POOR, pitiful evolution. Sure, over the aeons it has managed to compile a
noteworthy list of accomplishments鈥攂eetles, butterflies, birds of paradise
and Beethoven, to name a few. But when David Liu contemplates the primitive
building blocks it has to work with鈥攑roteins constructed from just 20
amino acids and DNA stitched together from an alphabet of four measly chemical
letters鈥攈e can鈥檛 help but lament what might have been. 鈥淭here are millions
of compounds it could play with and what does it settle on? Chemicals that are
rather uninteresting,鈥� jokes the Harvard University biochemist. 鈥淚magine what it
could do with some real tools.鈥�
We may soon find out. In a bold undertaking, Liu and like-minded researchers
have decided to alter the chemical building blocks of DNA, RNA and
proteins鈥攖he molecular trinity on which biology is built. By creating
nucleic acids (DNA and RNA) with an expanded genetic alphabet, and organisms
with handmade amino acids in their proteins, they hope to open up a new angle of
attack on some of science鈥檚 most intriguing questions. How did life begin? Could
evolution have chosen a different path? What alien biology might be lurking
below the surface of Mars or the moons of Jupiter?
The practical impact of extending life鈥檚 tool kit could be enormous, too.
Imagine building bacteria that can produce or destroy any chemical you care to
name. Or a protein that biologists could track on its journey through the cell.
Or DNA with all the chemical wizardry of a protein. Some researchers even think
they can teach drugs, pigments or plastics the deepest of life鈥檚
secrets鈥攈ow to adapt, reproduce and evolve into better forms. Say goodbye
to evolution. It鈥檚 time for a revolution.
Advertisement
快猫短视频s have been tinkering with the basic chemicals of life for decades,
but traditionally they have simply juggled the order of the building blocks that
nature provided. Evolutionary experts have long argued that natural selection
has chosen the optimal materials and codes, so any substantial change to these
basic elements are bound to be detrimental to the cell, if not deadly. 鈥淧eople
were convinced the amino acid set used by proteins was essentially immutable,鈥�
says Jeffrey Tze-Fei Wong from the Hong Kong University of Science and
Technology at Kowloon. So when Wong decided, in the 1980s, to test whether cells
could make use of a new amino acid, his colleagues assured him the experiment
would fail.
But Wong proved them wrong. He swapped the natural amino acid tryptophan for
a poisonous derivative called fluorotryptophan, and started feeding the toxic
brew to various microbes. Most perished, but he found one strain of the
bacterium Bacillus subtilis that grew, albeit poorly, by incorporating
low levels of the new amino acid into its proteins. Wong then introduced random
DNA mutations into the bacteria and selected those that grew slightly better, in
the hope that he could help them evolve a tolerance to the unnatural diet.
Amazingly, after a few generations, a mutant derivative he dubbed HR15 was not
only able to thrive on high levels of fluorotryptophan, but was actually
poisoned by tryptophan.
The strange finding in this obscure organism went largely unnoticed until
Andrew Ellington and his colleagues at the University of Texas at Austin
announced they had also weaned the microbiologist鈥檚 favourite bacterium,
Escherichia coli, onto fluorotryptophan (快猫短视频, 1 August
1998, p 14). Life鈥檚 ability to use new protein building blocks could no longer
be dismissed as a fluke.
These feats of what Ellington calls 鈥渦nnatural selection鈥� were encouraging,
but Liu and others have even bigger ambitions. Rather than swap one amino acid
for another, they would like to expand the protein tool kit by several amino
acids and introduce an element of design, putting the amino acids they want
exactly where they choose. They would also like the new amino acids to confer
some special property on the protein鈥攆luorescence, for instance, to make
it easy to track, or new powers to make or break chemical bonds.
To achieve that kind of control requires some re-engineering of the machinery
that transforms genetic code into protein. The process begins with the cell
making an RNA transcript of a DNA gene. This RNA message is shuttled out of the
cell鈥檚 nucleus to a protein factory called a ribosome. There, the RNA鈥檚 sequence
of chemical letters鈥攖he bases adenine, cytosine, guanine and uracil (A, C,
G and U)鈥攊s deciphered.
The translators of the code are molecules called transfer RNAs. At one end of
a tRNA molecule is a codon鈥攁 run of three bases鈥攖hat binds to the
three complementary bases on a messenger RNA molecule, so that every A pairs
with a U and C pairs with a G. Attached to the other end of each tRNA is an
amino acid鈥攁 protein building block. A particular codon encodes a
particular amino acid, so the ribosome 鈥渞eads鈥� the code by grabbing the
appropriate tRNA to match the next three-base sequence of mRNA and hooking up
its amino acid to the growing protein chain
(see Diagram).
Unusual ingredients
By making tRNA molecules that carry completely novel amino acids, Peter
Schultz of the Scripps Research Institute in La Jolla managed to generate some
unusual new proteins. He took advantage of the fact that three of the possible
64 codons鈥攖he base sequences UAG, UAA and UGA鈥攄on鈥檛 code for any
amino acid. They鈥檙e called stop codons, and normally terminate the
protein-building process. Schultz engineered a stop codon into the middle of a
gene, then created a tRNA that would read it and add a synthetic amino acid to
the growing protein chain. Using this technique in a test tube, Schultz and
other researchers have now made dozens of artificial proteins.
This system was up and running when Liu, then a graduate student in Schultz鈥檚
lab, walked into his boss鈥檚 office and asked to be given an ambitious project.
鈥淗e said `OK鈥�,鈥� remembers Liu. 鈥渀I want you to move unnatural amino acids into
living cells.鈥� My mouth just fell open.鈥� Schultz had produced the strange tRNAs
by a tricky chemical synthesis that bacteria can鈥檛 mimic. The challenge, Liu
realised, was to get the cell to create them all by itself.
Bacteria attach amino acids to tRNA using enzymes called aminoacyl-tRNA
synthetases (aaRS). So to add a new amino acid to a living cell鈥檚 repertoire
would require a new tRNA and an aaRS that could attach the new amino acid
without disrupting the loading of natural tRNAs. Rather than start from scratch,
Liu identified a natural tRNA-enzyme pair from yeast that doesn鈥檛 interact with
any tRNA or enzyme from E. coli. With a little tinkering, he persuaded
this tRNA to read an E. coli stop codon (Proceedings of the
National Academy of Sciences, vol 96, p 4780). More recently, he found
another suitable tRNA-enzyme pair from the bacterium Methanococcus
jannaschii.
Then the team started generating mutants of the yeast and M.
jannaschii that would produce new versions of the enzymes capable of
attaching novel amino acids to the tRNA. So far they鈥檝e made modest
progress鈥攁n enzyme that will insert one unnatural amino acid for every 99
natural ones. 鈥淚t鈥檚 a tricky thing switching the specificity of an enzyme,鈥� says
Thomas Magliery, a graduate student in Schultz鈥檚 lab who has taken over some of
Liu鈥檚 work. 鈥淓volution did it over millions of years and we鈥檝e only been working
on it a short time.鈥�
But even if the researchers improve the specificity of their enzyme, using
stop codons has its limits. There are only three to play with and, more
importantly, fouling their normal function would mess up the entire workings of
the cell.
One solution, devised by Masahiko Sisido and colleagues at Okayama University
in Japan, is to use a tRNA that reads a four-letter codon
(快猫短视频, 5 February, p 15).
Recently they showed that, at least in a test
tube, the strange tRNA could be used to put either of two synthetic amino acids
into a protein (Journal of the American Chemical Society, vol 121, p
12194). Researchers in Schultz鈥檚 lab are also looking for suitable tRNAs that
recognise four-letter codons to use in E. coli.
But there are also problems with four-letter codons. Even if you choose a
rare sequence, it will inevitably occur naturally somewhere in the genetic code.
This would allow the strange tRNAs to insert unnatural amino acids where they鈥檙e
not wanted. And sometimes a normal three-base tRNA might bind to the first three
letters of the four-base codon, shifting the rest of the chain out of
register. So to make proteins with pinpoint accuracy, researchers have
come up with a much more radical approach鈥� creating completely new DNA
letters.
Several researchers have tweaked existing bases to avoid disrupting the
overall structure of the DNA chain. Such modifications can upset the normal base
pairing, so they are useful for introducing mutations. But the first real
success at creating completely new bases that actually increase the coding power
of DNA came from Steve Benner, now at the University of Florida, Gainesville,
when he devised a new base pair, which he dubbed iso-C and iso
-G.
He created derivatives of the bases guanine and cytosine by switching the
chemical groups involved in forming hydrogen bonds. These bonds form when
electrons are shared between bases. They hold the DNA double helix together, and
guide the assembly of DNA by forming temporary attachments within enzymes called
polymerases. To his surprise, rather than shunning these base impostors, Benner
found that some polymerases were able to accept them.
He used his creation to make what he calls 鈥渢he 65th codon鈥�: iso-C,
A and G. Given a tRNA carrying an iso-G, the ribosome would correctly
read this codon and insert an unnatural amino acid. In fact, his imaginary base
works so well it raises the question of why it wasn鈥檛 picked by evolution. 鈥淚f
we found life on Mars using iso-C, iso-G and C and G, it
wouldn鈥檛 really surprise me,鈥� says Benner. However, for life on Earth,
iso-G does pose a slight problem. It pairs with T (the DNA version of the
RNA base U) as well as iso-C, making it a little too promiscuous to be
faithfully replicated inside a natural cell.
More recently, Floyd Romesberg at the Scripps Research Institute has found a
totally new way to expand the code, by creating bases that can pair only with
themselves鈥攁nd he has done it by playing with large greasy molecules that
barely look as though they could fit into a DNA double helix.
Back to bases
In a flurry of papers over the past year, he has announced ever more reliable
versions of these bases. He dubbed one of the earlier models PICS. Unfortunately
this base is incorporated into DNA roughly a hundred times less specifically and
rapidly than a natural base. But his latest invention, called 3MN, is
incorporated just ten times less efficiently than a natural base (Journal of
the American Chemical Society, vol 122, p 8803). Romesberg says this shows
rapid progress. 鈥淎 couple of years ago, we wouldn鈥檛 have believed we could be
that close to a natural base,鈥� he says.
And while one new base may seem like a minor accomplishment, Romesberg points
out that in a three-letter code it would allow the creation of 61 completely
novel codons for new amino acids鈥攅nough to satisfy the curiosity of most
chemists for a long time to come. But, just as Liu found problems with unnatural
tRNAs, getting real cells to accept the new bases will be the next big
challenge.
Other researchers think that the most productive way to investigate new forms
of life isn鈥檛 simply to pump up the coding power of DNA and RNA, helping them
manufacture a greater range of proteins, but to allow nucleic acids to bypass
proteins entirely. The modern genetic code allows a division of labour, where
RNA and DNA hold the code for proteins, but proteins ultimately do the cell鈥檚
chemical work. But many biologists believe that early in life鈥檚 development
there would have been no proteins. In the primordial world, so the theory goes,
something rather like RNA would have been both code and chemical powerhouse
rolled into one.
For more than a decade, biologists have been performing test-tube
versions of evolution that attempt to recreate this 鈥淩NA world鈥�. While nucleic
acids have emerged with some interesting powers, such as the ability to chop up
their neighbours, so far they have failed to develop the ability to replicate
themselves, or even the ability to build their own precursors. Bruce Eaton,
president of the company Invenux in Denver, Colorado, thinks this is because the
primitive RNAs probably had many more chemical variants than the four modern
bases. 鈥淓verything now in proteins was probably once available in RNA,鈥� he
says.
To reconstruct one of these primitive and more powerful RNAs, he added a new
chemical group called a pyridine to the RNA base uracil. Pyridines are known to
bind metals and accelerate chemical reactions. He then looked for modified RNAs
that performed a reaction to create molecules with carbon rings like those found
in bases. RNAs capable of this reaction had never been found before, but experts
think they may have played a vital role in developing life. In 1997, he reported
finding not one, but five different families of these new 鈥渟uper鈥� RNAs capable
of creating carbon rings.
Carlos Barbas at the Scripps Research Institute has pulled off a similar
trick with a DNA base, giving it the same side chain as the amino acid
histidine. This has allowed him to create one of the smallest DNA enzymes known,
a 12-base sequence that can cleave molecules of RNA (Journal of the American
Chemical Society, vol 122, p 2433). Barbas has already created DNA bases
that carry the side chains of nearly all of the 20 natural amino acids. Using
these, he thinks it should eventually be possible to create a self-sustaining
鈥渆cology鈥� of DNA molecules that work just like proteins. Throw a membrane around
that, he says, and you have something that looks very much like a living
cell.
How much further can these souped-up codes be pushed? If evolution taught DNA
and RNA to encode protein, and if biochemists can coax them to carry out
catalytic reactions, why can鈥檛 they be taught to encode鈥攁nd
evolve鈥攕omething entirely different? Liu thinks they can. His idea is to
use DNA as a template-cum-assembly platform for entirely new chemicals鈥攊n
essence, to write a new genetic code for plastic, paint or penicillin.
The scheme Liu envisions is attaching a skeleton structure for some
interesting chemical鈥攕ay, a b-lactam antibiotic, the family to which
penicillin belongs鈥攖o one strand of DNA. Then a series of tRNA-like
molecules will bind to codons in that strand, each one bringing in chemical
groups, in place of the usual amino acid, that react with and 鈥渄ecorate鈥� the
skeleton (see 鈥淎rtificial translation鈥�).FIG-mg22584701.JPG
The tRNA molecule would then fall away, leaving the newly modified drug,
still attached to the DNA, which can then be tested for its antibiotic
abilities. And because the DNA code determines exactly which tRNA modified the
skeleton, each drug can be resynthesised simply by DNA replication, attaching
the skeleton again, and putting in another mix of tRNAs. The drug can then
undergo another round of testing without first determining its
structure鈥攁n enormous time-saver for medicinal chemists. Furthermore, by
bringing in enzymes to mutate and reproduce the DNA molecules, a simplified
version of evolution can be re-enacted in a test tube. In this way, Liu hopes
that one successful drug in a million could easily be plucked out.
Eaton鈥檚 company is working on a scheme that is arguably even more ambitious.
It is also using tRNA-like molecules attached to drug synthesis reagents, but
this time the skeleton is attached to one of his modified, catalytic RNAs. In
this scheme, the RNA not only encodes the new drug, it also acts as an enzyme to
aid in its synthesis. So rounds of in vitro evolution will not only create new
molecules and enzymes, but a very simple one-step drug synthesis.
This process might mimic the early days of evolution, says Eaton. 鈥淭his is a
new way of looking at this. RNAs might have carried around the molecules they
produced.鈥� That, he speculates, may be how amino acids first became attached to
tRNAs鈥攖hey were part of the molecules that had created them.
But even as the work zooms along, researchers in this field realise that once
news gets out some people will worry about these wonder molecules or
super-bacteria escaping into the environment. 鈥淭he concern is that if we make
these superbugs, we鈥檒l soon be living on the Planet of the Apes,鈥� says
Romesberg. But he points out that unlike natural-born killers such as HIV or
Ebola virus, this brave new life鈥攆or all its unusual powers鈥攚ill be
remarkably fragile outside the safety of biology laboratories. On this planet at
least, it will always be completely dependent on chemicals that only humans can
provide. Life as we know it is not about to be outdone.