THE IMMUNE SYSTEM of mammals has evolved to its present remarkably efficient
state over thousands of millions of years. Antibodies are a key part of
this defence mechanism. These complex protein molecules react specifically
with single antigens, which very often are glycoproteins on the surface
of an invading virus, bacterium or other foreign organism.
Humans and other mammals can produce a vast number of different antibodies
– perhaps as many as 100 million. Now biotechnologists have developed a
method of using antibodies as enzymes. Enzymes are natural catalysts that
accelerate reactions, but living organisms produce only a limited range
of them. The new ‘artificial enzymes’ – catalytic antibodies which some
researchers nicknamed abzymes – offer tremendous advantages to the chemicals
and pharmaceuticals industries. They share the advantages of enzymes over
artificial catalysts, but can also speed up reactions for which there are
no naturally occurring enzymes. Artificial catalysts, which are often heavy
metals such as platinum or palladium, speed up many reactions but they are
scarce and expensive. Enzymes are not only more efficient and more specific
than such catalysts; they also work at ambient temperatures and atmospheric
pressure, thereby saving energy. They are also a renewable resource. Abzymes
have all these qualities, and, like enzymes, biochemists can constantly
improve them by new techniques of protein engineering.
Eventually, abzymes may outnumber the few thousand natural enzymes several
times. Enzymes are used in relatively few specialised areas of technology,
but abzymes could become the catalysts for a vast array of reactions. If
they do, one outcome will be safer and more pleasant working environments
and a more important role for biotechnology.
Advertisement
Both enzymes and antibodies are large, complex proteins in the form
of chain-like molecules. Their three-dimensional structures depend on the
exact arrangement of amino acid sub-units along their chains. Protein engineers
are now learning how these sequences of amino acids determine the twists
and turns of the molecules, which look rather like tangled balls of wool.
They are especially interested in the process of ‘molecular recognition’
and the structure of the active site – the part of the antibody that reacts
with its antigen, or, in the case of an enzyme, with its substrate (a substrate
is the specific molecule or compound with which the enzyme reacts).
These two types of molecule – enzymes and antibodies – work rather differently.
An antibody remains bound to its complementary antigen for a relatively
long period, acting as a marker that other elements of the immune system
recognise and home in on. An enzyme, on the other hand, binds only briefly
with its substrate molecule before releasing the products of its transformation
and grappling with another. Enzymes have the task of speeding up the conversion
of their substrates into the products of the reactions that they catalyse.
Linus Pauling first put forward the theory of enzyme action as long
ago as 1948. He said that enzymes work by somehow matching the high-energy,
short-lived complex that is the intermediate stage between the substrate
and the product of a reaction. This hypothetical configuration, in which
some bonds are being made and others broken, is called a transition state.
It holds the key to our understanding of how enzymes and abzymes work. Pauling’s
theory opened the way for the creation of antibodies that could work as
enzymes.
In 1969, the biochemist William Jencks of Brandeis University in Massachusetts
developed the theory further. He said that an enzyme molecule complements
the transition state of the reaction it catalyses, both structurally and
in the arrangement of its electrical charges. In other words, the enzyme
molecule acts as a template which moulds the molecule of substrate into
a new shape, that of a transition state. The transition state does not revert
to being a substrate molecule, because as the reaction goes along, the balance
of energy swings in favour of product molecules, which are at a lower energy.
Reaching the transition state is the difficult part of the reaction – in
energy terms, it is the top of the ‘hill’ between the two ‘valleys’ of substrate
and product that lie on either side (see Figure 1).
By the mid-1980s, biochemists had learnt a lot about how enzymes work,
and some of them began to realise that they could use antibodies as enzymes.
First, they would have to choose a reaction and predict the exact structure
of its transition state. Then they would make the transition state in the
laboratory. If they injected this ‘mimicked’ transition state into a mouse,
the mouse’s immune system would react by making an antibody against it.
Provided the mimicked transition state was close enough to the real one,
the scientists should be able to use the antibody to catalyse their reaction
in much the same way as an enzyme would. This, then, would be their abzyme.
The scientists could perpetuate the making of the antibody by the usual
technique of making monoclonal antibodies, which was devised by Cesar Milstein
and George Kohler at the Medical Research Council’s Laboratory of Molecular
Biology in Cambridge in 1975. This technique involves fusing the antibody-making
cells with fast-growing mouse cells grown in the laboratory, to form a hybrid
cell culture called a hybridoma. The hybridoma will continue to produce
the antibody indefinitely. Scientists could then test these antibodies to
see if they work as catalysts.
At least three research teams in the US and one in Britain are developing
abzymes in this way. The teams are led by Richard Lerner at the Research
Institute of the Scripps Clinic at La Jolla, California; by Stephen Benkovic
at Pennsylvania University; by Peter Schultz at the University of California
at Berkeley, and by Mike Blackburn and Dennis Burton at the University of
Sheffield. In fact, they all used a slightly more complicated procedure
than that of Milstein and Kohler.
The reason for this is that a mouse produces many different antibodies
to a single antigen, and only a small proportion of those antibodies produced
against antigens in the form of mimicked transition states act as catalysts.
But this does mean that researchers can select the antibody with the greatest
catalytic activity and learn the ground rules for increasing the activity
of their catalysts by protein engineering.
Scientists at the Scripps Institute and at Berkeley reported simultaneously
at the end of 1986 their first successes in making and testing abzymes.
Both groups used them to catalyse the hydrolysis of carboxylic esters. Esters
are formed from an alcohol and an acid and are the organic equivalent of
common salt (sodium chloride) which is formed from an alkali and an acid.
But unlike salt, esters can easily be converted back into their component
alcohol and acid. This reaction, hydrolysis, is brought about simply by
adding water. Chemists add a catalyst to speed up the reaction.
The Berkeley group used an existing antibody and an existing ‘mimicked’
transition state for the reaction they were studying. The scientists at
the Scripps Institute produced new antibodies. Both teams made abzymes that
speeded up the hydrolysis reaction; the Scripps group increased the pace
by about 1000 times and the Berkeley group by about 1500 times.
This was still many thousands of times less than the accelerations achieved
by natural enzymes. But since 1986 scientists have managed to improve the
performance of the most effective abzymes to the point where they will accelerate
some reactions to rates that are only about a hundredth of those achieved
by natural enzymes.
The first reactions to be catalysed by abzymes involved the breaking
of weak bonds such as those in esters and carbonates – ‘the nursery slopes
of catalytic chemistry’, as Blackburn at Sheffield puts it. Now, the research
teams are moving on to more ambitious abzymes that will break more robust
bonds, such as the extremely strong nitrogen-nitrogen triple bond. Some
researchers think that they might eventually develop abzymes as pharmaceutical
and research tools that will break peptide bonds – those that link the amino
acids that form the building blocks of proteins – with exquisite precision.
In a third generation of applications, the pharmaceuticals and chemicals
industries might eventually manufacture abzymes, not as end products in
themselves, but to catalyse reactions such as the making of carbon-oxygen
and carbon-nitrogen bonds. These are key steps in the manufacture of materials
such as paints and plastics.
But this is looking beyond what chemists can do. At present, industry
cannot use abzymes on such a large scale because the products of the reactions
they catalyse accumulate and, in turn, inhibit the reactions. Biotechnologists
will have to find ways of overcoming this problem by removing products as
fast as they are formed.
Meanwhile, Peter Schultz’s group at Berkeley has already increased the
chemical receptivity of some abzymes by building new chemical functions
into their reactive sites. This goes one better than nature by allowing
the abzymes to tackle reactions that are beyond the scope of natural antibodies.
Schultz and his colleagues added important groups of molecules, such as
those that will lose or gain electrons easily, to the active sites of abzymes.
Lerner and his group added metal atoms to active sites, thereby persuading
some abzymes to catalyse the hydrolysis of peptide bonds. Blackburn is seeking
to enhance the activity of abzymes by subtly altering the structure of the
‘mimicked’ transition states.
Other members of the four research teams are looking at ways of using
techniques of protein engineering to improve the performance of their abzymes.
Some laboratories are beginning to use a technique called site-directed
mutagenesis. This process deliberately alters known sequences of antibody
genes. In this way protein engineers can, for example, build flexible protein
loops into active sites, thereby making the abzymes more useful because
they will react with a greater variety of substrates. Although the extreme
specificity of antibodies could be an advantage in some reactions – in drugs,
for example – abzymes that will react with a greater variety of substrates
might be put to other uses, such as the removal of pollutants.
Not everyone is so enthusiastic about the future of abzymes: those who
are most cautious believe that abzymes will always work many times more
slowly than natural enzymes, even though some abzymes can enhance the rate
of a reaction by a hundred million times. The cautious also think that in
any case, events may overtake these developments. In the future scientists
may construct synthetic genes to act as the blueprints for completely new
enzymes. This point of view assumes that before long, protein engineering
will become an exact science, and researchers will fully understand the
rules according to which the sequence of amino acids in a protein molecule
determines how the protein folds, and therefore, in the case of an enzyme,
how it works.
Enthusiasts, on the other hand, believe that the timescale for learning
how to design enzymes to order will be long rather than short. They also
believe that new techniques are about to make it possible to screen millions
more antibodies for catalytic activity, and with great speed. These techniques
will accelerate the development of abzymes so much that they will at least
enjoy a long and prosperous reign before the coming of custom-built enzymes,
and may never have to abdicate at all.
Dennis Burton, who is on sabbatical leave at the Scripps Institute,
believes that abzymes will continue to hold advantages over engineered enzymes,
because protein engineers can make only comparatively crude changes in the
shape of a protein. The smallest alteration they can make in the reactive
site of an enzyme is to substitute one amino acid for another. This involves
reshaping an area some 1.5 angstroms (an angstrom is one-tenth of a nanometre)
long, whereas the reactive sites in abzymes can be modified on a scale about
one-fifth of this.
Many natural enzymes have evolved in such a way that they have become
less specific in the choice of substrates they catalyse. This has, in turn,
given organisms a greater choice in the range of things they can metabolise
for energy or food. Antibodies, on the other hand, have become more and
more specific, so that they can identify particular antigens and avoid reacting
with ‘self’ antigens, causing autoimmune diseases such as some forms of
arthritis. While abzymes that are less specific will be valuable for bulk
industrial processes, for pharmaceutical purposes they will have to be highly
specific.
Biotechnologists envisage that clinical medicine will be among the first
areas to benefit from abzymes. The group known as protease abzymes, for
example, will break peptide bonds with great precision in reactions that
natural enzymes either do not perform or do not perform selectively enough.
Two possible applications for such proteases could be in dissolving blood
clots and removing scar tissue. Researchers might also develop abzymes as
antiviral agents that will attack specific peptides in the enzymes or coat
proteins of a virus.
Chemists such as Blackburn see their industrial counterparts of the
future using abzymes increasingly for carrying out key steps in making fine
chemicals. This would be especially useful where such chemicals must have
all their molecules either ‘left-handed’ or ‘right-handed’. Enzymes or abzymes
make this specification possible because they can discriminate completely
between such molecules. And as new biopharma ceuticals come on the market
that are based on the chemical messengers called lymphokines (such as interferon)
the ability to control the ‘handedness’ of molecules will become more and
more important. Scientists are already developing natural enzymes for industrial
use in, for example, the LINK collaborative research programme between industry
and academia on biotransformations at the universities of Warwick, Kent
and Exeter. Abzymes will give them even more scope.
Until recently, the potential use of abzymes in medicine was limited
by the fact that the only way of making them was in mouse hybridomas. Abzymes
produced in this way are mouse tissue, so they are liable to be recognised
and rejected as foreign when used to treat people. But a team led by Greg
Winter at the Medical Research Council’s Laboratory for Molecular Biology
in Cambridge has developed a technique for ‘humanising’ antibodies. This
involves grafting human antibody material around the reactive site of a
mouse antibody, so that most of the antibody is human rather than mouse.
It greatly reduces the risk of rejection.
Winter and his colleague Sally Ward, also at the Cambridge laboratory,
pioneered a revolutionary technique – further developed by Lerner’s team
at the Scripps Institute with Dennis Burton and Angray Kang from Sheffield
– for making fragments of antibodies outside the body. These fragments,
called Fab (fragment antigen binding) fragments, each make up one arm of
a Y-shaped antibody molecule. Each arm has two components, a heavy and a
light chain (see Figure 2). To work as a catalyst, the abzyme must have
heavy and light chains, which are both necessary for strong bonding, but
it does not need the complete antibody molecule.
With this new technique, scientists no longer have to inject mice with
antigens each time they want to make a new antibody. Instead, they extract
the genes, first for the light and then for the heavy chains, from animal
or human antibody-producing cells. To do this they use a method called the
polymerase chain reaction (PCR), which uses enzymes to copy genes many times
over. Then they assemble the genes into ‘libraries’ of about a million genes
each for heavy and light chains, which they store in bacterial viruses
(see Figure 3). By combining the genes for light and heavy chains at random,
the scientists can make Fab fragments in enormous numbers. The genes for
the fragments are then expressed in bacteria, which means that the bacteria
develop the antibodies.
In this way, scientists can grow millions of different antibodies in
bacterial cultures. They can screen them for catalytic action many times
as quickly and easily as by using mouse hybridomas. Once they find antibodies
that have the catalytic action they need, scientists can produce them in
bacterial cultures which grow quicker and are much cheaper than cultures
of cells from mammals. It will be as easy to make human antibodies as rodent
antibodies in this way, so scientists will no longer have to ‘humanise’
them.
According to the Scripps team, this ability to screen large numbers
of antibodies means that researchers will be able to pick out the ones that
will catalyse reactions that are particularly tricky – for example, those
where the mechanism of the reaction is not well defined, or where it is
difficult to make an analogue of the transition state. They will also be
able to compare several different ways, or look for the best combination
of ways, to catalyse a chemical reaction. They see the breakthrough as important
primarily because it will enable scientists to develop and produce catalytic
antibodies much more quickly, easily and cheaply. Meanwhile, Andrew Hiatt
at the Scripps Clinic has successfully made complete antibodies in tobacco
plants, after transferring genes for heavy and light chains into two parent
plants. The antibodies made in this way appear to be normal.
Hiatt estimates that within four or five years it will be possible to
grow antibodies as cash crops and to produce them 10,000 times as cheaply
as antibodies made in hybridomas. He sees plants as the ideal medium for
growing catalytic antibodies, as well as those designed to neutralise pollutants
and others for use in medicine.
As Blackburn and Burton show, we can expect to see universities whose
immunology and organic chemistry departments interact making their own abzymes.
Two major projects within the LINK Biotransformations Programme are already
being planned in which Sheffield and other universities in Britain will
work with industrial partners to develop the scientific base of abzymes
as rapidly as possible. Now that techniques for mass-producing antibodies
cheaply are beginning to match the exquisite specificity provided by thousands
of millions of years of evolution, biotechnology could become the dominant
technology of the 21st century.
John Newell is editor of science, industry and medicine for the BBC’s
World Services.
