, f 21x. .3?“ a: z
. 3‘53 w. 5.3.)? n
31:21.3... . , V... A, _ _

..._2
1r..7_

_. 7; i1 1

 

 

W‘SIS ‘ l

This is to certify that the

dissertation entitled

GENETIC ENGINEERING 0F XYLOSE ISOMERASE
THERMOZYMES FOR ENHANCED ACTIVITY, STABILITY, AND UTILITY

presented by
DINLAKA SRIPRAPUNDH

has been accepted towards fulfillment
of the requirements for

Doctoral Food Science

degree in

y ‘sz 7- M

 

 

 

Date ,2'13‘ 2002

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

GENE

GENETIC ENGINEERING OF XYLOSE ISOMERASE THERMOZYMES FOR
ENHANCED ACTIVITY, STABILITY, AND UTILITY

By

Dinlaka Sriprapundh

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Food Science and Human Nutrition

2002

 

GEI

isome
identi‘

less

ABSTRACT

GENETIC ENGINEERING OF XYLOSE ISOMERASE THERMOZYMES FOR
ENHANCED ACTIVITY, STABILITY, AND UTILITY

By

Dinlaka Sriprapundh

Molecular determinants responsible for high thermostability of the xylose
isomerase from hyperthermophilic eubacterium Thermotoga neapolitana (TNXI) were
identified by comparative thermostability and site-directed mutagenesis studies with the
less thermostable counterpart enzyme from thermophilic eubacterium
T hermoanaerobacterium thermosulfitrigenes (TTXI). Despite their highly similar
structures and amino acid sequences (70.4 % identity), no obvious differences in the
enzyme structures can explain the differences in TNXI’s stability compared to that of
'ITXI except for a few additional prolines and fewer Asn+Gln in TNXI. TNXI has 2
additional prolines in the Phe59 loop (Pr058 and Pro62). This loop helps forming another
enzyme subunit’s active site. When the 2 prolines in TNXI were substituted with the
corresponding amino acids present in its less thermostable counterpart, TTXI, all mutant
enzymes showed significant loss in thermostability compared to the wild-type TNXI.
These data confirmed the hypothesis that prolines indeed play important roles in TNXI
thermostability by reducing its entropy of unfolding.

TNXI’s active site was engineered to improve its catalytic efficiency toward
glucose. The TNXI V185T mutant derivative was three times more efficient in glucose
isomerization than the wild-type TNXI. Although this mutant derivative was highly

thermostable and highly active at 97°C, it was less than 10% as active at 60°C and

 

 

required net
evolution at
low temper:
low tempera
exhibited dr
low pH as
mutation rel;
active site of
pHs. TNXI l
resulth fror
neighboring }
conformation
Bloch:
awmmaou
higher cataln
gm“? Iherm.
aClll’lty. Tl'lls
glucose Wilde]

Process based,

required neutral pH to work. To customize this TNXI mutant derivative, a directed
evolution approach was applied to the TNXI V185T to improve its activity on glucose at
low temperature and low pH. After two successive rounds of random mutagenesis and
low temperature/low pH activity screening, a new mutant, TNXI 1F1, was obtained that
exhibited dramatic improvement of glucose isomerase activity at low temperature and
low pH as compared to TNXI V185T. TNXI 1F 1 (V185T/L282P/F186S) with one
mutation relatively distant (L282P) from the active site and the other (F1868) within the
active site of the enzyme, was more active than TNXI V185T over all temperatures and
pHs. TNXI 1F] was also more stable than TNXI and TNXI V185T and this may have
resulted from additional H-bond formation between Ser186’s sidechain and the
neighboring L229 residue’s mainchain structure. This H-bond would strengthen local
conformation and the affinity of E231 co-ordination with the structural metal.
Biochemical properties and fructose productivities of TNXI 1F] and GensweetTM,
a commercially available glucose isomerase were also compared. TNXI lFl displayed
higher catalytic efficiencies on glucose at low or high temperature and pH ranges and had
greater thermal stability than GensweetTM despite having similar temperature optima of
activity. This greater thermal stability together with the superior kinetic parameters on
glucose render TNXI lFl an excellent candidate for the industrial glucose isomerization

process based on the lifetime fructose productivity estimation.

Copyright by
Dinlaka Sriprapundh
2002

Dedicated to my family, Mr. Udom and Mrs. Achara Sriprapundh

ml

 

 

 

J. Grcgt
guidance
Motenoi
research

DTS. IOhI

th‘enic
lhroughr
Kalli-aim
COmPUIt
slide- m.-
M. Kelly
Ontne p,

immobil

OPPOnUn

I

ACKNOWLEDGEMENTS

I would like to express my sincere appreciation and deepest gratitude to Professor
J. Gregory Zeikus, who is my scientific advisor and philosophical mentor, for his
guidance, encouragement, and patience through my graduate school years. I cannot thank
him enough for all the support he has provided me for the successful completion of this
research project. I would also like to thank all of the members of my guidance committee,
Drs. John Linz, James Pestka, and Gale Strasburg, for their advice and invaluable time.

I am grateful to my laboratory members especially Dr. Claire Vieille, Maris
Laivenieks, and Su-il Kang for all their support, friendship, advice, and encouragement
throughout my graduate program. I would also like to acknowledge the help from Dr.
Kaillathe Padmanabhan for computer aided molecular modeling in Silicon Graphic
Computer System, insightful discussion of protein structure, and assistant in poster and
slide- making as well as other computer-related troubles. I also should thank Dr. Robert
M. Kelly and Kevin Epting at the North Carolina State University for collaboration work
on the project and providing facilities and advice for the DSC measurement and enzyme
immobilization.

I owe my success to my former advisor Dr. Sirirat Rengpipat for giving me the
opportunities, her enormous support and encouragement through tough times. Without
her assistance, this work could not have been completed.

I am most grateful to my best friend Paweena Limjaroen for her dedication and

encouragement. I could not have undertaken this endeavor without her help and

vi

 

UEJC

while

understanding. My deepest appreciation goes to my family and friends who supported me

while accomplishing this work.

vii

 

L151

L157

(TIA
Liter.

 

l CHAP‘
‘ Role of
l Themu:
‘ .\\ lOsc

Xylem ,

TABLE OF CONTENTS

page

LIST OF TABLES .................................................................................... x

LIST OF FIGURES ................................................................................... xi

CHAPTER I

Literature Review ..................................................................................... l
Xylose isomerase and high fructose corn syrup ......................................... 2
X1 structure, sequence homology and reaction mechanisms ........................... 5
Thermozymes and their thermal stability ................................................ 13
Genetic and protein engineering of XIs ................................................. 16

Thermostability mutations ....................................................... 16
Active site mutations .............................................................. 20

Directed evolution of enzymes ........................................................... 22
References .................................................................................. 30

CHAPTER II

Role of Proline in Xylose Isomerase Thermal Stability: Comparative

Thermostability of a Thermophilic Thermoanaerobacterium thermosulfurigenes

Xylose Isomerase (TTXI) and a Hyperthermophilic Thermotoga neapolitana

Xylose Isomerase (TNXI) .......................................................................... 36
Abstract ...................................................................................... 37
Introduction ................................................................................. 38
Materials and Methods ..................................................................... 41
Results ....................................................................................... 45
Discussion ................................................................................... 52
References ................................................................................... 57

CHAPTER HI

Role of Substrate Binding Pocket in Thermotoga neapolitana Xylose Isomerase

(TNXI) on Glucose Isomerase Activity .......................................................... 59
Abstract ...................................................................................... 60
Introduction ................................................................................. 6 1
Materials and Methods ..................................................................... 63
Results ....................................................................................... 67
Discussion ................................................................................... 70
References ................................................................................... 73

CHAPTER IV

Directed Evolution of Thermotoga neapolitana Xylose Isomerase: High Activity

on Glucose at Low Temperature and Low pH ................................................... 74
Abstract ...................................................................................... 75
Introduction ................................................................................. 77

viii

 

CHAI
Bioch-
an Ind
Ewfl\

CHAP
Role 0

CH AP’
COnclu

page

Materials and Methods ..................................................................... 80

Results ........................................................................................ 85

Discussion .................................................................................... 99

References ................................................................................. 105
CHAPTER V

Biochemical Comparison of Stability, Activity, and Fructose Synthesis by
an Industrial Glucose Isomerase (GENSWEETW) versus a Laboratory-

Evolved Thermo-Acid Stable Xylose Isomerase (TNXI lFl) ............................... 108
Abstract ..................................................................................... 109
Introduction ................................................................................ 1 10
Materials and Methods ................................................................... 1 13
Results ...................................................................................... 1 17
Discussion ................................................................................. 13 1
References ................................................................................. 1 36

CHAPTER VI .

Role of Metals in TNXI Stability at Extremely High Temperatures ........................ 138
Abstract ..................................................................................... 139
Introduction ................................................................................ 1 40
Materials and Methods ................................................................... 142
Results ...................................................................................... 143
Discussion ................................................................................. 147
References ................................................................................. 149

CHAPTER VII
Conclusions and Directions for Future Research .............................................. 150

ix

.11

 

 

 

CHAPT

Tbkl.

CHAPT
Lmkl.
luHeZ.

IuMe3.

CHAPT'

IaMcl,

CHAPT,
lame].
CHAPT;

Tdble 1.

LIST OF TABLES

CHAPTER I

Table 1. Examples of enzymes that were successfully optimized using directed

evolution ................................................................................ 28-29
CHAPTER II
Table 1. Oligonucleotides and DNA templates used for site-directed mutagenesis ........ 42
Table 2. Effect of mutations on enzyme precipitation temperatures .......................... 50
Table 3. Catalytic parameters of TTXI (at 65°C) and TNXI (at 80°C) and of their

mutant derivatives ......................................................................... 51
CHAPTER HI
Table 1. Oligonucleotides and DNA templates used for site-directed

mutagenesis ................................................................................. 64
Table 2. Catalytic parameters of TTXI (at 65°C) and TNXI (at 80°C) and of their

mutant derivatives ......................................................................... 69
CHAPTER IV
Table 1. Kinetic parameters of TNXI and its mutant derivatives .............................. 91
CHAPTER V
Table 1. Comparison of glucose isomerization kinetic parameters of

TNXI 1F 1 vs. GensweetTM .............................................................. 125
Table 2. Comparison of thermal activity and stability properties of

TNXI lFl vs. GensweetTM ............................................................ 132
Table 3. Comparison of fructose production parameters of

TNXI lFl vs. GensweetTM ............................................................. 134
CHAPTER VI
Table 1. Melting temperatures (Tm) of TNXI and its mutant derivatives

in the presence (holo-enzyme) and absence (apo-enzyme) of

5 mM MgSO4 and 0.5 mM COCl2 as determined by DSC ......................... 146

 

 

 

 

CHA
Figur

Figure
Figure
Figure
FJSUre.

CHAPT

Figure 1

‘EUre 2

LIST OF FIGURES

“Images in this dissertation are presented in color”

CHAPTER I
Figure 1. Illustration of the reactions carried out by xylose isomerases (XIs) ................ 3
Figure 2. Starch processing into high fructose corn syrup (HFCS) ............................ 4

Figure 3. Three-dimensional structure of a subunit of Thermotoga neapolitana
xylose isomerase (type 11 XI) ............................................................ 7

Figure 4. Homology alignment of xylose isomerase amino acid sequences from
different organisms by the ClustalW analysis ..................................... 9-10

Figure 5. Schematic representation of the active site of TNXI ................................ 1 1

Figure 6. A metal-assisted hydride-shift mechanism of xylose in the

Streptomyces rubiginosus xylose isomerase (type I XI) ............................. 14
CHAPTER 11
Figure 1. Three-dimensional structure of TNXI’s Phe59 loop region ........................ 46

Figure 2. Effect of temperature on the specific activities of TFXI and TNXI and
their Phe59 loop mutant derivatives .................................................. 47

Figure 3. Inactivation curves of TTXI and TNXI and their Phe59 loop mutant
derivatives at 85°C (A) (TTXI and derivatives) and 95°C (B)
(TNXI and derivatives) ................................................................. 48

Figure 4. Three-dimensional structure of the Phe59 loop mutations in TTXI ............... 54

Figure 5. Van der Waals contacts between Lys61 and Pro62 in 'ITXI Ala62Pro

mutant derivative ........................................................................ 55
CHAPTER HI
Figure 1. Activity and stability of TNXI and its active site mutant
derivatives ................................................................................. 68
Figure 2. Superposition of TTXI and TNXI active sites ....................................... 71

xi

 

 

CHAP

Figure

Figure

Figure

CHA]

CHAPTER IV
Figure l. A) Effect of temperature on the specific activities of TNXI and its
mutant derivatives on glucose at pH 7.0 and

B) Ln (specific activity) versus 1/Temperature ................................... 87-88

Figure 2. Effect of pH on the specific activities of TNXI and its mutant
derivatives on glucose at 80°C ......................................................... 89

Figure 3. Inactivation curves of TNXI and its mutant derivatives at A) 80°C,

pH 7.0 and B) 80°C, pH 5.5 ........................................................ 92-93
Figure 4. Three—dimensional model of the TNXI 1F] monomer showing the

positions of mutations V185T, F1868, and L282P .................................... 96
Figure 5. Three-dimensional model of part of the area surrounding TNXI 3A2

and 1F1’s Leu282Pro mutation ......................................................... 97
Figure 6. Three-dimensional model of part of the TNXI 1F 1 active site showing .

hydrogen bonds among 8186, L229, and E231 ....................................... 98

CHAPTER V

Figure 1. Effect of temperature on the specific activities of TNXI 1F1 and
GensweetTM on glucose at pH 7.0 .................................................... 1 18

Figure 2. Effect of pH on the specific activities of TNXI 1F1 and GensweetTM
on glucose at pH 7.0 ................................................................... 1 19

Figure 3. Inactivation curves of TNXI 1F1and GensweetTM at 60°C (pH 7.0) (A)
and at 60°C (pH 5.5) (B) .......................................................... 121-122

Figure 4. Thermal unfolding of TNXI 1F1 (A) and GensweetTM (B) in the
presence of 5 mM MgSO., and 0.5 mM CoClz followed by DSC ............... 123

Figure 5. Estimated fructose productivity of TNXI 1F] and GensweetTM in
different conditions .................................................................... 126

Figure 6. Experimental fructose conversion of TNXI 1F1 and GensweetTM
at pH 7.0 and 5.5 at 80°C (A) and at 60°C(B) ...................................... 128

Figure 7. Browness of syrups from experimental fructose conversion
of TNXI 1F 1 and GensweetTM ........................................................ 129

xii

 

 

C HAPTEF

Figure 1.T
it

Figure 3: T
lit‘

CHAPTER VI

Figure 1. Thermal unfolding of the holo-forms of TNXI and its mutant derivatives
in the presence of 5 mM MgSO4 and 0.5 mM CoClz followed by DSC ........ 144

Figure 2: Thermal unfolding of the apo-forms of TNXI and its mutant derivatives
followed by DSC ........................................................................ 145

xiii

CHAPTER I

LITERATURE REVIEW

XYLOSE ISOMERASE AND HIGH FRUCTOSE CORN SYRUP

Xylose isomerase (XI) (D-xylose ketol isomerase; EC 5.3.1.5), generally known
as glucose isomerase (GI), catalyzes the interconversion of D-xylose to D-xylulose in
vivo and D-glucose to D-fructose in vitro (Figure 1). Interconversion of xylose to
xylulose provides a nutritional requirement in bacteria that thrive on decaying plant
materials and also aids in the bioconversion of hemicelluloses to ethanol. Bioconversion
of renewable biomass to fermentable sugars and ethanol is important in view of the rapid
depletion of fossil fuels. Isomerization of glucose to fructose is of commercial
importance in the production of high fructose corn syrup (HFCS) (Bhosale et al., 1996).
Xylose isomerase is one of the most important industrial enzymes in the food industry
and also one of the three highest tonnage value enzymes, amylase and protease being the
other two (Bhosale et al., 1996).

The production of HFCS (Figure 2) from cornstarch comprises three major
processes: (1) liquefaction of starch by OL-amylase, (2) saccharification of starch slurry by
the combined action of amyloglucosidase and debranching enzymes, and (3)
isomerization of glucose by X1 (or G1). The final product is a corn syrup containing a
mixture of glucose and fructose with a greater sweetening capacity than that of sucrose.
An equilibrium mixture of glucose and fructose (1:1) is 1.3 times sweeter than sucrose
and 1.7 times sweeter than glucose. The sweetening capacity of glucose is 70 to 75 ‘70 that
of sucrose, whereas fructose is twice as sweet as sucrose (Barker, 1976). The price of

HFCS is 10 to 20 % lower than that of sucrose on the basis of its sweetening power.

 

 

 

H H O H H O H- C-OH
0 OH H OH
H OH OH H
xylopyranose xylulofuranose
CH20H H
H H o H CH2OH o 'H- C-OH
01 OH H OH
H OH OH H

glucopyranose fructofuranose

Figure 1: Illustration of the reaction carried out by xylose isomerases (XIs)

   
   

ED 42 % Fructose

, Glucose
1 isomerase

‘iJLI .2.',- '

9

NaOH
Rig“ Deionize

[a -Amy1ase I Flucoamylase

WW

 

 

 

         
   
     

 

 

Liquefaction Saccharification Isomerization

pH 6.0 pH 4-2 — 4-6 pH 8.0 — 8.2
95-105°c 60°C 55 — 60°C

90 min 48'96 hr Streptomyces,
Bacillus licheniformis ASPergillus mm, Bacillus,

B. .rtearmhermnnhilus Aspergillus ”i8” Actinnnlanes

 

Figure 2: Starch processing into high fructose corn syrup (HFCS)

HFCS is preferred by the food industry since it does not pose the problem of
crystallization, as is the case with sucrose. Moreover, fructose plays an important role as
a diabetic sweetener because it is only slowly reabsorbed by the stomach and does not
influence the glucose level in blood.

The major uses of HFCS are in the beverage, baking, canning, and confectionery
industries. The annual world consumption of HFCS was estimated to have reached 10
million tons (dry weight) in 1995 (de Raadt et al., 1994). In 1996, HFCS has almost
completely replaced sucrose in beverage market in the United States. Beside its industrial
importance, xylose isomerase also serves as an interesting model for studying structure-
function relationships by advanced biochemical and genetic engineering techniques such

as site-directed mutagenesis and directed evolution.

XI STRUCTURE, SEQUENCE HOMOLOGY, AND REACTION MECHANISMS

To elucidate structure-function relationships of xylose isomerases, the genes
encoding XIs (xylA) sequences from various organisms have been compared. XIs can be
classified into two groups based on their amino acid sequences and a 40-50 residue N-
terminal extension (Vangrysperre et al., 1988). Class I XIs contain about 390 amino acids
and consist of those from Streptomyces spp., Actinoplanes spp., Ampullariella spp.,
Arthrobacter spp., and Thermus thermophilus. The enzymes from Escherichia coli,
Bacillus spp., Lactobacillus spp., Lactococcus spp., Thermoanaerobacterium
thennosulfurigenes, and Thermotoga spp. contain approximately 440 amino acids and are

grouped as class II XIs. The enzymes being studied extensively in this work

(Thermoanaerobacterium thermosulfurigenes XI: TTXI and Thermotoga neapolitana XI:
TNXI) belong to the class H XIs. The XI monomer consists of an eight—stranded parallel
B-barrel surrounded by eight helices with an extended C-terminal tail that provides
extensive contacts with a neighboring monomer (Figure 3). The active site pocket is
defined by an opening in the barrel of which the entrance is lined with hydrophobic
residues while the bottom of the pocket consists mainly of glutamate, aspartate, and
histidine residues coordinated to two bivalent cations (MgZT, Mn”, or Coz+)(Whitlow et
al., 1991). Three type I and five type H XI’s xylA sequences were aligned based on their
sequence homology (Fig 4). In spite of the low homology between classes I and II
enzymes, the amino acids involved in the substrate (H100, T140, E231, K233, D338) and
metal ions binding (E231, E267, H270, D295, D306, D308, D338), as well as in catalysis
(H100, D103, D338), are completely conserved (Fig. 5). Thus, the essential structure at
the catalytic center of XIs appears to be analogous in class H XIs so far compared.

The subunit structure and amino acid composition of XI reveal that it is a tetramer
or a dimer of similar or identical subunits associated with non-covalent bonds and is
devoid of interchain disulfide bonds. Most known XIs are homotetramers with molecular
masses of approximately 45 to 50 kDa per subunit, although some XIs have been found
to be dimeric. The dissociation and unfolding of the tetrameric XI from Streptomyces sp.
Strain NCIM 2730 revealed that the tetramer and the dimer are the active species whereas
the monomer is inactive (Ghatge et al., 1994).

The native Thermotoga neapolitana homotetrameric xylose isomerase is also

eXpressed as a catalytically active and thermostable dimer in E. coli (Hess et al., 1998).

 

Figure 3: Three-dimensional structure of a subunit of Thermotoga neapolitana xylose

isomerase (type 11 XI).

 

 

l
-
1“
re! ;
Figure 4: Homology alignment of xylose isomerase amino acid sequences from different 3 T-
Li organisms by the ClustalW analysis (http://www.ebi.ac.uk/clustalw). The sequences are f ‘

T.m., Thermotoga maritima; T.n., Thermotoga neapolitana:

 

T.t., Thermoanaerobacterium thennasulfurigenes; B.s., Bacillus subtilis;

lllil
1;

B.l., Bacillus licheniformis; E.c., Escherichia coli; A.m., Actinoplanes missouriensis;

' l

S.r., Streptomyces rubiginosus; Ar, Arthrobactor sp. (*) means amino acids are identical,

1" “5 7|" m tn
" . .

‘1

(2) means conserved substitutions, and (.) means semi-conserved substitutions.

‘ H} l'

' In 7" (I! m In .‘
L! 1') " i h

"

ii
In I

; .

.‘I' IN ”I “y
I) I

‘l

 

w m y m w w a a a
H 3 n F-m n 5 3

v m > m w w a a a
H B n F'm n 3 3 H B n H‘m n S 3

y m y m m w 6 6 H

> m > m m m a a a
H B n H‘m n D S

> m y m w m H 6 a
H 3 n H‘m n D 3

y m > m m m H a H
H 8 0 »:m n S B

------- MAEFFPEIPKIQFEGKESTNPLAFRFYDPNEVIDGKPLKDHLKFSVAFWHTFV
------- MAEFFPEIPKVQFEGKESTNPLAFKFYDPEEIIDGKPLKDHLKFSVAFWHTFV
------- MNKYFENVSKIKYEGPKSNNPYSFKFYNPEEVIDGKTMEEHLRFSIAYWHTFT
MAQSHSSSINYFGSANKVVYEGKDSTNPLAFKYYNPQEVIGGKTLKEHLRFSIAYWHTFT

--------- MFFRNIGMIEYEGADSENPYAFKYYNPDEFVGGKTMKEHLRFAVAYWHTFD
——————— MQAYFDQLDRVRYEGSKSSNPLAFRHYNPDELVLGKRMEEHLRFAACYWHTFC
———————————————————————————— MSVQATREDKFSFG-~——---——--——LWTVG
———————————————————————————— MNYQPTPEDRFTFG—--—--—-——-—-LWTVG
———————————————————————————— MSVQPTPADHFTFG-—-—-——-—-——-LWTVG

* . *

NEGRDPFGDPTAERPWNRFSDPMDKAFARVDALFEFCEKLNIEYFCFHDRDIAPEGKTLR
NEGRDPFGDPTADRPWNRYTDPMDKAFARVDALFEFCEKLNIEYFCFHDRDIAPEGKTLR
ADGTDQFGKATMQRPWNHYTDPMDIAKARVEAAFEFFDKINAPYFCFHDRDIAPEGDTLR
ADGTDVFGAATMQRPWDHYKG-MDLAKMRVEAAFEMFEKLDAPFFAFHDRDIAPEGSTLK
ADGKDPFGDGTMFRAWNRLTHPLDKAKARAEAAFEFFEKLGVPYFCFHDVDIVDEGATLR

WNGADMFGVGAFNRPWQQPGEALALAKRKADVAFEFFHKLHVPFYCFHDVDVSPEGASLK
WQARDAFGDATRT ------------- ALDPVEAVHKLAEIGAYGITFHDDDLVPFGSDAQ
WQGRDPFGDATRR ------------- ALDPVESVRRLAELGAHGVTFHDDDLIPFGSSDS
WTGADPFGVATRK ------------- NLDPVEAVHKLAELGAYGITFHDNDLIPFDATEA

'k it . . . *‘kt *.

ETNKILDKVVERIKERMKDSNVKLLWGTANLFSHPRYMHGAATTCSADVFAYAAAQVKKA
ETNKILDKVVERIKERMKDSNVKLLWGTANLFSHPRYMHGAATTCSADVFAYAAAQVKKA
ETNKNLDTIVAMIKDYLKTSKTKVLWGTANLFSNPRFVHGASTSCNADVFAYSAAQVKKA
ETNQNLDMIMGMIKDYMRNSGVKLLWNTANMFTNPRFVHGAATSCNADVFAYAAAQVKKG
ETFTYLDQMSSFLKEMMETSHVQLLWNTANMFTHPRYVHGAATSCNADVYAYAAAKVKKG
EYINNFAQMVDVLAGKQEESGVKLLWGTANCFTNPRYGAGAATNPDPEVFSWAATQVVTA
TR----DGIIAGFKKALDETGLIVPMVTTNLFTHPVFKDGGFTSNDRSVRRYAIRKVLRQ
ER----EEHVKRFRQALDDTGMKVPMATTNLFTHPVFKDGGFTANDRDVRRYALRKTIRN
ER----EKILGDFNQALKDTGLKVPMVTTNLFSHPVFKDGGFTSNDRSIRRFALAKVLHN

*.* *..* . i i .
. .. . . . ..

LEITKELGGEGYVFWGGREGYETLLNTDLGLELENLARFLRMAVEYAKKIGFTGOFLIEP
LEITKELGGEGYVFWGGREGYETLLNTDLGFELENLARFLRMAVDYAKRIGFTGQFLIEP
LEITKELGGENYVFWGGREGYETLLNTDMEFELDNFARFLHMAVDYAKEIGFEGQFLIEP
LETAKELGAENYVFWGGREGYETLLNTDLKFELDDLARFMHMAVDYAKEIGYTGQFLIEP
LDIAKELGAENYVFWGGREGYETLLNTDMKLELENLSSFYRMAVEYAREIGFDGQFLIEP
MEATHKLGGENYVLWGGREGYETLLNTDLRQEREQLGRFMQMVVEHKHKIGFQGTLLIEP
MDLGAELGAKTLVLWGGREGAEYDSAKDVSAALDRYREALNLLAQYSEDRGYGLRFAIEP
IDLAVELGAETYVAWGGREGAESGGAKDVRDALDRMKEAFDLLGEYVTSQGYDIRFAIEP
IDLAAEMGAETFVMWGGREGSEYDGSKDLAAALDRMREGVDTAAGYIKDKGYNLRIALEP

.." . * *i**** ”k i. . s t. . .**
a o. o a o . . o -

KPKEPTKHQYDFDVATAYAFLKNHGLDEYFKFNIEANHATLAGHTFQHELRMARILGKLG
KPKEPTKHQYDFDVATAYAFLKSHGLDEYFKFNIEANHATLAGHTFQHELRMARILGKLG
KPKEPTKHQYDFDVANVLAFLRKYDLDKYFKVNIEANHATLAFHDFQHELRYARINGVLG
KPKEPTAHQYDTDAATTIAFLKQYGLDNHFKLNLEANHATLAGHTFEHELRMARVHGLLG
KPKEPTKHQYDFDAATTIAFLETYGLKDHFKLNLEANHATLAGHTFEHELRVAALHDMLG
KPQEPTKHQYDYDAATVYGFLKQFGLEKEIKLNIEANHATLAGHSFHHEIATAIALGLFG
KPNEPRGDILLPTAGHAIAFVQELERPELFGINPETGHEQMSNLNFTQGIAQALWHKKLF
KPNEPRGDILLPTVGHALAFIERLERPELYGVNPEVGHEQMAGLNFPHGIAQALWAGKLF
KPNEPRGDIFLPTVGHGLAFIEQLEHGDIVGLNPETGHEQMAGLNFTHGIAQALWAEKL

**.it i. i i i .. i . . i

SIDANQGDLLLGWDTDQFPTNIYDTTLAMYEVIKAGG ----- FTKGGLNFDAKVRRASYK
SIDANQGDLLLGWDTDQFPTNVYDTTLAMYEVIKAGG ----- FTKGGLNFDAKVRRASYK
SIDANTGDMLLGWDTDQFPTDIRMTTLAMYEVIKMGG ----- FDKGGLNFDAKVRRASFE
SVDANQGHPLLGWDTDEFPTDLYSTTLAMYEILQNGG ----- LGSGGLNFDAKVRRSSFE
SIDANQGDLLLGWDTDEFPTDLYSAVLAMYEILKAGG ----- FKTGGINFDAKVRRPSFA
SVDANRGDAQLGWDTDQFPNSVEENALVMYEILKAGG ----- FTTGGLNFDAKVRRQSTD
HIDLNGQHGPKFDQDLVFGHGDLLNAFSLVDLLENGP-DGAPAYDGPRHFDYKPSR-TED
HIDLNGQNGIKYDQDLRFGAGDLRAAFWLVDLLES ------ AGYSGPRHFDFKPPR-TED

. * * . *

HIDLNGQRGIKYDQDLVFGHGDLTSAFFTVDLLENGFPNGGPKYTGPRHFDYKPSR-TDG

i .ti’ * i
n

53
53
53
6O
51
53
19
19
19

113
113
113
119
111
113
66
66
66

173
173
173
179
171
173
122
122
122

233
233
233
239
231

233

182
182
182

293
293
293
299
291
293
242
242
242

348
348
348
354
346
348
300
295
301

'1‘

 

 

 

 

 

 

>m3wmmm'av-3ra
H a n H‘m n S B

>m>mww~3~3~a
H S 0 Few n D 3

VEDLFIGHIAGMDTFALGFKIAYKLAKDGVFDKFIEEKYRSFKEGIGKEIVEGKTDFEKL 408
VEDLFIGHIAGMDTFALGFKVAYKLVKDGVLDKFIEEKYRSFREGIGRDIVEGKVDFEKL 408
PEDLFLGHIAGMDAFAKGFKVAYKLVKDRVFDKFIEERYASYKDGIGADIVSGKADFRSL 408
PDDLIYAHIAGMDAFARGLKVAHKLIEDRVFEDVIQHRYRSFTEGIGLEIIEGRANFHTL 414
DEDLFHAHIAGMDTYAVGLKVASRLLEDKALDQVIEERYESYTKGIGLEIKEGRTDLKKL 406
KYDLFYGHIGAMDTMALALKIAARMIEDGELDKRIAQRYSGWNSELGQQILKGQMSLADL 408
YDGVWESAKANIRMYLLLKERAKAFRADPEVQEALAASKVAELKTPTLNPGEGYAELLAD 360
FDGVWASAAGCMRNYLILKERAAAFRADPEVQEALRASRLDELARPTA-—ADGLQALLDD 353
YDGVWDSAKANMSMYLLLKERALAFRADPEVQEAMKTSGVFELGETTLNAGESAADLMND 361

‘ : * .:. :
EEYIIDKED-IELPSGK—QEYLESLLNSYIVKTIAELR —————— 444
EEYIIDKET—IELPSGK-QEYLESLINSYIVKTILELR ------ 444
EKYALERSQ-IVNKSGR—QELLESILNQYLFAE ----------- 439
EQYALNHKS—IKNESGR-QEKLKAILNQYILEV ----------- 445
AAYALENDH-IENQSGR-QERLKATVNRYLLNALREAPAGKETH 448
AKYAQEHHLSPVHQSGR-QEQLENLVNHYLFDK ----------- 440
RSAFEDYDADAVGAKGFGFVKLNQLAIEHLLGAR ---------- 394
RSAFEEFDVDAAAARGMAFERLDQLAMDHLLGARG --------- 388

SASFAGFDAEAAAERNFAFIRLNQLAIEHLLGSR ---------- 395

t .

10

CE Ti... 6
C .. -

5...:

 

 

62865 2m $233 .985 meson :32: 98 @053 2?: 3:03 SEES:

.559? as 88593 05 53, new 32 ESE 53> masoEBE 83.60% .395. mo 26 268 05 mo 5335858 ocmfionom ”m 8sz

a

 

 

 

 

 

 

 

 

 

 

 

NMNA

ll

 

 

tetr;
nati'

don

for c
hu\e
grout
grou;
(Whiz
Valid,
ilenkn
of sub:
Slums:-
with Cc
Other 1];
101'. and f
iiCllthe
150mm £184
will the (
Obsm'ed .
is also me

The recombinant TNXI existed as a dimer as well as a tetramer with the ratio of dimer to
tetramer of approximately 20:1. Although the structural features of the recombinant and
native forms of TNXI differ in the degree of subunit assembly, their functional prOperties
do not differ. The dimer is a catalytically viable and stable form of the enzyme.
Typically, two divalent cations (MgZT, Mn2+, or C0,“) per monomer are required
for catalytic activity and stability of XIs. Two distinct metal binding sites, M1 and M2,
have been identified in X13: (1) the metal in site M1 is coordinated to four carboxylate
groups; (ii) the metal in site M2 is coordinated to one imidazole and three carboxylate
groups. These were initially referred to as the structural and catalytic metals, respectively
(Whitlow et al., 1991, and Marg and Clark, 1990). But these assumptions are no longer
valid, since later studies showed that both metals are directly involved in catalysis
(Jenkins et al., 1992 and Allen et al., 1994). Metal specificity depends on both the nature
of substrate and on the enzyme type. Thermus aquaticus XI, a type I enzyme isomerizes
glucose most efficiently in the presence of Mn2+, but its activity toward xylose is highest
with Co2+ (Lehmacher and Bisswanger, 1990). The type II Bacillus coagulans XI, on the
other hand, isomerizes xylose most effectively in the presence of Mn2+, while its activity
toward fructose is highest with Co2+ (Marg and Clark, 1990). In our case, the three metals
activate the TNXI. At any concentration, C02+ is the best activating metal. Glucose
isomerase activity in the presence of Mg2+ is approximately 40 % of the activity observed
with the Co2+ enzyme. Poorly active, the Mn2+ enzyme show only 16 % of the activities
observed with the Coz+ enzyme (Vieille et al., 2001). The thermal stabilization of TNXI
is also metal-specific: the Mn2+ enzyme is significantly more stable than the Co2+ and

Mg2+ enzymes at 101°C.

12

 

r3}
198
enz3
The”!
inter.
in [hi
intem'
3 met;
al., 19‘
(Lee er
for the
molecule
differenci

Vt‘l'SUS X \‘1

THERMC

lSOlaled ”Om

lllllgllf’ “mull“

A mechanism of xylose isomerase catalysis was identified based on results of x-
ray crystallographic studies on Arthrobacter or Streptomyces enzymes (Farber et al.,
1989; Collyer et al., 1990) and biochemical properties exhibited by thermophilic
enzymes obtained by site-directed mutagenesis of the xylA gene from
Thermoanaerobacterium thermosulfurigenes (Lee et al., 1990). The enzymatic
interconversion of aldose to ketose by xylose isomerases involves binding of the substrate
in the ring form, substrate ring opening, isomerization of the linear intermediate,
intermediate ring closure, and release of the product. The isomerization step proceeds by
a metal ion-assisted hydride-shift mechanism (Figure 6) (Farber et al., 1989: Collyer et
al., 1990; Lee et al., 1990), and this step, rather than ring opening, is rate determining
(Lee et al., 1990). D-xylose and D-glucose have identical atomic configuration, except
for the presence of an additional —CHzOH group at the C-6 position in the glucose
molecule. This extra hydroxymethyl group must therefore be responsible for the
differences in the catalytic efficiency exhibited by xylose isomerase toward glucose

versus xylose.

THERMOZYMES AND THEIR THERMAL STABILITY

Thermozymes are enzymes that evolved in thermophiles (organisms thriving at
50—80°C) and hyperthermophiles (organisms thriving at 80°C or above). Most archea and
some bacteria are thermophiles and hyperthermophiles. These organisms have been
isolated from all types of terrestrial and marine environments. Thermozymes developed

unique structure-function properties above 70°C (Vieille and Zeikus, 1996; Vieille and

13

 

*\

Figure
rublgim
COFFEspi

D308 an

 

 

OH OH H K183 OH OH H l K183

 

 

Figure 6: A metal—assisted hydride-shift mechanism of xylose in the Streptomyces
rubiginosus xylose isomerase (type 1 X1). In type II XIs, Both Mn2+ are replaced by Co“.
Corresponding residues of D257 and K183 in the Thermotoga neapolitana XI (TNXI) are

D308 and K233, respectively.

14

 

Zhfll
pol}
e\ce
inch
and;
therr
10pm:
once
Uiltl
and t
conce

reactu

SOlUbl:

used n1

COUnter
StllJCtUr
they 5h.

s.»

C0Unt

er?
menh02\

Zeikus, 2001). They have already been used in molecular biology (e.g. Taq DNA
polymerase), starch-processing (e.g. a-amylases, glucose isomerases) industry, and are
excellent catalytic candidates for several additional applications that require high stability
including organic syntheses, diagnostics, waste treatment, pulp and paper manufacture,
and animal feed (Vieille et al., 1996). Intrinsically stable and active at high temperatures,
thermozymes offer major biotechnological advantages over meSOphilic enzymes
(optimally active at 25-50°C) or psychrophilic enzymes (optimally active at 5-25°C): (i)
once expressed in mesophilic hosts, thermozymes are easier to purify by heat treatment,
(ii) their thermostability is associated with a higher resistance to chemical denaturants,
and (iii) performing enzymatic reactions at high temperatures allow higher substrate
concentrations, lower viscosity, fewer risks of microbial contamination, and higher
reaction rates.

Protein stability has been actively studied for several decades starting with small,
soluble, and monomeric enzymes (e.g. lysozyme and ribonuclease)(Dill, 1990). Having
access to thermozymes allows us to determine what protein stabilization mechanisms are
used in nature to gain extreme stability. Thermozymes are relatively similar to mesophilic
enzymes: (1) their amino acid sequences are 40-85 % similar to those of their mesophilic
counterparts (Vieille et al., 1995 and Burdette et al., 1996), (ii) their three-dimensional
structures are superimposable (Davies et al., 1993 and Fujinaga et al., 1993), and (iii)
they share the same catalytic mechanisms (Vieille et al., 1995 and Voorhorst et al.,
1995). Therefore, their increased stability (as compared with their mesophilic
counterparts) must be a result of differences in specific amino acid sequences. Since

thermozymes are optimally active under more severely denaturing conditions than their

15

 

enZ}:
efficr
oxera

unde-

xleSc

mesophilic counterparts, they need to be more rigid. This increased rigidity is essential
for preserving their active center structures and protect them from unfolding. Such
rigidity is demonstrated by lower hydrogen exchange rates and by lower susceptibility to
proteolytic degradation and chemical denaturant or thermal unfolding (Veronese et al.,
1984, Wrba et al., 1990, Kanaya and Itaya, 1992). Observations from large groups of
enzymes indicate that stabilizing substitutions tend to improve the enzymes’ packing
efficiency (through cavity filling and increase in core hydrophobicity), and to increase
overall enzyme rigidity through helix stabilization, electrostatic interaction optimization
and conformational strain reduction.

Well-known and studied mechanisms of protein thermostabilization (Vieille and
Zeikus, 1996; Vieille and Zeikus, 2001) are (i) hydrophobic interactions, (ii) packing
efficiency and reduction in solvent-accessible hydrophobic surface, (iii) aromatic
interactions, (iv) salt bridges, (v) disulfide bridges, (vi) hydrogen bonds, (vii) metal
binding, (viii) reduction of conformational strain, (ix) reduction of the entropy of
unfolding by proline substitution, (x) helix stabilization, (xi) stabilization of loops (xii)
intersubunit interactions and oligomerization, (xiii) resistance to covalent destruction, and

(xiv) post-translational modifications.
GENETIC AND PROTEIN ENGINEERING OF Xls
Thermostability mutations

A cluster of aromatic amino acid residues is present in the active site pocket of

xylose isomerases isolated from different sources and the hydrophobic interactions

16

 

 

(111101
mum
t’l al.
residt

them;

heat I
mutan
import
subsUt
by 43J
protrud
reducer
tl1501105
increase
TINSAJ
fullCllOn
Unchang
Wildqypc
A
(199]) by
residues_ 5

from Chem

among these aromatic amino acids were postulated to be one of the important forces that
maintains the monomers associated into active dimers (Collyer et al., 1990 and Whitlow
et al., 1991). Meng et a1. (1993) explored the thermostability of mutants of aromatic
residues (Trp48, PheS9, Trpl38, Phel44, and Trp187) in Thermoanaerobacterium
thermosulfurigenes XI) with Co2+ as a co-factor.

The Trp187His and Phel44Lys mutant enzymes were no longer resistant to the
heat treatment at 75°C, which was one of the purification steps of the wild type and
mutant enzymes. This indicates that the hydrophobic character of Trp187 and Phe144 is
important in maintaining the enzyme structure at high temperature. The Trp138
substitutions with smaller hydrophobic residues had increased the thermostability at 85°C
by 43-91 % without affecting the enzymes’ activities. Since the indole group of Trp138
protrudes into the active site cavity, replacement of Trp138 with Phe, Met, or Ala
reduced the area of active site hydrophobic surface and therefore enhanced
thermostability. Trp48Arg mutation did not change the activity of the enzyme but it
increased the stability of the enzyme by 60 %. The enhancement of thermostability in the
Trp48Arg mutant enzyme was brought about because Arg48 could presumably fulfill the
function of hydrogen bonding to Asp338 and thus leave the active site structure
unchanged. The Phe59His mutant was relatively less stable with shorter half-live than the
wild-type enzyme.

A different approach to improving thermostability was adopted by Quax et al.
(1991) by converting lysine residues in Actinoplanes missouriensis XI (AMXI) to Arg
residues. Special attention was given to preventing deleterious effects that could result

from chemical modification of xylose isomerase by sugar components of high fructose

l7

 

C011
AM

inte

inter
inert
impt‘
of th
uhicl

explu.

MI li

solutic

isomer;
Compdr.
Actinop.
thermW
thennmL
(ii A.
Tc

1 ii) A

the

corn syrup, and in particular, from non-enzymatic glycation of key amino groups. Each
AMXI subunit contains 20 lysine residues and two of them are located at the dimer/dimer
interface, Lys252 and Ly5294 (Gly304 and Lys340 in TNXI, respectively).

Lys252 is buried in the XI tetramer and is also involved in electrostatic
interactions across the interdimer interface. The mutation Lys252Arg resulted in a 30 ‘70
increased in thermostabilty of the Mg-enzyme in solution with no loss in activity. More
importantly, the half-life of the immobilized mutant at 70°C was 3-fold longer than that
of the wild type. Lys252 would be accessible for glycation when the dimer dissociates,
which would inactivate the enzyme by preventing reassociation of the dimer. This would
explain why the immobilized mutant enzyme is more resistant to glucose inactivation. .

Lys294 forms a salt bridge with the M1 ligand Asp256 (Asp308 in TNXI) and the
M2 ligand Asp292 (Asp338 in TNXI). Lys294Arg had decreased thermostability in
solution but very little after immobilization.

Chang et al. (1999) studied the structures of highly thermostable type I xylose
isomerases from Thermus thermophilus (TthXI) and Thermus caldophilus (TcaXI)
compared to those of less thermostable XIs from Arthrobacter B3728 (AXI) and
Actinoplanes missouriensis (AMXI). Analyses of various factors that may affect protein
thermostability indicate that the possible structural determinants of the enhanced
thermostability of TcaXI/TthXI over AXI/AMXI are
(i) An increase in ion pair networks: the total ion pairs per tetramer of

TcaXI/TthXI/AXI/AMXI are 150/ 1 63/93/ 1 07, respectively.

(ii) A decrease in the large intersubunit cavities: there is a clear correlation between

the decreased number of large internal cavities and thermostability. The total

18

 

lllll

m)

(iii)

(iV)

volume and surface area of cavities are 389/373/896/591 A3 and
938/887/ 1702/ 1 136 A2 for TcaXI/TthXI/AXI/AMXI, respectively. All cavities in
TcaXI/TthXI tetramers are small, whereas some of the cavities in AXI/AMXI
tetramers are very large. The volume and surface area of the largest cavities in
TcaXI/TthXI/AXI/AMXI are 38/30/193/122 A3 and 74/64/304/207 A3.

A removal of potential deamidation and isoaspartate formation sites: Deamidation
and isoaspartate formation have been found to be one of the important processes
leading to irreversible heat denaturation of protein at neutral pH (Ahern and
Klibanov, 1985 and Aswad, 1990). They occur frequently when the sequences
Asn-Gly, Asn-Ser, and Asp-Gly lie in highly flexible regions of the polypeptide
chain. The analysis indicated that they are much fewer in TcaXI/TthXI (2/1)
compared to AXI/AMXI (7/7).

Shortened loops and Proline residues: TcaXI and TthXI have a significantly
shortened loop due to deletions, compared with AXI and AMXI. Compared with

AXI and AMXI, TcaXI and TthXI have approximately five more proline residues.

A site-directed mutagenesis study of His219 residue of Streptomyces rubiginosus

XI (His270 in TNXI) and its effect on activity and thermostability was done by Cha et al.

(1994). This residue is conserved in all xylose isomerases. The three dimensional

structure of the enzyme revealed that His219 is part of the octahedral coordination sphere

of M2, One of the two metal ions (MnZT) in the active site. Substitutions of His219 with

Ser, Glu, and Asn resulted in enzymes with the kcat values of only 0.3—0.5 % of that of the

wild-type enzyme. The Km values of these mutant enzymes increased by 30-40 fold over

19

 

the \V

actix ii

enZ} it
stable
et‘t‘ecti
interat
betu e
structt
carbo
“3511
Coord
medst

1’6 Spry

reflex
enZ\j

Cfih

the wild-type value. The His219Lys mutant enzyme did not exhibit any measurable
activity.

Thermal denaturation studies indicated that the His219Ser and His 219Asn mutant
enzymes are 5-8°C less stable, whereas Hi5219Glu and Hi5219Lys are 13-24°C less
stable than the wild-type enzyme. In the His219Ser structure, a water molecule
effectively replaced the Ne-2 atom of the imidazole ring of His219 and mediated the
interaction between Mn2+ at the M2 site and Ser219. A similar water-mediated interaction
between the metal ion and Asn219 was observed in the His219Asn mutant enzyme
structure. On the other hand, no direct or water mediated interactions between the
carboxyl group of Glu219 and the metal were observed. Whereas octahedral coordination
was maintained for the metal at the M2 site in His219Ser and Hi5219Asn, a pentahedral
coordination with the metal at the M2 site was observed in His219Glu. Metal activation
measurements supported the observation that metal binding is perturbed and is

responsible for thermal lability of His219 mutant enzymes.

Active site mutations

Most active site mutations reduce or destroy activity but some are also relevant to
enzyme thermostability (Hartley et al., 2000). However, Meng et al. (1990) successfully
used site-directed mutagenesis to switch the substrate preference of
Themloanaerobacterium thermosulfurigenes XI (TTXI) from xylose to glucose by
redesigning its substrate binding pocket. In the Arthrobacter enzyme, the structure of the
enzyme complex with the six carbon competitive inhibitor, sorbitol, indicated that the

C6-hydroxymethyl group of the substrate (glucose) is oriented toward the bottom of the

20

 

\kfll
'Ihrl

sntul

deem

for x

substrate-binding pocket and is adjacent (3.4-3.7 A) to the residues Met87, Thr89, and
Va1134. These residues are highly conserved among X15 and correspond to Trpl38,
Thr140, and Va1185, respectively, in the TTXI. Each of these residues was replaced with
smaller amino acids to prove that they are part of the substrate-binding pocket in TTXI.
Replacement of Trp138 with Phe, a smaller residue, produced an enzyme that had a
higher catalytic efficiency (kw/Km) for glucose than the wild-type enzyme due to both a
decrease in Km and an increase in km. On the other hand, this mutation increased the Km
for xylose and decreased the catalytic efficiency for this substrate. This was suggested to
be due to a more spacious pocket and this increased the freedom of movement of the
xylose molecule, decreasing its binding efficiency. The enlargement of the binding
pocket also decreased binding energy between the enzyme and the transition state
resulting in the decrease of km.

The Vall85Thr mutant enzyme had a slightly lower Km and a higher kcat for
glucose. Substitution with a Ser residue, which has a smaller sidechain but otherwise is
equivalent to Thr, did not improve the glucose catalytic efficiency. Likewise, replacement
with Ala did not significantly change either Km or kw. Thus these data suggested that
Va1185 does not hinder glucose binding, but the Thr substitution may provide an
additional hydrogen bonding, presumably to the C6-OH group of glucose.

Replacement of Thr140 with Ser increased the KM for both xylose and glucose
and resulted in lower catalytic efficiency for glucose. It can be concluded that Thr140
hydrogen bonds to the substrate but does not strictly hinder the binding of glucose. The
double mutant enzymes Trp138Phe/Vall85Thr and Trp138Phe/Vall85Ser had a higher

catalytic efficiency for glucose than the wild-type enzyme of 5- and 2-fold, respectively.

21

 

'Ihc
rest

isor

I)lF

llbfar
appro
”56d 5
[hmllg

Vumu

They also exhibited 1.5-and 3-fold higher catalytic efficiency for glucose than for xylose,
respectively, thus they can be called the real “glucose isomerase” rather than xylose

isomerase.

DIRECTED EVOLUTION OF ENZYMES

Directed evolution, also termed evolutionary engineering, has recently emerged as
a key technology for biomolecular engineering and generating impressive results in the
functional adaptation of enzymes to artificial, non-natural environments (Reetz and
Jaeger, 1999; Pluckthun et al., 2000; and Wintrode and Arnold, 2000). The approach is
highly attractive because its principles are simple and do not require detailed knowledge
of structure, function, or mechanism. Essentially like natural (Darwinian) evolution,
directed evolution comprises the iterative implementation of (1) the generation of a
“library” of mutated genes, (2) its functional expression, and (3) a sensitive assay to
identify individuals showing the desired properties, either by selection or by screening.
After each round, the genes of improved variants are deciphered and subsequently serve
as parents for another round of optimization (Brakmann, 2001).

A series of experimental strategies have been developed for generating mutant
libraries in the laboratory which differ in diversity. They can be divided into two main
approaches, random mutagenesis and recombination. Random mutagenesis is a widely
used strategy, which target whole genes. This may be achieved by passing cloned genes
through mutator strains (Cox, 1976; and Greener et al., 1996), by treating DNA with

various chemical mutagens (Shortle and Nathans, 1978; Kadonaja and Knowles, 1985;

22

 

and Dc'
19921.
emerg-
per nt
alterat
suhsu
11111111
and t
perm
lli'ldl‘

leId

COW

rec‘
emI
Seq-
Pare

al..

and Deshler, 1992), or by error-prone PCR (Leung et al., 1989; and Cadwell and Joyce,
1992). Due to its simplicity and versatility, error-prone PCR (random PCR mutagenesis)
emerged as the most common technique, which can result in mutation rate as high as 2 %
per nucleotide position. The mutation rate may also be adjusted to lower values with
alterations of PCR conditions. However, only a limited number of amino acid
substitutions is accessible by this method since this reaction biases the distribution of
mutation type in favor of transitions (A(—>G and T<——>C) over transversions (A/GHC/T),
and because multiple substitutions within a single codon are extremely rare. Complete
permutation of a single amino acid position (saturation mutagenesis) may enable the
finding of non-conservative replacements that are inaccessible by random PCR
mutagenesis (Miyazaki and Arnold, 1999).

Recombination of DNA represents an alternative or additional approach for
generating genetic diversity that is based on mixing and concatenation of genetic material
from a number of parent sequences. Recombination may be advantageous in
concentrating beneficial mutations which have arisen independently and may be additive,
and likewise, in concentrating deleterious mutations which subsequently might be more
efficiently purged from the population by selection (Zeyl and Bell, 1997; and Moore and
Maranas, 2000). DNA shuffling was the first technique introduced for random in vitro
recombination of gene variants created by random mutagenesis (Stemmer, 1994). It
employs the PCR reassembly of whole genes from a pool of short overlapping DNA
sequences (SO-300 bp) generated by random enzymatic fragmentation of different
parental genes. Alternative protocols include staggered extension process (StEP) (Zhao et

al., 1998) and random-priming recombination (Shao et al., 1998). StEP recombination is

23

 

a PC
of p
stru
prir
for

€\lt

0f

the

CU

ph
bu

EX

a PCR reaction with very short annealing and extension steps that promote the formation
of premature extension products. The truncated strands may anneal randomly to a parent
strand, thus combining the different sequences from different parental strands. Random-
priming recombination, as an alternative to DNA shuffling, produces random fragments
for reassembly by annealing of short, random primers to a certain template gene and
extension by a polymerase.

In directed enzyme evolution, diversity is created on the DNA level, but selection
or screening acts on the level of the encoded protein. Therefore, functional expression of
the DNA libraries is a necessary prerequisite for the detection of improved enzyme
variants. The most common approaches for recombinant protein expression employ the
cellular transcription/translation machineries of well-established organisms such as E.
coli, S. cerevisiae, or B. subtilis. Alternatively, a physical link between genotype and
phenotype may be established by generating fusions between the protein of interest and a
bacteriophage coat protein. Following intracellular assembly, recombinant phages
express the protein variants on their surface while enclosing the appertaining genetic
information within their genome (Johnson and Ge, 1999; and Smith and Petrenko, 1997).

The most challenging step in directed evolution experiments is to develop a
screening or selection scheme that is sensitive to the desired properties. Selection can be
used either in vivo or in vitro. In vivo selection is most often achieved by genetic
complementation of hosts that are deficient in a certain pathway or capacity. In vitro
enrichment procedures that are detached from cell survival may also be termed selection.
These techniques have been developed for the biopanning of phage-displayed peptide

libraries by binding to a ligand that is immobilized on an appropriate column matrix

24

 

enr

alt:

&5\

Oil

*3
E;

mitt
mo
COT.

6E2

dire

Oil:

Tegl

lFlt;

(Brakmann, 2001). Recently, the approach has also been applied to the selective
enrichment of phage-displayed functional enzyme libraries. Screening is an important
alternative to selection. It enables a better control of the applied constraints, and also is
more versatile, predominantly in unnatural environments, or with unnatural substrates. It
usually requires that the mutant libraries are diluted and well-distributed which can be
achieved by conventional plating of transfonnants on agar plates or filter membranes.
This time-consuming step is sometimes accelerated using robotic systems. Common
assays are based on visual or spectroscopic detection, for example formation, alteration,
or destruction of colors or fluorescence characteristics. The determination of the optical
parameters can also be accomplished by using automatic plate-readers, which enable a
normalization of measured values to respective cell densities and may also be used to
monitor the reaction kinetics. It should be noted that it is important to choose selective
constraints that precisely reflect the desired property since the first law of directed
enzyme evolution is “you get what you screen for” (Schmidt-Dannert and Arnold, 1999).
During the past few years, many enzymes have been successfully improved by
directed evolution (Table 1). The narrow range of substrates accepted by natural enzymes
often prevents their use in new synthetic and commercial applications. Thus, by far most
results were efficient tuning of catalytic efficiency toward non-natural substrates.
Thermostability of enzymes and enantioselectivity of specific bioconversions have also
been improved by these approaches. These examples showed that directed evolution is a
powerful and reliable tool for improving biocatalysts in reasonably short periods of time.
The only directed evolution approach (or random mutagenesis) performed on

xylose isomerase reported to date was by Lonn et al. (2002). The thermophilic Thermus

25

N

r.

\\

th

 

re

211‘:

21d

ter

thermophilus XI was subjected to one round of random PCR mutagenesis. It was
screened for xylose isomerase activity at lower temperature than optimal by expression of
the mutated genes in E. coli and replica—plated on McConkey agar plates, complemented
with 1 % xylose followed by incubation at 30°C for 2 days. Three transformant colonies
that were deeper red (suggesting higher XI activity) than the wild—type were selected and
further characterized. Three amino acid substitutions were identified as Phe163Leu in
domain I (C-terminal tail), and Glu372Gly/Val379Ala in domain 11 ([Ol/Bls barrel). These
mutant enzymes showed improved catalytic rate constants (km) by up to nine times on
both xylose and glucose with up to 26 times higher Km values on xylose but relatively
unchanged for glucose. All enzyme variants’ relative activities on xylose are higher than
the wild-type at low temperatures. These results suggested that amino acid substitutions
distant from the catalytic center could lead to cold adaptation. There is a close
relationship between molecular flexibility and function. Thermophilic enzymes are rigid
and require high temperatures in order to gain sufficient molecular flexibility for activity.
Their molecular structure must therefore be balanced between the requirements for
stability and dynamics. They suggested that the sequence changes underlying the
adaptation of T. thermophilus XI variants to temperatures lower than their optimal
temperature, allow a higher degree of flexibility in areas that move during catalysis.
Kinetic analysis demonstrated that the increase in the relative activity in the enzyme
variants for xylose at low temperatures was indeed caused by an increase in kcal and not
by a decrease in the Km value. This suggests that the mutant enzymes did not acquire

higher affinity for the substrate than the wild-type enzyme at lower temperatures.

26

 

1n
and with
Ilienm in )2
directed m
COUlp‘dl’lWl
Tllr’mluum.
themiostah
glucose at
biochemica.
commercial.

industrial up

In this study, the molecular determinants that are responsible for thermostability
and activity toward glucose of a xylose isomerase from a hyperthermophilic eubacterium
Thermotoga neapolitana (TNXI) will be identified. The method used in such work is site-
directed mutagenesis based on the amino acid sequence and three dimensional structure
comparisons between TNXI and a xylose isomerase from a thermophilic eubacterium
Thermoanaerobacterium thermosulfurigenes (TTXI). Furthermore, the highly
thermostable TNXI will be engineered by directed evolution to improve its activity on
glucose at low temperature and pH. Finally, the resulting laboratory-evolved enzyme’s
biochemical properties and fructose productivity will be compared to those of the
commercially available glucose isomerase, GensweetTM, to further ascertain its utility in

industrial applications.

27

Table 1: Examples of enzymes that were successfully optimized using directed evolution.

 

 

 

 

activity in organic
solvent

+ screening

Target enzyme Target property Change evolved Approach Reference
subtilisin E activity in organic ~ 170-fold increase error-prone PCR + Chen &
solvents in 60% screening Arnold
dimethylformamide (1993);
Arnold &
Chen (1994)
B-lactamase activity towards 32,000-fold greater DNA shuffling + Stemmer
new substrate resistance to selection (1994)
cefotaxime
para-nitrobenzyl activity towards 60- 150 fold increase error-prone PCR Moore &
esterase pNB esters; and DNA shuffling Arnold

(1996): Moore
et al. (1997);

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Arnold &
Moore (1998)
B-galactosidase activity towards 66-fold increased DNA shuffling + Zhang et al.
new substrate; activity; 1000-fold screening (1997)
substrate increase in substrate
specificity specificity
aminoacyl-tRNA aminoacylation of 55-fold increase in DNA shuffling + Liu et al.
synthetase a modified tRN A activity selection (1997)
aspartate activity towards B 105 increase DNA shuffling + Yano et al.
aminotransferase branched amino selection (1997)
and 2-oxo acids
lipase enantioselectivity increase in error-prone PCR + Reetz et al.
in hydrolysis of p- enantiomeric excess screening ( 1997)
nitrophenyl 2- from 2% to 81%
methyldecanoate
pNB esterase thermostability 14 0C increase in Tm error-prone PCR, Giver er al.
+ increased activity DNA shuffling + (1998)
at all temperatures screening
subtilisin E thermostability 17 °C increase in Tm error-prone PCR, Zhao &
+ increased activity DNA shuffling + Arnold (1999)
at all temperatures screening
subtilisin BPN’ activity at 10°C 2-fold increase chemical Taguchi et al.
mutagenesis + (1998)
screening
cephalosporinases activity towards 270-540-fold DNA shuffling of Crameri et al.
moxalactam increased resistance homologous genes (1998)
+ selection
kanamycin thermostability increase 20°C DNA shuffling + Hoseki
nucleotidyl screening/selection et al. (1999)

 

28

 

.1wl w. l_.l

_ll

 

transferase

 

 

 

 

 

 

 

B. remove fructose active without FBP random Allen &
stearothermophilus 1,6 bisphosphate mutagenesis + Holbrook
LDH requirement (FBP) screening (2000)
phospholipase A, thermostability increase Tm by 1 1°C error prone PCR + Song & Rhee
without screening (2000)
compromising
activity
TEM-l [3- activity towards 20,000 fold increase high frequency Zaccolo et a1.
lactamase cefotaxime random ( 1999)
mutagenesis
myoglobin peroxidase activity 25-fold increase error-prone PCR + Wan et al.
screening (1998)
hydantoinase enantioselectivity inverted error-prone PCR + May et al.
+ total activity enantioselectivity, 3 screening (2000)
x increase in total
activity
xylose isomerase activity at low improved kcan at low error prone PCR + Lonn er
temperature temperatures screening a1.(2002)

 

 

 

 

 

29

 

10.

ll.

12.

13.

14.

REFERENCES

. Allen, K. N ., Lavie, A., Glasfeld, A., Tanada, T. N., Gerrity, D. P., Carlson, S. C.,

Farber, G. K., Petsko, G. A., and Ringe, D. (1994) Isotopic exchange plus substrate
and inhibitor kinetics of D-xylose isomerase do not support a proton-transfer
mechanism. Biochemistry 33, 1481-1487.

Allen, S. J., and Holbrook, J. J. (2000) Production of an activated form of Bacillus
stearothermophilus L-2-hydroxyacid dehydrogenase by directed evolution. Protein
Eng. 13, 5-7.

Arnold, F. H., and Chen, K. (1994) US patent 5,316,935.
Arnold, F. H., and Moore, J. C. (1998-1999) US patents 5,741,691 and 5,906,930.
Barker, S. A. (1976) Pure fructose syrups. Process Biochem. 11, 20-25.

Bhosale, S. H., Rao, M. B., and Deshpande, V. V. (1996) Molecular and industrial
aspects of glucose isomerase. Microbial. Reviews 60, 280- 300.

Brakmann, S. (2001) Discovery of superior enzymes by directed molecular evolution.
Chembiochem. 2, 865-871.

Burdett, D. S., Vieille, C., and Zeikus, J .G. (1996) Cloning and expression of the gene
encoding the Thermoanaerobactor ethanolicus 39E secondary-alcohol dehydrogenase
and biochemical characterization of the enzyme. Biochem. J. 316, 115-122.

Cadwell, R. C., and Joyce, G. F. (1992) Randomization of genes by PCR
mutagenesis. PCR Methods Appl. 2, 28-33.

Cha, J., Cho, Y., Whitaker, R. D., Carrell, H. L., Glusker, J. P., Karplus, P. A., and
Batt, C. A. (1994) Perturbing the metal site in D-xylose isomerase: effect of
mutations of His-220 on enzyme stability. J. Biol. Chem. 269, 2687-2694.

Chang, C., Park, B. C., Lee, D. -S., and Suh, S. W. (1999) NCBI Protein Database
CAB 4930288.

Chen, K., and Arnold, F. H. (1993) Tuning the activity of an enzyme for unusual
environments: sequential random mutagenesis of subtilisin E for catalysis in
dimethylforrnamide. Proc. Natl. Acad. Sci. USA 90, 5618-5622.

Collyer, C. A., Henrick, K., and Blow, D.M. (1990) Mechanism for aldose-ketose
interconversion by D-xylose isomerase involving ring opening followed by a 1,2-

hydride shift. J. Mol. Biol. 212, 21 1-235.

Cox, E. C. (1976) Bacterial mutator genes and the control of spontaneous mutation

3O

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Ann. Rev. Genet. 10, 135-156.

Crameri, A., Raillard, S. A., Berrnudez, E., and Stemmer, W. P. C. (1998) DNA
shuffling of a family of genes from diverse species accelerates directed evolution.
Nature 391, 288-291.

Davies, G. J., Gamblin, S. J ., Littlechild, J. A., and Watson, H. C. (1993) The
structure of a thermally stable 3-phosphoglycerate kinase and a comparison with its
mesophilic equivalent. Protein Struct. F unct. Genet. 15, 283-289.

de Raadt, A., Ebner, M., Ekhart, C. W., Fechter, M., Lechner, A., Stobl, M., and
Stutz, A. E. (1994). Glucose isomerase (E.C.5.3. 1.5) as a reagent in carbohydrate
synthesis: success and failures with the isomerisation of non-natural derivatives of D-
glucose into the corresponding 2-ketoses. Catalysis Today 22, 549-561.

Deshler, J. O. (1992) A simple method for randomly mutating cloned DNA fragments
by using chemical mutagens and the polymerase chain reaction. Gen. Anal. Techn.
Appl. 9, 103-106.

D111, K. A. (1990) Dominant forces in protein folding. Biochemistry 29, 7133-7155.

Farber, G. K., Glasfeld, A., Tiraby, G., Ringe, D., and Petsko, G. A. (1989)
Crystallographic studies of the mechanism of xylose isomerase. Biochemistry 28,
7289-7297.

Fujinaga, M., Berthet,-Colominas, C., Yaremchuk, A. D., Tukalo, M. A., and Cusak,
S. (1993) Refined cr stal structure of the seryl-tRNA synthetase from Thermus
thermophilus at 2.5 resolution. J. Mol. Biol. 234, 222-233.

Ghatge, M. S., Phadatare, S. U., Bodhe, A. M., Deshpande, V. V. (1994) Unfolding
and refolding of glucose/xylose isomerase from Streptomyces sp. NCIM 2730.
Enzyme Microb. Technol. 16, 323-327.

Giver, L., Gerhenson, A., Freskgard, P. O., and Arnold, F. H. (1998) Directed
evolution of a thermostable esterase. Proc. Natl. Acad. Sci. USA 95, 12809-12813.

Greener, A., Callaghan, M., and Jerpseth, B. (1996) An efficient random mutagenesis
technique using an E. coli mutator strain. Methods Mol. Biol. 57, 375-385.

Hartley, B. S., Hanlon, N., Jackson, R. J., and Rangarajan, M. (2000) Glucose
isomerase: insights into protein engineering for increased thermostability. Biochim.
Biophys. Acta. 1543, 294-335.

Hess, M. J., Tchemajenko, V., Vieille, C., Zeikus, J. G., and Kelly, R. M. (1998)

Thermotoga neapolitana homotetrameric xylose isomerase is expressed as a
catalytically active and thermostable dimer in Escherichia coli. Appl. Environ.

31

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Microbiol. 64, 2357-2360.

Hoseki, J., Yano, T., Koyama, Y., Kuramitsu, S., and Kagamiyama, H. (1999)
Directed evolution of thermostable kanamycin-resistance gene: a convenient selection
marker for Thermus thermaphilus. J. Biachem.(Takya) 126, 951-956.

Jenkins, J., Janin, J., Rey, F., Chiadmi, M., Van Tilbeurgh, H., Lasters, I., DeMaeyer,
M., VanBelle, D., Wodak, S. J ., Lauwereys, M., Stanssens, P., Mrabet, N. T.,
Snauwaert, J ., Matthyssens, G., and Lambeir, G. A. -M. (1992) Protein engineering
of xylose (glucose) isomerase from Actinaplanes missouriensis .1. Crystallography
and site directed mutagenesis of metal binding sites. Biochemistry 31, 5449-5458.

Johnson, K., Ge, L. (1999) Phage display of combinatorial peptide and protein
libraries and their application in biology and chemistry. Curr. Top. Microbial.
Immunol. 243, 87-105.

Kadonja, J. T., and Knowles, J. R. (1985) Nucleic Acids Res. 13, 1733-1745.

Kanaya, S., and Itaya, M. (1992) Expression, purification, and characterization of a
recombinant ribonuclease H from Thermus thermaphilus HB8. J. Biol. Chem. 267,
10184-10192.

Lehmacher, A., and Bisswanger, H. (1990) Comparative kinetics of D-xylose and D-
glucose isomerase activities of the D-xylose isomerase from Thermus aquaticus H88.
Biol. Chem. 371, 527-536.

Lee, C., Bagdasarian, M., Meng, M., and Zeikus, J. G. ( 1990a) Catalytic mechanism
of xylose (glucose) isomerase from Clastridium thet'mosulfurafenes. J. Biol. Chem.
265, 19082-19090.

Leung, D. W., Chen, E., and Goeddel, D. V. (1989) Systems for heterologous gene
expression. Technique 1, 11-15.

Liu, D. R., Magliery, T. J., Pastemk, M., and Schultz, P. G. (1997) Engineering a
tRNA and aminoacyl-tRNA synthetase for the site-specific incorporation of unnatural
amino acids into proteins in vivo. Proc. Natl. Acad. Sci. USA 94, 10092-10097.

Lonn, A., Gardonyi, M., van Zyl, W., Hahn-Hagerdal, B., and Otero, R. C. (2002)
Cold adaptation of xylose isomerase from Thermus thermaphilus through random
PCR mutagenesis. Eur. J. Biochem. 269, 157-163.

Marg, G. A., and Clark, D. S. (1990) Activation of the glucose isomerase by divalent
cations: evidence for two distinct binding sites. Enzyme Microb. Technol. 12, 367-

373.

May, 0., Nguyen, P., and Arnold, F. H. (2000) Inverting enantioselectivity by

32

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

directed evolution of hydantoinase for improved production of L-methionine. Nat.
Biotechnol. 18, 317-320.

Meng, M., Lee, C., Bagdasarian, M., and Zeikus, J. G. (1991) Switching substrate
preference of thermophilic xylose isomerase from D-xylose to D-glucose by
redesigning the substrate binding pocket. Proc. Natl. Acad. Sci. USA 88, 4015-4019.

Meng, M., Bagdasarian, M., and Zeikus, J. G. (1993) Thermal stabilization of xylose
isomerase from Thermoanaerabacterium thermosulfurigenes. Proc. Natl. Acad. Sci.
USA 90, 8459-8463.

Miyazaki, K., and Arnold, F. H. ( 1999) Exploring nonnatural evolutionary pathways
by saturation mutagenesis: rapid improvement of protein function. J. Mal. Eval. 49,
716-720.

Moore, G. L., and Maranas, C. D. (2000) Modeling DNA mutation and
recombination for directed evolution experiments. J. Theor. Biol. 205, 483-503.

Moore, J. C., and Arnold, F. H. (1996) Directed evolution of a para—nitrobenzyl
esterase for aqueous-organic solvents. Nat. Biotechnol. 14, 458-467.

Moore, J. C., Jin, H. M., Kuchner, O., and Arnold, F. H. (1997) Strategy for the in
vitro evolution of protein function: enzyme evolution by random recombination of
improved sequences. J. Mal. Biol. 272, 336-347.

Pluckthun, A., Schaffitzel, C., Hanes, J ., and Jermutus, L. (2000) In vitro selection
and evolution of proteins. Adv. Prat. Chem. 55, 367-403.

Quax, W. J., Mrabet, N. T., Luiten, R. G. M., Schuurheizen, P. W., Stanssens, P., and
Lasters, I. (1991) Enhancing the thermostability of glucose isomerase by protein
engineering. Biotechnology 9, 738-742.

Reetz, M. T., Zonta, A., Schimossek, K., Liebeton, K., and Jaeger, K. E. (1997)
Directed evolution of an enantioselective lipase. Angew. Chem. Int. Ed. Eng]. 36,
2830-2832.

Reetz, M. T., and Jaeger, K. E. (1999) Directed evolution of enantioselective
enzymes for organic chemistry. Top. Curr. Chem. 200, 32-57.

Schmidt-Dannert, C., and Arnold, F. H. (1999) Directed evolution of industrial
enzymes. Trends Biotechnol. 17, 135-136.

Shao, Z., Zhao, H., Giver, L., and Arnold, F. H. (1998) Random priming in vitro

recombination: an effective tool for directed evolution. Nucleic Acids Res., 26, 681-
683.

33

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

Shortle, D., and Nathans, D. (1978) Local mutagenesis: a method for generating viral
mutants with base substitutions in preselected regions of the viral genome. Proc. Natl.
Acad. Sci USA 75, 2170-2174.

Smith, G. P., and Petrenko, V. A. (1997) Phage display. Chem. Rev. 97, 391-410.

Song, I. K., and Rhee, J. S. (2000) Simultaneous enhancement of thermostability and
catalytic activity of phospholipase A(l) by evolutionary molecular engineering. Appl.
Environ. Microbiol. 66, 890-894.

Stemmer, W. (1994) Rapid evolution of a protein in vitro by DNA shuffling Proc.
Natl. Acad. Sci. USA, 91, 10747-10751.

Taguchi, S., Ozaki, A., and Momose, H. (1998) Engineering of a cold-adapted
protease by sequential random mutagenesis and screening system. Appl. Environ.
Microbiol. 64, 492-495.

Vangrysperre, W., Callens, M., Kersters-Hilderson, H., and De Bruyne, C. K. ( 1988)

Evidence for an essential histidine residue in D-xylose isomerases. Biochem. J. 250,
153-160.

Veronese, F. M., Boccu, E., Schiavon, O., Grandi, C., and Fontana, A. (1984) General
stability of thermophilic enzymes: studies on 6-phosphogluconate dehydrogenase
from Bacillus stearathermophilus and yeast. J. Appl. Biochem. 6, 39-47.

Vieille, C., Hess, M., Kelly, R. M., and Zeikus, J. G. (1995) xylA cloning and
sequencing and biochemical characterization of xylose isomerase from Thermataga
neapolitana. Appl. Environ. Microbiol, 61, 1867—1875.

Vieille, C., Burdette, D. S., and Zeikus, J. G. ( 1996) Thermozymes. Biatechnol.
Annu. Rev. 2, 1-83.

Vieille, C. and Zeikus, J. G. (1996) Thermozymes: identifying molecular
determinants of protein structural and functional stability. Trends Biotechnol. 14,
183-190.

Vieille, C., and Zeikus, J. G. (2001) Hyperthermophilic enzymes: sources, uses, and
molecular mechanisms for thermostability. Microbial. Mol. Biol. Rev. 65, 1-43.

Vieille, C., Sriprapundh, 8., Kelly, R. M., and Zeikus, J. G. (2001) Xylose
isomerases from Thermataga. Methods Enzymal. 330, 245-249.

Voorhorst, W. G. B., Eggen, R. I. L., Luesink, E. J., and de Vos, W. M. (1995)
Characterization of the celB gene coding for B-glucosidase from the
hyperthermophilic archaeon Pyracaccusfuriasus and its expression and site—directed
mutation in Escherichia cali. J. Bacterial. 177, 7105-71 1 1.

34

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

Wan, L., Twitchett., M. B., Eltis, L. D., Mauk, A. G., and Smith, M. (1998) In vitro
evolution of horse heart myoglobin to increase peroxidase activity. Proc. Natl. Acad.
Sci. USA 95, 12825-12831.

Whitlow, M., Howard, A. J., Finzel, B. C., Poulos, T. L., Winbome, E., and
Gilliland, G. L. ( 1991) A metal-mediated hydride shift mechanism for xylose

isomerase based on the 1.6 A Streptomyces rubiginasus structures with xylitol and
D-xylose. Proteins 9, 153—173.

Wintrode, P. L., Miyazaki, K., and Arnold, F. H. (2000) Cold adaptation of a
mesophilic subtilisin-like protease by laboratory evolution. J. Biol. Chem. 275,
31635-31640.

Wrba, A., Schweiger, A., Schultes, V., Jaenicke, R., and Zavodsky, P. (1990)
Extremely thermostable D-glyceraldehyde-3-phosphate dehydrogenase from the
Eubacterium Thematoga maritima. Biochemistry 29, 7584-7592.

Yano, T., Oue, S., and Kagamiyama, H. (1998) Directed evolution of an aspartate
aminotransferase with new substrate specificities. Proc. Natl. Acad. Sci. USA 95, .
5511-5515.

Zaccolo, M., and Gherardi, E. (1999) The effect of high-frequency random
mutagenesis on in vitro protein evolution: a study on TEM-l beta-lactamase. J. Mal.
Biol. 285, 775-783.

Zeyl, C., and Bell, G. (1997) The advantage of sex in evolving yeast populations.
Nature 388, 465-468.

Zhang, J-H., Dawes, G., and Stemmer, W. P. C. (1997) Directed evolution of a
fucosidase from a galactosidase by DNA shuffling and screening. Proc. Natl. Acad.
Sci. USA 94, 4504-4509.

Zhao, H., Giver, L., Shao, Z., Affholter, J ., and Arnold, F. H. (1998) Molecular
evolution by staggered extension process (StEP) in vitro recombination. Nat.

Biotechnol., 16, 258-261.

Zhao, H., and Arnold, F. H. (1999) Directed evolution converts subtilisin E into a
functional equivalent of therrnitase. Protein Eng. 12, 47-53.

35

CHAPTER II

ROLE OF PROLINE IN XYLOSE ISOMERASE THERMAL STABILITY:
COMPARATIVE THERMOSTABILITY OF A THERMOPHILIC
Thermaanaerobacterium thermasulfurigenes XYLOSE ISOMERASE (TTXI) AND
A HYPERTHERMOPHILIC Thermatoga neapolitana XYLOSE ISOMERASE

(TNXI)

36

ABSTRACT

Xylose isomerases (Xls) from Thermaanaerabacterium thermasulfurigenes
(TTXI) and Thermataga neapolitana (TNXI) are 70.4% identical in their amino acid
sequences and have a nearly superimposable crystal structure. Nonetheless, TNXI is
much more thermostable than TTXI. Except for a few additional prolines and fewer
Asn+Gln in TNXI, no other obvious differences in the enzyme structures can explain the
differences in their stability. TNXI has 2 additional prolines in the Phe59 loop (Pr058 and
Pro62). Mutations Gln58Pro, Ala62Pro, and Gln58Pro/Ala62Pro in TTXI and their
reverse counterpart mutations in TNXI were constructed by site-directed mutagenesis.
Surprisingly, only the Gln58Pro mutation enhanced thermostability of TTXI (43 %
longer half-life at 85°C), whereas, the Ala62Pro and Gln58Pro/Ala62Pro mutations both
dramatically lowered the TTXI’s stability. Analysis of the three-dimensional (3D)
structure of TTXI and the predicted structure of its Ala62Pro mutant derivative indicated
that a steric hindrance between Pro62-C8 and Lys6l-CB (2.92 A) is responsible for the
reduced thermostability of the mutant enzyme compared to the native TTXI. All the
reverse counterpart mutations destabilized TNXI thus confirming that these 2 prolines

play important roles in TNXI’s thermostability.

37

INTRODUCTION

Xylose isomerase (D-xylose ketol isomerase; EC 5.3.1.5) (X1) is an intracellular
enzyme found in a number of bacteria that utilize xylose as carbon substrate for growth
(Chen, 1980). XI converts D-xylose to D-xylulose in viva and also catalyzes the
conversion of D-glucose to D-fructose in vitro (Takasaki et al., 1969). This latter activity
is used in industry for the production of high fructose corn syrup (HFCS) and xylose (i.e.,
glucose) isomerase is one of the largest volume commercial enzymes used today (Lee
and Zeikus, 1991). Therrnostable XIs with neutral or slightly acidic pH optima have a
potential for industrial applications with the advantages of faster reaction rates, higher
fructose concentrations at equilibrium, decreased viscosity of substrate and product
streams, and less problems of by-products formation (Lee and Zeikus, 1991). Two XI
groups have been identified; type I enzymes are shorter than type II enzymes by about 50
amino acids at their N-terminus (Vangrysperre et al., 1990).

Two type II XIs have been studied extensively in our laboratory: one from a
thermophile, Thennoanaerabacterium thermasulfurigenes (TTXI) and the other from a
hyperthermophile, Thermataga neapolitana (TNXI). The genes encoding these enzymes
(xylA genes) were cloned, sequenced, and expressed in Escherichia cali (Lee et al.,
1990a, 1990b, Meng et al., 1991, and Vieille et al., 1995). TNXI has a higher turnover
number, a lower Km for glucose, and is more thermostable than any other known type II
xylose isomerases (Vieille et al., 1995). XIs are significantly stabilized and activated in
the presence of divalent cations, especially by Mg“, Co“, and Mn”. Although TNXI

and TTXI are highly similar (70.4% amino acid sequence identity), TNXI is significantly

38

 

“1016 ll
for II
Tle‘1
the dif

in TX.

al.. 1‘:
lExllli
enho]
residt
does
inter.
three
and

muta
addn
unto
lodri

91581

(

Allen
Serif)
stahih.‘

compo!

more thermostable than TTXI with an optimum temperature for activity of 95°C (85°C
for TTXI) and melting temperature in the presence of 0.5 mM Co2+ of 126°C (82°C for
TTXI) (Vieille et al., 1995). No obvious differences in the enzyme structures can explain
the differences in their stabilities except for a few additional prolines and fewer Asn+Gln
in TNXI (as seen from its alignment with TTXI).

Proteins can be stabilized by decreasing their entropy of unfolding (Matthews et
al., 1987). Prolines, with their pyrrolidine ring, can only adopt a few configurations. They
restrict the configurations allowed for the preceding residue, and they can decrease the
entropy of a protein’s unfolded state. Substituting Pro for another carefully chosen
residue can thus increase protein stability, provided that the newly introduced proline
does not create volume interferences, and does not destroy stabilizing non-covalent
interactions. The effect of prolines on protein stabilization has been studied by site-
directed mutagenesis. The two T4 lysozyme mutants Ala82Pro (Matthews et al., 1987)
and Ile3Pro (Dixon et al., 1992) illustrate the importance of carefully selecting the
mutation location. With mutation Ala82Pro, the two conditions listed above were
addressed, and the mutation stabilized the protein mainly by decreasing its entropy of
unfolding. In mutant Ile3Pro, the substitution eliminated a hydrogen bond and the
hydrophobic interactions created by lie. The degree of enthalpic destabilization was
greater than the entropy gained resulting in destabilization of the mutant enzyme. In
Allen et al.’s study of Aspergillus awamari glucoamylase, three proline mutants:
Ser30Pro, Asp345Pro, and G1u408Pro were constructed. The Ser3OPro mutation
Stabilized the protein because the mutation site allowed a residue conformation

compatible with a proline, and because residue 29 (a valine) could adopt one of the

39

 

 

conlor
Gil-11.1
conio

stahil

Ptolll

proli

conformations allowed for the residue preceding a proline (Matthews et a1, 1987). The
G1u408Pro mutation was destabilizing because the mutation site did not allow any
conformation required for proline, whereas the Asp345Pro mutation did not affect
stability due to the fact that the substitution destroyed the Ol-helix dipole in that region.
Prolines 58 and 62 in TNXI are present in a large loop in which 14 hydrophilic residues
surround Phe59. The corresponding TTXI residues are Gln58 and Ala62. The Phe59 loop
participates in building the neighboring subunit’s active site (Farber et al., 1989: Whitlow
etaL,l99l)

In the present report we test the hypothesis that substituting Gln58 and Ala62 with

prolines will stabilize TTXI, and that the reverse mutations will destabilize TNXI.

40

MATERIALS AND METHODS

Bacterial strains and chemicals: E. cali strain BL21 (DE3) (Novagen, Madison, WI.)
was used to overexpress the recombinant T. thermasulfurigenes and T. neapolitana xylA
genes cloned in the pET23a and pET22b+ vectors, respectively (Novagen). Media and
growth conditions were the same as described by Vieille et al. (1995). Medium

components and all other chemicals were reagent grade.

Site-directed mutagenesis and other DNA techniques: A11 DNA manipulations were
performed using established protocols (Sambrook et al., 1989; Ausubel et al., 1993).
Point mutations were introduced into the T. neapolitana and T. thermasulfurigenes xylA
genes using the QuikChangeTM site-directed mutagenesis kit (Stratagene, La Jolla, CA).
Mutagenic Oligonucleotides (Table 1) were synthesized by the Macromolecular Structure
Facility, Department of Biochemistry, Michigan State University. Mutations were
verified by DNA sequencing using the Thermosequenase sequencing kit (United States

Biochemical, Cleveland, OH).

Protein purification: Recombinant enzymes were purified using the procedure of
Vieille et al. (1995) followed by two additional steps. Partially purified enzymes were
applied to a DEAE-Sepharose column (2.5x15 cm) equilibrated with 50 mM MOPS (pH
7.0) containing 5 mM MgSO., and 0.5 mM CoClz (buffer A), and enzymes were eluted
using 500 m1 of 0—250 mM NaCl gradient in buffer A. The pooled fractions from the

DEAE-Sepharose

41

Table 1: Oligonucleotides and DNA templates used for site-directed mutagenesis. Bold

and underlined nucleotides are the mutation sites.

 

DNA templates

 

Gln58Pro/Ala62Pro

5’-GGAACAGATCCAI I IGGCAAACC-3’
3’-CCTTGTCTAGGTAAACCGI l lGG—S’

Mutations Oligonucleotides
TTXI:
5’-GGAACAGATCCATTTGGCAAAGC-3’
01°58!“ 3’-CCTI‘GTCTAGGTAAACCGTITCG-5’ “XI g°°°
Ala 62Pro 5’-GATCAAT'ITGGCAAACCTACCATGC-3’ TTXI we
3’-CTAGTTAAACCG’I'ITGGATGGTACG-5’ g

TTXI (Ala62Pro)
gene

 

 

TNXI:

ProS8Gln

Pro62Ala

ProSSGln/Pro62Ala

 

 

5’ -GGGAAGGGATCAGTTCGGAGACCC-3’
3’-CCCTTCCCTAGTCAAGCCTCTGGG-5’

5’-CCTTCGGAGACGCAACGGCCGATC-3’
3’ -GGAAGCCTCTGCGTTGCCGGCTAG-S’

5’-AGTTCGGAGACGCAACGGCCGATC-3’
3’ -TCAAGCCTCTGCGTTGCCGGCTAG-5’

 

TNXI gene

TNXI gene

TNXI (Pr058Gln)
gene

 

 

42

 

column were concentrated in a stirred ultrafiltration cell (MW cut-off 30 kDa)(Amicon,
Beverly, MA), dialyzed twice against buffer A, applied to a Polybuffer column
(Pharrnacia, Uppsala, Sweden) equilibrated with 25mM histidine-HCl (pH 6.2), and
eluted using a pH (6.0 to 4.0) gradient according to the manufacturer’s instructions. The
active fractions were pooled, concentrated in a stirred ultrafiltration cell (Amicon), and
dialyzed twice against buffer A. Concentrated and homogenous enzymes were dispensed

and stored frozen at -70°C.

Xylose isomerase assays: XI activity was routinely assayed with glucose as the
substrate. The enzyme (0.06 mg/ml) was incubated in 50 mM MOPS (pH 7.0 at room
temperature) containing 1 mM CoClz and 1 M glucose at 60°C for 20 min. The reaction
was stopped by cooling the tubes in ice. The amount of fructose produced was
determined by the cysteine-carbazole-sulfuric acid method (Dische and Borenfreund,
1951). To determine the effect of temperature on XI activity, the assay mixtures were
incubated at the temperatures of interest in a Perkin-Elmer Cetus GeneAmp PCR system
9600 (Perkin Elmer, Norwalk, CT) for 20 min. To determine the kinetic parameters,
assays were performed in the presence of either 80 to 1,400 mM glucose or 20 to 900
mM xylose. The amounts of fructose and xylulose produced were determined as above.
Absorbances were measured at 537 nm and 560 nm for xylulose and fructose,
respectively. One unit of isomerase activity is defined as the amount of enzyme that

produced 1 umole of product per min under the assay conditions.

43

Thermostability assays: The time course of irreversible thermoinactivation was
measured by incubating the enzyme (0.5-5 mg/ml) in 10 mM MOPS buffer (pH 7.0)
containing 50 11M CoClg (buffer B) at 85°C (TTXI derivatives) or 95°C (TNXI
derivatives) in a Perkin-Elmer Cetus GeneAmp PCR system 9600 for various amounts of
time, and by determining the residual glucose isomerase activity at 65°C (TTXI
derivatives) or 80°C (TNXI derivatives). The first order rate constant, k, of irreversible
thermoinactivation, was obtained by linear regression in semi-log coordinates of residual

activity. Enzyme half-life was calculated from the equation: tum = ln 2/k.

Heat-induced enzyme precipitation: Heat-induced enzyme precipitation was monitored
from 25°C to 100°C by light scattering (71. = 580 nm) using protein solutions (0.2 mg/ml)
in buffer B. Absorbance measurements were conducted in 0.3 m1 quartz cuvettes (path
length = 1.0 cm), using a Gilford Response spectrophotometer (Corning, Oberlin, OH)
equipped with a Peltier cuvette heating system. The increasing thermal gradient was
10°C min]. The temperature of 50% precipitation was the temperature at which the OD

at 580 nm equals half of the difference between the baseline and the maximum ODs.

Analysis of TTXI and TNXI three-dimensional (3D) structures: Enzymes were
visualized on an IRIS-4D25 computer (Silicon Graphics Computer System, Mountain
View, CA) using the INSIGHT II graphic program (Biosym Technologies, San Diego.
CA). Proteins Data Bank (PDB) files (#lAOC for TTXI and #1AOE for TNXI) were

obtained from the Protein Data Bank website (www.rcsb.org/pdb).

 

44

RESULTS

TNXI Pr058 and Pro62 are present in a large loop in which fourteen hydrophilic
residues surround Phe59 (Figure 1). According to crystallographic data, Phe59 (Phe26 in
Actinoplanes XI) participates in the architecture of the neighboring subunit’s active site
(Farber et al., 1989; Whitlow et al., 1991). The corresponding residues in TTXI, Gln58
and Ala62, have backbone dihedral angles ([-66.57, -l4.49] and [-57.91, 138.87],
respectively) allowed for prolines (Nicholson et al., 1988). With backbone dihedral
angles of (-139.73, -175. 10) and (-93.32, 176.20), respectively, TTXI residues Asp57 and
Lys61 are in extended conformations, the most common conformation for residues
preceding prolines (Nicholson et al., 1988). This structural information suggests that
mutations Gln58Pro and Ala62Pro would not create unfavorable backbone conformations
and that they could stabilize TTXI.

Mutation Gln58Pro did not affect the optimum temperature for TTXI activity (i.e.,
85°C), whereas mutation Ala62Pro and double mutation Gln58Pro/Ala62Pro decreased it
by 7°C and 12°C, respectively (Figure 2A). Mutation Gln58Pro stabilized TTXI at 85°C:
the enzyme’s half-life was extended from 69 to 99 min (a 43% increase). With half-lives
of 6.2 and 21 min at 85°C, respectively, Ala62Pro and Gln58Pro/Ala62Pro mutant TTXIs
were significantly less thermostable than the wild-type enzyme (Figure 3A). These
surprising results indicated that Pro in position 62 destabilized TTXI. Similar results were
obtained in precipitation experiments. Mutation Gln58Pro increased TTXI’s temperature

of 50% precipitation by 6°C, whereas mutations Ala62Pro and Gln58Pro/Ala62Pro

decreased it by approximately 4°C and 3°C, respectively (Table 2).

45

 

Figure 1: Three-dimensional structure of TNXI’s Phe59 loop region. Only parts of the
tetramer’s subunits A and C are shown. Subunit C, shown in yellow ribbon, is interacting
with subunit A’s Phe59 loop. Residues in subunit A’s Phe59 loop are colored based on

their hydrophilicity (Blue-hydrophilic, Red-hydrophobic).

46

 

 

 

Specific activity (U/mg)
Specific activity (U/mg)

 

 

 

 

 

 

40 50 60 70 80 90 100 50 60 70 80 90 100
Temperature (°C)

Figure 2: Effect of temperature on the specific activities of TTXI and TNXI and their

Phe59 loop mutant derivatives. The substrate used was glucose. (A) TTXI and its mutant

derivatives. Symbols: —l:l—: TTXI; ---<>---: Gln58Pro; ----O----: Ala62Pro; "um-A

Gln58Pro/Ala62Pro. (B) TNXI and its mutant derivatives. Symbols: —El—: TNXI;

<> Pr058Gln; --------O: Pro62Ala; --------A. Pr058Gln/Pro62Ala.

47

Ln (residual activity)
“.3
O/

 

 

 

Ln (residual activity)
Ln (residual activity)

 

 

 

 

 

 

-2 I T l l r
0 25 50 75 100 0 50 100 150

 

Time (min)

Figure 3: Inactivation curves of TTXI and TNXI and their Phe59 loop mutant derivatives

(A) TTXI and its mutant derivatives at 85°C. Symbols: —Cl—: TTXI; -----'<>

Gln58Pro; --------:O Ala62Pro; ----A----. Gln58Pro/Ala62Pro. Half-lives of TTXI,
Gln58Pro, Ala62Pro, and Gln58Pro/Ala62Pro are 69.3, 99.0, 6.2, and 21.0 min,

respectively. (B) TNXI and its mutant derivatives at 95°C. Symbols: —Cl—: TNXI;

mom; Pr058Gln; --------:O Pro62Ala; A. Pr058Gln/Pro62Ala. Half-lives of TNXI,

Pr058Gln, Pro62Ala, and ProS8Gln/Pro62Ala are 69.3, 49.5, 7.1, and 7.1 min,

respectively.

48

d,

The reverse counterpart mutations in TNXI (Pr058G1n and Pro62Ala, and double
mutation Pr058Gln/Pro62Ala) decreased the optimum temperature for TNXI activity
(Figure 28). Not surprisingly, thermoinactivation curves revealed that mutation ProS8Gln
decreased TNXI’s half-life from 69.3 min to 49.5 min (29% decrease). Both mutations
Pro62Ala and Pr058Gln/Pro62Ala decreased TNXI’s half-life to 11.6 min (83%
decrease), which confirmed that these mutations were all destabilizing as we expected
(Figure 3B). Again, these results were confirmed by precipitation experiments. The
temperature of 50% precipitation of both Pro62Ala and Pr058Gln were 95.7°C, 1.1°C
lower than that of the wild-type TNXI. Double mutant Pro62Ala/Pr058Gln precipitated at
an even lower temperature than each single mutant enzyme, suggesting that these
destabilizing effects are additive.

The kinetic features of the wild-type and mutant xylose isomerases with glucose
and xylose as substrates were determined at 65°C for the TTXI series and at 80°C for the
TNXI series (Table 3). With the exceptions of mutations Ala62Pro and
Gln58Pro/Ala62Pro in TTXI, the mutations in TTXI and TNXI Phe59 loops did not
significantly alter the enzymes’ catalytic properties. The Ala62Pro mutation increased
both TTXI’s Vmax and Km on glucose, leaving its catalytic efficiency on glucose almost
unchanged. This mutation had a much stronger effect on TTXI activity on xylose. It
increased its affinity for xylose and its catalytic efficiency approximately 2.9 times.
Mutation Gln58Pro/Ala62Pro had a more pronounced effect on TTXI activity on glucose.
A 2.6-fold increase in its Km for glucose decreased its catalytic efficiency almost 3-fold.
These results indicate that mutations Ala62Pro and Gln58Pro/Ala62Pro altered TTXI’s

catalytic features and at the same time destabilized the enzyme.

49

Table 2: E

of three in

 

Table 2: Effect of mutations on enzyme precipitation temperatures. Results are the means

of three independent experiments.

 

 

 

 

Denaturation temperature for 50%
Enzymes
precipitation (°C)

Wild-type TTXI 83.7 i 0.2
Gln58Pro 89.9 i 0.6
Ala62Pro 79.9 i 1.0
Gln58Pro/Ala62Pro 80.8 i 0.7
Wild-type TNXI 96.8 i 0.4
ProS8Gln 95.7 i 0.4
Pro62Ala 95.7 r 0.6
ProS8Gln/Pro62Ala 94.0 t 0.3

 

 

50

 

 

Caz—92-238532 of 3 33. 3_u:_o> EEC.— oLCC szc 32.6 swash: 32.536 9.03 33:5 56> wen EM .mEoEtono Eoc:..i.dz:

09F: .76 acne:— 05 was $237.91 80315.53? «:93»: :95 mo flan AUoOm any CAZF ficm abowc .3 CALLE~O 33.42552» 9.3.2339 um. 03:8

 

 

 

 

 

 

 

 

 

 

 

w.mm finev ad n mm 5; a fix Ev Wm; ad a md 5.3 a 0.8. £<NeoEEmeoE
wdw OAS. ed a 0.3 wd H No 06 mdmo ed a N2 _.: n 2.2 «730on
fine @me md a md wd n :. ~.v wfivv ed H mm 5.2 n ~62 mammot—
vdo QM: N; a 0.3 n; a Q: Wm Owen N4 a _.2 «How u v.2: 59F 090-25,
mam 932 car. a m.~m 3: a Sb v; o6; ed a Nd Qmm n vdom Summon—$95350
Qwv odS ed on v.2 o._ a 5.2 3 920 ed a v.2 NE a _.mm_ Samoa?
v.2 cams mg n _.m_ Ow a Maw m.m 0.3m v6 n as «.2 a odfi 856:0 1
ma: odes wd a 3; fin a q; _.v odcv md a md H2 u 3.: an: 090-259 5
Avis-e c-2873
be to. REE na> A2810. :3 .3. ans: §> A28 ex
ex} ex} moauam
one—xx 08020

 

 

 

 

8208598 E02032: 00:: mo 9808 05 0S £30m 285850
5802820322 05 8 3% 5823 REE 05 mo mafia 8050 $585 35890 0.63 329’ §e> 28 av. £88598 5385005

ooh: 0o 885 20 on £33. 83388 cause use no one 608 as szn o5 6.8 5 Ext 0o 8388 2:36 ”m 23.

The
different I
enzymes :
stability 2
identified
thermostu
stabiliza'i
T10 Obvic
therm0>t
COmp-are

1
"- Thes
engineel
entropy
pTOIEin.
0f the c
adOpt 0:
Should :
non‘Cm
backbor

dihedral

DISCUSSION

The high degree of similarity between TTXI and TNXI and their significantly
different thermostabilities and thermophilicities make comparative studies of these
enzymes attractive for understanding the key molecular features responsible for their
stability and activity differences (Vieille et al., 1995). Multiple factors have been
identified (Vieille and Zeikus, 1996) that can be responsible for a protein’s high
thermostability. They include packing efficiency, hydrophobic interactions, loop
stabilization, reduction of entropy of unfolding, electrostatic interactions, etc. In our case,
no obvious differences in 'I'TXI and TNXI structures could explain their different
thermostabilities except for a few additional prolines and fewer Asn+Gln in TNXI
compared with TTXI.

Two additional prolines are found in a large “Phe59 loop” region in TNXI (Figure
1). These prolines are substituted with Gln and Ala in TTXI. Proteins can be stabilized by
engineering proline into selected sites thereby decreasing the proteins’ conformational
entropy of unfolding (Allen et al., 1998). In order for a proline substitution to stabilize a
protein, at least four criteria have to be considered: (i) the mutation site should allow one
of the conformations allowed for proline; (ii) the preceding residue should be able to
adopt one of the conformations allowed for the residue preceding a proline; (iii) proline
should not create volume interferences; and (iv) proline should not destroy stabilizing
non-covalent interactions (Nicholson et al., 1988). TTXI Gln58 and Ala62 have
backbone dihedral angles allowed for prolines, and Asp57 and Lys61 have backbone

dihedral angles allowed for residues preceding prolines. Despite these first two

52

confomiati‘or
As seen in F
proline p)TTC
the wild-t}?t
coxalent inte
Gln58Pro. 'dl
reduction of
destabilizing
and it is not
eliminate an)
mutation HlOt
ring (C5 ator
CarbOn atom:
Waals interac
The unfavoral
[O [Oca] Conft
EifftEct [he dCllt
defiabilizing
destabilizgmOn
decrease in Um
1“ COnc
[0 identify an

(o a _
clueose’, 15mm

conformational requirements being satisfied, only mutation Gln58Pro stabilized TTXI.
As seen in Figure 4A, the conformation of Gln58’s sidechain is very close to that of the
proline pyrrolidine ring. No volume interference is created by the Gln58Pro mutation. In
the wild-type TTXI structure, Gln58 is not involved in any potentially stabilizing, non-
covalent interactions (not shown). No enthalpic destabilization is expected with mutation
Gln58Pro, and the stabilization it provides to TTXI probably entirely results from a
reduction of the entropy of unfolding. On the other hand, the Ala62Pro mutation had a
destabilizing effect on TTXI. In TTXI, Ala62’s sidechain points toward a large cavity,
and it is not in close vicinity of any other residues. So the Ala62Pro mutation does not
eliminate any stabilizing non-covalent interactions. Detailed analysis of the Ala62Pro
mutation modeled into the 'I'TXI structure (Figure 5) suggests that Pro62’s pyrrolidine
ring (C5 atom) is in close contact (within 2.92 A) with Lys6l’s sidechain (CB atom).
Carbon atoms have Van der Waals radii of 1.70-1.78 A in protein. Optimal Van der
Waals interactions between 2 carbon atoms would take place at approximately 3.4-3.5 A.
The unfavorable Van der Waals contact between Pro62-C5 and Lys6l-CB probably leads
to local conformational changes. Not only are these changes destabilizing, they also
affect the active site structure and the enzyme’s interaction with the substrate. The overall
destabilizing nature of mutation Ala62Pro indicates that the conformational
destabilization of the native enzyme more than cancels the benefits of a potential
decrease in unfolding entropy.

In conclusion, our data show that genetic engineering approaches can be utilized
to identify amino acid residues responsible for extreme thermal stability of xylose

(glucose) isomerase thermozymes. Both proline58 and 62 appear to stabilize TNXI’s

53

.5on E 2a £53852 #6:?» E m_ moo. omega m.< zcanzm 6:3 E 2m 852mm: 88 9:8 0 255.6 do. E mm D

:Sfism dimes—xx 5:938 Amy ”Exams—O 5:82: 2v AXE. E 225:8 n82 omega 2: Co 23023 65558662.; ”v Eswfi

V“

 

54

 

Figure 5: Van der Waals contacts between Lys61 and Pro62 in TTXI Ala62Pro mutant
derivative. Carbon atom Van der Waals radii were arbitrarily fixed at [.7 A. Lys6l and

Ala62 are in yellow. Pro62 is in green.

55

 

 

Phe59 lOOl

mutation.

Phe59 loop; whereas TTXI’s thermal stability can be enhanced by the Gln58Pro

mutation.

56

 

I<J

Allen. .\
dud/m ’7'
Protein

. Ausubel

and Stru
Greene 1

-. Chen. \\

. Creightt

Freemar

. Dische.

detectioi

. Dixon. F

(1992) S
$101.81},

Farber. (
Crystallt
7289‘721

Lee. C., .
0f Xylose
265, 190

Ilie. C.. 1
Bagdasar
€105“?le

2638-264

leeca

glUCose i
ThenNOQ

from 5] [e
SCI. L'S‘q

. Meng M

PrefErenC
redesignn

REFERENCES

Allen, M. J ., Coutinho, P. M., and Ford, C. F. (1998) Stabilization of Aspergillus
awamori glucoamylase by proline substitution and combining stabilizing mutations.
Protein Eng. 11, 783-788.

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J.A.,
and Struhl, K. (1993) Current Protocols in Molecular Biology. (Struhl, K., ed.),
Greene Publishing & Wiley-lnterscience, New York.

Chen, W. (1980) Glucose isomerase. Process Biochem. 15, 30—35.

Creighton, T. E. (1993) Protein: Structures and Molecular Properties, 2nd ed., W.H.
Freeman and Company, New York.

Dische, Z. and Borenfreund, E. (1951) A new spectrophotometric method for the
detection and determination of keto sugars and trioses. J. Biol. Chem. 192, 583—587.

Dixon, M. M., Nicholson, H., Shewchuk, L., Baase, W. A., and Matthews, B. W. .
(1992) Structure of a hinge-bending bacteriophage T4 lysozyme mutant, Ile3Pro. J.
Mol. Biol. 227, 917-933.

Farber, G. K., Glasfeld, A., Tiraby, G., Ringe, D., and Petsko, G. A. (1989)
Crystallographic studies of the mechanism of xylose isomerase. Biochemistry 28,
7289-7297.

Lee, C., Bagdasarian, M., Meng, M., and Zeikus, J. G. (19903) Catalytic mechanism
of xylose (glucose) isomerase from C lostridium thermosulfurofenes. J. Biol. Chem.
265, 19082-19090.

Lee, C., Bhatnagar, L., Saha, B. C., Lee, Y. E., Takagi, M., Imanaka, T.,
Bagdasarian, M., and J. G. Zeikus. (1990b) Cloning and expression of the

Clastridium thermosulfurogenes glucose isomerase. Appl. Environ. Microbiol. 56,
2638-2643.

10. Lee, C. and Zeikus, J. G. (1991) Purification and characterization of thermostable

ll.

12.

glucose isomerization from Clostridium thermosulfurogenes and
Thermoanaerobacter strain B6A. Biochem. J. 273, 565-571.

Matthews, B. W., Nicholson, H., and Becktel, W. (1987) Enhanced protein stability
from site-directed mutations that decrease the entrOpy of unfolding. Proc. Natl. Acad.
Sci. USA 84, 6663-6667.

Meng, M., Lee, C., Bagdasarian, M., and Zeikus, J. G. (1991) Switching substrate

preference of thermophilic xylose isomerase from D-xylose to D-glucose by
redesigning the substrate binding pocket. Proc. Natl. Acad. Sci. USA 88, 4015-4019.

57

 

, Nichol

thermo
336. of

. Sambrt

labom

Harbor

. Takasu

Agrit'.

. Vangr)

R.. and
carbox

'. Vieille

seq uen
neapol

. Vieille

determ
183-19

. Whitlo

Gillilar
isomer;
D‘X}'IO:

13.

14.

15.

16.

17.

18.

19.

Nicholson, H., Becktel, W. J ., and Matthews, B. W. (1988) Enhanced protein
thermostability from designed mutations that interact with a-helix dipoles. Nature
336, 651-656.

Sambrook, J ., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: a
laboratory manual (2nd ed.), Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, New York.

Takasaki, Y., Kosugi, Y., and Kanbayashi, A. (1969) Glucose isomerase enzymes.
Agric. Biol. Chem. 33, 1527-1534.

Vangrysperre, W., van Damme, J ., Vandekerckhove, J ., De Bruyne, C. K., Cornelis,
R., and Kersters-Hilderson, H. (1990) Localization of the essential histidine and
carboxylate group in D-xylose isomerases. Biochem. J. 265, 699—705.

Vieille, C., Hess, M., Kelly, R. M., and Zeikus, J. G. (1995) xylA cloning and
sequencing and biochemical characterization of xylose isomerase from Thermotoga
neapolitana. Appl. Environ. Microbiol. 61, 1867-1875.

Vieille, C., and Zeikus, J. G. (1996) Thermozymes: identifying molecular
determinants of protein structural and functional stability. Trends Biotechnol. 14,
183-190.

Whitlow, M., Howard, A. J., Finzel, B. C., Poulos, T. L., Winbome, E., and
Gilliland, G. L. (1991) A metal-mediated hydride shift mechanism for xylose

isomerase based on the 1.6 A Streptomyces rubiginosus structures with xylitol and
D-xylose. Proteins 9, 153-173.

58

CHAPTER III

ROLE OF SUBSTRATE BINDING POCKET IN Thermotoga neapolitana XYLOSE

ISOMERASE (TNXI) ON GLUCOSE ISOMERASE ACTIVITY

59

ABSTRACT

The active site of xylose isomerase from the thermophilic
Thermoanaerobacterium thermosulfurigenes (TTXI) has been previously engineered to
improve its catalytic efficiency toward glucose and increase its thermostability. The same
mutations were introduced into xylose isomerase from a hyperthermophilic Thermotoga
neapolitana (TNXI), which shares 70.4 % amino acid sequence identity. Similar trends
were observed, but to different extent. The TNXI Va1185Thr was the most efficient
mutant derivative with a 3.1 fold increase in its catalytic efficiency toward glucose
whereas the double mutation Trp138Phe/Va1185Thr was the best at increasing glucose
catalytic efficiency in TTXI. With a maximal activity at 97°C of 45.4 U/mg on glucose,
this TNXI mutant derivative is the most active type 11 XI ever reported. This “true”
glucose isomerase engineered from a native xylose isomerase has now comparable

kinetic properties on glucose and xylose.

60

INTRODUCTION

Xylose isomerase (D-xylose ketol-isomerase; EC 5.3.1.5) converts xylose to
xylulose during xylose metabolism in various microorganisms (Chen, 1980). This
enzyme also catalyzes the interconversion of glucose and fructose in vitro and has been
used as an industrial biocatalyst for the production of high fructose corn syrup. It displays
lower kcat and higher Km values for glucose than those for xylose. Specificity of enzymes
toward their substrates is determined in part by molecular residues that provide for
binding of the substrate and that maintain substrates steric configuration in the active site.
The enzymatic interconversion of xylose/glucose (aldoses) to xylulose/fructose (ketose‘s)
by xylose isomerases involves binding of the substrate in the cyclic form, linearization of
substate, isomerization of the linear intermediate, product ring closure, and release of the
cyclic product. The isomerization step proceeds by a metal ion mediated hydride shift
mechanism, and this step, rather than substrate ring opening, is rate-determining (Lee et
al., 1990). Xylose and glucose have identical atomic configuration, except for the
presence of an additional hydroxy-methyl (CHzOH) group at the C-6 position in the
glucose molecule. This extra hydroxymethyl group must therefore be responsible for the
differences in the catalytic efficiency exhibited by xylose isomerase toward glucose
versus xylose. The orientation of glucose as a substrate in the active site of the
thermophilic enzyme from Thermoanaerobacterium thermosulfurigenes (TTXI) is such
to position the C-6 end of hexose toward the HileO residue in the substrate-binding

pocket. The residues in the thermophilic enzyme that are in contact with the C6-OH group

61

of the substrate equivalent to those in the sorbitol-bound xylose isomerase from
Arthobacter (Collyer et al., 1990) are Trpl38, Thrl40, Va1185, and G1u23 1.

In a previous study, genetic engineering was used to improve TTXI activity on
glucose (not the natural substrate): the enzyme’s substrate-binding pocket was enlarged,
and the water-accessible surface area in the active site was altered (Meng et al., 1991 and
1993). Substituting Trpl38 with Phe, a smaller residue, decreased TTXI’s Km for glucose
and increased its catalytic efficiency (kw/Km) on glucose 2.6 times. This mutant enzyme
was less active than the wild-type TTXI on xylose. Interestingly, this mutation doubled
'I'I‘Xl’s half-life at 85°C (Meng et al., 1993). Substituting Va1185 with Thr, a polar
residue, improved TTXI’s catalytic efficiency on glucose. This improvement was thought
to result from the creation of an additional hydrogen bond to glucose’s C6-OH group. The
double mutation Trp138PheNa1185Thr vastly improved TTXI’s catalytic efficiency on
glucose (5.7 times).

Since TTXI and the xylose isomerase from the hyperthermophilic Thermotoga
neapolitana (TNXI) are more than 70 % identical in their amino acid sequences and have
a nearly superimposable three-dimensional structure, the enhancement of glucose’s
catalytic efficiency would be expected from those substitutions in the TNXI. In this
study, we introduced these mutations in TNXI to determine if TNXI activity and stability
could be further enhanced by genetic engineering. All mutations were characterized by
their kinetic parameters, optimum temperatures, half-lives at 85°C or 95°C, in order to

compare their thermal stability and activity.

62

MATERIALS AND METHODS

Bacterial strains and chemicals: E. coli strain BL21 (DE3) (Novagen, Madison, WI.)
was used to overexpress the recombinant T. thermasulfurigenes and T. neapolitana xylA
genes cloned in the pET23a and pET22b+ vectors, respectively (Novagen). Media and
growth conditions were the same as described by Vieille et al. (1995). Medium

components and all other chemicals were reagent grade.

Site-directed mutagenesis and other DNA techniques: A11 DNA manipulations were
performed using established protocols (Sambrook et al., 1989; Ausubel et al., 1993).
Point mutations were introduced into the T. neapolitana and T. thermosulfurigenes xylA
genes using the QuikChangeTM site-directed mutagenesis Kit (Stratagene, La Jolla, CA).
Mutagenic Oligonucleotides (Table 1) were synthesized by the Macromolecular Structure
Facility, Department of Biochemistry, Michigan State University. Mutations were

verified by DNA sequencing using the Thermosequenase sequencing kit (United States

Biochemical, Cleveland, OH).

Protein purification: Recombinant enzymes were purified using the procedure of Vieille
et al. (1995) followed by two additional steps. Partially purified enzymes were applied to
a DEAE-Sepharose column (2.5x15 cm) equilibrated with 50 mM MOPS (pH 7.0)
containing 5 mM MgSO4 and 0.5 mM COC12 (buffer A), and enzymes were eluted using
500 ml of 0—250 mM NaCl gradient in buffer A. The pooled fractions from the DEAE-

Sepharose column were concentrated in a stirred ultrafiltration cell (MW cut-off 30 kDa)

63

Table l: Oligonucleotides and DNA templates used for site-directed mutagenesis. Bold

and underlined nucleotides are the mutation sites.

 

 

Mutations Oligonucleotides DNA templates
Trpl38Phe 5’-GAAGCTCCTCTTTGGTACTGC—3’ TNXI gene
3’-CTTCGAGGAGAAACCATGACG-5’
Val 1 85Thr 5’-GAAGGGTACACC'ITCTGGGGTG-3’ TNXI gene
3’-C'ITCCCATGTGGAAGACCCCAC-5’
, , TNXI
Trp138PheNa1185Thr 5 -GAAGCTCCTC_TTTGGTACTGC-3 (v “85“")
3’-CTI‘CGAGGAGAAACCATGACG-5’ a gene

 

 

 

 

 

64

(Amicon, Beverly, MA), dialyzed twice against buffer A, applied to a Polybuffer column
(Pharmacia, Uppsala, Sweden) equilibrated with 25mM histidine-HCI (pH 6.2) ,and
eluted using a pH (6.0 to 4.0) gradient according to the manufacturer’s instructions. The
active fractions were pooled, concentrated in a stirred ultrafiltration cell (Amicon), and
dialyzed twice against buffer A. Concentrated; homogenous enzymes were dispensed and

stored frozen at -70°C.

Xylose isomerase assays: XI activity was routinely assayed with glucose as the
substrate. The enzyme (0.06 mg/ml) was incubated in 50 mM MOPS (pH 7.0 at room
temperature) containing 1 mM CoClz and 1 M glucose at 60°C for 20 min. The reaction
was stopped by cooling the tubes in ice. The amount of fructose produced was
determined by the cysteine-carbazole-sulfuric acid method (Dische and Borenfreund,
1951). To determine the effect of temperature on XI activity, the assay mixtures were
incubated at the temperatures of interest in a Perkin-Elmer Cetus GeneAmp PCR system
9600 (Perkin Elmer, Norwalk, CT) for 20 min. To determine the kinetic parameters,
assays were performed in the presence of either 80 to 1,400 mM glucose or 20 to 900
mM xylose. The amounts of fructose and xylulose produced were determined as above.
Absorbances were measured at 537 nm and 560 nm for xylulose and fructose,
respectively. One unit of isomerase activity is defined as the amount of enzyme that

produced 1 ttmole of product per min under the assay conditions.

Thermostability assays: The time course of irreversible thermoinactivation was

measured by incubating the enzyme (0.5—5 mg/ml) in 10 mM MOPS buffer (pH 7.0)

65

containing 50 ttM COC12 (buffer B) 95°C in a Perkin-Elmer Cetus GeneAmp PCR system
9600 for various amounts of time, and by determining the residual glucose isomerase
activity at 80°C. The first order rate constant, k, of irreversible thermoinactivation, was
obtained by linear regression in semi-log coordinates of residual activity. Enzyme half-

life was calculated from the equation: tum = In 2/k.

Heat-induced enzyme precipitation: Heat-induced enzyme precipitation was monitored
from 25°C to 100°C by light scattering (7t = 580 nm) using protein solutions (0.2 mg/ml)
in buffer B. Absorbance measurements were conducted in 0.3 ml quartz cuvettes (path
length = 1.0 cm), using a Gilford Response spectrophotometer (Corning, Oberlin, OH)
equipped with a Peltier cuvette heating system. The increasing thermal gradient was
10°C min“. The temperature of 50% precipitation was the temperature at which the OD

at 580 nm equals half of the difference between the baseline and the maximum ODs.

Analysis of TNXI three-dimensional (3D) structures: Enzymes were visualized on an
IRIS-4D25 computer (Silicon Graphics Computer System, Mountain View, CA) using
the INSIGHT II graphic program (Biosym Technologies, San Diego, CA). Proteins Data
Bank (PDB) file (#1AOE for TNXI) was obtained from the Protein Data Bank website

(www.rcsb.orgfi_3db).

66

RESULTS

The Va1185Thr mutant and the double mutant Trp138PheNa1185Thr TNXI
derivatives had the same optimum temperature of activity (97°C) as the wild-type TNXI
(Figure 1A). Interestingly, mutation Va1185Thr almost doubled TNXI specific activity on
glucose at 97°C (from 26.0 U/mg to 45.6 U/mg). Although the Trp138Phe mutation
decreased TNXI optimum temperature to 87°C, the mutant enzyme still retained almost
80% of its activity at 97°C. The stability of these mutant enzymes was studied at 95°C
(Figure 1B). With a half-life of 69.3 min, the Va1185Thr mutation did not affect TNXI
stability at 95°C. The Trp138Phe mutant (half-life of 87 min) and the
Trp138PheNa1185Thr double mutant (half-life of 99 min) enzymes were 25% and 43%
more stable, respectively, than the wild-type TNXI.

When kinetic parameters were concerned, these three TNXI mutant enzymes
showed the same trends as their equivalents in TTXI (Table 2). They all showed
improved catalysis on glucose and poorer catalysis on xylose. Most of the catalytic
efficiency increases were due to lower Km’s for glucose. Major differences existed
though, between TTXI and TNXI derivatives. Whereas Trpl38Phe is the best mutation in
term of increasing TTXI catalytic efficiency on glucose, its effect on TNXI catalytic
efficiency on glucose is only marginal. On the other hand, the mutation Vall85Thr is
better at increasing TNXI catalytic efficiency on glucose, than at increasing TTXI’s.
These differences are surprising, knowing that TTXI and TNXI active site structures are

almost completely superimposable.

67

 

 

50 2.5
B

a) 40- - 2 A
E e»
:> '5
>~ 30- —1.5 ‘3
.‘2: El ct:
> u—
.8 g
S 20- - 1 '53
95 :2,
0
<1.) C:
:2“ 104 -o.5 ”1

0

 

 

 

 

 

50 6O 70 80 90 1001100 20 4O 60 80 100120

Temperature (°C) Time (min)
Figure 1: Activity and stability of TNXI and its active site mutant derivatives. (A) Effect

of temperature on the specific activities of TNXI and its active site mutant derivatives.

The substrate used was glucose. Symbols: —Cl—: TNXI; ---<>---: Trp138Phe; --------:O

Vall85Thr; -------.A Trpl38Phe/Va1185Thr. (B) Inactivation curves of TNXI and its

active site mutant derivatives at 95°C. Symbols: same as in (A). Half-lives of TNXI,
Trp138Phe, Va1185Thr, and Trp138Phe/Va1185Thr are 69.3, 87.0, 69.3, and 99.0 min,

respectively.

68

:8: _e e 082 sec 9.5.

 

 

 

 

 

 

 

 

 

 

 

ofim 5me wd H 0.2 0; H 93: NE v.05 m.o H o.m_ _.m H mév Emw_~a>\0gmwm_&._i
aém vanes 0.0 H 92 m._ H 0.2 «.2 swan: c; H Nom 2: H meme EH33;

mdm wdvv ed H Em n; H 92 fie odan Wm H 52 Yam H v52 23min;

VS ng a; H 0.3 h; H 0.: Wm Owen N._ H _.m_ mdm H 1.2: RE 095-253
QNM own 02 Wm H on odm o3 0.2 Nam H mm .Ehhw:0>\0amwm_fih
vdm o: 03 n; H 2 5.0 own 0.: cs H 3 mar—.325»

0.2 owe ~.N_ : H 8 92 o3 v.2 v.5 H 3 .05528...

«.3 co: 0.; Na H N_ w.m own ”9 as H o: .CCL. 090-33

$28.3 958 $2873 $52
$5 .30. A25 av. p-331 9:5 em
33 59> 23 ae> 8:555
083x 08050

 

 

 

 

05250098 5030355 00:: $0 5005 05 0.8 530M .8533

:0502.50552 05 9 £50 .5023 355 05 mo w5=m 00050 nwsefi 005030 0.83 85? §e> 53 ev— .3:05._0nx0 505.035

8.5 0e 28.: 2: ea ease .eeéeéee Ease :2: 0e e5 608 3 02h e5 6.8 3 it .8 85888 e536 ”N 23.

69

DISCUSSION

The key molecular structure-function feature previously studied in
Thermoanaerobacterium thermosulfurigenes xylose isomerase (TTXI) is residues in the
active sites. Meng et al., (1991) have shown that mutation Trp138Phe significantly
increased TTXI catalytic efficiency on glucose (2.7 fold increase). This increase was
suggested to result from a better accommodation of glucose in the substrate-binding site
because Trp was substituted by a smaller residue, Phe. This substitution was also shown
to double TTXI’s half-life at 85°C (Meng et al., 1993). The higher thermostability may
be explained as the consequence of a reduction of the water-accessible hydrophobic
surface area. Another mutation that has been shown to increase TTXI’s catalytic
efficiency on glucose was Va1185Thr. This increase was attributed to additional hydrogen
bonding of Thr to glucose’s C6—OH.

When these mutations were introduced into TNXI, we observed the same trends
in terms of catalytic features. All the mutant enzymes, including the double mutant
Trp138PheNa1185Thr, had a higher catalytic efficiency for glucose than the wild-type
TNXI. These increases mainly resulted from a much lower Km for glucose than that of
TNXI. On the other hand, all these mutant TNXIs showed a lower catalytic efficiency for
xylose like they also did in TTXI. W138F was the mutation that had the most significant
effect on TTXI’s catalytic efficiency on glucose. Here, TNXI V185T mutant derivative
had the highest catalytic efficiency on glucose. This result came as a surprise since the
two XIs’ active sites were almost completely superimposable. As seen in figure 2,

Trpl38 and Val l 85 are almost completely superposed in TTXI and TNXI. The

70

 

Figure 2: Superposition of TTXI and TNXI active sites. Single letter code is used for
amino acid residues. TTXI residues are in yellow. TNXI residues are in red. Metal site I
(i.e., structural metal, M1) and metal site II (i.e., catalytic metal, M2) are represented by
crosses. Both metal sites are occupied by C02+in both enzymes. Co2+ in M2 of TTXI and

TNXI are 1.23 A apart.

7]

conformations of the neighboring residues are also extremely conserved. The most
significant difference between TTXI and TNXI active sites is the position of metal 11
(i.e., the catalytic metal, a Co2+ in both enzymes): the two cations are 1.23 A apart. This
shift of the catalytic metal might affect the enzyme-substrate binding properties and/or
the catalytic metal’s reactivity during catalysis, thus explaining why there was a
difference in mutation effects on TTXI and TNXI catalytic activities on glucose. All
TNXI mutant derivatives were as stable as (Va1185Thr) or more stable than the wild-type
TNXI (Trp138Phe and Trp138Phe/Va1185Thr). These mutations did not increase TNXI
stability to the same extent as they did in TTXI (Meng et al., 1993). Mutation Trp138Phe
might have the same stabilizing potential in TNXI as in TTXI, but another TNXI
molecular feature probably becomes limiting for stability before the full stabilization
potentially provided by the mutation can be reached. These findings prove that it is
possible to further stabilize hyperthermophilic proteins. To our best knowledge, with a
maximal activity on glucose at 97°C of 45.4 Units/mg, TNXI Va1185Thr mutant
derivative is the most active type II xylose isomerase ever reported.

In conclusion, our data show that genetic engineering approaches can be utilized
to identify amino acid residues responsible for high catalytic activity of xylose (glucose)
isomerase thermozymes. TNXI’s thermal stability was further enhanced by the
Trpl38Phe substitution. The Va1185Thr mutation significantly enhanced TTXI’s and
TNXI’s Vmax and kcm/Km for glucose isomerization to fructose. This significant catalytic
enhancement of glucose isomerase activity was made possible in large part by the
template enzyme’s naturally evolved function as xylose isomerase. Genetic engineering

altered the active site only to better accommodate glucose as the substrate.

72

10.

REFERENCES

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, IA,
and Struhl, K. (1993) Current Protocols in Molecular Biology. (Struhl, K., ed.),
Greene Publishing & Wiley-Interscience, New York.

Chen, W. (1980) Glucose isomerase. Process Biochem. 15, 30-35.

Collyer, C. A., Henrick, K., and Blow, D.M. (1990) Mechanism for aldose-ketose
interconversion by D-xylose isomerase involving ring opening followed by a 1,2-
hydride shift. J. Mol. Biol. 212, 211-235.

Dische, Z. and Borenfreund, E. (1951) A new spectrophotometric method for the
detection and determination of keto sugars and trioses. J. Biol. Chem. 192, 583-587.

Lee, C., Bagdasarian, M., Meng, M., and Zeikus, J. G. (1990a) Catalytic mechanism
of xylose (glucose) isomerase from Clostridium thermosulfurofenes. J. Biol. Chem.
265, 19082-19090.

Lee, C., Bhatnagar, L., Saha, B. C., Lee, Y. E., Takagi, M., Imanaka, T.,
Bagdasarian, M., and J. G. Zeikus. (1990b) Cloning and expression of the
Clastridium thermosulfurogenes glucose isomerase. Appl. Environ. Microbiol. 56,
2638-2643.

Meng, M., Lee, C., Bagdasarian, M., and Zeikus, J. G. (1991) Switching substrate
preference of thermophilic xylose isomerase from D-xylose to D-glucose by
redesigning the substrate binding pocket. Proc. Natl. Acad. Sci. USA 88, 4015-4019.

Meng, M., Bagdasarian, M., and Zeikus, J. G. (1993) Thermal stabilization of xylose
isomerase from Thermoanaerobacterium thermosulfurigenes. Proc. Natl. Acad. Sci.
USA 90, 8459-8463.

Sambrook, J ., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: a
laboratory manual (2nd ed.), Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, New York.

Vieille, C., Hess, M., Kelly, R. M., and Zeikus, J. G. (1995) xylA cloning and

sequencing and biochemical characterization of xylose isomerase from Thermotoga
neapolitana. Appl. Environ. Microbiol. 61, 1867-1875.

73

CHAPTER IV

DIRECTED EVOLUTION OF Thermotoga neapolitana XYLOSE ISOMERASE:

HIGH ACTIVITY ON GLUCOSE AT LOW TEMPERATURE AND LOW pH

74

ABSTRACT

The Thermotoga neapolitana xylose isomerase (TNXI) is extremely thermostable
and highly active at 95°C and above. Its mutant derivative, TNXI V185T, was the most
active type 11 XI previously reported, with a catalytic efficiency (kw/Km) of 25.1 5'1 mM'1
toward glucose at 80°C (pH 7.0). To further optimize this enzyme’s potential industrial
utility, two rounds of random mutagenesis and low temperature/low pH activity screening
were performed using the TNXI Vl85T-encoding gene as the template. Mutants TNXI
3A2 (V185T/L282P) and IF] (V185T/L282P/F186S) were obtained after rounds one and
two of random mutagenesis, respectively. TNXI lFl was more active than 3A2, which. in
turn was more active than TNXI V185T at all temperatures and pHs tested. TNXI 3A2
and 1F1 high activities at low temperatures were due to significantly lower activation
energies (57 and 44 kJ/mole, respectively) than that of TNXI and V185T (87 kJ/mole).
This observation suggested that TNXI 3A2 and lFl’s increased activity at low
temperature is a consequence of their increased flexibility in the active sites. Although
3A2 was more active than TNXI and V185T, its kinetic stability (based on the enzymes’
half life in different incubation conditions) was inferior to those of TNXI V185T possibly
due to unfavorable van der Waal contacts of Pr0282’s pyrrolidine ring with neighboring
mainchain atoms. This would, in turn, lead to conformational changes and eventually
destabilize the enzyme. Unlike TNXI 3A2, 1Fl is more kinetically stable than TNXI and
TNXI V185T. lFl’s enhanced stability is thought to be a result of additional H-bond
formation between Ser186’s sidechain and the neighboring L229 residue’s mainchain

structure. This, in turn, strengthens local conformation and the affinity of E231 co-

75

ordination with the structural metal, hence restoring the thermostability lost in 3A2. We
showed here that low temperature/low pH activity of a hyperthermostable enzyme could

be enhanced without costs to extreme thermal stability by directed enzyme evolution.

76

INTRODUCTION

Xylose isomerase (XI) (EC.5.3.1.5) is an intracellular enzyme found in bacteria
that can utilize xylose as a carbon substrate for growth (Chen, 1980). Due to its ability to
use glucose as substrate and convert it to fructose, X1 is often referred to as glucose
isomerase, and it is widely used in the industrial production of high fructose corn syrup
(HFCS) (Meng et al., 1993; Bhosale et al., 1996). X1 is one of the three highest tonnage
value enzymes, amylase and protease being the other two (Bhosale et al., 1996). Most
industrially used XIs are isolated from mesophilic organisms (e.g., Streptomyces spp. and
Actinoplanes spp.). The reaction temperature used in the current industrial glucose
isomerization process is limited to 60°C because of by-product and color formations that
occur at high temperature and alkaline pH, and because the isomerases themselves are not
highly thermostable (Lee and Zeikus, 1991; Vieille and Zeikus, 2000). Thermostable XIs
with neutral or slightly acidic pH optima have a potential for industrial applications.
Performing isomerization at higher temperature than 60°C and neutral/slightly acidic pH
with thermostable XI would allow faster reaction rates, higher fructose concentrations at
equilibrium, higher process stability, decreased viscosity of substrate and product
streams, and reduced by-products formation (Lee and Zeikus, 1991; Vieille and Zeikus,
2000).

The XI from the hyperthermophile Thermotoga neapolitana (TNXI) has been
studied extensively in our laboratory. The gene encoding TNXI (xylA) was cloned,
sequenced, and overexpressed in Escherichia coli (Vieille et al., 1995). TNXI’s active

site was engineered by site-directed mutagenesis to increase its activity on glucose

77

(Sriprapundh et al., 2000). The TNXI Vall85Thr (V185T) mutant derivative is more
active, more glucose-efficient, and as stable as the wild-type TNXI. It was also the most
active type II XI ever reported. Although TNXI V185T is highly thermostable and highly
active at 97°C, it is very poorly active (10 % of maximal activity) at the current industrial
isomerization temperature (60°C) and it requires a neutral pH for optimal activity.

Rules for engineering protein activity and stability by rational design are likely to
be protein-specific, and any such design effort would require prior detailed structural
information. Numerous and intensive site-directed mutagenesis studies have probed this
issue. Despite these efforts, considerable disagreement remains over which forces
dominate stabilization mechanisms, and no generally applicable rules have been
established (Giver et al., 1998; Vieille and Zeikus, 2001). Although protein chemists
continue to elucidate the relationships between the sequence, structure, and function of
proteins, the extensive knowledge that is necessary for the application of rational
engineering approaches is available for only a tiny fraction of known enzymes. Directed
evolution, on the other hand, has proved to be useful for modifying enzymes in the
absence of such knowledge (Kuchner and Arnold, 1997). In directed evolution, the
process of natural evolution is accelerated in a test tube for selecting proteins with the
desired properties (Moore and Maranas, 2000). A typical experimental cycle of directed
evolution begins with the creation of a library of mutated genes. Among the methods that
introduce mutations randomly along the entire length of a gene (Leung et al., 1989,
Stemmer, 1994, Zhao and Arnold, 1997, Shao et al., 1998, Zhao et al., 1998, and
Ostermeier et al., 1999), error-prone PCR has been used the most extensively. The

mutated genes are then ligated into an expression vector and transformed into suitable

78

bacterial cells. A screening procedure is next employed to identify the few transforrnants
expressing proteins/enzymes with improved properties. Random mutagenesis and
screening are repeated several times depending on the extent to which the properties of
the protein should be altered and on the effects of mutations observed in each generation.
Interest in engineering enzymes using directed evolution has grown significantly in the
past few years. It has been used to increase enzyme thermostability, activity on novel
substrates, substrate specificity, and enantioselectivity. For example, six generations of
random mutagenesis, recombination, and screening stabilized Bacillus subtilis p-
nitrobenzyl esterase significantly (>14°C increase in Tm) without compromising its
catalytic activity at lower temperature (Giver et al., 1998).

Here we use the TNXI V185T-encoding gene as the template for directed
evolution to develop an enzyme active at 60°C and acidic pH. We show that activity can
be increased significantly at low temperature and acidic pH without cost to the enzyme

thermal stability.

79

MATERIALS AND METHODS

Random mutagenesis: Random mutations were introduced into the TNXI V185T-
encoding gene cloned between the NdeI and HindIII restriction sites of pET23a(+). PCR
was performed with primers 5’-CGACTCACTATAGGGAGAC-3’ and 5’-
GGTGGTGCTCGAGTGCG-3’ encoding sequences upstream of the Mid site and
downstream of the HindIII, respectively, in pET23a(+). The reaction mixture contained
100 ng plasmid DNA, 50 pmol of each primer, 50 mM KCl, 10 mM Tris-HCl (pH 8.3),
1.5 mM MgC12, 0.4 mM dCTP, 0.4 mM dTTP, 0.08 mM dATP, 0.08 mM dGTP, and 2.5
Units Taq DNA polymerase (Roche, Nutley, NJ) in a 50 ul reaction volume. Cycling
parameters were 36 cycles of 95°C for 45 sec, 50°C for 45 sec, and 72°C for 3 min.
Amplification of the 1.4-kb product was checked by running a small aliquot of the
reaction on a 1 % agarose gel. The PCR product was purified using the Geneclean III kit
(BiolOl, Carlsbad, CA) and cloned back into the NdeI and HindIII sites of pET23a(+)
using standard molecular biological techniques (Ausubel et al, 1993). For the second

round of random mutagenesis, the gene encoding TNXI 3A2 was used as the template.

Construction of a mutant library: The plasmids resulting from random mutagenesis
were transformed into electrocompetent E. coli HB101(DE3) cells (XI-deficient) created
using the ADE3 lysogenation kit (Novagen, Madison, WI). Transformants were plated on
Luria-Bertani (LB) agar containing 100 ttg/ml ampicillin. After 16 hr of growth, single
colonies were picked with sterile toothpicks and transferred into 24-well plates, each well
containing 2 ml of LB plus 100 ug/ml ampicillin. Plates were then incubated overnight at

37°C on a shaker at 175 rpm to allow for cell growth. One hundred fifty microliters of

80

each culture were transferred to sterile 96-well plates. These plates were used to quantify
bacterial growth by reading absorbance of the bacterial suspensions (OD595) in a
microplate reader (Dynatech, McLean, VA), before being stored at 4°C to save the
original cultures. The rest of the cultures were pelleted by centrifugation at 1,000 g for 10
min and resuspended in 200 ul of 50 mM MOPS (pH 7.0) containing 5 mM MgSO.; and
0.5 mM CoClg (i.e., buffer A). Bacterial suspensions were incubated with 50 pl of a 1 %
(w/v) lysozyme solution at 37°C for 1 hr before being subjected to 3 freeze/thaw cycles
(5 min in a dry ice-ethanol bath and 5 min in a 50°C water bath) to break the cells and
release the enzymes into the supernatant. Cell-free crude extracts were then obtained by

centrifugation at 1,000 g for 10 min and stored at 4°C for further use.

Screening the mutant library for increased activity on glucose at 60°C and low pH:
The crude extracts were assayed. for glucose isomerization in two conditions: 60°C (pH
7.0) and 80°C (pH 5.2). Assays were performed in microtiter plates with 150 pl of 100
mM MOPS (pH 7.0) or 100 mM sodium acetate (pH 5.2) containing 1 mM CoClg, 0.4 M
glucose, and 10 til of crude extract. The plates were incubated at 60°C (pH 7.0) or 80°C
(pH 5.2) for 10 min and placed on ice to stop the reactions. The fructose produced was
assayed using the resorcinol-ferric ammonium sulfate-hydrochloric acid method (Schenk
and Bisswanger, 1998). Ten microliters of each reaction were transferred to a new set of
microtiter plates and mixed with 40 ul of distilled water and 150 pl of a freshly prepared
1:] mixture (v/v) of solution A (0.05 % resorcinol in ethanol) and solution B (0.216 g of
FeNH4(SO4)2.12HzO in IL HCl). The plates were incubated in an 80°C water bath for 30

min to develop the color. The OD490 was measured with a microplate reader (Dynatech)

81

with 0-2.5 mM fructose as standards. A crude extract of HB101(DE3)pET23a(+) was
used as the negative control on each plate. Crude extracts of HB101(DE3) expressing
TNXI V185T and TNXI 3A2 were the positive controls in mutagenesis rounds one and
two, respectively. Mutants with potentially higher activity on glucose than the positive
control were selected on the basis of increases in both OD490 and OD490/OD595 relative to
the positive controls in the two rounds of mutagenesis. Mutants showing increased
activity were screened a second time using crude extracts prepared from 5 ml cultures.
These crude extracts were prepared as described above, before being heat-treated at 80°C

for 15 min and centrifuged.

Oligonucleotide synthesis and DNA sequencing: PCR primers were synthesized by the
Macromolecular Structure Facility, Department of Biochemistry and Molecular Biology
at MSU. DNA sequences were determined either manually using the Thermosequenase
kit (USB, Cleveland, OH) or automatically at the MSU Genomics Technology Support

Facility.

Protein Purification: Recombinant enzymes were purified as described (Vieille et al.,
1995), followed by an additional ion-exchange chromatography step. Partially purified
enzymes were applied to a DEAE-Sepharose column (2.5 x 15 cm) equilibrated with
buffer A, and enzymes were eluted using a 500 ml linear 0—300 mM NaCl gradient in
buffer A. The pooled fractions from the DEAE-Sepharose column were concentrated in a

stirred ultrafiltration cell (30 kDa MW cut-off) (Amicon, Beverly, MA) and dialyzed

82

twice against buffer A. Concentrated, homogenous enzymes were dispensed and stored

frozen at -70°C.

Glucose isomerase assays: TNXI and its mutants were assayed routinely with glucose as
the substrate. The enzyme (1-1.5 mg/ml) was incubated in 100 mM MOPS (pH 7.0) [or
100 mM sodium acetate (pH 5.5)] containing 1 mM CoClz and 0.4 M glucose at 80°C for
10 min. The reaction was stopped by transferring the tube to an ice bath. The amount of
fructose produced was determined by the resorcinol-ferric ammonium sulfate-
hydrochloric acid method (Schenk and Bisswanger, 1998). To determine the effect of
temperature on activity, the enzymes were incubated in the reaction mixture at the
temperatures of interest in a heated water (45-95°C) or oil bath (95-110°C) for 10 min.
The effect of pH on activity was determined using the routine assay described above
except that the MOPS buffer was substituted with 100 mM sodium acetate (pH 4.3-5.8),
100 mM PIPES (pH 6.1-7.0), or 100 mM EPPS (pH 7.2-8.1). All pHs were adjusted at
room temperature, and the ApKa/At’s for acetate, PIPES, and EPPS (0, -0.0085, and -
0.011, respectively) (USB, Cleveland, OH) were taken into account for the results. To
determine the kinetic parameters, assays were performed in 50 mM MOPS (pH 7.0)
containing 10-1,500 mM glucose and 1 mM CoClg. One unit of glucose isomerase
activity is defined as the amount of enzyme that produces 1 pmol of fructose per minute

under the assay conditions.

Thermal inactivation assays: To obtain the apo-enzymes (metal-free enzymes), the

purified enzymes were incubated overnight at 4°C in 50 mM MOPS (pH 7.0) containing

83

10 mM EDTA. They were then dialyzed twice against 50 mM MOPS (pH 7.0) containing
2 mM EDTA, and they were finally dialyzed twice against 50 mM MOPS (pH 7.0)
without EDTA. CoClz (0.5 mM) was added to the apo-enzymes and equilibrated at 4°C
overnight before thermoinactivation assays. The time course of irreversible
thermoinactivation was measured by incubating the enzymes (0.1-0.2 mg/ml) in either 10
mM MOPS (pH 7.0) or 10 mM sodium acetate (pH 5.5) at various temperatures for
different periods of time in a heated water bath. Residual glucose isomerase activity was
measured at 80°C as described above. The first order rate constant, k, of irreversible
thermoinactivation was obtained by linear regression in semi-log coordinates. Enzyme

half-lives were calculated from the equation: too) = ln2/k.

Analysis of three-dimensional (3D) structures of TNXI and its variants: Enzymes
were visualized on an IRIS-4D25 computer (Silicon Graphics Computer System,
Mountain View, CA) using the INSIGHT 11 graphic program (Biosym Technologies, San
Diego, CA). The TNXI pdb file (#lAOE) was obtained from the Protein Data Bank

(www.rcsb.org/pdb).

84

RESULTS

Construction of mutant TNXI libraries and screening for activity on glucose at low
temperature and low pH: TNXI V185T is optimally active at 95°C - 97°C, but its
activity at 60°C does not exceed 10% of its optimal activity (Sriprapundh et al., 2000). It
retains only 20% of its optimal activity at pH 5.2. To increase this enzyme’s activity at
60°C and at acidic pH, and to gain insight into the factors determining the effects of
temperature and pH on activity, we subjected the TNXI V185T-encoding gene to
sequential random mutagenesis and to low temperature/low pH activity screening.
Random mutations were introduced into the gene by error-prone PCR. The PCR
conditions used were suggested to yield an average of 1-2 mutations per gene, conditions
deemed optimal for the improvement of specific properties by mutagenesis and screening
(Arnold and Moore, 1997). After the first round of random mutagenesis, 1,000
transforrnants were screened for their activity on glucose at low temperature (60°C, pH
7.0) and at low pH (pH 5.2, 80°C). Thirty mutants were identified that showed
significantly higher activity (> 30% increase) than TNXI V18ST in both screening
conditions. The phenotype of these mutants was tested again with heat-treated crude
extracts prepared from 5 ml cultures. Higher activity on glucose was confirmed in only
eleven out of the thirty crude extracts. XI expression level in these eleven crude extracts
was checked by SDS-PAGE. Ten crude extracts showed higher XI content than the TNXI
Vl85T control (data not shown). These ten mutants were discarded. The remaining
mutant, TNXI 3A2, was purified to homogeneity. Once it was verified that TNXI 3A2

was significantly more active than TNXI V185T at 60°C and at pH 5.2, the gene

85

encoding TNXI 3A2 was used as the template in a second round of error-prone PCR and
activity screening at low temperature and low pH. A library of ~l,500 transforrnants was
screened using TNXI 3A2 as the positive control. A single mutant, TNXI 1F1, showed
80% and 40% increases in activity on glucose at 80°C (pH 5.2) and 60°C (pH 7.0),
respectively, based on assays with heat-treated crude extracts. TNXI 3A2 and IF] were
then purified to homogeneity. Their catalytic properties were studied in function of

temperature and pH, and their thermostability was determined.

Effects of temperature and pH on TNXI 3A2 and IF 1 activities: The effect of
temperature on 3A2 and IF] glucose isomerase activities is shown in Figure 1A in
comparison to the activities of TNXI and TNXI V185T. Both 3A2 and lFl show
significantly higher specific activity on glucose than TNXI and TNXI V185T at all
temperatures. At their optimal temperatures of activity (i.e., 90°C for lFl and 95°C for
3A2), both mutants are ~ 3-fold more active than TNXI V185T. Activation energies
(Ea’s) for activity on glucose were calculated from the linear regressions shown in Figure
1B, using the equation A = Aoe'E‘VRT. Whereas TNXI V185T shows the same activation
energy as TNXI (i.e., 87 kJ/mole), 3A2 and lFl show significantly decreased Ea’s (57
and 44 kJ/mole, respectively). These lower Ea’s explain why 3A2 and 1F1 are as much as
7.3 and 12.3 times more active, respectively, than TNXI at 60°C, but only 4.2 and 4.8
times more active, respectively, than TNXI at 90°C.

The effect of pH on the activities of TNXI and its mutant derivatives is shown in
Figure 2. 3A2 and IF] show significantly increased specific activity on glucose

compared to TNXI and TNXI V185T over the entire active pH range. The activity

86

50

 

A . A
O
40" A .
A
EB
E o
a 30- A ‘
5’ o
'5
.3 A
«3
5
M.—
'6'
0
a.
U)

 

 

 

 

120

Temperature (°C)

Figure 1: A) Effect of temperature on the specific activities of TNXI and its mutant
derivatives on glucose at pH 7.0. (a): TNXI; (0): TNXI V185T; (O): TNXI 3A2: (A):

TNXI 1F].

87

 

Ln(specific activity)

 

 

 

'1 l l l I
0.0026 0.0027 0.0028 0.0029 0.003 0.0031

l/Temperature ( l/ K)

Figure 1: B) Ln (specific activity) versus l/Temperature. All linear regressions had r2

values above 0.97. (D): TNXI; (O): TNXI V185T; (O): TNXI 3A2; (A): TNXI lFl.

Activation energies (E) for activity of TNXI, TNXI Vl85T, TNXI 3A2, and TNXI 1F]

are 87, 87, 57, and 44 kJ/mole, respectively.

88

4O

 

 

 

 

A
A
A
A
30~
A
go A
\ O
:3 C
E O . O A
:g 20- ‘
8 A O
.2
I":
g c
U)
107 ‘A' . .. .
. . . I I. I i\-
/ 3‘. I v
C A/
0_ .l“-
I I I
4 5 6 7 8 9

pH

Figure 2: Effect of pH on specific activities of TNXI and its mutant derivatives on

glucose at 80°C. (El): TNXI; (O): TNXI V185T; (O): TNXI 3A2; A TNXl 1F].

89

increase is so significant that 3A2 and lFl are more active at pH 5.5 than TNXI and

TNXI Vl 8ST are at pH 7.0.

Kinetic parameters of TNXI 3A2 and IF]: The kinetic parameters on glucose of TNXI
V185T, 3A2, and IF 1 were compared in different conditions of temperature and pH
(Table 1). In all conditions tested, TNXI 3A2 and IF] showed higher Km and Vmax values
than TNXI V185T did. At pH 7.0 (both at 60°C and 80°C), TNXI 3A2 and lFl ’s Vmax
values increased more significantly than their Km’s for glucose, yielding important
increases in catalytic efficiencies (up to 2.3 fold for lFl at 60°C [pH 7.0]). At 80°C (pH
5.5), the increases in TNXI 3A2 and lFl’s Vmax’s do not compensate for the major
increases in their Km for glucose (i.e., 3.0 fold for 3A2 and 4.6 fold for lFl). In these
conditions, TNXI 3A2 and lFl show catalytic efficiencies that are approximately half
that of TNXI Vl8ST. At 60°C (pH 5.5), TNXI 3A2’s increase in Vnm does not
compensate for a poor glucose affinity (high Km), resulting in a lower catalytic efficiency
than that of TNXI V185T. Unlike TNXI 3A2, 1F1 has a higher catalytic efficiency on
glucose than TNXI V185T does due to a dramatic increase (5 fold) in its Vmax that
surpasses the increases in its Km (3.7 fold) in these conditions. Its 5-fold increase in Vmax
makes lFl a 1.7 fold more active enzyme at 60°C (pH 5.5) than TNXI V185T is at 80°C

(pH 7.0).

Thermal stability of TNXI 3A2 and IF]: To determine whether the mutations present
in 3A2 and IF] affected the kinetic stability of the mutated enzymes, the residual

activities of 3A2 and lFl were measured after heat treatment at 80°C (pH 7.0) and 80°C

90

.5833 5202820222 05 9 9%

5623 325 05 mo mafia 886 smack: 8:350 203 82m> ee> 98 5% $258598 Eoufimuufi 025 we £808 2: 8a 333m.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we 53 md H od— 3. H min 53 mum 5o H v.2 o.m H «.8 7: 5mg
3 new no H 3 ha H 33m m.~_ NS ho H ~12 Wm H mém N<m HXZH
Wm m2 «6 H w.m 03 H :3 me 5 0o H QM vi H mdm HQ _> 39E.
9-28% :3 9:52 :25 p.263 A 73 ESE A25
EMNnov— «aux KQE> EM Eggs; «3 Kfl~=> EM
3 me 00% 3 mg 00%
c6 ma: _.v H mwm 9g 2% 98. EN H 3% E. H can F: CmZF
H QEN
c.» mmfi to H Eva no H mdw— mom M62 ed H 3am YN H W? N<m CnZH
0.2 com v.on.2 _.mHm.om fimm vow fioH m: 2 Hm.m~ H3?» CmZP
9.283 :3 353 A25 Fife be 3.52 9:5
evince .5. 55 ex £3. .3 §> ex eiém
3 me boom 3 mg 9.8

 

 

 

 

e.mo>zm>tou ESE: m: 98 Eng mo E80583 38:0. H. 2an

91

 

 

 

 

5 _
A
I
4 5 Q ‘
”>1 li‘Q
:‘é "
s I A ‘
To' U a
g 4 - O . ‘ A
o O A
*- fl
°\° I] 9 . O
E“ [:1 O
._i
3.5 _,
El 3
3
Cl
3 I I I I '
0 100 200 300 400 500
Time (min.)

Figure 3: A) Inactivation curves of TNXI and its mutant derivatives at 80°C (pH 7.0).
Symbols used are the same as in Figure l. Half-lives of TNXI, TNXI V185T, TNXI 3A2,

and TNXI lFl are 1.6 hr, 3.8 hr, 4.5 hr, 6.7 hr, respectively

92

 

 

 

 

 

B
I
4.5
E
Q
E
:3
2
§
if, A
5 0
El
0
3 I I I
O 100 200 300 400

Time (min.)

Figure 3: B) Inactivation curves of TNXI and its mutant derivatives at 80°C (pH 5.5).
Symbols used are the same as in Figure 1. Half-lives of TNXI, TNXI V185T, TNXI 3A2,

and TNXI 1F1 are 1.3 hr, 2.3 hr, 1.7 hr, and 3.0 hr, respectively.

93

(pH 5.5) for various lengths of time (Figure 3). Stability experiments performed with the
metal-free enzymes in 10 mM MOPS (pH 7.0) containing 0.5 mM CoClz showed that
3A2 and lFl (with “/2 of 4.5 hr and 6.7 hr, respectively) were kinetically more stable
than TNXI (ti/2 of 1.6 hr) and TNXI V185T (II/2 of 3.8 hr). At pH 5.5, 1F] (tug of 3.0 hr)
remained more stable than TNXI (ti/2 of 1.3 hr) and TNXI V185T (ti/2 of 2.3 hr); 3A2

was less stable (tug of 1.7 hr).

Amino acid substitutions in TNXI 3A2 and IF]: The mutations present in 3A2 and
IF] were identified by DNA sequencing. In addition to Va1185Thr already present in
TNXI V185T, 3A2 contained a single additional mutation, Leu282Pro. The Leu282Pro
mutation is located in helix 0t7 of the (Oi/[3)3-barrel structure, at approximately 12-14 A
from the catalytic center (Figure 4). Helix a7 itself is located at the surface of a monomer
and at the interface of the dimer. Neither Leu nor Pro’s sidechain can form hydrogen
bonds with neighboring residues. Whenever a proline occurs in a peptide chain, it
interrupts a—helices and creates a kink or bend (Lehninger, 1970). Detailed analysis of the
Leu28Pro mutation modeled into the TNXI structure (Figure 5) suggests that Pr0282’s
pyrrolidine ring (C7, and C5) is in close contact (in some cases ~1.7A) with mainchain
atoms of residues Phe278 and Gln279. With van der Waal’s radii of 1.87 and 1.35 A for
C and O atoms, respectively, in proteins, optimal van der Waal interactions between
carbon atoms of Pro282 sidechain and the mainchain C and O atoms of residues Phe278
and Gln279 would take place at approximately 3.2 A to 3.7 A. The unfavorable van der

Waal contacts (clashes) probably lead to local conformational rearrangements. These

94

changes might, in turn, affect the active site structure and dynamics, the enzyme’s

interaction with the substrate, and probably inter-subunit interactions within the dimer.

1F1 contains the same two mutations as 3A2, plus mutation Phe186Ser. This last
mutation is located in the active site, adjacent to mutation Vall85Thr (Figure 4). Serine’s
sidechain is much less bulky than that of the original Phe. Residue 186’s sidechain points
into the active site cavity, and it is close to the bulky sidechains of residues Tyr184,
Phe228, Phe262, and Leu229. The Phe186Ser mutation probably leads to a
rearrangement of the neighboring residues. This change in local packing may in turn be

responsible for the large increase in low temperature activity of mutant 1F].

95

 

Figure 4: Three-dimensional model of the TNXI lFl monomer showing the positions of

mutations V1 8ST, F1863, and L282P.

96

62 E 550% 2m 2658an RC Nwmoi use A<V Nwmsoq .coocm Em: E :39? mm mwménm 8:28»

do 2:855 2583mm .:o_§=E oimwmzoq m. _ n: 98 N<m HXZH wEvcsotsm 8%.. 2: mo :3 mo 3on EcoficoEmwbuEF um oSwE

 

97

 

Figure 6: Three-dimensional model of part of the TNXI lFl active site showing

hydrogen bonds among S186 (red), L229 (pink) and E231 (blue). E23lco-ordinates Co”

(purple ball) at the structural site (M1).

98

DISCUSSION

Thermostable enzymes are generally barely active at low temperature, but they
are as active at their optimal growth temperature as their mesophilic counterparts (Zeikus
and Brock, 1971; Varley and Pain, 1991). It was shown here that activity of a
hyperthermostable enzyme at low temperature and low pH can be improved without a
loss in its extreme stability. Our results have shown that the quality of the library of
random TNXI mutants was sufficient to isolate mutants with increased activity at low
temperature and low pH. Using two sequential rounds of random mutagenesis, we were
able to obtain a TNXI mutant derivative, 1F 1, showing high activity at low temperature
and low pH. lFl is not only more active overall than its parental enzymes, but it is also
more active especially at low temperature than its parental enzymes. Since 1F 1 is more
active at low temperature with a lower temperature optimum but more stable than the
wild-type enzyme, we suggest that the molecular determinants of this enzyme’s activity
and thermal stability are in fact, not the same. This has also been previously observed in
the study of Bacillus subtilis p-nitrobenzyl esterase, in which the laboratory-evolved
mutant enzyme had a 14°C increase in Tm but still maintained its catalytic activity at low

temperature (Giver et al., 1998).

Recent studies (Aguilar et al., 1997; Zavodsky et al., 1998; Kohen et al., 1999)
suggested that it might be the reduced flexibility of thermostable enzymes that impairs
their catalytic activity at low temperatures. Particularly striking is the potential of single
point mutations to significantly increase low temperature activity. Recent studies of

psychrophilic enzymes have suggested that, despite the many differences observed

99

between mesophilic and psychrophilic enzymes, single amino acid substitutions may be
capable of conferring most psychrophilic characteristics (Somero, 1995; Feller and
Gerday, 1997). Other studies using random mutagenesis and screening/selection
succeeded in increasing the activity (by 3 fold at 20°C for Pyrococcus furiosus ,8-
glucosidase and by 17 fold at 37°C for Sulfolobus solfataricus indole glycerol phosphate
synthase) of hyperthermophilic enzymes at mesophilic temperatures with changes in
temperature optima (Merz et al., 2000; Lebbink et al., 2000). One study even increased
the catalytic efficiency of a mesophilic subtilisin at 10°C by 100 % (Taguchi et al.,

1999).

In our study, TNXI 1F 1 was obtained with 4.5 and 2.2 fold increases in Vm at
60°C (pH 7.0) and 80°C (pH 5.5), respectively, with only 5°C lower optimal temperature
compared to those of TNXI V185T. The Arrhenius plot of activity of TNXI and its
mutants (Figure 1B) revealed that TNXI and TNXI V185T require higher levels of
activation energy for their catalytic activities than either TNXI 3A2 or lFl. The
difference is more pronounced with lFl with an approximately 2-fold decrease in the E3
of activation compared with TNXI and TNXI Vl85T. The reduction of Ea of activity
observed in TNXI 3A2 and IF] suggested improved dynamics and flexibility in the
active site of the enzymes even at low temperature thus their activities at low temperature
are vastly enhanced. Although TNXI V185T has improved catalytic efficiency on glucose
compared to TNXI due to improved glucose binding affinity and higher catalytic rate
(Sriprapundh et al., 2000), its 13a of activity remained similar to TNXI’s suggesting that
its active site dynamics and flexibility remained unchanged. This observation is in good

agreement with the assumption of cold-adapted thermophilic enzymes by Lonn et al.

100

(2002) that mutations underlying the adaptation of enzymes to temperatures lower than
their optima allow a higher degree of flexibility in areas that move during catalysis. This,
in turn, reduces the free energy of activation compared with the wild type enzymes. The
higher flexibility in areas that move during catalysis increases the km, of the reactions
catalyzed by the cold-adapted enzymes. A study of lactate dehydrogenases cold-
adaptation (Fields and Somero, 1998) also found that mutations that increase flexibility in
regions of the enzyme involved in catalytic conformational changes may reduce energy
barriers to these rate-goveming shifts in conformation and thereby increase kw. TNXI
1F 1 has higher km and Km values than TNXI V185T does. This observation was
rationalized in terms of localized increases in conformational flexibility; mutations that
reduce the energetic barriers between different active site conformations (thus allowing
for more rapid interconversion among them) will lead to higher values of km. These same
mutations, however, will allow the enzyme to populate conformations that bind substrate
poorly more easily, leading to increases in Km. Our present study provides obvious
support for this hypothesis.

Our sequential random mutagenesis and screening approach with TNXI resulted
in the identification of amino acid residues or local structural conformations that are
critical for thermostability, metal-binding affinity, and low temperature/low pH catalytic
activity. Two mutations were identified in TNXI lFl in addition to Vl85T, namely
L282P and F186S. Leu282 is at the inter subunit interface of the enzyme dimer. While
the L282P mutation improved the enzyme’s low temperature and low pH activities, the
detailed analysis of the modeled 3D structure of 3A2 revealed unfavorable van der Waal

contacts between Pr0282’s pyrrolidine ring and the enzyme’s backbone structure. These

101

unfavorable contacts probably lead to local conformational rearrangements and make the
enzyme less stable (as observed in the shortened half-life at pH 5.5 compared to TNXI
and TNXI V185T). The second mutation, F1868, is located in the active site, adjacent to
Thr185. Since serine’s sidechain is considerably smaller than phenylalanine’s, this
mutation would create a cavity or increase mobility in the active site of the enzyme
resulting in a dramatic improvement of 1F] ’8 low temperature activity. A potential extra
strengthening H-bond between Ser186’s sidechain and the Leu229 mainchain 0 (<3.2 A)
(Figure 6) is relatively close to E231. This E231 sidechain coordinates metal ion at the
structural site (M1). This proposed extra H-bond in the F1868 mutation might explain the
increased kinetic stability of TNXI 1F 1 that even surpassed that of TNXI.

There are a few evidences that demonstrated the effect of positions where
mutations occur on activity and stability of laboratory and naturally evolved enzymes. A
study of psychrophilic enzymes revealed that amino acid substitutions distant from the
catalytic center or in the major substrate-binding site of enzymes could lead to cold-
adaptation (Feller and Gerday, 1997). In the studies done by Lebbink et al. (2000), all
mutants containing subunit interface substitutions were less stable and had lower
temperature optima than the wild-type Pyrococcus fim’osus fi-glucosidase, suggesting
that subunit interfaces also play an important role in thermoadaptation. Our results
showed that even without selective pressure to maintain thermostability, it is possible to
obtain a mutant thermozyme with a mutation in the active site that has comparable
thermostability with the wild-type enzyme while its low temperature and low pH activity

are vastly enhanced.

102

The only directed evolution approach (or random mutagenesis) performed on
xylose isomerase that has been reported to date was by Lonn et al. (2002). The
thermophilic type I Thermus thermophilus XI was subjected to one round of random PCR
mutagenesis and screening for xylose isomerase activity at lower temperature than
optimal. Three amino acid substitutions were identified as F 163L in the C-terminal tail

domain, and E37ZGN 379A in the (oz/13):; barrel domain. These mutant enzymes showed

improved catalytic rate constants (km) by up to nine times on both xylose and glucose
with up to 26 times higher Km values on xylose but relatively unchanged for glucose. All
enzyme variants’ relative activities on xylose are higher than the wild-type at low
temperatures with lower thermostability. Kinetic analysis demonstrated that the increase
in the relative activity in the enzyme variants for xylose at low temperatures was indeed
caused by an increase in km and not by a decrease in the Km value. This suggests that the
mutant enzymes did not acquire higher affinity for the substrate than the wild-type
enzyme at lower temperatures. These results as well as ours suggested that amino acid
substitutions either distant from or inside the catalytic center can lead to cold adaptation.
The main difference between their work and ours was that we, in fact, were able to
enhance the thermostability of our mutant enzyme while increasing its activity at low
temperatures.

An alignment of different XIs (not shown) revealed that neither Pr0282 nor
Ser186 is present in any known XI. Because it is in the active site, mutation F186S could
potentially have been rationally designed based on the structures of TNXI, TNXI V185T,
and 3A2 in the presence and absence of substrate, and on modeling. In contrast, mutation

L282P, in the middle of an (at-helix, and 12 to 14 A from the active site was completely

103

unpredictable, and could only be obtained through directed evolution. With a vast
improvement in specific activity at 60°C, at pH 5.5, a higher catalytic efficiency on
glucose than TNXI V185T in all conditions tested, and thermostability comparable to that
of TNXI V185T, 1F1 could be an interesting candidate for industrial applications. Further
study of lFl ’s potential usefulness in conditions used in the industrial production of high
fructose syrup in comparison with a commercially available glucose isomerase is

underway.

ACKNOWLEDGEMENT

This work was supported by the National Science Foundation, grant no. BBS-01 15754.,

104

10.

ll

12.

13.

REFERENCES

. Aguilar, C. F., Sanderson, F., Moracci, M., Ciaramella, M., Nucci, R., Rossi, M., and

Pearl, L.H. (1997) Crystal structure of the B-glycosidase from the hyperthermophilic
archeon Sulfolobus solfataricus: resilience as a key factor in thermostability. J. Mol.
Biol. 271, 789-802.

Arnold, F. H., and Moore, J. C. (1997) Optimizing industrial enzymes by directed
evolution. Adv. Biochem. Eng. Biotechnol. 58, 1-14.

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D.D., Seidman, J. G., Smith, J. A.,
and Struhl, K. (1993) Current Protocols in Molecular Biology. (Struhl, K., ed.),
Greene Publishing & Wiley-Interscience, New York

Bhosale, S. H., Rao, M. B., and Deshpande, V. V. (1996) Molecular and industrial
aspects of glucose isomerase. Microbiol. Reviews 60, 280- 300.

Chen, W. (1980) Glucose isomerase. Process Biochem. 15, 30-35.

Feller, G., and Gerday, C. (1997) Psychrophilic enzymes: molecular basis of cold
adaptation. Cell. Mol. Life Sci. 53, 830-841.

Fields, P. A., and Somero, G. N. (1998) Hot spots in cold adaptation: localized
increases in conformational flexibility in lactate dehydrogenase A4 orthologs of
Antarctic notothenioid fishes. Proc. Natl. Acad. Sci. USA. 95, 11476-11481.

Giver, L., Gershenson, A, Freskgard, P. O., and Arnold, EH. (1998) Directed
evolution of a thermostable esterase. Proc. Natl. Acad. Sci. U.S.A. 95, 12809-1281.

Hess, J. M., Tchemajenko, V., Vieille, C., Zeikus, J. G., and Kelly, R. M. (1998)
Thermotoga neapolitana homotetrameric xylose isomerase is expressed as a

catalytically active and thermostable dimer in Escherichia coli. Appl. Environ.
Microbiol. 64, 2357-2360.

Kohen, A., Cannio, R., Bartolucci, S., and Klinman, J. P. (1999) Enzyme dynamics
and hydrogen tunneling in a thermophilic alcohol dehydrogenase. Nature 399, 496-9.

. Kuchner, O. and Arnold, F. H. (1997) Directed evolution of enzyme catalysts Trends

Biotechnol. 15, 523-530.
Lebbink, J. H. G., Kaper, T., Bron, P., van der Oost, J ., and de Vos, W. M. (2000)
Improving low-temperature catalysis in the hyper-thermostable Pyrococcus furiosus

beta-glucosidase CelB by directed evolution. Biochemistry 39, 3656-3665.

Lee, C. and Zeikus, J. G. (1991) Purification and characterization of thermostable
glucose isomerization from Clostridium themzosulfurogenes and

105

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Thermoanaerobacter strain B6A. Biochem. J. 273, 565-571.
Lehninger, A. L. ( 1970) Biochemistry (Worth Publishers, Inc., New York), 1st Ed.

Leung, D., Chen, R., and Goeddel, D. (1989) Systems for heterologous gene
expression. Techniques 1, 11-15.

Lonn, A., Gardonyi, M., van Zyl, W., Hahn-Hagerdal, B., and Otero, R. C. (2002)
Cold adaptation of xylose isomerase fi'om Thermus thermophilus through random
PCR mutagenesis. Eur. J. Biochem. 269, 157-163.

Meng, M., Bagdasarian , M., and Zeikus , J. G. (1993) Thermal stabilization of xylose
isomerase from Thermoanaerobacterium thermasulfurigenes. Proc. Natl. Acad. Sci.
U.S.A. 90, 8459-8463.

Meng, M., Lee, C., Bagdasarian, M., and Zeikus, J. G. (1991) Switching substrate
preference of thermophilic xylose isomerase from D-xylose to D-glucose by
redesigning the substrate binding pocket. Proc. Natl. Acad. Sci. U.S.A 88, 4015-4019.

Merz, A., Yee, M., Szadowski, H., Pappenberger, G., Crameri, A., Stemmer, W. P.,
Yanofsky, C., and Kirschner, K. (2000) Improving the catalytic activity of a
thermophilic enzyme at low temperatures. Biochemistry 39, 880-889.

Ostermeier, M., Nixon, A., and Benkovic, S. (1999) Incremental truncation as a
strategy in the engineering of novel biocatalysts. Bioorg. Med. Chem. 7, 2139-
2144.

Schenk, M. and Bisswanger, H. (1998) A microplate assay for D-xylose/D-glucose
isomerase. Enz. Microbiol. T echnol. 22, 721-723.

Shao, Z., Zhao, H., Giver, L., and Arnold, F. H. (1998) Random priming in vitro
recombination: an effective tool for directed evolution. Nucl. Acids Res. 26, 681-683.

Somero, G. N. (1995) Proteins and temperature. Annu. Rev. Physiol. 57, 43-68.

Sriprapundh, D., Vieille, C., and Zeikus, J. G. (2000) Molecular determinants of

xylose isomerase thermal stability and activity: analysis of thermozymes by site-
directed mutagenesis. Protein Eng. 13, 259-265.

Stemmer, W. (1994) Rapid evolution of a protein in vitro by DNA shuffling. Proc.
Natl. Acad. Sci. USA 91, 10747-10751.

Taguchi, S., Ozaki, A., Nonaka, T., Mitsui, Y., and Momose, H. (1999) A cold-

adapted protease engineered by experimental evolution system. J. Biochem.
126, 689-693.

106

27.

28.

29.

30.

31.

32.

33.

Varley, P. G., and Pain, R. H. (1991) Relation between stability, dynamics, and

enzyme activity in 3-phosphoglycerate kinases from yeast and Thermus thermophilus.
J. Mol. Biol. 220, 531-538.

Vieille, C., Hess, M., Kelly, R. M., and Zeikus, J. G. (1995) xylA cloning and

sequencing and biochemical characterization of xylose isomerase from Thermotoga
neapolitana. Appl. Environ. Microbiol. 61, 1867-1875.

Wintrode, P. L., Miyazaki, K., and Arnold, F. H. (2000) Cold adaptation of a
mesophilic subtilisin-like protease by laboratory evolution. J. Biol. Chem. 275,
31635-31640.

Zavodsky, P., Kardos, J ., Svingor, A. and Petsko, G. A. (1998) Adjustment of
conformational flexibility is the key event in the thermal adaptations of proteins.
Proc. Natl. Acad. Sci. U.S.A. 95, 7406-7411.

Zeikus, J. G., and Brock, T. D. (1971) Protein synthesis at high temperatures:
aminoacylation of tRN A. Biochim. Biophys. Acta. 228, 736-45.

Zhao, H., and Arnold, F. H. (1999) Directed evolution converts subtilisin E into a A
functional equivalent of thermitase. Protein Eng. 12, 47-53.

Zhao, H., Giver, L., Shao, Z., Affholter, J., and Arnold, F. H. (1998) Molecular

evolution by staggered extension process (StEP) in vitro recombination. Nat.
Biotechnol. 16, 258-261.

107

CHAPTER V

BIOCHEMICAL COMPARISON OF STABILITY, ACTIVITY, AND FRUCTOSE
SYNTHESIS BY AN INDUSTRIAL GLUCOSE ISOMERASE (GENSWEETW)

VERSUS A LABORATORY-EVOLVED THERMO-ACID STABLE XYLOSE

ISOMERASE (T NXI lFl)

108

ABSTRACT

Biochemical properties and fructose productivities of a laboratory-evolved xylose
isomerase from hyperthermophilic Thermotoga neapolitana (TNXI lFl) and GensweetTM
(Genencor, Rochester, NY), a commercially available glucose isomerase from a
genetically modified strain of Streptomyces rubiginosus were compared. TNXI lFl
displayed higher catalytic efficiencies on glucose at low or high temperature and pH
ranges and had greater thermal stability than GensweetTM despite having similar
temperature optima for activity. This greater thermal stability together with the superior
kinetic parameters on glucose render TNXI lFl a genuine candidate for the industrial
glucose isomerization process based on the lifetime fructose productivity estimation. At
high temperature and neutral to alkaline pH, the Maillard browning reaction is a major
concern in the resulting syrups. This was overcome by using TNXI lFl for fructose

production at 90°C and pH 5.5-6.5.

109

INTRODUCTION

The production of high fructose corn syrup (HFCS) using immobilized glucose
isomerase (G1) is considered one of the largest commercial enzymatic processes
(Klibanov, 1983a; and Bhosale et al., 1996). The last step of the process is the enzymatic
isomerization of glucose into a mixture typically containing 42 % fructose with 51 %
glucose and 7 % oligosaccharides (Visuri and Klibanov, 1987). A costly fructose
enrichment step is then typically employed to increase the fructose concentration to 55 %
level to give the same sweetness level as sucrose (at the same concentration of solids) for
its major use (Bucke, 1981). The enzyme-catalyzed isomerization of glucose into fructose
is performed in industrial bioreactors at 55-60°C where the half-life of G1 is on the order
of several weeks. Even a modest increase in the half—life of the enzyme will substantially
reduce the cost of the HFCS production (Klibanov, 1983b).

Since most industrially employed GIs exhibit temperature optima in the range of
80-90°C (Hartley et al., 2000), only insufficient operational stability precludes their use
at these higher temperatures, which would be highly beneficial. Not only thermostability
of commercial enzymes is the main limiting factor of such application at high
temperature, but in high substrate concentrations, the Maillard browning reaction of the
enzyme with reducing sugars (e.g. glucose and fructose) is also the dominant reason for
enzyme inactivation. The Maillard reaction was considerably faster than other
inactivation mechanisms (Visuri et al., 1999). Theoretically, the Maillard reaction can be
retarded at low temperature or low pH. Hence the use of highly thermostable GIs at

higher temperature and lower pH than the current isomerization conditions would result

110

in more operational stability with reduced browning reactions. The proposed process
would be ideal since previous liquefaction and saccharification steps were also performed
at high temperature and low pH and minimal adjustment of isomerization condition
would be required.

Xylose isomerase (XI) from the hyperthermophilic eubacterium Thermotoga
neapolitana is one of the most thermostable characterized XIs (Vieille et al., 2001). The
gene encoding the enzyme was cloned, sequenced, and expressed in Escherichia coli
(Vieille et al., 1995). Its active site was also engineered by site-directed mutagenesis to
increase its activity on glucose rather than its natural substrate, xylose (Sriprapundh et al.,
2000). The TNXI V185T mutant derivative is more active, more glucose efficient, and
more stable than the wild-type TNXI. This enzyme’s activity on glucose at low
temperature and low pH was recently further improved by directed evolution resulting in
the TNXI 1F 1 derivative (V185T/LZ82P/F186S) with higher overall activity on glucose
throughout the temperature and pH ranges compared to TNXI V185T (Sriprapundh et al.,
submitted). Despite its higher activity at low temperature, TNXI 1Fl remains relatively
as stable as TNXI V185T and is more stable than the wild-type TNXI. With such vast
improvement in every biochemical aspect of TNXI IF], it would be interesting to
compare its utility against a commercially available glucose isomerase to determine
whether it can be genuinely considered for industrial applications.

In this study, the effect of temperature and pH on specific activity on glucose

together with kinetic parameters and thermal stability in different incubation conditions
were compared between TNXI 1F] and GensweetTM, a commercially available glucose

isomerase from a genetically modified strain of Streptomyces rubiginosus. Furthermore.

11]

their fructose production in various combinations of pHs and temperatures as well as
lifetime fructose productivities of each enzyme under these conditions were studied. We

show that TNXI 1F] has superior properties as an industrial glucose isomerase.

112

MATERIALS AND METHODS

Enzyme source: GensweetTM SGI, a xylose isomerase derived from a genetically
modified strain of Streptomyces rubiginosus, was provided by Dr. Jay Shetty of Genencor
International (Rochester, NY) as a kind gift. SDS-gel electrophoresis showed that the

enzyme was pure.

Protein Purification: TNXI 1F 1 was purified using the procedure of Vieille et al. (1995)
followed by an additional ion-exchange chromatography step. Partially purified enzyme
was applied to a DEAE-Sepharose column (2.5x15 cm) equilibrated with buffer A, and
the enzyme was eluted using a 500 ml linear 0—300 mM NaCl gradient in buffer A. The
pooled fractions from the DEAE-Sepharose column were concentrated in a stirred
ultrafiltration cell (MW cut-off 30 kDa) (Amicon, Beverly, MA) and dialyzed twice
against buffer A. Concentrated, homogenous enzyme was dispensed and stored frozen at

-70°C.

Glucose isomerase assays: TNXI 1F] and GensweetTM were assayed routinely with
glucose as the substrate. The enzyme (1-1.5 mg/mL) was incubated in 100 mM MOPS
buffer (pH 7.0) [or 100 mM sodium acetate buffer (pH 5.5)] containing 1 mM CoClg and
0.4 mM glucose at 80°C for 10 min. The reaction was stopped by transferring the tube to
an ice bath. The amount of fructose produced was determined by the modified resorcinol-
ferric ammonium sulfate-hydrochloric acid method (Schenk and Bisswanger, 1998). To

determine the effect of temperature on the activity of TNXI 1F] and GensweetTM, the

113

holo-enzyme was incubated in the reaction mixture at the temperatures of interest in a
heated water bath (45-95°C) or a heated oil bath (95-1 10°C) for 10 min. The effect of pH
on activity was determined using the routine assay described above except that the MOPS
buffer was substituted with 100 mM sodium acetate buffer (pH 4.3-5.8), 100 mM PIPES
buffer (pH 6.1-7.0), or 100 mM EPPS buffer (pH 7.2-8.l). All pHs were adjusted at room
temperature, and the ApKa/At’s for acetate, PIPES, and EPPS (0, -0.0085, -0.01 1,
respectively) (USB, Cleveland, OH) were taken into account for the results. To determine
the kinetic parameters, assays were performed in the presence of 10-1,500 mM glucose,
50 mM MOPS (pH 7.0) and 1 mM CoClz. One unit of glucose isomerase activity is
defined as the amount of enzyme that produces 1 pmol of fructose per minute under the

assay conditions.

Thermal inactivation assays: To obtain the apo-enzymes (metal-free enzymes), the
purified enzymes were incubated overnight at 4°C in 50 mM MOPS buffer (pH 7.0)
containing 10 mM EDTA. They were then dialyzed twice against 50 mM MOPS buffer
(pH 7.0) containing 2 mM EDTA, and they were finally dialyzed twice against 50 mM
MOPS buffer (pH 7.0) without EDTA. CoClz (0.5 mM) was added to the apo-enzyme
and equilibrated at 4°C overnight before the thermoinactivation was initiated. The time

course of irreversible thermoinactivation was measured by incubating the enzymes (0.1-
0.2 mg/ml) in either 10 mM MOPS buffer (pH 7.0) or 10 mM sodium acetate (pH 5.5) at

various temperatures for different periods of time in a heated water bath. Residual

glucose isomerase activity was measured as described above at 80°C. The first order rate

114

constant, k, of irreversible thermoinactivation was obtained by linear regression in semi-

log coordinates. Enzyme half-lives were calculated from the equation: tum = ln2/k.

Differential Scanning Calorimetry (DSC): DSC experiments were performed on a
Nanocal differential scanning calorimeter (Calorimetry Sciences Corp., Provo, UT) using
a scan rate of 1°C/min. Samples were scanned from 25°C to 100°C. The apo-enzymes
were scanned against 50 mM MOPS (pH 7.0). Enzymes containing both Mg2+ and Co”

were dialyzed against buffer A, then scanned against the dialysis buffer as control.

Fructose production experiments: TNXI 1F1 and GensweetTM (50 ug) were incubated
in 1 ml reaction in capped 1.5 ml tubes containing 2.5 M glucose, 5 mM MgSO4, and 50
mM of either MOPS (pH 7.0) or sodium acetate (pH 5.5) at various temperatures for up
to 24 hours. The reactions were then stopped on ice and were assayed for fructose
produced by the method described above. Browness of resulting syrups was monitored by

maximal absorbance at 425 nm.

115

Modeled fructose productivity: Lifetime fructose productivity of TNXI lFl and
GensweetTM were estimated using the one phase inactivation mathematical model
(Bandlish et al., 2002). Kinetic and inactivation data for the soluble enzymes were used
to eliminate variations due to non-optimal immobilization protocols and the potential
influence of mass transfer limitations on immobilized enzyme kinetics. The final
equations presented a derivative of parameters for GI kinetics that also considered the

equilibrium between glucose and fructose:

are; lGlul [kid
[E0] [GIU]+Km [le

 

P_ = [Glu] [1-exp(-th)][k_cfl1
[Ea] [GIUHKm [kn]

Using these equations, enzyme productivities (kg fructose per kg enzyme) were
calculated using representative data for the soluble enzymes. The calculations assumed 3
M glucose feed, a representative of an industrial HFCS production (Pedersen 1993;
Godfrey and West, 1996). Enzyme productivity, P, is defined as the total amount of

glucose converted to fructose per unit amount of enzyme during a period of time.

116

RESULTS

Effect of temperature and pH on TNXI 1F] and GensweetTM activities: The effect of
temperature on TNXI 1F] and GensweetTM glucose isomerase activities is shown in
Figure 1. Both enzymes have comparable specific activities on glucose in the temperature
range of 45-75°C. The optimal temperatures of glucose isomerase activity of TNXI 1F]
and GensweetTM are 90°C and 85°C, respectively. Although they have relatively the same
activity at low temperature, TNXI 1F] is much more active at its optimal temperature of
activity with 47.6 U/mg compared with 30.9 U/mg of GensweetTM. Figure 2 demonstrated
the effect of pH on glucose isomerase activities at 80°C of TNXI 1F 1 and Gensweet'lM.
The two enzymes have comparable activity over the pH range of 4.6 to 8.1 and retain

more than 70 % of their optimal activities in the pH range of 6.1 to 8.0.

Thermal stability of TNXI 1F 1 and Gensweetm: To determine the thermal stability of
TNXI lFl and GensweetTM, the residual activities of TNXI 1F] and GensweetTM were
measured after heat treatment at 60°C, pH 7.0 and 5.5 for various lengths of time

(Figures 3, A and B). Investigations of the metal-free enzymes in buffer at saturated Coy

concentration showed that at both pHs, TNXI 1Fl is far more superior in term of thermal
stability when compared to GensweetTM with half-lives of 1 15.5 and 38.5 hr at pH 7.0 and
5.5, respectively, compared to 2.9 and 1.7 hr for GensweetTM. The TNXI lFl is more
stable than GensweetTM by 1.6 and 1.36 orders of magnitudes at 60°C, pH 7.0 and 5.5,
respectively. These results suggest that although GensweetTM and TNXI 1F 1 have almost

the same temperature of optimal activity (Topt), only TNXI lFl has extreme

117

50

 

 

 

 

 

40-

a:

E

El 30‘ '

Zr ' . 0

IE -

a

“ /

5% 20-

8 .

0..

U)
10- .

O
O

OI I l I
40 60 80 100

Temperature (C)

Figure 1: Effect of temperature on the specific activities of TNXI IF] (I) and

GensweetTM (O) on glucose at pH 7.0.

118

40

 

30-

ml

Specific activity (U/mg)

10m

 

 

 

 

Figure 2: Effect of pH on the specific activities of TNXI IF] (I) and GensweetTM (O) on

glucose at 80°C.

119

thermostability. GensweetTM, on the other hand, is thermophilic but obviously not
extremely thermostable.

TNXI 1F] and Gensweetm’s melting temperatures (Tm) were determined by DSC
in the presence and absence of metals (Figure 4, and Table 2). With the holo-enzymes,
both scans revealed one thermal transition, at 107.3°C and 934°C, for the TNXI 1F] and
Gensweetm, respectively. With a higher Tm of ~14°C, TNXI 1F1 holo-enzyme is much
more stable than that of GensweetTM which is in good agreement with the results obtained
from inactivation experiments. DSC of apo-enzymes of both enzymes also revealed one
thermal transition 784°C and 761°C for the TNXI 1F 1 and Gensweetm, respectively.

The extent of Tm difference observed in apo-enzymes was not as pronounced as that in

holo-enzyme forms

120

 

Ln (%residual activity)

 

 

 

 

1 I I I

0 20 40 60 80

Time (hr)

Figure 3: (A) Inactivation curves of TNXI 1F] and GensweetTM at 60°C (pH 7.0).

Symbols used are the same as in Figure 1.

121

 

 

 

 

5 _
B
4.5
E
g 4-1
E3:
'5
3
E
3
2 3.5-
E
C
.—l
3d
2.5 I I r
O 10 20 30 40

Time (hr)

Figure 3: (B) Inactivation curves of TNXI 1F1 and GensweetTM at 60°C (pH 5.5).

Symbols used are the same as in Figure 1.

122

100 ewMeg—enmmmm----w-__....-a"--- W.H..

O\\IOO\O
COCO

MHC (KcaI/mole*K)
r—‘ N DJ A kit
O O O O O

 

 

 

 

O

l l . i

70 80 90 100 110 120 130.

 

Te mpe rature (°C)

Gensweet B

 

450
400 .
350 i
300 i
250 >

p—

£11

0
l

MHC (Kcal/mole*K)
'— N
O O
o c

U:
C

 

 

 

O

I I

70 80 90 100 110 120

O\
O

 

 

 

 

 

Te mpe rature (°C)

 

Figure 4: Thermal unfolding of (A) TNXI 1F 1 and (B) GensweetTM in the presence of 5

mM MgSO4 and 0.5 mM COC12 followed by DSC.

123

Kinetic parameters of TNXI lFl and Gensweetm: The kinetic parameters on glucose
of TNXI 1Fl and GensweetTM were compared in different conditions (Table 1). In all
comparable conditions, GensweetTM has higher Km and Vmax than TNXI lFl (except at
80°C, pH 5.5 in which TNXI lFl’s Vmax is higher than that of GensweetTM). The
difference in Km of GensweetTM and TNXI lFl is more pronounced than that of Vmax
resulting in worse catalytic efficiency (kw/Km) on glucose for GensweetTM than TNXI
lFl. The TNXI 1F1’s superiority of glucose catalytic efficiency on glucose is more noted
at pH 7.0 than at pH 5.5 and at higher (95°C) or lower (60°C) than both enzymes’

optimal temperatures.

Modeled fructose productivity: The lifetime fructose productivity of both enzymes was
estimated using the one phase inactivation model at various combinations of temperatures
and pHs. The modeled time course of fructose productivity of TNXI 1F] and GensweetTM
at various conditions is shown in (Figure 5 and Table 3). Fructose productivity of
GensweetTM at 80°C cannot be generated because no residual activity was detected at
80°C in buffer pH 7.0 and 5.5 after just 10 minutes. At 60°C, GensweetTM produced a
maximum amount of 1.3 and 0.4 kilogram (kg) fructose per gram (g) enzyme at pH 7.0
and 5.5, respectively. With lifetime fructose production of 30.5 and 4.4 kg fructose/g

enzyme at pH 7.0 and 5.5, respectively, TNXI 1F1 yielded approximately 24 and 12-fold

124

85:53 82 dz...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

he: mww Y: N60 04» Nwm m8 wwh 7.: Enz.—r
0; wow_ 0. :u ofio: 0d *2: m.mm _.©mw_ anokwmcomv
0000
Tum Oma_ aSm O.Nm 0.0 was; m.mm QVBN 7.: CHE
Nam wabm mane vdav YN 5N2 hdm ode 53026.50
Doom
CZ DZ OZ DZ 0.x. CONN _.mm w..vmm T: CHE
DZ DZ DZ DZ DZ DZ DZ CZ Euooamcomv
Uooo
h6~ mwOM ode céwfi DZ DZ DZ DZ 7: USAF
m; wwwm $60 méwNN DZ DZ DZ *DZ 552.550
Uoma
nemwfic pee... £83 Ea} 5%me re} £83 Eaves. ease
o.» ma 3 ma

 

 

2.803800 .m> _ n: Caz... mo 880883 0:35 nonmwtoaofl 883w mo cemwaafioo L 2an

125

_-. W, _ , ,,, ._ .. . a 1
+Gensweefi00lpH 7.0 l
—O-—Genswee160CIpH 5.5 !
+TNXI 1F1 GOO/pH 7.0
—D—TNX| 1F1 SOC/pH 5.5
—I—TNX| 1F1 soc/pH 7.0

 

4000

3000

 

2000

Productlvlty (g Fructose/g Enzyme)

 

 

 

 

 

1 3654 7306 10959 1461 1

 

 

 

 

Time (min.)
35000
+ GenswEt-BOCIpH 7.0
30000 , '
—o— Gensw eet 60C/pH 5.5

—I- TNXI 1 F1 60ClpH 7.0
,--D— INXI1F1 SOC/pH 5.5
l —l— INXl1F1 BOOpH 7.0
20000 —D— TNXI 1 F1 BOCJpH 5.5
15000

10000

Productivity (9 Fructose/g Enzyme)

5000

 

 

  

 

 

0 E'
1— O) (O (V) 0 w to N O) h v ‘- w to t') O N
w h (D m 0') N V- O') 00 Ix (D (O N ‘- O)
N m w ‘- g N O N In w r- l~ O 0') LO
(0 (D O, n O) 0) to O) N (D N (D 0') N
v— F v- N N N (*) (fl 0') V V V In
Time (min.)

 

l
1

Figure 5: Estimated fructose productivity of TNXI 1F] and GensweetTM in different

co nditions.

126

increases in fructose production compared to GensweetTM under the same conditions. The
main reason for superior fructose productivity by TNXI 1F1 over GensweetTM is mainly
due to its higher thermostability. It is also important to note that at 60°C the fructose
production of GensweetTM reached the maximum points before 24 hr whereas it took
approximately 15 days at pH 5.5 and more than 30 days at pH 7.0 for TNXI 1F1 to reach
its maximum production. At 80°C, TNXI 1F1 produced 4.5 and 2.4 kg fructose/g enzyme
at pH 7.0 and 5.5, respectively. The fructose production for TNXI 1F1 at 80°C, both pH

7.0 and 5.5, reached the maximum points at approximately 2.5 days.

Fructose production experiments: Fructose production by TNXI 1F1 vs. GensweetTM
with 45 % glucose syrup at various combinations of temperatures and pHs was performed
to study the effect of both temperature and pH on fructose conversion ratio (compared to
glucose) and potential browning reactions with each enzyme (Figure 6, 7 and Table 3).
To prove that higher isomerization yield of fructose may be achieved by increasing the
reaction temperature, TNXI 1F1 and GensweetTM were incubated with glucose syrups
(pH 7.0 or 5.5) at 60, 80, and 90°C for up to 24 hr. At both pH, increases in fructose
conversion was observed to be proportional to higher temperature for up to at least 24 hr
in all cases with one exception. The syrup incubated with TNXI 1F1 at 90°C, pH 7.0
showed a higher conversion percentage for up to 6 hr after which the conversion rate
remain relatively constant and its fructose conversion percentage was surpassed by that of

the syrup incubated with TNXI 1F1 at 80°C, pH 7.0. The explanation for this event might

be due to TNXI lFl’s relatively short half-life at high temperature above 85°C. The

127

 

 

OJ
U!

00
O

   
 

; —I-TNXI 1F1 BOO/pH 7.0 '

:—D—TNXI 1F1 80C/pH 5,5 l
.» +Gensweet 8OC/pH 7.0

l —O— Gensweet 80C/pH 5.5 i

N
(11
Hip—i

% Fructose
.1 N
01 o

l
l

 
  
    

  
 

 

 

 

 

 

  
  

10
l
5
0 l t l | I l l I l l T f J
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (hr)
25 l B
a, 3:. z . A-
a) 15 _ ~—I—TNX11F160C/pH7.0
.3 —D-TNXI 1F1 60C/pH 5.5
.3. +Genswee160C/pH 7.0
I; 10 . —O—Gensweet SEQ/r1146;

(J1

 

 

 

0: i l l 1 1T 1 l I in
0 2 4 6 81012141618202224

Time (hr)

 

 

Figure 6: Experimental fructose conversion of TNXI 1F 1 and GensweetTM at pH 7.0 and

5.5. at 80°C (A) and at 60°C (B).

128

 

  
   

 

 

 

f: 1
l 2-0 —u- TNX|1F1 pH 7.0 90c
l -D— TNXI1F1 pH 5.5 soc
1 5 +TNXI 1F1 pH 7.0 soc
a ' —O—TNXI 1F1 pH 5.5 80C
3 + Gensweet pH 7 0 80C
0 1.0 , figs-censweet pH 5.5 800
0.5
0.0 . . . i . . .
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (hr)

 

Figure 7: Browness of syrups from experimental fructose conversion of TNXI 1F1 and
GensweetTM. The time course of browning reactions was monitored by maximal

absorbance at 425 nm.

 

reactions were performed at two different pHs, 7.0 and 5.5, to also investigate the
feasibility of prevention of browning the syrups that occurs at high temperature and pH.
The brownness of the resulting syrups was monitored at the maximum wavelength of
absorbency, which is 425 nm. As expected, the brownness of the syrups was observed to
be most pronounced at the highest temperature tested (90°C) and pH (7.0). Syrups
resulting from reactions at low pH (5.5) have dramatically fewer problems with
browning.

Fructose conversion percentage was compared between TNXI IF] and
GensweetTM. At 80°C, TNXI 1F1 compared favorably with Gensweet with slightly higher
conversion percentage at 24 hr (32 % for TNXI 1F1 vs. 28 % for GensweetTM), which is
possibly due to GensweetTM’s less stable nature at high temperature. However, at 60°C,
TNXI 1F1 was obviously better than Gensweet at converting glucose to fructose with

higher conversion rate throughout the time course of 24 hr.

130

DISCUSSION

The initial goal of this study was to compare biochemical and kinetic parameters
as well as productivities of a laboratory-evolved xylose isomerase TNXI IF] and a
commercially available glucose isomerase, GensweetTM to ascertain that TNXI 1F1 can
be genuinely considered for industrial glucose isomerization. Table 2 summarizes key
properties of the two enzymes. Not surprisingly, TNXI 1F1 compares favorably with
GensweetTM in every aspect. The key factor that distinguishes the two enzymes was
shown to be their thermal stability difference. Although TNXI 1F] and GensweetTM have
almost the same apparent temperature optima (e.g., 90°C and 85°C, respectively),
GensweetTM is much less thermostable than TNXI 1F1 by more than one order of
magnitude at 60°C at pH 7.0 or 5.5.

A mathematical model derived to account for the effect of temperature on
reversible enzyme kinetics, inactivation rates, and the glucose-fructose chemical
equilibrium (Bandlish et al., 2002) was used to estimate their lifetime fructose
productivity. Because Km, km, and k1) are based on soluble enzyme data, the effect of
immobilization is not taken into account. However, these estimates provide useful
information concerning the potential of the enzyme for HFCS production under optimal
conditions. TNXI 1F1 has the lifetime fi'uctose productivity at 60°C, pH 7 .0 of 30.5 kg
fructose/ g enzyme whereas Gensweet”, which reached its maximum fructose conversion
in less than a day due to its limited thermal stability, produced only 1.3 kg fructose/g

enzyme. TNXI 1F1 ’s estimated greater fructose productivity mainly resulted fi'om both

131

Table 2: Comparison of thermal activity and stability properties of TNXI 1F1 vs.

 

 

 

 

 

 

 

 

 

GensweetTM
Properties GensweetTM TNXI 1F1
Tom (°C) 85 90
Optimal pH 7.5 6.7
Tm of Holo-enzyme (°C) 93.4 107.3
Tm of Apo-enzyme (°C) 76.1 78.4
Tm at 60°C/pH 7.0 (hr) 2.9 115.5
TI/z at 60°C/pH 5.5 (hr) 1.7 38.5

 

 

132

 

greater thermal stability and better catalytic efficiency on glucose compared to those of
Gensweet”.

Experimental fructose conversion was performed with 45 % glucose syrups
incubated with 50 ug of either TNXI 1F1 or GensweetTM in various conditions to
simulate industrial conditions and also to study the effect of temperature and pH on
browning reactions resulting from interactions of enzymes with reducing sugars. TNXI
1F1 has a slight edge in term of fructose conversion ratio in every condition tested
compared to GensweetTM with maximal conversion observed at 80°C, pH 7.0 (32 %). The
browness of resulting syrups was monitored up to 24 hr at maximal absorbance of 425
nm. Browning of syrup occurs much more pronounced at 90°C, pH 7.0 and can be greatly
reduced by either lowering the pH, the reaction temperature or both. It should also be
noted that at 60°C, browness of the resulting syrup is marginal. When the enzyme
concentration in the experiment was increased from 50 g to 100g and 1 mg, not only did
we see more pronounced browning, but precipitates were also observed in the resulting
syrups (data not shown).

A study of a glucose isomerase from Streptomyces rubiginosus by Visuri et al.
(1999) suggested that in an industrial process, glucose isomerase inactivation is caused
mainly by a Maillard-type browning reaction between the enzyme and the reactive
substrates glucose and fructose resulting in inactive glycated protein complexes. From
our data, we can speculate that using TNXI 1F1 at 60°C at pH 7.0 or even at slightly
higher temperature or lower pH would result in sufficiently high fi'uctose yield and
productivity as well as reduced concern on Maillard browning reaction. Further detailed

study of immobilized TNXI 1F1 and GensweetTM on the production of high fructose

133

8555.8 .2 dz...

 

 

 

 

 

 

 

 

 

 

 

 

.d v _.N OZ mnmv own Wm oo
_.o V a6 2 Ommom we: o8 oc
mo age. 2 3mm VNM: Wm cm
0; 92 cm move wvcm o8 ow
56 5o DZ 02 OZ Wm om
NN WE we Q2 Q2 o6 oo 7.: HXZH
_.o v n; OZ vom wmm Wm o0
mo v ad 2 £2 00: ox. an
_.o Wm _ _ DZ OZ Wm ow
: 02 NN DZ .32 OS ow 3583300
C: m BEANS onEhNam onPSNcm A V
Ea E .58 80 2: 80 2: 3.0 2: o
Amsmwwmmcv 0 8885 \ aha. a we mean w» e... 2% :a amea. eEEm
omowoam e\e
53:36. 2&0on 33:03:65 535355

 

 

 

 

 

 

 

 

 

5863980 .m> _ n: 59E. mo £80888 cote—68m 882E mo acmEquU “m 2an

134

syrup in different conditions in underway to ascertain the potential use of TNXI 1F1 in

industrial application.

ACKNOWLEDGEMENT

This work was supported by the National Science Foundation, grant no. BES-
0115754. We express our gratitude to Dr. Jay Shetty of Genencor International

(Rochester, NY) for providing GensweetTM for this work.

135

ll.

12.

13.

REFERENCES

Bandlish, R. K., Hess, J. M., Epting, K. L., Vieille, C., and Kelly, R. M. (2002)
Glucose-to-fructose conversion at high temperatures with xylose (glucose) isomerases

from Streptomyces murinus and two hyperthermophilic Thermotoga species.
Biotechnol. Bioeng. 80, 185-194.

Bentley, L. S. and Williams, EC. (1996) Starch conversion. In Industrial Enzymology
(Godfrey, T and West, S. I. eds), pp. 339-357. Stockton Press, New York.

Bhosale, S. H., Rao, M. B., and Deshpande, V. V. (1996) Molecular and industrial
aspects of glucose isomerase. Microbial. Reviews 60, 280- 300.

Bucke, E. (1981) Enzymes and Food Processing (G. G. Birch, N. Blakebrough, and
K. J. Parker, eds), 51-72. Applied Science Publishers, London.

Hartley, B. S., Hanlon, N., Jackson, R. J ., and Rangarajan, M. (2000) Glucose
isomerase: insights into protein engineering for increased thermostability. Biachim.
Biophys. Acta. 1543, 294-335.

Klibanov, A. M. (19833) Approaches to enzyme stabilization. Science 219, 722.

Klibanov, A. M. (1983b) Stabilization of enzymes against thermal inactivation.
Advan. Appl. Microbial. 29, 1-28.

Pedersen, S. (1993) Bioprocess technology Tanaka A., Tosa T., Kobayashi T., Eds.
Marcel Dekker, New York 16, 185-208.

Schenk, M. and Bisswanger, H. (1998) A microplate assay for D-xylose/D-glucose
isomerase. Enz. Microbial. Technol. 22, 721-723.

. Sriprapundh, D., Vieille, C., and Zeikus, J. G. (2000) Molecular determinants of

xylose isomerase thermal stability and activity: analysis of thermozymes by site-
directed mutagenesis. Protein Eng. 13, 259-265.

Sriprapundh, D., Vieille, C., and Zeikus, J. G. (2002) Directed evolution of
Thermotoga neapolitana xylose isomerase: high activity on glucose at low
temperature and low pH. Protein Eng. (submitted)

Vieille, C., Hess, M., Kelly, R. M., and Zeikus, J. G. (1995) xylA cloning and

sequencing and biochemical characterization of xylose isomerase from Thermotoga
neapolitana. Appl. Environ. Microbial. 61, 1867-1875.

Vieille, C., Sriprapundh, 8., Kelly, R. M., and Zeikus, J. G. (2001) Xylose
isomerases from Thermotoga. Methods Enzymol. 330, 245-249.

136

14. Visuri, K. and Klibanov, A. M. (1987) Enzymatic production of high fructose corn
syrup (HFCS) containing 55% fructose in aqueous ethanol. Biotechnol. Bioeng. 30,
917-920.

15. Visuri, K., Pastinen, 0., Wu, X., and Makinen, K. (1999) Stability of native and
cross-linked crystalline glucose isomerase. Biotechnol. Bioeng. 64, 377-380.

137

 

CHAPTER VI

ROLE OF METALS IN TNXI STABILITY AT EXTREMELY HIGH

TEMPERATURES

138

ABSTRACT

TNXI and its directed evolution mutants’ melting temperatures (Tm) were
determined by DSC in the presence and absence of metals. With the holo-enzymes, all
scans except TNXI 1F1 revealed two thermal transitions. Surprisingly, TNXI 1F1
showed only one thermal transition, at a slightly lower temperature than those of the
wild-type TNXI and TNXI V185T second thermal transitions showing that 1F1 lacks the
first thermal transition common in the wild-type TNXI, TNXI V185T, and TNXI 3A2
mutant. DSC of apo-enzymes of all TNXIs revealed only one thermal transition. The
result suggested that the additional mutation that occurs in TNXI 1F1 mutant, Phe186Ser,
is responsible for altered metal-binding property and also strongly enhances the metal
requirement of TNXI as seen in a very low Tm of TNXI 1F1 apo-enzyme compared to
those of the wild-type TNXI and TNXI V185T relative to their holo-enzymes. With a Tm
difference of 289°C (approximately 2-fold higher that of TNXI) between apo- and holo-
forms of TNXI 1F1, its extreme thermal stability is even more strongly metal—dependent

than the wild-type TNXI.

139

INTRODUCTION

Xylose isomerase from hyperthermophilic eubacterium T hermotoga neapolitana
(TNXI) is optimally active at a temperature of 95°C or above. Its xylA gene was cloned,
sequenced, and expressed in Escherichia cali, which yielded a recombinant XI with
catalytic characteristics identical to those of the native enzyme (Vieille et al., 1995).
Further study by gel filtration chromatography showed that the recombinant enzyme was
both a homodimer and a homotetramer, with the dimer being the more abundant form
(Hess et al., 1998). The ratio of dimer to tetramer was approximately 20:1, based on total
protein assay data (Bradford, 1976). The two forms had comparable stabilities when they
were thermoinactivated at 95°C. Differential scanning calorimetry (DSC) revealed
thermal transitions at 99 and 109.5°C for both forms suggesting that the association of the
subunits into the tetrameric form may have little impact on the stability and biocatalytic
properties of the enzyme (Hess et al., 1998).

Typically for X18, two divalent cations (Mg2+, (302+, and Mn2+) per monomer are
required for catalytic activity and stability (Hess et al., 1998) . The three metals activate
TNXI with C02+ being the best activating metal. Activity of the Mg2+-TNXI and the
Mn2+-TNXI are approximately 40 and 16 %, respectively, of the activity observed with
the Coz+-TNXI. Also, the stabilization provided by metals to TNXI is metal specific: the
Mn2+-TNXI is significantly more stable than the C0931 and Mg2+-TNXI (Vieille et al.,
2001).

The active site of TNXI has been previously engineered to improve its catalytic

efficiency toward glucose (Sriprapundh et al., 2000). The TNXI V185T is the most

140

efficient site-directed mutant with a 3.1 fold increase in its catalytic efficiency toward
glucose and comparable kinetic properties on glucose and xylose. To further optimize
this enzyme’s potential industrial utility, directed evolution (sequential random
mutagenesis and low temperature/low pH activity screening) was successfully applied to
the TNXI V185T-encoding gene to obtain enzymes that have high glucose isomerase
activity at low temperature and low pH. The best mutant enzyme, TNXI 1F1 (containing
V185T, L282P, and F186S mutations), was dramatically more active than TNXI V185T
at all temperatures and pHs tested. TNXI 1F1 is also more kinetically stable than TNXI
and TNXI V185T. TNXI 1F1 ’5 enhanced stability is thought to be a result of additional
H-bond formation between Ser186’s sidechain and the neighboring L229 residue’s
mainchain structure. This, in turn, strengthens local conformation and the affinity of
E231 co-ordination with the structural metal, hence improving the thermostability of the
mutant enzyme.

In this study, the thermodynamic stability of TNXI, TNXI V185T and its directed
evolution mutant derivatives, TNXI 3A2 and IF] were followed by DSC in both apo-
and holo- forms to compare and contrast their unfolding behaviors with regard to metal
ions requirement for their thermostability. We show here that increased thermal stability
of xylose isomerase is associated with metal binding especially in TNXI 1F1 where the

apo-enzyme was 30°C less stable than the holo-enzyme.

141

MATERIALS AND METHODS

Protein Purification: TNXI and its mutant derivatives were purified using the procedure
of Vieille et al. (1995) followed by an additional ion-exchange chromatography step.
Partially purified enzyme was applied to a DEAE-Sepharose column (2.5x15 cm)
equilibrated with buffer A (50 mM MOPS pH 7.0, 5 mM MgSO4, 0.5 mM CoClz), and
the enzyme was eluted using a 500 ml linear 0—300 mM NaCl gradient in buffer A. The
pooled fractions from the DEAE-Sepharose column were concentrated in a stirred
ultrafiltration cell (MW cut-off 30 kDa) (Amicon, Beverly, MA) and dialyzed twice
against buffer A. Concentrated, homogenous enzyme was dispensed and stored frozen at

-70°C.

Differential Scanning Calorimetry (DSC): DSC experiments were performed on a
Nanocal differential scanning calorimeter (Calorimetry Sciences Corp., Provo, UT) using
a scan rate of 1°C/min. Samples were scanned from 25°C to 100°C. To obtain the apo-
enzymes (metal-free enzymes), the purified enzymes in buffer A were incubated
overnight at 4°C in 50 mM MOPS buffer (pH 7.0) containing 10 mM EDTA. They were
then dialyzed twice against 50 mM MOPS buffer (pH 7.0) containing 2 mM EDTA, and
they were finally dialyzed twice against 50 mM MOPS buffer (pH 7.0) without EDTA.
The apo-enzymes were scanned against 50 mM MOPS @H 7.0). Enzymes containing
both Mg2+ and Co2+ were dialyzed against buffer A, then scanned against the dialysis

buffer as control.

142

RESULTS

TNXI mutants’ melting temperatures (Tm) were determined by DSC in the
presence and absence of metals (Table 1). With the holo-enzymes (Figure 1), all scans
except TNXI 1F1 revealed two thermal transitions. The TNXI wild-type and TNXI
V185T went through thermal transitions at 101°C and then 110°C and 114.5°C,
respectively. TNXI 3A2 went through these transitions at significantly lower
temperatures (86.6°C and 101°C). Surprisingly, 1F1 showed only one thermal transition,
at a slightly lower temperature than those of the wild-type TNXI and TNXI V185T
second thermal transitions suggesting that 1F1 lacks the first thermal transition common
in the wild—type TNXI, TNXI Vl85T, and TNXI 3A2 mutant.

DSC of apo-enzymes (Figure 2) of all TNXIs revealed only one thermal transition
at 96.5°C, 96.9°C, 84.4°C, and 784°C for the wild-type, TNXI V185T, TNXI 3A2, and
TNXI 1F1, respectively. The result suggested that the additional mutation that occurs in
TNXI 1F1 mutant, Phe186Ser, is responsible for altered metal-binding property and also
strongly enhances the metal requirement of TNXI as seen in a very low Tm of TNXI 1F1
apo-enzyme compared to those of the wild-type TNXI and TNXI V185T relative to their
holo-enzymes. A Tm difference of 289°C between TNXI 1F1 apo- and holo- form was
observed. This high Tm difference is approximately 2-fold higher than those observed in

TNXI, TNXI V185T, and TNXI 3A2 (135°C, 176°C, and 16.6°C, respectively).

143

 

 

 

 

 

 

100 , q , .
90 . —TNX1 WT i
—-—TNXIV185T;
80 ------ TNXI3A2
,_. . '--—-TNXllFl
it 70 _
"a . \
E. 50 E [’1
Q 40 i: ’
I a
2 30 1 w
20 l \\
10 . ‘
0 I I l l l ll
70 80 90 100 110 120 130
Temperature (°C)

Figure 1: Thermal unfolding of the holo-forms of TNXI and its mutant derivatives in the

presence of 5 mM MgSO.; and 0.5 mM CoClz followed by DSC.

144

 

 

 

 

 

 

 

 

 

 

 

 

120
V ~ _ —##‘ i l
——TNXI WT g
100 —TNXIV185T
:4 4_TN?<I_3'§_3 ,
E 80‘ 1
° l
E
g 60
¥
U l
I 40 l -
2 J
20 '
O 1 r
60 70 80 110120
Temperature (°C)
—TNXI 1F1
X
'K
s 4 -
E
a
3 3
¥
‘5’ 2
E
1 _
0 I T l l l
60 70 8O 90 100 110 120 ;
Temperature (°C) j

Figure 2: Thermal unfolding of the apo-forms of TNXI and its mutant derivatives

followed by DSC.

145

 

Table l: Melting temperatures (Tm) of TNXI and its mutant derivatives in the presence

(holo-enzyme) and absence (apo-enzyme) of 5 mM MgSO4 and 0.5 mM CoClz as

 

 

 

determined by DSC.
Holo-enzyme Apo-enzyme
Enzyme
Tml (°C) Tm2 (°C) Tm (°C)

TNXI wild-type 101.2 1 10 96.5
TNXI V185T 101.3 114.5 96.9
TNXI 3A2 86.6 10] 84.4
TNXI 1F1 N/A 107.3 78.4

 

 

 

 

146

 

DISCUSSION

Unfolding of TNXI and its mutant derivatives were followed by DSC. With the
holo-enzymes, all scans except TNXI 1F1 revealed two thermal transitions, which is
normal for TNXI (Hess et. al., 1998). These transitions could correspond either to the
existence of two phases in TNXI inactivation (Hess and Kelly, 1999) or to the presence in
the solution of a heterogeneous population of enzymes containing Mg2+ and Co2+ as
ligand (s)(Hess et al., 1998). Surprisingly, TNXI 1F1 showed only one thermal transition,
at a slightly lower temperature than those of the wild-type TNXI and TNXI V185T
second thermal transitions. This suggests that TNXI 1F1 probably lacks the first thermal
transition common in the wild-type TNXI, TNXI V185T, and TNXI 3A2 mutant. DSC of
apo-enzymes of all TNXIs showed only one thermal transition.

The result suggested that the additional mutation that occurs in TNXI 1F1
mutant, Phe186Ser, is responsible for altered metal-binding property and also strongly
enhances the metal requirement of TNXI as seen in a very low Tm of TNXI 1F1 apo-
enzyme compared to those of the wild-type TNXI and TNXI V185T relative to their
holo-enzymes. A Tm difference of 289°C was observed between TNXI 1F1 apo- and
holo- form. This high Tm difference is approximately 2-fold higher than those observed in
TNXI, TNXI V185T, and TNXI 3A2 (135°C, 176°C, and 16.6°C, respectively). This
evidently demonstrated that TNXI lFl’s extreme thermal stability is strongly metal-
dependent, even more than those of TNXI, TNXI Vl85T, and TNXI 3A2. This is not
surprising since the mutation is located in the active center close enough form a hydrogen

bond to the mainchain carbonyl group of the Leu229 residue (Sriprapundh et al., in

147

preparation). The neighboring G1u23] residue’s sidechain co-ordinates (forms a hydrogen
bond with) a structural metal (metal I) of the enzyme. So, the mutation would strengthen
the local conformation that, in turn, affects the binding affinity of the structural metal and
ultimately stabilize the enzyme.

The significance of cation binding in XI stability has not yet been examined
closely (Vieille et al., 2001). How ever some information is available on this issue. Site—
directed mutagenesis has been used to partially fill a metal binding site with sidechain of
an amino acid. These mutations to both metal-binding sites, M1 and M2, resulted in
destabilized XIs. In Streptomyces rubiginosus XI, the mutation of Hi3220 affected metal
binding at M2, which is turn was responsible for destabilization (Cha et al., 1994). A
similar observation has been made for Escherichia cali XI: mutation of His271 (ligand to
M2) significantly destabilized the enzyme (Batt et al., 1990). Other point mutations
triggering conformational changes in active site residues have also been found to
destabilize XIs (Varsani et al., 1993). Metal ions probably act to lock the active site in a

stable conformation, which is lost as soon as the metal leaves the active site.

148

REFERENCES

. Batt, C. A., Jamieson, A. C., and Vandeyar, M. A. (1990) Identification of essential
histidine residues in the active site of Escherichia cali xylose (glucose) isomerase.
Proc. Natl. Acad. Sci. USA 87, 618-622.

. Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding. Anal.
Biochem. 72, 248-254.

. Cha, J., Cho, Y., Whitaker, R. D., Carrell, H. L., Glusker, J. P., Karplus, P. A., and
Batt, C. A. (1994) Perturbing the metal site in D-xylose isomerase: effect of
mutations of His-220 on enzyme stability. J. Biol. Chem. 269, 2687-2694.

. Hess, J. M., Tchemajenko, V., Vieille, C., Zeikus, J. G., and Kelly, R. M. (1998)
Thermotoga neapolitana homotetrameric xylose isomerase is expressed as a

catalytically active and thermostable dimer in Escherichia cali. Appl. Environ.
Microbial. 64, 2357-2360.

Sriprapundh, D., Vieille, C., and Zeikus, J. G. (2000) Molecular determinants of
xylose isomerase thermal stability and activity: analysis of thermozymes by site-
directed mutagenesis. Protein Eng. 13, 259-265.

. Varsani, L., Cui, T., Rangarajan, M., Hartley, B. S., Goldberg, J ., Collyer, C., and
Blow, D. M. (1993) Arthrabacter D-xylose isomerase: protein-engineered subunit
interfaces. Biochem. J. 291, 575-583.

Vieille, C., Hess, M., Kelly, R. M., and Zeikus, J. G. (1995) xylA cloning and
sequencing and biochemical characterization of xylose isomerase from Thermotoga
neapolitana. Appl. Environ. Microbial. 61, 1867-1875.

. Vieille, C., Sriprapundh, S., Kelly, R. M., and Zeikus, J. G. (2001) Xylose
isomerases from Thermotoga. Methods Enzymal. 330, 245-249.

Vieille, C., Epting, K. L., Kelly, R. M., and Zeikus, J. G. (2001) Bivalent cations and

amino acid composition contribute to the thermostability of Bacillus licheniformis
xylose isomerase. Eur. J. Biochem. 268, 6291-6301.

149

 

CHAPTER VII

CONCLUSIONS AND DIRECTIONS FOR FUTURE RESEARCH

150

 

From these research studies, we were able to identify molecular determinants
responsible for the xylose isomerase enzyme from hyperthermophilic eubacterium
T hermataga neapolitana (TNXI)’s high thermostability at extremely high temperature by
comparative thermostability and site-directed mutagenesis studies with less thermostable
xylose isomerase fi'om thermophilic eubacterium Thermaanaerabacterium
thermasulfurigenes (TTXI). Despite their highly similar structures and amino acid
sequences (70.4 % identity), no obvious differences in the enzyme structures can explain
the differences in its stability compared to that of TTXI except for a few additional
prolines and fewer Asn+Gln in TNXI. TNXI has 2 additional prolines in the Phe59 loop
(Pr058 and Pro62) which helps forming another enzyme subunit’s active site. When these
2 prolines in TNXI were substituted with the corresponding amino acids present in its
less thermostable counterpart, TTXI, all the mutant enzymes showed significant loss in
thermostability compared to the wild-type TNXI. These data confirmed the hypothesis
that prolines indeed play important role in protein thermostability by reducing its entropy
of unfolding. Introduction of these two prolines in TTXI resulted in one mutant enzyme
(TTXI Q58P) with improved thermostability and the other (A62P), surprisingly with
significantly lower thermostability. Analysis of the three-dimensional (3D) structure of
TTXI and the predicted structure of its A62P mutant derivative indicated that a steric
hindrance between Pro62-C8 and Lys61-CB (2.92 A) is responsible for the reduced
thermostability of the mutant enzyme compared to the native TTXI.

We also successfully engineered TNXI’s active site by site-directed mutagenesis
to better fit glucose which is not its natural substrate. The V185T mutation in TNXI is the

most efficient mutant derivative with a 3.1-fold increase in its catalytic efficiency toward

151

 

glucose as a consequence of an improvement in glucose binding affinity (low Km) and a
faster catalytic rate of the reaction (higher km). This engineered “glucose” isomerase
from a native xylose isomerase has now comparable kinetic properties on glucose and
xylose.

To further optimize this enzyme’s potential industrial utility, directed evolution
(sequential random mutagenesis and low temperature/low pH activity screening) was
successfully applied to the TNXI V185T-encoding gene to obtain enzymes that have high
glucose isomerase activity at low temperature and low pH. The best mutant enzyme,
TNXI 1F1 (containing Vl85T, L282P, and F186S mutations), was dramatically more
active than TNXI V185T at all temperatures and pHs tested. Its high activities at low
temperatures were due to significantly lower activation energies (44 kJ/mole) compared
to that of TNXI and Vl85T (87 kJ/mole). TNXI 1F1 is also more kinetically stable than
TNXI and TNXI V185T. TNXI lFl’s enhanced stability is thought to be a result of
additional H-bond formation between Ser186’s sidechain and the neighboring L229
residue’s mainchain structure. This, in turn, strengthens local conformation and the
affinity of E231 co-ordination with the structural metal, hence improving the
thermostability of the mutant enzyme. Since the enzyme’s low temperature activity can
be significantly enhanced without loss in its thermostability, we suggest that the
molecular determinants of this enzyme’s activity and thermal stability are in fact, not the
same.

TNXI 1F1 was also subjected to a comparative study of biochemical properties
and fructose productivities with a commercially available glucose isomerase, GensweetTM

to further ascertain its properties to be used for industrial applications. TNXI 1F1

152

displayed higher catalytic efficiencies on glucose at low or high temperature and pH

ranges and had greater thermal stability than GensweetTM despite having similar

temperature optima for activity. This greater thermal stability together with the superior

kinetic parameters on glucose render TNXI 1F1 a genuine candidate for the industrial

glucose isomerization process based on the lifetime fructose productivity estimation.

1)

2)

3)

Directions of future research should include;

Directed evolution of TNXI 1F 1 to obtain a mutant enzyme that has improved
thermostability at extremely high temperature at lower pH. This would facilitate
using such enzyme in HFCS production at very high glucose isomerization
temperature (90-95°C) to obtain directly a 55 % fructose syrup without an
additional fructose enrichment step currently employed.

TNXI 1F1 immobilization studies. Since Xylose isomerase are used in the
industry as an immobilized form, studies of its properties, stability, and utility are
encouraged.

Detailed study of role(s) of metals in thermostability of TNXI 1F1. Unlike TNXI
and TNXI V185T, TNXI lFl holo-enzyme (containing Mg2+ and Co”) showed
only one thermal transition followed by DSC. It would be interesting to address

why it behaves such way and what is responsible for such event.

153

 

11 111111111111 1

3 1293 02467 4933