npg
© 2015 Nature America, Inc. All rights reserved.
news and views (i) The mRNA encoding MRPL18 has a short upstream ORF (uORF) of nine codons, which is followed by the ‘authentic’ AUG and then by the CUG identified by the authors as the initiator of MRPL18(cyto) (Fig. 1). This arrangement is conserved in mice. The role of such uORFs in regulating translation of downstream sequences was initially identified in the classic experiments on translation of GCN4 in yeast7. Although reinitiation can occur downstream of the uORF, the site of reinitiation is influenced by the activity of translation initiation factors. Such a mechanism has also been established in mammalian cells, for mRNAs encoding proteins such as ATF4 (ref. 8). In the case of the MRPL18 mRNA, phosphorylation of eIF2α slows translational initiation, thus leading to skipping of reinitiation at the second AUG and permitting some reinitiation at the subsequent CUG to synthesize MRPL18(cyto). A key point is that eIF2 can be phosphorylated by several kinases, each responding to a different type of stress9. Thus MRPL18(cyto) may be generated in response to many types of stress beyond heat shock. The relationship between stress, slowed translation initiation and the selective initiation of translation is evident from the observation that hyperactivation of translation, by means of a constitutively activated TORC1 complex, suppresses translation of heat-shock proteins10, possibly owing to reduced production of MRPL18(cyto). (ii) The phosphorylated MRPL18(cyto) associates with 80S ribosomes to yield a ‘hybrid’ ribosome that is essential for translation of the HSP70-encoding mRNA. The relatively new concept of ‘specialized’ ribosomes, tailored for the translation of specific mRNAs, has now been demonstrated in several instances11 and
may prove to be an important element in the overall regulation of translation. (iii) It has been known for some time that HSP70-encoding mRNA is translated in a manner that is cap independent but dependent on the 5′ untranslated region (5′ UTR), perhaps through an internal ribosome entry site (IRES)12, although that has been disputed10. It now seems likely that the presence of MRPL18(cyto) permits the hybrid ribosome to bypass normal cap-dependent initiation, presumably by interacting with some structure in the 5′ UTR to effect translation. Future studies determining the location of the MRPL18(cyto) on the cytoplasmic ribosome may suggest the mechanism by which this occurs. The conceptual novelty of this story is the realization that any of the mRPs might be coopted for regulatory functions in the cytoplasm or in the nucleus. Although the unusual structure of mammalian mitochondrial ribosomes involves far more protein-protein interactions than occur in cytoplasmic ribosomes, the mRPs are still predominantly RNA-binding proteins. Thus there should be many opportunities to adapt them as regulatory elements at a variety of levels in the nuclear and cytoplasmic RNAome. Need we remind the reader that plants have an additional source of such regulators in the form of the proteins making up the chloroplast ribosome? MRPL18 is closely related to RPs whose ancestry can be traced over three billion years and which are present in all ribosomes; they have recently been given the structural name uL18 (ref. 13). uL18 is one of the two proteins that associate with 5S rRNA in the ‘crown’ of the large ribosomal subunit. However, this does not hold true for human mitochondria. The recent
high-resolution structure of the large subunit of the human mitochondrial ribosome14 has revealed that MRPL18 associates with a mitochon drially encoded tRNAVal, which substitutes for 5S rRNA. Perhaps MRPL18 was selected for the regulatory role unveiled by Zhang et al.5 because of its unique way of interacting with a tRNA. In summary, Zhang et al.5 have shown that the regulated mistranslation of a mitochondrial ribosomal protein can lead to the formation of a new hybrid ribosome that permits the enhanced translation of one class of mRNAs even when the translation of most mRNAs is repressed. Perhaps more importantly, it is now apparent that the mitochondrial ribosomal proteins may serve two masters. COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Jäkel, S. & Gorlich, D. EMBO J. 17, 4491–4502 (1998). 2. Fukasawa, Y. et al. Mol. Cell. Proteomics 14, 1113–1126 (2015). 3. Hällberg, B.M. & Larsson, N.G. Cell Metab. 20, 226–240 (2014). 4. Richter, K., Haslbeck, M. & Buchner, J. Mol. Cell 40, 253–266 (2010). 5. Zhang, X. et al. Nat. Struct. Mol. Biol. 22, 404–410 (2015). 6. Sonenberg, N. & Hinnebusch, A.G. Cell 136, 731–745 (2009). 7. Hinnebusch, A.G. Microbiol. Rev. 52, 248–273 (1988). 8. Lu, P.D., Harding, H.P. & Ron, D. J. Cell Biol. 167, 27–33 (2004). 9. Wek, R.C., Jiang, H.Y. & Anthony, T.G. Biochem. Soc. Trans. 34, 7–11 (2006). 10. Sun, J., Conn, C.S., Han, Y., Yeung, V. & Qian, S.B. J. Biol. Chem. 286, 6791–6800 (2011). 11. Xue, S. & Barna, M. Nat. Rev. Mol. Cell Biol. 13, 355–369 (2012). 12. Rubtsova, M.P. et al. J. Biol. Chem. 278, 22350–22356 (2003). 13. Ban, N. et al. Curr. Opin. Struct. Biol. 24, 165–169 (2014). 14. Brown, A. et al. Science 346, 718–722 (2014).
Light-driven Na+ pumps as next-generation inhibitory optogenetic tools Przemyslaw Nogly & Jörg Standfuss The first structures of a light-driven sodium pump provide insight into the mechanism of ion transport and selectivity. Genetic manipulation of rat neuronal cells and of Caenorhabditis elegans worms demonstrates the utility of such pumps for optogenetic applications. Retinal-binding proteins have revolutionized neurobiology, serving as tools to analyze nerve Przemyslaw Nogly and Jörg Standfuss are at the Laboratory of Biomolecular Research, Paul Scherrer Institute, Villigen, Switzerland. e-mail: [email protected]
action and to control animal behavior by light1. Key for the development of the modern optogenetics field were the discovery and characterization of light-gated channelrhodopsins2,3 and the light-driven chloride pump halorhodopsin4. There has been much debate as to whether a light-driven
nature structural & molecular biology volume 22 number 5 may 2015
sodium pump that could serve as a tool for silencing nerve action might exist. Following the discovery of the sodium pump Kr2 in the marine bacterium Krokinobacter eikastus5, the article by Gushchin et al.6 in this issue of Nature Structural & Molecular Biology and the concurrently published article by Kato et al.7 in Nature 351
news and views 120 mM Cl–
4 mM K+
150 mM Na+
Na+
Cl–
Extracellular fluid
Axonal cytosol
npg
© 2015 Nature America, Inc. All rights reserved.
Na+
4 mM Cl– HR (hyperpolarization)
12 mM Na+ ChR (depolarization)
K+
140 mM K+ Kr2 (hyperpolarization)
Figure 1 Retinal-binding proteins used as optogenetic tools. Retinal (red sticks) isomerizes after illumination (yellow bolts). This triggers translocation of ions (arrows) and changes the regulatory concentration gradients (colored boxes) across the neuronal membrane17. HR, halorhodopsin18; ChR, channelrhodopsin12; Kr2, sodium-proton pump6.
provide the first structural insights into this fascinating light-driven sodium pump. Kato et al.7 report structures of the protein at neutral and acidic conditions, and prove its use in opto genetic applications. Gushchin et al.6 report three structures at different acidic pHs, including one at the near-atomic resolution of 1.4 Å. Both studies present a detailed biochemical characterization including mutagenesis to elucidate the mechanism of Na+ transport through the membrane and to engineer variants that preferentially pump K+ to further expand the optogenetic toolbox. Optogenetics is the use of light to activate genetically modified neurons through manipulation of ionic gradients across their cellular membranes. Chosen as Method of the Year 2010 by Nature Methods8 and as Breakthrough of the Decade by Science9, optogenetics has been successfully applied to studying the function of neuronal responses, for example by direct stimulation of basal ganglia cells to study Parkinson’s disease10. The stimulation is achieved by specifically targeting light-sensitive ion channels and pumps into a set of neurons to trigger or suppress action potentials. The advantages of using light as an effector are manifold11: it is noninvasive, it can be precisely targeted with high spatial and temporal precision, it can be 352
used simultaneously at multiple wavelengths and locations, and it can report the presence or activity of specific molecules in native tissues. Retinal-binding proteins are the most successful light switches that have been used in optogenetics (Fig. 1). Activation of the protein channelrhodopsin, for example, results in an inflow of cations, which is followed by membrane depolarization and nerve action. Halorhodopsin can be used to counter the depolarization of the membrane by pumping Cl– toward the cytosol. In particular, the structure of a channelrhodopsin12 led to the development of a diverse toolbox for optogenetic applications. This includes redshifted variants that allow activation deep within tissues13 and transformation of the native cation channel into Figure 2 Kr2 cavities as part of the iontranslocation pathway. Secondary structure of Kr2 in gray ribbon representation. Lysine covalently binding retinal is represented with yellow sticks. Orange sticks represent residues Asp116 and Leu120, which have been implicated in the ion-transport mechanism. Arrows indicate the direction of the ion entry and release. The surfaces indicate cavities inside of the Kr2 structure with the ion-uptake and selectivity cavity in blue, the ion-conducting cavity in magenta and the ion-release cavity in green. Cavities are calculated with Hollow19.
an engineered chloride channel14,15. However, a much-sought-after tool for hyperpolarization of the membrane, which would enable efficient transport of cations out of the cell, was missing. Kr2 can be this protein and can complement the available optogenetic toolbox of retinalbinding proteins. Kato et al.7 have shown that this is possible by heterogeneously expressing Kr2 to manipulate rat cortical neurons and living C. elegans worms. In the cellular membranes of bacteria, archaea and unicellular eukaryota, retinal-binding proteins play a crucial part in establishing and maintaining the right ionic gradients. All retinal-binding proteins of this class share a common seven transmembrane helix (7TM) architecture that is also common in humans, in which similar proteins act as visual pigments and G protein–coupled receptors. The beststudied retinal-binding 7TM protein remains the light-driven proton pump bacteriorhodopsin (bR)16, which provides an important point of reference for the study of other proteins in this class. The structures presented by Gushchin et al.6 and Kato et al.7 have an overall high degree of similarity to bR, with the largest differences found in the presence of an additional N-terminal α-helix in the extracellular region and in binding of loop B-C over a hydrophilic cavity likely to be part of the Na+-release region (green cavity in Fig. 2). As in all retinal-binding 7TM proteins, the retinal in bR (and Kr2) is covalently linked to the protein via a Schiff base bond to a lysine side chain. Light-induced isomerization of retinal from the all-trans to the 13-cis conformation drives the translocation of protons from the protonated Schiff base link to a hydrophilic pathway further across the proteins. In halorhodopsin, in contrast, the inward transport of the negatively charged Cl– is enabled by stabilization of the positively charged protonated Schiff base, and retinal isomerization promotes its translocation along the hydrophilic path.
+ 1D Na 1D
Release
l
na
ti Re
D116
Conducting
L120 Uptake–selectivity Na+
1D
volume 22 number 5 may 2015 nature structural & molecular biology
npg
© 2015 Nature America, Inc. All rights reserved.
news and views It has been questioned whether Na+, owing to its positive charge and inability to bind covalently to the Schiff base like a proton, could be pumped with support of the retinal Schiff base and retinal isomerization. In Kr2, subtle differences in the amino acid composition of the environment surrounding the covalent Schiff base link of the retinal and along the transport pathway are responsible for the ability to pump Na+ against an osmotic gradient. The structure-based model of Na+ translocation proposed independently by the two groups relies on the deprotonation of the Schiff base after retinal isomerization. The proton is transferred to the neighboring Asp116 and causes it, or nearby Leu120, as shown in Gushchin et al.6, to flip away to enable passage of Na+ across a neutral Schiff base. Another interesting region was identified just at the beginning of the ion-translocation path. The ion-uptake cavity (blue cavity in Fig. 2) contains residues that affect the selectivity of the pump. Point mutations introduced in this region affect the transport of the ions, with one unusual case leading to preferential K+ transport, which
has not been observed for the wild-type protein. This observation adds another exciting opportunity to use K+ gradients to hyperpolarize the nerve cells in optogenetics. K+ is particularly interesting because it is present at high concentration in the cytosol and would potentially be less invasive for the studied system. Publication of these initial Kr2 structures has answered many questions but also has opened the way for a range of follow-up experiments. For example, it will be interesting to compare the effect of Na+ currents for the optogenetic control of diverse systems and to verify the extent to which engineered Kr2 variants pumping K+ will be useful. Engineering of further interesting variants will be facilitated by a more thorough experimental characterization of the pumping mechanism and localization of Na+-binding sites. Given the impact that structural data has had on unraveling the photocycle of retinal-binding proteins, we are looking forward to seeing structures of photoactivated Kr2 intermediates to provide further mechanistic insights into this promising optogenetic tool.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Deisseroth, K. Sci. Am. 303, 48–55 (2010). 2. Nagel, G. et al. Science 296, 2395–2398 (2002). 3. Nagel, G. et al. Proc. Natl. Acad. Sci. USA 100, 13940–13945 (2003). 4. Schobert, B. & Lanyi, J.K. J. Biol. Chem. 257, 10306–10313 (1982). 5. Inoue, K. et al. Nat. Commun. 4, 1678 (2013). 6. Gushchin, I. et al. Nat. Struct. Mol. Biol. 22, 390–395 (2015). 7. Kato, H.E. et al. Nature doi:10.1038/nature14322 (6 April 2015). 8. Nat. Methods 8, 1 (2011). 9. News Staff. Science 330, 1612–1613 (2010). 10. Kravitz, A.V. et al. Nature 466, 622–626 (2010). 11. Häusser, M. Nat. Methods 11, 1012–1014 (2014). 12. Kato, H.E. et al. Nature 482, 369–374 (2012). 13. Lin, J.Y., Knutsen, P.M., Muller, A., Kleinfeld, D. & Tsien, R.Y. Nat. Neurosci. 16, 1499–1508 (2013). 14. Berndt, A., Lee, S.Y., Ramakrishnan, C. & Deisseroth, K. Science 344, 420–424 (2014). 15. Wietek, J. et al. Science 344, 409–412 (2014). 16. Wickstrand, C., Dods, R., Royant, A. & Neutze, R. Biochim. Biophys. Acta 1850, 536–553 (2015). 17. Lodish, H.F. Molecular Cell Biology (W.H. Freeman, 2000). 18. Kolbe, M., Besir, H., Essen, L.O. & Oesterhelt, D. Science 288, 1390–1396 (2000). 19. Ho, B.K. & Gruswitz, F. BMC Struct. Biol. 8, 49 (2008).
A bumpy road for RNA polymerase II Luciana E Giono & Alberto R Kornblihtt The identification of a second regulatory checkpoint controlling RNA polymerase II elongation near the poly(A) site of protein-coding genes reveals an additional level of complexity in the modulation of eukaryotic transcriptional elongation and termination. A great deal of the regulation of eukaryotic gene expression relies on the control of transcription performed by RNA polymerase II (RNAPII). This process can be regulated at the level of initiation, elongation and termination, and each of these three enzymatic events is linked to differential phosphorylation of specific amino acid residues within the repeated heptamer of the C-terminal domain (CTD) of RNAPII’s largest subunit. Each of the 52 CTD repeats of the mammalian enzyme (or 26 in budding yeast) bears the consensus sequence Tyr-SerPro-Thr-Ser-Pro-Ser, in which all residues Luciana E. Giono and Alberto R. Kornblihtt are at the Instituto de Fisiología, Biología Molecular y Neurociencias (IFIBYNEUBA-CONICET) and Departamento de Fisiología, Biología Molecular y Celular, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria, Buenos Aires, Argentina. e-mail: [email protected]
except the prolines are subject to regulatory phosphorylation. A commonly accepted paradigm holds that Ser5 phosphorylation by the CDK7 kinase subunit of general transcription factor TFIIH is a hallmark of transcriptional initiation, whereas Ser2 phosphorylation meditated by the CDK9 subunit of the P-TEFb complex is needed for effective elongation. However, this simplistic scheme has been challenged by the subsequent demonstration of important roles for additional kinases such as CDK12 and CDK13 (ref. 1) and for other CTD residues such as Tyr1, Thr4 or Ser7 (ref. 2) as well as by evidence that hyperphosphorylation of both Ser2 and Ser5 after DNA damage can lead to inhibition of elongation3. One way to distinguish differential phos phorylation patterns that control polymerase initiation, elongation and termination decisions from those that are merely correlated with these events is to assess the effects of specific pharmacological kinase inhibitors. In a report in this issue, Murphy and colleagues4 monitor
nature structural & molecular biology volume 22 number 5 may 2015
the effects of two CDK9 inhibitors, DRB and KM05382, both genome wide and on individual model genes. Their genome-wide analysis was performed with the powerful global run-on sequencing (GRO-seq) technique, which allows global mapping of RNAPII densities. The authors found that most genes possess a CDK9-dependent checkpoint to elongation within 500 bp downstream of the transcription start site (TSS), regardless of whether RNAPII pauses proximally to the promoter. In other words, when CDK9 kinase activity was inhibited, RNAPII could proceed from the TSS to an early checkpoint but no further. Surprisingly, RNAPII molecules that had passed the early CDK9 checkpoint when the inhibitor was added continued to transcribe through the remainder of the gene body without apparent difficulty. The authors propose that CDK9 activity is necessary to facilitate RNAPII passage through the early checkpoint and that RNAPII molecules that have already negotiated this passage 353
© 2015 Nature America, Inc. All rights reserved.
news and views (i) The mRNA encoding MRPL18 has a short upstream ORF (uORF) of nine codons, which is followed by the ‘authentic’ AUG and then by the CUG identified by the authors as the initiator of MRPL18(cyto) (Fig. 1). This arrangement is conserved in mice. The role of such uORFs in regulating translation of downstream sequences was initially identified in the classic experiments on translation of GCN4 in yeast7. Although reinitiation can occur downstream of the uORF, the site of reinitiation is influenced by the activity of translation initiation factors. Such a mechanism has also been established in mammalian cells, for mRNAs encoding proteins such as ATF4 (ref. 8). In the case of the MRPL18 mRNA, phosphorylation of eIF2α slows translational initiation, thus leading to skipping of reinitiation at the second AUG and permitting some reinitiation at the subsequent CUG to synthesize MRPL18(cyto). A key point is that eIF2 can be phosphorylated by several kinases, each responding to a different type of stress9. Thus MRPL18(cyto) may be generated in response to many types of stress beyond heat shock. The relationship between stress, slowed translation initiation and the selective initiation of translation is evident from the observation that hyperactivation of translation, by means of a constitutively activated TORC1 complex, suppresses translation of heat-shock proteins10, possibly owing to reduced production of MRPL18(cyto). (ii) The phosphorylated MRPL18(cyto) associates with 80S ribosomes to yield a ‘hybrid’ ribosome that is essential for translation of the HSP70-encoding mRNA. The relatively new concept of ‘specialized’ ribosomes, tailored for the translation of specific mRNAs, has now been demonstrated in several instances11 and
may prove to be an important element in the overall regulation of translation. (iii) It has been known for some time that HSP70-encoding mRNA is translated in a manner that is cap independent but dependent on the 5′ untranslated region (5′ UTR), perhaps through an internal ribosome entry site (IRES)12, although that has been disputed10. It now seems likely that the presence of MRPL18(cyto) permits the hybrid ribosome to bypass normal cap-dependent initiation, presumably by interacting with some structure in the 5′ UTR to effect translation. Future studies determining the location of the MRPL18(cyto) on the cytoplasmic ribosome may suggest the mechanism by which this occurs. The conceptual novelty of this story is the realization that any of the mRPs might be coopted for regulatory functions in the cytoplasm or in the nucleus. Although the unusual structure of mammalian mitochondrial ribosomes involves far more protein-protein interactions than occur in cytoplasmic ribosomes, the mRPs are still predominantly RNA-binding proteins. Thus there should be many opportunities to adapt them as regulatory elements at a variety of levels in the nuclear and cytoplasmic RNAome. Need we remind the reader that plants have an additional source of such regulators in the form of the proteins making up the chloroplast ribosome? MRPL18 is closely related to RPs whose ancestry can be traced over three billion years and which are present in all ribosomes; they have recently been given the structural name uL18 (ref. 13). uL18 is one of the two proteins that associate with 5S rRNA in the ‘crown’ of the large ribosomal subunit. However, this does not hold true for human mitochondria. The recent
high-resolution structure of the large subunit of the human mitochondrial ribosome14 has revealed that MRPL18 associates with a mitochon drially encoded tRNAVal, which substitutes for 5S rRNA. Perhaps MRPL18 was selected for the regulatory role unveiled by Zhang et al.5 because of its unique way of interacting with a tRNA. In summary, Zhang et al.5 have shown that the regulated mistranslation of a mitochondrial ribosomal protein can lead to the formation of a new hybrid ribosome that permits the enhanced translation of one class of mRNAs even when the translation of most mRNAs is repressed. Perhaps more importantly, it is now apparent that the mitochondrial ribosomal proteins may serve two masters. COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Jäkel, S. & Gorlich, D. EMBO J. 17, 4491–4502 (1998). 2. Fukasawa, Y. et al. Mol. Cell. Proteomics 14, 1113–1126 (2015). 3. Hällberg, B.M. & Larsson, N.G. Cell Metab. 20, 226–240 (2014). 4. Richter, K., Haslbeck, M. & Buchner, J. Mol. Cell 40, 253–266 (2010). 5. Zhang, X. et al. Nat. Struct. Mol. Biol. 22, 404–410 (2015). 6. Sonenberg, N. & Hinnebusch, A.G. Cell 136, 731–745 (2009). 7. Hinnebusch, A.G. Microbiol. Rev. 52, 248–273 (1988). 8. Lu, P.D., Harding, H.P. & Ron, D. J. Cell Biol. 167, 27–33 (2004). 9. Wek, R.C., Jiang, H.Y. & Anthony, T.G. Biochem. Soc. Trans. 34, 7–11 (2006). 10. Sun, J., Conn, C.S., Han, Y., Yeung, V. & Qian, S.B. J. Biol. Chem. 286, 6791–6800 (2011). 11. Xue, S. & Barna, M. Nat. Rev. Mol. Cell Biol. 13, 355–369 (2012). 12. Rubtsova, M.P. et al. J. Biol. Chem. 278, 22350–22356 (2003). 13. Ban, N. et al. Curr. Opin. Struct. Biol. 24, 165–169 (2014). 14. Brown, A. et al. Science 346, 718–722 (2014).
Light-driven Na+ pumps as next-generation inhibitory optogenetic tools Przemyslaw Nogly & Jörg Standfuss The first structures of a light-driven sodium pump provide insight into the mechanism of ion transport and selectivity. Genetic manipulation of rat neuronal cells and of Caenorhabditis elegans worms demonstrates the utility of such pumps for optogenetic applications. Retinal-binding proteins have revolutionized neurobiology, serving as tools to analyze nerve Przemyslaw Nogly and Jörg Standfuss are at the Laboratory of Biomolecular Research, Paul Scherrer Institute, Villigen, Switzerland. e-mail: [email protected]
action and to control animal behavior by light1. Key for the development of the modern optogenetics field were the discovery and characterization of light-gated channelrhodopsins2,3 and the light-driven chloride pump halorhodopsin4. There has been much debate as to whether a light-driven
nature structural & molecular biology volume 22 number 5 may 2015
sodium pump that could serve as a tool for silencing nerve action might exist. Following the discovery of the sodium pump Kr2 in the marine bacterium Krokinobacter eikastus5, the article by Gushchin et al.6 in this issue of Nature Structural & Molecular Biology and the concurrently published article by Kato et al.7 in Nature 351
news and views 120 mM Cl–
4 mM K+
150 mM Na+
Na+
Cl–
Extracellular fluid
Axonal cytosol
npg
© 2015 Nature America, Inc. All rights reserved.
Na+
4 mM Cl– HR (hyperpolarization)
12 mM Na+ ChR (depolarization)
K+
140 mM K+ Kr2 (hyperpolarization)
Figure 1 Retinal-binding proteins used as optogenetic tools. Retinal (red sticks) isomerizes after illumination (yellow bolts). This triggers translocation of ions (arrows) and changes the regulatory concentration gradients (colored boxes) across the neuronal membrane17. HR, halorhodopsin18; ChR, channelrhodopsin12; Kr2, sodium-proton pump6.
provide the first structural insights into this fascinating light-driven sodium pump. Kato et al.7 report structures of the protein at neutral and acidic conditions, and prove its use in opto genetic applications. Gushchin et al.6 report three structures at different acidic pHs, including one at the near-atomic resolution of 1.4 Å. Both studies present a detailed biochemical characterization including mutagenesis to elucidate the mechanism of Na+ transport through the membrane and to engineer variants that preferentially pump K+ to further expand the optogenetic toolbox. Optogenetics is the use of light to activate genetically modified neurons through manipulation of ionic gradients across their cellular membranes. Chosen as Method of the Year 2010 by Nature Methods8 and as Breakthrough of the Decade by Science9, optogenetics has been successfully applied to studying the function of neuronal responses, for example by direct stimulation of basal ganglia cells to study Parkinson’s disease10. The stimulation is achieved by specifically targeting light-sensitive ion channels and pumps into a set of neurons to trigger or suppress action potentials. The advantages of using light as an effector are manifold11: it is noninvasive, it can be precisely targeted with high spatial and temporal precision, it can be 352
used simultaneously at multiple wavelengths and locations, and it can report the presence or activity of specific molecules in native tissues. Retinal-binding proteins are the most successful light switches that have been used in optogenetics (Fig. 1). Activation of the protein channelrhodopsin, for example, results in an inflow of cations, which is followed by membrane depolarization and nerve action. Halorhodopsin can be used to counter the depolarization of the membrane by pumping Cl– toward the cytosol. In particular, the structure of a channelrhodopsin12 led to the development of a diverse toolbox for optogenetic applications. This includes redshifted variants that allow activation deep within tissues13 and transformation of the native cation channel into Figure 2 Kr2 cavities as part of the iontranslocation pathway. Secondary structure of Kr2 in gray ribbon representation. Lysine covalently binding retinal is represented with yellow sticks. Orange sticks represent residues Asp116 and Leu120, which have been implicated in the ion-transport mechanism. Arrows indicate the direction of the ion entry and release. The surfaces indicate cavities inside of the Kr2 structure with the ion-uptake and selectivity cavity in blue, the ion-conducting cavity in magenta and the ion-release cavity in green. Cavities are calculated with Hollow19.
an engineered chloride channel14,15. However, a much-sought-after tool for hyperpolarization of the membrane, which would enable efficient transport of cations out of the cell, was missing. Kr2 can be this protein and can complement the available optogenetic toolbox of retinalbinding proteins. Kato et al.7 have shown that this is possible by heterogeneously expressing Kr2 to manipulate rat cortical neurons and living C. elegans worms. In the cellular membranes of bacteria, archaea and unicellular eukaryota, retinal-binding proteins play a crucial part in establishing and maintaining the right ionic gradients. All retinal-binding proteins of this class share a common seven transmembrane helix (7TM) architecture that is also common in humans, in which similar proteins act as visual pigments and G protein–coupled receptors. The beststudied retinal-binding 7TM protein remains the light-driven proton pump bacteriorhodopsin (bR)16, which provides an important point of reference for the study of other proteins in this class. The structures presented by Gushchin et al.6 and Kato et al.7 have an overall high degree of similarity to bR, with the largest differences found in the presence of an additional N-terminal α-helix in the extracellular region and in binding of loop B-C over a hydrophilic cavity likely to be part of the Na+-release region (green cavity in Fig. 2). As in all retinal-binding 7TM proteins, the retinal in bR (and Kr2) is covalently linked to the protein via a Schiff base bond to a lysine side chain. Light-induced isomerization of retinal from the all-trans to the 13-cis conformation drives the translocation of protons from the protonated Schiff base link to a hydrophilic pathway further across the proteins. In halorhodopsin, in contrast, the inward transport of the negatively charged Cl– is enabled by stabilization of the positively charged protonated Schiff base, and retinal isomerization promotes its translocation along the hydrophilic path.
+ 1D Na 1D
Release
l
na
ti Re
D116
Conducting
L120 Uptake–selectivity Na+
1D
volume 22 number 5 may 2015 nature structural & molecular biology
npg
© 2015 Nature America, Inc. All rights reserved.
news and views It has been questioned whether Na+, owing to its positive charge and inability to bind covalently to the Schiff base like a proton, could be pumped with support of the retinal Schiff base and retinal isomerization. In Kr2, subtle differences in the amino acid composition of the environment surrounding the covalent Schiff base link of the retinal and along the transport pathway are responsible for the ability to pump Na+ against an osmotic gradient. The structure-based model of Na+ translocation proposed independently by the two groups relies on the deprotonation of the Schiff base after retinal isomerization. The proton is transferred to the neighboring Asp116 and causes it, or nearby Leu120, as shown in Gushchin et al.6, to flip away to enable passage of Na+ across a neutral Schiff base. Another interesting region was identified just at the beginning of the ion-translocation path. The ion-uptake cavity (blue cavity in Fig. 2) contains residues that affect the selectivity of the pump. Point mutations introduced in this region affect the transport of the ions, with one unusual case leading to preferential K+ transport, which
has not been observed for the wild-type protein. This observation adds another exciting opportunity to use K+ gradients to hyperpolarize the nerve cells in optogenetics. K+ is particularly interesting because it is present at high concentration in the cytosol and would potentially be less invasive for the studied system. Publication of these initial Kr2 structures has answered many questions but also has opened the way for a range of follow-up experiments. For example, it will be interesting to compare the effect of Na+ currents for the optogenetic control of diverse systems and to verify the extent to which engineered Kr2 variants pumping K+ will be useful. Engineering of further interesting variants will be facilitated by a more thorough experimental characterization of the pumping mechanism and localization of Na+-binding sites. Given the impact that structural data has had on unraveling the photocycle of retinal-binding proteins, we are looking forward to seeing structures of photoactivated Kr2 intermediates to provide further mechanistic insights into this promising optogenetic tool.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Deisseroth, K. Sci. Am. 303, 48–55 (2010). 2. Nagel, G. et al. Science 296, 2395–2398 (2002). 3. Nagel, G. et al. Proc. Natl. Acad. Sci. USA 100, 13940–13945 (2003). 4. Schobert, B. & Lanyi, J.K. J. Biol. Chem. 257, 10306–10313 (1982). 5. Inoue, K. et al. Nat. Commun. 4, 1678 (2013). 6. Gushchin, I. et al. Nat. Struct. Mol. Biol. 22, 390–395 (2015). 7. Kato, H.E. et al. Nature doi:10.1038/nature14322 (6 April 2015). 8. Nat. Methods 8, 1 (2011). 9. News Staff. Science 330, 1612–1613 (2010). 10. Kravitz, A.V. et al. Nature 466, 622–626 (2010). 11. Häusser, M. Nat. Methods 11, 1012–1014 (2014). 12. Kato, H.E. et al. Nature 482, 369–374 (2012). 13. Lin, J.Y., Knutsen, P.M., Muller, A., Kleinfeld, D. & Tsien, R.Y. Nat. Neurosci. 16, 1499–1508 (2013). 14. Berndt, A., Lee, S.Y., Ramakrishnan, C. & Deisseroth, K. Science 344, 420–424 (2014). 15. Wietek, J. et al. Science 344, 409–412 (2014). 16. Wickstrand, C., Dods, R., Royant, A. & Neutze, R. Biochim. Biophys. Acta 1850, 536–553 (2015). 17. Lodish, H.F. Molecular Cell Biology (W.H. Freeman, 2000). 18. Kolbe, M., Besir, H., Essen, L.O. & Oesterhelt, D. Science 288, 1390–1396 (2000). 19. Ho, B.K. & Gruswitz, F. BMC Struct. Biol. 8, 49 (2008).
A bumpy road for RNA polymerase II Luciana E Giono & Alberto R Kornblihtt The identification of a second regulatory checkpoint controlling RNA polymerase II elongation near the poly(A) site of protein-coding genes reveals an additional level of complexity in the modulation of eukaryotic transcriptional elongation and termination. A great deal of the regulation of eukaryotic gene expression relies on the control of transcription performed by RNA polymerase II (RNAPII). This process can be regulated at the level of initiation, elongation and termination, and each of these three enzymatic events is linked to differential phosphorylation of specific amino acid residues within the repeated heptamer of the C-terminal domain (CTD) of RNAPII’s largest subunit. Each of the 52 CTD repeats of the mammalian enzyme (or 26 in budding yeast) bears the consensus sequence Tyr-SerPro-Thr-Ser-Pro-Ser, in which all residues Luciana E. Giono and Alberto R. Kornblihtt are at the Instituto de Fisiología, Biología Molecular y Neurociencias (IFIBYNEUBA-CONICET) and Departamento de Fisiología, Biología Molecular y Celular, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria, Buenos Aires, Argentina. e-mail: [email protected]
except the prolines are subject to regulatory phosphorylation. A commonly accepted paradigm holds that Ser5 phosphorylation by the CDK7 kinase subunit of general transcription factor TFIIH is a hallmark of transcriptional initiation, whereas Ser2 phosphorylation meditated by the CDK9 subunit of the P-TEFb complex is needed for effective elongation. However, this simplistic scheme has been challenged by the subsequent demonstration of important roles for additional kinases such as CDK12 and CDK13 (ref. 1) and for other CTD residues such as Tyr1, Thr4 or Ser7 (ref. 2) as well as by evidence that hyperphosphorylation of both Ser2 and Ser5 after DNA damage can lead to inhibition of elongation3. One way to distinguish differential phos phorylation patterns that control polymerase initiation, elongation and termination decisions from those that are merely correlated with these events is to assess the effects of specific pharmacological kinase inhibitors. In a report in this issue, Murphy and colleagues4 monitor
nature structural & molecular biology volume 22 number 5 may 2015
the effects of two CDK9 inhibitors, DRB and KM05382, both genome wide and on individual model genes. Their genome-wide analysis was performed with the powerful global run-on sequencing (GRO-seq) technique, which allows global mapping of RNAPII densities. The authors found that most genes possess a CDK9-dependent checkpoint to elongation within 500 bp downstream of the transcription start site (TSS), regardless of whether RNAPII pauses proximally to the promoter. In other words, when CDK9 kinase activity was inhibited, RNAPII could proceed from the TSS to an early checkpoint but no further. Surprisingly, RNAPII molecules that had passed the early CDK9 checkpoint when the inhibitor was added continued to transcribe through the remainder of the gene body without apparent difficulty. The authors propose that CDK9 activity is necessary to facilitate RNAPII passage through the early checkpoint and that RNAPII molecules that have already negotiated this passage 353
