Download Protein synthesis: Twenty three amino acids and

Dispatch
R563
Protein synthesis: Twenty three amino acids and counting
Michael Ibba*, Constantinos Stathopoulos† and Dieter Söll†‡
The genetic code can be interpreted during translation
as 21 amino acids and three termination signals. Recent
advances at the interface of chemistry and molecular
biology are extending the genetic code to allow
assignment of new amino acids to existing codons,
providing new functional groups for protein synthesis.
Addresses: *Center for Biomolecular Recognition, Department of
Medical Biochemistry and Genetics, Laboratory B, The Panum Institute,
Blegdamsvej 3c, DK-2200, Copenhagen N, Denmark. †Departments of
Molecular Biophysics and Biochemistry, and ‡Chemistry, Yale
University, New Haven, Connecticut 06520-8114, USA.
E-mail: [email protected]
Current Biology 2001, 11:R563–R565
0960-9822/01/$ – see front matter
© 2001 Elsevier Science Ltd. All rights reserved.
From the days when the genetic code was first broken to
its near-universal validation in the genomic age, the riddle
has persisted as to why only twenty amino acids were
selected for ribosomal protein synthesis and what governed the selection of this particular set. These twenty
amino acids are phenomenally versatile building blocks,
providing proteins which permit life to flourish in a wide
variety of physically and chemically demanding habitats,
from below 0°C to over 100°C. Even though over 120
additional amino acids are found in natural proteins [1],
they are all outside the genetic code, being instead formed
by post-translational modification of the 20 canonical amino
acids of ribosomally made proteins. Thus, protein synthesis relies universally on 61 sense codons that encode 21
amino acids — including the nonstandard amino acid
selenocysteine [2] — and three termination codons. Now,
just as the integration of chemistry and molecular biology
first facilitated the breaking of the genetic code in the
1960s, the application of these same disciplines four
decades later has finally enabled the extension of the
genetic code to include new amino acids [3,4].
Chemists and molecular biologists have long sought to alter
the rules governing protein synthesis with the aim of
inserting novel amino acids into proteins. Why other amino
acids present in the cell, such as ornithine, have been
excluded from the standard repertoire of coded protein
synthesis remains a mystery. Two classes of biomolecules
are entrusted with ensuring the fidelity of translation:
transfer RNAs (tRNAs), which carry the messenger
(m)RNA-encoded amino acids to the ribosome, and aminoacyl-tRNA synthetases, which acylate — ‘charge’ — the
tRNA with the correct (cognate) amino acid. The specificity of this process is extremely high — the error frequency
is in the range 10–4–10–5 [5] — as each mischarging error
will lead to the formation of a mutant polypeptide.
The fidelity of the acylation process is ensured by
sophisticated editing mechanisms associated with many
aminoacyl-tRNA synthetases, in which incorrectly activated
or charged amino acids are removed (reviewed in [6]).
Although these editing activities are a major barrier to the
incorporation of nonstandard amino acids into proteins, it
has been known for many years that chinks can be found
in the cell’s armor. For instance, the leucine analog
5′,5′,5′-trifluoroleucine was shown to be charged onto tRNA
and incorporated into proteins by Escherichia coli [7], while
Bacillus subtilis growth could be made dependent on the
tryptophan analog 4-fluorotryptophan to the virtual exclusion of the natural amino acid [8]. Furthermore, the in vivo
production of proteins containing selenomethionine is a
key tool of modern X-ray crystallography [9].
It was recognized early on that protein synthesis is at the
interface of chemistry and biology, and attempts to deviate
from the constraints of the genetic code in vitro began with
the combined chemical/enzymatic generation of tRNAs
acylated with desired amino acid derivatives [10]. The idea
that nonsense codons are ideal signals for amino acid misincorporation mediated by suppressor tRNAs is well established [11]. When these suppressor tRNAs are chemically
or enzymatically aminoacylated they can be used for the
incorporation of synthetic amino acids at specific positions
in proteins, but only in vitro or ex vivo [12,13].
The long search for an in vivo approach to site-specific cotranslational insertion of an unusual amino acid — effectively an expansion of the cellular genetic code — has
come to a fruition in the work of Wang et al. [3]. They
engineered E. coli for in vivo incorporation of O-methyl-Ltyrosine in response to an amber (UAG) nonsense codon
positioned in-frame in the gene encoding dihydrofolate
reductase. To this end it was necessary to introduce an
orthogonal tRNA–aminoacyl-tRNA synthetase pair into
E. coli, involving an engineered aminoacyl-tRNA synthetase
that recognizes only the cognate orthogonal tRNA and a
non-standard amino acid.
To achieve the desired properties, the tRNA–aminoacyltRNA synthetase pair was imported from another organism, Methanococcus jannaschii, whose tyrosyl-tRNA synthetase
does not charge E. coli tRNAs. The tRNA used was a suppressor tRNATyr which is not charged by E. coli aminoacyltRNA synthetases. In vivo selection was used to obtain an
RNA–protein pair that interact as little as possible with
R564
Current Biology Vol 11 No 14
Figure 1
M. jannaschii tRNATyr
CUA
E. coli tRNATyr
Random mutagenesis
No
aminoacylation
M. jannaschii mutRNATyr
(Library)
CUA
Wild-type
M. jannaschii TyrRS
Negative selection
Mutagenesis of
active site residues
M. jannaschii mutRNATyr
CUA
that cannot be used by
E. coli AARSs
Positive selection
PCR-random
mutagenesis
Positive selection
Pool of mutant
M. jannaschii TyrRSs
Orthogonal
M. jannaschii mut tRNATyr
CUA
Aminoacylation
Addition of non-natural AA
Positive selection
DNA shuffling
Further selection
Orthogonal tRNA–synthetase pair
inside the cells
Current Biology
Schematic representation of the selection of an orthogonal
tRNA–aminoacyl-tRNA pair for the in vivo site-specific incorporation of
O-methyl-L-tyrosine in E. coli (see text for details). Separate selections
were first applied to obtain a pool of mutant tyrosyl-tRNA synthetase
variants and an orthogonal derivative of tRNATyrCUA (a variant of
tRNATyr where the anticodon has been changed to generate an amber
suppressor tRNA). These two components were then brought together
and used to select particular tyrosyl-tRNA synthetase mutants that
preferentially recognize O-methyl-L-tyrosine rather than tyrosine. TyrRS,
tyrosyl-tRNA synthetase; AARS, aminoacyl-tRNA synthetase; AA,
amino acid.
E. coli tRNAs or tyrosine, but show the best interaction
within the orthogonal pair and with the tyrosine analog
(Figure 1). The goal was successfully achieved: the analog
incorporation was over 99% faithful [3].
The successful generation of an orthogonal tRNA–aminoacyl-tRNA synthetase pair encompassed the use of a heterologous tRNA and synthetase and a suppressor tRNA
(principles that have also been employed successfully and
extended to eukaryotic systems by RajBhandary and colleagues [14,15]). Perhaps the key factor here, however,
was the initial abolition of the original activities of the
tRNA–synthetase by mutation, which then gave the
subsequent selection of orthogonal activities and specificities a real chance to succeed. In effect, Wang et al. [3]
ensured that they minimized the sequence space that had
to be sampled at each step of their selection procedure.
Another key point in this approach was the use of tyrosyltRNA synthetase, which does not have any known editing
activity. The principles of this approach are not specific to
a particular amino acid, and so can be extended to other
synthetase–tRNA pairs on a case-by-case basis for the
selection (from combinatorial libraries) of many
structurally diverse amino-acid analogs. It is an interesting
coincidence that tyrosyl-tRNA synthetase, the crucial
synthetase used in this study, was the first enzyme to be
subjected to structure-directed protein engineering twenty
years ago [16].
In another recent study [4], an elegant genetic system was
used to sample all E. coli codons in vivo to determine
which may be used for ‘encoding’, not only their canonical
amino acid, but also cysteine. This was based on a sensitive selection for ‘suppression’ of a mutation of the codon
for an essential cysteine in the active site of thymidylate
synthase. Mutations that caused such a phenotype were
shown to map to the valS gene, which encodes valyltRNA synthetase. The mutations were found to affect
residues in the editing site of the enzyme, so that the
mutant enzymes were less efficient at correcting the misacylated Cys–tRNAVal. Aminobutyric acid, long known to
be recognized by valyl-tRNA synthetase, was then fed to
the growth medium in the hope it would be incorporated
instead of valine. This was the case, 24% of the valines in
the E. coli proteins being replaced by aminobutyric acid.
While the observed incorporation of a significant amount
of aminobutyric acid into protein was not surprising in
light of the earlier knowledge, it was most revealing that
the reason for the genetically selected facile incorporation
was found to be in the reduced editing activity of an
aminoacyl-tRNA synthetase. In effect, the approach taken
by Döring et al. [4] makes a virtue of editing activities,
seen as one of the limitations in the selection of candidates
for orthogonal pair generation by Wang et al. [3], as a
means of selecting for novel enzymatic activities.
The immediate impact of these recent breakthroughs is
that they provide new tools for protein engineering. The
potential of such tools has long been appreciated from
in vitro studies and can now be fully exploited in vivo. The
broader finding that the amino acids of the genetic code
are mutable, as predicted by Wong [17], has wider implications for our understandings of the evolution of this aspect
of protein synthesis. As Döring et al. [4] point out, the
infiltration of the genetic code by an additional amino acid
now shows how a less stringent priomordial code could
Dispatch
have been restricted during evolution — an ancestral
valyl-tRNA synthetase could in effect settle on a particular
substrate from several choices. On the other hand, Wang
et al. [3] reveal just how amenable the genetic code is to
accommodating new amino acids, indicating how the
genetic code may have evolved from encoding a handful of
amino acids to the twenty routinely used today.
A striking example of the ability of the protein synthesis
machinery to find different ways to accommodate amino
acids can still be seen in the case of cysteine. To date, four
unrelated routes have been identified for the synthesis of
Cys-tRNA: the conventional cysteinyl-tRNA synthetase
[18], a dual-specificity prolyl-cysteinyl-tRNA synthetase
[19], an unclassified evolutionarily restricted cysteinyltRNA synthetase [20] and a non-editing valyl-tRNA synthetase mutant [4]. In all but the last case, cysteine is
specifically attached to cysteine-specific tRNA (tRNACys)
— the variety of synthesis mechanisms does not impart
any variety to the genetic code. While this raises the obvious
question as to why so many routes service the genetic code
with cysteine, to which there is currently no obvious
answer, the question also arises as to why the catalytic
promiscuity observed for cysteine has not been exploited
to expand the genetic code to include more amino acids. A
heretical answer would be that it has, and we simply have
not yet found the ingenuity with which to look for new
meanings in the genetic code.
Another increasingly observed evolutionary trend, lateral
gene transfer, would seem to provide a more credible
answer as to why the genetic code is almost universal and
confined to twenty amino acids. Rampant lateral gene
transfer has been proposed as a key event in the evolution
of the first organisms, with translation being one of the
first cellular processes to become ‘fixed’ [21]. Once translation was established, its universal use of twenty amino
acids would allow organisms to take full advantage of gene
transfer to sample a large gene pool; any organism not conforming to this universal code would thus be at a disadvantage. Such an explanation extols the potential advantages
of a universal genetic code, without in any way requiring
twenty to be a magic number, a fact now finally disproved.
Acknowledgements
The authors would like to thank Uttam RajBhandary, William Whitman, and
Carl Woese for stimulating discussions. MI is the recipient of an Investigator
Fellowship from the Alfred Benzon Foundation. Work in the authors’ laboratories was supported by grants from the Novo Nordisk Foundation (MI), the
Danish Research Agency (MI), the National Institute of General Medical Sciences (DS), National Aeronautics and Space Administration (DS), and the
Department of Energy (DS).
References
1. Uy R, Wold F: Posttranslational covalent modification of proteins.
Science 1977, 198:890-896.
2. Böck A, Forchhammer K, Heider J, Leinfelder W, Sawers G, Veprek B,
Zinoni F: Selenocysteine: the 21st amino acid. Mol Microbiol 1991,
5:515-520.
R565
3. Wang L, Brock A, Herberich B, Schultz PG: Expanding the genetic
code of Escherichia coli. Science 2001, 292:498-500.
4. Döring V, Mootz HD, Nangle LA, Hendrickson TL, de Crécy-Lagard V,
Schimmel P, Marliere P: Enlarging the amino acid set of Escherichia
coli by infiltration of the valine coding pathway. Science 2001,
292:501-504.
5. Loftfield RB, Vanderjagt D: The frequency of errors in protein
biosynthesis. Biochem J 1972, 128:1353-1356.
6. Ibba M, Söll D: The renaissance of aminoacyl-tRNA synthesis.
EMBO Rep 2001, 2:382-387.
7. Fenster ED, Anker HS: Incorporation into polypeptide and charging
on transfer ribonucleic acid of the amino acid analog
5′′,5′′,5′′-trifluoroleucine by leucine auxotrophs of Escherichia coli.
Biochemistry 1969, 8:269-274.
8. Wong JT: Membership mutation of the genetic code: loss of
fitness by tryptophan. Proc Natl Acad Sci USA 1983,
80:6303-6306.
9. Hendrickson WA, Horton JR, LeMaster DM: Selenomethionyl
proteins produced for analysis by multiwavelength anomalous
diffraction (MAD): a vehicle for direct determination of
three-dimensional structure. EMBO J 1990, 9:1665-1672.
10. Hecht SM, Alford BL, Kuroda Y, Kitano S: ‘Chemical aminoacylation’
of tRNA’s. J Biol Chem 1978, 253:4517-4520.
11. Inokuchi H, Hoben P, Yamao F, Ozeki H, Söll D. Transfer RNA
mischarging mediated by a mutant Escherichia coli
glutaminyl-tRNA synthetase. Proc Natl Acad Sci USA 1984,
81:5076-5080.
12. Noren CJ, Anthony-Cahill SJ, Griffith MC, Schultz PG: A general
method for site-specific incorporation of unnatural amino acids
into proteins. Science 1989, 244:182-188.
13. Nowak MW, Kearney PC, Sampson JR, Saks ME, Labarca CG,
Silverman SK, Zhong W, Thorson J, Abelson JN, Davidson N, et al.:
Nicotinic receptor binding site probed with unnatural amino acid
incorporation in intact cells. Science 1995, 268:439-442.
14. Drabkin HJ, Park HJ, RajBhandary UL: Amber suppression in
mammalian cells dependent upon expression of an Escherichia
coli aminoacyl-tRNA synthetase gene. Mol Cell Biol 1996,
16:907-913.
15. Kowal AK, Köhrer C, RajBhandary UL: Twenty-first aminoacyl-tRNA
synthetase-suppressor tRNA pairs for possible use in site-specific
incorporation of amino acid analogues into proteins in eukaryotes
and in eubacteria. Proc Natl Acad Sci USA 2001, 98:2268-2273.
16. Winter G, Fersht AR, Wilkinson AJ, Zoller M, Smith M: Redesigning
enzyme structure by site-directed mutagenesis: tyrosyl tRNA
synthetase and ATP binding. Nature 1982, 299:756-758.
17. Wong JT: Evolution of the genetic code. Microbiol Sci 1988,
5:174-181.
18. Li T, Graham DE, Stathopoulos C, Haney PJ, Kim HS, Vothknecht U,
Kitabatake M, Hong KW, Eggertsson G, Curnow AW, Lin W, et al.:
Cysteinyl-tRNA formation: the last puzzle of aminoacyl-tRNA
synthesis. FEBS Lett 1999, 462:302-306.
19. Stathopoulos C, Li T, Longman R, Vothknecht UC, Becker HD, Ibba
M, Söll D: One polypeptide with two aminoacyl-tRNA synthetase
activities. Science 2000, 287:479-482.
20. Fabrega C, Farrow MA, Mukhopadhyay B, de Crécy-Lagard V,
Ortiz AR, Schimmel P: An aminoacyl tRNA synthetase whose
sequence fits into neither of the two known classes. Nature 2001,
411:110-114.
21. Woese C: The universal ancestor. Proc Natl Acad Sci USA 1998,
95:6854-6859.