structure- and sequence-analysis inspired engineering of proteins for enhanced thermostability

COSTBI-1113; NO. OF PAGES 7

Structure- and sequence-analysis inspired engineering ofproteins for enhanced thermostabilityHein J Wijma, Robert J Floor and Dick B Janssen

Available online at www.sciencedirect.com

Protein engineering strategies for increasing stability can be

improved by replacing random mutagenesis and high-

throughput screening by approaches that include

bioinformatics and computational design. Mutations can be

focused on regions in the structure that are most flexible and

involved in the early steps of thermal unfolding. Sequence

analysis can often predict the position and nature of stabilizing

mutations, and may allow the reconstruction of thermostable

ancestral sequences. Various computational tools make it

possible to design stabilizing features, such as hydrophobic

clusters and surface charges. Different methods for designing

chimeric enzymes can also support the engineering of more

stable proteins without the need of high-throughput screening.

Addresses

Department of Biochemistry, Groningen Biomolecular Sciences and

Biotechnology Institute, University of Groningen, Nijenborgh 4, 9747 AG

Groningen, The Netherlands

Corresponding author: Janssen, Dick B ([email protected])

Current Opinion in Structural Biology 2013, 23:xx–yy

This review comes from a themed issue on Engineering and design

Edited by Florian Hollfelder and Stefan Lutz

0959-440X/$ – see front matter, Published by Elsevier Ltd.

http://dx.doi.org/10.1016/j.sbi.2013.04.008

IntroductionStability under practical conditions is one of the most

critical properties when biotechnological applications of

proteins are explored [1]. High thermostability is often

accompanied by good performance under unfavorable

conditions, such as the presence of cosolvents. Improved

thermal stability may also correlate with higher expres-

sion yields in heterologous hosts, improved long-term

survival under mild conditions, increased ability to

remain active in non-aqueous solvents, and a longer

half-life under harsh industrial process conditions [2].

For medicinal proteins, thermostability may correlate

with high serum survival times [3]. In protein engineer-

ing, many function-gaining mutations, such as modified

enzyme selectivity, diminish stability. As a result, ther-

mostable proteins can tolerate a larger number of

mutations than mesostable proteins [4], and give better

results when used as a starting point in protein engin-

eering [5,6]. For these reasons, and because thermostable

variants often are not available from the natural

Please cite this article in press as: Wijma HJ, et al.: Structure- and sequence-analysis inspired engine

10.1016/j.sbi.2013.04.008

www.sciencedirect.com

biodiversity, increasing thermostability is an important

goal of protein engineering.

This review focuses on recent developments in protein

engineering aimed at enhancing thermostability. A num-

ber of well-established techniques in protein thermosta-

bilization by mutagenesis have been developed,

providing important insight in the structural features that

govern thermostability. These principles and examples

have been reviewed elsewhere [4,7–12]. On the other

hand, directed evolution has emerged as a powerful

approach, especially in the absence of structural infor-

mation [8,11]. However, when high-throughput screening

is not possible, random approaches that are frequently

used in directed evolution may become inefficient. This

triggers a quest for more focused methods, for example,

by incorporating information provided by bioinformatics

and structural analysis. The first section of this review

highlights recent progress in identifying target positions

for mutagenesis and substitutions that enhance stability.

The next sections describe recent advances in protein

stabilization by employing sequence comparisons, com-

putational design, loop grafting and chimera building.

Locating critical sites for stabilityimprovementMutations that strongly enhance protein thermostability

often appear to cluster in a particular region of the protein,

whereas similar substitutions introduced elsewhere in the

protein may have a negligible effect on thermostability

[7]. The explanation for this phenomenon is that the

strongly stabilizing mutations are located at a spot where

the protein starts to unfold [7]. Mutations that improve

local structural stability also improve global stability of a

protein, especially when they are positioned in a region

that unfolds at a critical step in a kinetically controlled

unfolding pathway. However, the pathways of unfolding

are usually unknown, making a knowledge-driven

approach difficult. Yet, several methods have been

explored to predict which positions are most likely to

contribute to thermostability and should be targeted by

mutagenesis [13–17,18�].

An approach that employs crystallographic information to

identify target positions is the B-fit method [13], in which

residues with the highest crystallographic B-factors are

selected for mutagenesis. The rationale is that highly

flexible residues are more likely to unfold in an early stage,

and that substitutions at these positions have a higher

chance to stabilize the folded state. Several successes

ering of proteins for enhanced thermostability, Curr Opin Struct Biol (2013), http://dx.doi.org/

Current Opinion in Structural Biology 2013, 23:1–7



[email protected]


http://www.sciencedirect.com/science/journal/0959440X

2 Engineering and design


Table 1

Recent examples of stabilized proteins

Targeted protein TappM increase

(8C)

Part of protein targeted

by mutagenesis (%)

Origin of stabilizing mutations Reference

Suspected critical

region targeted

Esterase +9a 1 Consensus library [14]

Epoxide hydrolase +21a 3 NDT codon library [15]

a-Amino ester hydrolase +7a 3 Consensus design [16]

Heme peroxidase +4 3 Computational design [17]

Xylanase +2 4 Computational design [20]


Haloalkane dehalogenase +18 4 Random and site-directed [24]

b-Glucosidase +7 3 Site-directed mutagenesis [25]

No suspected critical

region targeted

Thioredoxin +32 NA Ancestral method [28]

Triosephosphate isomerase +8 8 Ancestral method [31]

b-Amylase +9 4 Ancestral method [32]

Acylphosphosphatase +9 5 Computational design [33]

GTPase +9 4 Computational design [33]

Methionine amino peptidase +18 2d Computational design [37��]


Cellobiohydrolase +9a 10 Computational and consensus [42]

Methyl parathion hydrolase +12 3 Computational design [41]

Sesquiterpenoid synthase +45 2 Computational design [45�]

Baeyer–Villiger monooxygenase +16b NA Chimeric enzyme [46�]

Aldo-keto reductase – NA Chimeric enzyme [49]

Polyol dehydrogenases +12b NA Chimeric enzyme [50]

Cellulase +8.5a,c NA SCHEMA [53,54��]

Cytochrome P450 +14.5a,c NA Chimeric enzyme [56��]

Green fluorescent protein +5 NA Circularization [59]

a On the basis of 50% residual activity after heat treatment, values of which can differ extremely from the T appM measured by other techniques [18�].

b Compared to the melting temperature of the thermolabile parent of the chimera.c Compared to the melting temperature of the thermostable parent of the chimera.d Only residues near void in the hydrophobic core were targeted.

have been reported (Table 1) [14–17]. A recent investi-

gation of an esterase that was improved by the B-fit method

showed that the observed increase in temperature at which

50% of the enzyme was inactivated by heating was in fact

due to improved refolding after heat treatment [18�].Whereas the mutant esterase can refold after heating

(95 8C), it unfolds at an even lower temperature than

the wild-type esterase, suggesting that the B-fit approach

can also yield substitutions that increase reversibility of

unfolding rather than enhancing thermostability.

Besides crystallographic B-factors, which may be influ-

enced by crystal contacts and solvent conditions, various

theoretical methods can be used to predict flexible

regions that may serve as targets for stabilization. A

recent example is provided by the stabilization of xyla-

nase, for which flexible residues were located either by

MD simulations at different temperatures or by molecu-

lar mechanics based flexibility predictions [19] (http://

www.flexweb.asu). Using the molecular mechanics

based flexibility analysis, computationally designed

point mutations were introduced at 8 positions in the

sequence of the 185 aa protein, which resulted in an

improvement of the T appM by 2 8C and a 15-fold longer

half-life at 50 8C [20]. Employing the results of MD

simulations and computational design, five different

mutations were incorporated in the same xylanase, which


10.1016/j.sbi.2013.04.008


gave a 4 8C higher apparent melting temperature and a

30-fold longer half-life [21].

Constraint Network Analysis (CNA) can provide insight

in the unfolding mechanisms of proteins, including the

identification of residues involved in critical unfolding

steps [22��,23]. In CNA, hydrogen bonds act as con-

straints of the structure and residues that are connected

by a continuous network of hydrogen bonds are con-

sidered to be folded. The strength of each H-bond is

calculated after which the H-bonds are gradually removed

to simulate an increasing temperature. The CNA simu-

lations indicated that the disappearance of particular H-

bonds signals the abrupt disintegration of the original

interaction cluster into much smaller clusters, which

corresponds to melting of the protein. The H-bonding

residues involved in the collapse of the cluster of citrate

synthase, isopropylmalate dehydrogenase, and thermoly-

sin-like protease agreed with the residues that are known

to be critical for thermostability improvement.

Recently, two groups reported that the mutation of resi-

dues at a tunnel to the enzyme active site provided highly

stabilized variants [24,25]. For a haloalkane dehalogenase

it was possible to increase the T appM by 19 8C using this

method [24] yielding an enzyme variant that was also less

inhibited by DMSO, which was attributed to more difficult



http://www.flexweb.asu/

http://www.flexweb.asu/



Engineering of proteins for enhanced thermostability Wijma, Floor and Janssen 3


entrance to the active site. For a b-galactosidase, mutation

of tunnel residues yielded an enzyme with a 10 8C increase

of the optimal temperature for catalytic activity. The

authors concluded that selective targeting of the active

site entrance tunnel should be a broadly applicable strategy

to stabilize enzymes [24]. A tunnel to an active site may

need to be flexible in order to allow substrate to enter the

active site, which can explain the success of this method.

The ancestral and the consensus methodsThe ancestral stabilization method allows for the stabil-

ization of proteins without the use of structural infor-

mation. The method is based on the hypothesis that far

ancestors of current organisms were thermophiles that

possessed proteins which were more thermostable than

extant homologs. Thus, mutations for improving stability

are derived from predicted sequences of common ances-

tors of a protein family. This method has been developed

from 1990 onwards [26] and was successfully used to

stabilize several proteins. Recent progress includes a

study on the stabilization of the B-subunit of the DNA

gyrase, in which the ancestral method was compared to

the consensus method [27]. The consensus method

replaces amino acids by residues that are most frequently

present in multiple sequence alignments. The con-

structed ancestral protein was 4–5 8C more stable

(T appM ) compared to the protein created on basis of the

consensus method. The ancestral method was also used

for recreating a thioredoxin that should have occurred in

organisms that existed about a billion years ago, instead of

millions of years ago as explored used in other studies

[28]. The proposed recreated precambrian thioredoxin

variants were 32 8C more stable in T appM than modern

thioredoxins and were more active under acidic con-

ditions which supposedly existed in the precambrian era.

The half-life of activity at 85 8C of a thermostable cellulase

was increased by 14-fold using the consensus method.

Interestingly, most protein sequences that were used to

create the consensus alignment originated from mesostable

organisms, but they could still contribute to the stability of a

thermostable protein [29]. Consensus mutations were also

shown to enhance the evolvability of a Kemp eliminase.

Such mutations were included alongside functional

mutations to counteracting destabilizing mutations that

were selected because they contributed to the function,

in this case the Kemp elimination [30�]. Without such

stabilizing mutations, it was found to be impossible to

evolve a computationally designed Kemp eliminase.

It is possible to increase the success rate of the ancestral

and consensus methods by including structural infor-

mation. As an example, consensus mutations combined

with structural information on B-factors and inter-sub-

unit disulphide bonds, lead to a 7 8C increase in the half-

life of activity of an a-amino ester hydrolase [16]. For the

ancestral method, a study on the stabilization of a


10.1016/j.sbi.2013.04.008


triosephosphate isomerase introduced two new

parameters (residues that are correlated to other residues

and residues that were nearly invariant and therefore

might have ‘hidden correlations’ to other residues) to

exclude residues from being targeted for mutagenesis

[31]. Using this approach 90% of the predicted mutations

were found to be stabilizing, leading to a final variant that

was 8 8C more stable in T appM , whereas a variant which was

designed without use of these extra parameters was 2 8Cless stable. A study on the stabilization of a b-amylase

combined the ancestral method with structural infor-

mation and introduced a parameter to account for the

extent to which the residues surrounding the target

residue are conserved [32]. This parameter correlated

well with the stability effects of substitutions.

Computational design to stabilize enzymesWith computational protein design, the effect of

mutations is calculated based on the 3D-structure of

the protein. An appropriate energy function implemented

in the algorithm is used to discover mutations that

stabilize the folded protein structure. The use of com-

putational design can reduce the screening effort required

to find a desired stability enhancement by orders of

magnitude as compared to random directed evolution

methods. Two different approaches are currently

employed in the computational design of thermostable

enzymes. Mutations are either selected on the basis of

their predicted effect on the overall folding energy

(DGFold), or computational tools are applied to improve

particular features that stabilize proteins.

Several methods can improve stabilizing features in a

protein, such as loops, hydrophobic clusters, and surface

charges. In all these approaches, the energy function

within the software is used to verify that the designed

protein structure is lower in energy than the native

structure. One approach is the computational design

of a favorable network of positive and negative charges

on the protein surface. This way, it was possible to

increase the T appM of an acylphosphosphatase and a

GTPase by +9 8C [33]. By developing a design algorithm

that used the Rosetta energy function [34] to find the

most favorable positions to introduce mutations that

increase the net charge of a protein (supercharging

[35]), it was possible to increase the heat-survival of

an antibody [36]. The stabilized antibodies were also

monomeric in solution, unlike the native antibody which

formed aggregates. A commonly used computational

approach involves improving the hydrophobic packing

interactions between buried protein residues

[20,21,37��,38]. Software to detect hydrophobic networks

was used to analyze hydrophobic clusters in different

xylanases. Introduction of promising mutations into the

target xylanase gave three mutations which increased the

T appM by 8 8C [38]. An algorithm to introduce hydrophobic

mutations to fill protein voids was experimentally







demonstrated to give a T appM increase of 18 8C when applied

to a methionine amino peptidase [37��].

The second approach for using computational design to

enhance stability does not aim to improve one particular

type of interaction but generates all types of mutations

that should be favorable for stability according to the

implemented energy function. This work is mostly done

with energy functions that are specifically parameterized

to predict differences in DGFold due to point mutations.

Examples of suitable algorithms include FoldX [39],

PoPMuSiC [40], and PreTherMut [41]. Application of

FoldX in combination with consensus design resulted in

an improvement of the T appM of a cellobiohydrolase by 9 8C

[42]. For a feruloyl esterase, the T appM did not increase by

mutations that were designed by PoPMuSiC, but the

half-life of the enzyme at 50 8C was improved 10-fold

[43]. With PreTherMut, the T appM of a methyl parathion

hydrolase was improved by 12 8C [44]. Recently, using a

similar algorithm, the stability (T appM ) of a sesquiterpenoid

synthase was increased by a phenomenal 45 8C [45�].However, the resulting enzyme gave more side-products

and the catalytic activity was reduced.

Chimeric enzymes, truncation andcircularizationThermostability can also be enhanced by making chimeric

proteins consisting of subdomains of different proteins.

This way, favorable features from different parents can be

combined. For example, chimeric variants of Baeyer–Vil-

liger monooxygenases (BVMOs) were obtained by com-

bining subdomains from BVMOs with different stabilities.

Some chimeric enzymes were 10–15 8C more stable that

the thermolabile parent enzymes, while retaining or even

improving the substrate selectivity of a thermolabile

BVMOs [46�]. Several stabilizing chimeras have been

created between different xylanases [47] and between

xylanases and glucanases [48]. A chimera created by graft-

ing loops from an archeal enzyme onto a human aldo-keto

reductase showed that it is possible to combine favorable

features, in this case stability and activity, of archeal and

human proteins [49]. Such a stabilized chimera can even be

obtained without using a thermostable variant, as demon-

strated by combining domains of two mesostable polyol

dehydrogenases [50]. The resulting chimera was 12 8Cmore stable than its parents. Finally, a chimeric enzyme

can also be constructed by proteins with a low sequence

similarity (<30%), as was shown by replacing parts of a

(ba)8-barrel protein by parts of flavodoxin-like proteins

[51��]. The replacement was possible since the structure of

flavodoxin-like proteins resembles a part of the (ba)8-

barrel. The resulting chimera between the two proteins

adopted a fold strongly resembling the (ba)8-barrel scaf-

fold and was thermostable.

A sophisticated method for creating such chimeras is the

SCHEMA approach [52], which uses an algorithm to


10.1016/j.sbi.2013.04.008


calculate the most favorable crossover positions between

subdomains. The calculation selects positions that mini-

mize the number of interactions that are broken when

combining subdomains. The method was applied to the

stabilization of a cellulase by recombination of an enzyme

from three different parents [53,54��]. By analyzing and

modeling the thermostabilities of 54 characterized mem-

bers of a family of fungal cellobiohydrolases, one of the

subdomains appeared critical for stability and its replace-

ment lead to an 8.5 8C increase in apparent melting

temperature (T appM ). In another study, the SCHEMA

method was combined with an active learning algorithm,

which analyzed data from a training set of chimeras to

predict a large collection of stabilizing chimeras [55]. This

method revealed a negative correlation between the iso-

electric point of a chimeric enzyme and its long-term

stability, which was used to construct enzymes with

enhanced long-term stability. Recently, a chimeric

P450 enzyme was created which was 14.5 8C more stable

compared to its parent [56��]. This thermostable enzyme

was created by modeling the protein fitness landscape and

constructing the most stabilizing variants. Bayesian stat-

istics were used to predict, to a reasonable extent, the

mean and variance fitness of any sequence.

Other possible stabilization methods are truncation and

cyclization. It is well known that the removal of thermo-

labile parts or even whole domains from a protein can

significantly improve its stability [57]. Recent achieve-

ments include the stabilization of an endo-b-1,4-gluca-

nase by removing a cellulose binding domain, which in

this case is not needed for biotechnological use of the

enzyme. This truncation resulted in a threefold increase

in half-life of the protein at 65 8C, while the catalytic

activity increased. Circularization of the proteins

sequence can also enhance stability. The N-terminus

and C-terminus are often the most flexible part of a

protein and easily become the targets of proteolytic

enzymes. Therefore, connecting the C-terminus and

N-terminus can stabilize a protein. Such a circulized

protein can be created by introducing the protein of

interest between parts of a naturally split intein from

the dnaE gene [58]. This method was used to circulize a

GFP mutant, resulting in a 5 8C more stable protein [59].

This circulized protein was also highly resistant to

proteolysis.

Conclusions and outlookEven though directed evolution is a highly successful

method for improving protein stability, there is a demand

for additional powerful techniques, especially in case of

proteins that cannot be expressed and assayed in high-

throughput format [60]. Sequence-based bioinformatics

approaches and computational methods exploiting

protein structures are especially successful. When inte-

grated into directed evolution protocols, these newer

method can boost the efficiency and reliability of directed







Figure 1

14

12

8

10

6

4

2

00 10 20 30 40

Obtained increase in TMapp (ºC)

Num

ber

of p

rimar

y pu

blic

atio

ns

2007

- 2

012

Current Opinion in Structural Biology

TappM improvements obtained by protein engineering of enzymes from

2007 to 2012. Primary publications were searched on Web of Science

using different key word combinations to find protein engineering studies

aimed at increasing enzyme thermostability. If a publication was a

follow-up study in which an earlier stabilized enzyme was further

improved, the additional stabilization is plotted.

evolution protocols. For example, the B-fit approach and

the consensus method assist the design of small sets of

variants (or smart libraries) from which improved

enzymes may be obtained by screening only small num-

bers. Furthermore, when three-dimensional structures

are available, energy calculations may identify substi-

tutions that improve hydrophobic cores, or introduce

surface charges that prevent aggregation. This again leads

to smaller numbers of mutants to be tested experimen-

tally, and a high abundance of desired variants among

those sets.

An important remaining challenge is to obtain large

improvements of thermostability. Currently, the increase

in thermostability by mutagenesis typically amounts to 2–15 8C in T app

M (Figure 1). For many applications this is

acceptable because even a modest increase often corre-

sponds to a >10-fold longer life-time (e.g. [20,21,61]).

However, the differences in T appM between natural

enzymes that display hyperthermostability and their

mesostable homologs are much larger, indicating that

there is still a lot of room for improving engineering

techniques.

References and recommended readingPapers of particular interest, published within the period of review,have been highlighted as:

� of special interest�� of outstanding interest

1. Bornscheuer UT, Huisman GW, Kazlauskas RJ, Lutz S, Moore JC,Robins K: Engineering the third wave of biocatalysis. Nature2012, 485:185-194.

2. Ferdjani S, Ionita M, Roy B, Dion M, Djeghaba Z, Rabiller C,Tellier C: Correlation between thermostability and stability


10.1016/j.sbi.2013.04.008


of glycosidases in ionic liquid. Biotechnol Lett 2011, 33:1215-1219.

3. Gao D, Narasimhan DL, Macdonald J, Brim R, Ko MC, Landry DW,Woods JH, Sunahara RK, Zhan CG: Thermostable variants ofcocaine esterase for long-time protection against cocainetoxicity. Mol Pharmacol 2009, 75:318-323.

4. Tokuriki N, Tawfik DS: Stability effects of mutations and proteinevolvability. Curr Opin Struct Biol 2009, 19:596-604.

5. Besenmatter W, Kast P, Hilvert D: Relative tolerance ofmesostable and thermostable protein homologs to extensivemutation. Proteins 2007, 66:500-506.

6. Bloom JD, Labthavikul ST, Otey CR, Arnold FH: Protein stabilitypromotes evolvability. Proc Natl Acad Sci USA 2006, 103:5869-5874.

7. Eijsink VGH, Bjørk A, Gaseidnes S, Sirevag R, Synstad B, Van denBurg B, Vriend G: Rational engineering of enzyme stability. JBiotechnol 2004, 113:105-120.

8. Eijsink VGH, Gaseidnes S, Borchert TV, Van den Burg B: Directedevolution of enzyme stability. Biomol Eng 2005, 22:21-30.

9. Magliery TJ, Lavinder JJ, Sullivan BJ: Protein stability bynumber: high-throughput and statistical approaches to one ofprotein science’s most difficult problems. Curr Opin Chem Biol2011, 15:443-451.

10. Kamerzell TJ, Middaugh CR: The complex inter-relationshipsbetween protein flexibility and stability. J Pharm Sci 2008,97:3494-3517.

11. Bommarius AS, Broering JM, Chaparro-Riggers JF, Polizzi KM:High-throughput screening for enhanced protein stability. CurrOpin Biotechnol 2006, 17:606-610.

12. Polizzi KM, Bommarius AS, Broering JM, Chaparro-Riggers JF:Stability of biocatalysts. Curr Opin Chem Biol 2007, 11:220-225.

13. Reetz MT, Carballeira JD, Vogel A: Iterative saturationmutagenesis on the basis of B factors as a strategy forincreasing protein thermostability. Angew Chem Int Ed Engl2006, 45:7745-7751.

14. Jochens H, Aerts D, Bornscheuer UT: Thermostabilization of anesterase by alignment-guided focussed directed evolution.Protein Eng Des Sel 2010, 23:903-909.

15. Gumulya Y, Reetz MT: Enhancing the thermal robustness of anenzyme by directed evolution: least favorable starting pointsand inferior mutants can map superior evolutionary pathways.Chembiochem 2011, 12:2502-2510.

16. Blum JK, Ricketts MD, Bommarius AS: Improved thermostabilityof AEH by combining B-FIT analysis and structure-guidedconsensus method. J Biotechnol 2012, 160:214-221.

17. Kim SJ, Lee JA, Joo JC, Yoo YJ, Kim YH, Song BK: Thedevelopment of a thermostable CiP (Coprinus cinereusperoxidase) through in silico design. Biotechnol Prog 2010,26:1038-1046.

18.�

Augustyniak W, Brzezinska AA, Pijning T, Wienk H, Boelens R,Dijkstra BW, Reetz MT: Biophysical characterization of mutantsof Bacillus subtilis lipase evolved for thermostability: factorscontributing to increased activity retention. Protein Sci 2012,21:487-497.

The authors investigated the mechanism of stabilization of hyper-stabi-lized esterase variants [13]. The originally observed 45 8C increase in 50%heat-survival assays is entirely due to improved refolding. The authorspropose that improving refolding is an attractive mechanism to stabilizeenzymes.

19. Rader AJ, Hespenheide BM, Kuhn LA, Thorpe MF: Proteinunfolding: rigidity lost. Proc Natl Acad Sci USA 2002, 99:3540-3545.

20. Joo JC, Pohkrel S, Pack SP, Yoo YJ: Thermostabilization ofBacillus circulans xylanase via computational design of aflexible surface cavity. J Biotechnol 2010, 146:31-39.

21. Joo JC, Pack SP, Kim YH, Yoo YJ: Thermostabilization ofBacillus circulans xylanase: computational optimization of



http://refhub.elsevier.com/S0959-440X(13)00075-4/sbref0005








































































unstable residues based on thermal fluctuation analysis. JBiotechnol 2011, 151:56-65.

22.��

Radestock S, Gohlke H: Protein rigidity and thermophilicadaptation. Proteins 2011, 79:1089-1108.

The authors show that Constraint Network Analysis is a powerful tool toinvestigate protein unfolding mechanisms and predict changes in meltingtemperature. See also reference [23]. The results of the analysis providenew insights into protein unfolding.

23. Rathi PC, Radestock S, Gohlke H: Thermostabilizing mutationspreferentially occur at structural weak spots with a highmutation ratio. J Biotechnol 2012, 159:135-144.

24. Koudelakova T, Chaloupkova R, Brezovsky J, Prokop Z,Sebestova E, Hesseler M, Khabiri M, Plevaka M, Kulik D, KutaSmatanova I, Rezacova P, Ettrich R, Bornscheuer UT,Damborsky J: Engineering enzyme stability and a resistanceto an organic cosolvent by modification of residues in theaccess tunnel. Angew Chem Int Ed Engl 2013, 52:1959-1963.

25. Lee HL, Chang CK, Jeng WY, Wang AH, Liang PH: Mutations inthe substrate entrance region of b-glucosidase fromTrichoderma reesei improve enzyme activity andthermostability. Protein Eng Des Sel 2012, 25:733-740.

26. Malcolm BA, Wilson KP, Matthews BW, Kirsch JF, Wilson AC:Ancestral lysozymes reconstructed, neutrality tested, andthermostability linked to hydrocarbon packing. Nature 1990,345:86-89.

27. Akanuma S, Iwami S, Yokoi T, Nakamura N, Watanabe H,Yokobori S, Yamagishi A: Phylogeny-based design of a B-subunit of DNA gyrase and its ATPase domain using a smallset of homologous amino acid sequences. J Mol Biol 2011,412:212-225.

28. Perez-Jimenez R, Ingles-Prieto A, Zhao ZM, Sanchez-Romero I,Alegre-Cebollada J, Kosuri P, Garcia-Manyes S, Kappock TJ,Tanokura M, Holmgren A, Sanchez-Ruiz JM, Gaucher EA,Fernandez JM: Single-molecule paleoenzymology probes thechemistry of resurrected enzymes. Nat Struct Mol Biol 2011,18:592-596.

29. Anbar M, Gul O, Lamed R, Sezerman UO, Bayer EA: Improvedthermostability of Clostridium thermocellum endoglucanaseCel8A by using consensus-guided mutagenesis. Appl EnvironMicrobiol 2012, 78:3458-3464.

30.�

Khersonsky O, Kiss G, Rothlisberger D, Dym O, Albeck S,Houk KN, Baker D, Tawfik DS: Bridging the gaps in designmethodologies by evolutionary optimization of the stabilityand proficiency of designed Kemp eliminase KE59. Proc NatlAcad Sci USA 2012, 109:10358-10363.

The authors enhance the evolvability of an enzyme by co-introducingconsensus mutations. Without these stabilizing mutations it was impos-sible to increase the activity by directed evolution.

31. Sullivan BJ, Nguyen T, Durani V, Mathur D, Rojas S, Thomas M,Syu T, Magliery TJ: Stabilizing proteins from sequencestatistics: the interplay of conservation and correlation intriosephosphate isomerase stability. J Mol Biol 2012, 420:384-399.

32. Yamashiro K, Yokobori S, Koikeda S, Yamagishi A: Improvementof Bacillus circulans beta-amylase activity attained using theancestral mutation method. Protein Eng Des Sel 2010, 23:519-528.

33. Gribenko AV, Patel MM, Liu J, McCallum SA, Wang C,Makhatadze GI: Rational stabilization of enzymes bycomputational redesign of surface charge-chargeinteractions. Proc Natl Acad Sci USA 2009, 106:2601-2606.

34. Das R, Baker D: Macromolecular modeling with rosetta. AnnuRev Biochem 2008, 77:363-382.

35. Lawrence MS, Phillips KJ, Liu DR: Supercharging proteins canimpart unusual resilience. J Am Chem Soc 2007, 129:10110-10112.

36. Miklos AE, Kluwe C, Der BS, Pai S, Sircar A, Hughes RA,Berrondo M, Xu J, Codrea V, Buckley PE, Calm AM, Welsh HS,Warner CR, Zacharko MA, Carney JP, Gray JJ, Georgiou G,Kuhlman B, Ellington AD: Structure-based design of


10.1016/j.sbi.2013.04.008


supercharged, highly thermoresistant antibodies. Chem Biol2012, 19:449-455.

37.��

Borgo B, Havranek JJ: Automated selection of stabilizingmutations in designed and natural proteins. Proc Natl Acad SciUSA 2012, 109:1494-1499.

The paper describes an algorithm that automatically detects voids in thehydrophobic core of a protein and then suggests mutations to fill thosevoids and improve protein stability.

38. Kim T, Joo JC, Yoo YJ: Hydrophobic interaction networkanalysis for thermostabilization of a mesophilic xylanase. JBiotechnol 2012, 161:49-59.

39. Van Durme J, Delgado J, Stricher F, Serrano L, Schymkowitz J,Rousseau F: A graphical interface for the FoldX forcefield.Bioinformatics 2011, 27:1711-1712.

40. Dehouck Y, Grosfils A, Folch B, Gilis D, Bogaerts P, Rooman M:Fast and accurate predictions of protein stability changesupon mutations using statistical potentials and neuralnetworks: PoPMuSiC-2.0. Bioinformatics 2009, 25:2537-2543.

41. Tian J, Wu N, Chu X, Fan Y: Predicting changes in proteinthermostability brought about by single- or multi-sitemutations. BMC Bioinformatics 2010, 11:370.

42. Komor RS, Romero PA, Xie CB, Arnold FH: Highly thermostablefungal cellobiohydrolase I (Cel7A) engineered using predictivemethods. Protein Eng Des Sel 2012, 25:827-833.

43. Zhang SB, Wu ZL: Identification of amino acid residuesresponsible for increased thermostability of feruloyl esteraseA from Aspergillus niger using the PoPMuSiC algorithm.Bioresour Technol 2011, 102:2093-2096.

44. Tian J, Wang P, Huang L, Chu X, Wu N, Fan Y: Improving thethermostability of methyl parathion hydrolase fromOchrobactrum sp. M231 using a computationally aidedmethod. Appl Microbiol Biotechnol. 2013, 97:2997-3006.

45.�

Diaz JE, Lin CS, Kunishiro K, Feld BK, Avrantinis SK, Bronson J,Greaves J, Saven JG, Weiss GA: Computational design andselections for an engineered, thermostable terpene synthase.Protein Sci 2011, 20:1597-1606.

This report shows that computational design can be used to gain anextreme increase in melting temperature of an enzyme, although at theexpense of catalytic activity and selectivity.

46.�

van Beek HL, de Gonzalo G, Fraaije MW: Blending Baeyer–Villiger monooxygenases: using a robust BVMO as a scaffoldfor creating chimeric enzymes with novel catalytic properties.Chem Commun (Camb) 2012, 48:3288-3290.

The authors show they can create selective and stable chimeric enzymesby introducing favorable activity from a thermolabile enzyme into athermostable scaffold.

47. Liu LW, Zeng LY, Wang SY, Cheng J, Li XQ, Song AD, Wu K,Chen HG: Activity and thermostability increase of xylanasefollowing transplantation with modules sub-divided fromhyper-thermophilic CBM9_1-2. Process Biochem 2012, 47:853-857.

48. Liu L, Wang L, Zhang Z, Guo X, Li X, Chen H: Domain-swappingof mesophilic xylanase with hyper-thermophilic glucanase.BMC Biotechnol 2012, 12:28.

49. Campbell E, Chuang S, Banta S: Modular exchange ofsubstrate-binding loops alters both substrate and cofactorspecificity in a member of the aldo-keto reductasesuperfamily. Protein Eng Des Sel 2012, 26:181-186.

50. Wulf H, Mallin H, Bornscheuer UT: Protein engineering of athermostable polyol dehydrogenase. Enzyme Microb Technol2012, 51:217-224.

51.��

Shanmugaratnam S, Eisenbeis S, Hocker B: A highly stableprotein chimera built from fragments of different folds. ProteinEng Des Sel 2012, 25:699-703.

The authors create a stabilizing chimeric enzyme by combining proteinswith a very low sequence identity. This article shows that successfulcombinations between such proteins are possible.

52. Otey CR, Landwehr M, Endelman JB, Hiraga K, Bloom JD,Arnold FH: Structure-guided recombination creates anartificial family of cytochromes P450. PLoS Biol 2006, 4:e112.

































































































































53. Heinzelman P, Snow CD, Smith MA, Yu X, Kannan A, Boulware K,Villalobos A, Govindarajan S, Minshull J, Arnold FH: SCHEMArecombination of a fungal cellulase uncovers a singlemutation that contributes markedly to stability. J Biol Chem2009, 284:26229-26233.

54.��

Heinzelman P, Snow CD, Wu I, Nguyen C, Villalobos A,Govindarajan S, Minshull J, Arnold FH: A family of thermostablefungal cellulases created by structure-guided recombination.Proc Natl Acad Sci USA 2009, 106:5610-5615.

The authors use the SCHEMA method to predict and create stabilizingchimeras. The developed model correlates reasonably well with themeasured stability.

55. Romero PA, Stone E, Lamb C, Chantranupong L, Krause A,Miklos A, Hughes RA, Fechtel B, Ellington AD, Arnold FH,Georgiou G: SCHEMA-designed variants of human arginase I &II reveal sequence elements important to stability andcatalysis. ACS Synth Biol 2012, 1:221-228.

56.��

Romero PA, Krause A, Arnold FH: Navigating the protein fitnesslandscape with Gaussian processes. Proc Natl Acad Sci USA2013, 110:E193-E201.


10.1016/j.sbi.2013.04.008


The authors present a successful method to model the protein-fitnesslandscape using Bayesian statistics, and use this method to construct a15 8C more stable P450.

57. Wang Y, Yuan H, Wang J, Yu Z: Truncation of the cellulosebinding domain improved thermal stability of endo-beta-1,4-glucanase from Bacillus subtilis JA18. Bioresour Technol 2009,100:345-349.

58. Iwai H, Lingel A, Pluckthun A: Cyclic green fluorescent proteinproduced in vivo using an artificially split PI-PfuI intein fromPyrococcus furiosus. J Biol Chem 2001, 276:16548-16554.

59. Zhao Z, Ma X, Li L, Zhang W, Ping S, Xu MQ, Lin M: Proteincyclization enhanced thermostability and exopeptidase-resistance of green fluorescent protein. J. Microbiol Biotechnol2010, 20:460-466.

60. Lutz S: Beyond directed evolution — semi-rational proteinengineering and design. Curr Opin Biotechnol 2010, 21:734-743.

61. Korkegian A, Black ME, Baker D, Stoddard BL: Computationalthermostabilization of an enzyme. Science 2005, 308:857-860.







































structure- and sequence-analysis inspired engineering of proteins for enhanced thermostability

Documents