CLIN. CHEM. 40/5, 688-704
(1994)
Enzymes in Diagnostics: Achievements and Possibilitiesof
Recombinant DNA Technology
Erhard
Kopetzki,’
Klaus Lehnert,
and Peter Buckel
from an industrial point of view, the scope and
possibilities of recombinant DNA technology for “diagnostic
enzyme” production and application. We describe the construction of enzyme-overproducing
strains and show how to
simplify downstream processing, increase product quality
We discuss,
and process profitability, improve diagnostic enzyme properties, and adjust enzymes to harsh assay conditions. We also
consider some safety and environmental aspects of enzyme
production. Other aspects of diagnostic enzymes that we
cover are the facilitation of enzyme purification by attachment
of short amino acid tails, the introduction of tails or tags for
site-specific conjugation or oriented immobilization, the construction of bi- or multifunctional enzymes, and the production
of enzyme-based diagnostic tests as demonstrated by the
homogeneous immunoassay system of CEDL#{174}
tests. We
use as examples of diagnostic enzymes glucose-6-phosphate dehydrogenase (EC 1.1.1.49), glucose oxidase (EC
1.1.3.4), alkaline phosphatase (EC 3.1.3.1), o-glucosidase
(EC 3.2.1.20), pyruvate oxidase (EC 1.2.3.3), creatinase (EC
3.5.3.3), and -galactosidase
(EC 3.2.1.23).
Indexing Terms: recombinant
enzymes/enzyme
zyme immunoassay/enzyme
amplification/enzyme
purification/en-
immobilization
The term “diagnostic enzymes” refers to enzymes used
directly or as a component of an assay system for the
determination
of many different
substances.
These substances range from simple chemicals such as phosphates, ammonia, ethanol, and acetic acid; and metabolites such as glucose, urea, creatinine,
and ATP; to
toxic substances such as pesticides, herbicides, and
heavy metals; proteins such as viral antigens, antibodies, and serum proteins, and DNA drugs; vitamins;
and
hormones. The determination
of enzyme activity is an
additional
use. According
to the target-specific
biomolecule that we use, we differentiate between DNA-, immuno-, and enzyme diagnostics.
Enzymatic
AnalysIs
Diagnostic
enzymes are important
for analysis
of
of blood, serum, and urine for monitoring
and
diagnosing diseases (1, 2). In most cases the enzymatic
reactions are followed by measuring spectrophotometrically the oxidation/reduction
of nicotinamide
adenine
samples
Boehringer
Mannheim
GmbH, Department
of
Nonnenwald 2, D-82377 Penzberg, Germany.
1 Corresponding
author. Fax +49 08856 602659.
Received June 28, 1993; accepted
688
February
CLINICAL CHEMISTRY, Vol. 40, No. 5,
Genetics,
11, 1994.
1994
Am
coenzymes [NAD(P)/NAD(P)H
+ H]
or
by using colorimetric reactions as the end point. Ammonia, for instance, can be determined
with a single enzyme system by using the glutamate
dehydrogenase
reaction (Table 1, reaction
1). A coupled enzymatic reaction of hexokinase
and glucose-6-phosphate
dehydrogenase (G6P-DH) can be used to measure glucose (Table
1, reaction 2).2 A system that involves more than two
enzymes is exemplified for the determination of creatinine (Table 1, reaction
3). Coupled enzymatic
indicator
systems
are being used increasingly
for the determination of blood a-amylase and creatine kinase to diagnose
pancreatic
and myocardial diseases,
respectively.
dinucleotide
Enzyme Immunoassays
Enzyme immunoassays
(EIAs), which have found increasing use since 1970, are based on both immunological and enzymatic
reactions [see (3)]. Depending
on the
test used, EIAs are classified as homogeneous
or heterogeneous.
Heterogeneous assays are characterized by the use of
a solid-phase support for immobilization
of antigens or
antibodies
and at least one step to separate reactants
and reaction
products.
The principle of a direct sandwich ELISA and competitive
ELISA is exemplified
in
the quantification
of blood antibodies and antigens
induced, for example, by viral infections such as AIDS and
hepatitis
B (Fig. 1, heterogeneous
EIA). In heterogeneous assays enzymes function as easily detectable
labels that are usually linked chemically to antibodies,
antigens,
or haptens. Enzymatic catalysis of the label
results in an amplificationeffect.
Currently,peroxidase,
-galactosidase,
and alkaline phosphatase
are the most
commonly used enzyme labels.
In homogeneous
assays, the analyte-specific
signal is
generated
competitively
by the analyte-modulated
rate
of enzyme catalysis of an enzyme label (kinetic assay).
In principle,
the enzyme-label
activity is inhibited
(a)
directly by “site-specific” attachment
of a hapten-anti2NoJ
abbreviations:
EIA, enzyme immunoassay;
PCR,
chain reaction; NASBA, nucleic acid sequence-based
amplification; t-PA, tissue-type plasminogen activator; IBs, inclusion bodies; G6P-DH, glucose-6-phosphate dehydrogenase; kbp,
kilobase pair(s); GOD, glucose oxidase; ngd, N-glycosylation defective Saccharomyce8 cerevisio.e mutant; SDS-PAGE, sodium dodecyl
polymerase
sulfate-polyacrylamide gel electrophoresis; PLAP, placental alkaline phosphatase; GPI, glycosyl-phosphatidylinositol;
PyOD, pyruvate oxidase; 3D, three-dimensional;
G1cDH, glucose dehydrogenase; ED, enzyme donor; and BA, enzyme acceptor.
homogeneous
Emit”
technique
CA) is presented
schematically
EIA).
Table 1. Enzymatic analysIs.
1. Ammonia
2-Oxoglutarate
+ NH4
NAD(P)H
+
hi8ss
(Syva Co., Palo Alto,
in Fig. 1 (homogeneous
dehydrogenase
L-glutamate + NAD(P)
2.
+
H20
Glucose
hexotdnase
Glucose + ATP
glucose 6-phosphate+ ATP
Glucose6-phosphate+ NAD(P)
,oste
deliydrogenase
6-phosphogluconate+ NAD(P)H + H
3. Creatinine
Creatinine
creatine
+
+
Sarcosine
H20
H20
+
omalininase
‘creatine
omae
urea
+
sarcosine
H20 aarcoelnooxidase glycine+ formaldehyde
+ H202
H202 + 4-AAP + TBHB PeloxJdasev
benzoquinoneiminedye + HBr
+ 2
H20
4-MP, 4-aminoantipynne; TBHB,2,4,6-tribromo-3-hydroxybenzoic
acid.
1. Heterogeneous
Immobilized Enzymes
Immobilized
enzymes have mainly found application
in enzyme-electrode
and dry reagent products. These
devices are most commonly used for glucose monitoring,
with glucose oxidase being used as the enzyme.
Enzyme electrodes (4) consist of an immobilized
enzyme layer combined
with an electrode
(transducer).
The electrochemically
active compound generated
by
the immobilized enzyme layer in the presence of a substrate analyte is measured potentiometrically
or amperometrically
(5).
Disposable dry reagent devices such as test strips and
dipsticks provide all the reagents needed, including sensitive biomolecules
such as antigens,
antibodies,
and
enzymes, in a dry, reconstitutable
“ready-to-use”
form.
Since its invention
by Free and Free (6), the use of
immobilized enzymes has been described for nearly all
commonly measured analytes in clinical chemistry.
EtA
Enzyme-Based
Direct aandwich essay to, antibody
Competave assay Ion antibody
-)
-(
‘-<
-</
/
,,__i:
>
.<u
j;#{174}
/
-#{174}I/
-(
/
v)
.
Direct sandwich assay to, antigen
2.
-)
-C
‘V
Competitive assay ton antigen (v)
Homogeneous EtA
.(v
Inactive enn
active enzyme
‘V
EMIT (Enzyme.Monfloned Immuno Test)
Fig. 1. Enzyme immunoassays.
body complex to the enzyme label or (b) indirectly by
hapten-antibody-modifled
substrates, inhibitors, or cofactors. Free sample analyte competes for the bound
antibody
and releases
the inhibited
enzyme reaction,
thus providing
a signal (e.g., #{163}4lmin)
that is proportional to the sample artalyte
concentration.
The key
features
of these assay types are speed, easy automation, and the lack of separation and washing steps. The
Signal and Target Amplification
Enzymes are used for signal and target amplification
to improve the sensitivity
and speed of immunoassays
or
to enable the multiplication
of nucleic acid fragments.
Signal amplification
by a second enzyme depends on a
coupled substrate or cofactor cycle that is catalytically
activated by an enzyme label (7, 8). The principle of
enzyme amplification
is illustrated
in Fig. 2, in which
alkaline phosphatase
is used as the primary
enzyme
label for the AMPAK
system (Novo BioLabs, Cambridge, UK). In this system, NADP,
an alkaline phosphatase substrate, is dephosphorylated
to NAD,
which
activates
a coupled redox cycle involving
an alcohol
dehydrogenase
and diaphorase reaction for the continuous generation of a purple formazan product.
In contrast to signal amplification,
target amplification involves multiplication
of the target molecules.
Currently,
target amplification
is possible only for nucleic acids. The PCR (polymerase
chain reaction)
and
NASBA”
(nucleic acid sequence-based
amplification;
Cangene Corp., Mississauga,
Canada) systems represent such in vitro amplification
processes. In the case of
PCR (9, 10) a specific double-stranded
DNA sequence is
exponentially
multiplied
by enzymatic
synthesis after a
series of cycles involving template denaturation,
primer
annealing,
and the extension of hybridized
primers with
use of a thermostable
DNA polymerase
(Fig. 2, PCR). In
addition to thermal cycling methods, homogeneous,
isothermal transcription-based
amplification
systems have
been developed. For instance,
according to the NASBA
system (Fig. 2), single-strandedRNA is transcribed
reversely by using reverse
transcriptase
and a targetspecific oligonucleotide
primer that contains
the T7
RNA polymerase
promoter sequence as a “5’-overhang”
(11). The RNA strand of the RNA/DNA
hybrid is degraded with ribonuclease H (RNase H) before the second
DNA strand is synthesized by reverse transcriptase
and
CLINICAL CHEMISTRY, Vol. 40, No. 5,
1994
689
AMPAK 8ystem
1. Signal amplification:
Ethanol
NADP
pi
Formazan
2.
INT-violet
Target amplification:
PCR (Polymerase
Chain Reaction)
5.... #{149}:iiiiiiiiii
III 1111111111111111111
NASBA (NucleicAcidSequence Based Amplification)
DNAtarget
3
5,
denaturation
cycle 1
Reverse
Transkflptase
primer annealing
#{149}flT
4
rrn*
p
II
primer annealing
[1
JJ,
primerelongation
RNA
DNA
IIIIIIIIIl1IIIIlllIIlIIlIt
primer elongation
,Jj,
RNaseH
U
RNAtarget
RNA
p
tIIIIIlItI,I,IItItIItlIIIlII
DNA
U
primer annealing
T
cycle 2
TTTP
,I&LU
-
DNA
II
Tr
primer elongation
DNA
DNA
17RNA
Polymerase
5#{149} amplthed
RNA ANA
3’4
(
U
(anbsens
II
DNA
IIlIIIlIIlIIIIIIIIIIII
cycles 3-30
RNA
RNaseH
-
U
primer annealmg
Reverse
________
DNA
Transkr#{231}tase
IllIllIltIl
thermal
cyclIng amplIfication,
about 107-fold
Isothermal
amplIficatIon,
DNA ___________
DNA
about 10’-fold
Fig. 2. Signal and target amplification.
a second target-specific
primer.
The generated
DNA
molecule is then transcribed
by T7 RNA polymerase
and, with this RNA synthesis,
the system enters the
cyclic phase.
The number of commercially
available enzymes is
690
CLINICAL
CHEMISTRY,
Vol. 40, No. 5, 1994
somewhat limited, and each enzyme may or may not be
ideal for a special technique or test system. Thus there
is a need to screen for new enzymes or to improve the
properties
of known enzymes either chemically
by dorivatization
(12) or genetically
by genetic engineering.
Only a few published
reports describe the production
and improvement
of diagnostic
enzymes by recombinant
DNA technology
[e.g., y-glutamyltransferase
(13)]; in
contrast,
many papers describe the modification
of enzymes used in industrial processes, such as glucose
isomerase
(14), penicillin G acylase (15), chymosin, and
various
proteases
and lipases (16), or enzymes of therapeutic potential,
such as prourokinase,
tissue-type
plasminogen
activator (t-PA), and superoxide dismutase
(17). Here, we review mainly our own work on enzymes
of potential use in diagnostics.
Gene Cloning for Diagnostic
Enzyme
Production
Diagnostic
enzymes are usually needed in relatively
small quantities
(i.e., <10 kg/year worldwide);
very
few (e.g., glucose oxidase; GOD) are produced in quantities >100 kg/year. Industrial
enzymes,
which are
used, for instance, in food processing
(e.g., a-amylase,
glucoamylase,
glucose isomerase,
and chymosin),
and
enzymes
used in the detergent
industry as additives to
washing
powders (e.g., proteases
and lipases), are produced in quantities of several 100 tons of enzyme protein per year worldwide
(18). The amount of enzyme
needed depends on the enzyme’s physical properties,
such as specific activity and stability, and2n the number of tests performed.
In contrast to
strial enzymes, which are rarely chemically
pu1. or singleprotein preparations,
diagnostic
enzymes are usually of
higher purity and free of impurities
that would interfere with a diagnostic
assay.
About 60 to 70 enzymes are manufactured
for diagnostic use (19). In accordance with the variety
of catalytic activities
needed, the biological sources of these
enzymes are numerous
and include plants, animal tissues, and microorganisms.
Examples
are horseradish
for the isolation of peroxidase, calf intestine for iIkaline
phosphatase,
and Aspergillus
niger for GOD. Enzymes
are typically found in the cell at concentrations
of <1%
of the total soluble cell protein. Extensive purification
is
therefore necessary to (a) produce an electrophoretically
pure enzyme (>100-fold)
and (b) reduce the contaminating enzyme impurities
present in crude cell lysates to
acceptable
levels [-300-40
000-fold, (20)]. In addition,
because the amount of enzymes
in plant and animal
sources may vary according
to season, the production of
a diagnostic
enzyme of reproducibly constant quality is
form or (b) as insoluble and inactive protein aggregates,
often referred to as an indusion body (IBs). The latter
requires that the lBs be easily renatured in vitro to an
active enzyme. Enzymes with a high level of expression
require only a two- to threefold purification
to become
electrophoretically
pure. Furthermore,
by combining
a
high level of expression with high cell density fermentation of microorganisms,
one can obtain several grams of
enzyme per liter of culture medium, thus demonstrating
that high-level expression and high volumetric yield are
the basis for efficient and economical enzyme production.
Several factors that support the use of recombinant-derived enzymes are listed in Table 2.
To ifiustrate
some of these points, we compare the
production of yeast a-glucosidase
(maltase) from commercial baker’s yeast paste with a yeast strain overexpressing recombinant
a-glucosidase
(21). The recombinant yeast strain produces -25 times more a-glucosidase
than is present in commercial
baker’s yeast paste. As a
result, the steps needed for classical purification of yeast
a-glucosidase
can be reduced from -10 to 4. An overview
of the economic and environmental
aspects of using a
recombinant yeast strain (see Table 3) shows the advantages of replacing naturally derived enzymes with recombinants as source materials.
-
difficult.
Recent advances in recombinant
DNA technology have
made it possible to clone numerous genes of interest and
to manipulate
bacterial, yeast, insect, and mammalian
cells to overproduce the desired enzyme. For economic
reasons, microorganisms
such as Escherrchia
coli and
yeasts are the organisms of choice, since they can be
grown easily and rapidly to high cell density on cheap
nutrients. During the last decade, many enzymes have
been expressed in microorganisms
at levels 10 to 100
times higher than those in the natural host cell. In some
cases the levels of expression have been up to 50% of the
total cellular protein. Such hyperexpressed
enzymes may
be synthesized
(a) in a soluble and enzymatically
active
Gene Cloning Techniques
In principle, it is possible to isolate any gene of interest. A variety of recombinant techniques exists for gene
cloning. In general,
one starts by isolating
DNA or
mRNA from the natural-source
host cell of the enzyme in
order to synthesize a genomic or cDNA library. A cDNA
library is usually required if the gene of interest
comes
from a eukaryotic cell whose genes are interrupted
by
nontranslated
intervening sequences (introns) (22). The
Table 2. Reasonsfor utilizingrecomounant-derived
enzymes.
Enzyme source
1. The natural enzyme source material is limited or expensive.
2. The natural enzyme source material is a pathogenIc organism
or may be contaminated with a pathogen (e.g., a virus).
3. The natural enzyme source material contains only low enzyme
concentrations or contains enzyme impurities that are difficultto
remove and require extensive purification.
4. The natural enzyme source undergoes seasonal variations.
Enzyme quality
1. A high enzyme concentration in the source simplifies enzyme
purification in general and in many instances improves the
enzyme purity and hence the quality.
Economic and environmental aspects
1. The enzyme can be fermented routinely in unlimited amounts.
2. Standard systems (host-vector systems) and procedures can
be applied to enzyme expression, fermentation, and
purification.
3. The costs of enzyme manufacture can be decreased.
4. A high enzyme content in the source leads to a tremendous
reduction in the amount of biomass to be processed and hence
of the chemicals and other materials required for fermentation
and isolation. Furthermore the waste material produced is
greatly reduced.
CLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
691
DNA fragment
Table 3. Comparison
of yeast a.giucosldase
production from commercial baker’s yeast vs a
recombinant yeast strain.
Recombinant
Baker’s yeast
Purification steps
Produced
Choosing the Right Strategy and Host-Vector System
The choice of the host-vector
system plays a key role
in the successful expression and production of a recom-
waste material
Yeast cell debris, tons
1100
25
12
25
0.5
90
133
0
5
3700
45000
50
9000
44#{176}
Ammonium sulfate, tons
Potassium phosphate, tons
Alumina adsorber, tons
Filtrationaid, tons
Energy
Water, m3
Electric energy, kWh
Reduction in raw matedal, energy, and waste:
gene
yeast
10
236
0.4
Biomass, tons
Specific activity, kU/g protein
encoding
the enzyme
can
>90%
then
be identified
by
screening the library with antibody or DNA probes, complementation
of mutants, or by activity-screening
using
one of the strategies listed in Table 4.
In the past,
we and others have isolated many genes
enzymes
with the aid of degenerated oligonucleotides
based on partially
determined
enzyme amino acid sequences. Amino acid sequences were
obtained by Edman degradation from the NH2 terminus
of the purified enzyme or, if necessary,
from separated
peptide fragments
created
by specific enzymatic
or
chemical fragmentation.
More recently, the PCR technique
has had a great
impact on gene cloning procedures
in general. This includes the generation of specific DNA probes for screening; the addition of new sequences,
e.g., single restriction sites, to the ends of DNA fragments
for subcloning,
the introduction
of DNA alterations
(e.g., insertions,
deletions,
and substitutions);
and the recombining
of
DNA sequences at any desired junction. If the DNA and
protein sequences of an enzyme are known, it is relatively easy to amplify and clone directly an engineered
DNA sequence by PCR. This means, e.g., the addition of
a singular restriction
site at both ends of the desired
encoding
diagnostic
Table 4. Common gene isolation strategies.
1. Complementation of homologous or corresponding
heterologous mutants (e.g., E. coil or yeasts and other fungi)
a growth-deficient phenotype.
Cross-hybridization (colony hybridization) with a DNA probe
consisting of an already cloned gene with homology to the
gene of interest.
Screening the library with synthetic degenerated oligonucleotide
probes designed on the basis of known protein sequences.
Detection of the desired gene expressed as a fusion protein
(e.g., with -gaIactosidase, phage Agtl 1 system) by antibody
recognition.
DetectIon of the desired gene by activity screening, assuming
that the native promoter is active in the host cell employed.
Various PCR techniques based on known protein or DNA
with
2.
3.
4.
5.
6.
sequences.
692
CLINICAL
CHEMISTRY, Vol. 40, No. 5, 1994
to simplify the insertion
in an appropriate expression cassette,
to replace heterologous
by homologous
signal sequences,
or to delete membranespanning segments (23).
binant
enzyme. To choose an appropriate
expression
system, one should consider the hierarchical order of the
source organism for the natural enzyme, the natural
localization
of the enzyme (within or outside the cell)
and, in addition, the structural
organization
and enzymatic and physical properties of the enzyme molecule to
be overexpressed.
For a checklist,
which is not comprehensive, see Table 5.
The success in using a particular host-vector
system
for heterologous
gene expression
is not predictable.
However, as a general rule, for the expression of a biologically active enzyme one should choose a host cell
that is relatedevolutionarilyas closely as possible to the
natural host and that is economical to use. At present,
only microbial
host-vector
systems fulfill the criterion
of economic
reduction of enzymes in bulk quantities.
Insect- an
ammalian-based
host-vector
systems
may
prove usefut only for enzymes of high value when production fails in microbial cells and the natural source
material is limited in supply or not acceptable for safety
reasons.
With regard to the possible toxicity of a foreign enzyme to a host cell, it is advisable
to use regulated
rather than constitutive
expression of the gene. There-
Table 5. ConsIderations in choosing the right
expression strategy and appropriate host-vector
system.
Natural source organism
1. Higher eukaryote or microbial eukaryote (fungi, etc.)
2. Prokaryote
Natural localization of the enzyme within or outside the cell
1. Cytoplasm
2. Cell organelle
3. Membrane-associated
4. Periplasm or outside the cell
Enzymatic and physical properties
1. Enzyme activity, e.g., toxicity due to interference with the host
cell metabolism
2. Stability, e.g., protease sensitivity
Structuralorganization of the enzyme molecule
1. Molecular mass
2. Number of subunits (homomer, heteromer)
3. Disulflde bonds (almost solely in extracellular proteins)
4. Posttranslational modification
Proteolytic protein maturation (preproprotein)
Glycosylation
Acetylation, sutfation, phosphorylation, myristylation,
carboxylation, hydroxylation, etc.
5. Cofactor(s)
fore, strong and strictly regulated promoters that can be
induced for maximal protein synthesis
after growth of
the culture to the appropriate cell density are preferred.
In contrast to several proteins of pharmacological
interest, such as erythropoietin,
for which proper biological activity depends on a specific carbohydrate
structure, most posttranslationally
modified proteins do not
require
their naturally
occurring
posttranslational
modifications
for enzymatic activity. Hence, it may be
possible to choose a host cell that cannot perform these
modifications
even if the natural enzyme is glycosylated.
For natural, cytoplasmically
localized enzymes it is
appropriate to start expression studies with the easiest
and most developed bacterial
gene expression system, the
E. coli system, for which a variety of expression vectors,
host strains and, in addition, technology for high-density
fermentation
are available. E. coli has been used successfully for the overexpression
of a broad range of cytoplasmic enzymes of prokaryotic origin, including diagnostic
enzymes such as G6P-DH from Leuconostoc
dextranicus
(24), creatinase from Pseudomonasputida
(25), galactose
dehydrogenase
from P. fluorescens (26), and Taq polymerase from Thermus
aquaticus
(27). However, naturally secreted enzymes, particularly
those of eukaryotic
origin, give rise to the formation of large insoluble and
inactive protein aggregates [for a review, see (28)]. The
formation of lBs is not a phenomenon
observed solely
with eukaryotic proteins expressed
in E. coli. Even homologous E. coli proteins such as /3-galactosidase (29) or
homologous Saccharomyces
cerevisiae
enzymes such as
catalase T and alcohol dehydrogenase
I (E. Kopetzki, in
preparation) give rise to lB formation when expressed at
high levels. Protein aggregation
is assumed to arise when
(a) the rate of protein synthesis exceeds the rate of proper
protein folding (30); (b) the cellular environment,
for
instance, the redox potential, is not favorable for disulfide
bond formation; or (c) helper proteins, “chaperones” (31),
are necessary for proper folding of a biologically
active
enzyme but are not present in the heterologous host cell
(32).
For the production
of a naturally secreted enzyme,
a
host cell with a naturally
high capabifity for secretion
may be superior to E. coli, especially
if the enzyme is of
eukaryotic origin. Currently, the most promising host for
heterologous
secretion of enzymes in bulk quantities are
fungi-based expression-secretion
systems, including various yeast species (e.g., Saccharomyces,
Kluyveromyces,
Hansenula,
Pichia, Schizosacchammyces,
and Schwanniomyces), and Aspergillus
and Trichoderma.
However,
there is no guarantee of successful expression-secretion
of an active heterologous
enzyme
per se, even if several
different gene expression
systems are tested.
In most
cases of unsuccessful heterologousgene expression the
recombinant
enzyme
is neverthelessexpressed
at high
levels within the cell, but is insoluble and biologically
inactive. In this situation, in vitro renaturation
of the
mature
protein (encoded by an engineered
structural
gene that does not code for prepro sequences such as
signal sequences) may be an economic alternative
in the
future. Recently,
this approach was successfully
applied
of several naturally secreted complex
eukaryotic
proteins
of pharmacological
interest,
e.g.,
t-PA and t-PA variants, which have up to 17 disulfide
bonds, and prourokinase
(32), interferons,
interleukins,
and colony stimulating
factors.
We use several of the strategies discussed above, including gene cloning, gene overexpression,
and, if indicated, renaturation
of protein aggregates,
as examples
of strategies
that can be used for the production of diagnostic enzymes.
to the production
Examples
Glucose-6-Phosphate
Dehydrogenase
G6P-DH is used as an indicator enzyme at the end of
coupled enzymatic
reactions of spectrophotometrically
recorded assays (Table 1, reaction 2) and as an enzyme
label in homogeneous
EIAs (33). The enzyme from Leuconostoc catalyzes the oxidation of glucose 6-phosphate
with NAD
or NADP
as cofactor, in contrast to yeast
G6P-DH, which is specific for NADP.
G6P-DH from
Leuconostoc
is a cytoplasmic enzyme of prokaryotic origin consisting of two identical subunits with an apparent molecular mass of -55 kDa per subunit, and it does
not undergo any posttranslational
modifications.
On the
basis of this information, E. coli was used as the host for
the expression of G6P-DH.
A G6P-DH gene of L. dextranicus
was isolated from a
genomic gene pool by using as a probe a mixture of
degenerated
synthetic oligonucleotides
derived (a) from
an active-site
region of human erythrocyte
G6P-DH
showing high homology to sequenced peptides of L. mesenteroides (34) and (b) from the NH2-terminal
amino
acid sequence of L. dextranicus
G6P-DH, which was
determined
by Edman
degradation
(35). The gene library was constructed
in plasmid pUC18 by using L.
dextranicus
DNA that had been partially digested with
Bell and size-fractionated
(3-4 kbp). Because of the degeneracy of the genetic code, it is not possible to deduce
a unique nucleotide
sequence
for the hybridization
probe. We therefore
designed
“partially
degenerated
probes” (36) from amino acid sequences with the least
ambiguity in the corresponding
codons and taking into
consideration
the codon usage of the Lactobacfflaceae
(Fig. 3). The G6P-DH gene was then isolated by colony
hybridization
(37), characterized
by restriction enzyme
analysis and DNA sequence determination,
and then
expressed in E. coli by using the natural L. dextranicus
promoter and terminator.
Surprisingly,
the G6P-DH gene was expressed constitutively in E. coli, with the enzyme constituting
up to
50% of the total protein and being enzymatically
active
and completely soluble. Overexpression
of G6P-DH was
expected to lead to serious interference with the host cell
metabolism,
because glucose 6-phosphate,
the G6P-DH
substrate, is a key metabolite
for various biosynthetic
pathways. Compared with L. mesenteroides,
the natural
source organism,
the recombinant
E. coli strain
produces
100-fold more G6P-DH with respect to the specific enzymatic activity. This high level of expression of
-
CLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
693
I
R
active site region
ftQ
I
Gin Glu His Phe
IC&AGM CAT m
amino acid seguence
-
I
Phe Glu Asn Ala Phe Asp Asp Asn Gin Leu Phe Arg lie Asp His Tyr Leu Gly Lys Glu Met Vat Gin Asn
iTt GAGMT GCCm OATOATMT CAGITO TIT AGGATAGATCATTAT UG 006 AAGGAGATGGTGCAGMT
C A C T C C C C ACA
CCA
I
C C CCA
A A A
A A C
dene
GGCC
CM OAkCAT m 3
CAC
mosttrequmaty
usedcodonsI,
Lactobodeae
CM GM CAT iTt ‘3’
degenerated
pcoOe
active site region
C.fl5&
flr,k4l
A
T
T
G
C
C
5’. ITT OAAMC OCT iTt OATGATMC CM TTA
MTGCC
MT
ITO
CiT
m
C
T
T
T
C
C
C
CGT AlT OATCATTAT TTA GGTAAAGM ATGOTt CM MC ‘3
AGA
TTGGGA
MT
CGA
CIT
CGC
5’ iTt GM MC OCT iTt GATOATMC CM TTA iTt COT AU GATCATTAT TTA GGTAM GM ATOOTTCAAMC
T
A A
H2N
Fig. 3. Design of synthetic oiigonucleotide probes used for isolation of the
G6P-DH made it possible to establish
a very efficient
and simple purification
process consisting
of only enzyme precipitation
(crystallization)
from dialyzed cell
lysates (24).
Glucose Oxidase
The enzyme GOD from A. niger has found wide application (a) in the preservation of food (e.g., in the removal
of oxygen from beverages,
as an antioxidant,
and in the
removal of residual glucose from dried egg products), (b)
as a source of hydrogen peroxide, and (c) as a means for
detecting
and quantifying
glucose in industrial
solutions and in body fluids such as blood and urine (38).
GOD catalyzes the oxidation of -D-g1ucose
to D-glucono-1,5-lactone
and the concomitant
reduction of oxygen to hydrogen peroxide.
The enzyme from A. niger is naturally secreted and is
apparently
targeted to peroxisomes (39) but, under certain conditions,
is also secreted into the culture medium
(40). The fungal GOD is glycosylated, consists of two
identical
subunits
(150 kDa each), and contains one
tightly but not covalently bound flavin adenine dinucleotide (FAD) cofactor and one disulfide bridge per subunit. In the last few years several groups have isolated
the A niger GOD gene and reported on expression and
production studies of GOD in transformed A niger, A.
nidulans, and Saccharomyces
cerevisiae
(41-43). The
GOD gene was cloned from A niger genomic and cDNA
libraries
in a way similar to that described above for
G6P-DH.
Homologous
and heterologous
expression and secretion of GOD were studied in Aspergillus
and Saccharomyces. A plasmid containing
the GOD gene, including
the natural
GOD promoter with its regulatory
properties, was reintroduced
intoA niger to increase the GOD
gene dosage by genomic
integration.
The best transformants produced about six times as much GOD as the
parent strain (43). However, the natural host organism
did not prove to be ideal for GOD overexpression.
First,
694
CLINICAL CHEMISTRY, Vol. 40, No. 5,
1994
L. dextranicus G6P-DH gene.
GOD expression and secretion depends on the presence
of glucose and a pH of the culture medium
>4.5. GOD
production leads to gluconic acid and hydrogen peroxide
formation and hence complicates
cell culture. These
problems can probably be circumvented
by substituting
the natural GOD promoter with another potent and
glucose-independent
Aspergillus
promoter.
Second,
GOD is usuallycell associated,
thus making purification
more difficult.
Third, purified Aspergillus
GOD is often
contaminated
with host cell enzymatic
impurities
such
as catalase, amylase, and cellulase, which interfere with
its application.
To overcome these disadvantages,
GOD expression
and secretion were studied in S. cereuisiae. The mature
GOD coding sequence was fused in frame with the yeast
a-factorprepro sequence or the natural GOD signal
sequence and then inserted into episomal replicating
2-gm DNA-based multicopy vectors under the control of
a potent and tightly regulated
yeast promoter (42).
Derepression
of the hybrid promoter leads to the secretion of large amounts of active GOD into the culture
medium (up to 3 g/L) in a fermenter (44). However,
compared with the GOD enzyme extracted from A. niger, biochemical
characterization
revealed
that the
yeast-derived
GOD was very heterogeneous
in size as a
result of extensive and varying N-linked glycosylation.
The different average carbohydrate
content of -70% vs
14% for the yeast- and A niger-derived
GOD, respectively, leads to a threefold lower specific activity by
weight for the yeast-derived
GOD. Hyperglycosylation
of the yeast GOD significantly
increased the thermal
and pH stability of the enzyme, and the increase in the
molecular
mass permitted
simple purification
based
only on cross-flow filtration of GOD that was free of
detectable catalase, amylase, and cellulase impurities.
The yeast-derived
enzyme proved useful in food processing, but was not accepted as a substitute
for the A
niger-derived
enzyme in established
diagnostic applications, mainly because of the reduced specific activity.
To take advantage of the high secretion capability of
yeast, we engineered
an appropriate yeast mutant strain
deficient in N-linked
outer chain glycosylation
(hyperglycosylation) to produce an A. niger GOD of diagnostic
quality. For reviews of the carbohydrate structure of natural yeast mannoproteins,
see (45-47). Briefly, asparagine-linked
core oligosaccharides
(Man9GlcNAc2Gluc3)
are added and trimmed
(Man8GlcNAc2)
in the endoplasmic reticulum. Then, proteins destined for secretion, such
as invertase and acid phosphatase,
are elongated by the
addition of an outer chain containing up to 150 mannose
residues during transit through the Golgi apparatus.
Glycosylation-defective
mutants were screened by using [3Hlmannose suicide selection and natural invertase
to identify mutants defective in N-glycosylation
(48). In
practice, the yeast cells were mutagenized
and then
incubated
with [3H]mannose.
Glycosylation-defective
mutants
with a decreased ability to incorporate radioactive mannose were enriched during storage at -80#{176}C
for several weeks. One of these N-glycosylation-defective (ngd) mutants, ngd29, was used to construct hyperglycosylation-deficient
yeast host strains suitable for
transformation
by
auxotrophic
complementation
(URA3 gene) and LEU2-d gene-based plasmid amplification (49) via mating, tetrad dissection,
and spore
analysis. GOD expression and secretionwere studied in
the yeast strain BMY3-9A (MATa ngd29 ura3-52 leu21,112 his3-i)
transformed
with plasmid YEpL-GOD
(48). This episomal replicating yeast vector contains the
GOD coding sequence,
including
the authentic
GOD
signal
sequence inserted into the glucose-repressible
and maltose-inducible
a-glucosidase
expression cassette
(21). Transformed
BMY3-9A cells express and secrete
active GOD as efficiently
as wild-type yeast cells. The
GOD was homogeneous
in size (-70 kDa) and was indistinguishable
by SDS-PAGE
from the A. niger-derived enzyme (Fig. 4).
Increased
thermostabiity
was even found for the
core-like
glycosylated
yeast-derived
GOD, as demonstrated by differential
scanning calorimetry,
which indicated increased melting points (Tm, Table 6). Furthermore, the yeast ngd29 mutant-derived
GOD enzyme
was almost as stable as chemically
cross-linked A. ni123
Fig.4. SDS-PAGEof purifiedA. niger (lane 1), yeastwild-type (lane
2), and GOD derived from yeast hyperglycosylation-defective
ngd29-mutant (lane 3).
Size markers are indicated at the left.
Table 6. Enzymatic and biochemical characterization of
differentially glycosylated GODs.
Specific activity (kU/g product)
Molecular mass, kDa
Carbohydrate content, %
DSC, TmValue, ‘C
Catalase impurities
Wild-type
Yeast
A. niger
yeast
ngd29 mutant
225
70
13
69
70-140
71
228
68
+
nd
70
12
75
-
DSC, differential scanning calorimetry; nd, not determined.
ger-derived enzyme (H. Schmuck, personal communication), and proved valuable for diagnostic applications.
These results show that GOD is an excellent example
for demonstrating
the value of (a) different host organisms in improving production process (enzyme yield)
and enzyme quality (free of interfering enzyme impurities) and (b) genetically
modified host strains (mutants)
in improving the enzymatic properties of an enzyme by
altered posttranslational
modification.
Human Placental Alkaline Phosphatase
(PLAP)
Besides E. coli (3-galactosidase and horseradish
peroxidase, calf intestine alkaline phosphatase
is most commonly used as an enzyme label in heterogeneous
immunoassays.
Alkaline
phosphatases
are metalloenzymes
that hydrolyze phosphate esters at basic pH and are present in nearly all living organisms. Compared with almost all known microbial
alkaline
phosphatases,
the alkaline phosphatases of animal and human tissues exhibit
a several-fold higher specific activity
and are generally
more stable (50). Since human PLAP was the first eukaryotic alkaline
phosphatase
to be cloned and sequenced, we used this enzyme for microbial expression
studies. The PLAP gene was isolated from a human placental cDNA expression library constructed in the bacteriophage vector Agtll with the aid of PLAP-specific antibodies induced against CNBr fragments
of the enzyme
as a probe (51). PLAP is a homodimeric,
membraneassociated glycoenzyme of the cell surface, which is also
often found in serum. The PLAP gene encodes a prepro
PLAP protein with two posttranslational
cleaved signal
sequences: (a) a signal sequence of 22 amino acids at the
NH2 terminus
and (b) a hydrophobic
recognition
sequence of 29 amino acids at the COOH terminus
required
for glycosyl phosphatidylinositol
modification
and anchoring [GPI anchor (52)] of PLAP to the cytoplasmic
membrane [see (53, 54) and references cited therein].
Taking into account the naturally
occurring posttranslational
changes (secretion, glycosylation,
and attachment
to the cytoplasmic
membrane via a GPI anchor), we chose S. cerevisiae
as a microbial host that, in
principle, has the ability to perform these types of modification. For the expression
and secretion studies we
replaced the GOD gene with several PLAP gene constructions in plasmid YEpL-GOD. The coding sequence
of the mature PLAP gene was fused with several secretory leader-encoding
sequences.
These included
the
a-factor prepro and invertase signal sequences of yeast
CLINICAL
CHEMISTRY, Vol. 40, No. 5,
1994
695
as well as the natural PLAP and the already successfully used A. niger GOD signal sequence.In addition, we
investigatedthe requirement
of the hydrophobic C-terminal signal sequence
(GPI anchor) for secretion
and
proper acquisition
of enzyme activity. However, all attempts to express active PLAP in yeast failed for unknown reasons, although
inactive and insoluble PLAP
protein of the expected size was detected by SDS-PAGE
and immunoblotting
in the SDS-extracted
pellet fraction of lysed yeast cells.
After obtaining thesenegative results we changed our
strategy.
We synthesized the mature PLAP protein as
inactive lBs in E. coli and tried to renature the inactive
PLAP protein to enzymatic activityin vitro. The gene
coding for the mature Met-PLAP (without the NH2- and
COOH-terminal
signal
sequences)
was engineered and
amplified
by PCR and ligated into an E. coli expression
vector.Overexpression of the mature PLAP gene was
achieved
only with the aid of very tightly regulated
promoters such as the T5-N25 hybrid promoter (55).
However, our attempts to restore enzymatic PLAP activity by renaturation
of isolated lBs according to published procedures [for a review, see (56)] failed, even after
systematic
variation
of the factors influencing
folding,
formation
of disulfide
bonds,
and protein association.
Nevertheless,
enzymatically
active PLAP could be recovered by a new modified renaturation
process (57). The
main operations of this renaturation
process are listed in
Table 7.Currently,
this strategyseems very promising,
not only for the production of eukaryotic
alkaline
phosphatases, but also for enzymes that cannot be expressed
in microbial host organisms in the active form.
Enzyme Extensions and Fused Enzymes
If the hurdles of gene cloning and gene expression of
an enzyme have been overcome, it is feasibleto alter the
nature of an enzyme molecule in a modest amount of
time by genetic engineering
techniques
to give it special
properties such as (a) a fused polypeptide
tail to aid
purification,
covalent modification,
or immobilization
and (b) fusion of a second structural gene (enzyme).
Tails and Tags for Purification
The concept of purification
fusions is based on (a) the
affinity of the polypeptide
tag for a biospecific ligand,
enabling
efficient purification
of the fusion protein by
affinity chromatography,
or (b) changed physical properties of the fused polypeptide that make efficient purification possible by covalent chromatography,
hydrophobic
Table 7. SchematIc presentation of the renaturatlon
process for mature PLAP.
1.
Isolation
of lBs.
2. Solubllization
of lBs in denaturants (6 mol/L guanidinium
chloride or 6-8 mol/L urea).
3. Reduction of disulfide
bondswith 1,4-dithioerythritol.
4. Removal of excess 1 ,4-dithioerythritolby dialysis.
5. Dilution in renaturation buffer.
6. Protein purificationof active PLAP.
696
CLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
interaction
chromatography,
ion-exchange
chromatography, or metal chelate affinity chromatography.
Fusion tails can be added at the DNA level either to
the NH2 or COOH terminus or to both termini of an
enzyme. An enzymatic or chemical cleavage site can be
included to allow the removal of the tail, if necessary, to
restore enzyme function. Various fusion tails have been
used for the efficient recovery of proteins, often by onestep purification from crude cell extracts. These fusion
tail-ligand
systems include the following
For use in affinity chromatography1) Whole enzymes that bind to immobilized substrates,
substrate analogs, or inhibitors,
e.g., 3-galactosidasep-aminophenyl-f3-D-thiogalactoside
and glutathione-Stran8ferase-glutathione.
2) Domains of proteins, e.g., with affinity to IgG, carbohydrates, or biotin, such as Protein A-IgG, cellulase
cellulose-binding-domain--cellulose,
and transcarboxylase-biotin-binding-domain-biotin.
3) Epitopes that are recognized
by immobilized
monoclonal antibodies.
For ion-exchange
chromatography4) A cluster of charged amino acids, e.g., polyCArg) or
poly(Asp).
For metal chelate affinity chromatography5) A cluster of histidine residues.
For hydrophobic interaction chromatography6) A cluster of hydrophobic amino acids, e.g., poly(Phe).
For covalent chromatography7) One or several cysteine residues.
The advantages
and disadvantages
of the different
tails and their potential in laboratory and industrialscale applications have been the topic of several reviews
(58-61).
We therefore focus on the choice and application of those tails that may be most useful for the economic production of diagnostic enzymes.
Generally,
the tails most attractive for the production
of diagnostic enzymes (a) do not interfere with expression; (b) bind with high affinity
and selectivity
to a
cheap adsorbent;
(c) do not interfere with the enzymatic
activity or substrate
specificity; (d) alter the physical
properties of the enzyme molecule as little as possible,
and hence (e) can be left on after purification.
These
criteria are only more or less fulfilled by the available
tails.
Hydrophobic and covalent tails usually do not have
the selectivity needed for efficient recovery of a protein,
especially for one-step purification.
Affinity tails based
on an antigenic
epitope suitable for immunoaflinity
purification may, for economic reasons, be attractive only
for laboratory-scale
production or high-value diagnostic
enzymes.
The eight-amino
acid Flag peptide
AspTyrLysAspAspAspAspLys
developed by Hopp et al. (62)
is an excellent example of this approach. The first four
residues of the Flag sequence are recognized
by a calcium-dependent
monoclonal
antibody;
the elution of the
NH2-terminal-tailed
enzyme
is achieved at decreased
calcium concentrations.
The residual residues of the
Flag peptide provide an enzymatic
cleavage site for enterokinase.
Charged tails may be advantageous
in the
lower-cost techniques
based
on ion-exchange,
are therefore studying thisapproach.
and we
Poly(Arg) Tails to Yeast a-Glucosidaseand A. niger GOD
A positively charged
tail of six arginine
residues was
attached to the COOH terminus
of a-glucosidase
(21,63).
This raised the isoelectric
point of natural a-glucosidase
from 5.6 to >8 for a-glucosidase-(Arg)6,
permitting
efficient one-step
cation-exchange
purification.
Furthermore, the (Arg)6 tail could be removed by treatment
with
carboxypeptidase
B. However, this approach
proved limited, mainly because
of partial proteolytic degradation
of
the tail by endo- and carboxypeptidases
already acting on
the tail during cell growth and during protein purification. This was observed
in both E. coli and wild-type
yeast cells. In addition, the instabifity
in E. coli of plasmids with a poly(Arg)-tail
has been noted by other groups
(58,64).
Yeast mutant strains lacking endoproteinases
A
and B and carboxypeptidases
Y and S proved useful for
cytoplasmic
expression
of a-glucosidase-(Arg)6
and permitted
the isolation of undigested
a-glucosidase-(Arg)6
(A. Grossmann, personal
communication).
Similar results were obtained for a COOH-terminal
(Arg)6-tail
fused to A. niger GOD (see above). Secretion
of GOD-(Arg)6 in wild-type
yeast strains leads to partial
to complete removal of the (Arg)6 tail, this removal
probably being caused by proteases
of the yeast secretory pathway that are needed for natural prepro a-factor
processing.These include KEX2 endoprotease,which is
specific for adjacent basic amino acids (Lys and Arg),
and KEX1, a carboxypeptidase
that cleaves COOH-terminal Arg and Lys residues (65). This problem may be
completely solved by the construction
and use of a yeast
mutant
strain lacking the proteases
KEX1 and KEX2.
In summary,
the use of tails consisting
of Arg or Lys
residues with a positive charge may usually be difficult.
Poly(His) Tails Designedfor Metal AffinityPurification
Pyruvate Oxidase (PyOD) and GOD
of
In our hands, metal chelate tails [poly(His)-tails]
designed for metal affinity chromatography
proved to be
most suitable for the recovery of several diagnostic enzymes. Metal affinity chromatography
of proteins is
based mainly on the ability of surface-accessible
His and
Cys residues to bind to immobilized
(chelated) first-row
transition metals such as Ni2, Zn2, and Cu2 (66, 67).
Commercially
available affinity resins containing iminodiacetate
or nitrilotriacetate
charged
with Zn2 or
Ni2, respectively,
are most commonly used as adsorbent. The usefulness of affinity tails composed of several
adjacent His residues was first demonstrated
for mouse
dihydrofolate
reductase (68).
A general strategy of engineering
an optimum metal
chelate tail is demonstrated
here for the cloned Lactobacillus plantarum
PyOD gene expressed in E. coli (H.
Burtscher,
personal
communication).
This enzyme is
used in various spectrophotometric
assays, e.g., for the
estimation
of transaminases,
pyruvate
kinase, lactic
acid, glycerol, free fatty acids, ADP, phosphate,
and
potassium.
To eliminate
traces of contaminating
and
interfering
enzyme impurities,
the enzyme requires extensive purification,
even if it is produced by a recombinant E. coli clone.
First, COOH-terminal
poly(His) tails ranging
in size
from two to six residueswere tested.A tailof three to four
His residues proved to be optimal with respect to binding
and elution under physiologicalconditions.PyOD-poly(His) enzymes were eluted by displacement
with histidine or imidazole at neutral pH. Increasing
amounts of
imidazole
or histidine
were needed
in relation to the
increasing
poly(His)chain length [e.g., -30 mmol/L and
40 mmol/L imidazole for PyOD-(His)3
and PyOD-(His)4,
respectively].
However, COOH-terminal
extensions
interfered with the enzymatic activity of PyOD and caused
about a fivefold decrease
in specific enzymatic
activity.
Hence, in a second step, a (His)3 tail was engineered
to
the NH2-terminus
of PyOD. The Met(His)3-PyOD
retained the same enzymatic
properties
as the natural
PyOD and was accessible to efficient metal affinity purification. However, the Met(His)3-PyOD
enzyme was not
as efficiently expressed
as the natural PyOD. Therefore,
in a third step, the NH2-terminal
part of Met(His)3-PyOD
was varied. An optimum
NH2-terminal
poly(His) tail
with respect to efficient translation
initiation
(overexpression), accessibility
to affinity purification,
and specific enzymatic
activity was constructed by insertion
of
three additional
His residues adjacent to the Met residue
at amino acid position 3 of natural
PyOD (Table 8).
Poly(His)-tagged
GOD is another example of the recovery of a secreted
enzyme
by the poly(His) tail fusion
technique.
The attachment
of four His residues to the
COOH terminus of GOD had no effect on expression or
secretion,
and the enzymatic propertiesof GOD-(His)4
approach those of wild-type
GOD. GOD-(His)4 was recovered directly from the culture medium to >98% homogeneity by using either Zn2-iminodiacetate
or Ni2nitrilotriacetate
adsorbents and an imidazole gradient
ranging from 0 to 100 mmol/L for elution.
Tails and Tags for Site-Specific
Immobilization
Besides being useful in enzyme purification,
affinity
tails offer, in principle,
the possibility of site-specific and
oriented enzyme immobilization.
This may be advantageous with respect to (a) high reproducibility
of enzyme
immobilization,
(b) elimination
of losses of enzyme activity caused by random (especially
covalent) enzyme
immobilization,
and (c) build-up of defined and oriented
enzyme
layers, e.g., for sensors or optical devices. How-
Table 8. Poly(Hls) tails engineered for metal affinity
purIfication of PyOD.
PyOD modification
N-MetValMetLys
C
N-Met----N-Met-(HIs)
------C
N-MetValMet(Hls)
C
Localization of
poly(HIs)
Wild-type
COOH terminus
NH2 terminus
NH2 terminus
(insertion)
Specific
acffvfty,
don
level
100
20
60
+
100
+++
CLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
897
ever, noncovalent
enzyme immobilization
(e.g., to the
surface of a ligand-coated
solid support) or enzyme coupling (e.g., for the preparation
of enzyme conjugates)
based only on the affinity of the purification
tails discussed above may not, in most cases, be stable enough
for diagnostic
uses with the exception of high-affinity
antigen (tail)-antibody
complexes.
Covalent purification
and site-specific immobilization
via a genetically
engineered
cysteine affinity tag was
recently demonstrated
for Bacillus subtilis glucose dehydrogenase
(G1cDH) expressed in E. coli (69). Natural
G1cDH, both in free and immobilized
forms,
has been
used for the estimation
of glucose. A single surfaceaccessible
active
cysteine
residue was introduced
at
amino acid position 44 by site-directed
mutagenesis of
cysteine-less tetrameric
GIcDH without loss of enzymatic activity. GlcDHcys44 immobilized
on thiopropylSepharose
via reversible
disulfide
bridge
formation
served as an initial
model system for studying
sitedirected covalent fixation. However, additional studies
are required
to demonstrate
the new possibilities
arising from site-specific
and oriented enzyme immobilization and their potential for use in new diagnostic components and devices.
Genetically Fused Antigen-Marker
Enzyme Conjugates
A component
of all ELISAs is either an antigenenzyme conjugate
or an antibody-enzyme
conjugate.
These conjugates are synthesized by chemical crosslinking reactions,
which give rise to several difficulties.
Generally,
conjugates
are heterogeneous
mixtures
with
respect to (a) the kind and number of conjugated
molecules (e.g., enzyme-antigen,
enzyme-enzyme,
antigenantigen, and more complex home- and heteropolymers)
and (b) the points where they are cross-linked.
Therefore, it is often necessary to isolate the conjugate of the
desired stoichiometry,
e.g., by size fractionation.
Furthermore, the functional
sites of a conjugate
(the active
site of the enzyme or the antigen-antibody
binding site)
can be more or less inactivated
by random chemical
cross-linking.
Even more potent marker enzymes such
as luciferase may be restricted
in application because of
their sensitivity
to inactivation
by established
chemical
cross-linking
procedures
(70). For these reasons
it
would be advantageous
to synthesize
enzyme-antigen
conjugates
directly by genetic means.
Competitive
ELISAs (Fig. 1) based on genetically
fused marker enzyme-antigen
conjugates
[e.g., fl-galactosidase-interferon-2a(antigen),
Rlknhine phosphataseproinsulin,and luciferase-Protein
Al have been studied
by Mosbach
and co-workers (71-73).
The luciferaseProtein A conjugate serves as a universal
marker enzyme conjugate
that can be attached noncovalently
to
any IgG molecule for use in bioluminescent
immunoassays. In the studies of Mosbach et al., the marker enzyme-antigen
conjugates
were synthesized
in E. coli
and proved functionally
active with respect
to enzymatic activity and antigen-antibody
binding. The detection limits for human
interferon-2a,
insulin,
and luciferase-Protein
A-IgG conjugate
were -50 /.Lg/L, 2.5
698
CLINICAL CHEMISTRY, Vol. 40, No. 5,
1994
p.gfL, and 9 x iO-’ mol, respectively.
The sensitivity
obtained for the estimation of insulin was close to that of
commercial insulin ELISAs, thus indicating the potential of this approach. However, the marker enzymes
used are not ideal with respect to subunit structure,
size,
and specific activity
(E. coli 3-galactosidase:
homotetramer, 465 kDa, 600 kU/g; E. coli alkaline phosphatase:
homodimer, 95 kDa, 50 kU/g, and Vibrio harveyi luciferase:
heterodimer).
From the enzymatic and genetic point of view, a
marker enzyme should be (a) small, (b) stable, and (c)
monomeric, and (d) should exhibit a high specific activity. Hence, the test sensitivity
and test convenience
may
be further
improved
if more suitable marker enzyme
genes become available,and if the technology needed for
their expression and, if necessary,
enzyme
renaturation,
is developed.
In this context, homodimeric
eukaryotic
1k2l1ine phosphatases and monomeric horseradish peroxidase (74) are promising enzyme candidates
because
of their high stability
and specific activity (up to 3000
kU/g and 2000 kU/g, respectively).
MultifunctionalEnzymes Made by Gene Fusion
In enzymatic analysis most assays are based on two or
more enzymes (Table 1). Therefore, with respect to enzyme production, it seems economically
advisable and
attractive to join the “free” serially operating enzymes
by genetic means and then produce only one multienzyme protein instead of two or several. The feasibility of
this concept is supported by naturally
occurring hi- and
multifunctional
enzymes (75) such as tryptophan synthase and the two multifunctional
subunits of the yeast
fattyacid synthetase
complex.
The most suitable order for two synthetically
fused
structural
genes at either the NH2 or COOH terminus
with respect to the encoded bifunctional
enzyme must be
determined
experimentally.
In this system the translational stop codon of the first gene is removed and replaced by a short peptide linker. The peptide linker may
be designed according
to the proposals
of Argos (76),
which rely on naturally occurring
interdomain
linkers
of proteins of known three-dimensional
structure.
Depending on the subunit structure of the natural enzymes
(monomeric,
dimeric,
oligomeric,
or heteromeric),
the
theoretically
expected subunit interactions
of an artificial bifunctional
enzyme can range from monomeric to
highly complex aggregated
structures as shown schematically
in Table 9. It should be noted that many
known enzymes are oligomeric.
Artificial
hi- and trifunctional
enzymes
have been
constructed
and compared with the unfused
enzymes
catalyzing the same sequential
reaction by Bulow and
Mosbach (77). In particular,
tetrameric E. coli $-galactosidase was fused to monomeric E. coli galactokinase
[/3-galactosidase-galactokinase
(78)1 and dimeric Pseudomonas fluorescens
galactose dehydrogenase
[/3-galactosidase-galactose
dehydrogenase
(79)1 and to both
enzymes [/3-galactosidase-galactose
dehydrogenase-galactokinase
(80)]. The formation of complex aggregates
as outlined
in Table 9 was expected for the (3-galacto-
Table 9. Expected subunit structure of a bifunctional
enzyme In relation to the subunit structure of the unfused
enzymes
Fused
Unfused enzymes
1. Monomer(a) +
enzymes
Monomer
monomer(p)
2. Monomer(a)
+
Dimer
ImprovedEnzymes for DiagnosticUses
dimer($::)
3.
Dimer(a::a)
+
dimer(p::)
4. Tetramer(a4) +
monomer($)
5. Tetramer(a4) +
Unear
polymer
Tetramer
(a-)4
Polymer
dimer($::)
The a and $ represent monomeric subunits of two differentenzymes.
sidase-galactose
and galactosidase-gahybrid enzymes.
However, both artificial enzymes assembled to give soluble complexes consisting
mainly of six and four subunits for galactosidase-galactose
dehydrogenase
and
four and eight subunits for galactosidase-galactose
dehydrogenase-galactokinase;
the authors also noted the
potential for formation of larger aggregates
in vitro.
The hybrid enzymes proved to be enzymatically
active
and displayed the two or three fused enzymatic
activities with specific activities
(corrected because of the
increased molecular mass) corresponding
to -50-100%
of those of the unfused enzymes. Furthermore,
the physical enzymatic properties such as Km values, pH optima,
and thermostabifity
were hardly affected. Under certain
circumstances
(e.g., low initial substrate
concentrations) the kinetic enzymatic behavior of the sequentially
operating hybrid enzymes was significantly
superior to
an identical system composed of the separate enzymes
with respect to (a) the time needed to reach a new
steady-state
rate and (b) a higher steady-state
rate per
se. These effects have been explained by substrate channeling, which may rely on favorable spatial orientation
of the active centers of the fused enzymes
or on the
formation of higher local intermediate
concentrations
caused by proximity of the serially operating enzymes.
In principle, the use of an artificial bifunctional
enzyme in enzymatic analyses has been demonstrated
for
the estimation
of lactose, with a coupled reaction catalyzed with fused (3-galactosidase
and galactose dehydrogenase. However, from a practical point of view, general
application
of this approach may be limited
for two
reasons: (a) Most enzymes are composed of two or more
subunits and, when fused to a second enzyme,
may undergo complex subunit interaction and, hence, aggregation followed by precipitation,
especially
during long
storage. (b) To compensate for the differing properties of
the individual enzymes, e.g., specific activity, substrate
lactose
dehydrogenase
dehydrogenase-galactokinase
affinity, pH optima, and activity
loss during
storage,
optimized serially operating analytical enzyme systems
do not usually contain the individual
enzymes in equal
(stoichiometric)
amounts. Nevertheless,
this approach
may be useful in special instances
and, in particular,
substrate or electron channeling may offer new possibilities for the development
of new biosensor devices.
Many diagnostic enzymes that are in use today are
only more or less ideal with respect to their physical
properties (e.g., heat, detergent,
and salt stability;
pH
and temperature
optima), and their enzymatic behavior
(e.g., catalytic activity, substrate specificity) under the
necessary
and often predetermined
test-kit conditions.
Frequently,
optimized assay conditions
are based on
complex reaction procedures due to the lack of enzymes
that enable the direct generation of an easily detectable
product or to the need for additional materials, including enzymes. Examples of such conditions are (a) sample
pretreatment
(e.g., with detergents,
salts, lipase, or esterase), (b) enzyme stabilization
(e.g., by sugars, polyalcohols, bovine serum albumin, or salts), and (c) suppression of interfering
reactions (e.g., by using ascorbate
oxidase or glutamate dehydrogenase
to remove ascorbic
acid or ammonia, respectively).
These problems underline the continuing need for improved enzymes.
To evaluate novel diagnostic enzymes, one should determine whether an existing enzyme
or gene could be
improved by chemical or genetic means or whether it is
more promising to screen for a novel enzyme that exhibits the desired properties. Screening
for a novel enzyme is indicated
when the desired
enzyme properties
differ tremendously
from those of any known enzyme.
For instance, the principleof DNA amplification
by PCR
was demonstrated
with the aid of the thermosensitive
E.
coli DNA polymerase
I (Klenow fragment),
and this
enzyme
was therefore
added after each cycle (81).
Hence, a thermostable
DNA polymerase
that could
withstand
repeated
exposure
to high temperatures
(-95#{176}C)
was needed. It was more practical to screen
thermophilic
microorganisms
for such enzymes than to
try to improve the Kienow polymerase. Several suitable
thermostable
DNA polymerases
have since been isolated and characterized,
and their genes have been
cloned and overexpressed
in E. coli, e.g., Taq DNA polymerase from T. aquaticus (27) and Vent
DNA polymerase from Thermococcus
litoralis (New England Biolabs, Mississauga,
Canada).
Enzyme
screening
is the more time-consuming
and
more costly approach, but is sometimes
the method of
choice. However,
when relatively
small changes are
needed, chemical enzyme derivatization
and immobilization or the redesigning
of an existing gene may be the
better alternative.
Usually, chemical intra- and intermolecular cross-linking
of enzymes, includingthe evaluation of soluble polymer-stabilized
enzyme conjugates,
may be indicated,
if only the physical properties of an
enzyme need to be changed, e.g., the heat, pH, salt, or
detergent
resistance
(12). Chemical stabilization
of enCUNICAL CHEMISTRY, Vol. 40, No. 5,
1994
699
zymes always means an additional step of downstream
processingand, hence, an increase in price.
In most instances, results comparable with those obtained by chemical
stabilization
may be obtained by
genetic means using random or site-directed
mutagenesis and subsequent screening to select for an improved
enzyme.
The genetic engineering
approach also offers
the construction
of novel catalytic
activities
such as
modulated or new substrate specificities,
different pH
and temperature
optima, or an increased rate of catalysis. The improvement
of a diagnostic enzyme by redesigning an existing gene is demonstrated
below for homodimeric
creatinase
(creatine amidinohydrolase)
from
P. putida (25).
Creatinase
is used in a new sequential
enzymatic
creatinine
determination
which, in contrast to the nonenzymatic Jaff#{233}
reaction(82), is more reliable with respect to reproducibility
and interference
of serum fluctuations (83). The serial enzymatic reaction is shown in
Table 1, reaction
3. Lipase, cholate, and other detergents are necessary
to clariIr the reaction mixture,
especially when the assay is carried out with lipemic sera.
However, the lipase/detergent
system included destabilizes natural creatinase.
Under the predetermined
assay
conditions,
creatinase
tends to denature and then to
form aggregates
that interfere with the assay. Since the
instability
of creatinase could be overcome by chemical
cross-linking
(R. Schmuck, personal
communication),
we sought to determine
whether this would be possible
by genetic means, too.
The creatinase
gene was cloned with the aid of a
plaque-screening
test based on an ELISA and then subcloned and overexpressed in E. coli (25). Random plasmid mutagenesis
and a creatinase
activity plate assay
were used for the isolation of more detergent-resistant,
mutant creatinase
enzymes. Briefly, E. coli cells carrying the creatinase
expression
plasmid
were exposed to
nitrosoguanidine,
and the mutagenized
plasmid
DNA
was isolated and reintroduced
into unmutagenized
E.
coli cells (84). The colonies so derived were transferred
to nitrocellulose
filters and lysed with chloroform/toluene vapor. The filters with the bound cell lysates were
then exposed to increased
concentrations
of the destabilizing agent (e.g., cholate) or a destabilizing
physical
parameter
(e.g., increased temperature)
under predetermined test conditions. After this pretreatment
the filters were placed upside down on an agar plate containing appropriate
staining
reagents
for detecting
creatinase
activity (25). The pretreatment
and selection
steps were adjusted so that colonies that expressed the
wild-type creatinase
were negative and hence only improved mutant
enzymes
were visible. After repeated
screening,
three
positive
colonies
remained
out of
>50 000. Enzymatic characterization
established
that
the isolated mutant creatinase
enzymes were indeed
more detergent resistant.
For example, the remaining
creatinase
activity of the wild-type and the single-mutant creatinase
SM-3051 (Ala’#{176}9Val’#{176}#{176})
was 27% and
84%, respectively,
after 1 h of incubation
under standard assay conditions.
-
700
CLINICAL CHEMISTRY, Vol. 40, No. 5,
1994
Further improved creatinase
mutants
were obtained
(a) by combining different stabilizing
mutations
at the
DNA level and (b) by repetition of the mutagenesis
and
activity-screening
procedure with a stabilized
creatinase mutant already isolated. These creatinase doublemutants (e.g., DM-414-4: Ala’#{176}9 Val’#{176}9;
Val355
Met355) were as stable as the chemically
cross-linked
creatinase.
The three-dimensional
(3D) structure of creatinase
has been determined
(85), thus enabling the interpretation of the stabilizing
amino acid exchanges
in the
context of the structural model. This encouraged us to
use creatinase as a model system to study the value of
predicted stabilizing amino acid substitution(s)
deduced
with the aid of the 3D structure. However, all creatinase
mutants designed on the assumption that a more spacefilling amino acid in an area of creatinase of low electron
density might be stabifizing
(e.g., by additional hydrophobic interactions)
were only as stable as the wild-type
enzyme (L. RUssmann, personal communication).
To summarize,
the success of random mutagenesis
and screening for a mutant enzyme depends mainly on
the feasibility
of setting up an efficient discriminating
selection and screening system for the desired phenotype. In contrast, 3D structure-based
enzyme engineering is essentially
more complex. Our understanding
of
the factors that stabilize proteins or allow for efficient
catalysis is somewhat limited, although good progress is
being made (86, 87).
-
Genetically Engineered p-Galactosidase
Homogeneous
Immunoassay
-
for a New
System
A novel homogeneous
immunoassay
system, CEDIA,
has been developed on the basis of the well-known a-cornplementation
reaction of E. coli /3-galactosidase
(88, 89)
(see Fig. 5). Active tetrameric
f3-galactosidase
is formed
by spontaneous
association
of a short NH2-terminal
/3-galactosidase
fragment of -70-90
amino acids (the a-polypeptide, also termed enzyme donor, ED) and a (3-galactosidase monomer with a small deletion of up to -50 amino
acids near the N}12 terminus
(enzyme acceptor, EA). Ligands (haptens) can be chemically attached to the a-peptide in such a way that they do not interfere with association and enzymatic activity of /3-galactosidase.
This
assembly of active /3-galactosidase
can be inhibited by a
ligand (hapten)-speciflc
antibody. Therefore, one can set
up a homogeneous
assay based on the competition
between the ligand (analyte) in the sample and the EDligand (hapten) conjugate for the antibody. Within 5-15
miii this leads to the development of a colorimetric signal
(rate of substrate hydrolysis by the enzyme) that is directly proportional to the amount of analyte present in
the sample.
In the CEDIA test, genetic engineering
was used for
the generation and selection of appropriate ED-EA pairs
with respect to (a) favorable kinetics of formation of active /3-galactosidase,
(b) enabling site-directed
chemical
coupling of a ligand to the ED fragment, and (c) the
stability of both /3-galactosidase
fragments.In particular,
numerous
EAs were prepared that varied in size and in
chromogenic
Basic Reaction: 13-galactosidasecomplementation with ED-ligand conjugate
substrate
w
complementation
+
subunit association
colour
EA
ED-ligand
monomer
conjugate
inactive
inactive
tetrameric
enzyme
active
No analyte:
-<
-<
-<
-<
Y YY
immobilized antibody
plus EA monomer
-(#{149}1
“C
Y YY
no active
enzyme
formed
addition of substrate
and ED-ligand conjugate
Sample with analyte:
WV
.<
WV
-<.
-.
-#{247}
‘(
YY
immobilized antibody
plus EA monomer
.
free ligand
addition of substrate
and ED-ligand conjugate
Y
anti-ligand
V
EA-monomer
antibody
ED-ilgand
conjugate
**
active enzyme
Fig.5. Principleof homogeneous
immunoassay
basedon engineered/3-galactosidase
(CEDIA).
the position of the deleted amino acidsin the NH2-terminal region
of the /3-galactosidase
monomer. Numerous
ED polypeptides
were constructed
by site-directed mutagenesis (90) containing
either a single cysteine
or lysine
residue at various amino acid positions suitable for sitedirected coupling of ligand (hapten)-maleimide
or ligand
(hapten)-N-hydroxysuccinimide
ester derivatives,
respectively. Functional
EAs were then preselected by cornCLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
701
plementation
in vivo (E. coli expressing
an appropriate
ED gene). Furthermore,
two E. coli overproducers
were
constructed
for the production of the EA and the ED
/3-galactosidase
fragment. The small ED fragment could
not be purified from E. coli host cells directly, because of
proteolytic degradation
inside the cells. Hence, the small
ED fragment
was coexpressed with the large EA fragment to form activeand stable/3-galactosidase
in vivo.
After purification of the active /3-galactosidase,
the EA
polypeptide was recovered by denaturation of the enzyme
complex with 6 molIL urea.
This technology has found commercial application in
clinical laboratories for the estimation of a wide range of
analytes
(91), including hormones (e.g., thyroxine and
cortisol), vitamins
(e.g., B12 and folate), and therapeutic
drugs (e.g., phenytoin and digoxin).
Concluding Remarks
Enzymes derived by recombinant
DNA technology
are becoming
more common both in industry, where
they are used as biocatalysts,
and in medicine, where
they are used as therapeutic agents and diagnostic tools.
The advantages
of such enzymes include potential increased enzyme yield (up to 50% of the total cellular
protein), simplified downstream
processing,
increased
product quality, process profitability
and safety (e.g.,
substitution
for pathogenic
natural source material).
Use of these enzymes also has environmental
advantages, such as reduced raw material, energy, and waste.
In those cases where a microbial host fails to express
the enzyme efficiently or to secrete it in an enzymatically active form, in vitro renaturation
of insoluble aggregated protein (e.g., from E. coli) is a possible alternative. This strategy is attractive
for two of the most
commonly
used enzyme
labels, i.e., mammalian
alkaline phosphatases
and plant peroxidases.
It remains to be seen whether diagnostic
enzymes
will be as important
in the future as in the past. Research is now focusing on the development
of sensors,
especially immunosensors,
that permit sensitive determination of analytes
over a wide dynamic
range
in a
matter
of seconds to minutes
in a homogeneous
test
format. This determination
may be accomplished,
e.g.,
by opticalor electrochemical
signal detection of nonenzymatic reporter groups, with the use of such techniques as surface plasmon resonance (92), evanescent
wave (93), and electrochemiluminescence
(94). Furthermore, miniaturization
of already established
commercially available tests may reduce the quantities
of
diagnostic enzymes needed.
Nevertheless,
clinical DNA-based
diagnosis is a new
area that holds promise for the use of new diagnostic
enzymes,
indudung
thermostable
polymerase8
and ligases for thermal cycling methods such as PCR and LCR
[ligase chain reaction
(95)], and several jointly acting
enzymes (e.g., reverse transcriptase,
RNase H, T7-RNApolymerase,
and DNA polymerase)
for isothermal
cycling methods such as NASBA and 3SR [self-sustained
sequence replication, (96)1.
702
CLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
We thank our colleagues at Boehringer Mannheim for comments and for providing unpublished information. Special thanks
are due to D. Ambrosius, H. Burtscher, A. Grossmann, M. Jarsch,
G. Kreile, U. Michaelis, L. RUssmann, R. Rudolph, R. Schmuck, G.
Schumacher, and R. looper for reading the manuscript.
References
1. Bergmeyer HU, ed. Methods of enzymatic analysis, 3rd ed.
Weinheim and New York: VCH Publishers, Vol. 1-10, 1983.
2. Warburg 0, Christian W. Pyridin, der wasserstoffubertragende
Bestandteil von Garungsfermenten
(Pyridin-Nucleotide). Biochem
Z 1936;287:291-328.
3. GoslingJP. A decade of development in immunoassay methodology [Review]. Cliii Chem 1990;36:1408-27.
4. Leland C, Clark JR. The enzyme electrode. In: Turner APF,
Karube I, Wilson GS, eds. Biosensors: fundamentals and applications. Oxford: Oxford University Press, 1987:1-12.
5. Bartlett PN, Whitaker RG. Strategies for the development of
amperometric enzyme electrodes. Biosensors 1988;3:359-79.
6. Free AB, Free HM. A simple test for urine bilirubin. Gastroenterology 1953;24:414-21.
7. Bates DL. Enzyme amplification in diagnostics. Trends Biotechnol 1987;5:204-9.
8. Johannsson A, Bates DL. Amplification by second enzymes. In:
Kemeny DM, Challacombe SJ, eds. ELISA and other solid phase
immunoassays. Chichester: John Wiley & Sons, 1988:85-106.
9. Mullis KB, Faloona FA. Specific synthesis of DNA in vitro via
a polymerase-catalyzed
chain reaction. Methods Enzymol 1987;
155:335-50.
10. Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR
protocols: a guide to methods and applications. San Diego: Academic Press, 1990.
11. Malek L, Darasch 5, Davey C, Henderson G, Howes M, Lens P,
et al. Application of NASBA” isothermal nucleic acid aniplification method to the diagnosis of HIV-1 [Abstract]. Clin Chem
1992;38:458.
12. Glazer AN, DeLange RJ, Sigman DS. Chemical modification
of proteins. Selected methods and analytical procedures. In: Work
TS, Work E, eds. Laboratory techniques in biochemistry and
molecular biology, Vol. 4. Amsterdam: North-Holland Publishing
Company, 1976:1-206.
13. Siest G, Oster T, Visvikis A, Thioudellet C, Angele C, Schiele
F, et al. Application of gene transfer technologies to the production
of enzyme reference materials. Example of y-glutaxnyltransferase
[Review]. Clin Chem 1993;39:1573-89.
14. Quax WJ, Mrabet NT, Luiten RGM, Schuurhuizen PW, Stanssons P, Lasters I. Enhancing the thermostability
of glucose
isomerase by protein engineering.
Biotechnology (NY) 1991;9:
738-42.
15. Francetic 0, Deretic V, Max:janovic N, Glisin V. Penicillin
acylases: their function, structures, genes and commercial applications. Biotech-Forum 1988;5:90-4.
16. Arbige MV, Pitcher WH. Industrial enzymology: a look towards the future. Trends Biotechnol 1989;7:330-5.
17. Flohe L, GOnzlerWA. Enzymes in therapy. In: Gerhartz W, ed.
Enzymes in industry. Production and applications. Weinheim and
New York: VCH Publishers, 1990:178-84.
18. Cheetham PSJ. The applications of enzymes in industry. In:
Wiseman
A, ed. Handbook
of enzyme biotechnology.
Chicheater:
John Wiley & Sons, 1985:274-379.
19. Krelle GB. Enzymes in analysis and medicine, survey. In:
Gerhartz W, ad. Enzymes in industry. Production and applications.
Weunheim and New York: VCH Publishers, 1990:151-3.
20. Marriage HJ, Chowdrey PE. Manufacture
of diagnostic enzymes. Biochem Soc Trans 1990;18:240-2.
21. Kopetzki E, Buckel P, Schumacher G. Cloning and characterization of baker’s yeast a-glucosidase:
overexpression
in a yeast
strain devoid of vacuolar proteunases. Yeast 1989;5:11-24.
22. Winnacker EL. From genes to clones: introduction to gene
technology. Weinheim and New York: VCH Publishers, 1987.
23. Higuchi R. Using PCR to engineer DNA. In: Erlich HA, ad.
PCR technology: principles and applications for DNA amplification. New York: Stockton Press, 1989:61-70.
24. Jarsch M, Lang G. Klonierung und Uberexpression
von Glucoee-6-phosphat
Dehydrogenase
aus Leuconostoc dextranicus.
European patent application no. 0489523, 1992.
25. Schumacher G, Hilscher W, Mollering H, Siedel J, Buckel P.
Production and improvement of enzymes by genetic engineering.
In: Galteau MM, Siest G, Henny J, eds. Biologie prospective.
Comptes rendus du 7 Colloque de Pont-a-Mousson. Paris: John
Libbey Eurotext,
1989:57-64.
26. Buckel F, Zehelein E. Expression of Pseudomonas fluoreacens
D-ga]actose dehydrogenase in E. coil. Gene 1981;16:149-59.
27. Lawyer FC, Stoffel RK, Saiki K, Myambo R, Druminond R,
Gelfand DI. Isolation, characterization, and expression in Escherichia coli of the DNA polymerase gene from Thermus aquaticus.
J Biol Chem 1989;264:6427-37.
28. Marston
FAO. The purification
of eukaryotic polypeptides
synthesized in Escherichia coli [Review]. Biochem J 1986;240:112.
29. Cheng YSE. Increased cell buoyant densities of protein overproducing Escherichia coli cells. Biochem Biophys Res Commun
1983;111:104-11.
30. Kiethaber T, Rudolph R, Kohier HH, Buchner J. Protein
aggregation in vitro and in vivo: a quantitative model of the
kinetic competition between folding and aggregation [Review].
Biotechnology
(NY) 1991;9:825-9.
31. Ellis RJ, van der Vies SM. Molecular chaperones. Ann Rev
Biochem 1991;60:321-47.
32. Buchner J, Rudolph R. Routes to active proteins from transformed microorganisms.
Curr Opin Biotechnol 1992;2:532-8.
33. Chang J, Gotcher 5, Gushaw JB, Gadsden RH, Bradley CA,
Stewart TC. Homogeneous enzyme immunoassay
for theophylline
in serum and plasma. Clin Chem 1982;28:361-7.
34. Bhadbhade MM, Mams MJ, Flynn TG, Levy HR. Sequence
identity between a lysine-containing peptide from Leuconostoc
mesenteroides glucose-6-phosphate dehydrogenase and an active
site peptide from human erythrocyte glucose-6-phosphate dehydro-
genase. FEBS Lett 1987;211:243-6.
35. Hunkapil1er MW, Hewick EM, Dreyer WJ, Hood LE. Highsensitivity sequencing with a gas-phase sequenator. Methods Enzymol 1983;91:399-413.
36. Suggs SV, Wallace RB, Hirose T, Kawashima EH, Itakura K.
Use of synthetic oligonucleotides as hybridization probes: isolation
of cloned cDNA sequences for human 132-microglobulin. Proc Nati
Acad 8th USA 1981;78:6613-7.
37. Grunstein M, Hogneas DS. Colony hybridization: a method for
the isolation of cloned DNAs that contain a specific gene. Proc Nati
Acad Sci USA 1975;72:3961-5.
38. Richter G. Glucose oxidase. In: Godfrey T, Reichelt J, eds.
Industrial enzymology. The application of enzymes in industry.
New York: Macmillan, 1983:428-36.
39. van Dijken JP, Veenhuis M. Cytochemical
localization of
glucose oxidase in peroxisomes of Aspergillus niger. Eur J Appl
Microbiol Biotechnol 1980;9:275-83.
40. Mischak H, Kubicek CP, R#{246}hr
M. Formation and location of
glucose oxidase in citric acid producing mycelia of Aspergillus
niger. Appl Microbiol Biotechnol 1985;21:27-31.
41. Kriechbaum M, Heilmann HJ, Wientjes FJ, Hahn M, Jany
KD, Gassen HG, et al. Cloning and DNA sequence analysis of the
glucose oxidase gene from Aspergillus niger NRRL-3. FEBS Lett
1989;255:63-6.
42. Frederick KR, TungJ, Emerick RS, Masiarz FR, Chamberlain
SH, Vasavada A, et al. Glucose oxidase from Aspergillus niger.
Cloning, gene sequence, secretion from Saccharomyces cerevisiae
and kinetic analysis of a yeast-derived enzyme. J Biol Chem
1990;265:3793-802.
43. Whittington
H, Kerry-Williams
5, Bidgood K, Dodsworth N,
Peberdy J, Dobson M, et al. Expression of the Aspergillus niger
glucose oxidase gene in A. niger, A. nidulans and Saccharomyces
cerevisiae. Curr Genet 1990;18:531-6.
44. De Baetselier A, Vasavada A, Dohet P, Ha-Thi V, De Beukelaer M, Erpicum T, et al. Fermentation of a yeast producing A.
niger glucose oxidase: scale-up, purification and characterization
of the recombinant enzyme. Biotechnology (NY) 1991;9:559-61.
45. Ballou CE. The yeast cell wall and cell surface. In: Strathern
JN, Jones EW, Broach JR, eds. The molecular biology of the yeast
Saccharomyces
cerevisiae. Metabolism and gene expression. Cold
Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1982:
335-60.
46. Kukuruzinska
MA, Bergh MLE, Jackson BJ. Protein glycosylation in yeast. Ann Rev Biochem 1987;56:915-44.
47. Tanner W, Lehie L. Protein glycosylation in yeast [Review].
Biochim Biophys Acta 1987;906:81-99.
48. Lehle L, Lehnert K, Kopetzki E. HefewirtsstSmme mit Defekten in der N-Glycosylierung. European patent application no.
93112338.4, 1993.
49. Erhart E, Hollenberg CF. The presence of a defective LEU2
gene on 2z DNA recombinant plaamids of Saccharomyces cerevisine is responsible for curing and high copy number. J Bacteriol
1983;156:625-35.
50. McComb RB, Bowers GN Jr, Posen 5, eds. Alkaline
phos-
phatases. New York and London: Plenum Press, 1979.
51. Millan JL. Molecular cloning and sequence analysis of human
placental alkaline phosphatase. J Biol Chem 1986;261:3112-5.
52. Low MG, Saltiel AR. Structural and functional roles of glycosyl-phosphatidylinositol
in membranes. Science
1988;239:288-75.
53. Millan JL. Oncodevelopmental expression and structure
alkaline phosphatase genes. Anticancer Res 1988;8:995-1004.
54. Mi]lan JL. Oncodevelopmental alkaline phosphatases:
of
in
search for a function. Proc Chin Biol Res 1990;344:453-75.
55. Stuber D, Matile H, Garotta G. System for high-level production in Escherichia coli and rapid purification of recombinant
proteins: application to epitope mapping, preparation of antibodies, and structure-function
analysis. Immunol Methods 1990;4:
121-52.
56. Rudolph R. Renaturation of recombinant, disuliide-bonded
proteins from “inclusion bodies.” In: Tschesche H, ad. Modern
methods in protein and nucleic acid research. Berlin and New
York: Walter de Gruyter, 1990:149-71.
57. Michaelis U, Rudolph R, Jarsch M, Kopetzki E, Burtacher H,
Schumacher G. Verfahren zur Gewinnung rekombinanter, biologisch aktiver, eukaryontischer cilkcdiRcher Phosphatase.
European
patent application no. 93112289.9, 1993.
58. Ford CF, Suomunen I, Glatz CE. Fusion tails for the recovery
and purification of recombinant proteins [Review]. Protein Express
Purif 1991;2:95-107.
59. Hammond PM, Atkinson T, Sherwood RF, Scawen MD. Manufacturing new-generation proteins: part I. Pharmacol Technol mt
1991;14:24-32.
60. Sassenfeld HM. Engineering proteins for purification [Review]. Trends Biotechnol 1990;8:88-93.
61. Ljungquist C, Breitholtz A, Brink-Nilsson H, Mob T, Uhlen
M, Nilsson B. Immobilization and affinity purification of recombinant proteins using histidine peptide fusions. Eur J Biochem
1989;186:563-9.
62. Hopp TP, Prickett KS, Price VL, Libby RT, March CJ, Cerretti
DP, et al. A short polypeptide marker sequence useful for recombinant protein identification and purification. Biotechnology (NY)
1988;6:1204-10.
63. Kopetzki E, Schumacher G, Buckel P. Control of formation of
active soluble or inactive insoluble baker’s yeast a-glucosidase
F!
in Escherichia coli. Mol Gen Genet 1989;216:149-55.
64. Skerra A, Pfitzinger I, Pluckthun A. The functional expression
of antibody Fv fragments in Escherichia coli: improved vectors and
a genetically applicable purification technique. Biotechnology
(NY)
1991;9:273-8.
65. Jones EW. Three proteolytic systems in the yeast Saccharomyces cerevisiae [Munireview]. J Biol Chem 1991;266:7963-6.
66. Porath J, Carlsson J, Olsson I, Beifrage G. Metal chelate
affinity chromatography
a new approach of protein fractionation.
Nature 1975;258:598-9.
67. Arnold FH. Metal-affinity separations: a new dimension in
protein processing [Review]. Biotechnology (NY) 1991;9:151-6.
68. Hochuli E, Bannwarth W, D#{246}beli
H, Gem R, Stiiber D.
Genetic approach to facilitate purification of recombinant proteins
with a novel metal chelate adsorbent. Biotechnology (NY) 1988;6:
1321-5.
69. Persson M, Bulow L, Mosbach K. Purification and site-specific
immobilization
of
genetically
engineered glucose dehydrogenase
on thiopropyl-Sepharose.
FEBS Lett 1990;270:41-4.
70. Kricka U. Chemiluminescent and bioluminescent
[Review]. Clin Chem 1991;37:1472-81.
techniques
71. Peterhans A, Mecklenburg M, Meussdoerffer F, Mosbach K. A
simple competitive enzyme-linked
iminunosorbent
assay using
antigen-3-galactosidase fusions. Anal Biochem 1987;163:470-5.
72. Lindbladh C, Persson M, Billow L, Stahl 5, Mosbach K The
design of a simple competitive ELISA using human proinsulinCLINICAL CHEMISTRY, Vol. 40, No. 5, 1994
703
alkaline phoephatase conjugates prepared
Biophys Res Commun 1987;149:607-14.
73. Lindbladh C, Mosbach K, Billow L.
cally fused Protein A/luciferase conjugate
cent immunoassays.
J Immunol Methods
74. Smith AT, Santama N, Dacey 5,
by gene fusion. Biochem
of a genetifor use in biolumines1991;137:199-207.
Edwards M, Bray RC,
Thorneley RNF, et al. Expression of a synthetic gene for horseradish peroxidase C in Escherichia coli and folding and activation of
the recombinant enzyme with Ca2 and heme. J Biol Chem
1990;265:13335-43.
75. Bisswanger H, Schminke-Ott E, eds. Multifunctional
proteins.
New York John Wiley & Sons, 1980.
76. Argos P. An investigation of oligopeptides linking domains in
protein tertiary structures
and possible candidates for general
gene fusion. J Mol Biol 1990;211:943-58.
77. Billow L, Mosbach K Multienzyme systems obtained by gene
fusion. Trends Biotechnol 1991;9:226-31.
78. Billow L, Ljungcrantz P, Mosbach K Preparation of a soluble
bifunctional
821-3.
enzyme
Preparation
by gene fusion. Biotechnology
(NY) 1985;3:
79. Ljungcrantz P, Carisson H, Mansaon MO, Buckel P, Mosbach
K, Billow L. Construction
of an artificial bifunctional enzyme,
p-galactosidaae/ga]actose
dehydrogenase, exhibiting efficient galactose channeling. Biochemistry
1989;28:8786-92.
80. Ujungcrantz F, Billow L, Mosbach K Construction and characterization of a recombinant tripartite enzyme, galactose dehydrogenase/-gaIactosidase/ga1actokinase.
FEBS Lett 1990;275:
91-4.
81. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich
HA, et al. Enzymatic amplification of p-globin genomic sequences
and restriction site analysis for diagnosis of sickle cell anemia.
Science 1985230:1350-4.
82. Butler AR. The Jaff#{233}
reaction. Identification of the coloured
species. Clin Chim Acta 1975;59:227-32.
83. Siedel J, Mollering H, Ziegenhorn J. Sensitive color reagent
for the enzymic determination of creatinine [Abstract]. Clin Chem
1984;30:968-9.
84. Talmadge K, Gilbert W. Construction of plasmid vectors with
704
CUNICAL CHEMISTRY, Vol. 40, No. 5, 1994
unique Pst I cloning sites in a signal sequence coding region. Gene
1980;12:235-41.
85. Hoefiken 1W, Knof SI, Bartlett PA, Huber R, Moellering H,
Schumacher G. Crystal structure determination, refinement and
molecular model of creatine amidinohydrolase
from Pseudomonas
putida. J Mol Biol 1988;204:417-33.
86. Nosoh Y, Sekiguchi T. Protein engineering for thermostability. Trends Biotechnol 1990;8:16-20.
87. Langone JJ, ad. Molecular design and modeling concepts and
applications. Part A. Proteins, peptides, and enzymes. Methods
Enzymol
1991;202.
88. Henderson DR, Friedman SB, Harris JD, Manning WB, Zoccoli M. CEDIA’, a new homogeneous immunoassay system. Clin
Chem 1986;32:1637-41.
89. Khanna FL, Dworschack RT, Manning WB, Harris JD. A new
homogeneous enzyme immunoassay using recombinant enzyme
fragments. Chin Chim Acta 1989;185:231-40.
90. Henderson DR. Methods for protein binding enzyme complementation assays. US patent no. 4708929, 1987.
91. Kaspar P, Khanna FL. Convenient therapeutic drug monitoring [Applic Note]. Eur Clin Lab 1991;10:24.
92. Cullen DC, Brown RGW, Lowe CR. Detection of immunocomplex formation via surface plasmon resonance on gold-coated
diffraction gratings. Biosensors 1988;3:211-25.
93. Bluestein BI, Walczak IM, Chen SY. Fiber optic evanescent
wave immunosensors for medical diagnostics. Trends Biotechnol
1990;8:161-8.
94. Blackburn GF, Shah HP, Kenten JH, Leland J, Kamin RA,
Link J, et al. Electrochemiluminescence detection for development
of immunoassays
and DNA probe assays for clinical diagnostics.
Clin Chem 1991;37:1534-39.
95. Barany F. The ligase chain reaction [Review]. PCR Methods
Applic 1991;1:5-16.
96. Fahy E, Kwoh DY, Gingeras TR. Self-sustained
sequence
replication (3SR): an isothermal transcription-based
amplification
system alternative to PCR [Review]. PCR Methods Applic 1991;1:
25-33.