close

Вход

Забыли?

вход по аккаунту

?

High-Pressure Liquid Chromatography (HPLC) of Proteins [New Analytical Methods (29)].

код для вставкиСкачать
High-pressure Liquid Chromatography (HPLC) of Proteins
New Analytical
Methods (29)
By Cerhard Seipke," Hubert Miillner, and Ulrich Crau
The application of high-pressure liquid chromatography (HPLC) to proteins has undergone
a dramatic development in recent years. Nowadays its many variants expand the repertoire
of high-performance analysis methods available to the protein chemist, which, until now,
have been dominated by electrophoretic techniques. The advent of gene technology has
resulted in a renaissance of protein chemistry. The new analytical and preparative problems
that have thereby emerged are often ideally solved by HPLC methods. HPLC has long since
ceased to be solely a laboratory technique; HPLC systems are now being developed for the
separation of proteins-particularly those of great pharmaceutical interest-on a 100-g
scale. The range of applications of analytical and preparative HPLC will be illustrated by
two examples of pharmaceutical importance-insulin and interleukin 2.
1. Introduction
Liquid chromatography has become one of the most important biochemical separation methods since its discovery
more than eighty years ago. A differential distribution of
biopolymers between a stationary and a mobile phase can
be achieved by exploiting differences in molecular size,
ionic properties, solubility, and polar or nonpolar character as well as specific interactions with other molecules
(e.g., enzyme-inhibitor interactions). Earlier, however, the
packing materials available precluded rapid high-resolution separations, so that these methods were mainly employed at low pressure for preparative purposes. The analytical sector was dominated by electrophoretic methods.
High-resolution chromatography requires microparticulate column packings with a narrow particle size distribution.".21 Although this concept was known very early, its
realization in practice took decades since it was necessary
to solve several technical problems:
- Pump systems providing constant flow rates at high
pressures had to be developed.
- The "soft" gels previously used had to be replaced by
-
pressure-stable materials.
Columns had to be reproducibly and evenly packed
with these fine particles.
Once the instrumental requirements had been fulfilled,
the application of HPLC in biochemistry underwent a dramatic development. Within a few years a wide range of
new analytical methods have become available to protein
chemists. The applications of HPLC (Table 1) are not restricted to analysis, however. The new techniques are being
increasingly employed preparatively as well-both in the
laboratory and in industrial production.
2. The Construction and Operation of an H P L C
Unit
HPLC is of itself not a new method, but rather a refinement of known methods combined with greater instrumen[*] Dr. G . Seipke, Dr. H. Miillner, Dr. U. Grau
Pharma Forschung Biochemie der Hoechst A C
Postfach 800320, D-6230 Frankfurt am Main 80 (FRG)
Angun, Chem. In8 Ed.
En$. 25 11986) 535-552
tal complexity and precision. The successful application of
HPLC as a quantitative analytical method requires great
care in the choice and purity of the solvent, a highly constant flow rate, and continual checks of column performance.
Table I . Some applications of HPLC in the chemistry of amino acids, peptides, and proteins.
Analysis
-
-
Amino acid analysis
Purity control of synthetic peptides, optlmization of reaction conditions
Identification of multiple or mutant forms of proteins
Molecular weight determination
Analytical peptide mapping
Sequence analysis (identification of degradation products)
Chemical or enzymatic modification (process control and optimization,
product characterization)
Determination of enzyme specificity
Metabolism studies, identification of prohormones
Preporailfie seporarion
- Isolation of natural products from extracts
- Purification of products obtained from chemical synthesis or gene technology
- Purification of modified peptides or proteins, for example, radioactively
labeled substances
- Separation of diastereomers
- Isolation of protein fragments for sequence studies
A typical HPLC apparatus (Fig. 1) consists of a highpressure pumping system, a sample applicator (injector
and preinjector), a column, and a detector. It is also possible to include a fraction collector and/or a microcomputer
for data processing. Automation is desirable if the instrument is to be employed for routine analytical o r preparative purposes.
2.1. Pumps
All HPLC pumps on the marketf3]are capable of delivering both organic solvents and aqueous buffer solutions
with equally high precision. The pumps must provide a
constant flow with little pulsation. These requirements are
ideally fulfilled by a non-reciprocating pump resembling a
large hypodermic syringe, whose delivery is free of pulsation and almost independent of the back pressure of the
column and the viscosity of the solvent. Its use is relatively
restricted, however, because of its design, which involves a
fixed, relatively large internal volume.
0 VCH Verlagsge.sellschaft mbH. 0-6940 Weinheim, 1986
0570-0833/86/0606-0535 % 02 50/0
535
dients are sufficient for most applications. A third solvent
can be useful for the application of larger volumes of sample in preparative work or for the addition of a second,
more strongly eluting solvent at the end of a gradient in
reversed-phase chromatography." ' I
2.3. Eluents
71solvent flow
__
information f l o w
___--
optional
Fig I . Basic deslgn of an HPLC system
The delivery systems most frequently in use employ
either short-stroke piston pumps or membrane piston
pumps. Complex electronic controls are employed to ensure that the delivery volume set is maintained under
changing back pressure (altered compressibility). Optimized pump design reduces the pulsation caused by the
movement of the pistons. Further smoothing is usually
achieved by hydraulic damping.
2.2. Gradient Formers
The composition of the mobile phase can be varied during the analysis by mixing two or more solvents continuously (gradient elution). This technique predominates in
the separation of proteins by adsorption chromatography.
Gradient elutions yield peaks of equal sharpness throughout the whole course of elution. I n contrast to isocratic
elution, in which the solvent composition remains unchanged, gradient elution results in the same detection sensitivity for strongly retarded components as well as a
shorter analysis time.Ia71 Proteins bind cooperatively to
more than one portion of the stationary p h a ~ e . [ ~Only
- ' ~ the
simultaneous effect of a corresponding number of molecules of the "displacing agent" (for example, salt in ionexchange chromatography, organic solvent in reversedphase chromatography) is able to displace them. From this
it can be deduced that there will be an increasingly concave dependence of retention on the elution composition
with increasing molecular ~ e i g h t . l ~ . " In
' ~ practice, if the
displacer concentration is slowly increased, extremely
strong adsorption (that is, minimal migration) of the proteins is initially observed. As the concentration is further
increased, adsorption and thus retention strongly decreases within a range that becomes narrower with increasing molecular weight and whose position can be extremely
different from protein to p r ~ t e i n . [ ~ -Gradient
"'~
elution is
therefore essential for the separation of complex mixtures
of proteins, but its use places great demands on the precision of such delivery systems.
In the great majority of cases the solvents are mixed by
employing two separate, adjustable pumps. Binary gra536
The choice of eluents depends, above all, on the separation principle (see Section 3); further limitations are set by
the compatibility of the eluents with the stationary phase
(see Section 2.5) and with the materials from which the
HPLC equipment is constructed. Warnings are often given
against the use of buffer solutions containing halides. In
practice, however, it is possible to avoid corrosion of the
pumps and columns."21 The choice of buffer components
and organic solvents is also dictated by the method and
required sensitivity of detection (see Section 2.6). The viscosity must also be taken into account since it not only
contributes to the back pressure of the column, but also
influences the separation efficiency of a column (measured
in theoretical plates A"'. I3l).
Filtration of all eluents through a 0.2- or 0.45-pm membrane filter not only protects the pumps, but also improves
the long-term stability of the separation columns. Subsequent thorough degassing removes dissolved oxygen,
which would make sensitive detection at low wavelengths
impossible and could lead to problems when mixing gradients (formation of air bubbles). Sparging with helium is
very effective and does not significantly affect the composition of the mobile
In the case of pump designs
that are sensitive to dissolved helium, degassing can be
performed under vacuum or ultrasonically.
2.4. Sample application
The analysis of complex protein mixtures and the reequilibration of the columns requires a relatively long
time. Automatic injection of the samples ensures economical use of the expensive setup. As in the case of manual
application, the choice is between fixed sample size (sample loop) and variable sample size (syringe).
2.5. Columns
The chromatographic column is the centerpiece of every
HPLC unit. Its design and the type and quality of its packing are decisive for good separation and minimal dispersion of the sample components. In general, steel columns
with minimal dead-volume, fixed end-fittings are employed. In addition, economically priced cartridge systems
are available, in which ready-packed exchange cartridges
are compressed in a corresponding module. Radially compressible plastic cartridges offer the additional advantages
that wall effects[I5l are avoided and that it is possible to
repack the matrix after the formation of cracks.
Most of the commercially available analytical steel columns have a diameter of 4.0-4.6 mm and a length of 1030 cm. Larger columns are sometimes employed for gel filtration, and, in the case of adsorptive techniques, signifiAngcw. Chem. Inr. Ed. Engl. 25 (1986) 535-552
cantly smaller columns have been employed with SUCcess,lIo- I')l The resolution of a column (proportional to
Noi['I) is, i n fact, not very greatly dependent on its length.
However, from the special elution behavior of macromolecules (see Section 2.2) and simple trial separations, the
conclusion should not be drawn that long columns generally fail to offer advantages for the separation of proteins
using gradient
According to the gradient theory
of Snvder et al.,"' this is only true when short elution times
(i.e., steep gradients) are chosen. However, when difficult
separation problems require longer elution times (i.e., a
shallower gradient), then there is a drastic increase in the
influence of column length on the performance of the sysIf short columns can be employed, they have the
advantage of less back pressure (higher long-term stability)
and less dilution of the sample.
Theoretically, the column diameter has no influence o n
the separation performance of the column if the flow rate
is suitably adjusted.["I Because, in addition, the wall effect[Is1is of less significance for larger diameters, there is a
great potential for increasing the loading capacity for preparative work. Reducing the diameter leads to smaller
peak volumes and, hence, to lower detection limits;[211the
"micro-bore'' technique (column diameter < 1 mm) has
found only occasional application in the protein field so
farl'i.
241 and, like the use of short columns mentioned
above, requires increased care with respect to possible
peak broadening by the other system components. Suitable
prepacked columns are only available with packing materials of low pore diameter (ca. 100 A), which are less suitable
than packing materials with larger pore-size particles
(>300 A) for proteins with molecular weights of more
than 20000.[25-271
As far as the particle size of the packing is concerned, a
compromise must be reached between the highest possible
resolution, which improves as the particle size decreases,
and the long-term stability of the column packing. At present, the minimum particle diameter is 5 g m for the large
pore-size packings needed for the separation of high-molecular-weight biological
Silica gel largely fulfills the most important criteria for a
packing material : very small particle size, high mechanical
stability, and high porosity. However, it is only stable at
pH values between 2 and 7. Organic polymers are suitable
for use with basic buffers, but are not yet available with
similar small particle sizes and with equivalent pressure
stabilities. An easily exchangeable precolumn is highly recommended when working with complicated mixtures (such
as cell homogenates). Filtrations or irreversible adsorptions take place at this stage, thus protecting the main column.
Many protein separations can be performed at ambient
temperature. However, the temperature affects the resolution, retention, and recovery in all adsorptive tech-
the quality of the column packing, the performance of the
whole system, and, in particular, the recovery rate of the
proteins. Various model mixtures have been proposed for
this purpose in the literature.'""
2.6. Detectors
UV detection is by far the most commonly applied
method of detection in protein separation."21 The peptide
bond absorbs at h = 210-230 nm, the absorbance being independent of the amino acid
At 210 nm,
moreover, the secondary structure is without influence.134'
The sensitivity is 10- to 20-fold higher than at A=250280 nm, where only the aromatic amino acids absorb.
However, working at lower wavelengths restricts the
choice of solvent, places greater demands on its purity,
and requires the effective exclusion of oxygen. It is
scarcely possible to avoid a drifting baseline in gradient
elution. Programmable detectors (e.g., diode arrays) can
scan the whole available spectrum at various points
(usually the peak maxima) in a very short time. This can be
of particular value in the analysis of short peptides (e.g.,
from enzymatic or chemical degradation^)."^-^"
In particular cases, the intrinsic fluorescence can be employed with sufficient sensitivity for the identification of
peptides rich in aromatic amino
Fluorescent detection with on-line, post-column derivatization is of more
universal
With ~-phthalaldehyde[~"'
or fluorescamine,['" derivatives are formed that absorb at higher
wavelengths and thus allow unhindered detection in the
presence of strongly UV-absorbing buffers. For preparative work, only an aliquot is divefted for detection by
means of a switching valve (splitting technique). A cysteine-specific method has also been de~cribed.'~''
3. Principles of Separation
3.1. Gel Filtration
Gel filtration occupies a special place among chromatographic separation methods. The separation does not take
place on the basis of interaction between the sample, the
stationary phase, and the mobile phase, but depends on
"sorting" according to molecular size. In ideal gel filtration, proteins are separated on the basis of their distribution between mobile liquid phase and the liquid situated in
the pores of the matrix.["31The elution volume available for
separation thus lies between two limiting volumes: the volume of liquid between the particles ( V J and the total diffusion volume as the sum of V<,and the pore volume V,.
The peak capacity is, therefore, strictly limited. The elution
volume V, of a substance peak is given by
niques,'2x'7"l
Commercially available ready-packed columns, which
are delivered with a test chromatogram, are employed
most often. However, columns that are specially supplied
for protein chromatography are seldom tested by the manufacturers with appropriate sample mixtures under relevant elution conditions. In practice each user has to test
Anyrii . C'lienr. Irir. Ed. Enyl. 25 llY861 535-552
The distribution coefficient K r , is virtually 0 for very large
molecules. These cannot diffuse into the pores and are,
therefore, eluted most rapidly. For small molecules K , , is
1 ; they migrate unhindered into all the available pores and
are thus subjected to the greatest delay.
537
In general, a linear relationship between the distribution
coefficient and the logarithm of the molecular weight exists in the range Ko=0.15-0.80. Gel filtration can, therefore, be employed for the characterization of proteins
when interactions between the protein molecules and
the surfaces of the particles can be excluded and the protein is regular in form and not aggregated. Nonideal column behavior can be investigated by means of simple
te~ts.1~~1
The resolution of the column depends on its dimensions,
the density of the packing, the pore volume, the pore-size
distribution, and the peak width. Since the separation only
takes place within a limited elution volume, efforts must be
made to restrict the volumes of the individual peaks as
much as possible.
The peak width is associated with the molecular transport between the phases and reaches its optimum at small
particle sizes,14s1low flow rate and viscosity of the mobile
phase, and high diffusion coefficient for the protein. From
these relationships it can be understood that the limits for
sample volume and concentration are lower (ca. 1 0 0 ~ g
and 10 FL per mL of column volume) than is the case for
adsorption
All the other parameters mentioned are fixed properties
of the column packing. Spherical particles can be packed
significantly more densely than can irregular ones, so that
V,,-the column volume that is not available for separation-only makes u p 37% of the total column
The maximum volume available for separation is achieved
by increasing the pore volume V,. Here a compromise has
to be reached between optimal pore volume and high mechanical ~ t a b i l i t y . ~A~ ~narrow
, ~ ~ ' pore-size distribution is
advantageous for high resolution, but the packing process
itself evidently also plays an important role in determining
the efficiency of the c01umn.I~~~
Of the organic polymer particles developed for gel filtration, the TSK-PW series (Biorad) is finding increasing use;
the hydrophobic interactions that are unavoidable in all
materials of this type (see Section 3.3) are apparently minimized, thereby allowing very efficient protein separations,
particularly in the high-molecular-weight range.14'l
Silica gels exhibit better separation performance in narrow molecular-weight range^.^^^.^^] The surface may be
treated with hydrophilic o r g a n o ~ i l a n e s [ ~or~ organic
~ ~ ' ] polymer~'~''to block silanol groups and prevent adsorption
effects. However, the treatment is never complete for steric
reasons,1531so that these materials always exhibit residual
ion-exchanger properties. This can be compensated for by
the presence of salts in the eluent.[441Too high a salt concentration should be avoided, however, since it would increase hydrophobic interactions. A new surface treatment
variant that imparts greater pH stability to the support is
also worthy of mention.lS4'
Silica gels are also very stable in the presence of chaotropic salts and detergents. It is possible to employ guanidinium chloride, sodium dodecyl sulfate, urea, and organic
solvents to denature the proteins before separation,
thereby preventing self-association and increasing solubility.[5Z-581
However, this can greatly increase the hydrodynamic radius of the protein, which results in displacement
of the fractionation range to lower molecular weight. Fig538
ure 2 shows the gel filtration chromatograms of two protein standards with two different eluents.
0
17.5
t lmini
-
35
Fig. 2. Gel permeation chromatography of protein standards. HPLC system:
Bio Sil TSK, 7.5 mm x 60 cm (Bio Rad). 0.7 mL/min, ambient temperature,
Abs?so=absorption at J.=280 n m ; t=retention time. (A): 20 mM phosphate
buffer, 0. I M sodium sulfate, pH 6.8, peaks In the order of elution: thyroglobulin ( M W 670000). y-globulin ( M W 158000), ovalbumin ( M W 44000), myoglobin ( M W 17000), vitamin BIZ ( M W 1350). MW=rnolecular weight (B):
20°h acetic acid/30% acetonitrlle, pH 3.0, peaks in the order of elution: myoglobin ( M W 17000). 17.8 min: cytochrome c (MW 12500), 19.1 min: insulin
(MW 5850), 24.9 min; insulin 6-chain ( M W 3400). 25.5 min; phenylazo-iPro-Leu-Gly-Pro-o-Arg ( M W 777). 28.7 min; sodium azide (MW 69, 32.7
min). Z = benzyloxycdrbonyl.
3.2. Ion-Exchange Chromatography
Because of its amphoteric character, a protein can exist
in either cationic or anionic form depending on the pH. A
characteristic parameter of proteins is, therefore, the isoelectric point at which the net charge of the protein is zero.
The isoelectric point determines the pH of the medium and
the choice of the type of ion exchanger. The strength of
binding is determined by the number of cooperatively effective ionic groups on the protein, the charge density of
the ion exchanger (theoretical capacity), and the ionic
strength of the mobile phase. Thus, elution of the proteins
can be brought about by a gradient of increasing ionic
strength, a pH gradient, or a combination of both. The user
has a wide range of possibilities available for influencing
retention and selectivity.
Silica gel and its derivatives were the most commonly
used ion-exchange supports for a long time and were used
in the great majority of systematic investigations of ionexchange HPLC of proteins.[591 Organic methacrylatebased polymers, which are only available in relatively large
pore sizes and which are partially hydrophobic, did not
prove to be as useful.[601Recently, however, hydrophilic organic polymers, such as the above-mentioned TSK-PW series, have become available and provide good results in
analytical and preparative separations, at a somewhat
lower flow rate and
Their stability in the alkaline p H range allows a wide selection of buffer solutions
and more efficient regeneration methods, thereby increasing the working lives of the columns, even when crude extracts are separated.[621
Angew. Chem Int. Ed. Engl. 25 13986) 535-552
Silica gels are available in a wide range of pore sizes and
with a large variety of organic ligands. Strongly acidic and
basic exchangers are usually characterized by higher resolution and detection sensitivity, because the better desorption leads to smaller peak volumes. The range of variation
of the pH of the mobile phase does not affect the degree of
ionization of the exchanger, so that the optimization problem is less ~ o m p l e x . A
~ particularly
~ ~ ~ " ~ ~ stable and efficient
masking of nonspecific binding sites on the silica gel surface and a high ion-exchange capacity is achieved when
the ionic ligands are cross-linked within a thin polymer
film .167-7111 The theoretical capacity in this case is proportional to the total surface area of the particles and, hence,
highest for small pore-size material. However, ion-exchangers with wide pores are more accessible to high-molecular-weight proteins and, hence, exhibit a higher effective
capacity. The optimization of peak sharpness also plays a
crucial role, because diffusion is as unhindered as possible. The best choice for most protein separation problems
is a pore size of 300 A, whereas for molecules with a molecular weight larger than 150000 a pore size of 1000 A or
more is preferred."6.27,"61
It is usual to elute by means of a salt gradient at a constant pH. Since the displacing power of a range of ions can
be completely different for different proteins,[291the exchange of one displacing ion for another can result in different selectivities. Thus, an astute choice of the salt type
can optimize the distribution of all sample components
and the resolution of early eluting components as well as
improve the recovery of particularly strongly adsorbed
The retention time of a protein increases as the difference between its isoelectric point and the p H of the buffer
increases.'"'] The selectivity can also be influenced by varying this parameter, since the shape of the titration curve is
characteristic for each protein. A good starting point is
often pH 7, which leads, on the average, to the best selectivity in anion- and cation-exchange chr~matography.''~]
A
different pH can be more advantageous for the separation
of a particular pair of proteins. This can often only be determined empirically and cannot necessarily be deduced
from the titration curves of the protein^.'^]
Other conditions are easier to optimize than the composition of the mobile phase. The total volume of the salt gradient is generally between 5 and 15 column volumes. A reduction of the flow rate and of the steepness of the gradient improve the resolution ~ o n s i d e r a b l y . ~ ' ~ .An
' ~ ] increase in the temperature also has a positive
The
load that present-day columns can accept is high; 2-3 mg
of protein per mL of packing volume can easily be separated without loss of r e s o l ~ t i o n . ~ Apart
' ~ . ~ ~from
]
a few exceptions, the recovery rates are greater than 90%.[29.7'l
3.3. Hydrophobic-Interaction Chromatography
Hydrophobic interactions play an important role in the
folding of proteins and in the stabilization of their tertiary
The nonpolar amino-acid side chains of soluble proteins are largely shielded inside the three-dimensional structure. Nonetheless, proteins possess a few hydrophobic points of contact on the surface, which lead to
A n y e t Clirm. Inr
Ed Enyl. 25 11986) 535-552
adsorption to a nonpolar stationary phase. This effect was
at first observed as abnormal retention behavior in conventional affinity chromatography and then later deliberately
exploited for the separation of proteins on aikylated agarose gels.173-751H ydrophobic adsorption of proteins on
HPLC gel filtration columns has been observed under
certain ~ o n d i t i o n s . ' ~ " Special
. ~ ~ ~ phases based on silica
ge1[77-8010r
organic polymerparticles'8'.sZ1werelaterdeveloped
for performing this type of separation by HPLC.
The hydrophobic interaction is strong in the presence of
high concentrations of i ~ n ~ .The
[ ~type
~ ,and
~ concentra~ ~
tion of the salt are decisive for the retention, which is
greater for salts exerting a larger salting-out effect on proteinS.f79.X2-841Th e strength of the hydrophobic interaction
on the other
increases with increasing
hand, the retention is reduced by the addition of small
quantities of organic solvent or chaotropic reagents (urea,
guanidinium c h l ~ r i d e ) . [ ~Theoretically,
~~~*~
the bonding
ought to be strongest at the isoelectric
in practice,
however, the p H can affect the retention of different proteins, depending on the matrix, to differing degrees and in
different ways, which is probably attributable to the effect
of charges on the stationary phase.[8' x31 The selectivity of
the separation can also be strongly affected by changing
the temperature and by addition of organic solvents to the
eluent,~30. 82.831
The type of salt is also important for the resolution.
Neutral salts with a high salting-out effect are preferred
for hydrophilic (weakly binding) proteins, whereas those
with a low salting-out effect are preferable for hydrophobic (strongly binding) proteins.'7xi The resolution can
also be improved by longer columns,'22.X21lower flow
and shallower gradients (longer gradient elution
and by the addition of chaotropic agents or organic solvents.[30.X21
Lower temperatures are also advantag e o ~ s , ' ~Denaturation
",~~~
caused by matrix contact is rare
in this
The sample load that the columns can
carry and the recovery rate are about the same as those in
ion-exchange chromatography.
3.4. Reversed-Phase Chromatography
Reversed-phase (RP) chromatography was introduced as
early as 19501s71
and has become the preferred method of
HPLC because of its wide scope of application.LXX1
It is basically a variant of hydrophobic chromatography (see Section 3.3). In both cases binding takes place between a nonpolar surface of the protein and the stationary phase,
thereby shielding the stationary phase from the polar
(aqueous) mobile p h a ~ e . [ * ~In- ~contrast
'~
to hydrophobic
chromatography, significantly more nonpolar, i.e., more
heavily alkylated, stationary phases, are employed here, so
that the proteins bind to the matrix even in aqueous buffers of low ionic strengths. Differential desorption is
achieved by applying a gradient with an increasing concentration of organic solvent. On the one hand, this lowers
the surface tension of the mobile phase and favors the formation of a solvent shell around the molecule that has
been d e ~ o r b e d ; ~ " - ~on
' ] the other hand, it favors desorption by means of direct competitive interaction with the
protein and the matrix. Several theoretical models are un539
"'1
der discussion for the quantitative description of the comalkyl ligand for the same density of
plex thermodynamic and kinetic phenomena involved in
The degree of alkylation, on the other hand, is critical for
the interaction between stationary phase, eluent compothe resolving power and the gentle elution of proteins,[Y2.93.
119-1221
nents, and sample molecules and particularly the relationtde-pore particles have the advantage
ship between retention and solvent ~ ~ m p ~ ~ i t i o n . [ ~ ~ that,
~ ' ~because
~
of the limited surface area, an optimal coatMany binding sites can participate in the interaction of a
ing density (ca. 3-4 pmol alkyl residue/m2) is obtained
protein with the
17.951 H ence: adsorption effects
even at a low carbon content (2-4% wt-% C).f120.1211
Th'IS
predominate rather than a distribution between stationary
allows the elution of proteins at low concentrations of orand mobile
During gradient elution, the relaganic s o I ~ e n t . ' ~l2'1 ~ .0~ctyl
~ , or octadecyl ligands lead to
tively more weakly binding solvent molecules first occupy
better masking of residual silanol groups than d o ultrafree contact sites on the matrix and the protein surfaces
short alkyl residues ( <C4).[12"lIt is recommended that silanot involved in binding. When a critical concentration is
nophilic interactions be further suppressed by treating the
10.')4.971 the protein is suddenly displaced from
alkylated silica gel with reactive silanes of small size
the matrix surface. This concept of a competitive interac("end-~apping").'~~.
""I
tion between matrix, protein, and solvent molecules is the
Electrostatic interactions with free silanol groups lead to
basis of a newer stoichiometric displacement model'x1that
very asymmetrical peaks or even to irreversible adsorption
explains very well the observed influences of molecular
of the proteins to the RP matrix. Since the dissociation of
size and solvent composition and is in agreement with the
the silanol groups is largely suppressed at low pH ( < pH
general theory of gradient elution.'41
4),[12'] the best results are generally obtained in this reThe retention of a sample molecule depends on the ratio
gion,[26.93.103, 116. 1241 Th us, satisfactory protein separations
I 16.
of the polar to nonpolar amino-acid side chains on the part
are possible even in dilute hydrochloric
of the surface that is involved in binding to the stationary
Various buffer substances, however, are able to change the
phase. In the case of smaller polypeptides ( < 2 0 amino
retention characteristics of proteins owing to ion-pair foracids), it is possible to calculate the retention time accumation.[92.93. I 1 I. 1251 E xamples include phosphoric acid,
rately in advance from the sum of the relative hydrophobiperchloric acid, perfluorinated carboxylic acids, and alkyl
cities of the individual amino a ~ i d s . [ ~ ~ -However,
'"'~
presulfonic acids and their salts. Phosphoric acid decreases[lz6I
dictions are difficult for some proteins on account of the
and perfluorinated carboxylic acids increase the retention
tertiary structure. Even in the case of B30 analogues of intime.[9', 1271 Cationic components in the eluent affect the
sulin (see Section 4.1 and Fig. 5), the incorporation of parmatrix properties directly. Ammonium, alkylammonium,
ticular amino acids leads to a change in the retention time
pyridinium, and morpholinium ions deactivate silanol
from the predicted value. In addition, under the usual elugroups by hydrogen bonding and electrostatic interact i o n ~ . [I Z~x 1 ~High
.
concentrations ( > 0. I M ) of chaotropic
tion conditions for RP-HPLC, possible conformational
changes must also be taken into account,['"'] which depend
salts (e.g., NaCIO,, Na2S0,) lead to increased peak sharpness and improved elution, particularly on account of the
on the composition of the eluent and other parameters. In
electrostatic shielding of the silanol groups.[100.1 2 9 - 1 3 1 1
extreme cases they can lead to multiple peaks (native and
denatured forms).i1"6-'091
Dilute trifluoroacetic acid has found the greatest application of al 1 buffer systems,[ 17- 19.83.95. 96.132- 1371 whereas
The behavior of a silica-gel-based matrix is not just influenced by the nonpolar ligands but also, to a consideralonger-chain perfluorinated carboxylic acids have been
employed much more ~ e l d o m l y . ' "lo4.
~ ~' 3 3 , l3*] Trifluoroble extent, by the free silanol groups. This leads to the folacetic acid is UV-transparent and easily evaporated and,
lowing phenomenon: with increasing organic solvent conhence, is also suitable for preparative work. It is particucentration, the retention of a given protein first decreases,
larly recommended, as is dilute formic acid,["'] for the sepas expected for nonpolar interactions, but then increases
aration of hydrophobic proteins or protein fragagain owing to the silanophilic cor~tribution.~~'~
Ionic comm e n t ~ . [ " ~ .On
' ~ ~the
] other hand, the elution of the strucponents in the eluent suppress these interactions, leading
tural proteins of polio virus was not possible with any of
to the formation of sharper peaks with higher resolution
these systems, but only with highly concentrated formic
and, in the case of certain ion-pair reagents, allowing
acid.[14o1
Pyridine acetate or pyridine formate are also volamilder elution conditions. Only when these factors were
tile buffers, which have been employed with great success
included in the optimization of the separation conditions
in the pH 3-5 range, that is, for work with acid-labile prowas it possible to overcome the initial difficulties111n1
exteins.1141-1451A d.isadvantage is that their low UV-transparperienced in the RP-HPLC of proteins.
ency necessitates detection at A = 280 nm or post-column
The HPLC user today finds himself confronted with a
derivatization (see Section 2.6). Eluents with even higher
bewildering array of stationary phases, in which silica gel
pH values include ammonium hydrogen carbonate (pH
derivatives predominate. Two wide-pore materials based
7.8)[l4'] and hexafluoroacetone/ammonia (pH 7.2)."481
on organic polymers have recently appeared on the marThe most frequently reported nonvolatile buffer is phosket.["'.'''] Their advantage lies in the possibility of effecI I 1491very often in the form of the tri- or
phate,[28.98.In'.
tive regeneration by washing with dilute sodium hydroxide
tetraalkylammonium salts[" I.
and with the addition
solution. Stationary phases with a pore size of ca. 300A
of sodium perchlorate.[''.
129-1'11 U sually the separations
are best suited for proteins with molecular weights greater
are performed below pH 4, more rarely at higher pH.11541
than 20000 with respect to loading capacity, resolution,
104. 114-1 161 Th
and recovery.[25.26.96.
Separations have also been reported with Tris buffer at pH
e retention of most pro7.5."5s1
teins is essentially independent of the chain length of the
*.
540
Angea,. Chem. Int. Ed. Engl 25 (1986) 535-552
The organic component must be sufficiently nonpolar to
effect desorption of the protein from the matrix; on the
other hand, it must be polar enough that the protein does
not precipitate. This virtually restricts the choice to acetonitrile and alcohols. Dioxane has been used in the purification of interferon."s61 The eluting capacity increases in
the order methanol, ethanol, acetonitrile, propanol, butanol.["] Changing or combining the solvents allows differSmall addient selectivities to be ~ b t a i n e d . ' ~ * .14''. ~ '1s7-1601
.
tions of methoxymethanol can improve the resolution."". l4')]Propanol is recommended for larger, hydrophobic proteins. Since the concentration of the organic
component in the eluate can be kept low, fewer problems
are to be expected with precipitation and denaturation.[I2'. ''I1 Occasionally, however, peak broadening is observed,"'"'. I '2.1481 which can be reduced by raising the temperature, provided that the proteins are sufficiently stable.'281 In general, both improvements[98' and deteriorations"O'.
II6.12'1
. the resolution have been observed on
in
raising the temperature. The resolution improves for reduced flow rates and shallower
The elution conditions of RP-HPLC not only reduce the
solubility of many proteins, but also can result in conformational changes.["" In the case of labile proteins this can
lead to complete denaturation with irreversible adsorption
or precipitation. This is not just a preparative problem, it
can also bring into question the accuracy of a quantitative
analysis. Examples are available for the successful isolation of functionally intact proteins by RP-HPLC in the
presence of stabilizing cofactors or by special treatment of
the fractions ~ b t a i n e d . " ~ 161-1631
~ - ' ~ ~In. some cases, furthermore, certain proteins are eluted with low yield or are initially inactivated."21Recent investigations[86. 122.164, lb51 of
the mechanism of this process have revealed that, depending on the protein, denaturation processes can occur at
various rates during column contact; the degree of denaturation depends on how long the protein is in contact with
both the matrix and the solvent. Desorption initiates a refolding to the natural tertiary structure. The kinds of peaks
obtained are determined by the relationship between the
kinetics of unfolding and refolding and the separation kinetics: a) individual peaks of narrow width are observed
when the protein is either especially stable (no denaturation) or the refolding is greatly inhibited (only denaturation product); b) mult~ple['06-'09~''2.1651
or extremely asymmetrical[?H.
112. 121. 122. 164. 1651
peaks are observed when native
and (several) denatured forms are present in equilibrium.
Strongly eluotropic solvents, hydrophilic ion-pair reagents,
reduced concentrations of the perfluorinated carboxylic
acids, or steeper gradients inhibit the denaturation on the
stationary phase.""', ' I 2 ] This leads to higher recovery rates
and a high degree of retention of the native biologically
active form,l?8.IIZ,I16.1661 High salt concentrations, on the
other hand, can have a negative effect[28."61since they intensify the interaction with the matrix. High flow rates and
temperatures d o not allow sufficient refolding to occur, so
that, in the extreme case, only the denatured form is
eluted''h4.1651 or irreversible binding takes place (memory
effect)." i2.1651 In those cases high resolutions are obtained,L'x'but the recovery is sometimes small."
RP-HPLC is unquestionably one of the most efficient
Anyew Chem. lnr. Ed. Enql. 25 (1986) 535-552
methods of separation. Its greatest potential lies in the separation of closely related proteins and protein derivatives,
since selectivity and resolution can be manipulated in
many different ways. However, the protein-chemical consequences of the elution conditions must be reevaluated in
each case.
3.5. Other Techniques
Adsorption to hydroxylapatite is a central component of
one system for the purification of monoclonal antibodi e ~ ; " however,
~~'
it can also be employed for other separation problems.[1681 Affinity ~ h r o m a t o g r a p h y " ~involves
~~
biospecific selection from complex mixtures and, thus,
places no great demands on elution techniques. The HPLC
version of this method, therefore, emphasizes high speed
rather than high resolution. Stationary phases are prepared
by the user himself by coupling the appropriate ligand to
g I a s ~ , [ 'silica
~ ~ ] geI,[l7I.1721 or organic poIymer~.["~]
Some ligands have a variety of applications; these include adenosine m o n ~ p h o s p h a t e , " protein
~~~
le~tins!"~~and the
pseudo-affinity ligand Cibacron Blue.['751Ready-to-use columns can be envisaged here. In view of the wide variety of
other specific ligands, the use of self-made columns is inevitable. However, the process can be simplified by the use
of a pre-packed, pre-activated column, which is then
treated by allowing a solution containing the ligand to
flow slowly through the
Chromatofocusing
can be carried out with commercially available ion exchangers and buffer solution^.^'^^. 17x1 Like its electrophoretic
counterpart, the separation depends strictly on the isoelectric point.
4. Insulin
The blood-sugar-lowering hormone insulin is a small
protein. It has a molecular weight of 5800 and consists of
two peptide chains: the A chain, with 21 amino acid residues, and the B chain, with 30. The two peptide chains are
cross-linked by two disulfide bridges. A third disulfide
bridge links the cysteine residues at positions 6 and I I
within the A chain."79- 1801
Insulin has a relatively rigid structure[lx']and therefore
exhibits relatively high stability. Consequently, it poses a
simple problem for HPLC analysis,L182-'861
and insulins
separated by HPLC generally remain biologically and immunologically a ~ t i v e . [ ' ~ ' - M
~ 'oreover,
~~
insulin is of eminent importance in pharmaceutical research. For these reasons, insulin is ideal for demonstrating the scope of application of HPLC of proteins.
In this connection, it is important for preparative and
analytical work that insulin is insoluble near to its isoelectric point at p H 5.4 and dissociates at acidic pH, whereas it
forms dimers and hexamers at neutral pH. The hexamer is
stabilized by zinc. Separations above the isoelectric point
are only successful with zinc-chelating and dissociating additives such as ethylenediaminetetraacetate (EDTA) and
alcohols,"901 urea,["ll or detergents.['"]
Insulin is nowadays produced by two important methods, isolation from animal pancreas"931and genetic engi541
neering. The insulin from the pancreases of three pigs satisfies the requirements of one diabetic for about ten days.
Insulin with human sequence can also be produced by
semisynthesis from porcine insulin. In contrast to this, the
production of human insulin by genetically altered microorganisms, whose carbon source is glycerin or simple sugars, is virtually ~ n l i m i t e d . " ~ ~ - ' ~ ' ~
4.1. Semisynthesis of Human Insulin
Human insulin only differs from porcine insulin in the
amino acid at position 30 of the B chain. Since lysine occurs at position B29, replacement of alanine (porcine insulin) by threonine (human insulin) is possible using the
enzyme trypsin (Fig. 3). Although, in an aqueous medium
at a neutral pH, trypsin catalyzes the cleavage of peptide
bonds at basic amino acids, at a lower pH and in a suitable
1
x,
1
7 :i-(ny%X,,
A-chain
iy
, kY,.l>
porcine insulin
I
I
22
2930
Arg
Lys-Ala
2
r%rxixxxxxxuirx*.tw~>
8-chain
?iBu
HC-CH3
-
so
Fig. 4. HPLC chromatogram of the crude reaction mixture obtained in the
semisynthesis of human insulin. HPLC system: Rad Pak C8, IOpm,
8 x 10 mm (Waters), 50 mM tetraethylammonium phosphate, 0.25 M sodium
perchlorate, 10% acetonitrile, pH =3.0 (buffer A), and 50 mM tetraethylammonium phosphate, 90% acetonitrile, pH = 3.0 (buffer B), gradient from 28%
to 40% B, 1.5 mL/min, ambient temperature, detection at A=ZlOnm; [ = r e tention time. 1-111: see text.
401
f .
timml
trypsin
Thr ltBu1 O t B u
29
25
t Lminl
0
human insulin ester
Lys-NH-CH-C-OtBu
br
10
14
00
02
04
06
08
l g l r e l hydrophobicity)
Fig. 3. Semisynthesis of human insulin from porcine insulin
medium, it can also catalyze the formation of peptide
bonds"981 or even directly catalyze the replacement of an
amino acid without the formation of an intermediate hydrolysis p r o d u ~ t . [ ' ~ ~ -This
' ~ ' l replacement is formally equivalent to aminolysis of the acyl-enzyme intermediate.
HPLC analysis of reaction mixtures suggests a direct exchange mechanism in human insulin semisynthesis (Fig.
4): After reaction with threonine tert-butyl ester, the O H
group of which is protected by a tBu group, des-(B30) insulin is not detectable; only the desired side-chain-etherified human insulin ester (III), nonconverted porcine insulin
(II), and des-(B23-B30) insulin (I) are in evidence. The latter is an unwanted by-product originating from proteolysis
at arginine (B22) and, for steric reasons, is apparently only
formed in small amounts.[2n2'
A series of B30 derivatives, whose behavior in reversedphase HPLC corresponds exactly with the change in hydrophobicity"OO1of the C-terminal region of the B chain
(Fig. 5 ) , have been produced by semisynthesis.[20'1However, two derivatives-thr(B30) insulin and tyr(B30) insulin-for which local structural changes (e.g., the pres542
-
10
12
1L
Fig. 5. Retention tune of various B30 deri\,atives of insulin as a function of
relative hydrophobicity of the amino acid at position B30 in the following
RP-HPLC systems: TFA: Nucleosil 300-7 protein-RP, 4 . 6 ~125 m m (Matrifluoroacetic acid, pH = 2. I,gradient from 30% to 40%
cherey-Nagel), 0. I%
acetonitrile, 0.8 rnL/min, ambient temperature. TEAP: Rad Pak C18, 5 pm,
8 x 100 mm (Waters), composition of buffers and other conditions as for Figure 4.
ence of an additional hydrogen bond) can be postulated,
d o not follow the linear relationship.
4.2. Production of Human Insulin by Gene Technology
Except for the final purification steps, the procedure
used to produce human insulin from microorganisms by
gene technology differs fundamentally from that used to
isolate it from pancreas. The insulin from pancreas is present in the biologically active, native spatial structure in all
isolation steps. The preliminary stages in gene technology,
on the other hand, involve inactive and generally poorly
soluble precursors, which, for instance in the E . coli cell,
are deposited in large quantity in the form of inclusion
bodies (Fig. 6). These precursors have to be converted into
the native structure by sophisticated protein chemistrysupported by appropriate chemical analysis (Fig. 6). The
proinsulin concept presented here exploits the natural
folding tendency of the physiological insulin precurs0r,1203.2n41
in which the A and B chain are connected by a
Angew. Chem. In(. Ed. Engl. 25 (1986) 535-552
m
E Coli
Enrichment
Fragmentation
Derivatization
Purification
proinsulinS-sulfonate
Folding
0
31.5
63
t [miniDroinsulin
Fig. 7. HPLC chromatogram of the reaction mixtures obtained in the preparation of human insulin by gene technology: after chemical fragmentation Of
the insulin precursor (A) and after further modification by oxidative sulfitolysis and fractional precipitation (8). HPLC system: PLRP-S 300 A,
4.6 x 250 mm (Latek), 0.1 M hexafluoroacetone/NH,, 2% butanol, 8% acetonitrile, pH =7.0 (buffer A), 0.1 M hexafluoroacetone/NH3, 9% butanol, 81%
acetonitrile, p H s 7 . 0 (buffer B), gradient from 20% to 35% buffer B i n 60
min, 0.8 mL/min, 40°C. 1-111: see text.
Purification
insulin-Arg2
Purification
u
insulin
Fig. 6. Left: Principal steps in the production of human insulin from genetically engineered microorganisms. Right: Electron micrographs of E. coli cells
(above) producing the insulin precursor and of the rhombohedra1 insulin
crystals (zinc-stabilized hexamer) obtained by this process (below).
peptide segment, the C peptide. The alternative concept,
namely, separate production of the A and B chains, results,
on recombination of these two chains, in a low yield and
numerous by-products. Nevertheless, this was historically
the first method for producing human insulin by gene
technology and was also successfully used on a technical
~caIe.[”~J
For the isolation of insulin, the following procedure
can be used: the microorganisms are harvested following
fermentation, and the insulin precursor is enriched. The
precursor is then cleaved chemically at methionine residues, thereby resulting in the formation of “preproinsu1in”-a proinsulin with an N-terminal extension consisting
of several amino acids. The six cysteine residues of preproinsulin, which are initially present in a form not precisely defined (E. coli lacks an enzyme system that specifically links the disulfides), are converted by sulfitolysis[20s1
to S-sulfonates. At this stage, the product is detectable by
HPLC for the first time. In the reaction mixtures available
prior to this, there are either too many by-products or the
molecules of interest become irreversibly adsorbed on the
reversed-phase support. This clearly demonstrates one of
the limitations of HPLC analysis (Fig. 7A).
The preproinsulin S-sulfonate is a markedly acidic molecule. For reasons of solubility, the acid-mobile phases
normally used in reversed-phase HPLC are therefore not
feasible. A neutral hexafluoroacetone system.”48’ in which
the preproinsulin S-sulfonate (I) can be analyzed separately even from a precursor fragment (11) that is chemiAngew Chem. Int. Ed. Engl.
2s (1986) 535-552
cally very similar, is used instead (Fig. 7B). In contrast to
fragments I and 11, fragment I11 does not contain a cysteine residue and therefore elutes as a sharp peak without
prior derivatization (Fig. 7A).
The preproinsulin S-sulfonate is not only an interrnediate with a key role for purification and chemical analysis; the naturally folded preproinsulin is also available
from it in high yield^.^^"^^^'^ The folding and correct linking of the two interchain disulfide bridges can be examined by fingerprint analysis. The following requirements
must be met:
- the formation of a limited number of fragments of in-
terest
- a minimum of nonspecific reactions
- the identification of individual fragments
-
a rapid and efficient method of separation and detection
Proteolytic fragmentation by Staphylococcus aureus protease V8 under special conditions in connection with
HPLC analysis fulfills these requirements.[z08’S . aureus
protease V8 cleaves C-terminally at glutamate; the theoretical-and
also experimentally verified-fragmentation
pattern is shown in Figure 8. Fragments I, 11, and 111 are of
particular importance for the analysis. Fragment 111 contains the C-terminal region of the B chain and is present
irrespective of whether preproinsulin S-sulfonate, disulfide-bridged preproinsulin or other species of the same sequence are present. It serves as internal standard for quantification purposes. In contrast, fragments I and I1 are only
found in full amount when their disulfide bridges are present to the full extent. By-products, such as incorrectly disulfide-bridged species, would produce new peaks at the
expense of peaks I and 11. Fingerprint analysis of preproinsulin S-sulfonate does not produce either fragment I
or 11. Preproinsulin S-sulfonate, reduced with an excess of
mercaptan, shows correspondingly low proportions of
543
VI
Fig. 8. Peptide fragments of preproinsulin formed by proteolysis with S. aureus protease V8. The prefix pre indicates that the proinsulin
is lengthened at the N-terminal end by five amino acids (white circles).
both fragments. For comparison, an S . aureus protease V8
fingerprint of correctly folded preproinsulin is shown in
Figure 9.
HPLC analysis is also essential in the next step, namely,
in following the enzymatic processing of preproinsulin to
insulin-Arg, and insulin-Arg (Fig. 10). Proteolysis of preproinsulin with trypsin occurs mainly at six arginine o r lysine residues (see Fig. 8). However, cleavage at arginine(A0) at the N-terminal end of the A chain, at arginine(B0) at the N-terminal end of the B chain, and at arginine( B32) are preferred kinetically. Cleavage at lysine(B29)
Fig. 9. HPLC analysis (fingerprints) of S. aureus protease V8 digests of (A)
preproinsulin S-sulfonate, (B) preproinsulin reduced with excess thiol, and
( C ) correctly folded preproinsulin. HPLC system: Rad Pak C 18, 5 pm,
8 x 100 mrn (Waters), buffers as in Figure 4, gradient from 20% to 45% buffer
B in 25 min. 1-111 correspond to the fragments in Figure 8.
544
yields des-Thr(B30) insulin in small amount, while cleavage at arginine(B22) to give des-( B23-30) insulin-as
found in semisynthesis-virtually does not occur, apparently for steric reasons.
0
tlminl
-
- 1
50
Fig. 10. HPLC‘ chromatogram of samples taken at various times alter incubation of preproinsulin with trypsin. HPLC system: Rad Pak C 18, Spm,
8 x 100 rnm (Waters), buffers, gradients, and other conditions as in Figure
4.
Angew. Chem. Int. Ed. Engl. 25 (1986) 535-552
(A)
Separation of des-Thr(B30) insulin, human insulin-Arg,,
and human insulin by reversed-phase HPLC is dependent
to a large extent on whether perchlorate is present in the
eluent. When a triethylammonium phosphate system without perchlorate is used as the mobile phase (Fig. ll), the
sequence of the elution of these species is the reverse of
that obtained when a tetraethylammonium phosphate system containing perchlorate is used (Fig. 10). However,
HPLC analysis of the product mixture obtained by cleavage of purified human insulin-Arg, with carboxypeptidase
B, which was performed using the perchlorate-free triethylammonium system (Fig. 1 I), affords distinctly less
sharp peaks than analysis with the system containing
perchlorate (see also Section 3.4).
87.97
99.0% fl
I
11
u'
*;
;starting material
40
t lminl
-0
N NaCl
I
;@lo
30
t
t-tg. I I HPLC' chromatogram of samples taken at various times after incubation of human insulin-Arg, with carboxypeptidase B. HPLC system: Rad Pak
C 18, 6 pm: buffer A: 0.25 M triethylammoniumphosphate, 2Ooh acetonitrile,
pH = 3.0: buffer B: 0.25 M triethylammonium phosphate, 50% acetonitrile,
pH=3.0, gradient from 19% to 32% buffer B in 40 min, 1.5 mL/min, 35°C.
4.3. Purity Assessment and Quality Control
In order to ensure optimum tolerance and minimum immunogenicity, all insulins used in therapy nowadays are of
high purity. In 1925, a few years after the discovery of insulin, a preparation containing 5 units per mg was used
therapeutically; in 1935, the insulin content was about 20
units per mg. The highly purified insulin available at present has 28 units per mg. Foreign proteins in the end product are consistently in the ppm range o r less and are only
detectable by radioimmunoassay. The same applies to
proinsulin, which is very efficiently separated by gel and
ion-exchange chromatography. Ion-exchange chromatography (Fig. 12) yields a product that, moreover, is largely
free of insulin derivatives formed during production and
storage. For instance, under the acidic conditions of ethanol extraction, the acid amide groups of asparagines-in
particular those of Asn(A21)-are hydrolyzed to aspartate,
the desamido i n ~ u l i n sor~ ~their
~ ~ ~ethyl esters being
formed.
HPLC permits identification of the individual products
and the isolation of high-purity fractions (Fig. 12). It may
likewise be used for quality control of the pure substance
(e.g., with respect to species identity) or of the finished
pharmaceutical as well as for checking the stability of insuAngeu Chem. Ini. Ed. Engl. 25 (1986) 535-552
0
l o
50
40
f[hl-
Fig. 12. Ion-exchange chromatography of crude, semisynthetic human insulin ( 8 ) and HPLC analysis of pooled fractions (A). The conditions of the
ion-exchange purification on a diethyIaminoethyl(DEAE)-cellulose matrix
with detergent-containing eluents and a sodium chloride gradient are described elsewhere 11921. HPLC system: see Figure 4.
lins to storage at different temperatures and over several
years.
Gel-permeation HPLC is nowadays also used routinely
for quality control and stability tests (Fig. 13); by this
method, high-molecular-weight insulin derivatives, such as
dimers originating as a result of transamidation, can be detected. For clinical therapy only highly purified insulins
are used-for instance, they must not contain more than
10 ppm of proinsulin or more than 1% of high-molecularweight derivatives.""'
Detection of the absence of porcine insulin in semisynthetic human insulin poses a somewhat complicated analytical problem. Although it is possible to separate human
and porcine insulin by HPLC, desamido(A21) human insulin, which can be a companion substance, coelutes with
porcine insulin. Detection can be successfully carried out,
however, at the preliminary stage of the human insulin es-
0 35%
19 83
0
4
8
12
16
t lminl
-
20
24
28
32
Fig. 13. Gel permeation HPLC of human inbulin-Arg2obtained by gene technology. HPLC system: Bio Sil TSK 125, 7.5 x 600 m m (Bio Rad), 20°/0 acetic
acid, 30% acetonitrile, pH = 3.0, 0.7 mL/min, ambient temperature, detection
at A = 280 nm, sample load 200 Gg.
545
ter,[i"'l or by the S . aureus protease V8 fingerprint method.
If porcine insulin is present in the human insulin, fragment
111 (Fig. 8) will exhibit a small after-peak. On the other
hand, if desamido(A21) human insulin is present, this after-peak will occur at fragment I1 (Fig. 14).
I1 111
I
I
,
0
10
20
right). Such infusion systems, in miniaturized form, are
now in clinical trial. Externally worn devices, by means of
which insulin is supplied subcutaneously via a catheter,
are commercially
while implants are still at
the development
With implants, insulin is delivered into the abdominal cavity or vascular system, the reservoir being refilled from time to time through the skin.
Early in the development of these systems, however, a
severe problem emerged: insulin is not stable in neutral solution under the conditions found in dosage systems. Denatured, biologically inactive insulin usually precipitates
within a few day~.[~'~.'''1
Acidic insulin solutions (pH = 3) are comparatively stable physically (i.e., against denaturation). Under the conditions in implants, however, a number of modified products
form within a refill cycle of four weeks; consequently, native insulin is ultimately only present in traces (Fig. 16).
The stability of insulin in acidic solutions in dosage devices, as shown by HPLC analysis, is accordingly insufficient for human
I
t [min] +
Fig. 14. HPLC chromatogram of S . aureus protease V8 digests of mixtures of
(A) human insulin contaminated with desamido(A21) insulin and (B) human
insulin contaminated with porcine insulin. HPLC system: see Figure 9.
4.4. Stability of Insulin in Dosage Devices
The pharmacokinetics of insulin in healthy individuals
is extremely complex. The pancreas excretes a small
amount of insulin as basal supply and larger amounts after
glucose intake. Although the blood sugar then increases
slightly, it settles down very rapidly at approximately
80 mg/dL. Substitution therapy with insulin administered
subcutaneously can only very inadequately imitate these
kinetics. Nowadays, a relatively constant blood-sugar level
of 150-200 mg/dL is considered to be ideal. Such a profile (Fig. 15) is only possible when insulin preparation
and food intake are well adapted to each other. Such
high blood-sugar values, however, are considered at least
partly responsible for various late complications of diabeteS.[21i-2131
The normal excretion of insulin in healthy individuals
can be simulated by means of an infusion system (Fig. 15,
Fig. 15. Comparison of average blood-sugar levels of twelve type I diabetics
under conventional therapy with Depot-H-Insulin Hoechstm (left, Prof. Skrabalo, Zagreb 1982) and of a different group of patients under therapy with
external infusion devices (right, adapted from Rizza el al. [223]). The range of
blood-sugar levels in healthy subjects is shown as shaded area.
546
0
25
50
tlminl-
Fig. 16. HPLC chromatograms of acid porcine insulin (pH = 3.0) at refill and
after passage through an implantable dosing device (240 FL/day) with 2 and
3 weeks of constant shaking at 37°C. The arrows indicate native (retention
time 24 min) and desamido(A21) insulin (retention time 25 min). The latter
is present in small amounts at / = O (arrow) and increases initially due to the
acidic medium: native insulin decreases continuously, with the formation of
several new species different from desamido(A21) insulin. HPLC system:
see Figure 4.
Adsorption can be regarded as the initial step in surface
denaturation of
Adsorption of insulin to hydrophobic interfaces can be prevented in neutral medium
by the addition of a small amount,of a surface-active substance.[' 19, 2201 Such an insulin solution is comparatively
stable not only physically but also chemically, on account
of the neutral pH. HPLC analyses have shown that insulin
delivered through a pump at 37°C undergoes hardly any
more changes than would be the case under conditions of
elevated temperature alone (Fig. 17). A specific derivative
eluting in HPLC with a retention time of 1.25 relative to
insulin is thereby formed and can be isolated by preparative HPLC in a butanol/ethanol/water system.[2081It is a
monomeric insulin, which is modified in the N-terminal region of the B chain, possibly at B3I2*l1and which has full
biological activity and an immunogenicity as low as that of
human insulin.[*'']
Angew. Chern. Inr. Ed. Engl. 25 (1986) 535-552
/
0
/
20
t lminl-
40
I
20
0
Fig. 17. HPLC chromdlograma of’J stabilized, neutral human insulin preparation (Hoe 21 PH) tested under identical conditions as described in Figure
16.
5. Further Applications of Analytical and
Semipreparative HPLC in Protein Chemistry
At the present state of development, all conventional column-chromatographical methods are transferable to the
HPLC version. Column materials and pumps also tolerate
buffer systems such as those used in gel filtration, ionexchange (Fig. 18), hydrophobic-interaction (Fig. 19), or
affinity chromatography (Fig. 20). The resolution attainable is, in many cases, comparable in every way to that of
the electrophoretic method. However, separation by HPLC
is not restricted to size and charge differences. Furthermore, HPLC enables a smooth transition to the preparative scale with good recovery and, except in gel filtration,
enrichment from very dilute solutions.
A good example is the isolation of colony stimulating
factor (CSF) from human urine by hydrophobic-interaction HPLC.12241Thymosin p4 could be detected in various
cell culture extracts after rough preenrichment and subsequent RP-HPLC.“46’ Several separation principles are generally combined ; direct recovery of inhibin from follicular
fluid was successfully carried out by gel filtration and RPHPLC,1’25’ and ion-exchange and reversed-phase HPLC
proved to be appropriate for isolating y-interferon.[z261
Semipreparative techniques are widely used for the charac-
0
tlminl
-
LS
h g . I X. laolation of glucose-6-phosphate dehydrogenase by ion-exchange
chromatography of a crude bacterial extract (300 pg). HPLC system: Protein
Pak 5 PW, 7.5 mm x 7.5 cm (Waters), linear gradient from 0.02 M Tris/HCI.
pH=8.5, to 0.02 M Tris/HCI, 0.5 M NaCI, pH=8.5, in 25 rnin, 5 mL/min,
detection at A=280 n m : 0.1 absorption units full. Recovery of activity and
protein was 100% and 81%. respectively (with permission of Millipore-Waters, Eschborn, FRG).
Angew Chem. i n ! . Ed. Engl. 2s (1986) 535-552
-
60
40
tlminl
Fig. 19. Separation of various Ca’*-binding proteins by hydrophobrc-inleraction chromatography. HPLC system: TSK-Phenyl-5PW. 7.5 mm x 7.5 cm
(LKB), gradient from 1.5 M Na2S04 in 0.1 M phosphate buffer, pH=7.0,
1 mM C a 2 + to 0.1 M phosphate buffer, pH=7.0, 1 m M C a Z + in 74 min as
indicated (---), 0.75 mL/min, detection at 1 = 2 0 6 mm. Parvalbumrn (PV),
troponin C (TNC), and calmoddins from bovine brain (a), tetrahymena pyriformis (b), and paramecium (c) were recovered with more than 90% (from
[258]with permission of Academic Press).
t
h
I
0.
5
10
15
20
25
n-
Fig. 20. Affinity chromatography of acetylcholinesterdse ( A C h t ) from a
crude extract (1 rnL). n = fraction number. HPLC system: Ultraaffmity-EP.
4.6 mm x 5.0 cm (Beckman), derivatized with procainamide (8 mg): sample
loading ( I mL) with 0.1 M potassium phosphate, pH=7.0: elution with
0.1 ~rl
decamethonium salt in 0.1 M potassium phosphate, pH=7.0, 0.15 mL/
min, detection at 1=280 n m ( 0 - - - 0 )or by measurement of AChE activity
( O - - - O ) . Fivefold enrichment with 75% recovery in 2.5 h. (With permission
of Beckman Instruments, Munich, FRG.)
terization of proteins. The speed, reliability, and sensitivity
of sequence analysis of proteins was decisively improved
by HPLC methods; nowadays, the primary structure of
proteins can be determined with sample sizes in the microgram range. Again, these amounts can be optimally obtained by HPLC.L’8~23~2271
The isolation of membrane-bound proteins is somewhat
more problematic; the resolutions customarily obtained
with water-soluble proteins are rarely achieved here. Successful use of ion-exchange chromatography in the presence of nonionic o r zwitterionic detergents has been de~ c r i b e d . ~ ’ ~Antigenicity
* ~ ’ ~ ~ ~ is preserved by such mild separating conditions.[2301This can be of particular importance
in isolating viral proteins, which are of use in obtaining
antisera for diagnosing viral infections or in developing
synthetic vaccine^.^'^'^'^^^ However, there are also reports
that, under drastic RP-HPLC conditions, the isolated pro547
teins were still suitable for producing antisera to native viral proteins even though they could no longer be identified
by antibodies to the native
HPLC is meanwhile also being widely used in the purification of radioactively labeled polypeptide hormones and
antibodies for radioimmunological determinations. The
high selectivity of RP-HPLC can be utilized to separate
molecules labeled at different positions and, consequently,
to optimize the reaction conditions, remove by-products,
and isolate individual species.[233,2341
A h'igh degree of specific activity and stability is thereby obtained. The purification of antibodies can be carried out equally efficiently
by h y d r o ~ y l a p a t i t e [ ' or
~ ~hydrophobic-interaction
I
chromat0graphy,'4~]or by ion-exchange chromatography alone[2351
or in combination with gel filtration.[2361
The purely analytical application of protein HPLC is
much less common. In the most recent European pharmacopeial directives, HPLC methods are included in the case
of insulin only. This will definitely change, however, when,
as a result of gene technology, a number of new pharmaceutically useful polypeptides and proteins become available. The insulin example shows what possibilities are offered by HPLC for process optimization and quality control, in particular for the separation of microheterogeneities caused by processing.
The wide range of applications in the diagnostic field
have still not been fully explored. Interesting examples are
the identification of abnormal insulins and proinsulins in
hyperinsulinemic patient^,["^.*^^^ the analysis of genetically dependent hemoglobin
the characterization of glycosylated hemoglobins caused by metabolic
d i s o r d e r ~ , [ ~the
~ " l separation of receptor i s ~ f o r m s , and
~~~'~
the determination of immunoglobulin E in serum by affinity c h r ~ m a t o g r a p h y . ' ~Certain
~~'
isoforms of transferases
may be connected with myocardial infarction
specific
growth factors were detected by HPLC in the sera of cancer patients.f2+'I In qualitative serum screening, electrophoresis, with its high sample throughput, is unparalleled;
HPLC, however, may be employed irrespective of the sample concentration and is particularly of advantage where
quantitative analyses are concerned.[239.240.2421
6. Preparative Separation of Proteins by HPLC
Large-pore silica gels with minimal carbon coat115.116.I 19-222.2451 are one of the basic prerequisites
of preparative RP-HPLC.
The T P material from Vydac (TP=totally porous) particularly meets the requirements for the preparative separation of
Due to the minimal carbon coating
and residual saturation of silanol groups by double bonding (i.e., a second derivatization with very short ligands),
this material permits selective separations to be performed,
and, despite low capacity, the yield obtained is good. The
spherical 15-20-pm particles can be dry-packed. With radially compressible plastic cartridges, a simple but efficient separation system is obtained.
For preparative purposes, volatile buffers are most commonly used as solvent systems. The trifluoroacetic acid/
acetonitrile system is most often used, with considerable
success.[17.12'. 1361 The counterion effect results in sharp
peaks, and the products can be obtained salt-free after
freeze drying. Relatively sensitive proteins can be separated by pyridine/acetic acid or by pyridine/formic
a ~ i d [at ~p H~ 4-5.
, ~ Here
~ ~ too,
~ the proteins can be isolated
salt-free. Although acetonitrile is one of the best eluents
(sharp peaks), the less toxic propanol, which is stronger in
eluotropic power, and mixtures of propanol and butanol
are being used to an ever increasing extent125."5.1661
in order to keep the impact of the denaturing organic phase on
the protein as low as possible.
Interleukin 2 will be used to exemplify that direct transference of an analytical separation to the preparative scale
is possible.
6.1. Interleukin 2
Interleukin 2 (IL-2) is an immunologically active protein
that, like y-interferon, interleukin 1, and CSF, is now attracting great attention as a result of the progress achieved
in gene technology. It plays a central role in the immune
system during an immune reaction (Fig. 21). IL-2 is the
proliferation factor that is necessary for the increase in activated T and B lymphocytes-the blood cells responsible
for specific immune reaction. When the body comes in
contact with a n antigen, e.g., with a virus-transformed cell
or fragments thereof, the antigen is first eliminated by the
macrophages. The macrophages digest the antigen and express parts of it o n their surface, together with other essential indicator proteins. At the same time, they release a signal protein, interleukin 1, which activates the lymphocytes.
By means of interleukin 1 and the antigen on the macrophage surface, the T-helper lymphocytes specific to this
antigen are activated; they react to this activation by production of a number of messenger substances, including
interleukin 2. Interleukin 2 binds to the lymphocytes,
which, subsequent to activation by the antigen and interleukin 1, express the receptor for interleukin 2. The binding of IL-2 to the receptor drives the cells to proliferation
as long as IL-2 is present.
The cell division initiated by IL-2 binding leads to the
specific defense reaction of the immune system. The number of reactive cells is increased by a factor of 100-1000.
IL-2 is consequently an immunostimulator and is of interest for a whole series of therapeutic applications. Studies
ing[92,93.
548
Fig. 21. Role OF interleukrn 2
in
immune response (adapted from 12591).
Angew. Chern. In[. Ed Engl. 25 (1986) S35-552
performed at the National Cancer Institute o f the United
States of America have shown that IL-2 can also be successfully used in the treatment of tumors.
Human IL-2 was initially enriched in trace amounts
from supernatants of activated lymphocytes and partially
purified by RP-HPLC.[249-25'1
Elucidation of the nucleotide
sequence by Tuniguchhi et al.[2521
made possible the production of human IL-2 by gene technology. The protein consists of 133 amino acids with a molecular weight of about
15 000. Two of the three cysteine residues form a disulfide
bridge[2531
(Fig. 22), which is essential for biological activity. After the nucleotide sequence-and hence the amino
acid sequence-had been elucidated, human IL-2 was
cloned and expressed~252~254.2551
and has ever since been
available in therapeutic quantities. IL-2 has also been obtained from yeast, as well as from E. coli. The specific activity of the natural molecule glycosylated at Thr3 (Fig. 22)
is identical to that of glycosylated 1L-2 from yeast and to
that of nonglycosylated IL-2 from E. coli. It is clear from
this that glycosylation is not essential to the activity.
tography of 20 pg of IL-2 led to an elution of approximately 18 pg, i.e., a recovery of 90%.
The conditions of the analytical separation were first
transferred to a semipreparative scale (2-cm column). After
optimization of flow rate and elution gradient, we were already able to isolate IL-2 in milligram amounts. For preparative separation, we used radially compressible plastic
cartridges (5.7 cm x 30.0 cm) filled with large-pore Vydac
material. Despite the larger particles ( I 5-20 pm), the quality of separation of IL-2 on such a cartridge is comparable
to that on analytical columns (Fig. 23). About 60 mg of
high-purity pyrogen-free interleukin 2 was obtained per
run.
_1
I
j
'
I!
!'
CHO
0
20
tlminl
-
125
Fig. 23. Analytical and preparative HPLC separation of recombinant IL-2
from E. coli. (A) Analytical HPLC of an extract obtained with acetic acid/
butanol. HPLC system: Hipore CIS, 4.6 mm x 25.0 cm with guard column
(Bio Rad); eluent A : 0.1% trifluoroacetic acid in water; eluent B: 0.1% trifluoroacetic acid/80% acetonitrile in water; gradient from 35% to 85% B in
60 min, 1.5 mL/min, detection at A1=210 nm. (B) Preparative HPLC of 60 mg
crude IL-2. HPLC system: Prep Pak radially compressed cartridge,
5.7 cm x 30.0 cm (Waters) with T P 218, 15-20 w, 300 (Vydac), gradient as
described for (A), 50 rnL/rnin, detection with a 1/50 split. (C) Analytical
HPLC of the material obtained in (B). HPLC system as in (A). The curved
baseline results from an overlapping of the n-x* transition of the free acid
with the x-n* transition of the CF,COY anion [257].
A
80
Fig. 22. Primdry amucture of IL-2 with the essential disulfide bond and the
glycosylation site (CHO).
The physical properties of the nonglycosylated protein,
however, complicate purification by traditional methods
since, in E. coli, the protein is present as an aggregate in
inclusion bodies. It is insoluble in conventional buffers
and can therefore be relatively easily enriched. This is
achieved by washing the material with buffers of different
ionic strength, with and without detergents o r organic solvents; foreign proteins and other cell constituents are
thereby flushed out and IL-2 is enriched. This raw material
is then extracted with a mixture of acetic acid and alcohols. Such an extract contains u p to 4Ooh IL-2, which is
subsequently purified directly by RP-HPLC. Due to the
considerable hydrophobicity of IL-2, the protein is only
eluted after all impurities.
The initial studies using an analytical column and the
trifluoroacetic acid/acetonitrile solvent system gave excellent separations and showed no memory effect (i.e., elution
of protein, strongly bound in denatured form, in the subsequent runs) by precipitated IL-2, despite the large concentration of acetonitrile required for the elution. RechromaAngew. Chem.
Int.
Ed. Engl. 25 (1986) 535-552
It is well known that, in chromatography, the recovery
of hydrophobic proteins can be increased when a coating
of silica gel with relatively short alkyl chains, such as C3,
C,, o r C6, is used. The proteins are less strongly bound and
also often require a smaller proportion of organic phase
for the elution.
The advantage of shorter alkyl chains could also be demonstrated in the case of IL-2. On CI8material, the yield
decreased substantially after a single overload. The initial
separating characteristics and initial yield could only be
reestablished by regeneration with 60% acetic acid/guanidinium hydrochloride. In contrast, the C, material was relatively insensitive to overloading. Accordingly, for preparative, routine purification work, C, material should be given
preference to C I Smaterial.
6.2. Other Protein Purifications on a Preparative Scale
In contrast to IL-2, which can be obtained by means of
preparative HPLC in pure and highly active form, many
other proteins undergo partial unfolding and denaturation
upon RP-HPLC owing to the pH conditions, organic
eluents, and adsorption effects. By careful choice of the
stationary phase and elution system and by appropriate
549
further treatment of the fractions, it is also possible, however, to purify enzymes such as trypsin and alkaline phosphatase without loss of a ~ t i v i t y . [ ~ ~For
~ . ' ~instance,
'~
a preparation of trypsin purified on RP-18 is free of other proteolytic activities to the extent that it can be used for highly
selective peptide cleavages, as in sequence analysis.
Whereas the preparative separation of proteins with a
molecular weight greater than 20000 is still limited to a
few examples, the isolation and purification of antibodies
with a molecular weight of around 180000 by high-pressure ion-exchange and hydroxylapatite chromatography
has become generally accepted. A complete system for purifying monoclonal antibodies is available commercially
under the name MAPS.""]
7. Summary and Outlook
Within the last few years, HPLC versions of all conventional chromatographic techniques have been utilized in
protein chemistry. Efficient separation systems have been
developed and are now being used increasingly in fields
that previously were a domain of electrophoretic methods.
While the advantages of electrophoresis lie in the huge
number of parallel analyses possible (e.g., in serum screening), HPLC impresses by the speed of the individual analyses with direct detection of the components. Newly developed supporting materials made of extremely small
nonporous particles (< 2 pm) promise even further improvements. The high-resolution separations can also be
extended to preparative HPLC. The development here is
aimed at offering, at as reasonable a price as possible, stationary phases with relatively coarse particles, which can
easily be packed into large columns by the user and can be
operated with considerable long-term stability.
A crucial problem is the biocompatibility of stationary
and mobile phase. Gel filtration and ion-exchange chromatography are carried out under approximately physiological conditions. Separations on reversed phase, however, still involve the risk of protein denaturation, even if
considerable efforts have been made to reduce this problem, e.g., by using detergents. Solutions of this kind are
often specific for a particular protein and can rarely be
generalized. Hydrophobic-interaction chromatography
separates according to criteria similar to those of RPHPLC but, in principle, under milder conditions. By this
means, higher recovery rates of biologically active proteins
are achieved. For this reason, considerable further progress can be expected in this relatively new field.
We wish to thank Dr. H.-H. Schone for his interest in
these studies and for helpful discussions. We also wish to
thank C. Miiller and R . Lohfink for their help in preparing
the manuscript.
Received: March 3, 1986 [A 579 IE]
German version: Angew. Chem. 98 (1986) 530
Translated by Dr. Frank Hompson, Saarbnicken,
and by C. McConaghy, Hoechst AG, Frankfurt.
J. 3.5 (1941) 1388.
121 J . C . Giddings. Dynamics of Chromatography. Marcel Dekker, New
York 1965.
131 H. Funke, LaborPraxis 8 (1984) 18.
[ I { A. J. P. Martin, R. L. M. Synge, Biochem.
550
141 L. R. Snyder, M. A. Stadalius, M. A. Quarry, Anal. Chem. 55 (1983)
1412A.
[ S ] N. K. Boardman, S. M. Partridge, Biochem. J. 59 (1955) 543.
[6] H. P. Jennissen, J. Chromatogr. 159 (1978) 71.
[7] W. Kopaciewicz, M. A. Rounds, J. Fausnaugh, F. E.Regnier, J. Chmmntogr. 266 (1983) 3.
[S] X. Geng, F. E. Regnier, J. Chromatogr. 296 (1984) 15
[9] M. A. Stadalius, H. S. Gold, L. R. Snyder, J. Chromatogr. 296 (1984)
31.
[lo] J. M. Di Bussolo, J. R. Cant, J. Chromatogr. 327 (1985) 67.
[ I I ] M. Hermodson, W. C. Mahoney, Methods €nzymol. 91 (1983) 382.
1121 J. Luiken, R. van der Zee, G. W. Welling, J. Chromatogr. 284 (1984)
482.
[I31 L. R. Snyder, J. J. Kirkland: Introduction to Modern Liquid Chromatography. Wiley-Interscience, New York 1979.
[ 141 L. R. Snyder, J. Chromatogr. Sci. 21 (1983) 65.
[IS] P. A. Bristow, J. H. Knox, Chromatographia 10 (1977) 279.
[I61 G. Vanecek, F. E. Regnier, Anal. Biochem. 109 (1980) 345.
[I71 J. D. Pearson, N. T. Lin, F. E. Regnier, Anal. Biorhem. 124 (1982)
217.
[I81 J. W. Crdbb, L. M. G. Heilrneyer, J. Chrornafogr. 296 (1984) 129.
[I91 R. M. Moore, R. R. Walters, J. Chromotogr. 317 (1984) 119.
[20] F. E . Regnier, Science 222 (1983) 245.
1211 M. A. Stadalius, M. A. Quarry, L. R. Snyder, J. Chromatogr. 327 (1985)
93.
1221 N. T.Miller, B. L. Karger, J . Chromutogr. 326 (1985) 45
[23] E. C. Nice, C. J. Lloyd, A. W. Burgess, J. Chromotogr. 296 (1984)
153.
1241 R. van der Zee, G. W. Welling, J . Chrornafogr 325 (1985) 187.
[25] K. J. Wilson, E. van Wieringen, S. Klauser, M. W. Berchtold, G. J.
Hughes, J. Chromatogr. 237 (1982) 407.
[26] R. V. Lewis, A. Fallon, S. Stein, K. D. Gibson, S. Udenfried, Anal. Biochem. 104 (1980) 153.
[27] G. Vanecek, F. E. Regnier, Anal. Biochem. 121 (1982) 156.
[28] K. A. Cohen, K. Schellenberg, K. Benedek, B. L. Karger. B. Grego, M.
T. W. Hearn, Anal. Biochem. 140 (1984) 223.
[29] W. Kopdciewicz, F. E. Regnier, Anal. Biochem. 133 (1983) 251.
[30] S. C. Goheen, S . C . Engelhorn, J . Chromatogr. 317 (1984) 5 5 .
1311 F. E. Regnier, Methods Enzymol. 91 (1983) 137.
1321 R. P. W. Scott: Liquid Chromatography Detector.s, Elsevier, New York
1977.
[33] C. J. R. Thorne in H. L. Kornberg (Ed.). Techniques in L!fe Sciences.
B 104, Elsevier, Amsterdam 1978, p. I .
1341 W. B. Gratzer in G . D. Fasrnan (Ed.): Poly-a-amino acid.s, Bio/ogica/
Macromolecu1e Series, Marcel Dekker, New York 1967. p. 177.
1351 J . L. Di Cesare, Chromatogr. N e w d 9 (1981) 16.
1361 C. Y. Yang, H. J. Pownall, A. M. Gotto, A n d Biochem. 145 (1985)
67.
[37] L. Servillo, G. Colonna, C . Balestieri, R. Ragone, G. Irace, Anal. Biochem. 126 (1982) 251.
[38] T. D. Schlabach, T. C. Wehr, Anat. Biochem. 127 (1982) 222.
[39] G. Schwedt: Chemrsche Reaktionsdetektoren f i r die schnelle Nussigkeitsrhromatographi. Hfithig Verlag, Heidelberg 198 I .
[40] J. R. Benson, Anal. Biochem. 71 (1976) 489.
[41] P. Bohlen, S. Stein, J Stone, S. Udenfried, A n d Biorhem. 67 (1975)
438.
[42] H. Nika, J. Chromotogr. 297 (1984) 261.
1431 G. K. Ackers, Adv. Protein Chem. 24 (1970) 343.
1441 E. Pfannkoch, K. C. Lu, F. E. Regnier, H. G. Barth, J. Chromotogr. Sci.
18 (1980) 430.
1451 G. Guiochon, M. Martin, J. Chromatogr. 326 (1985) 3
1461 T. V. Alfredson, C. T. Wehr, L. Tallman, F. Klink, J. Liq. Chromatogr. 5
(1982) 489.
[47] M. Okazaki, Y. Ohno, I. Hara, J Chromatogr. 221 (1980) 257.
[48] Y. Kato, K. Komiya, H. Sasaki, 7. Hashimoto, J . Chromatogr. I93
(1980) 311.
149) S. C . Goheen, R. S. Matson, J . Chromatogr. 326 (1985) 235.
[SO] F. E. Regnier, R. Noel, J . Chromatogr. Sci. 14 (1976) 316.
[Sl] H. Engelhardt, D. Mathes, J. Chromatogr. 142 (1977) 31 I .
I521 Y. Kato, K. Komiya, H. Sasaki, T. Hashimoto, J. Chromatogr. 190
(1980) 297.
1531 P. Roumeliotis, K. K. Unger, J Chromatogr. 149 (1978) 21 1.
1541 R. W. Stout, J. J. De Stefano, J. Chromatogr. 326 (1985) 63.
1551 Y. Kato, K. Komiya, H. Sasaki, T. Hashimoto, J. Chromatogr. 193
(1980) 29.
1561 Y. Kato, K. Komiya, H. Sasaki, T. Hashimoto, J . Chromatogr. 193
(1980) 458.
1571 J . E . Rivier, J . Chromatogr. 202 (1980) 211.
[58] B. S. Welinder in W S. Hancock (Ed.): Handbook q/ HPLCfor the
Separation ofAmino Acids. Peptides and Proteins, Vol. 11. CRC Press,
Boca Raton, FL, USA 1984, p. 413.
[89] F. E. Regnier, Anal. Biochem. 126 (1982) I .
I601 0. Mikes, I n t . J. Pept. Prorein Res. 14 (1979) 393.
Angew. Chem. In[. Ed. Engl. 25 (1986) 835-552
[61] Y . Kato, K. Nakamurd, T. Hashimoto, J Chromafogr. 294 (1984) 207.
[62] T. Kadoya. T. laobe, Y . Amano, Y. Kato, K Nakamura, T. Okuyama,
J . Lrq Cliromafogr. 8 (1988) 638.
1631 K. Nakamura. Y . Kato, J . Chromarogr. 333 (1985) 29.
[64] J. Richey, Am. Lab. 14 (1982) 104.
1651 L. Fdgerstam. L Soderberg, L. Wahlstrom, Protides B i d Fluids 30
(1982) 621.
1661 K. M. Gooding. M. N. Schmuck, J . Chromafogr. 327 (1985) 139.
1671 S. H. Chang, R. Noel, F. E. Regnier, Anal. Chem. 48 (1976) 1839.
1681 A. J. Alpert, F. E. Regnier, J . Chromarogr 185 (1979) 375.
I691 S. Gupta, E. Pfannkoch, F. E. Regnler, Anal. Biochem. 128 (1983)
196.
1701 Y. Kato. K. Komtya, T. Hashimoto, J . Chromafogr. 246 (1982) 13.
1711 Y Kato. K. Nakamurd, T. Hashimoto, H R C & CC. J . High Resolut.
Chromutogr. Chromotogr. Cornmun. 8 (1985) 184.
[72] C. Tanford. Science 200 (1978) 1012.
1731 2. Er-el, Y . Zaidenzaig, S. Shaltiel, Biochem. Biophys. Res. Commun. 49
(1972) 383.
[74] H. A. Arfmann, S. Shaltiel, Eur. J . Biochem. 70 (1976) 269.
[78] S . Shaltiel, Methods Enzyntol. 104 (1984) 69.
[76] H. Engelhardt, D. Mathes, Chromorographia 14 (1981) 328.
[77] Y . Kato. T. Kitamura, T. Hashimoto, J . Chromatogr. 266 (1983) 49.
1781 J . L. FdUsnaUgh, E. Pfannkoch, S. Gupta, F. E. Regnier, Anal. Biochem.
137 ( 1984) 464.
1791 N. T. Miller. B. Feibush, B. L. Karger, J . Chromatogr. 316 (1984) 519.
1801 D. L. Gooding, M. N . Schmuck, K. M. Gooding, J . Chromatogr 296
(1984) 107.
[81] Y. Kato, T. Kitamura, T. Hashimoto, J. Chromafogr. 292 (1984) 418.
1821 Y. Kato, T. Kitamura, T. Hashimoto, J . Chromafogr. 298 (1984) 407.
1831 J . L. Fausnaugh, L. A. Kennedy, F. E. Regnier, J . Chrornatogr. 317
(1984) 141.
1841 W. R. Melander, D. Corradini, C. Horvath, J . Chromarogr. 317 (1984)
67.
[88] W. R. Melander, C. Horvath, Arch. Biochem. Biophys. 183 (1977) 393.
[86] R. H. Ingraham, S. Y . M. Lau, A. K. Taneja, R. S. Hodges, J. Chromaroyr 327 (1988) 77.
1871 G. A. Howard, A. J. P. Martin, Brochem J . 46 (1950) 532.
1881 R. E. Majors, J Chromufogr. Sci. I8 (1980) 485.
(891 C. Horvath, W. Melander, I. Molnar, J . Chromafogr. 125 (1976) 129.
[90] C. Horvath, W. Melander, I. Molnar, Anal. Chem. 49 (1977) 142.
1911 C. Horvath, W. Melander, J. Chromotogr. Sci: 15 (1977) 393.
1921 W. S. Hancock, J. T. Sparrow in C. Horvath (Ed.): High-PerJormance
Liquid Chromofography: Advances and Perspectives. Yo/. 3. Academic
Press, New York 1983, p. 49.
[93] M. T. W. Hearn, Methods Enzymol. 104 (1984) 190.
1941 R. A. Barford, B. J. Sliwinski, A. C. Breyer, H. L. Rothbart, J. Chromatrqr 235 (1982) 281.
[98] W. C. Mahoney, Biochim. Brophys. Acta 704 (1982) 284.
[96] J . 0. Pearson, W. C. Mahoney, M. A. Hermodson, F. E. Regnier, J .
Chromatogr. 207 (1981) 325.
1971 M. T. W. Hearn, B. Grego, J . Chromatogr. 255 (1983) 128.
[98] M. J . O’Hare, E. C. Nice, J . Chromarogr. 171 (1979) 209.
[99] S:J. Su, B. Grego, B. Niven, M. T. W. Hearn, J . Lrq. Chroma/ogr. 4
(1981) 1745.
[loo] J. L. Meek, Z. L. Rossetti, J . Chromatogr. 211 (1981) IS.
[loll T. Sdsagawa, T. Okuyama, D. C. Teller, J. Chromatogr. 240 (1982)
329.
[I021 E. Keshavarz, S. Nakai, Biochrm. Biophys Acra 576 (1979) 269.
[I031 S. Terabe, H. Nishi, T. Ando, J. Chromafogr. 212 (1981) 295.
I1041 E. C. Nice, M. W. Capp, N. Cooke, M. J. O’Hare, J. Chromatogr 218
(1981) 569.
[I081 A. J Sadler, R. Micanovic, G. E. Katzenstein, R. V. Lewis, C . R. Middaugh, J Chromufogr. 317 (1984) 93.
[I061 K. Benedek, S. Dong, B. L. Karger, J . Chromatogr. 317 (1984) 227.
[I071 A R. Kerlavage, C. J. Weitzmann, T. Hasan, B. S. Cooperman, J. Chromarogr. 266 (1983) 225.
[I081 S. A. Cohen, K. Benedek, S. Dong, Y. Tapuhi, B. L. Karger, Anal.
Chum. 56 ( 1984) 2 17.
11091 P. G . Stanton. B. Grego, M. T. W. Hearn, J . Chromafogr. 296 (1984)
I 89.
[ I 101 W. S. Hancock, C . A. Bishop, M. T. W. Hearn, FEBS Left. 72 (1976)
139.
[ I I I] W. S. Hancock, C. A. Bishop, R. L. Prestidge, D. R. K. Harding, M. T.
W. Hearn, Science 200 (1978) 1168.
[I 121 Y. Kato, T. Kitamura, T. Hashimoto, 1.Clrromatogr. 333 (1985) 93.
[ I 131 Technical bulletin: Polymeric HPLC-columns for proreins and peplides.
Polymer Laboratories, 1985 (Latek GmbH).
[ I 141 F . V Warren, B. A. Bidlingmeyer, J. Lrq. Chromafogr. 8 (1985) 619.
[ I 181 R. V Lewis, D. De Wald, J. Liq. Chromatogr. 5 (1982) 1367.
[I161 M. J. O’Hare, M. W. Capp, E C . Nice, N. H C. Cooke, B. G. Archer,
Anal. Biochem. 126 (1982) 17.
[ I 171 E. c‘. Nice, M. W. Capp, M. J. O’Hare, J. Chromatogr. 185 (1979)
41.
Anyeu. Chem. I n r . Ed. Engl. 25 (1986) 535-552
[ I IS] J. D. Pearson, F. E. Regnier, J . Liq. Chromatogr. 6 (1983) 497.
[ I 191 N. H. C. Cooke, B. G. Archer, M. J . O’Hare, E. C. Nice, M. W. Capp, J .
Chromatogr. 255 (1983) I 18.
11201 H Engelhardt, H. Muller, Chromafographia 19 (1984) 77.
(1211 M. T. W. Hearn, B. Grego, J. Chromotogr. 282 (1983) 841.
11221 M. T. W. Hearn, B. Grego, J . Chromarogr. 296 (1984) 61.
[I231 1. Molnar, C. Horvath, J . Chromarogr. 142 (1977) 263.
[I241 M. T. W. Hearn, 8 . Grego, W. S. Hancock, J . Chromafogr. 185 (1979)
429.
[I251 M . T. W. Hearn, Adu. Chromafogr. 18 (1980) 59.
[126] W. S. Hancock, C. A. Bishop, L. J. Meyer, D. R. K. Harding, M T. W.
Hearn, J . Chromatogr. 161 (1978) 291.
[I271 W. C. Mahoney, M. A. Hermodson, J . Biol. Chem. 255 (1980) I 1 199.
[I281 J. E. Rivier, J . Liq. Chromafogr. 1 (1978) 343.
[I291 J. B. Shelton, J. R. Shelton, W. A. Schroder, J . Liq. Chromutogr. 4
(1981) 1381.
[I301 U. Certa, G. van Ehrenstein, Anal. Biochem. 118 (1981) 147.
[I311 G. Szepesi, M. Gazdag, J. Chromnfogr. 218 (1981) 597.
[I321 M. Kehl, F. Lottspeich, A. Henschen, Hoppe-Seyleri 2. Ph.wro1. Chem.
363 (1982) 1501.
[I331 H. P. J . Bennett, C. A. Browne, S. Solomon, J. Liq Chromatogr. 3
(1980) 1353.
[134] C. A. Browne, H. P. J. Bennett, S. Solomon, Anal. Biochem. 124 (1982)
201.
[I381 J. J. L‘ltalien, J . E. Strickler, Anal. Biochem. 127 (1982) 198.
[I361 K. Titani, T. Sasagawa, K. Resing, K. A. Wdlsh, Anal. Biochem. 123
(1982) 408.
[I371 H. P. Nlck, R. E. H. Wettenhall, M. T. W. Hearn, F. J . Morgan, Anal.
Biochem. 148 (1988) 93.
[I381 M. van der Rest, P. Fietzek, Eur. J. Biochem. 125 (1982) 491.
11391 G. E. Gerber, H. G. Khorana, Methods Enzymol. 88 (1982) 56.
[I401 J. Heukeshoven, R. Dernick, J. Chromarogr. 326 (1985) 91.
[I411 M. Rublnstein, Anal. Biochem. 98 (1979) I .
[I421 B. N. Jones, R. V. Lewis, S. PIabo, K. Kojima, S. Kimuro, S Stein, J .
Liq. Chromafogr 3 (1980) 1373.
[I431 M. Green, K. H Brackmann, Anal. Biochem. 124 (1982) 209.
[I441 A. Fallon, R. V. Lewis, K. D. Gibson, Anal. Bi<Jchem.I10 (1981) 318.
11451 K. J. Wilson, A. Honegger, R. P. Stolzel, G. J. Hughes, Brochem. J. 19Y
(1981) 31.
[I461 E. Hannappel, W. Leibold, Prorrdes B i d . Fluids 32 (1984) 1093.
[I471 B. Grego, G. S. Baldwin, J. A. Knessel, R. J. Simpson, F. J. Morgan. M .
T. W. Hearn, J . Chromatogr. 297 (1984) 21.
[I481 G. E. Tarr, J . W. Crabb, Anal. Biochem. 131 (1983) 99.
[I491 S. W. Herring, R. K. E m s , J . Chromatogr 266 (1983) 249.
[I501 W. S. Hancock, J. D. Capra, W. A. Bradley, J. T. Sparrow, J . Chromafogr. 206 (1981) 59.
[I811 W. S. Hancock, J. T. Sparrow, J . Chromafogr. 206 (198 I ) 71.
[152] R. J. Ferris, C. A. Cowgill, R. R. Trdut, Brochemisfry 23 (1984) 3434.
11531 J. Spiess, J. E. Rivier, C . Rivier, M. Vale, Proc. Nafl.Acotl. Scr. USA 78
(1981) 6517
[I541 W. Richter, P. Schwandt, J. Neurochem. 36 (1981) 1279.
[ISSl M. W. Berchthold, K. .I. Wilson, C . W. Heizmann, Biochemi.s/rv 21
(1982) 6552.
[I561 E. Rinderknecht, B. H O’Connor, H. Rodriguez, J . Brol. Chem. 259
(1984) 6790.
[I871 B. Larsen, B. L. Fox, M. F. Burke, V. J. Hruby, lnt. J . Pept. Prorein Re.s.
13 (1979) 12.
[IS81 J. A. Smith, M. 1. O’Hare, J . Chromalogr. 299 (1984) 13.
[IS91 H. Gazdag, G. Szepesi, J . Chromatogr 218 (1981) 603.
I1601 R. van der Zee, S. Welling-Wester, G. W. Welling, J. Chromatop- 206
(1983) 577.
[I611 A. R. Kerlavage, T. Hasan, B. S. Cooperman, J . Brol. Chem. 258 (1983)
6313.
I1621 G. s. Baldwin, B. Grego, M. T. W. Hearn, J. A. Knessel, F. J. Morgan,
R. J. Simpson, Proc. Null. Acad. Sci. USA 80 (1983) 5276.
[I631 M. P. Strickler, 1. Kintzios, M. J. Gemski, J . Liq. Chromatogr. 5 (1982)
1921.
11641 S. A. Cohen, K. Benedek, Y. Tapuhi, 1. C. Ford, B. L. Karger, Anal.
Biochem. 144 (1988) 278.
11651 M. T. W. Hearn, A. N. Hodder, M. 1. Aguilar, J . Chromatogr. 327
( 1985) 47.
I1661 J. P. Chang, W. R. Melander, C. Horvath, J . Chromatogr. 318 (1985)
1I .
[I671 T. L. Brooks, A. Stevens, I n / . Lab 15 (1988) No. 9, p. 72.
I1681 S. van der Wal. J. F. K. Huber, Anal. Biochem. I05 (1980) 219.
11691 R. R. Wdlters, Anal. Chem. 57 (1988) 1099A.
[I701 T. M. Phillips, W. D. Queen, N. S. Moore, A. M. Thompson, J . Chronrorogr. 327 (1985) 213.
11711 S. Ohlson, L. Hansson, P. 0. Larsson, K. Mosbach, FEES Leff. 93
(1978) 8.
[I721 J. R. Sportsman, G. S. Wilson, Anal. Chem. 52 (1980) 2013.
11731 M. E. Robers, M. W. Adlard, G. Saunders, G. Holt, J. Chromatogr. 326
(1985) 163.
55 I
[I741 A. Borchert, P. 0. Larsson, K. Mosbach, J . Chromatogr. 244 (1982)
49.
[I751 C. R. Lowe, M. Glad, P. 0. Larsson, S. Ohlson, D. A. P. Small, T. Atkinson, K. Mosbach, J . Chromatogr. 215 (1981) 303.
[ I 761 Fast Affinity Chromatography (Bulletin 5933) Beckman Instruments,
San Ramon, CA, USA, 1985.
(1771 G. Wagner, F. E. Regnier, Anal. Biochem. 126 (1982) 37.
[I781 T. W. Hutchens, H. E Dunaway, P. K. Besch, J . Chromatogr. 327
(1985) 247.
[I791 F. Sanger, H. Tuppy, Biochem. J . 49 (1951) 481.
[I801 F. Sanger, E. 0. P. Thompson, Biochem. J . 53 (1953) 366.
[ I S l ] G. Bentley, E. Dodson, G. Dodson, D. Hodgkin, D. Mercola, Nature
London 261 (1976) 166.
[I821 A. McLeod, S. P. Wood, J . Chromatogr. 285 (1984) 319.
[I831 H. S. Tager, Diabetes 33 (1984) 693.
[I841 B. S. Welinder, S. Linde in W. S. Hancock (Ed.): Handbook of HPLC
f o r the Separation of Amino Acids. Peptides and Proteins, Vol. 11. C R C
Press, Boca Raton, FL, USA 1984, p. 357.
[I851 E. P. Kroeff, R. E. Chance in J. L. Guerinian (Ed.): Proc. o f t h e FDA
Workshop on Insulins. Growth Hormone and Recombinant DNA Technology, The U. S . Pharmacopeial Convention Inc., Rockville, MD,
USA 1982, p. 148.
B. S. Welinder, F. H. Andresen, ibid., p. 163.
S. Linde, B. Hansen, 0. Sonne, J. J. Holat, J. Gliemann, Diabetes 30
(1981) I .
B. H. Frank, D. E. Peavy, C. S. Hooker, W. C. Duckworth, Diabetes32
(1983) 705.
B. S. Welinder, S. Linde, B. Hansen, J . Chromatogr. 265 (1983) 301.
K. H. Joergensen, DBP 1940 130 C2, Novo Terapeutisk Laboratorium
1983.
R. D. Cole, Fed. Proc. Fed. A m . Soc. Exp. Biol. 18 (1959) 206.
M. Zoltobrocki, DOS 2659568 C3, Hoechst AG 1981.
W. Burgermeister, F. Enzmann, H. H. Schone in A. Hasselblatt, F. von
Bruchhausen (Eds.): Handbuch der experimentellen Pharmakologie.
Springer, Berlin 1975, p. 715.
D. V. Goeddel, D. G. Kleid, F. Bolivar, H. C. Heyneken, D. G. Yansura, R. Crea, T Hirose, A. Kraszewski, K. ItakUrd, A. D. Riggs, Proc.
Natl. Acad. Sci. USA 76 (1979) 106.
R. E. Chance, E. P. Kroeff, J. A. Hoffmann in J. L. Guerigian (Ed.):
Insulins. Growth Hormone and Recornbinant DNA Technology. Raven
Press, New York 1981, p. 71.
I. S. Johnson, Science 219 (1983) 632.
F. Wengenmeyer, Angew. Chem. 95 (1983) 874; Angew. Chem. Int. Ed.
Engl. 22 (1983) 842.
I. Fruton, Adu. Enzymol Relat. Areas Mol. Biol. 53 (1982) 239.
H.-D. Jakubke, P. Kuhl, A. Konnecke, Angew. Chem. 97 (1985) 79; Angew. Chem. I n / . Ed. Engl. 24 (1985) 8 5 .
a) R. Obermeier, G. Seipke, U. Grau in K. G. Petersen, K. J. Schluter,
L. Kerp (Eds.): Neue Insuline. Freiburger Graphische Betriebe, Freiburg 1982, p. 32; b) J. Markussen, ibid., p. 38.
R. Obermeier, G. Seipke, Process Biochem. 19 (1984) 29.
1. D. Young, F. H. Carpenter, J . Biol. Chem. 236 (1961) 743.
D. F. Steiner, Trans. N. Y. Acad. Sci. 30 (1967) 60.
D. F. Steiner, J. L. Clark, C. Nolan, A. H. Rubenstein, E. Margoliash,
B. Aten, P. E. Oyer in S. Baba, T. Kaneko, N. Yanaihara (Eds.): Proind i n . Insulin. C-Peptide. Excerpta Medica, Amsterdam 1979, p 267.
.
. 1. M. Swan, Nature London 180 (1957) 643.
[206] G. H. Dixon, A. C. Wardlaw, Nature London 188 (1960) 721.
[207] B. H. Frank, Eur. Pat. Appl. 0037255, Eli Lilly and Company 1981.
[208] U. Grau, Diabetes 34 (1985) 1174.
[209] F. H. Carpenter, A. Crdmbach, J . B i d . Chem. 237 (1962) 404.
[2lOI United Slates Pharmacopeia. 21st Revision, U. S. Pharmacopeial Convention Inc., Rockville, MD, USA 1985, p. 536.
[21 I] H. Ross, G. Bernstein, H. Rifkin in M. Ellenberg, H. Rifkin (Eds.):
Diabetes Meflitus. Theory and Practice, Medical Examination Publishing Corp., New York 1983, p. 907.
[212] M. Brownlee, H. Vlassara, A. Cerami, Ann. Intern. Med. 101 (1984)
527.
(2131 P. Rastein in M. Ellenberg, H. Rifkin (Eds.): Diabetes Meflitus, Theory
and Practice, Medical Examination Publishing Corp., New York 1983,
p. 941.
I2141 J. C. Pickup, D. Rothwell, Med. Biol. Eng. Comput. 22 (1984) 385.
[215] K. Irsigler, H. Kritz, R. Lovett (Eds.): Diabetes Treatment with Implantable Insulin Infusion Systems. Urban und Schwanenberg, Munich
1983.
[216] K. Irsigler, H. Kritz, Diabetes 28 (1979) 196
[217] W. D. Lougheed, H. Woulfe Flanagan, J. R. Clement, A. M. Albisser,
Diabetologia 19 (1980) I .
1218) U . Grau, Diabetologia 28 (1985) 458.
552
[2191 U. Grau in K. D. Hepp, R. Renner (Eds.): Continuous Insulin Infusion
Therapy. Experience from One Decade, Schattauer, Stuttgart 1985,
p. 33.
[2201 U. Grau, G. Seipke, R. Obermeier, H. Thurow in K. G. Petersen, K. J.
Schluter, L. Kerp (Eds.): Neue Insuline, Freiburger Graphische Betriebe, Freiburg 1982, p. 441.
I2211 J. Brange, L. Langkjaer, S. Havelund, E. Ssrensen, Diabetologia 25
(1983) 143.
[2221 U. Grau, P. Jahrling, Diabetologia. submitted for publication: H. P.
Neubauer, Hoechst AG, unpublished results.
[2231 R. A. Rizza, J. E. Gerich, M. W. Haymond, R. E. Westland, L. D. Hall,
A. H. Clemens, I. J. Service, N. Engl. J . Med. 303 (1980) 1313.
[2241 P. V. Byrne, J . Chromatogr. 283 (1984) 459.
12251 M. Dobos, H. G. Burger, M. T. W. Hearn, F. J. Morgan, Mol. Cell.
Endocrinol. 31 (1983) 187.
[2261 D. Novick, D. G Fischer, Y. Mory, M. Revel, M. Rubinstein, Protides
Biol. Fluids 32 (1984) 1049.
[2271 M. W. Hunkapiller, J. E. Strickler, K. J . Wilson, Science 226 (1984)
304.
[2281 P. Lundahl, E. Greijer, H. Lindblorn, L. G. Fagerstam, J . Chromatogr.
297(1984) 129.
12291 H. Ikigai, T. Nakae, Y. Kato, J . Chromatogr. 322 (1985) 212.
12301 P. Lambotte, J. van Snick, T. Boon, J . Chromatogr. 297 (1984) 139.
[2311 G. W. Welling, G. Kroen, K. Slopsema, S. Welling-Wester, J. Chromatogr. 326 (1985) 173.
P321 R. van der Zee, G. W. Welling, J . Chromatogr. 327 (1985) 377.
I2331 B. H. Frank, M. J. Beckage, K. A. Willey, J . Chromatogr. 266 (1983)
239.
[2341 B. S. Welinder, S. Linde, B. Hansen, 0 Sonne, J. Chromatogr. 298
(1984) 41.
[2351 J. R. Deschamps, J. E. K. Hildreth, D. Deer, J. T. August, Anal. Biochem. 147 (1985) 451.
[2361 M. J . F. Schmerr, K. R. Goodwin, H. D. Lehmkuhl, R. C. Cutlip, J .
Chromatogr. 326 (1985) 225.
[237] S. Seino, A. Funakoshi, 2. 2. Fu, A. Vinik, Diabetes 34 (1985) I.
[238] T. H. J. Huismans, J. B. Wilson, Protides Biol. Fluids 32 (1984) 1029.
12391 A. Kutlar, F. Kutlar, J. B. Wilson, M. G . Headlee, T. H. J. Huismans,
Am. J . Hematol. 17 (1984) 39.
[240] E. C. Toren, D. N. Vacik, P. B. Mockridge, J . Chromatogr. 266 (1983)
207.
[2411 D. M. Boyle, R. D. Wiehle, N. A. Shahabi, J. L. Wittcliff, J . Chromatogr. 327 (1985) 369.
12421 T. M. Phillips, N. S. Moore, W. D. Queen, A. M. Thompson, J . Chromntogr 327 (1985) 205.
I2431 S. W. Rabkin, P. Desjarduis, Clin. Chim. Acta 138 (1984) 245.
[244] E. S. Kimball, W. H. Bohn, K. D. Cockley, T. C. Warren, S. A. Sherwin,
Cancer Res. 44 (1984) 3613.
[245] L. E. Henderson, R. Sowder, S. Oroszlan in T. Liu, A. Schechter, R.
Heinrikson, P. Condliffe (Eds.): Chemical Svnthesis and Seyuenzing of
Peptides and Proreins. Elsevier/North-Holland, Amsterdam 1981,
p. 251.
L246j J. E. Rivier, R. McClintock, R. Galyean, H. Anderson, J . Chromatogr.
288 (1984) 303.
[247] H. J. Friesen, S. Stein, M. Evinger, P. C. Familletti, J. Moscherd, J .
Meienhofer, J. Shively, S. Pestka, Arch. Biochem. Biophys. 206 (1981)
432.
I2481 S. A. Rosenberg, M. T. Lotze, L. M. Muul, S. Leitman, A. E. Chang, S.
E. Ettingshausen, Y. L. Matory, J. M. Skibber, E. Shiloni, J. T. Vetto, C.
A. Seipp, C. Simpson, C. M. Reichert, N. Engl. J. Med. 313 (1985)
1485.
[249] K. Welte, C. Y. Wang, R. Mertelmann, V. Salvatore, S. P. Feldman, M.
A. S. Moore, J . Exp. Med. 156 (1982) 454.
12501 E. M. Kniep, B. Kniep, W. Grote, H. S. Conradt, D. A. Monner, P. F.
Miihlradt, Eur. J . Biochem. 143 (1984) 199.
12511 A. Godard, J. Naulet, M. A. Peyrat, H. Vie, J. F. Moreau, J. D. Bignon,
J. P. Soullilou, J . Immunol. Methods 70 (1984) 233.
[252] T. Taniguchi, M. Matsui, T. Fujita, C. Takaoka, N. Kashima, R. Yoshimoto, J. Hamuro, Nature London 302 (1983) 305.
[253] R. J. Robb, R. M. Kutny, M. Panico, H. R. Morris, V. Chowdry, Proc.
Nut. Acad. Sci. USA 81 (1984) 6486.
(2541 R. Devos, G. Plaetinck, M. Cheroute, C. Simons, W. Degrave, J. Tavernier, E. Remaut, W. Fiers, Nucleic Acids Res. I 1 (1983) 4307.
I2551 K. Kato, T. Yamada, K. Kawahara, H. Onda, T. Asano, H. Sugino, A.
Kakinuma, Biochem. Bioph.y.7. Res. Commun. 130 (1985) 692.
[256] M. P. Strickler, M. J. Gemski, B. P. Doctor, J . Liy. Chromatogr. 4
(1981) 1765.
12571 G. Winkler, P. Wolschan, P. Briza, F. X. Heinz, C. Kunz, J . Chromatogr 347 (1985) 83.
[258] D. Guerini, J. Krebs, Anal. Biochem. I50 (1985) 178.
[259] H. Wagner, M. Rollinghoff, Behring Inst. Mitt. 67 (1980) 2.
Angew. Chem. Int. Ed. Engl. 25 (1986) 535-552
Документ
Категория
Без категории
Просмотров
4
Размер файла
2 108 Кб
Теги
analytical, hplc, high, protein, method, pressure, chromatography, liquid, new
1/--страниц
Пожаловаться на содержимое документа