often rely on numerous interactions rather conformational strain reduction, and loop shortening [70]. On the other
than a particular interaction type and frequently lead to highly compact hand, the structural requirements of thermal stability could impose
thermostable proteins. On the other hand, the sequence-based mecha­ challenges in terms of protein folding and functionality [58].
nisms are found in thermostable proteins with no pronounced structural
differences with their mesophilic homologs, instead, their thermal sta­
bility is rooted in a small number of apparently strong interactions. 2.2. Factors involved in the protein thermal stability
Unlike organisms that originated in extreme conditions which employ
the structure-based mechanisms of thermal stability, the sequence-based One approach to engineering the thermal stability of protein is to
mechanisms seem to be more prevalent in organisms that originally identify differences by comparing the structures and sequences of ho­
evolved as mesophiles but then recolonized and adapted to hot envi­ mologous proteins of hyperthermophiles, thermophiles, mesophiles, and
ronments [59]. psychrophiles. Some studies show that the function of thermophiles at
During the evolutionary process, the secondary and tertiary struc­ high temperatures is not dependent on covalent modifications or
tures of proteins are often better conserved than the primary structures structural motifs or different amino acids. Instead, it seems that thermal
[60]. Hyperthermophilic, mesophilic and psychrophilic enzymes differ stability is related to the combination of many minor structural alter­
only in the range of activation and stability temperatures. Several ations and modulation of the canonical forces, which are acquired by
studies showed that the sequence homologies of these proteins are from exchange of some amino acids for others [71,72].
40 to 85 % identical, their 3D structures are superimposable and their
catalytic mechanisms are the same. For example, transcription of the 2.2.1. Hydrophobic effect
arginine biosynthetic genes in bacteria is regulated by an arginine The primary driving force behind protein folding is the hydrophobic
repressor (ArgR) [61]. Sequence alignments of two ortholog proteins, effect. The hydrophobic patterns for proteins are due to the burial of
thermophilic and mesophilic ArgR, display 31 % sequence identity and hydrophobic residues in the protein core, leading to better packing and
some changes in amino acid types, as shown in Fig. 1. But these two to increased van der Waals interactions [73,74]. The tighter packing is
thermophilic and mesophilic proteins show similar tertiary structures dependent on the proportions of alanine, isoleucine, valine and proline
which are superimposed (Fig. 2) [61,62]. As illustrated in Fig. 2, the in the hydrophobic core [8,75,76]. Introduction of a methyl group in­
most structural changes in these two ortholog proteins belong to the creases the hydrophobic effect within the protein core and improves the
loops. It can be seen that the length of loop regions and the B-factor protein thermal stability. It is estimated that each additional methyl
values, which describes the thermal motions, are decreased in the group buried during protein folding enhances 1.3±0.5 fold of protein
thermophile protein. On the other hand, protein dynamics seem to play thermal stability [77]. In addition, studies demonstrate that the
a significant role in thermal stability. conserved similarity in the core of mesophilic and hyperthermophilic
The 3D structures of thermophilic enzymes are more rigid than protein homologs is higher than in their solvent-exposed regions [78].
mesophilic homologs at 25 to 40 ◦ C, and their rigidity is essential for Deeper analysis of hydrophobic cores in thermophilic proteins suggests
maintenance of thermal stability [53]. Several lines of evidence suggest that hydrophobic forces at higher temperatures increase the stability of
that the unfolding process of psychrophilic, mesophilic and thermo­ proteins [79]. In contrast, the weakening of the hydrophobic effect is
philic proteins is triggered by the fluctuation of the more flexible regions mainly related to cold denaturation [80,81]; thus, the hydrophobicity of
of loops undergoing low activation energy motions [63–68]. The aromatic and hydrophobic residues decreases in psychrophilic proteins,
structural comparison of these proteins (ArgR) also points to the higher whereas the hydrophobic amino acids increase in thermophile proteins
rigidity of thermophilic proteins, particularly on their loops, which is [3]. Mutagenesis of the hydrophobic core usually changes the stability of
depicted in Fig. 2 based on the B-Factors [22]. Hence, thermophilic a protein due to interruption of the well-packed core residue side chains.
enzymes are suitable candidates for understanding the molecular Thus substitution of a small hydrophobic residue by a large one may
mechanisms of protein thermal stability, the temperature constraints of introduce unfavorable contacts and vice versa which will introduce a
enzyme function, and even the enzyme evolution process [53,57,69]. In cavity, may lead to instability [82]. Therefore, it is clear from the above
that thermal stability depends significantly upon hydrophobic effects.

Fig. 1. The pairwise sequence alignment of arginine repressor (ArgR) protein in Geobacillus stearothermophilus as a thermophile versus Escherichia coli K12 as a
mesophile is colored based on identity (A) and residue type (B). The thermophilic and mesophilic ArgR display 31 % of sequence identities, the difference between
amino acid compositions lead to different intramolecular interactions.

644
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

Fig. 2. A) Structural alignment of ArgR in Geobacillus stearothermophilus as a thermophilic protein (PDB ID:1B4B, red) versus ArgR in Escherichia coli K12 as a
mesophilic protein (PDB ID:1XXA, blue). The superimposable tertiary structures illustrate that secondary and tertiary structures of proteins are conserved more than
the primary structure. In addition, the percentage of flexible turn and coil structure is higher in mesophilic ArgR. B) The structural alignment is colored based on the
B-Factor. The lower B-factor displays rigid loops in thermophile rather than mesophile ArgR. (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)

2.2.2. Electrostatics interactions and hydrogen bonds bonds in hyperthermophilic proteins have a role against thermal
Some studies show that the content of charged amino acids such as unfolding and denaturation by increasing the kinetic stability by raising
glutamic acid, lysine, and arginine are higher in thermophilic proteins. the activation energy barrier between the folded to the unfolded states
Thus, this will lead to extra electrostatic interactions like salt bridges [94]. Removing a natural disulfide bond interferes with protein folding
and ionic networks [83–85]. Electrostatic interactions are basically and decreases thermal stability because of higher conformational en­
divided into three types of interactions: charge-charge, charge-dipole, tropy [78]. It is interesting to note that the de novo engineering of di­
and dipole-dipole [86]. Fully ionized, oppositely charged residues create sulfide bonds in flexible regions of the protein increases thermal stability
a charge-charge interaction known as a salt bridges [87]. while disulfide bridges in fully structured regions have no impact [8].
Salt bridges in proteins are mostly exposed to the aqueous solvent.
Such exposed bridges will contribute very little to the thermal stability 2.2.4. Aromatic interactions
of proteins, for the oppositely charged residue-residue interaction can be Thermal stability is a cumulative effect of all weak protein in­
replaced by residue-water interaction. On the other hand, the direct teractions. Weak interactions involve the hydrophobicity, hydrophobic
interactions between the charged surface residues and solvent that in­ clusters and aromatic clusters [95]. Aromatic interactions are one of the
crease stability play important roles in the formation of ion network at factors in the structural stability of proteins and include π-π, cation-π and
high temperatures [88]. However, if a salt bridge is buried in the folded aryl‑sulfur interactions [96]. Thermally stable proteins have additional
protein, it will contribute very significantly to its stability [89]. The aromatic clusters located on the surface of the protein or near the pro­
impact of electrostatic interactions on thermal stability is an enigmatic tein's active site. Protein stability is increased by 0.6–1.3 kcal/mol by a
issue. There is no direct relationship between the number of ionized pair of aromatic interactions [78,97]. The aromatic residues such as Trp,
residues, ion pairs, or ion-pair networks and the thermal stability of the Tyr and Phe, and His increase the thermal stability of proteins [95]. The
protein. Instead, the enhanced stabilities are related to optimization of free energy of interaction Phe-Phe in a helix has been reported as be­
electrostatic interactions due to the optimized positioning of amino tween − 0.1 and − 0.8 kcal/mol. Furthermore, Leu-Tyr and Phe-Met
acids [78]. The presence of salt bridge networks together with hydrogen have the same magnitude of free energy as hydrophobic interactions
bonds abundant in thermally stable proteins contributes to their struc­ (see above) [82]. In addition, the number of π–π clusters, pairwise π–π
tural rigidity [90]. Conversely, they are absent in psychrophilic proteins interactions and cation–π interactions, especially those involving argi­
due the flexibility necessary for protein activity [4]. nine is increased in the hyperthermophilic proteins [98]. An interesting
Hydrogen bonds are formed within and between polypeptide chains example of aromatic interactions is provided by the so-called tryptophan
and with the surrounding aqueous medium [86]. The energy of forma­ zipper. These are β-hairpins which are stabilized by Trp-Trp interactions
tion of a single hydrogen bond buried in the protein is estimated at alone. They show Tm values varying from 48.5 to 78 ◦ C depending upon
approximately 0.6 kcal/mol, contributing to the total thermodynamic the peptide length [99].
stability of proteins [91]. Besides, introduction of a hydrogen bond into
a protein structure enhances the thermal stability of the protein. For 2.2.5. Reduction of the entropy of unfolding
example, replacing the buried Ala170 with serine, thereby forming an Proteins are stabilized when their entropy of unfolding decreases.
internal hydrogen bond, was shown to increases the Tm of Bacillus Rigid residues like proline reduce the conformational entropy of
neutral proteinase by 0.7 ◦ C, in good agreement with the general esti­ unfolding [100]. The highest mainchain conformational entropy be­
mate [92]. longs to the glycine residue, while the proline residue with its pyrroli­
dine ring has the lowest. During the folding process, therefore,
2.2.3. Disulfide bridges restriction of glycines requires much more energy than that of the pro­
Disulfide bridge formation decreases the entropy of the unfolded line residue which constrains available conformation of the preceding
state (U) of proteins and causes protein stabilization [71]. The mutation residue [71]. Therefore, replacing glycine by proline or mutating other
or biochemical interruption of intramolecular disulfide bonds can amino acids to proline can increase protein stability. The most
decrease the Tm of thermophilic proteins [93]. It seems that disulfide remarkable stability increase is observed when such substitution occurs

645
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

in reverse turns or at the N-terminal caps of helices. However, such constituted by motions occurring at a broad range of timescales from
substitutions may actually decrease the overall stability if they replace pico- to milli-seconds and length scales from Å (angstrom) to several
too many favorable interactions with unfavorable ones [19,101]. nanometers, corresponding to local bond fluctuations and long-range
inter-domain motions. Various motional modes of protein are acti­
2.2.6. Compactness vated consecutively as a function of temperature and exhibit varying
The protein's radius of gyration (Rg) and solvent-accessible surface degrees of coupling to solvent dynamics [112]. In general, protein
area (SASA) are described as compactness in protein structure [102]. flexibility underlies protein function and increases the conformational
Compactness enhances protein thermal stability, which in turn depends entropy of the folded state. Hence, it seems that structural fluctuations in
upon an increased number of atoms buried and so hidden from solvent, thermophilic proteins are related to the entropic part of free energy,
loop shortening, eliminating unnecessary cavities and optimizing the which can enhance thermal stability [113]. Recent findings show that
packing of side chains in the interior of the protein all contribute to hyperthermophilic enzymes are often inactive in the temperature range
density increase [83]. The more density of non-covalent interactions of 20–37 ◦ C because these enzymes lack sufficient flexibility to show full
such as van der Waals contacts, H-bonds, and electrostatic interactions activity [114]. Several thermophilic enzymes have been characterized
all contribute to an increase the protein compactness [59]. In addition, as more efficient biocatalysts than their mesophilic homologs at 37 ◦ C.
stabilization of the N- and C- termini of proteins and loops by H-bonds This suggests that these enzymes have local flexibility in their active
and ion pairs can also improve the protein compactness [103]. sites compared to their overall rigidity [53]. Rigidity protects proteins
from unfolding at high temperatures, proteolytic degradation and
2.2.7. α-Helix and β-sheets chemical denaturation. Experimental techniques such as nuclear mag­
Generally, α-helices and β-sheets are the most secondary structures in netic resonance (NMR) spectroscopy, hydrogen/deuterium (H/D) ex­
thermophilic proteins, while α-helices are less stable than β-sheet. Psy­ change, and neutron scattering show that the local flexibility and
chrophilic proteins which require flexibility at low temperature in order rigidity distribution of thermophilic proteins is similar to that of their
to function, tend to have a higher α-helical content in their structures mesophilic homologs, when the comparison is made at the respective
[104]. On the other hand, some thermophilic proteins like β-lactamases optimal temperatures [114]. Additionally, high-resolution ultrasonic
are richer in β-sheet structure, which is more stable than an α-helix spectroscopy, pressure perturbation calorimetry (PPC) and time-
[105]. The flexibility in α-helices in thermophilic proteins is lower and is resolved fluorescence spectroscopy are other experimental techniques
dependent on the increased quantity of buried methyl groups on the for studying protein flexibility [115]. Protein flexibility is described as
inner face of the helix that improves its packing [106]. Positively static or dynamic where static refers to the structural diversity and
charged residues like lysine near C-terminal and negatively charged number of fluctuating conformations in an equilibrium ensemble, while
residues like glutamic acid near N-terminal ends stabilize the inherent dynamic refers to how quickly these conformations interchange [55,56].
α-helix dipole. The antiparallel arrangement of adjacent helices provides Thermal flexibility refers to the amplitude of residue motions as a
5–7 kcal/mol of stabilization energy to those proteins in which the function of temperature and is linked to the activation energy, kinetic
arrangement exists [107]. Facchiano et al. [108] have shown that stability and cooperativity of unfolding [56]. It has been recently re­
β-branched residues like valine, isoleucine and threonine in α-helix ported that large kinetic barriers to unfolding may be a common strategy
decreased thermal stability. Hence, mutagenesis of these residues with used by hyperthermophiles to withstand extreme conditions [116]. In
the side chain not well accommodated in the α-helix may be a strategy to protein engineering, two strategies are employed to increase the ther­
increase protein stability. modynamic stability. One strategy stabilizes the folded state with the
addition of favorable interactions and/or the removal of unfavorable
2.2.8. Amino acids composition interactions. The other strategy destabilizes the unfolded state by
DNA sequencing studies demonstrate no significant differences be­ decreasing the chain entropy of the unfolded state [54].
tween amino acid compositions in the sequence of mesophilic and
thermophilic proteins. The closer analysis of the amino acid composition 4. Rigidifying of flexible structures for improvement of thermal
of proteins however indicates that polar residues are rare and charged stability
residues are frequent on the surface of thermophilic proteins [109].
Thermophile proteins have a significant fraction of the Lys, Arg, and Glu Many computational and experimental studies show that global
residues and a small fraction of the Ala, Asp, Asn, Gln, Ser, and His conformational rigidity enhances the thermostability of proteins
residues on their external surfaces [24,69,110]. In comparison, the compared to their mesostable homologs [113,114,117–120].
amino acid compositions of the interior of proteins from mesophilic and Conversely, structural analysis demonstrates that psychrophilic proteins
thermophilic organisms are the same. Generally, it seems that the resi­ are more flexible than their mesophilic homologs or orthologs. There­
dues of Ile, Val, Tyr, Trp, Arg, Glu, and Leu is characteristic to the fore, improving the rigidity at appropriate sites in a protein can pre­
thermophilic organisms [111]. sumably increase its thermal stability. As there are fewer contacts with
Overall, while the individual mechanisms contributing to the ther­ other residues in highly flexible regions of a protein, local perturbations
mal stability of proteins are fairly well understood, as exemplified of the conformation are observed within the interconnected network of
above, the way in which these individual small contributions lead to a non-covalent interactions. As a result, these flexible residues can trigger
global net impact on thermal stability is complex, protein-specific and protein unfolding due to thermal fluctuation [24]. Hence, rigidifying
generally poorly understood. In particular, further studies are needed to flexible regions is one of the effective strategies to enhance thermal
elucidate how the cooperativity or anti-cooperativity of the molecular stability. This strategy involves two steps. First, finding the flexible sites
interaction networks influence thermal stability and permit adaptation and then rigidifying them.
to extreme conditions.
4.1. Loops
3. Flexibility and rigidity in thermophilic proteins
Loops in proteins tend to be irregular structures that connect the
The concepts of flexibility and rigidity are pivotal for interpreting the more regular α-helix and β-sheet regions of folded polypeptide and are
relationship between function, structure and thermal stability. As usually located on the surfaces of proteins. Unlike α-helices and β-sheets
polypeptide chains, proteins possess numerous degrees of internal mo­ whose dynamics are dominated by small amplitude local fluctuations on
tions which carry thermal energy according to the equipartition theo­ a pico-to-nano-second timescale, loops frequently exhibit larger-
rem. Consequently, proteins exhibit very rich internal dynamics, amplitude dynamics on a micro- to milli-second timescale and which

646
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

are often functionally relevant [121]. Up to half of residues in enzyme hydrophobic, more negatively charged residues, and shorter lengths
active sites and up to one-third of protein sequences are found in loop were observed in fast loops. Static loops were characterized by more
structures. Allosteric regulation, signaling processes, and molecular hydrophobicity, increased atomic contact, and larger lengths, while
recognition often involve loop structures [122]. The exposed loops at the slow loops tended to lie between these two classifications [123].
surface of the protein interact with solvent and very often, ligands. The
most flexible parts of a folded protein tend to be associated with loops,
although they may be as rigid as α-helices or β-strands because of their 4.2. Prediction and identification of flexible loops
specific intramolecular interactions with neighboring amino acids, or
possibly because they comprise few residues. Flexible loops are frequently undetectable or at best ill-defined in the
In the evolutionary process, sequence changes occur more frequently electron density maps obtained by X-ray crystallography. As a result, the
in loop regions [123]. Fig. 3 displays the flexible loops and changes in experimental characterization of the structural dynamics of loops is
amino acid types in the structure of the CsaA protein (characterized as a often a challenging task. NMR, single-molecule fluorescence spectros­
chaperone protein) in Thermus thermophilus a thermophile (PDB ID: copy and small-angle X-ray or neutron scattering (SAXS/SANS) are a few
1GD7, in red) [124] versus the near isostructural tRNA-binding protein experimental techniques that provide applicable structural and
in E. coli as a mesophile (PDB ID: 3ERS, in blue) [125]. As shown in dynamical information about these flexible loops [121,127,128]. In
Fig. 3, because of the distinct amino acid sequences between these particular, NMR spectroscopy enables accessing loop dynamics at
proteins, different intermolecular interactions are observed particularly atomic resolution on a wide range of timescales including pico- to nano-
in the loop regions. In CsaA, as the thermophilic protein with a distinct second (sub-τc, τc: rotational correlation time of protein) and micro- to
amino acid composition, the distance between loops is shorter, which milli-second timescales [129]. More recently, loop motions occurring at
leads to a more compact structure due to more possible non-covalent nano- to micro-second timescales (supra-τc) are becoming increasingly
interactions. accessible to NMR methods [130–132]. However, these methods are not
Brüschweiler and co-workers [126] have performed NMR-validated applicable to all proteins, especially when protein size and solubility
MD simulations of 38 different proteins and classified their loops ac­ compromise the sensitivity of experiments. Therefore, computational
cording to the timescale of their dynamics as fast (correlation time < 10 techniques are required to complement structural studies. The integra­
ns), slow (correlation time 10–500 ns) and static (static over the course tion of multiple experimental and computational techniques has proved
of the MD trajectory, >500 ns). Three residue types: Met (neutral), Asp useful in accessing protein motions over a broad range of time and
(negatively charged), and Lys (positively charged) showed a clear pro­ length scales that are not accessible to any individual technique, even in
pensity for fast loops. The charged residue of Arg had a preference for the context of highly flexible, multi-domain or intrinsically disordered
both slow and static loops, while Glu had a preference for static loops. proteins [133–138]. The computational methods used for studying
Moreover, polar residues of Tyr, Thr, and His, which are involved in protein dynamics are classified as knowledge-based, physics-based, and
forming and breaking of hydrogen bonds, were found in slow loops. The hybrid approaches. For example, Monte Carlo and molecular dynamics
residues of Cys, Tyr, Ile, Phe, Lue and hydrophobic amino acids with (MD) simulation methods may be used to study the thermodynamic and
aromatic sidechains appeared in static loops. On the other hand, less kinetic properties of loops. Enhanced sampling MD methods are proper
tools to research loop dynamics [139]. In addition, Machine/Deep-

Fig. 3. Structural comparison of the CsaA protein in Thermus thermophilus as a thermophile (PDB ID: 1GD7, in red) versus the structurally near identical tRNA-binding
protein in Escherichia coli as a mesophile (PDB ID: 3ERS, in blue). Hydrophobic, polar, positive and negative charged residues in their loops of these proteins are
shown in light gray, green, red, and blue, respectively. The shorter distances in the loops of thermophile protein display non-covalent interactions and lead to more
rigidity. The loop distances and composition of amino acids of each loop are displayed with the squares. (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)

647
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

Learning (ML/DL) approaches could help structural bioinformatics in 4.3. Rigidifying flexible loops
further investigation [140,141]. The atomic B-factor (B-value) is a
parameter derived from X-ray data and indicates the average atomic Kumwenda et al. [146] showed that the surfaces and flexible surface
displacement of the atom in the crystal structure of a protein. Flexible loops have a significant role in the stability of proteins. Protein thermal
and thermo-unstable residues have higher B-factors (Fig. 2B) [18]. In the stability is enhanced when the penetration of water molecules is avoided
review article published by Sun et al. [54], the limitation of using the B- by increasing the density of solvent-protein hydrogen bonds. Therefore,
factor approach to study flexibility is discussed. The protein flexibility in the number of charged and polar residues is greater than apolar residues
solution is different from that in a crystal structure generally obtained in the flexible surface loops of thermophilic proteins. Hyperthermophilic
around 110 K, especially in the flexible loops [142]. MD simulation is a proteins show large hydrophilic patches in contact with water. It seems
powerful technique for studying protein kinetics and thermodynamics. that protein solubility at higher temperatures relates to these patches.
In addition, the flexibility of proteins at a specific temperature can be For hyperthermophilic proteins like other proteins, flexibility increases
investigated by these tools [143]. The overall flexibility of the protein is above ambient temperature. In the case of mesophiles, this will lead to
measured via the average root mean square fluctuation (RMSF) values denaturation, whereas, with hyperthermophiles, it leads to optimal ac­
during the MD trajectory. RMSF measures the average amplitude of each tivity before eventual denaturation. The flexibility is correlated with
atom's motions relative to a mean reference position. The higher value of higher exposure to solvent and arises by forming extended hydrogen-
RMSF represents higher flexibility [24]. Higher RMSF values, like the bond clusters [147,148]. Structural sequence alignments indicate that
higher B-values, indicate flexible residues usually due to their having exposed loop regions have been deleted in thermophilic sequences
fewer contacts with other residues. Hence, rigidifying these residues by rather than their mesophilic homologs. Therefore, it seems that loop
increasing the number of contacts or removing the residues completely, deletion is an evolutionary strategy for thermal stability gain [149].
can enhance their thermal stability [144,145]. As depicted in Fig. 4, the Furthermore, the H-bond forming residues particularly Arg and Tyr are
substitution of His489 by Asp (H489D) in the luciferase enzyme intro­ more frequently found on the surface of thermophilic proteins. Inspired
duced a new salt bridge between the C-terminal and N-terminal do­ by these nature-based strategies, loop size optimization and loop
mains, leading to protein rigidity and improved thermal stability. It is anchoring via substituting residues with Arg, Tyr, and Lys have been key
worth mentioning that some of these flexible loops are critical compo­ approaches in thermostability engineering [24,25,40,70,150–152].
nents of protein conformation and are involved in many biological The average size of the most abundant turn and short loop structures
functions, including their folding after biosynthesis. Therefore, amino in natural proteins are 3–10 residues, and there are predominantly rich
acid mutation in these loops often alter molecular function, if not render in proline, and glycine [153,154]. Pro and Gly play significant roles
the polypeptide insoluble. during chain compaction early in folding. Gly decreases the activation
energy of conformational alterations during protein folding while cis
prolyl bonds in Pro increase the barrier height. Therefore, loop

Fig. 4. Structural change of luciferase enzyme due to the

H489D mutation. Luciferase in native and H489D
mutated forms after 50 ns molecular dynamics simula­
tions are shown in purple and green, respectively. These
two structures were fitted based on the backbone. The
dashed line depicts the changes in the distance between
residues Arg112 and Asp489. The radius of gyration (Rg)
values for native and H489D mutated forms are 26.4 and
25.0 Å, respectively. (For interpretation of the references
to colour in this figure legend, the reader is referred to the
web version of this article.)

648
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

formation is accelerated by Gly and is significantly slower around trans for selecting the extraordinary mutant. For this purpose, flow cytometry,
prolyl bonds in Pro [153]. Hence, loop shortening by insertion of proline cell-surface display systems, and robotics have been developed,
residue or deletion of Gly is an efficient method of rigidifying protein although at a high cost. Reetz [178], Packer and Liu [179], Wang et al.
structure [38,155–157]. [180], and Bunzel et al. [181] describe the challenging issues and future
An alternative thermal stability engineering strategy for particular perspectives of the directed evolution approach.
protein families is protein-protein contact optimization or the formation
of higher-order oligomeric structures. Some studies show that psychro­ 5.2. Ancestral sequence reconstruction and consensus sequences
philic proteins have lower oligomerization forms than mesophilic and
thermophilic proteins [3,158]. Protein-protein contacts occur often Ancestral sequence reconstruction (ASR) and consensus design
during loop-loop interactions [3]. It is however notable that there is methods are two evolution-based methods widely used for stability en­
usually a trade-off between protein stability and flexibility/activity, so gineering. Both of them start with the multiple-sequence alignment of
that the optimization of thermal stability through protein rigidification relevant homolog sequences [57,182]. ASR uses statistical models for
may reduce protein activity. Consequently, the number of successful amino acid substitution by inferring hypothetical ancestral protein se­
examples of a protein stabilized by rigidifying one of more flexible loops quences based on the evolutionary relationships of related proteins,
is still limited because determining the best candidates for mutagenesis while the sequences derived by alignment of homologous thermally
which do not impair protein function is often a formidable challenge in stable proteins are utilized for the consensus sequence approach [183].
this strategy. ASR is presented as a worthwhile approach because ancestral proteins
often display desirable thermostabilities or unique characteristics. The
5. Genetic modifications for engineering thermally stable general workflow of ASR involves several steps which are data gathering
proteins and alignment of homologs, assessment of phylogeny and ancestral se­
quences, and finally, characterization of evolutionary intermediates by
Most enzymes need to be modified and optimized before being uti­ experimental methods. Ancestral sequences are reconstructed based on
lized in therapeutics and in industries as catalysts. However, enzymes the phylogenetic relationship between modern homologous sequences
derived from natural environments have evolved to function in their and by using a statistical model of amino acid substitution and calcu­
native conditions and cannot always meet the demands of the artificial lating sequences at internal nodes of the phylogenetic tree. The number
production environments, including high substrate specificity, catalytic and diversity of the input sequences, and the quality of the sequence
activity and environmental compatibility. Protein engineering aims to alignment will affect the phylogenetic tree derived from the evolu­
overcome these natural limitations by improving various functional tionary model used, and will influence the outcome of ASR. Hence,
aspects of proteins and industrial enzymes. Thermal stability is the main protein engineers should be aware of how to select these sequences
issue. Therefore, a specific subdivision in protein engineering is ther­ when performing ASR [184–186]. Generally, the ASR method is a
mally stable protein engineering. Up to now, several strategies have probabilistic method that finds functionally important main sub­
been used for engineering thermally stable proteins, including random stitutions in protein families by comparing the structures and functions
mutagenesis [159,160], rational design experiments [161], loop- of existing proteins and ancestral intermediates and thus provides clues
structure modification [162], directed evolution [163–165], ancestral for the rational design of new variants [167,187]. To date, several ASR
sequence reconstruction [166], consensus sequence alignment [167], studies have resulted in production of thermally stable prokaryotic
structure-assisted protein tailoring and computational modeling ap­ proteins as well as eukaryotic protein variants with improved perfor­
proaches [168–170]. These strategies can open new horizons in pro­ mance and good catalytic efficiency, making them particularly valuable
ducing more thermally stable proteins and biocatalysts from their for biotechnological applications [188–193]. These days, ASR and
mesophile counterparts. consensus design are increasingly popular due to the availability of
bioinformatics tools and the decreasing cost of gene synthesis. In addi­
5.1. Directed evolution tion, these strategies provide ways to investigate the relationship be­
tween sequence and function from the evolutionary standpoint.
To obtain new thermally stable catalysts for biotechnological appli­
cations, most studies aim to understand the factors influencing the sta­ 5.3. Semi-rational and rational design
bility of proteins upon mutation as well as predicting the change in
stability thereby induced [171]. Over the past two decades, directed In the library of mutants generated during directed evolution, the
evolution has been a powerful technique for enzyme engineering, as it workload of screening is very large on account of the large number of
does not need prior structural, functional, or mechanistic information mutants produced. Therefore, to achieve the required improvements in
about the target protein [172]. Following such a strategy, the natural stability and catalytic efficiency, researchers turned their attention to
evolutionary process (Darwinian evolution) is imitated in vitro under rational and semi-rational design methods, which are based on a holistic
controlled, targeted selective pressure [16]. Early methods primarily understanding of the structural features and functional characteristics of
relied on random mutagenesis and this requires effective screening for proteins [194,195]. Rational design creates a novel protein by making
the desired activity [173]. Arnold [174] launched the idea of directed specific alterations predicted with the help of modern computation. This
evolution of enzymes and developed efficient strategies for engineering approach relies upon our understanding of protein structure, function
new enzymes by multiple generations of random mutagenesis and and dynamics so that a rational design approach is likely to focus on
recombination combined with screening for improved variants. In regions that are better understood [196]. Also, despite advances in X-ray
addition, she proposed remodeling the natural enzyme by replacing crystallography and the large number of tertiary structures that have
amino acid residues [175]. Liu et al. [13] summarized the techniques been determined, the total number of well-defined enzyme structures is
used in directed evolution strategy. Random mutagenesis is a powerful still far from adequate to facilitate subsequent development of rational
technology for generating proteins, enzymes, or entire metabolic path­ design. In recent years, this deficiency has been partly compensated by
ways and even genomes with the desired features [176]. developing molecular simulation technology which can feed into a
Oligonucleotide-directed mutagenesis, error-prone polymerase chain rational design strategy. When there is no experimental information
reaction (epPCR), and DNA shuffling are the strategies for random about a three-dimensional structure, homologous structures can be used
mutagenesis [177]. Although directed evolution with randomized as a template to predict the structure of the unknown enzyme [197].
mutagenesis is a suitable procedure, it requires both an extensive mutant Rational and semi-rational design methods can significantly reduce the
library from which to start and high-throughput screening is essential size of the mutation library and hence improve the efficiency of protein

649
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

engineering. Recently, there has been significant interest in high speed such that thermally stable proteins can be designed by optimizing
and intelligent tools for determining the detailed structural and func­ packing and idealizing backbone conformation [223]. Most proteins
tional roles of proteins, which significantly improves the definition of created de novo have a high content of regular secondary structure, a
“weak spots” or specific areas of enzymes such as flexible loops which high contact order, and minimal loops all of which favor greater ther­
may lead to significant loss of activity at higher temperatures [198]. modynamic stability [224].
Several successful auxiliary methods have been proposed to predict
and improve potential heat-resistant protein mutants [54,199]. The 6. Conclusion
majority of strategies used in the rational design of thermal stability are
divided into three. The computationally based (estimating the effect of Defining the structural basis of protein thermal stability represents a
varying an amino acid on the overall energy of the protein structure), the key topic in structural biology and in protein chemistry. Improving
sequence-based (detecting the dominant features of thermophilic pro­ protein thermal stability is an important goal in the biopharmaceutical
teins), and the structure-based (observing the various stabilizing in­ and biotechnological sector. The balance between protein function and
teractions with the protein structure) strategies. Which, singly or stability in a crowded cell medium and the prevention of aggregation are
together allow a specific site to be accurately chosen for mutagenesis driving forces in protein evolution. The mutation of a small sequence of
[196]. Thus, in a rational design strategy, residues are selected as the a few residues, or even a single residue, in a non-functional loop can
targets using the above methods, then site-specific mutagenesis of the affect thermal stability. Thermophilic proteins and their mesophilic
key residues is performed and the functional properties of the target homologs show high similarity in their sequences and their folded
mutants screened [200]. Some of the computational tools developed to structures are generally largely superimposable. DNA sequencing has
support protein engineering design are given here. For example, the demonstrated that thermophilic proteins have a higher content of hy­
Rosetta 3 [194,201–203] is the prevalent toolkit for prediction of drophobic residues like Val, Ile, and Pro, which enhance their rigidity.
structural perturbations in mutations [204]. In addition, SDM [205], Charged residues capable of forming salt bridges are abundant in ther­
FoldX [206], HotSpot Wizard [207], YASARA [208], PoPMuSiC [209], mophiles. However, the type of amino acid preference differs from one
AUTO-MUTE [210], FireProtDB [211], CUPSAT [212], I-Mutant [213], thermophilic protein family to next. The loop structures in thermophilic
Prethermut [214], HoTMuSiC [215], Caver analyst [216] and MoDEL proteins have a profound effect upon the overall stability of the protein
[217] are other resources for protein engineering. Also, constraint and its interactions with other subunits, its binding of cofactors, and its
network analysis (CNA) computes proteins' local and global flexibility/ activity. During evolution, loop mutagenesis provided novel enzyme
rigidity properties by the graph theory-based rigidity analysis of thermal activities. Hence, loop alteration is one possible strategy for modifying a
unfolding simulation [218]. FIRST (Floppy inclusion and rigid substructure protein to satisfy our need. The folded protein can be stabilized by
topography) can also identify the rigid and flexible regions of the protein adding favorable interactions, removing unfavorable interactions, or
conformation [82]. more generally decreasing the entropy of the unfolded state. However,
our limited knowledge of the relationship between thermal stability and
5.4. De novo design of thermally stable proteins the intramolecular interactions that stabilize proteins may lead to un­
successful mutagenesis. Local perturbation in flexible loops is observed
During the evolution process, the secondary and tertiary structures of because of the fewer contacts between the amino acids. These loops,
proteins are more conserved than the primary sequence [219]. A protein therefore, can trigger protein unfolding. Computational and experi­
is a “biomolecular machine”, whose function depends on the specifically mental studies reveal that the overall rigidity of a protein conformation
folded 3D structure and its inherent dynamic properties. 30 years ago, increases thermostability. Consequently, rigidifying flexible loops at
the rational or de novo design of proteins were considered impossible appropriate sites of proteins should enhance thermal stability. Loop
because of the lack of a knowledge of the fundamental details of protein deletion or shortening by insertion of proline or deletion of Gly, coupled
folding. However, in recent years, it has become possible to design with loop anchoring by increasing the number of salt bridges, hydrogen
proteins with predictable function. Our knowledge of the fundamental bonds, or hydrophobicity, are good strategies to enhance the overall
principles of protein folding, thanks to the ever-expanding database of rigidity of a protein conformation leading to better compactness and a
crystallographic structures and the physicochemical information pro­ decreased number of internal cavities. Selection, design and engineering
vided by the site-directed mutagenesis of native proteins has allowed the of a functional biocatalyst by identifying the sites responsible for ther­
design of proteins de novo. The concept of de novo protein design and mal stability requires new tools and strategies. In recent years, the
production is the creation of smaller and much more stable proteins that methodological advances of enzyme engineering, especially in rational
are resistant to chemical, enzymic and thermal insult [220]. Based on or de novo enzyme design provide powerful and impressive approaches
Woolfson's opinion, protein engineering is different from protein design to manipulate thermostable proteins. The advances in structural bioin­
in that the native protein structures are engineered or redesigned for formatics spurred on by the availability of the enormous 3D structural
novel functions including those that have not yet been identified in the information have led to the publication of a variety protein design al­
biosphere. Ideally therefore, these designed proteins should have the gorithms. For example, the recently introduced AlphaFold method [225]
same (or similar) structure, dynamics and stability as a native protein enables highly accurate protein structure prediction and promises rev­
yet catalyze a distinct reaction [221]. In addition, however, de novo olutionary changes in protein engineering including engineering for
protein design has enabled the formation of completely new protein improved thermal stability. This review article has highlighted novel
structures based on the physical principles of intramolecular and inter­ strategies for the identification of flexible loops and the intramolecular
molecular interaction [222]. forces contributing to protein rigidity. It is worthwhile reiterating that
David Baker's experiments demonstrate that most soluble de novo protein activity is related in no small measure to protein flexibility,
designed proteins expressed in E. coli remain folded at 95 ◦ C. Therefore, whereas protein integrity is related to protein rigidity. However, chal­
it seems that de novo proteins are a little more stable than natural lenges remain in a full understanding of the concepts of flexibility and
counterparts. In his opinion, the need for a specific protein function rigidity necessary for the successful design of thermally stable proteins
outweighed the need for greater thermostability during the evolutionary but advanced methods in computational biology like Rosetta and SDM
process. Therefore, selection for protein function has come at the cost of are already making an impact on the de novo design of proteins to solve
thermal stability. Alternatively, understanding the forces that stabilize this issue and produce new biocatalysts.
proteins must be at least qualitatively precise in order to allow the
efficient engineering stable proteins. The methods used in RosettaDesign
are based on iterative sidechain optimization and backbone relaxation

650
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

Declaration of competing interest [28] M. Matsumura, W.J. Becktel, M. Levitt, B.W. Matthews, Stabilization of phage T4
lysozyme by engineered disulfide bonds, Proc. Natl. Acad. Sci. 86 (17) (1989)
6562.
The authors report no conflict of interest. [29] J. Zhao, J. Chen, H. Wang, Y. Guo, K. Li, J. Liu, Enhanced thermostability of D-
psicose 3-epimerase from clostridium bolteae through rational design and
Acknowledgments engineering of new disulfide bridges, Int. J. Mol. Sci. 22 (18) (2021).
[30] G. Li X. Fang F. Su Y. Chen L. Xu Y. Yan M. Kelly Robert , Enhancing the
thermostability of Rhizomucor miehei lipase with a limited screening library by
Any support of the Research Council of the University of Tehran is rational-design point mutations and disulfide bonds, Appl. Environ. Microbiol. 84
gratefully acknowledged. (2) e02129-17.
[31] M. Imani, S. Hosseinkhani, S. Ahmadian, M. Nazari, Design and introduction of a
disulfide bridge in firefly luciferase: increase of thermostability and decrease of
References pH sensitivity, Photochem. Photobiol. Sci. 9 (8) (2010) 1167–1177.
[32] A.A. Dombkowski, K.Z. Sultana, D.B. Craig, Protein disulfide engineering, FEBS
[1] P.L. Privalov, Stability of proteins small globular proteins, in: C.B. Anfinsen, J. Lett. 588 (2) (2014) 206–212.
T. Edsall, F.M. Richards (Eds.), Advances in Protein Chemistry, Academic Press, [33] M.Z. Kamal, S. Ahmad, T.R. Molugu, A. Vijayalakshmi, M.V. Deshmukh,
1979, pp. 167–241. R. Sankaranarayanan, N.M. Rao, In vitro evolved non-aggregating and
[2] S. Yadav, F. Ahmad, A new method for the determination of stability parameters thermostable lipase: structural and thermodynamic investigation, J. Mol. Biol.
of proteins from their heat-induced denaturation curves, Anal. Biochem. 283 (2) 413 (3) (2011) 726–741.
(2000) 207–213. [34] K. Imada, M. Sato, N. Tanaka, Y. Katsube, Y. Matsuura, T. Oshima, Three-
[3] F. Pucci, M. Rooman, Physical and molecular bases of protein thermal stability dimensional structure of a highly thermostable enzyme, 3-isopropylmalate
and cold adaptation, Curr. Opin. Struct. Biol. 42 (2017) 117–128. dehydrogenase of thermus thermophilus at 2.2 Å resolution, J. Mol. Biol. 222 (3)
[4] G. Feller, C. Gerday, Psychrophilic enzymes: hot topics in cold adaptation, Nat. (1991) 725–738.
Rev. Microbiol. 1 (3) (2003) 200–208. [35] S.J. Crennell, G.O. Hreggvidsson, E. Nordberg Karlsson, The structure of
[5] J.G. Zeikus, C. Vieille, A. Savchenko, Thermozymes: biotechnology and rhodothermus marinus Cel12A, a highly thermostable family 12 endoglucanase,
structure–function relationships, Extremophiles 2 (3) (1998) 179–183. at 1.8Å resolution, J. Mol. Biol. 320 (4) (2002) 883–897.
[6] A. Goncearenco, B.G. Ma, I.N. Berezovsky, Molecular mechanisms of adaptation [36] J. Qiu, H. Yang, Y. Shao, L. Li, S. Sun, L. Wang, Y. Tan, Z. Xin, Enhancing the
emerging from the physics and evolution of nucleic acids and proteins, Nucleic activity and thermal stability of a phthalate-degrading hydrolase by random
Acids Res. 42 (5) (2014) 2879–2892. mutagenesis, Ecotoxicol. Environ. Saf. 209 (2021), 111795.
[7] K.B. Zeldovich, I.N. Berezovsky, E.I. Shakhnovich, Protein and DNA sequence [37] M. Lehmann, D. Kostrewa, M. Wyss, R. Brugger, A. D'Arcy, L. Pasamontes, A.P.G.
determinants of thermophilic adaptation, PLoS Comput. Biol. 3 (1) (2007), e5. M. van Loon, From DNA sequence to improved functionality: using protein
[8] S. Kumar, A.K. Dangi, P. Shukla, D. Baishya, S.K. Khare, Thermozymes: adaptive sequence comparisons to rapidly design a thermostable consensus phytase,
strategies and tools for their biotechnological applications, Bioresour. Technol. Protein Eng. Des. Sel. 13 (1) (2000) 49–57.
278 (2019) 372–382. [38] A. Kheirollahi, K. Khajeh, A. Golestani, Rigidifying flexible sites: an approach to
[9] A. Broom, Z. Jacobi, K. Trainor, E.M. Meiering, Computational tools help improve improve stability of chondroitinase ABC I, Int. J. Biol. Macromol. 97 (2017)
protein stability but with a solubility tradeoff, J. Biol. Chem. 292 (35) (2017) 270–278.
14349–14361. [39] A. Barzegar, A.A. Moosavi-Movahedi, J.Z. Pedersen, M. Miroliaei, Comparative
[10] Z. Xu, Y.K. Cen, S.P. Zou, Y.P. Xue, Y.G. Zheng, Recent advances in the thermostability of mesophilic and thermophilic alcohol dehydrogenases: stability-
improvement of enzyme thermostability by structure modification, Crit. Rev. determining roles of proline residues and loop conformations, Enzym. Microb.
Biotechnol. 40 (1) (2020) 83–98. Technol. 45 (2) (2009) 73–79.
[11] R. Sinha, P. Shukla, Current trends in protein engineering: updates and progress, [40] H. Yu, Y. Yan, C. Zhang, P.A. Dalby, Two strategies to engineer flexible loops for
Curr. Protein Pept. Sci. 20 (5) (2019) 398–407. improved enzyme thermostability, Sci. Rep. 7 (1) (2017) 41212.
[12] R.K. Singh, J.K. Lee, C. Selvaraj, R. Singh, J. Li, S.Y. Kim, V.C. Kalia, Protein [41] N.K. Tafreshi, S. Hosseinkhani, M. Sadeghizadeh, M. Sadeghi, B. Ranjbar,
engineering approaches in the post-genomic era, Curr. Protein Pept. Sci. 19 (1) H. Naderi-Manesh, The influence of insertion of a critical residue (Arg356) in
(2018) 5–15. structure and bioluminescence spectra of firefly luciferase *, J. Biol. Chem. 282
[13] Q. Liu, G. Xun, Y. Feng, The state-of-the-art strategies of protein engineering for (12) (2007) 8641–8647.
enzyme stabilization, Biotechnol. Adv. 37 (4) (2019) 530–537. [42] J. Damnjanović, H. Nakano, Y. Iwasaki, Deletion of a dynamic surface loop
[14] F.H. Arnold, Protein engineering for unusual environments, Curr. Opin. improves stability and changes kinetic behavior of phosphatidylinositol-
Biotechnol. 4 (4) (1993) 450–455. synthesizing streptomyces phospholipase D, Biotechnol. Bioeng. 111 (4) (2014)
[15] O. Buß, J. Rudat, K. Ochsenreither, FoldX as protein engineering tool: better than 674–682.
random based approaches?, Computational and structural, Biotechnol. J. 16 [43] H. Narayanan, F. Dingfelder, A. Butté, N. Lorenzen, M. Sokolov, P. Arosio,
(2018) 25–33. Machine learning for biologics: opportunities for protein engineering,
[16] S. Lutz, S.M. Iamurri, Protein engineering: past, present, and future, Methods Mol. developability, and formulation, Trends Pharmacol. Sci. 42 (3) (2021) 151–165.
Biol. 1685 (2018) 1–12. [44] J. Greenhalgh, A. Saraogee, P.A. Romero, Data-driven protein engineering,
[17] B. Kuhlman, P. Bradley, Advances in protein structure prediction and design, Nat. Protein Eng. (2021) 133–151.
Rev. Mol. Cell Biol. 20 (11) (2019) 681–697. [45] V. Pongsupasa, P. Anuwan, S. Maenpuen, T. Wongnate, Rational-design
[18] H. Yu, H. Huang, Engineering proteins for thermostability through rigidifying engineering to improve enzyme thermostability, in: F. Magnani, C. Marabelli,
flexible sites, Biotechnol. Adv. 32 (2) (2014) 308–315. F. Paradisi (Eds.), Enzyme Engineering: Methods and Protocols, Springer, US,
[19] M.-L. Liao, G.N. Somero, Y.-W. Dong, Comparing mutagenesis and simulations as New York, NY, 2022, pp. 159–178.
tools for identifying functionally important sequence changes for protein thermal [46] R. Liu, L. Liang, M.P. Lacerda, E.F. Freed, C.A. Eckert, Chapter 10 - advances in
adaptation, Proc. Natl. Acad. Sci. 116 (2) (2019) 679. protein engineering and its application in synthetic biology, in: V. Singh (Ed.),
[20] Y. Li, Conformational sampling in template-free protein loop structure modeling: New Frontiers and Applications of Synthetic Biology, Academic Press, 2022,
an overviwe, Comput. Struct. Biotechnol. J. 5 (6) (2013), e201302003. pp. 147–158.
[21] J. Reeb, B. Rost, Secondary structure prediction, in: S. Ranganathan, M. Gribskov, [47] S. Roda, A. Robles-Martín, R. Xiang, M. Kazemi, V. Guallar, Structural-based
K. Nakai, C. Schönbach (Eds.), Encyclopedia of Bioinformatics and Computational modeling in protein engineering. A must do, J. Phys. Chem. B 125 (24) (2021)
Biology, Academic Press, Oxford, 2019, pp. 488–496. 6491–6500.
[22] B.M. Nestl, B. Hauer, Engineering of flexible loops in enzymes, ACS Catal. 4 (9) [48] F. Ahmad, On the estimation of stability parameters from heat-induced
(2014) 3201–3211. conformational transition curves of proteins, J. Iran. Chem. Soc. 1 (2) (2004)
[23] N. Mollania, K. Khajeh, B. Ranjbar, S. Hosseinkhani, Enhancement of a bacterial 99–105.
laccase thermostability through directed mutagenesis of a surface loop, Enzym. [49] A.A. Saboury, F. Karbassi, Thermodynamic studies on the interaction of calcium
Microb. Technol. 49 (5) (2011) 446–452. ions with alpha-amylase, Thermochim. Acta 362 (1–2) (2000) 121–129.
[24] M. Rahban, N. Salehi, A.A. Saboury, S. Hosseinkhani, M.H. Karimi-Jafari, [50] A.A. Saboury, A.A. Moosavi-Movahedi, Derivation of the thermodynamic
R. Firouzi, N. Rezaei-Ghaleh, A.A. Moosavi-Movahedi, Histidine substitution in parameters involved in the elucidation of protein thermal profiles, Biochem.
the most flexible fragments of firefly luciferase modifies its thermal stability, Educ. 23 (3) (1995) 164–167.
Arch. Biochem. Biophys. 629 (2017) 8–18. [51] A.A. Saboury, M.S. Atri, M.H. Sanati, A.A. Moosavi-Movahedi, K. Haghbeen,
[25] M. Mortazavi, S. Hosseinkhani, Design of thermostable luciferases through Effects of calcium binding on the structure and stability of human growth
arginine saturation in solvent-exposed loops, Protein Eng. Des. Sel. 24 (12) hormone, Int. J. Biol. Macromol. 36 (5) (2005) 305–309.
(2011) 893–903. [52] F. Karbassi, K. Haghbeen, A.A. Saboury, M. Rezaei-Tavirani, B. Ranjbar,
[26] C.-W. Lee, H.-J. Wang, J.-K. Hwang, C.-P. Tseng, Protein thermal stability Calorimetric, spectrophotometric and circular dichroism studies on the impact of
enhancement by designing salt bridges: a combined computational and sodium dodecyl sulphate on the mushroom tyrosinase structure, Biologia-
experimental study, PLOS ONE 9 (11) (2014), e112751. Bratislava 59 (3) (2004) 319–326.
[27] J. Chen, H. Yu, C. Liu, J. Liu, Z. Shen, Improving stability of nitrile hydratase by [53] C. Vieille, G.J. Zeikus, Hyperthermophilic enzymes: sources, uses, and molecular
bridging the salt-bridges in specific thermal-sensitive regions, J. Biotechnol. 164 mechanisms for thermostability, Microbiol Mol Biol Rev 65 (1) (2001) 1–43.
(2) (2013) 354–362. [54] Z. Sun, Q. Liu, G. Qu, Y. Feng, M.T. Reetz, Utility of B-factors in protein science:
interpreting rigidity, flexibility, and internal motion and engineering
thermostability, Chem. Rev. 119 (3) (2019) 1626–1665.

651
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

[55] A.G. Quezada, N. Cabrera, Á. Piñeiro, A.J. Díaz-Salazar, S. Díaz-Mazariegos, [84] L. Boucher, S. Somani, C. Negron, W. Ma, S. Jacobs, W. Chan, T. Malia,
S. Romero-Romero, R. Pérez-Montfort, M. Costas, A strategy based on thermal G. Obmolova, A. Teplyakov, G.L. Gilliland, J. Luo, Surface salt bridges contribute
flexibility to design triosephosphate isomerase proteins with increased or to the extreme thermal stability of an FN3-like domain from a thermophilic
decreased kinetic stability, Biochem. Biophys. Res. Commun. 503 (4) (2018) bacterium, Proteins: Struct., Funct., Bioinf. 90 (1) (2021) 270–281, [Link]
3017–3022. org/10.1002/prot.26218, n/a(n/a).
[56] A.G. Quezada, A.J. Díaz-Salazar, N. Cabrera, R. Pérez-Montfort, Á. Piñeiro, [85] R. Das, M. Gerstein, The stability of thermophilic proteins: a study based on
M. Costas, Interplay between protein thermal flexibility and kinetic stability, comprehensive genome comparison, Funct. Integr. Genomics 1 (1) (2000) 76–88.
Structure 25 (1) (2017) 167–179. [86] H.-X. Zhou, X. Pang, Electrostatic interactions in protein structure, folding,
[57] R. Furukawa, W. Toma, K. Yamazaki, S. Akanuma, Ancestral sequence binding, and condensation, Chem. Rev. 118 (4) (2018) 1691–1741.
reconstruction produces thermally stable enzymes with mesophilic enzyme-like [87] E. Alsop, M. Silver, D.R. Livesay, Optimized electrostatic surfaces parallel
catalytic properties, Sci. Rep. 10 (1) (2020) 15493. increased thermostability: a structural bioinformatic analysis, Protein Eng. Des.
[58] M. Miotto, P.P. Olimpieri, L. Di Rienzo, F. Ambrosetti, P. Corsi, R. Lepore, G. Sel. 16 (12) (2003) 871–874.
G. Tartaglia, E. Milanetti, Insights on protein thermal stability: a graph [88] A. Irimia, F.M.D. Vellieux, D. Madern, A. Karshikoff, G. Tibbelin, R. Ladenstein,
representation of molecular interactions, Bioinformatics 35 (15) (2019) T. Lien, N.-K. Birkeland, G. Zaccaı¨, The 2.9Å resolution crystal structure of malate
2569–2577. dehydrogenase from Archaeoglobus fulgidus: mechanisms of oligomerisation and
[59] I.N. Berezovsky, E.I. Shakhnovich, Physics and evolution of thermophilic thermal stabilisation, J. Mol. Biol. 335 (1) (2004) 343–356.
adaptation, Proc. Natl. Acad. Sci. U. S. A. 102 (36) (2005) 12742–12747. [89] D.E. Anderson, W.J. Becktel, F.W. Dahlquist, pH-induced denaturation of
[60] T. Modi, P. Campitelli, I.C. Kazan, S.B. Ozkan, Protein folding stability and proteins: a single salt bridge contributes 3–5 kcal/mol to the free energy of
binding interactions through the lens of evolution: a dynamical perspective, Curr. folding of T4 lysozyme, Biochemistry 29 (9) (1990) 2403–2408.
Opin. Struct. Biol. 66 (2021) 207–215. [90] S. Kumar, C.J. Tsai, B. Ma, R. Nussinov, Contribution of salt bridges toward
[61] J. Ni, V. Sakanyan, D. Charlier, N. Glansdorff, G.D. Van Duyne, Structure of the protein thermostability, J. Biomol. Struct. Dyn. 17 (Suppl. 1) (2000) 79–85.
arginine repressor from Bacillus stearothermophilus, Nat. Struct. Biol. 6 (5) [91] C. Nick Pace, J.M. Scholtz, G.R. Grimsley, Forces stabilizing proteins, FEBS Lett.
(1999) 427–432. 588 (14) (2014) 2177–2184.
[62] G.D. Van Duyne, G. Ghosh, W.K. Maas, P.B. Sigler, Structure of the [92] V.G. Eijsink, G. Vriend, [Link] der Zee, [Link] den Burg, G. Venema, Increasing
oligomerization and L-arginine binding domain of the arginine repressor of the thermostability of the neutral proteinase of Bacillus stearothermophilus by
Escherichia coli, J. Mol. Biol. 256 (2) (1996) 377–391. improvement of internal hydrogen-bonding, Biochem. J. 285 (Pt 2) (1992)
[63] K. Kumar, K. Patel, D.C. Agrawal, J.M. Khire, Insights into the unfolding pathway 625–628 (Pt 2).
and identification of thermally sensitive regions of phytase from aspergillus Niger [93] E.A. Toth, C. Worby, J.E. Dixon, E.R. Goedken, S. Marqusee, T.O. Yeates, The
by molecular dynamics simulations, J. Mol. Model. 21 (6) (2015) 163. crystal structure of adenylosuccinate lyase from pyrobaculum aerophilum reveals
[64] N.C. Benson, V. Daggett, Dynameomics: large-scale assessment of native protein an intracellular protein with three disulfide bonds, J. Mol. Biol. 301 (2) (2000)
flexibility, Protein Sci. 17 (12) (2008) 2038–2050. 433–450.
[65] Y. Ueda, H. Taketomi, N. Gō, Studies on protein folding, unfolding, and [94] R. Ladenstein, B. Ren, Protein disulfides and protein disulfide oxidoreductases in
fluctuations by computer simulation. II. A. Three-dimensional lattice model of hyperthermophiles, FEBS J. 273 (18) (2006) 4170–4185.
lysozyme, Biopolymers 17 (6) (1978) 1531–1548. [95] B.S. Ibrahim, V. Pattabhi, Role of weak interactions in thermal stability of
[66] M.G. Pikkemaat, A.B.M. Linssen, H.J.C. Berendsen, D.B. Janssen, Molecular proteins, Biochem. Biophys. Res. Commun. 325 (3) (2004) 1082–1089.
dynamics simulations as a tool for improving protein stability, Protein Eng. Des. [96] M.L. Waters, Aromatic interactions in peptides: impact on structure and function,
Sel. 15 (3) (2002) 185–192. Biopolymers 76 (5) (2004) 435–445.
[67] T. Lazaridis, I. Lee, M. Karplus, Dynamics and unfolding pathways of a [97] N. Kannan, S. Vishveshwara, Aromatic clusters: a determinant of thermal stability
hyperthermophilic and a mesophilic rubredoxin, Protein Sci. 6 (12) (1997) of thermophilic proteins, Protein Eng. Des. Sel. 13 (11) (2000) 753–761.
2589–2605. [98] S. Hait, S. Mallik, S. Basu, S. Kundu, Finding the generalized molecular principles
[68] L. Julió Plana, A.D. Nadra, D.A. Estrin, F.J. Luque, L. Capece, Thermal stability of of protein thermal stability, Proteins: Struct., Funct., Bioinf. 88 (6) (2020)
globins: implications of flexibility and heme coordination studied by molecular 788–808.
dynamics simulations, J. Chem. Inf. Model. 59 (1) (2019) 441–452. [99] A.G. Cochran, N.J. Skelton, M.A. Starovasnik, Tryptophan zippers: stable,
[69] B.G. Ma, A. Goncearenco, I.N. Berezovsky, Thermophilic adaptation of protein monomeric beta -hairpins, Proc. Natl. Acad. Sci. U. S. A. 98 (10) (2001)
complexes inferred from proteomic homology modeling, Structure 18 (7) (2010) 5578–5583.
819–828. [100] R.S. Prajapati, M. Das, S. Sreeramulu, M. Sirajuddin, S. Srinivasan,
[70] A. Ruggiero, G. Smaldone, L. Esposito, N. Balasco, L. Vitagliano, Loop size V. Krishnamurthy, R. Ranjani, C. Ramakrishnan, R. Varadarajan, Thermodynamic
optimization induces a strong thermal stabilization of the thioredoxin fold, FEBS effects of proline introduction on protein stability, Proteins 66 (2) (2007)
J. 286 (9) (2019) 1752–1764. 480–491.
[71] W.F. Li, X.X. Zhou, P. Lu, Structural features of thermozymes, Biotechnol. Adv. 23 [101] K. Watanabe, T. Masuda, H. Ohashi, H. Mihara, Y. Suzuki, Multiple proline
(4) (2005) 271–281. substitutions cumulatively thermostabilize bacillus cereus ATCC7064 Oligo-1,6-
[72] S. Akanuma, M. Yamaguchi, A. Yamagishi, Comprehensive mutagenesis to glucosidase, Eur. J. Biochem. 226 (2) (1994) 277–283.
identify amino acid residues contributing to the difference in thermostability [102] D.N. Ivankov, N.S. Bogatyreva, M.Y. Lobanov, O.V. Galzitskaya, Coupling
between two originally thermostable ancestral proteins, PLOS ONE 16 (10) between properties of the protein shape and the rate of protein folding, PloS one 4
(2021), e0258821. (8) (2009) e6476-e6476.
[73] D. Maes, J.P. Zeelen, N. Thanki, N. Beaucamp, M. Alvarez, M.H.D. Thi, [103] P. Mahanta, A. Bhardwaj, K. Kumar, V.S. Reddy, S. Ramakumar, Structural
J. Backmann, J.A. Martial, L. Wyns, R. Jaenicke, R.K. Wierenga, The crystal insights into N-terminal to C-terminal interactions and implications for
structure of triosephosphate isomerase (TIM) from Thermotoga maritima: a thermostability of a (β/α)8-triosephosphate isomerase barrel enzyme, FEBS J. 282
comparative thermostability structural analysis of ten different TIM structures, (18) (2015) 3543–3555.
Proteins: Struct., Funct., Bioinf. 37 (3) (1999) 441–453. [104] K.K. Singh, L. Ray, Extremozymes: biocatalysts from extremophilic
[74] I.N. Berezovsky, The diversity of physical forces and mechanisms in microorganisms and their relevance in current biotechnology, in: S.K.N, B.
intermolecular interactions, Phys. Biol. 8 (3) (2011), 035002. B. Mishra, S. Mohapatra, D. Samantaray (Eds.), Environmental and Agricultural
[75] S. Timr, D. Madern, F. Sterpone, Chapter six - protein thermal stability, in: Microbiology, 2021, pp. 293–311.
B. Strodel, B. Barz (Eds.), Progress in Molecular Biology and Translational [105] S. Vijayakumar, S. Vishveshwara, G. Ravishanker, D.L. Beveridge, Differential
Science, Academic Press, 2020, pp. 239–272. stability of beta-sheets and alpha-helices in beta-lactamase: a high temperature
[76] M. Bueno, L.A. Campos, J. Estrada, J. Sancho, Energetics of aliphatic deletions in molecular dynamics study of unfolding intermediates, Biophys. J. 65 (6) (1993)
protein cores, Protein Sci. 15 (8) (2006) 1858–1872. 2304–2312.
[77] C.N. Pace, Contribution of the hydrophobic effect to globular protein stability, [106] A. Paiardini, R. Sali, F. Bossa, S. Pascarella, "Hot cores" in proteins: comparative
J. Mol. Biol. 226 (1) (1992) 29–35. analysis of the apolar contact area in structures from hyper/thermophilic and
[78] V. Renugopalakrishnan, R. Garduño-Juárez, G. Narasimhan, C.S. Verma, X. Wei, mesophilic organisms, BMC Struct. Biol. 8 (1) (2008) 14.
P. Li, Rational design of thermally stable proteins: relevance to [107] R.P. Sheridan, R.M. Levy, F.R. Salemme, Alpha-helix dipole model and
bionanotechnology, J. Nanosci. Nanotechnol. 5 (11) (2005) 1759–1767. electrostatic stabilization of 4-alpha-helical proteins, Proc. Natl. Acad. Sci. U. S.
[79] C. Nguyen, J.T. Young, G.G. Slade, R.J. Oliveira, M.E. McCully, A dynamic A. 79 (15) (1982) 4545–4549.
hydrophobic Core and surface salt bridges thermostabilize a designed three-helix [108] A.M. Facchiano, G. Colonna, R. Ragone, Helix stabilizing factors and stabilization
bundle, Biophys. J. 116 (4) (2019) 621–632. of thermophilic proteins: an X-ray based study, Protein Eng. 11 (9) (1998)
[80] C.L. Dias, T. Ala-Nissila, J. Wong-ekkabut, I. Vattulainen, M. Grant, M. Karttunen, 753–760.
The hydrophobic effect and its role in cold denaturation, Cryobiology 60 (1) [109] T.B. Mamonova, A.V. Glyakina, O.V. Galzitskaya, M.G. Kurnikova, Stability and
(2010) 91–99. rigidity/flexibility—two sides of the same coin? Biochim. Biophys. Acta, Proteins
[81] J.R. Grigera, A.N. McCarthy, The behavior of the hydrophobic effect under Proteomics 1834 (5) (2013) 854–866.
pressure and protein denaturation, Biophys. J. 98 (8) (2010) 1626–1631. [110] H.K. Liang, C.M. Huang, M.T. Ko, J.K. Hwang, Amino acid coupling patterns in
[82] H.P. Modarres, M.R. Mofrad, A. Sanati-Nezhad, Protein thermostability thermophilic proteins, Proteins 59 (1) (2005) 58–63.
engineering, RSC Adv. 6 (116) (2016) 115252–115270. [111] X.X. Zhou, Y.B. Wang, Y.J. Pan, W.F. Li, Differences in amino acids composition
[83] K. Manjunath, K. Sekar, Molecular dynamics perspective on the protein thermal and coupling patterns between mesophilic and thermophilic proteins, Amino
stability: a case study using SAICAR synthetase, J. Chem. Inf. Model. 53 (9) Acids 34 (1) (2008) 25–33.
(2013) 2448–2461.

652
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

[112] R. Lewandowski Józef, E. Halse Meghan, M. Blackledge, L. Emsley, Direct [144] M.T. Reetz, P. Soni, L. Fernández, Knowledge-guided laboratory evolution of
observation of hierarchical protein dynamics, Science 348 (6234) (2015) protein thermolability, Biotechnol. Bioeng. 102 (6) (2009) 1712–1717.
578–581. [145] M.T. Reetz, J.D. Carballeira, A. Vogel, Iterative saturation mutagenesis on the
[113] A. Karshikoff, L. Nilsson, R. Ladenstein, Rigidity versus flexibility: the dilemma of basis of B factors as a strategy for increasing protein thermostability, Angew.
understanding protein thermal stability, FEBS J. 282 (20) (2015) 3899–3917. Chem. Int. Ed. 45 (46) (2006) 7745–7751.
[114] S. Radestock, H. Gohlke, Protein rigidity and thermophilic adaptation, Proteins [146] B. Kumwenda, D. Litthauer, O.T. Bishop, O. Reva, Analysis of protein
79 (4) (2011) 1089–1108. thermostability enhancing factors in industrially important thermus bacteria
[115] T.J. Kamerzell, C. Russell Middaugh, The complex inter-relationships between species, Evol. Bioinformatics Online 9 (2013) 327–342.
protein flexibility and stability, J. Pharm. Sci. 97 (9) (2008) 3494–3517. [147] F. Sterpone, C. Bertonati, G. Briganti, S. Melchionna, Key role of proximal water
[116] M.L. Bagarolo, M. Porcelli, E. Martino, G. Feller, G. Cacciapuoti, Multiple in regulating thermostable proteins, J. Phys. Chem. B 113 (1) (2009) 131–137.
disulfide bridges modulate conformational stability and flexibility in [148] Z. Mou, Y. Ding, X. Wang, Y. Cai, Comparison of protein-water interactions in
hyperthermophilic archaeal purine nucleoside phosphorylase, Biochim. Biophys. psychrophilic, mesophilic, and thermophilic fe-SOD, Protein Pept. Lett. 21 (6)
Acta, Proteins Proteomics 1854 (10, Part A) (2015) 1458–1465. (2014) 578–583.
[117] D. Suplatov, V. Voevodin, V. Švedas, Robust enzyme design: bioinformatic tools [149] M.J. Thompson, D. Eisenberg, Transproteomic evidence of a loop-deletion
for improved protein stability, Biotechnol. J. 10 (3) (2015) 344–355. mechanism for enhancing protein thermostability, J. Mol. Biol. 290 (2) (1999)
[118] A.J. Rader, R.M. Yennamalli, A.K. Harter, T.Z. Sen, A rigid network of long-range 595–604.
contacts increases thermostability in a mutant endoglucanase, J. Biomol. Struct. [150] S. Kumar, C.-J. Tsai, R. Nussinov, Factors enhancing protein thermostability,
Dyn. 30 (6) (2012) 628–637. Protein Eng. Des. Sel. 13 (3) (2000) 179–191.
[119] R. Jaenicke, Do ultrastable proteins from hyperthermophiles have high or low [151] F. Kargar, M. Mortazavi, M. Torkzadeh-Mahani, S. Lotfi, S. Shakeri, Evaluation of
conformational rigidity? Proc. Natl. Acad. Sci. 97 (7) (2000) 2962. luciferase thermal stability by arginine saturation in the flexible loops, Curr.
[120] S. Radestock, H. Gohlke, Exploiting the link between protein rigidity and Proteomics 17 (1) (2020) 30–39.
thermostability for data-driven protein engineering, Eng. Life Sci. 8 (5) (2008) [152] H. Wang, L. Zhang, Y. Wang, J. Li, G. Du, Z. Kang, Engineering a thermostable
507–522. chondroitinase for production of specifically distributed low-molecular-weight
[121] A.G. Palmer, Enzyme dynamics from NMR spectroscopy, Acc. Chem. Res. 48 (2) chondroitin sulfate, Biotechnol. J. 16 (5) (2021), 2000321.
(2015) 457–465. [153] F. Krieger, A. Möglich, T. Kiefhaber, Effect of proline and glycine residues on
[122] P.D. Sun, C.E. Foster, J.C. Boyington, Overview of protein structural and dynamics and barriers of loop formation in polypeptide chains, J. Am. Chem. Soc.
functional folds, Chapter, Curr. Protoc. Protein Sci. 17 (1) (2004), 1711-17.1. 127 (10) (2005) 3346–3352.
[123] E. Papaleo, G. Saladino, M. Lambrughi, K. Lindorff-Larsen, F.L. Gervasio, [154] J. Dhar, P. Chakrabarti, Defining the loop structures in proteins based on
R. Nussinov, The role of protein loops and linkers in conformational dynamics and composite β-turn mimics, Protein Eng. Des. Sel. 28 (6) (2015) 153–161.
allostery, Chem. Rev. 116 (11) (2016) 6391–6423. [155] G. Yang, H. Yao, M. Mozzicafreddo, P. Ballarini, S. Pucciarelli, C. Miceli, Rational
[124] S. Kawaguchi, J. Müller, D. Linde, S. Kuramitsu, T. Shibata, Y. Inoue, D. engineering of a cold-adapted α-amylase from the Antarctic ciliate euplotes
G. Vassylyev, S. Yokoyama, The crystal structure of the ttCsaA protein: an export- focardii for simultaneous improvement of thermostability and catalytic activity,
related chaperone from thermus thermophilus, EMBO J. 20 (3) (2001) 562–569. Appl. Environ. Microbiol. 83 (13) (2017).
[125] M.A. Swairjo, A.J. Morales, C.C. Wang, A.R. Ortiz, P. Schimmel, Crystal structure [156] F. Haghdoust, M. Molakarimi, M. Mirshahi, R.H. Sajedi, Engineering aequorin to
of trbp111: a structure-specific tRNA-binding protein, EMBO J. 19 (23) (2000) improve thermostability through rigidifying flexible sites, J. Mol. Struct. 1240
6287–6298. (2021), 130575.
[126] Y. Gu, D.-W. Li, R. Brüschweiler, Decoding the mobility and time scales of protein [157] H. Yu, Y. Zhao, C. Guo, Y. Gan, H. Huang, The role of proline substitutions within
loops, J. Chem. Theory Comput. 11 (3) (2015) 1308–1314. flexible regions on thermostability of luciferase, Biochim. Biophys. Acta 1854 (1)
[127] E. Lerner, A. Barth, J. Hendrix, B. Ambrose, V. Birkedal, S.C. Blanchard, (2015) 65–72.
R. Börner, H.S. Chung, T. Cordes, T.D. Craggs, FRET-based dynamic structural [158] L.M. Zanphorlin, P.O. de Giuseppe, R.V. Honorato, C.C.C. Tonoli, J. Fattori,
biology: challenges, perspectives and an appeal for open-science practices, elife E. Crespim, P.S.L. de Oliveira, R. Ruller, M.T. Murakami, Oligomerization as a
10 (2021), e60416. strategy for cold adaptation: structure and dynamics of the GH1 β-glucosidase
[128] H.D.T. Mertens, D.I. Svergun, Combining NMR and small angle X-ray scattering from exiguobacterium antarcticum B7, Sci. Rep. 6 (1) (2016) 23776.
for the study of biomolecular structure and dynamics, Arch. Biochem. Biophys. [159] J. Xu, Y. Cen, W. Singh, J. Fan, L. Wu, X. Lin, J. Zhou, M. Huang, M.T. Reetz,
628 (2017) 33–41. Q. Wu, Stereodivergent protein engineering of a lipase to access all possible
[129] A.G. Palmer, H. Koss, Chapter six - chemical exchange, in: A.J. Wand (Ed.), stereoisomers of chiral esters with two stereocenters, J. Am. Chem. Soc. 141 (19)
Methods in Enzymology, Academic Press, 2019, pp. 177–236. (2019) 7934–7945.
[130] X. Xiang, A.L. Hansen, L. Yu, G. Jameson, L. Bruschweiler-Li, C. Yuan, [160] M.T. Reetz, S. Prasad, J.D. Carballeira, Y. Gumulya, M. Bocola, Iterative
R. Brüschweiler, Observation of sub-microsecond protein methyl-side chain saturation mutagenesis accelerates laboratory evolution of enzyme
dynamics by nanoparticle-assisted NMR spin relaxation, J. Am. Chem. Soc. 143 stereoselectivity: rigorous comparison with traditional methods, J. Am. Chem.
(34) (2021) 13593–13604. Soc. 132 (26) (2010) 9144–9152.
[131] D. Ban, Evolving role of conformational dynamics in understanding fundamental [161] T.D. Ngo, C. Oh, P. Mizar, M. Baek, K.-S. Park, L. Nguyen, H. Byeon, S. Yoon,
biomolecular behavior, mechanistic enzymology: bridging structure and function, Y. Ryu, B.H. Ryu, Structural basis for the enantioselectivity of esterase Est-Y29
Am. Chem. Soc. (2020) 57–81. toward (S)-ketoprofen, ACS Catal. 9 (1) (2018) 755–767.
[132] S. Pratihar, T.M. Sabo, D. Ban, R.B. Fenwick, S. Becker, X. Salvatella, [162] H. Tang, K. Shi, C. Shi, H. Aihara, J. Zhang, G. Du, Enhancing subtilisin
C. Griesinger, D. Lee, Kinetics of the antibody recognition site in the third IgG- thermostability through a modified normalized B-factor analysis and loop-
binding domain of protein G, Angew. Chem. Int. Ed. 55 (33) (2016) 9567–9570. grafting strategy, J. Biol. Chem. 294 (48) (2019) 18398–18407.
[133] F. Munari, N. Rezaei-Ghaleh, S. Xiang, W. Fischle, M. Zweckstetter, Structural [163] F.H. Arnold, A.A. Volkov, Directed evolution of biocatalysts, Curr. Opin. Chem.
plasticity in human heterochromatin protein 1β, PLOS ONE 8 (4) (2013), e60887. Biol. 3 (1) (1999) 54–59.
[134] S. Naudi-Fabra, M. Tengo, M.R. Jensen, M. Blackledge, S. Milles, Quantitative [164] K.E. Jaeger, T. Eggert, Enantioselective biocatalysis optimized by directed
description of intrinsically disordered proteins using single-molecule FRET, NMR, evolution, Curr. Opin. Biotechnol. 15 (4) (2004) 305–313.
and SAXS, J. Am. Chem. Soc. 143 (48) (2021) 20109–20121. [165] C.G. Acevedo-Rocha, F. Hollmann, J. Sanchis, Z. Sun, A pioneering career in
[135] N. Rezaei-Ghaleh, G. Parigi, M. Zweckstetter, Reorientational dynamics of catalysis: Manfred T. Reetz, ACS Catal. 10 (24) (2020) 15123–15139.
amyloid-β from NMR spin relaxation and molecular simulation, J. Phys. Chem. [166] C.D. Okafor, M.C. Pathak, C.E. Fagan, N.C. Bauer, M.F. Cole, E.A. Gaucher, E.
Lett. 10 (12) (2019) 3369–3375. A. Ortlund, Structural and dynamics comparison of thermostability in ancient,
[136] N. Rezaei-Ghaleh, G. Parigi, A. Soranno, A. Holla, S. Becker, B. Schuler, modern, and consensus elongation factor Tus, Structure 26 (1) (2018) 118–129,
C. Luchinat, M. Zweckstetter, Local and global dynamics in intrinsically e3.
disordered synuclein, Angew. Chem. Int. Ed. 57 (46) (2018) 15262–15266. [167] M.A. Spence, J.A. Kaczmarski, J.W. Saunders, C.J. Jackson, Ancestral sequence
[137] N. Rezaei-Ghaleh, F. Klama, F. Munari, M. Zweckstetter, Predicting the rotational reconstruction for protein engineers, Curr. Opin. Struct. Biol. 69 (2021) 131–141.
tumbling of dynamic multidomain proteins and supramolecular complexes, [168] B.E. Correia, Y.-E.A. Ban, D.J. Friend, K. Ellingson, H. Xu, E. Boni, T. Bradley-
Angew. Chem. Int. Ed. 52 (43) (2013) 11410–11414. Hewitt, J.F. Bruhn-Johannsen, L. Stamatatos, R.K. Strong, W.R. Schief,
[138] N. Rezaei-Ghaleh, F. Klama, F. Munari, M. Zweckstetter, HYCUD: a computational Computational protein design using flexible backbone remodeling and
tool for prediction of effective rotational correlation time in flexible proteins, resurfacing: case studies in structure-based antigen design, J. Mol. Biol. 405 (1)
Bioinformatics 31 (8) (2015) 1319–1321. (2011) 284–297.
[139] N. Salehi, M. Amininasab, R. Firouzi, M.H. Karimi-Jafari, Efficient construction of [169] J. Damborsky, J. Brezovsky, Computational tools for designing and engineering
a diverse conformational library for amyloid-β as an intrinsically disordered enzymes, Curr. Opin. Chem. Biol. 19 (2014) 8–16.
protein, J. Mol. Graph. Model. 88 (2019) 183–193. [170] F. Sesterhenn, C. Yang, J. Bonet, T. Cramer Johannes, X. Wen, Y. Wang, I. Chiang
[140] A. Barozet, P. Chacón, J. Cortés, Current approaches to flexible loop modeling, Chi, A. Abriata Luciano, I. Kucharska, G. Castoro, S. Vollers Sabrina, M. Galloux,
Curr. Res. Struct. Biol. 3 (2021) 187–191. E. Dheilly, S. Rosset, P. Corthésy, S. Georgeon, M. Villard, C.-A. Richard,
[141] K. Yoshida, S. Kawai, M. Fujitani, S. Koikeda, R. Kato, T. Ema, Enhancement of D. Descamps, T. Delgado, E. Oricchio, M.-A. Rameix-Welti, V. Más, S. Ervin, J.-
protein thermostability by three consecutive mutations using loop-walking F. Eléouët, S. Riffault, T. Bates John, J.-P. Julien, Y. Li, T. Jardetzky, T. Krey,
method and machine learning, Sci. Rep. 11 (1) (2021) 11883. E. Correia Bruno, De novo protein design enables the precise induction of RSV-
[142] P. Eastman, M. Pellegrini, S. Doniach, Protein flexibility in solution and in neutralizing antibodies, Science 368 (6492) (2020), eaay5051.
crystals, J. Chem. Phys. 110 (20) (1999) 10141–10152. [171] J.H. Lebbink, T. Kaper, P. Bron, J. van der Oost, W.M. de Vos, Improving low-
[143] S.A. Adcock, J.A. McCammon, Molecular dynamics: survey of methods for temperature catalysis in the hyperthermostable pyrococcus furiosus β-glucosidase
simulating the activity of proteins, Chem. Rev. 106 (5) (2006) 1589–1615. CelB by directed evolution, Biochemistry 39 (13) (2000) 3656–3665.

653
M. Rahban et al. International Journal of Biological Macromolecules 214 (2022) 642–654

[172] R.E. Cobb, N. Sun, H. Zhao, Directed evolution as a powerful synthetic biology [201] C.A. Rohl, C.E. Strauss, K.M. Misura, D. Baker, Protein structure prediction using
tool, Methods 60 (1) (2013) 81–90. Rosetta, Methods Enzymol. 383 (2004) 66–93.
[173] B. van den Burg, V.G.H. Eijsink, Selection of mutations for increased protein [202] F. DiMaio, A. Leaver-Fay, P. Bradley, D. Baker, I. André, Modeling symmetric
stability, Curr. Opin. Biotechnol. 13 (4) (2002) 333–337. macromolecular structures in Rosetta3, PLoS One 6 (6) (2011), e20450.
[174] K. Chen, F.H. Arnold, Enzyme engineering for nonaqueous solvents: random [203] E.H. Kellogg, A. Leaver-Fay, D. Baker, Role of conformational sampling in
mutagenesis to enhance activity of subtilisin E in polar organic media, Bio/ computing mutation-induced changes in protein structure and stability, Proteins
Technology 9 (11) (1991) 1073–1077. 79 (3) (2011) 830–838.
[175] F.H. Arnold, Directed evolution: creating biocatalysts for the future, Chem. Eng. [204] A. Leaver-Fay, M. Tyka, S.M. Lewis, O.F. Lange, J. Thompson, R. Jacak,
Sci. 51 (23) (1996) 5091–5102. K. Kaufman, P.D. Renfrew, C.A. Smith, W. Sheffler, I.W. Davis, S. Cooper,
[176] S. Chusacultanachai, Y. Yuthavong, Random mutagenesis strategies for A. Treuille, D.J. Mandell, F. Richter, Y.E. Ban, S.J. Fleishman, J.E. Corn, D.E. Kim,
construction of large and diverse clone libraries of mutated DNA fragments, in: S. Lyskov, M. Berrondo, S. Mentzer, Z. Popović, J.J. Havranek, J. Karanicolas,
Parasite Genomics Protocols, Springer, 2004, pp. 319–333. R. Das, J. Meiler, T. Kortemme, J.J. Gray, B. Kuhlman, D. Baker, P. Bradley,
[177] C. Neylon, Chemical and biochemical strategies for the randomization of protein ROSETTA3: an object-oriented software suite for the simulation and design of
encoding DNA sequences: library construction methods for directed evolution, macromolecules, Methods Enzymol. 487 (2011) 545–574.
Nucleic Acids Res. 32 (4) (2004) 1448–1459. [205] A.P. Pandurangan, B. Ochoa-Montaño, D.B. Ascher, T.L. Blundell, SDM: a server
[178] M.T. Reetz, Biocatalysis in organic chemistry and biotechnology: past, present, for predicting effects of mutations on protein stability, Nucleic Acids Res. 45 (W1)
and future, J. Am. Chem. Soc. 135 (34) (2013) 12480–12496. (2017) W229–w235.
[179] M.S. Packer, D.R. Liu, Methods for the directed evolution of proteins, Nat. Rev. [206] J. Schymkowitz, J. Borg, F. Stricher, R. Nys, F. Rousseau, L. Serrano, The FoldX
Genet. 16 (7) (2015) 379–394. web server: an online force field, Nucleic Acids Res. 33 (Web Server issue) (2005).
[180] Y. Wang, P. Xue, M. Cao, T. Yu, S.T. Lane, H. Zhao, Directed evolution: W382-8.
methodologies and applications, Chem. Rev. 121 (20) (2021) 12384–12444. [207] L. Sumbalova, J. Stourac, T. Martinek, D. Bednar, J. Damborsky, HotSpot wizard
[181] H.A. Bunzel, J.L.R. Anderson, A.J. Mulholland, Designing better enzymes: insights 3.0: web server for automated design of mutations and smart libraries based on
from directed evolution, Curr. Opin. Struct. Biol. 67 (2021) 212–218. sequence input information, Nucleic Acids Res. 46 (W1) (2018) W356–w362.
[182] F. Scossa, A.R. Fernie, Ancestral sequence reconstruction - an underused approach [208] H. Land, M.S. Humble, YASARA: a tool to obtain structural guidance in
to understand the evolution of gene function in plants? Comput. Struct. biocatalytic investigations, Methods Mol. Biol. 1685 (2018) 43–67.
Biotechnol. J. 19 (2021) 1579–1594. [209] Y. Dehouck, A. Grosfils, B. Folch, D. Gilis, P. Bogaerts, M. Rooman, Fast and
[183] M. Wilding, N. Hong, M. Spence, A.M. Buckle, C.J. Jackson, Protein engineering: accurate predictions of protein stability changes upon mutations using statistical
the potential of remote mutations, Biochem. Soc. Trans. 47 (2) (2019) 701–711. potentials and neural networks: PoPMuSiC-2.0, Bioinformatics 25 (19) (2009)
[184] N.M. Hendrikse, G. Charpentier, E. Nordling, P.O. Syrén, Ancestral diterpene 2537–2543.
cyclases show increased thermostability and substrate acceptance, FEBS J. 285 [210] M. Masso, I.I. Vaisman, Accurate prediction of stability changes in protein
(24) (2018) 4660–4673. mutants by combining machine learning with structure based computational
[185] A. Goldenzweig, M. Goldsmith, S.E. Hill, O. Gertman, P. Laurino, Y. Ashani, mutagenesis, Bioinformatics 24 (18) (2008) 2002–2009.
O. Dym, T. Unger, S. Albeck, J. Prilusky, Automated structure-and sequence- [211] J. Stourac, J. Dubrava, M. Musil, J. Horackova, J. Damborsky, S. Mazurenko,
based design of proteins for high bacterial expression and stability, Mol. Cell 63 D. Bednar, FireProtDB: database of manually curated protein stability data,
(2) (2016) 337–346. Nucleic Acids Res. 49 (D1) (2021) D319–d324.
[186] M. Musil, R.T. Khan, A. Beier, J. Stourac, H. Konegger, J. Damborsky, D. Bednar, [212] V. Parthiban, M.M. Gromiha, D. Schomburg, CUPSAT: prediction of protein
FireProtASR: a web server for fully automated ancestral sequence reconstruction, stability upon point mutations, Nucleic Acids Res. 34 (Web Server issue) (2006).
Brief. Bioinform. 22 (4) (2021), bbaa337. W239-42.
[187] G.K. Hochberg, J.W. Thornton, Reconstructing ancient proteins to understand the [213] E. Capriotti, P. Fariselli, R. Calabrese, R. Casadio, Predicting protein stability
causes of structure and function, Annu. Rev. Biophys. 46 (2017) 247–269. changes from sequences using support vector machines, Bioinformatics 21 (Suppl.
[188] T. Sanavia, G. Birolo, L. Montanucci, P. Turina, E. Capriotti, P. Fariselli, 2) (2005) ii54-8.
Limitations and challenges in protein stability prediction upon genome variations: [214] J. Tian, N. Wu, X. Chu, Y. Fan, Predicting changes in protein thermostability
towards future applications in precision medicine, Comput. Struct. Biotechnol. J. brought about by single- or multi-site mutations, BMC Bioinformatics 11 (2010)
18 (2020) 1968–1979. 370.
[189] S. Nakano, K. Kozuka, Y. Minamino, H. Karasuda, F. Hasebe, S. Ito, Ancestral L- [215] F. Pucci, R. Bourgeas, M. Rooman, Predicting protein thermal stability changes
amino acid oxidases for deracemization and stereoinversion of amino acids, upon point mutations using statistical potentials: introducing HoTMuSiC, Sci.
Commun. Chem. 3 (1) (2020) 1–11. Rep. 6 (1) (2016) 23257.
[190] A.G. Bart, K.L. Harris, E.M. Gillam, E.E. Scott, Structure of an ancestral [216] A. Jurcik, D. Bednar, J. Byska, S.M. Marques, K. Furmanova, L. Daniel,
mammalian family 1B1 cytochrome P450 with increased thermostability: P. Kokkonen, J. Brezovsky, O. Strnad, J. Stourac, A. Pavelka, M. Manak,
ancestral mammalian P450 1B1 structures, J. Biol. Chem. 295 (17) (2020) J. Damborsky, B. Kozlikova, CAVER analyst 2.0: analysis and visualization of
5640–5653. channels and tunnels in protein structures and molecular dynamics trajectories,
[191] Y. Gumulya, J.-M. Baek, S.-J. Wun, R.E. Thomson, K.L. Harris, D.J. Hunter, J. Bioinformatics 34 (20) (2018) 3586–3588.
B. Behrendorff, J. Kulig, S. Zheng, X. Wu, Engineering highly functional [217] T. Meyer, M. D'Abramo, A. Hospital, M. Rueda, C. Ferrer-Costa, A. Pérez,
thermostable proteins using ancestral sequence reconstruction, Nat. Catal. 1 (11) O. Carrillo, J. Camps, C. Fenollosa, D. Repchevsky, J.L. Gelpí, M. Orozco, MoDEL
(2018) 878–888. (Molecular dynamics extended Library): a database of atomistic molecular
[192] P. Babkova, E. Sebestova, J. Brezovsky, R. Chaloupkova, J. Damborsky, Ancestral dynamics trajectories, Structure 18 (11) (2010) 1399–1409.
haloalkane dehalogenases show robustness and unique substrate specificity, [218] D.M. Krüger, P.C. Rathi, C. Pfleger, H. Gohlke, CNA web server: rigidity theory-
ChemBioChem 18 (14) (2017) 1448–1456. based thermal unfolding simulations of proteins for linking structure, (thermo-)
[193] V.A. Risso, J.A. Gavira, J.M. Sanchez-Ruiz, Thermostable and promiscuous P stability, and function, Nucleic Acids Res. 41 (Web Server issue) (2013). W340-8.
recambrian proteins, Environ. Microbiol. 16 (6) (2014) 1485–1489. [219] Y. Chalopin, The physical origin of rate promoting vibrations in enzymes revealed
[194] E.C. Alley, G. Khimulya, S. Biswas, M. AlQuraishi, G.M. Church, Unified rational by structural rigidity, Sci. Rep. 10 (1) (2020) 17465.
protein engineering with sequence-based deep representation learning, Nat. [220] I.V. Korendovych, W.F. DeGrado, De novo protein design, a retrospective, Q. Rev.
Methods 16 (12) (2019) 1315–1322. Biophys. 53 (2020), e3.
[195] M.K. Tiwari, R. Singh, R.K. Singh, I.-W. Kim, J.-K. Lee, Computational approaches [221] D.N. Woolfson, A brief history of De novo protein design: minimal, rational, and
for rational design of proteins with novel functionalities, Comput. Struct. computational, J. Mol. Biol. 433 (20) (2021), 167160.
Biotechnol. J. 2 (3) (2012), e201204002. [222] X. Pan, T. Kortemme, Recent advances in de novo protein design: principles,
[196] S. Lutz, Beyond directed evolution—semi-rational protein engineering and methods, and applications, J. Biol. Chem. 296 (2021), 100558.
design, Curr. Opin. Biotechnol. 21 (6) (2010) 734–743. [223] D. Baker, What has de novo protein design taught us about protein folding and
[197] R.F. Alford, A. Leaver-Fay, J.R. Jeliazkov, M.J. O’Meara, F.P. DiMaio, H. Park, M. biophysics? Protein Sci. 28 (4) (2019) 678–683.
V. Shapovalov, P.D. Renfrew, V.K. Mulligan, K. Kappel, J.W. Labonte, M. [224] C. Yang, F. Sesterhenn, J. Bonet, E.A. van Aalen, L. Scheller, L.A. Abriata, J.
S. Pacella, R. Bonneau, P. Bradley, R.L. Dunbrack, R. Das, D. Baker, B. Kuhlman, T. Cramer, X. Wen, S. Rosset, S. Georgeon, T. Jardetzky, T. Krey, M. Fussenegger,
T. Kortemme, J.J. Gray, The Rosetta all-atom energy function for macromolecular M. Merkx, B.E. Correia, Bottom-up de novo design of functional proteins with
modeling and design, J. Chem. Theory Comput. 13 (6) (2017) 3031–3048. complex structural features, Nat. Chem. Biol. 17 (4) (2021) 492–500.
[198] I.V. Korendovych, Rational and semirational protein design, in: U.T. Bornscheuer, [225] J. Jumper, R. Evans, A. Pritzel, T. Green, M. Figurnov, O. Ronneberger,
M. Höhne (Eds.), Protein Engineering: Methods and Protocols, Springer, New K. Tunyasuvunakool, R. Bates, A. Žídek, A. Potapenko, A. Bridgland, C. Meyer, S.
York, New York, NY, 2018, pp. 15–23. A.A. Kohl, A.J. Ballard, A. Cowie, B. Romera-Paredes, S. Nikolov, R. Jain, J. Adler,
[199] Z. Xu, Y.-K. Cen, S.-P. Zou, Y.-P. Xue, Y.-G. Zheng, Recent advances in the T. Back, S. Petersen, D. Reiman, E. Clancy, M. Zielinski, M. Steinegger,
improvement of enzyme thermostability by structure modification, Crit. Rev. M. Pacholska, T. Berghammer, S. Bodenstein, D. Silver, O. Vinyals, A.W. Senior,
Biotechnol. 40 (1) (2020) 83–98. K. Kavukcuoglu, P. Kohli, D. Hassabis, Highly accurate protein structure
[200] T. Dinmukhamed, Z. Huang, Y. Liu, X. Lv, J. Li, G. Du, L. Liu, Current advances in prediction with AlphaFold, Nature 596 (7873) (2021) 583–589.
design and engineering strategies of industrial enzymes, Syst. Microbiol.
Biomanufacturing 1 (1) (2021) 15–23.

654