Biotechnol. Prog.

1092, 8, 469-478

409

REVIEW
Recombinant Human Insulin
Michael R. Ladisch' and Karen L. Kohlmann
Laboratory of Renewable Resources Engineering, A. A. Potter Engineering Center, Purdue University,
West Lafayette, Indiana 47907

Insulin is a well-characterized peptide that can be produced by recombinant DNA

technology for human therapeutic use. A brief overview of insulin production from
both traditional mammalian pancreatic extraction and recombinant bacterial and yeast
systems is presented, and detection techniques, including electrophoresis, are reviewed.
Analytical systems for insulin separation are principally based on reversed-phase
chromatography, which resolves the deamidation product(s) (desamido insulin) of insulin,
proinsulin, and insulin. Process-scale separation is a multistep process and includes
ion exchange, reversed-phase, and size exclusion chromatography. Advantages and/or
disadvantages of various separation approaches, as described by the numerous literature
references on insulin purification, are presented.

Contents
Introduction
Background
Insulin Structure
Traditional Insulin Production and
Purification
Production of Human Insulin
Conversion of Porcine Insulin
Insulin Produced by
Recombinant Methods
Two-Chain Method
Proinsulin Method
(Intracellular)
Proinsulin (Secreted)
Analytical Separation of Insulin
Reversed-Phase High-Performance
Liquid Chromatography
Affinity Chromatography
Electrophoresis of Insulins
Large-Scale Purification of Human
Insulin
Conclusion

469
470
470
470
471
471
472
472
473
474
474
474
474
474
475
476

Introduction
Insulin is a polypeptide hormone that is essential for
the supply of energy to the cells of the body (Barfoed,
1987). It has been estimated that the disease diabetes
mellitus (impaired insulin production and its complications) is the third largest cause of death in industrialized
countries after cardiovascular diseases and cancer (Barfoed, 1987). Produced by the @-cellsof the pancreas in
response to hyperglycemia, insulin is a potent hormone
directly or indirectly affecting every organ or tissue in the

* Author to whom correspondence should be addressed.

8756-7938/92/3008-0469$03.00/0

body. The main functions of insulin are to stimulate

anabolic reactions for carbohydrates, proteins, and fats,
all of which have the metabolic consequences of producing
a lowered blood glucose level (Norman and Litwack, 1987).
I t was estimated in 1986 that there were 60 million
diabetics in the world (Johnson, 1986)and that 12 million
Americans have diabetes (Barfoed, 1987). These numbers
are increasing dramatically. The US. population is
thought to be growing at a rate of 1%per year, but the
rate of growth of the insulin-using diabetic population is
5 4 % per year (Johnson, 1986).
Health problems relating to diabetes can be devastating.
Barfoed (1987) reports that, although acute symptoms
can be treated with insulin, vascular complications in the
later stages of the disease reduce the life expectancy of
diabetics by one-third. In addition, he reports that
diabetics are 25 times more likely to go blind and have
twice the risk of heart disease than nondiabetics. In the
US. the direct and indirect costs to the economy are
estimated to be over 5 billion dollars annually.
The administration of insulin as a diabetic treatment
has a long and interesting history. In 1922, Frederick
Banting and Charles Best successfully treated the first
human patient with an insulin preparation obtained from
animal pancreatic extractions (Barfoed, 1987). From 1922
to 1972 the only available insulin was purified from the
pancreases of pigs and cows. This insulin was quite
valuable in prolonging the lives of diabetics who otherwise
would have slowly died because glucose was unavailable
to their body cells. Improvements in the purification of
insulin during the 1970s were successful in producing a
stable drug with a predictable action time (Chance et al.,
1975; Dolan-Heitlinger, 1981; Barfoed, 1987). In 1950,
Novo Nordisk A/S (Bagsvaerd, Denmark) modified the
state of the insulin itself to obtain an insulin crystal with
a high zinc content which had a long action time, in
comparison to amorphous insulin that was quickly absorbed (Dolan-Heitlinger, 1981; Barfoed, 1987). The
additions of protamine and zinc delayed absorption since
insulin forms a stable hexamer in the presence of zinc,

0 1992 American Chemical Society and American Institute of Chemical Engineers

470

Michael R. Ladisch is Professor of Bioprocess and Agricultural Engineering and group leader of the Research and
Process Engineering Group in the Laboratory of Renewable
Resources Engineering at Purdue University. He received
his B.S. degree in chemical engineeringfrom Drexel University
in 1973 and M.S. and Ph.D. degrees in chemical engineering
from Purdue University 1974 and 1977, respectively. Dr.
Ladischs research interests include bioseparations, kinetics
of biochemical reactions, chemical reaction engineering, and
biomass conversion. He received the U. S. Presidential Young
Investigator Award in 1984 and the Van Lanen Award of the
BIOT Division of the American Chemical Society in 1990.
His work has resulted in numerous publications and several
patents. He was recently chairman of the committee on
bioprocess engineering (NRC), which studied research priorities and policy issues in bioprocess engineering.

Karen L. Kohilmann is currently a postdoctoral research

associate in the Laboratory of Renewable Resources Engineering (LORRE) at Purdue University. She obtained her
Ph.D. degree from the Food Science Department at Purdue
University in the area of food protein chemistry. Her thesis
research was in the study of the plasminogen/plasminogen
activator enzyme system in bovine milk and the relationship
of proteolytic enzymes to the gelation of milk. Prior to
graduate school she worked in the Central Research Group
of the American Home Foods division of the American Home
Products Corporation. Professionalaffiliationshave included
the American Dairy Science Association and the Institute of
Food Technologists. Her research interests lie in the development and improvement of protein separations.

with the rate of dissociation of the hexamer being the

rate-controlling step in the absorption of the insulin.
The development of recombinant DNA technologies
resulted in the production and use of the first recombinant
product, human insulin, in 1982. Then, as now, recombinant human insulin has two distinct advantages over
traditional insulin obtained from pancreatic extraction:
there is virtually an unlimited supply of recombinant

Biotechnol. Prog.., 1992, Vol. 8, No. 6

human insulin, and recombinant human insulin is identical

to human insulin. Recombinant human insulin is less
likely to cause immunological reactions during therapeutic
use than animal-derived insulin. Porcine insulin, when
administered over long periods of time, may result in
serious allergic reactions. Sequence variation in insulin
is most common a t positions 8-10 (middle of the A chain
disulfidebond), and these differencescan lead to antigenic
responses (Norman and Litwack, 1987). One impetus for
development of the recombinant human insulin process
was a perceived beef and pork shortage in the 1970swhich,
if it had actually occurred, would have restricted the
availability of the pancreatic tissue from which insulin
was extracted and, therefore, the amount of insulin
available (Hall, 1987). A t that time it was feared that the
insulin demand was increasing faster than the supply.
Americans continue to limit meat consumption, thereby
affecting the availability of pancreatic tissue.
Insulin was a particularly good choice to be the first
therapeutic protein to be produced with DNA technology,
simply because of the large amount of insulin needed. In
comparison to another recombinant product, the human
growth hormone which has a current usage of several
kilograms/year, the market for insulin is 2 orders of
magnitude higher (Prouty, 1991). Since being developed
in 1982, the amount of human insulin used has increased
dramatically. It is estimated that in the U.S. 73% of the
patients use human insulin (Crossley, 1991).

Background
Insulin Structure. Insulin is a well-defined peptide
with known amino acid sequence and structural characteristics (Watson et al., 1983; Norman and Litwack, 1987).
This hormone consists of two separate peptide chains
which are the A chain (21 amino acids) and the B chain
(30 amino acids) joined by disulfide bridges as indicated
in Figure 1. Proinsulin is the biologicalprecursor of insulin
and is a single peptide chain formed when the A and B
chains are connected by the C peptide (Figure 2). Human
and porcine insulins differ by only one amino acid, while
bovine and human insulins differ by three amino acids, as
indicated in Figure 1.
In a recent text on hormones, Norman and Litwack
(1987) detail the discovery of insulin sequence and
structure. Because of its small size, insulin has been an
ideal molecule for the study of peptide sequence and
structure. In 1955,the primary amino acid sequence was
found by F. Sanger, and in 1967, D. F. Steiner and R.
Chance determined the structure of proinsulin. Work
continued with the elucidation of the three-dimensional
structure through X-ray crystallography by D. CrowfootHodgkin in 1969. Using this technique, insulin was seen
to exist as a monomer, a dimer, or a hexamer. X-ray
crystallography remains an important technique in confirming the identities of different types of insulin. In the
structure-function relationship for insulin, three features
have been found to have been conserved: (1)the precise
positions of the three disulfide bonds; (2) the N- and
C-terminal regions of the A chain; and (3)the hydrophobic
residues of the B chain (Norman and Litwack, 1987).
Recombinant human insulin is chemically,physically,and
immunologically equivalent to pancreatic human insulin
and is biologically equivalent to both pancreatic human
insulin and purified pork insulin (Chance et al., 1981a).
Traditional Insulin Production and Purification.
Historically, insulin has been purified from animal tissue
by extraction procedures followed by chromatographic

471

Bhtechnol. Rog., 1992, Vol. 8, No. 6

Structures of Various lnsulins

I 1 I 2 13 14 [ 5 I 6 I 7 I8 19 I10 111 I12 I13 114 I15 [ 16 I17 I18 119 120 I21 122 I 23 I24 I25 I 26 I 27 I 28 I29 I30
A--Chain:
Ala-Gly-Val
Ala -Val
S

Gly-lle-Val-GluCln-Cys-Cys-Thr-Ser-lle-Cys-~r-Leu-T~-Gln-Leu-Glu-Asn-Tyr~s-Asn

Rabbil
Human

8.- Chaln:

Human

Phe-Val-Pisn-Gk-Hl~~u~~s-Gly-~~-~ls-L~~Va~Glu-~~-~eu-~~-L~uV~~~s-Gly-Glu-Ar~Cly-Ph~Ph~T~-~hr-~~Ly~-T

Rabbil
Ser

Ala

Sheep

Figure 1. Comparison of A and B chains of insulin from different sources. Reprinted with permission from Dolan-Heitlinger, J.
Recombinant DNA and Biosynthetic Human Insulin, A Source Book; Eli Lilly and Co.: Indianapolis, IN, 1982. Copyright 1982 Eli
Lilly and Co.

techniques (Dolan-Heitlinger, 1981; Barfoed, 1987). Frozen bovine or porcine pancreases are diced, extracted with
ethanol, and acidified to pH 2 with HC1 or H2S04. These
steps both inactivate and remove the enzyme trypsin,
which could degrade the insulin. Calcium carbonate is
then added to neutralize the solution, the extract is
concentrated by vacuum extraction a t low temperatures,
and the insulin is precipitated by salt addition. The
precipitate is redissolved in water and reprecipitated by
adjusting the pH to the isoelectric point of insulin.
Chromatography steps for further purification of the
insulin may includegel filtration followed by ion exchange
systems (Dolan-Heitlinger,1981). Application of insulin
to a gel filtration column gives a major peak composed of
insulin, desamido insulin(s), arginine insulin(s), insulin
ethyl ester(s), glucagon, pancreatic polypeptide, and
somatostatin. A small amount of larger molecular weight
material elutes ahead of the insulin peak and contains
proinsulin, and proinsulin intermediates, plus a mixture
of covalent insulin dimers. Proinsulin-like materials may
be immunogenic and therefore must be removed during
the purification procedure (Chance, 1972; Chance et al.,
1975). Gel filtration followed by ion exchange chromatography separates insulin from these other materials.
Production of Human Insulin. The desire to not be
restricted to animal tissue sources for insulin production
led to the interest in manufacturing human insulin.
Human insulin can be obtained by extraction from the
human pancreas, chemical synthesisfrom individualamino
acids, conversion of porcine insulin to human insulin
semisynthesis,or fermentation of genetically engineered
microorganisms (Barfoed,1987). Extraction from a human
pancreas is only feasible for research purposes, and
chemical synthesis (although it has been accomplished) is
currently not economical. The third and fourth methods
have been developed for commercial production of insulin
for therapeutic use and will be discussed.
Conversion of Porcine Insulin. Semisynthetic human insulin was commercially developed in the 1970s by
Novo Nordisk A/S and is produced by substituting the
B-30 alanine residue of porcine insulin with a threqnine
residue. Five steps result in human monocomponent
insulin (Barfoed, 1987). Insulin is first extracted from
frozen porcine pancreas glands. The second step of

SYNTHETIC
TRINUCLEOTIDE

TRANSFORM E COU

FUSION
PROTEIN

CnBr

C-PEPTIDE

,I(.:&:
ENZYMATIC

INSULIN

NH2

Figure2. Production of proinsulin in bacteria (modifiedto show

location of the C peptide). Reprinted with permission from
Watson, J. D.; et al. Recombinant DNA-A Short Course;W. H.
Freeman Co.: New York, 1983;pp 231-235. Copyright 1983 W.
H. Freeman Co.

processing entails conventional purification processes,

which have been described in the previous section. Next,

Biotechnol. Pmg., 1992, Vol. 8, No. 6

472

the purified porcine insulin is converted to human insulin

in a medium that contains only a small amount of water
and trypsin and a large quantity of organic solvent and
threonine ester. The trypsin hydrolyzes insulin a t LYSBBAlaBso, while a t the same time catalyzing the reverse
reaction in which the threonine ester displaces alanine
from position B-30 in the insulin molecule. This transpeptidation of porcine insulin to human insulin was optimized
to 97 % yield using solubletrypsin (Markussenet al., 1983).
Transpeptidation is also catalyzed by immobilizedtrypsin,
although the yield is lower a t 80% (Ueno and Morihara,
1987). This is followed by chromatographic purification
to reduce measurable levels of proinsulin and remove the
other reagents. Finally, the product is formulated and
then filled under sterile conditions, packaged, and distributed.
Insulin Produced by Recombinant Methods. Human insulin was the first animal protein to be made in
bacteria in a sequence identical to that of the human
pancreatic peptide (Watson et al., 1983). This was
accomplished by Eli Lilly and Co. (Indianapolis, IN) and
Genentech (San Francisco,CA). These companies worked
together to achieve expression of recombinant human
insulin in 1978 in Escherichia coli (E. coli) K-12 using
genes for the insulin A and B chains synthesized a t the
City of Hope National Medical Center (Chance et al.,
1981a). Scientists at Genentech cloned the genes in frame
with the @-galactosidasegene of plasmid pBR322; the
recombinant plasmid was amplified in E. coli (Chance et
al., 1981b). The first successfulexpressionwas announced
in 1978, with scale-up and approval by the appropriate
drug regulatoryagenciesachieved by 1982 (Johnson,1983).
Insulin's small size and the absence of methionine (Met)
and tryptophan (Trp) residues in the A and B chains were
critical elements in the decision to undertake the cloning
of this peptide, as well as in the achievement of rapid
development of the manufacturing process. The Met and
Trp residues produced as a consequence of engineering
and expression in E. coli are hydrolyzed by the reagents
used during the insulin recovery process. The presence
of these amino acids in insulin would have resulted in the
hydrolysis and destruction of the product.
Two-Chain Method. Although recombinant human
insulin is now produced in several ways, the first successful
method, illustrated schematically in Figure 3 (Watson et
al., 1983), was accomplished on a laboratory scale by
Genentech followed by successful scale-up by Eli Lilly
and Company (Chance et al., 1981a,b,c). Each insulin
chain was produced as a @-galactosidasefusion protein in
separate fermentations using E. coli transformed with
plasmids containing either the A or B insulin peptide
sequence. The products were intracellular and appeared
in prominent cytoplasmic inclusion bodies (Williams et
al., 1982). The method used to extract the peptides from
the inclusion bodies is proprietary information. Recombinant proteins produced in E. coli usually represent 1040% of the total protein (Burgess, 1987).
Once removed from the inclusion bodies, chemical
cleavage by CNBr a t the Met residue between the
@-galactosidase(abbreviated @-gal)and the A or B chain,
followed by purification, gave separate A and B peptides.
The peptides were then combined and induced to fold a t
a ratio of 2:l of A B chains (S-sulfonated forms) in the
presence of limited amounts of mercaptan in order to
obtain an active hormone (Chance et al., 1981c;Frank and
Chance, 1985). After 24 h, the yield was approximately
60% based on the amount of B chain used (Chance et al.,
1981a; Johnson, 1986). Goeddel et al. (1979) obtained

SYNTHETIC OLIGONUCLEOTIDESCODING
FOR A AND B CHAINS OF INSULIN
ATG

_.... .....
TGA
63 NUCLEOTIDES

ATG

TGA

90 NUCLEOTIDES

A-CHAIN PROTEIN

B-CHAIN PROTEIN

$ . / -

'L
777,

ACHAIN
.....................
(2, AMINO ACIDS)

NH2

..... ...

INSULIN

/
~

0 ACIDS)~

Figure 3. Production of insulin in bacteria by using synthetic

insulin genes. The bacteria produce hybrid proteins consisting
of the N-terminal portion of the &gal protein fused to either the
A or B chain of insulin. Reprintedwith permissionfrom Watson,
J. D.; et al. Recombinant DNA-A Short Course; W. H. Freeman
and Co.: New York, 1983; pp 231-235. Copyright 1983 W. H.
Freeman Co.

similar results with 20% of the total cellular protein

expressed as either the A or B chain fusion protein.
Subsequent folding of S-sulfonated chains gave 50-80 %
correct folding.
The large size of the @-galfusion protein limited yields
since the fusion protein of @-gal(- 1000 amino acids) and
insulin A or B chain (21 or 30 amino acids, respectively)
became detached from the cell's ribosome (premature
chain termination during translation)and thereforeyielded
incomplete insulin peptides (Burnett, 1983; Hall, 1987).
A key improvement to this approach was the use of the
tryptophan (Trp) operon in place of the lac operon (@-gal
system)to obtain a smaller fusion protein. The Trp operon
consists of a seriesof five bacterial genes which sequentially
synthesize the enzymes responsible for the anabolism of
tryptophan. One of these enzymes, Trp E, has only 190
amino acids compared to @-gal's1000 amino acids. The
Trp E gene followed by genes for the A or B chains of
insulin has the added advantage of enhancing fusion
protein production from 5-10% to 20-30% of the total
protein (Hall, 1987) since the Trp promoter is a strong
promoter in E. coli. This leads to a t least 10-fold greater

Biotechnol. Prog., 1992, Vol. 8, No. 6

475

A-chain
1004

ml
-

A-chain

191

tm

1Figure 4. Structure of insulin-codingplasmids: ApR,ampicillii

tetracycline resistance; Trp, tryptophan operon
resistance, TCR,
5'regulatory sequences. Reprinted with permissionfrom Burnett,
J. P. In Experimental Manipulation of Gene Expression;Inouye,
M., Ed.; Academic Press, Inc.: New York, 1983; pp 259-277.
Copyright 1983 Academic Press, Inc.
expression of polypeptide when compared to the lac (i.e.,
&gal) system (Burnett, 1983). The Trp operon is turned
on when the E. coli fermentation runs out of tryptophan
(Hall, 1987;Etienne-Decant, 1988). This characteristic is
beneficial during fermentation since cell mass can first be
maximized. Then, when appropriate, the cell's insulin
production system can be turned on by allowing the
fermentation media to become depleted in Trp. Since
the insulin fusion protein is an intracellular product (i.e.,
inclusion body), productivity is proportional to cell mass.
If the inclusion bodies were formed prematurely, cell
growth would stop, and therefore, total cell mass would
be lower than the maximum possible. Consequently, the
"Trp switch" is a very important practical tool in maximizing production.
The advantage of the Trp LE' system in terms of
potential productivity is clear from Figure 5. The smaller
Trp protein, like the large lac protein, is highly insoluble.
Consequently, the Trp E fusion protein also accumulates
intracellularly in the form of insoluble inclusion bodies,
which help to retard the proteolytic degradation of the
product during fermentation and initial recovery (Burnett,
1983). The structure of the insulin coding plasmid is shown
in Figure 4, and the relative molecular size of various
human insulin chimeric proteins is shown in Figure 5
(Burnett, 1983).
Details of the two-chain method of insulin production
are given in the literature. Separate fermentations of
specific recombinant E. coli strains are performed to
produce each chain with the inclusion bodies processed to
yield partially purified S-sulfonate forms. These are
subsequently used in the combination reaction (Frank and
Chance, 1983;Johnson, 1984;Johnson, 1986). Both of the
fermentations produce a chimeric protein composed of
the specific A or B chain linked through the amino acid
methionine (Met) to a Trp-LE' promoter peptide (Frank
and Chance, 1983).
After fermentation is completed, the cells are recovered
and disrupted. The cell debris is then separated from the
inclusion bodies, and the inclusion bodies are dissolved in
a solvent, although specifics are not known (Wheelwright,
1991). Inclusion bodies are sometimes dissolved in 6 M
guanidine HCl and 0.1 mM dithiothreitol (Burgess, 1987).
Next, the Trp-LE'-Met-A chain and the Trp-LE'-Met-B
chain undergo a CNBr cleavage to release the A and B

21

21

LE

proinsulin

Figure 5. Relative molecular size of various human insulin

chimeric polypeptides. The numbers indicate the number of
amino acids in each fragment. Reprinted with permission from
Burnett, J. P. In Experimental Manipulation of Gene Expression; Inouye, M., Ed.; Academic Press, Inc.: New York, 1983;pp
259-277. Copyright 1983 Academic Press, Inc.
insulin chains. Further modificationsof the A and B chains
include oxidative sulfitolysis, purification, and combination to produce crude insulin. This crude insulin is
subjected to ion exchange, size exclusion, and reversedphase high-performance liquid chromatography to produce
the purified recombinant human insulin (Frank and
Chance, 1983).
There is a recent patent (Bobbitt and Manetta, 1990)
on the chaotropic and sulfitolyticsolubilizationof inclusion
bodies of heterologous proteins. This particular process
includes sulfitolysis followed by a warming step, allowing
the protein S-sulfonate to precipitate in high purity: 95 95
pure compared to the starting material. This method seeks
to avoid improper disulfide bond formation or intermolecular cross-linking which can occur in the conventional
processing following cell lysis.
ProinsulinMethod (Intracellular). Human insulin can
also be made with recombinant microorganisms that
produce intact proinsulin instead of the A or B chains
separately. The proinsulin route is the current method of
choice for insulin production (Kroeff et al., 1989) and
entails one sequence of fermentation and purification steps
rather than two seta of sequences (i.e., one for the A chain
and one for the B chain). This approach is summarized
by Kroeff et al. (1989) with the basis of the recombinant
technology described by Watson (Watson et al., 1983),as
shown in Figure 2. Initially, mRNA is copied into cDNA,
and a methionine codon is chemically synthesized and
attached to the 5' end of the proinsulin cDNA. The cDNA
is inserted into a bacterial gene in a plasmid vector that
is introduced and then grown in E. coli. Proinsulin can
be released from the bacterial enzyme @gal) fragment
(or alternatively from the Trp-LE'-Met-proinsulin [Trp
proinsulin]) by destroying the methionine linker. The
proinsulin chain is subjected to a folding process to allow
intermolecular disulfides to form, and the C peptide can
then be cleaved with enzymes to yield human insulin
(Frank and Chance, 1983). In comparison, the two-chain
method previously described is more complex (Figure 3).
Human insulin, derived from proinsulin generated through
the Trp LE promoter, is discussed by Kroeff et al. (1989).
After recovery, the Trp proinsulin undergoes CNBr
cleavage to yield proinsulin. The proinsulin is subjected
to oxidative sulfitolysis, folding conditions in the presence
of a mercaptan, several purification steps, and then
enzymatic treatment to form the crude insulin. Ion

Biotechnol. Prog., 1992, Vol. 8, No. 6

474

exchange, reversed-phase, and size exclusion chromatography steps result in the purified recombinant human
insulin.
Proinsulin (Secreted). Villa-Komaroff etal. (1978)were
first to describe a secretion system for human proinsulin
in E. coli. Watson (1983) suggested that the ideal situation
for recombinant protein production would be to have large
amounts of the foreign protein efficiently secreted into
the medium by the bacteria. In the specificcase of insulin,
the recombinant protein could consist of j3-lactamase (an
enzyme that inactivates penicillin, which is naturally
secreted by bacteria into the culture media) and proinsulin.
Yeasts are also attractive for this type of system since
they secrete only a few of their own proteins. Therefore,
fewer extraneous proteins would need to be removed in
purification. Yeasts also are able to facilitatethe formation
of disulfide bonds. However, for glycosylated proteins,
yeasts tend to over glycosylate.
Novo Nordisk A/S used bakers yeast or Saccharomyces
cerevisiae to screte insulin as a single-chain insulin
precursor. Both the process and the product were
approved by the Danish Parliament in 1986 (Diers et al.,
1991). Diers et al. (1991) describe the unfolded peptide
as a leader or prosegment, next a Lys-Arg sequence
(recognizedby the processing enzyme), the B chain (amino
acids 1-29), a short peptide bridge, followed by the A chain
(amino acids 1-21). In this precursor, amino acid 29 of
the B chain of insulin is connected to amino acid 1of the
A chain by a short connecting peptide containing one basic
amino acid adjacent to the A chain. Human insulin is
produced through transpeptidation followed by hydrolysis
of the ester bond formed. Several chromatography steps
follow for further purification (Diers et al., 1991).
Another process for producing human proinsulin intracellularly in the yeast S. cereuisiae has recently been
described (Tottrup and Carlsen, 1990). Using this yeast
system in an optimized batch-fed fermentation, yields of
the fusion protein of superoxide dismutase-human proinsulin (SOD-PI) were reported to be 1500 mg/L. SODPI would be the starting material for the production of
recombinant human insulin; yields of the final product
have not been reported. Apparently, the expression of
heterologous polypeptides in yeast has sometimes been
lower than desirable. An Eli Lilly Co. patent (Beckage
and Ingolia, 1988) describes a process of aerobic culturing
of yeast, followed by anaerobic and then back to aerobic
conditions. This method is reported to result in a high
expression of product (produced intracellularly in S.
cerevisiae).

Analytical Separation of Insulin

Reversed-Phase High-Performance Liquid Chromatography. Reversed-phase high-performance liquid
chromatography (RP-HPLC) over alkylsilane supports
appears to be the method of choice in the analysis of insulin
(Tables I and 11) (Monch and Dehnen, 1978; Damgaard
and Markussen, 1979; OHare and Nice, 1979; Terabe et
al., 1979; Dinner and Lorenz, 1979; Kroeff and Chance,
1982; Lloyd and Corran, 1982; Rivier and McClintock,
1983;Parman and Rideout, 1983;McLeod and Wood, 1984,
Knip, 1984;Kalant et al., 1985; Smith and Venable, 1985;
Vigh, 1987). The high selectivity and high resolving
capabilities of RP-HPLC using a wide variety of stationary
and mobile phases allow the separation of insulin species
which differ by only one amino acid (Koreff et al., 1989).
The mechanism for the separation of insulin in RP-HPLC
systems is based on the hydrophobicities of insulin and
related compounds. The tendency of some proteins

(particularly large proteins) to bond so tightly to the

stationary phase that they are difficult to elute appears
to be the major limitation of the alkylsilane supports.
It is difficult to compare the analytical systems described
in Tables I and 11. Each particular system was designed
and optimized for a specific series of separations. In
general though, Smith et al. (1985) have summarized the
conditions affecting the RP-HPLC of insulin and related
compounds. Acetonitrile or methanol with various mixtures of aqueous buffers is usually used for the mobile
phase in insulin analysis. Capacity factors of insulin
decrease with increasing acetonitrile concentration up to
40% (Grego and Hearn, 1981). Since most polypeptides
and proteins strongly bind to alkylsilicasupports, chloride
salt may be substituted for phosphate salt in the mobile
phase. Most separations are carried out at ambient
temperatures, with optimal detection of insulin being
between 190 and 220 nm. All systems suffer, in varying
degrees, from interferences that are due to preservatives
or other insulin impurities (Smith et al., 1985).
With the advent of recombinant insulins, other analytical systems as well as modifications of the RP-HPLC
system have been described for the analytical separation
of insulin, proinsulin, C peptide, and insulin Aand B chains
from each other and from recombinant fusion proteins.
De Guevara (1985) isolated and quantitated the A and B
chains of insulin with a RP-HPLC system (Waters Radial
Pak with Bondapak c18 cartridge, Waters, Milford, MA)
using ion pairing with trifluoroacetic acid (TFA) as the
counterion for the A chain and formic acid for the B chain.
Kakita et al. (1981)used gel chromatography (Biogel P-30,
Bio-Rad, Richmond, CA) to separate proinsulin from the
C peptide with 1 M acetic acid as the eluting buffer.
Welinder (1984) looked a t the homogeneity of crystalline
insulin using RP-HPLC (Nucleosil 10 CIS) and highperformance gel permeation chromatography (HP-GPC)
(1-125 column, Waters). Welinder and Linde (1984)
reported the separation of insulin and insulin derivatives
using high-performance ion exchange chromatography
(HP-IEC) and gave a comparison to RP-HPLC. Ion
exchange chromatography using a 0-0.29 M NaCl salt
gradient over a DEAE-derivatized polymeric stationary
phase was recently reported by Ladisch et al. (1990). In
this ion exchange based separation, insulin was separated
from &galactosidase and from the insulin A chain.
Affinity Chromatography, Affinity chromatography
may also be used in the purification of fusion proteins. A
one-step affinity purification procedure which gives an
overallyield of 85-95 ?6 of hybrid protein with @-galactivity
using TPEG-Sepharose under high-salt conditions has
been described (Burgess, 1987). Nilsson et al. (1989)
describe a general gene fusion approach to facilitate
purification of recombinant proteins based on the fusion
of a gene of interest to a gene encoding a protein with a
strong affinity to a ligand (affinity fusion).
Smith et al. (1988) report the use of a technique called
chelating peptide immobilized metal ion affinity chromatography (CP-IMAC)to purify recombinant proinsulin.
In this method the proinsulin is expressed with a chelating
peptide on the NH2 terminus. The proinsulin can then
be affinity-purified with immobilized metal ions, and the
chelating peptide is removed chemically or enzymatically.
Electrophoresis of Insulins. Electrophoresis is a
commonly used method for detecting contaminating
proteins in a previously purified sample. However, unlike
many other proteins, the insulins present a special
challenge. Conventional electrophoretic techniques separate proteins down to molecular weights of about 1 2 OOO

Biotechnol. Rug., 1992, Vol. 8, No. 6

475

Table I. Isocratic &versed-Phase Chromatography Systems for Insulin

separation
system
comments
LiChrosorb RP-&*elution
bovine and porcine insulin
higher molecular weight
with acetonitrile in sulfuric
from monodesamido
proinsulin impurities
derivatives
retained until acetonitrile
acid buffer
concentration increased
to 26-27 %
bovine and porcine insulin
Nucleosil ODS,a-cUltrasphere
the second elution system had
ODS,'vd Hypersil ODs,".'
from monodesamido
the advantage of separating
LiChrosorb RP-18,' Spherisorb
derivatives
bovine and porcine insulin
ODS,'f and Zorbax TMSI
from human insulin
elution with (1)tartaric or acetic
acid with acetonitrile and
cetrimide and (2) acetonitrile
in acid phosphate buffer
bovine, porcine, and human
Zorbax TMS;g elution with
suggests that HPLC techniques
acetonitrile in an acid
insulin, monodesamido
can ala0 be used to determine
insulins and insulin dimers
phosphate buffer
potency of insulin crystals
and formulations
bovine, porcine, pancreatic
Vyda$ derivatized three ways:
all three columns and both
human, chicken, ovine, rabbit,
(1)C-18, (2) phenyl, (3) C-4;
buffers separated the insulins
elution with (1)TFA in
and rat insulins
acetonitrile and (2)
triethylammonium
phosphate in acetonitrile
Nucleosil 5C-1Rkelution with
hormones: (21) ACTH
separated bovine, porcine, and
analogues, (3) LH-RH
tartarate buffer-acetonitrile
human insulin with 29.2 %
analogues, and (4) insulins
containing sodium + butane
acetonitrile in the mobile phase;
sulfonate and sodium sulfate
slight changes in pH or buffer
concentration were not critical,
but consistent acetonitrile
concentration was critical
for retention times
ODS Ultrasphere;a,i isocratic
bovine, human, porcine,
use of insulin analogues allowed
comparison of separations
mouse, turkey, rodent insulins;
and shallow gradient elution
bovine proinsulins and a series
in acid phosphate buffer,
with predictions
of bovine insulin analogues
included chaotropic salts
bovine, porcine, recombinant
LiChrosorb RP-8band p
better separation of human
human, pancreatic human
Bondapak C-18;j elution with
insulin was obtained with the p
insulins; oxidized bovine
three systems: (1)ammonium
Bondapak C-18 column
sulfate acidified with
insulin A and B chains
sulfuric acid and mixed
with acetonitrile, (2)
tetramethylammonium
hydroxide, orthophosphoric
acid mixed with methanol, (3)
ethanolamine, orthophosphoric
acid mixed with acetonitrile
bovine and porcine
Nucleosil C-&kdisplacement
a proinsulin contamination target
insulin
chromatography with methanol
of 100 ppm was achieved
phosphate buffer and cetrimide
disp1acer

reference
Dinner and Lorenz, 1979

Lloyd and Corran, 1982

Kroeff and Chance, 1982

Rivier and McClintock, 1983

Terabe et al., 1979

McLeod and Wood, 1984

Kalant et al., 1985

Vigh et al., 1987

ODS = octadecylsilane. Merck, Darmstadt, FRG. Camlab, Great Britain. Anachem, Great Britain. e Shandon Southern, Great Britain.
Phase Separations, Great Britain. 8 Hitchin, Great Britain. * Separations Group, Hesparia, CA. Altex. j Waters, Milford, MA. Macherey
and Nagel, FRG.

Da, although modifications of conventional procedures

are reported to extend the separation range to Mr = 2500
(Fling and Gregerson, 1986). Proteins or peptides smaller
than this migrate and also may stain differently than larger
proteins. Insulin is a polypeptide (M,
= 5800) [A chain
M, = 2300,B chain Mr = 34001 (Sanger, 1949). Using the
techniques of Fling and Gregerson (19861,the B chain of
insulin shows an anomalously high molecular weight, while
the A chain does not stain due to either anomalous
migration, lack of banding, or inadequate fixing prior to
staining. There are clearly some difficulties involved in
the electrophoretic separation of insulin from ita component A and B chains. Insulin can, however, be separated
from proinsulin and monodesamido insulin using a polyacrylamide disc-gel electrophoresis system described by
Chance et al. (1968). In this particular system, however,
insulin and desalanine (des B-30)insulin have the same
net charge and migrate identically.

Large-Scale Purification of Human Insulin

Purification of a protein that has been overproduced in
E. coli requires a somewhat different methodology than
that used in conventional protein purification or techniques used for analytical purposes (Welinder and Linde,
1984). An exceedingly high degree of purity is required
as well as the freedom from contaminating solvents and
toxins (from E. coli) and the nutrients, metabolites,
catabolites, and host cell functional molecules which are
present as a consequence of cell growth (Johnson, 1986).
The recombinant protein must also be free of fusion protein
components. Furthermore, processing conditions must
be chosen to avoid proteolysis of hybrid proteins, especially
the larger proteins which are subject to proteolytic
degradation (Ullmann, 1984). In particular, insulin must
be separated from other forms of insulin (proinsulin,
desamido insulin, etc.) which might be immunogenic. A
large-scale purification scheme must meet all these re-

Blotechnol. Prog., 1992, Vol. 8, No. 6

476

Table 11. Gradient Reversed-Phase Chromatography Systems for Insulin

separation
system
comments
reference
ODS
Nucleosil
10-C18;avb
insulin and calibration proteins
separates proteins and
Monch and Dehnen, 1979
elution with phosphate
peptides
(cytochrome C, bovine serum
2-methoxyethanol and
albumin, aldolase, catalase,
isopropyl alcohol
chymotrypsinogen A,
2-methoxyethanol
ferritan, ovalbumin
p Bondapak C-18;celution
one of the first reports of
bovine and porcine insulin,
Damgaard and Markussen, 1979
with acetonitrile in an acid
the separation of insulin
porcine insulin from porcine
phosphate buffer
from monodesamido forms
monodesamido insulin
Hypersil ODS,d Nucleosilti-Cl8,b under optimal conditions all
O'Hare and Nice, 1979
(32) polypeptides and (9)
Spherisorb ODs,",' LiChrosorb
of the polypeptides could be
proteins
RP-lg and RP-8, and Zorbax
chromatographed; low pH
C8,g elution with acetonitrile
(<4.0) and high buffer
in an acid phosphate buffer
molarity were important
Zorbax TMS;h elution with
bovine, porcine, and human
separation at 40 O C gave
Smith et al., 1985
acetonitrile in an acid
insulins; desamido forms
better results than lower
sulfate buffer
of insulin
temperatures; HPLC
techniques and bioassay
potencies were in agreement
porcine proinsulin from
RPC-Md (ODs);"elution
a gradient system may better
Parman and Rideout, 1983
with acetonitrile or
detect spontaneous conversion
porcine insulin
methanol-acetonitrile
products in insulin and
and phosphate buffer
proinsulin preparations than
an isocratic system
Sherisorb S5 ODS2;'se elution
system useful for the separation
Knip, 1984
pancreatic peptides, insulin
with acetonitrile in an acid
of peptide hormones
and proinsulin
phosphate buffer
Macherey and Nagel, FRG. Waters, Milford, MA. Shandon, Runcorn, Great Britain. e Phase Separations,
0 ODS = octadecylsilane.
Queensferry, Great Britain. f Merck, Darmstadt, FRG. 8 Du Pont, Hitchin, Great Britain. h Du Pont Instruments, Wilmington, DE.

quirements, and in doing so, it has been estimated that

more than half of the processing costs are in the downstream purification relative to the actual fermentation
(Ullmann, 1984).
Inclusion bodies are formed in more than 80% of the
cases where proteins have been overproduced in E. coli
(Welinder and Linde, 1984). The E. coli fusion protein
8-galactosidase-insulin is no exception. Burgess (1987)
gives a procedure to release an insoluble fusion protein
involving lysis,centrifugation, denaturation, renaturation,
recipitation, and ion exchange chromatography steps.
Many proteins have been purified from insoluble inclusion
bodies by these methods of denaturation and renaturation.
More conventional purification procedures can be followed
for a secreted fusion protein such as 8-lactamase-insulin.
There are few detailed, published procedures for a
complete large-scale purification sequence for recombinant
insulin fromE. coli. However, Kroeff et al. (1989) describe
a multimodal chromatography preparative-scale system
with an integral RP-HPLC step developed to purify and
analyze recombinant human insulin produced from E. coli.
This RP-HPLC system is placed fairly late in the insulin
purification system since the majority of the impurities
(mainly from E. coli) are removed prior to this step by an
ion exchange step. A size exclusion separation follows the
reversed-phase step. The reversed-phase system is based
on a stationary phase of 10 pm Zorbax (Du Pont Instruments, Wilmington, DE) process-grade CS for the production-scale columns and 5-pm particles for the analytical
scale columns. Partially purified human insulin zinc
crystals prepared at Eli Lilly and Co. were the starting
material for this part of the purification sequence. The
insulin was applied in a water-rich mobile phase and then
eluted in a linear gradient of 0.25M acetic acid (eluent A)
to 60% aqueous acetonitrile (eluent B). An acidic mobile
phase is recommended sinceit provides excellentresolution
of insulin from structurally similar insulin-likecomponents
while promoting insulin solubility. The ideal pH is thought
to be in the region of 3.0-4.0, which is well below the
isoelectric pH of 5.4. Under mildly acidic conditions
insulin may deamidate to monodesamido insulin, but if

the RP-HPLC is done fairly rapidly (within a matter of

hours) the deamidation can be minimized. This RP-HPLC
system was successfullyused in a production-scale system
to purify recombinant insulin with a biological potency
equal to that obtained from the conventional purification
system, which employs ion exchange and size exclusion
chromatography (Kroeff et al., 1989).
Welinder and Linde (1984) investigated high-performance ion exchange chromatography (HP-IEC)for insulin
purification and compared this method to RP-HPLC for
insulin separation. HP-IEC allowed rapid fractionation
of crystalline insulin with no salt gradient. They found
HP-IEC gave good recovery and had fewer organic
contaminants than RP-HPLC, but needed to be performed
under dissociating conditions (7 M urea).

Conclusion
Insulin is a therapeutic protein which now has a history
of production using recombinant DNA methods. Improvements in genetic engineeringmethods have facilitated
both the production and the recovery of this high-value
peptide. The standard method for analysis of insulin and
insulin derivatives is RP-HPLC. Standard methods for
the purification of insulin produced by recombinant
methods include ion exchange, gel permeation, and
reversed-phase chromatography. While the current practice of these methods yields highly purified insulin, further
refinement in separation is an area of continuing research.
This reflects the increasing demand for therapeutic
proteins with minimal contamination and the need to
reduce the high cost of purification.
Notation
HCl
hydrochloric acid
CNBr
cyanogen bromide
RPreversed-phase high-performance liquid chromaHPLC tography
Met
methionine
TFA
trifluoroacetic acid
HPhigh-performancegel permeation chromatography
GPC

Bbtechnol. Rog.., 1992, Vol. 8, No. 6

HP-IEC
HzS04
Mr

high-performance ion exchange chromatography

sulfuric acid
molecular mass

Acknowledgment
The material is this work was supported b y the Purdue
University Agricultural Experiment Station, NSF Grant
BCS 8912150, and t h e Rohm and Haas Co. We t h a n k Dr.
Chuck Pidgeon (Purdue PharmacyDept.) andDr. Suzanne
Nielsen (Purdue Food Science Dept.) for their helpful
comments during review of this manuscript. We also wish
to t h a n k Dr. Ronald E. Chance,Dr. Eugene P. Kroeff, and
Dr. Walter F. Prouty (Eli Lilly and Co., Indianapolis, IN)
for helpful comments and their considerable efforts in
thoroughly reviewing this manuscript.

Literature Cited
Beckage,C. A,;Ingolia, T. D. Method, Vectors,and Transformanta
for High Expression of Heterologous Polypeptides in Yeast.
U.S.Patent 4,745,057,May 17, 1988.
Bobbitt, J. L.; Manetta, J. V. Purification and Refolding of
Recombinant Proteins. US.Patent 4,923,967,May 8, 1990.
Barfoed, H. C. Insulin Production Technology. Chem.Eng. Prog.
1987,83(lo),49-54.
Burgess, R. In Protein Engineering; Oxender, D. L., Fox, C. F.,
Eds.; Alan R. Liss, Inc.: New York, 1987;pp 71-82.
Burnett, J. P. In Experimental Manipulation of Gene Expression; Inouye, M., Ed.; Academic Press: New York, 1983;pp
259-277.
Chance, R. E. Amino Acid Sequences of Proinsulins and
Intermediates. Diabetes 1972,21(Suppl. 2), 461-467.
Chance, R. E.; Ellis, R. M.; Bromer, W. W. Porcine Proinsulin:
Characterization and Amino Acid Sequence. Science 1968,
161, 165-167.
Chance, R. E.; Root, M. A.; Galloway, J. A. In Immunological
Aspects of Diabetes Mellitus; Anderson, 0. O., Deckert, T.,
Nerup, J., Eds.; Acta Endocrinologica (Kbh.) (Suppl. 205):
Denmark, 1975;pp 185-196.
Chance,R. E.; Kroeff, E. P.; Hoffmann, J. A. InZnsulins, Growth
Hormone and Recombinant DNA Technology; Raven Press:
New York, 1981a;pp 71-85.
Chance, R. E.; Kroeff, E. P.; Hoffmann, J. A.; Frank, B. H.
Chemical, Physical, and Biologic Properties of Biosynthetic
Human Insulin. Diabetes Care 1981b,4 (2),147-154.
Chance, R. E.; Hoffmann, J. A.; Kroeff, E. P.; Johnson, M. G.;
Schmirmer, E. W.; Bromer, W. W.; Ross, M. J.; Wetzel, R. In
Peptides: Synthesis-Structure-Function;Rich, D. H., Gross,
E., Eds.; Pierce Chemical Co.: Rockford, IL, 1981c;pp 721728.
Crossley Patient Survey, 1991.
Damgaard, U.; Markussen, J. Analysis of Insulins and Related
Compounds by HPLC. Horm. Metab. Res. 1979,11,580-581.
De Guevara, 0.L. Identification and Isolation of Human Insulin
A and B Chains by High-PerformanceLiquid Chromatography.
J. Chromatogr. 1985,349,91-98.
Diers, I. V.; Rasmussen, E.; Larsen, P. H.; Kjaersig, 1.L. In Drug
Biotechnology Regulations (ScientificBasis and Practices);
Chiu, Y. H., Gueriguian, D. L., Eds.; MarcelDekker, Inc.: New
York, 1991;pp 167-177.
Dinner, A.; Lorenz, L. High-Performance Liquid Chromatographic Determination of Bovine Insulin. Anal. Chem. 1979,
51 (ll),1872-1873.
Dolan-Heitlinger, J. In Recombinant DNA and Biosynthetic
HumanZnsulin, A Source Book; Eli Lilly and Co.: Indianapolis,
IN, 1982.
Etienne-Decant, J. In Genetic Biochemistry: From Gene to
Protein; Ellis Horwood Limited: Chichester, U.K., 1988;pp
125-127.
Fling, S. P.; Gregerson, D. S. Peptide and Protein Molecular
Weight Determination by Electrophoresis Using a HighMolarity Tris Buffer System Without Urea. Anal. Biochem.
1986,155,83-88.

477

Frank, B. H.; Chance, R. E. Two Routes for Producing Human

Insulin Utilizing Recombinant DNA Technology. Munch.
Med. Wschr. 1983,125 (Suppl. l), 514-520.
Frank, B. H.; Chance, R. E. In Quo Vadis? Therapeutic Agents
Produced by Genetic Engineering; Joyeauk, A., Leygue, G.,
Morre, M., Roncucci, R., Schmelck, P. H., Eds.; Sanofi Group:
Toulouse-Labege, France, 1985;pp 137-146.

Goeddel,D.V.;Kleid,D.G.;Bolivar,F.;Heyneker,H.L.;Yansura,
D. G.;Crea, R.; Hirose, T.; Kraszewski, A.; Itakura, K.; Riggs,
A. D. Expression in Escherichia coli of Chemically Synthesized
Genes for Human Insulin. Proc. Natl. Acad. Sci. U.S.A. 1979,
76 (l),106-110.
Grego, B.; Hearn, M. T. J. Role of the Organic Solvent Modifier
in the Reversed Phase High-Peformance Liquid Chromatography of Polypeptides. Chromatographia 1981,14,589-592.
Hall, S. S. Znuisible Frontiers--TheRace to Synthesize a Human
Gene; Atlantic Monthly Press: New York, 1987.
Johnson, I. S. Human Insulin from Recombinant DNA Technology. Science 1983,219,632637,
Johnson, I. S. In Biotechnology and Biological Frontiers; The
American Association for the Advancement of Science
Publishers: Washington, DC, 1984;pp 46-56.
Johnson, R. D. The Processing of Biomacromolecules: A Challenge for the Eighties. Fluid Phase Equilib. 1986,29,109123.
Kakita, K.; Horino, M.; Tenku, A.; Nishida, S.; Mataumura, S.;
Matauki, M. Gel Chromatographic Separation of Human
C-Peptide and Proinsulin. J. Chromatogr. 1981,222,33-40.
Kalant, D.; Crawhall, J. C.; Posner, B. I. Separation of Biological
Variant Insulin Molecules from Different Speciesby ReversedPhase High-Performance Liquid Chromatography (HPLC).
Biochem. Med. 1985,34,230-240.
Knip, M. Analysis of Pancreatic Peptide Hormones by ReversedPhase High-Performance Liquid Chromatography. Horm.
Metab. Res. 1984,16,487-491.
Kroeff, E. P.; Chance, R. E. In Hormone Drugs, Proceedings of
the FDA-USP Workshop on Drug and Reference Standards
for Znsulins, Somatropins, and Thyroid-axis Hormones;U.S.
Pharmacopeial Convention,Inc.: Bethesda, MD, 1982;pp 148162.
Kroeff, E. P.; Owens, R. A.; Campbell, E. L.; Johnson, R. D.;
Marks, H. I. Production Scale Purification of Biosynthetic
Human Insulin by Reversed-Phase High-Performance Liquid
Chromatography. J. Chromatogr. 1989,461,45-61.
Ladisch, M. R.; Hendrickson, R. L.; Kohlmann, K. L. In Protein
Purification: From Molecular Mechanisms to Large Scale
Processes; ACS Symposium Series 427;American Chemical
Society: Washington, DC, 1990;pp 93-103.
Lloyd, L. F.; Corran, P. H. Analysis of Insulin Preparations by
Reversed-Phase High-Performance Liquid Chromatography.
J. Chromatogr. 1982,240,445-454.
Markussen, J.; Damgaard, U.; Pingel, M.; Snel, L.; Sorensen, A.
R.; Sorensen, E. Human Insulin (Novo): Chemistry and
Characteristics. Diabetes Care 1983,6 (Suppl. l), 231-235.
McLeod, A.; Wood, S. P. High-Performance Liquid Chromatography of Insulin. J. Chromatogr. 1984,285,319-331.
Monch, W.;Dehnen, W. J. High-Performance Liquid Chromatography of Polypeptides and Proteins on a Reversed-Phase
Support. J. Chromatogr. 1978,147,415-418.
Nilsson, B.; Anderson, S.; Uhlen, M. In Advances in Gene
Technology;Brew, K., Ed.; IRL Press Limited: McLean, VA,
1988; pp 122-123.
Norman, A. W.; Litwack, G. In Hormones;Academic Press: New
York, 1987;pp 264-317.
OHare, M. J.; Nice, E. C. Hydrophobic High-PerformanceLiquid
Chromatography of Hormonal Polypeptides and Proteins on
Alkylsilane-Bonded Silica. J . Chromatogr. 1979, 171, 209226.
Parman, A. U.; Rideout, J. M. Purification of Porcine Proinsulin
by High-Performance Liquid Chromatography. J. Chromatogr. 1983,256,283-291.
Prouty, W. F. Production Scale Purification Processes. In Drug
Biotechnology Regulations; Marcel Dekker, Inc.: New York,
1991;pp 221-262.

478

Rivier, J.; McClintock, R. Reversed-Phase High-Performance

Liquid Chromatography of Insulins from Different Species. J.
Chromatogr. 1983,268,112-119.
Sanger, F. Fractionation of Oxidized Insulin. Biochem. J. 1949,
44, 126-128.
Smith, D. J.; Venable, R. M.; Collins, J. Separation and
Quantitation of Insulins and Related Substances in Bulk
Insulin Crystals and in Injectables by Reversed-Phase High
Performance Liquid Chromatography and the Effect of
Temperature on the Separation. J. Chromatogr. Sci. 1985,
23,81-88.
Smith, M.C.; Furman, T. C.; Ingolia,T. D.; Pidgeon, C. Chelating
peptide-immobilized metal ion affinity chromatography. J.
Biol. Chem. 1988,263 (15),7211-7215.
Terabe, S.;Konaka, R.; Inouye, K. Separation of Some Polypeptide Hormones by High-Performance Liquid Chromatography.
J. Chromatogr. 1979,172,163-177.
Tottrup,H. V.; Carlsen,S.A. Process for the Production of Human
Proinsulin in Saccharomyces cerevisiae. Biotechnol. Bioeng.
1990,35,339-348.
Ueno, Y.; Morihara, K. Use of Immobilized Trypsin for Semisvnthesis of Human Insulin. Biotechnol. Bioeng. 1989,33,
126-128.
Ullmann, A. One-Step Purification of Hybrid Proteins which
have 8-Galactosidase Activity. Gene 1984,29,27-31.

Biotechnol. Prog.., 1992, Vol. 8, No. 6

Vigh, G.; Varga-Puchony, Z.; Szepesi, G.; Gazdag, M. SemiPreparative High-PerformanceReversed-Phase Displacement
Chromatography of Insulins. J. Chromatogr. 1987,386,353362.
Villa-Komaroff, L.; Efstratiadis, A.; Broome, S.; Lomedico, P.;
Tizard, R.; Naber, S. P.; Chick, W. L.; Gilbert, W. A Bacterial
Clone Synthesizing Proinsulin. Proc. Natl. Acad. Sci. U.S.A.
1978,75 (81, 3727-3731.
Watson, J. D.; Tooze, J.; Kurtz, D. T. Recombinant DNA-A
Short Course;Scientific American Books, W. H. Freeman Co.:
New York, 1983;pp 231-235.
Welinder, B. S.In CRC Handbook of HPLC for the Separation
of Amino Acids, Peptides and Proteins; Hancock, W. S., Ed.;
CRC Press: Boca Raton, FL, 1984; pp 413-419.
Welinder, B. S.;Linde, S. In CRC Handbook of HPLC for the
Separation of Amino Acids, Peptides and Proteins; Hancock,
W. S., Ed.; CRC Press: Boca Raton, FL, 1984;pp 357-362.
Wheelwright, S. M. Protein Purification; Oxford University
Press: New York, 1991;p 217.
Williams, D. C.; Van Frank, R. M.; Muth, W. L.; Burnett, J. P.
Cytoplasmic Inclusion Bodies in Escherichia coli Producing
Biosynthetic Human Insulin Proteins. Science 1982,215(5),
687-689.
Accepted July 3, 1992.
Registry No. Insulin, 9004-10-8.