npg
© 2014 Nature America, Inc. All rights reserved.
news and views authors use a combination of static and timeresolved NMR spectroscopy and small-angle X-ray scattering (SAXS) techniques. NMR lineshape analysis indicates that microsecond timescale dynamics are present along the CARDRING interface. NMR spin-relaxation experiments reveal motions of the ubiquitin activating (UBA) domain on the millisecond timescale. Using SAXS, the authors reveal the relationship of these motions to global conformation, with constitutively open and dimerization-deficient mutants allowing models of the closed, open and dimeric enzymatic states to be determined. Additionally, the CARD, in contrast to the other domains, is found to be highly mobile in the closed state, as determined by biochemical and NMR experiments. Finally, the incorporation of time-resolved SAXS measurements and stopped-flow injection enables separation and measurement of the various transition rates, thus demonstrating that opening of cIAP1 is the ratelimiting step for enzyme activation (Fig. 1b). The conformational change in cIAP1 before dimerization is rivaled in magnitude and rate by only a handful of other biological examples9,10. cIAP1 opening exposes over 8,000 Å2 of buried surface area, spans some 20 Å and occurs at a rate of 10 s–1. The result of a transition of this nature is ideal from a functional standpoint: rapid downstream signaling is enabled while rigid control is simultaneously maintained over a cellular process as important as apoptosis. cIAP1 has effectively been tuned to function as a binary switch through a
complex dynamic release mechanism to drive enzymatic function. This finding hints at the intriguing possibility of regulating this process via a small-molecule interaction. The discovery by Fairbrother and coworkers1 of a mechanism with such an intricate, multistage level of control over biological states as that of cIAP1 activation was enabled by a powerful and diverse arsenal of biophysical methods. These results reinforce three important observations relating to the biophysical study of macromolecules. First, the most functionally interesting aspects of biology often involve heterogeneous systems and complex motions, which can be masked or invisible within the confines of a crystal lattice. Although crystallographic models have undoubtedly provided a breadth of information about biological systems, a sophisticated understanding of enzymology requires dynamic information. Second, a complete understanding of biological motions requires measurements that are both large in scope and high in resolution on both temporal and spatial scales. In many cases, this necessitates the use of multiple complementary techniques for studying dynamics, as was exquisitely demonstrated for cIAP1. For example, NMR spin-relaxation experiments enable quantitative determination of conformational dynamics on timescales spanning over 12 orders of magnitude with atomic resolution, whereas time-resolved SAXS is more efficient than NMR at determining enzyme global conformation in the presence of multiple
states. Finally, the work supports the imperative to draw on the complete array of techniques available as well as to both improve existing methods and develop new ones. The desire for a mechanistic understanding of biology outstrips the limits of current biophysical tools. To enable the discovery of other new conformational mechanisms and inform pharmaceutical drug development, researchers must continue to push the boundaries of these techniques. The description of the dynamic activation pathway for cIAP1 is an elegant and foreshadowing application of these concepts. The regulation of downstream signaling through dynamics is an exciting and new perspective on apoptosis, and the possibility to develop modulators acting through dynamics rather than competition or inhibition of enzymatic function may be transformative in many aspects of drug discovery. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Phillips, A.H. et al. Nat. Struct. Mol. Biol. 21, 1068–1074 (2014). 2. Hunter, A.M., LaCasse, E.C. & Korneluk, R.G. Apoptosis 12, 1543–1568 (2007). 3. Vucic, D. & Fairbrother, W.J. Clin. Cancer Res. 13, 5995–6000 (2007). 4. Vince, J.E. et al. Cell 131, 682–693 (2007). 5. Varfolomeev, E. et al. Cell 131, 669–681 (2007). 6. Mace, P.D. et al. J. Biol. Chem. 283, 31633–31640 (2008). 7. Das, R. et al. EMBO J. 32, 2504–2516 (2013). 8. Dueber, E.C. et al. Science 334, 376–380 (2011). 9. Carr, C.M. & Kim, P.S. Cell 73, 823–832 (1993). 10. Krumbiegel, M., Herrmann, A. & Blumenthal, R. Biophys. J. 67, 2355–2360 (1994).
Optimizing membrane-protein biogenesis through nonoptimal-codon usage Alexey S Morgunov & M Madan Babu Two studies provide insights into the distinct strategies used by prokaryotes and eukaryotes to pause translation in order to facilitate cotranslational targeting of membrane proteins to the translocon. rates in fine-tuning protein expression. This implies that, in some cases, choosing nonoptimal (i.e., more slowly translated) codons at specific positions within the coding region might be advantageous. For example, a ‘ramp’ of nonoptimal codons at the beginning of mRNAs slows translation early on to prevent downstream ribosome ‘traffic jams’7. In addition, patterns of conserved codon clusters (both optimal and nonoptimal) are associated with the folding patterns of the encoded polypeptides8,9. Thus, rather than evolving toward maximal decoding efficiency through genome-wide codon optimization, organisms exploit codon variation to
functionally regulate local translation rates. Now two studies—one in eLife10 and one in this issue of Nature Structural & Molecular Biology11— suggest a role for nonoptimal-codon choice in facilitating efficient cotranslational targeting of membrane proteins to the translocons that mediate their insertion into the membrane. In a process that is conserved across all cellular systems, hydrophobic nascent peptides destined for membrane insertion are delivered to the translocon by the ubiquitous signal recognition particle (SRP) system12 (Fig. 1). Synthesis of excess polypeptide before the translocon is reached precludes efficient insertion, thus
nature structural & molecular biology volume 21 number 12 december 2014
1023
Redundancy in the genetic code means that most amino acids are encoded by multiple synonymous codons, each of which can be translated into the same amino acid at different rates1–4. In line with this, the rate at which individual codons are translated has been shown to vary substantially within the same gene3,5,6, thus suggesting a functional role for varying translation Alexey S. Morgunov and M. Madan Babu are at the Medical Research Council Laboratory of Molecular Biology, Cambridge, UK. e-mail: [email protected] or [email protected]
news and views a
SRP binding
Elongation
Translocation
paused
3′
mRNA 5′
? SRP
Ribosome with nascent peptide
Membrane Translocon
E. coli
b
S. cerevisiae
c
Elongation pause
d
Elongation pause
e
Elongation pause
Elongation arrest
Charged-tRNA supply and demand Anti-SD
npg
© 2014 Nature America, Inc. All rights reserved.
5′
SDLE
Anti-SD
3′
5′
TM1
SDLE
3′
5′
REST
3′
5′
Alu
3′
TM2
SRP Alu Proteins without long hydrophobic N tails
SRP
TM1
TM1 TM1
SRP
Figure 1 Mechanisms of translational pausing during membrane-protein synthesis rely on mRNA-encoded signals in both prokaryotes and eukaryotes. (a) Signal recognition particle (SRP) binding to the nascent peptide and a decreased elongation rate (or pausing) facilitate successful targeting to the translocon. (b,c) In prokaryotes, Shine-Dalgarno–like elements (SDLE) within the coding region of the mRNA base-pair with the anti–Shine-Dalgarno (anti-SD) element in 16S rRNA of the ribosome and pause elongation to promote SRP recognition either before peptides lacking long hydrophobic N-terminal tails emerge from the ribosome exit tunnel (b) or after the first transmembrane helix (TM) has emerged (c). (d,e) In eukaryotes (Saccharomyces cerevisiae), mRNA-encoded slowdown of translation (REST) elements result in lower local translation efficiencies (d), owing to nonoptimal-codon usage, thus facilitating SRP recruitment and leading to elongation arrest (e) mediated by the Alu domain of SRP.
highlighting the importance of timely targeting and the need for SRP-mediated pausing. In eukaryotes and several Gram-positive bacteria, the SRP contains an Alu domain, which is responsible for pausing polypeptide elongation until the ribosome successfully docks onto the translocon in the membrane13. Whether an analogous mechanism operates in organisms in which the SRP lacks an Alu domain, such as the bacterium Escherichia coli, has remained unclear. Fluman et al.10 analyzed in vivo translation rates inferred from genome-wide ribosome profiling data in E. coli6 and identified ‘strategically’ placed Shine-Dalgarno–like elements (SDLEs) within the coding regions that transiently slow down translational elongation. They observed that ribosomal pausing tends to occur either before or immediately after the emergence of the first transmembrane helix of many membrane proteins (Fig. 1b,c). Additional analyses suggested that such pauses are prevalent in proteins with a second, closely spaced transmembrane segment, perhaps to allow targeting before the polypeptide emerges from the ribosome. The importance of SDLEs was highlighted by experimentally introducing an SDLE into a membrane 1024
protein, which reduced the protein’s aggregation propensity upon overexpression. Whether these effects are directly connected to improved SRP recognition and targeting remains to be examined. Nevertheless, this analysis highlights the importance of mRNA-encoded features in membrane-protein biogenesis. Pechmann et al.11 propose a conceptually similar but mechanistically distinct strategy of translational pausing for membrane-protein targeting in yeast. By analyzing their earlier experimental data set of cotranslational interactions of the SRP with nascent polypeptides14, they report that the properties of the putative signal sequences on polypeptides are not the only determinants for SRP enrichment. They identified nonoptimal codon clusters (termed REST elements, for mRNA-encoded slowdown of translation) downstream of the region encoding the SRPbinding site that might promote local pausing of translation elongation. The increased ribosomal occupancy in REST elements, as observed from ribosome profiling data5, lends support to reduced translation rates in vivo. Furthermore, removal of the REST element
from a membrane protein resulted in inefficient translocation of the protein into the ER membrane (whether this effect is directly due to inefficient SRP recognition and targeting remains to be examined). On the basis of these observations, the authors propose that such REST elements may kinetically enhance SRP recognition and thus the efficiency of translocation of membrane proteins (Fig. 1d,e). The two studies10,11 provide complementary views to a fascinating story: both in a prokaryote and in a eukaryote, strategically positioned RNA elements (SDLE or REST) can cause local pausing of translation elongation, thereby playing a part in membrane-protein biogenesis. In addition, these papers also raise important questions that await investigation. The ribosome has been described as a hub for protein quality control, being involved cotranslationally in translocation, folding and degradation of the newly synthesized proteins15. The role of mRNA-based signals in controlling these processes is only now beginning to be appreciated. Not all membrane proteins are substrates of the SRP; for example, the tail-anchored membrane proteins are targeted via the guided entry of tail-anchored proteins
volume 21 number 12 december 2014 nature structural & molecular biology
npg
© 2014 Nature America, Inc. All rights reserved.
news and views (GET) pathway, wherein the nascent polypeptide chain interacts with the Bag6 complex cotranslationally16. Do nonoptimal codon clusters play a part in the translocation of all membrane proteins, including those that are not substrates of the SRP, or in the recruitment of other biogenesis factors12,15–17? It also remains to be investigated how codon choice is constrained by structural alterations in the ribosomal exit tunnel18, requirements for mRNA localization19,20 and translation by specialized ribosomes21. A further mechanistic point to resolve is whether such translational pausing is more relevant for the pioneering round of translation and what impact it has on subsequent rounds; in other words, is there a trade-off between the initial, correct targeting of the mRNA to the translocon and slower synthesis of the protein in the subsequent rounds of translation? Is translational pausing essential in all cellular conditions? Fluman et al.10 hypothesize that slow translation may be most crucial when membrane-protein biogenesis operates at saturating capacity, to balance the requirements for appropriate expression level and quality control. Such condition-dependent optimization may be widespread. With varying cellular conditions resulting in heterogeneous demands on the translation and targeting machinery, fine-tuning through variable translation rates may become essential to switch between translational responses. If nonoptimal codon choice can influence protein quality control and conditional translation programs, a need arises for feedback mechanisms. Hence, translation and protein biogenesis should perhaps be viewed in the context of a dynamic and regulated (in space and time) supply-demand relationship with cellular tRNA pools. Although metrics for calculating speciesspecific translation efficiency are continually
being perfected2,4,7,8,22, and the roles of tRNA expression, evolution and post-transcriptional modifications are under investigation1,23–25, there is still a lot to learn about how the pieces fit together. Although nonoptimal-codon usage adds an extra layer of complexity, it holds great potential for revealing new mechanistic insights into how redundancy in the genetic code is exploited for fine-tuning gene expression. Nonoptimal-codon usage is one aspect of the many hidden codes enabled by the redundancy of the genetic code that constrains the ability of protein-coding sequences to evolve26. It adds yet another dimension to the multiobjective optimization problem (i.e., Pareto optimization) that genomes encounter during evolution27. Thus, a fascinating aspect of nonoptimalcodon usage that needs further investigation is how it varies across the phylogenetic space, including considerations such as cross-species conservation, host-pathogen interactions and horizontally transferred genes (for example, pathogenicity islands in bacteria). Importantly, recognizing whether these processes are at play in humans can contribute to the understanding of pathologies linked to synonymous mutations28. Similar analyses in biotechnologically important species could allow a finer degree of control over the expression, fidelity and quality of the synthesized proteins. The role of nonoptimal-codon usage in the fascinating complexity of the genetic code is slowly unraveling, showing that sometimes slower is better. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. ACKNOWLEDGMENTS We thank M. Hegde, B. Santhanam, T. Flock, N.S. Latysheva, J.T. Harbrecht, G. Chalancon and
M. Torrent for their feedback. This work was supported by the UK Medical Research Council (MC_U105185859; A.S.M. and M.M.B.) and The Lister Institute of Preventive Medicine (M.M.B.). M.M.B. is supported as a Lister Institute Research Prize Fellow. 1. Gingold, H. & Pilpel, Y. Mol. Syst. Biol. 7, 481 (2011). 2. Dana, A. & Tuller, T. Nucleic Acids Res. 42, 9171–9181 (2014). 3. Gardin, J. et al. eLife 3, e03735 (2014). 4. Spencer, P.S., Siller, E., Anderson, J.F. & Barral, J.M. J. Mol. Biol. 422, 328–335 (2012). 5. Ingolia, N.T., Ghaemmaghami, S., Newman, J.R. & Weissman, J.S. Science 324, 218–223 (2009). 6. Li, G.W., Oh, E. & Weissman, J.S. Nature 484, 538–541 (2012). 7. Tuller, T. et al. Cell 141, 344–354 (2010). 8. Pechmann, S. & Frydman, J. Nat. Struct. Mol. Biol. 20, 237–243 (2013). 9. Zhang, G., Hubalewska, M. & Ignatova, Z. Nat. Struct. Mol. Biol. 16, 274–280 (2009). 10. Fluman, N., Navon, S., Bibi, E. & Pilpel, Y. eLife 3, e03440 (2014). 11. Pechmann, S., Chartron, J. & Frydman, J. Nat. Struct. Mol. Biol. 21, 1100–1105 (2014). 12. Akopian, D., Shen, K., Zhang, X. & Shan, S.O. Annu. Rev. Biochem. 82, 693–721 (2013). 13. Halic, M. et al. Nature 427, 808–814 (2004). 14. del Alamo, M. et al. PLoS Biol. 9, e1001100 (2011). 15. Pechmann, S., Willmund, F. & Frydman, J. Mol. Cell 49, 411–421 (2013). 16. Hegde, R.S. & Keenan, R.J. Nat. Rev. Mol. Cell Biol. 12, 787–798 (2011). 17. Ast, T., Cohen, G. & Schuldiner, M. Cell 152, 1134–1145 (2013). 18. Amunts, A. et al. Science 343, 1485–1489 (2014). 19. Weatheritt, R.J., Gibson, T.J. & Babu, M.M. Nat. Struct. Mol. Biol. 21, 833–839 (2014). 20. Sossin, W.S. & DesGroseillers, L. Traffic 7, 1581–1589 (2006). 21. Xue, S. & Barna, M. Nat. Rev. Mol. Cell Biol. 13, 355–369 (2012). 22. Qian, W., Yang, J.R., Pearson, N.M., Maclean, C. & Zhang, J. PLoS Genet. 8, e1002603 (2012). 23. Gingold, H. et al. Cell 158, 1281–1292 (2014). 24. Novoa, E.M. & Ribas de Pouplana, L. Trends Genet. 28, 574–581 (2012). 25. Yona, A.H. et al. eLife 2, e01339 (2013). 26. Weatheritt, R.J. & Babu, M.M. Science 342, 1325–1326 (2013). 27. Shoval, O. et al. Science 336, 1157–1160 (2012). 28. Sauna, Z.E. & Kimchi-Sarfaty, C. Nat. Rev. Genet. 12, 683–691 (2011).
Putting a finger in the ring John McCullough & Wesley I Sundquist Two complementary papers demonstrate that the homologous type II transmembrane proteins LAP1 and LULL1 adopt nucleotide-free AAA+ ATPase folds and donate arginine fingers to complete the active sites of Torsin AAA+ ATPases. Activated Torsin complexes appear to function in nuclear and endoplasmic reticulum membrane-remodeling processes, including a nuclear vesiculation pathway that carries large cellular and viral cargoes from the nucleus into the cytoplasm. ATP hydrolysis to remodel DNA, RNA and proteins1–3. AAA+ ATPases typically function as hexameric rings that pull substrate biopolymers through their central pores. This versatile activity is used to extract polypeptides from complexes or membranes, to unfold aggregated proteins and to refold substrates into high-energy states that can subsequently do work.
AAA+ ATPase activities can be regulated at the level of protein localization, substrate binding, autoinhibition and ring assembly. These regulatory mechanisms also allow AAA+ ATPases to act as binary switches, much as small GTPases regulate other cellular processes by alternating between GTP ‘on’ and GDP ‘off ’ states. In the small GTPases, GTPase-activating proteins (GAPs) help regulate the nucleotide state by
nature structural & molecular biology volume 21 number 12 december 2014
1025
True to their name, ATPases associated with a variety of cellular activities (AAA+ ATPases) power and regulate important pathways throughout the cell, using the energy of John McCullough and Wesley I. Sundquist are at the Department of Biochemistry, University of Utah, Salt Lake City, Utah, USA. e-mail: [email protected]
© 2014 Nature America, Inc. All rights reserved.
news and views authors use a combination of static and timeresolved NMR spectroscopy and small-angle X-ray scattering (SAXS) techniques. NMR lineshape analysis indicates that microsecond timescale dynamics are present along the CARDRING interface. NMR spin-relaxation experiments reveal motions of the ubiquitin activating (UBA) domain on the millisecond timescale. Using SAXS, the authors reveal the relationship of these motions to global conformation, with constitutively open and dimerization-deficient mutants allowing models of the closed, open and dimeric enzymatic states to be determined. Additionally, the CARD, in contrast to the other domains, is found to be highly mobile in the closed state, as determined by biochemical and NMR experiments. Finally, the incorporation of time-resolved SAXS measurements and stopped-flow injection enables separation and measurement of the various transition rates, thus demonstrating that opening of cIAP1 is the ratelimiting step for enzyme activation (Fig. 1b). The conformational change in cIAP1 before dimerization is rivaled in magnitude and rate by only a handful of other biological examples9,10. cIAP1 opening exposes over 8,000 Å2 of buried surface area, spans some 20 Å and occurs at a rate of 10 s–1. The result of a transition of this nature is ideal from a functional standpoint: rapid downstream signaling is enabled while rigid control is simultaneously maintained over a cellular process as important as apoptosis. cIAP1 has effectively been tuned to function as a binary switch through a
complex dynamic release mechanism to drive enzymatic function. This finding hints at the intriguing possibility of regulating this process via a small-molecule interaction. The discovery by Fairbrother and coworkers1 of a mechanism with such an intricate, multistage level of control over biological states as that of cIAP1 activation was enabled by a powerful and diverse arsenal of biophysical methods. These results reinforce three important observations relating to the biophysical study of macromolecules. First, the most functionally interesting aspects of biology often involve heterogeneous systems and complex motions, which can be masked or invisible within the confines of a crystal lattice. Although crystallographic models have undoubtedly provided a breadth of information about biological systems, a sophisticated understanding of enzymology requires dynamic information. Second, a complete understanding of biological motions requires measurements that are both large in scope and high in resolution on both temporal and spatial scales. In many cases, this necessitates the use of multiple complementary techniques for studying dynamics, as was exquisitely demonstrated for cIAP1. For example, NMR spin-relaxation experiments enable quantitative determination of conformational dynamics on timescales spanning over 12 orders of magnitude with atomic resolution, whereas time-resolved SAXS is more efficient than NMR at determining enzyme global conformation in the presence of multiple
states. Finally, the work supports the imperative to draw on the complete array of techniques available as well as to both improve existing methods and develop new ones. The desire for a mechanistic understanding of biology outstrips the limits of current biophysical tools. To enable the discovery of other new conformational mechanisms and inform pharmaceutical drug development, researchers must continue to push the boundaries of these techniques. The description of the dynamic activation pathway for cIAP1 is an elegant and foreshadowing application of these concepts. The regulation of downstream signaling through dynamics is an exciting and new perspective on apoptosis, and the possibility to develop modulators acting through dynamics rather than competition or inhibition of enzymatic function may be transformative in many aspects of drug discovery. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Phillips, A.H. et al. Nat. Struct. Mol. Biol. 21, 1068–1074 (2014). 2. Hunter, A.M., LaCasse, E.C. & Korneluk, R.G. Apoptosis 12, 1543–1568 (2007). 3. Vucic, D. & Fairbrother, W.J. Clin. Cancer Res. 13, 5995–6000 (2007). 4. Vince, J.E. et al. Cell 131, 682–693 (2007). 5. Varfolomeev, E. et al. Cell 131, 669–681 (2007). 6. Mace, P.D. et al. J. Biol. Chem. 283, 31633–31640 (2008). 7. Das, R. et al. EMBO J. 32, 2504–2516 (2013). 8. Dueber, E.C. et al. Science 334, 376–380 (2011). 9. Carr, C.M. & Kim, P.S. Cell 73, 823–832 (1993). 10. Krumbiegel, M., Herrmann, A. & Blumenthal, R. Biophys. J. 67, 2355–2360 (1994).
Optimizing membrane-protein biogenesis through nonoptimal-codon usage Alexey S Morgunov & M Madan Babu Two studies provide insights into the distinct strategies used by prokaryotes and eukaryotes to pause translation in order to facilitate cotranslational targeting of membrane proteins to the translocon. rates in fine-tuning protein expression. This implies that, in some cases, choosing nonoptimal (i.e., more slowly translated) codons at specific positions within the coding region might be advantageous. For example, a ‘ramp’ of nonoptimal codons at the beginning of mRNAs slows translation early on to prevent downstream ribosome ‘traffic jams’7. In addition, patterns of conserved codon clusters (both optimal and nonoptimal) are associated with the folding patterns of the encoded polypeptides8,9. Thus, rather than evolving toward maximal decoding efficiency through genome-wide codon optimization, organisms exploit codon variation to
functionally regulate local translation rates. Now two studies—one in eLife10 and one in this issue of Nature Structural & Molecular Biology11— suggest a role for nonoptimal-codon choice in facilitating efficient cotranslational targeting of membrane proteins to the translocons that mediate their insertion into the membrane. In a process that is conserved across all cellular systems, hydrophobic nascent peptides destined for membrane insertion are delivered to the translocon by the ubiquitous signal recognition particle (SRP) system12 (Fig. 1). Synthesis of excess polypeptide before the translocon is reached precludes efficient insertion, thus
nature structural & molecular biology volume 21 number 12 december 2014
1023
Redundancy in the genetic code means that most amino acids are encoded by multiple synonymous codons, each of which can be translated into the same amino acid at different rates1–4. In line with this, the rate at which individual codons are translated has been shown to vary substantially within the same gene3,5,6, thus suggesting a functional role for varying translation Alexey S. Morgunov and M. Madan Babu are at the Medical Research Council Laboratory of Molecular Biology, Cambridge, UK. e-mail: [email protected] or [email protected]
news and views a
SRP binding
Elongation
Translocation
paused
3′
mRNA 5′
? SRP
Ribosome with nascent peptide
Membrane Translocon
E. coli
b
S. cerevisiae
c
Elongation pause
d
Elongation pause
e
Elongation pause
Elongation arrest
Charged-tRNA supply and demand Anti-SD
npg
© 2014 Nature America, Inc. All rights reserved.
5′
SDLE
Anti-SD
3′
5′
TM1
SDLE
3′
5′
REST
3′
5′
Alu
3′
TM2
SRP Alu Proteins without long hydrophobic N tails
SRP
TM1
TM1 TM1
SRP
Figure 1 Mechanisms of translational pausing during membrane-protein synthesis rely on mRNA-encoded signals in both prokaryotes and eukaryotes. (a) Signal recognition particle (SRP) binding to the nascent peptide and a decreased elongation rate (or pausing) facilitate successful targeting to the translocon. (b,c) In prokaryotes, Shine-Dalgarno–like elements (SDLE) within the coding region of the mRNA base-pair with the anti–Shine-Dalgarno (anti-SD) element in 16S rRNA of the ribosome and pause elongation to promote SRP recognition either before peptides lacking long hydrophobic N-terminal tails emerge from the ribosome exit tunnel (b) or after the first transmembrane helix (TM) has emerged (c). (d,e) In eukaryotes (Saccharomyces cerevisiae), mRNA-encoded slowdown of translation (REST) elements result in lower local translation efficiencies (d), owing to nonoptimal-codon usage, thus facilitating SRP recruitment and leading to elongation arrest (e) mediated by the Alu domain of SRP.
highlighting the importance of timely targeting and the need for SRP-mediated pausing. In eukaryotes and several Gram-positive bacteria, the SRP contains an Alu domain, which is responsible for pausing polypeptide elongation until the ribosome successfully docks onto the translocon in the membrane13. Whether an analogous mechanism operates in organisms in which the SRP lacks an Alu domain, such as the bacterium Escherichia coli, has remained unclear. Fluman et al.10 analyzed in vivo translation rates inferred from genome-wide ribosome profiling data in E. coli6 and identified ‘strategically’ placed Shine-Dalgarno–like elements (SDLEs) within the coding regions that transiently slow down translational elongation. They observed that ribosomal pausing tends to occur either before or immediately after the emergence of the first transmembrane helix of many membrane proteins (Fig. 1b,c). Additional analyses suggested that such pauses are prevalent in proteins with a second, closely spaced transmembrane segment, perhaps to allow targeting before the polypeptide emerges from the ribosome. The importance of SDLEs was highlighted by experimentally introducing an SDLE into a membrane 1024
protein, which reduced the protein’s aggregation propensity upon overexpression. Whether these effects are directly connected to improved SRP recognition and targeting remains to be examined. Nevertheless, this analysis highlights the importance of mRNA-encoded features in membrane-protein biogenesis. Pechmann et al.11 propose a conceptually similar but mechanistically distinct strategy of translational pausing for membrane-protein targeting in yeast. By analyzing their earlier experimental data set of cotranslational interactions of the SRP with nascent polypeptides14, they report that the properties of the putative signal sequences on polypeptides are not the only determinants for SRP enrichment. They identified nonoptimal codon clusters (termed REST elements, for mRNA-encoded slowdown of translation) downstream of the region encoding the SRPbinding site that might promote local pausing of translation elongation. The increased ribosomal occupancy in REST elements, as observed from ribosome profiling data5, lends support to reduced translation rates in vivo. Furthermore, removal of the REST element
from a membrane protein resulted in inefficient translocation of the protein into the ER membrane (whether this effect is directly due to inefficient SRP recognition and targeting remains to be examined). On the basis of these observations, the authors propose that such REST elements may kinetically enhance SRP recognition and thus the efficiency of translocation of membrane proteins (Fig. 1d,e). The two studies10,11 provide complementary views to a fascinating story: both in a prokaryote and in a eukaryote, strategically positioned RNA elements (SDLE or REST) can cause local pausing of translation elongation, thereby playing a part in membrane-protein biogenesis. In addition, these papers also raise important questions that await investigation. The ribosome has been described as a hub for protein quality control, being involved cotranslationally in translocation, folding and degradation of the newly synthesized proteins15. The role of mRNA-based signals in controlling these processes is only now beginning to be appreciated. Not all membrane proteins are substrates of the SRP; for example, the tail-anchored membrane proteins are targeted via the guided entry of tail-anchored proteins
volume 21 number 12 december 2014 nature structural & molecular biology
npg
© 2014 Nature America, Inc. All rights reserved.
news and views (GET) pathway, wherein the nascent polypeptide chain interacts with the Bag6 complex cotranslationally16. Do nonoptimal codon clusters play a part in the translocation of all membrane proteins, including those that are not substrates of the SRP, or in the recruitment of other biogenesis factors12,15–17? It also remains to be investigated how codon choice is constrained by structural alterations in the ribosomal exit tunnel18, requirements for mRNA localization19,20 and translation by specialized ribosomes21. A further mechanistic point to resolve is whether such translational pausing is more relevant for the pioneering round of translation and what impact it has on subsequent rounds; in other words, is there a trade-off between the initial, correct targeting of the mRNA to the translocon and slower synthesis of the protein in the subsequent rounds of translation? Is translational pausing essential in all cellular conditions? Fluman et al.10 hypothesize that slow translation may be most crucial when membrane-protein biogenesis operates at saturating capacity, to balance the requirements for appropriate expression level and quality control. Such condition-dependent optimization may be widespread. With varying cellular conditions resulting in heterogeneous demands on the translation and targeting machinery, fine-tuning through variable translation rates may become essential to switch between translational responses. If nonoptimal codon choice can influence protein quality control and conditional translation programs, a need arises for feedback mechanisms. Hence, translation and protein biogenesis should perhaps be viewed in the context of a dynamic and regulated (in space and time) supply-demand relationship with cellular tRNA pools. Although metrics for calculating speciesspecific translation efficiency are continually
being perfected2,4,7,8,22, and the roles of tRNA expression, evolution and post-transcriptional modifications are under investigation1,23–25, there is still a lot to learn about how the pieces fit together. Although nonoptimal-codon usage adds an extra layer of complexity, it holds great potential for revealing new mechanistic insights into how redundancy in the genetic code is exploited for fine-tuning gene expression. Nonoptimal-codon usage is one aspect of the many hidden codes enabled by the redundancy of the genetic code that constrains the ability of protein-coding sequences to evolve26. It adds yet another dimension to the multiobjective optimization problem (i.e., Pareto optimization) that genomes encounter during evolution27. Thus, a fascinating aspect of nonoptimalcodon usage that needs further investigation is how it varies across the phylogenetic space, including considerations such as cross-species conservation, host-pathogen interactions and horizontally transferred genes (for example, pathogenicity islands in bacteria). Importantly, recognizing whether these processes are at play in humans can contribute to the understanding of pathologies linked to synonymous mutations28. Similar analyses in biotechnologically important species could allow a finer degree of control over the expression, fidelity and quality of the synthesized proteins. The role of nonoptimal-codon usage in the fascinating complexity of the genetic code is slowly unraveling, showing that sometimes slower is better. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. ACKNOWLEDGMENTS We thank M. Hegde, B. Santhanam, T. Flock, N.S. Latysheva, J.T. Harbrecht, G. Chalancon and
M. Torrent for their feedback. This work was supported by the UK Medical Research Council (MC_U105185859; A.S.M. and M.M.B.) and The Lister Institute of Preventive Medicine (M.M.B.). M.M.B. is supported as a Lister Institute Research Prize Fellow. 1. Gingold, H. & Pilpel, Y. Mol. Syst. Biol. 7, 481 (2011). 2. Dana, A. & Tuller, T. Nucleic Acids Res. 42, 9171–9181 (2014). 3. Gardin, J. et al. eLife 3, e03735 (2014). 4. Spencer, P.S., Siller, E., Anderson, J.F. & Barral, J.M. J. Mol. Biol. 422, 328–335 (2012). 5. Ingolia, N.T., Ghaemmaghami, S., Newman, J.R. & Weissman, J.S. Science 324, 218–223 (2009). 6. Li, G.W., Oh, E. & Weissman, J.S. Nature 484, 538–541 (2012). 7. Tuller, T. et al. Cell 141, 344–354 (2010). 8. Pechmann, S. & Frydman, J. Nat. Struct. Mol. Biol. 20, 237–243 (2013). 9. Zhang, G., Hubalewska, M. & Ignatova, Z. Nat. Struct. Mol. Biol. 16, 274–280 (2009). 10. Fluman, N., Navon, S., Bibi, E. & Pilpel, Y. eLife 3, e03440 (2014). 11. Pechmann, S., Chartron, J. & Frydman, J. Nat. Struct. Mol. Biol. 21, 1100–1105 (2014). 12. Akopian, D., Shen, K., Zhang, X. & Shan, S.O. Annu. Rev. Biochem. 82, 693–721 (2013). 13. Halic, M. et al. Nature 427, 808–814 (2004). 14. del Alamo, M. et al. PLoS Biol. 9, e1001100 (2011). 15. Pechmann, S., Willmund, F. & Frydman, J. Mol. Cell 49, 411–421 (2013). 16. Hegde, R.S. & Keenan, R.J. Nat. Rev. Mol. Cell Biol. 12, 787–798 (2011). 17. Ast, T., Cohen, G. & Schuldiner, M. Cell 152, 1134–1145 (2013). 18. Amunts, A. et al. Science 343, 1485–1489 (2014). 19. Weatheritt, R.J., Gibson, T.J. & Babu, M.M. Nat. Struct. Mol. Biol. 21, 833–839 (2014). 20. Sossin, W.S. & DesGroseillers, L. Traffic 7, 1581–1589 (2006). 21. Xue, S. & Barna, M. Nat. Rev. Mol. Cell Biol. 13, 355–369 (2012). 22. Qian, W., Yang, J.R., Pearson, N.M., Maclean, C. & Zhang, J. PLoS Genet. 8, e1002603 (2012). 23. Gingold, H. et al. Cell 158, 1281–1292 (2014). 24. Novoa, E.M. & Ribas de Pouplana, L. Trends Genet. 28, 574–581 (2012). 25. Yona, A.H. et al. eLife 2, e01339 (2013). 26. Weatheritt, R.J. & Babu, M.M. Science 342, 1325–1326 (2013). 27. Shoval, O. et al. Science 336, 1157–1160 (2012). 28. Sauna, Z.E. & Kimchi-Sarfaty, C. Nat. Rev. Genet. 12, 683–691 (2011).
Putting a finger in the ring John McCullough & Wesley I Sundquist Two complementary papers demonstrate that the homologous type II transmembrane proteins LAP1 and LULL1 adopt nucleotide-free AAA+ ATPase folds and donate arginine fingers to complete the active sites of Torsin AAA+ ATPases. Activated Torsin complexes appear to function in nuclear and endoplasmic reticulum membrane-remodeling processes, including a nuclear vesiculation pathway that carries large cellular and viral cargoes from the nucleus into the cytoplasm. ATP hydrolysis to remodel DNA, RNA and proteins1–3. AAA+ ATPases typically function as hexameric rings that pull substrate biopolymers through their central pores. This versatile activity is used to extract polypeptides from complexes or membranes, to unfold aggregated proteins and to refold substrates into high-energy states that can subsequently do work.
AAA+ ATPase activities can be regulated at the level of protein localization, substrate binding, autoinhibition and ring assembly. These regulatory mechanisms also allow AAA+ ATPases to act as binary switches, much as small GTPases regulate other cellular processes by alternating between GTP ‘on’ and GDP ‘off ’ states. In the small GTPases, GTPase-activating proteins (GAPs) help regulate the nucleotide state by
nature structural & molecular biology volume 21 number 12 december 2014
1025
True to their name, ATPases associated with a variety of cellular activities (AAA+ ATPases) power and regulate important pathways throughout the cell, using the energy of John McCullough and Wesley I. Sundquist are at the Department of Biochemistry, University of Utah, Salt Lake City, Utah, USA. e-mail: [email protected]
