P1: FDC/FGE P2: FDC
July 12, 2000 11:26 AR102 CHAP29
Annu. Rev. Biochem. 2000. 69:923–60
Copyright
c
2000 by Annual Reviews. All rights reserved
SYNTHESIS OF NATIVE PROTEINS
BY
CHEMICAL LIGATION
∗
Philip E. Dawson
1
and Stephen B. H. Kent
2
1
The Scripps Research Institute, La Jolla, California 92037; e-mail: dawson@scripps.edu;
2
Gryphon Sciences, South San Francisco, California 94080; e-mail: skent@gryphonsci.
com
Key Words chemical protein synthesis, thioester, protein, peptide, solid phase
synthesis, polymer-supported synthesis, protein engineering
■ Abstract In just a few short years, the chemical ligation of unprotected peptide
segments in aqueous solution has established itself as the most practical method for the
total synthesis of native proteins. A wide range of proteins has been prepared. These
synthetic molecules have led to the elucidation of gene function, to the discovery of
novel biology, and to the determination of new three-dimensional protein structures by
both NMR and X-ray crystallography. The facile access to novel analogs provided by
chemicalproteinsynthesishasledtooriginalinsightsintothemolecularbasisofprotein
function in a number of systems. Chemical protein synthesis has also enabled the
systematic developmentof proteins with enhanced potency and specificity as candidate
therapeutic agents.
CONTENTS
INTRODUCTION: Protein Science in the Postgenome Era ................... 924
DOMAINS: Building Blocks of the Protein World
......................... 925
CHEMICAL PROTEIN SYNTHESIS: The State of the Art in 1990
.............926
SYNTHETIC-PEPTIDE CHEMISTRY: Useful but Bounded
.................926
CHEMICAL LIGATION OF UNPROTECTED PEPTIDE SEGMENTS
.........929
NATIVE CHEMICAL LIGATION
.................................... 933
BIOCHEMICAL PEPTIDE LIGATION
................................ 935
Protein Splicing
................................................ 935
Conformationally Assisted Ligation
.................................. 938
SCOPE OF NATIVE CHEMICAL LIGATION FOR THE SYNTHESIS
OF PROTEINS
................................................. 939
∗
At the time of the invitation, as now, Stephen Kent is President and Chief Scientist at
Gryphon Sciences. Gryphon Sciences is focused on the development and sale of enhanced
protein therapeutics using chemical protein synthesis. The core technology of the company
is largely the subject matter of the chapter we have submitted.
0066-4154/00/0707-0923/$14.00
923
P1: FDC/FGE P2: FDC
July 12, 2000 11:26 AR102 CHAP29
924 DAWSON
¥
KENT
FOLDING SYNTHETIC PROTEINS ..................................940
CASE STUDIES IN THE APPLICATION OF CHEMICAL
PROTEIN SYNTHESIS
........................................... 944
Noncoded Amino Acids
.......................................... 944
Precise Covalent Modification
...................................... 945
Site-Specific Tagged Proteins
...................................... 945
Backbone Engineering
........................................... 946
Protein Medicinal Chemistry
....................................... 946
Rapid Access to Functional Gene Products
............................. 947
Structural Biology
.............................................. 947
CURRENT DEVELOPMENTS
...................................... 949
Expressed Protein Ligation
........................................ 949
Solid-Phase Protein Synthesis
...................................... 950
Membrane Proteins
............................................. 951
Glycoprotein Synthesis
........................................... 953
FUTURE DEVELOPMENTS
........................................ 954
Ligation Sites
.................................................. 954
Size of Protein Targets
........................................... 955
Chemical Synthesis of Peptide Segments
.............................. 956
SUMMARY AND CONCLUSIONS
................................... 956
INTRODUCTION: Protein Science in the Postgenome Era
An important current objective in biomedical research is to understand the molecu-
larbasis ofthe numerousand intricatebiological activitiesof proteins and therefore
to be able to predict and control these activities. The importance of this goal is
dramatically increased today because of the explosive success of the genome-
sequencing projects, which have revealed hundreds of thousands of new proteins,
but only as predicted sequence data (1). For the biologist, elucidation of the bi-
ological function of a predicted protein molecule is thus a challenge of great
significance. In the words of Freeman Dyson, “[In the post-genome era], proteins
will emerge as the big problem and the big opportunity. When this revolution oc-
curs, it will have a more profound effect than the Human Genome Project on the
future of science and medicine” (2).
For the past 20 years, most studies of the molecular basis of protein action have
been carried out by recombinant DNA-based expression of proteins in genetically
engineered cells (3). From its introduction, this powerful method revolutionized
the study of proteins by enabling the production of large amounts of proteins of
defined molecular composition and by allowing the systematic variation of the
amino acid sequence of proteins (4). Expression of proteins in engineered cells
is now a mature technology, and its scope and limitations are well understood:
(a) Small proteins (i.e. <30 kDa) are easier to express than large, multidomain
proteins; (b) folding of large-protein molecules can also be a challenge; (c) product
P1: FDC/FGE P2: FDC
July 12, 2000 11:26 AR102 CHAP29
CHEMICAL PROTEIN SYNTHESIS 925
heterogeneity is frequently a problem, caused by uncontrolled processing of the
nascent polypeptide in the cell; and (d) the overexpression of proteins that are toxic
to the cell, such as proteases, can be difficult (5).
Additionally, because the cell is used as a protein factory, such molecular bi-
ology studies are inherently limited to the 20 genetically encoded amino acids.
Efforts have been made to use cell-free synthesis to expand the repertoire of ri-
bosomal synthesis to include noncoded amino acids as building blocks (6, 7).
These attempts to incorporate other amino acids have had very limited success—
obtaining adequate amounts of pure protein from the cell-free translation systems
can be a significant challenge (8), and many unnatural amino acids are simply not
compatible with ribosomal polypeptide synthesis (9).
Chemical synthesis is an attractive alternative to biological methods of protein
production. The use of synthetic chemistry promises the unlimited variation of
the covalent structure of a polypeptide chain with the objective of understanding
the molecular basis of protein function. Chemistry also promises the ability to
systematically tune the properties of a protein molecule in a completely general
fashion.
This vision was one of the prime imperatives of organic chemistry in the time of
Emil Fischer at the beginning of the 20th century. In a 1905 letter to Adolf Baeyer,
Fischer wrote, “My entire yearning is directed toward the first synthetic enzyme.
If its preparation falls into my lap with the synthesis of a natural protein material,
I will consider my mission fulfilled” (10). In the decades since then, the challenge
of applying the methods of chemistry to the study of protein action has stimulated
numerous advances in synthetic methods. Historically, these advances included
the use of novel reversible protecting groups (11), novel activation methods for the
formation of covalent bonds (12), and even polymer-supported synthesis (13), all
of which sprang from the drive to apply the science of chemistry to the study of
proteins.
DOMAINS: Building Blocks of the Protein World
Because proteins are large molecules, applying chemical synthesis to them is a
considerable challenge. Furthermore, the biological functions of proteins orig-
inate in the tertiary structure of the protein molecule—that is, in the precise
three-dimensional folded structure of the polypeptide chain. The typical protein
molecule is ∼30 kDa in size and consists of two ∼15-kDa domains (14–16);
each domain has a polypeptide chain length of ∼130 (±40) amino acids (14–16).
Protein domains are defined as autonomous units of folding and, frequently, of
function (17, 18). As such, domains are the building blocks of the protein world.
The challenge confronting the chemist is, first, the total synthesis of folded do-
mains and then the ability to stitch these domains together to build complexprotein
molecules.
P1: FDC/FGE P2: FDC
July 12, 2000 11:26 AR102 CHAP29
926 DAWSON
¥
KENT
CHEMICAL PROTEIN SYNTHESIS: The State
of the Art in 1990
Since last reviewed in this journal (19), total chemical synthesis of native pro-
teins has made a number of important contributions to biomedical research. It is
notable that the Kent laboratory at the California Institute of Technology used total
chemical synthesis based on predicted gene sequence data to carry out pioneering
studies of human immunodeficiency virus 1 (HIV-1) protease enzyme (20). The
existence of this virally encoded aspartyl proteinase had been postulated based on
an analysis of viral nucleic acid sequence data, and molecular genetic studies had
indicated that its action in processing the gag-pol polyprotein was essential to the
viral life cycle (21). For this reason, the HIV-1 protease was, early on, proposed as
an important target for drug development. The first preparations of the enzyme of
defined molecular composition were produced by chemical synthesis (22), using a
highly optimized version of stepwise solid-phase peptide synthesis (19). This work
proved that the active form of the HIV-1 protease was a homodimer consisting of
two identical 99-residue polypeptide chains, and it showed that the chemically
synthesized enzyme accurately processed the putative cleavage sites in the viral
gag-pol translation product (22).
In a strikingly important contribution, total chemical synthesis was also used to
prepare large amounts of homogeneous enzyme for the determination of the orig-
inal crystal structures of the HIV-1 protease molecule (Figure 1). The structure of
the unliganded synthetic enzyme (23) corrected a seriously flawed low-resolution
structure (24) that had been obtained by using protein derived from recombinant
expression in Escherichia coli. Even more significantly, use of chemically synthe-
sized enzyme provided the first high-resolution cocrystal structures of the HIV-1
protease molecule complexed with substrate-derived inhibitors (25–27). These
structural data were made freely available to the research community and formed
the foundation for the successful worldwide programs of structure-based drug de-
sign (28) that led to the development of the highly effective protease inhibitor class
of acquired immune deficiency syndrome therapeutic agents (29).
SYNTHETIC-PEPTIDE CHEMISTRY: Useful but Bounded
Despite successful syntheses of the HIV-1 protease (22) and of a limited number
of other proteins (30–35), at the start of the decade of the 1990s total chem-
ical synthesis, by the standard methods of peptide chemistry of even a small
protein molecule remained a daunting task, often requiring large teams and tak-
ing years to complete, with no guarantee of success. The routine, reproducible
preparation of synthetic polypeptides of defined chemical structure was limited
to products of ∼50 amino acid residues (19; Figure 2). This size limitation ap-
plied equally to synthesis by solution or by solid-phase methods, but for differing
reasons.
P1: FDC/FGE P2: FDC
July 12, 2000 11:26 AR102 CHAP29
CHEMICAL PROTEIN SYNTHESIS 927
Figure 1 Crystal structures of chemically synthesized HIV-1 protease. These were the
original high-resolution structures (23, 25–27) of this protein and guided the subsequent
drug design programs. The synthetic protein preparation used for X-ray crystallography
contained
L-α-amino-n-butyricacidresiduesinplaceofthetwoCysresiduesineachsubunit.
(Left) Molscript representation of the synthetic enzyme in complex with the substrate-
derived inhibitor MVT101 (25). (Right upper panel) 2Fo-Fc electron density map for the
side chains of the unnatural amino acids used to replace the two Cys residues in each subunit
of the synthetic enzyme (23). (Lower panel) Side chains of the
L-α-amino-n-butyric acid
residues superimposed on the mercury atoms from Cys-containing enzyme (24) that has
been crystallized in the same space group. This shows that the side chains of the unnatural
amino acid have the same conformation as the natural Cys side chains. (Adapted from
References 23 and 25).
Figure 2 Historical progress in the size of synthetically accessible polypeptides.
