Genetic Code: Introducing Pyrrolysine Dispatch

Michael Ibba1 and Dieter Söll2 ATP-dependent attachment of lysine to tRNACUA, sug- gesting a mechanism by which the in-frame amber codon of mtmB1 could be translated as lysine. Monomethylamine methyltransferase of the archae- While this discovery of a new tRNA:aminoacyl-tRNA bacterium Methanosarcina barkeri contains a novel synthetase pair is unprecedented in its own right, it amino acid, pyrrolysine, encoded by the termination appears to be only the first step of a remarkable process. codon UAG. Initial studies suggest that pyrrolysine Solution of the crystal structure of monomethylamine may be co-translationally inserted during protein methyltransferase revealed that the UAG-encoded synthesis, probably by a mechanism analogous to residue is actually not lysine but pyrrolysine (Figure 1), that operating during selenocysteine incorporation. a new amino acid which — based on its structural context — appears to play a vital role in methylamine activation [4]. Taken together, these results provide Proteins derive much of their functional diversity from the tantalizing possibility that the in-frame amber three aspects of their structure: primary amino acid codon in mtmB1 is directly translated as pyrrolysine. sequence, post-translational modification of this There are two possible routes by which pyrrolysine sequence, and folding of the resulting polypeptide. could be introduced into monomethylamine methyl- Only the primary amino acid sequence is directly transferase: post-translational modification of lysine in determined by the genetic code, which allows for the the mature protein or co-translational insertion via incorporation of 20 standard amino acids during pyrrolysyl-tRNA formed by pre-translational modifica- protein synthesis. In addition, certain genetic contexts tion. While the present studies do not provide direct allow for the recoding of the termination codon UGA evidence to support either route, the finding of the as selenocysteine, giving a total of 21 naturally unusual tRNA:aminoacyl-tRNA synthetase pair needed occurring amino acids available for protein synthesis to generate an aminoacylated suppressor tRNA capable [1]. Efforts to increase the variety of protein structure of recognizing the UAG codon lends support to the and function have focused on expanding the genetic idea that pyrrolysine may be co-translationally inserted. code beyond these natural limits, to specifically This then raises the question of how pyrrolysyl- encode synthetic amino acids. To this end, compo- tRNACUA is synthesized? Again there are two possibil- nents of protein synthesis have been systematically ities, either free pyrrolysine is directly attached to tRNA redesigned, an approach which has already allowed or the pyrrolysyl moiety is synthesized by pre-transla- the site-specific incorporation of tyrosine analogues in tional modification of lysyl-tRNACUA. response to UAG (amber) termination codons [2,3]. The fact that lysine itself is a substrate for PylS Remarkably, recent work [4,5] suggests that nature in vitro would seem to argue against pyrrolysine being may have pre-empted aspects of this strategy; struc- directly attached to tRNA, given the implicit problem tural analysis of an archaeal methyltransferase has of substrate competition and subsequent errors in revealed that it contains the previously unknown protein synthesis. This problem could be circumvented amino acid pyrrolysine, which appears to be co-trans- if PylS first generates lysyl-tRNACUA and this is subse- lationally inserted during protein synthesis in response quently modified to give pyrrolysyl-tRNACUA. Compa- to a particular amber codon. rable indirect routes are used for the synthesis of The Methanosarcinaceae are among the most asparaginyl-, formylmethionyl-, glutaminyl- and seleno- metabolically versatile of the archaea, being able to cysteinyl-tRNAs from aminoacyl-tRNA precursors [8], thrive on a wide range of methanogenic substrates, including mono-, di- and trimethylamines [6]. An early Pyrrolysine step in catabolism of the methylamines is their NH activation by a specific methyltransferase, such as O 2 monomethylamine methyltransferase. Previous studies of the gene encoding Methanosarcina barkeri X OH monomethylamine methyltransferase (mtmB1) had N H revealed that it contained an in-frame amber codon O translated as lysine or a lysine derivative [7]. Examina- N tion of the region of the chromosome surrounding Lysine mtmB1 uncovered genes for a tRNA (tRNACUA) containing an amber anticodon and for a putative class II NH2 aminoacyl-tRNA synthetase (PylS). In vitro aminoacy- lation assays showed that PylS is able to catalyze the OH HN2 1 Department of Microbiology, The Ohio State University, 484 O West 12th Avenue, Columbus, Ohio 43210-1292, USA. E-mail: Current Biology [email protected] 2Departments of Molecular Biophysics and Biochemistry, and Chemistry, Yale University, New Haven, Figure 1. The structures of pyrrolysine and lysine. X indicates Connecticut 06520-8114, USA. E-mail: [email protected] a methyl, ammonium or hydroxyl group. Current Biology R465

Figure 2. Putative scheme for the co- ATP translational insertion of pyrrolysine at Lys Lys Pyl amber codons.

An amber suppressing tRNA (tRNACUA) is PylS PylB,PylC,PylD first charged with lysine by PylS to give CUA CUA Lys-tRNACUA. Lys-tRNACUA undergoes pre-translational modification to produce Elongation Pyl-tRNACUA which is then used for the CUA factor(s) translation on in-frame amber codons during ribosomal protein synthesis.

Protein Pyl Protein containing synthesis UAG-encoded pyrrolysine CUA 3’ GAU 5’ mRNA A site Ribosome Current Biology so there are numerous precedents. Srinivasan et al. replacement lowers the enzyme’s turnover number by [5] identified two additional genes of unknown over two orders of magnitude [13]. Fortunately, many function, pylB and pylC, which co-transcribed with of these questions concerning the incorporation and pylS and tRNACUA, and suggest that the products of function of pyrrolysine can be readily addressed these genes, and possibly that of another adjacent directly in Methanosarcina species, for which genetic gene, pylD, might participate in pyrrolysyl-tRNA tools enabling functional analyses have been devel- synthesis (Figure 2). oped (see for example [14]). The discovery of pyrrolysine raises all manner of The success of recent efforts to accommodate a questions, the most immediate of which concerns the variety of new amino acids within the protein synthe- distribution of this new amino acid. While mono- sis machinery [2,3,15] suggested that the genetic methylamine methyltransferase contains the only code might contain more information than the 21 example known to date, the presence of in-frame amber known amino acids [16]. The case of pyrrolysine would codons in the diverse genes for all three methanogen seem to provide a stunning confirmation of this pre- methyltransferases, and many other putative enzymes diction, and whatever the ‘where’, ‘how’ and ‘why’ of — such as methylases, transposases and various pyrrolysine incorporation, this amino acid represents open reading frames of unknown function, discussed a fascinating new addition to the natural set of protein in [6,9] — suggests the use of pyrrolysine might be building blocks. Let us hope there are many more widespread in the Methanosarcinaceae, and genome similar examples waiting to be discovered! sequence analyses indicate it might also be present in the bacterium Desulfitobacterium hafniense [5]. References These examples also raise a more fundamental 1. Commans, S. and Böck, A. (1999). Selenocysteine inserting tRNAs: question, namely how might in-frame UAG codons be an overview. FEMS Microbiol. Rev. 23, 335–351. decoded as pyrrolysine while the same codon main- 2. Wang, L., Brock, A., Herberich, B. and Schultz, P.G. (2001). Expand- ing the genetic code of Escherichia coli. Science 292, 498–500. tains its function in termination elsewhere? Seleno- 3. Wang, L., Brock, A. and Schultz, P.G. (2002). Adding L-3-(2-Naph- cysteine incorporation at in-frame termination UGA thyl)alanine to the genetic code of E. coli. J. Am. Chem. Soc. 124, codons poses a similar problem, which is overcome 1836–1837. through concerted interactions between specialized 4. Hao, B., Gong, W., Ferguson, T.K., James, C.M., Krzycki, J.A. and Chan, M.K. (2002). A novel UAG encoded residue in the structure of mRNA structures, elongation factors and accessory a methanogen methyltransferase. Science 296, 1462–1466. proteins (see for example [10], discussed in [11,12]). 5. Srinivasan. G., James, C.M. and Krzycki, J.A. (2002). Pyrrolysine Whether a similar mechanism exists for pyrrolysine is encoded by UAG in Archaea: charging of a UAG-decoding specialized tRNA. Science 296, 1459–1462. an intriguing question for future study, the answer to 6. Galagan, J.E., Nusbaum, C., Roy, A., Endrizzi, M.G., Macdonald, P., which would provide a means to assign specific FitzHugh, W., Calvo, S., Engels, R., Smirnov, S., Atnoor, D. et al. amber codons to the new amino acid. Beyond the (2002). The Genome of M. acetivorans Reveals Extensive Metabolic issues of where pyrrolysine is present and how it gets and Physiological Diversity. Genome Res. 12, 532–542. 7. James, C.M., Ferguson, T.K., Leykam, J.F. and Krzycki, J.A. (2001). there, the question of its biochemical necessity arises. The amber codon in the gene encoding the monomethylamine While the mechanism proposed by Hao et al. [4] methyltransferase isolated from Methanosarcina barkeri is trans- provides a ready explanation for its possible role in lated as a sense codon. J. Biol. Chem. 276, 34252–34258. 8. Ibba, M. and Söll, D. (2000). Aminoacyl-tRNA synthesis. Annu. Rev. mono- and dimethylamine activation, the specific func- Biochem. 69, 617–650. tionality imparted by pyrrolysine (and not found in 9. Deppenmeier. U., Johann, A., Hartsch, T., Merkl, R., Schmitz, R.A., the canonical amino acids) may not be suited for the Martinez-Arias, R., Henne, A., Wiezer, A., Bäumer, S., Jacobi, C. et al. (2002). The genome of Methanosarcina mazei: Evidence for products of other genes containing in-frame UAG lateral gene transfer between Bacteria and Archaea. J. Mol. Micro- codons. Again, selenocysteine provides an interesting biol. Biotechnol., 4, 453–461. parallel: studies of Escherichia coli formate dehydro- 10. Ringquist, S., Schneider, D., Gibson, T., Baron, C., Böck, A. and Gold, L. (1994). Recognition of the mRNA selenocysteine insertion genase showed that the conserved active site seleno- sequence by the specialized translational elongation factor SELB. cysteine is essential for formate oxidation — a cysteine Genes Dev. 8, 376–385. Dispatch R466

11. Söll, D. (1988). Genetic code: enter a new amino acid. Nature 331, 662–663. 12. Atkins, J.F. and Gesteland, R.F. (2000). The twenty-first amino acid. Nature 407, 463–465. 13.Axley, M.J., Böck, A. and Stadtman, T.C. (1991). Catalytic properties of an Escherichia coli formate dehydrogenase mutant in which sulfur replaces selenium. Proc. Natl. Acad. Sci. U.S.A. 88, 8450–8454. 14. Zhang, J.K., White, A.K., Kuettner, H.C., Boccazzi, P. and Metcalf, W.W. (2002). Directed mutagenesis and plasmid-based comple- mentation in the methanogenic archaeon Methanosarcina acetivorans C2A demonstrated by genetic analysis of proline biosynthesis. J. Bacteriol. 184, 1449–1454. 15. Döring, V., Mootz, H.D., Nangle, L.A., Hendrickson, T.L., Crécy- Lagard, V., Schimmel, P. and Marlière, P. (2001). Enlarging the amino acid set of Escherichia coli by infiltration of the valine coding pathway. Science 292, 501–504. 16. Ibba, M., Stathopoulos, C. and Söll, D. (2001). Protein synthesis: Twenty three amino acids and counting. Curr. Biol. 11, R563–R565.