REPORTS Rho-kinase inhibitor Y27632 (32) had no detectable effect on microtubule organization, delocalized myosin II, and again did not perturb anillin. Thus, Rho-kinase is required to localize myosin II to furrows in blebbistatin-arrested cells, which is broadly consistent with previous data (33), and other kinases, probably including aurora B, are required for midzone organization. Cyclin-dependent kinase (CDK) inhibitors and monastrol do not perturb the cytology of blebbistatin-arrested cells (Table 1). Thus, CDKs do not appear to be directly involved in maintaining the contractile ring after anaphase, although it is possible that they might influence C-phase timing. The target of monastrol, the mitotic kinesin Eg5, is thought to have a central role in the establishment of spindle bipolarity, but our data suggest that it is not responsible for bipolar organization of the midzone during cytokinesis. Our experiments emphasize the usefulness of fast-acting and reversible drugs in a dynamic cellular pathway such as cytokinesis. It is impossible to determine whether Eg5, for example, is required in cytokinesis with simple ablation experiments, because the ablation of Eg5 disrupts mitosis. New drugs that target guanosine triphosphatases, membrane dynamics, and mitotic motors will be useful in further dissecting the logic of cytokinesis. These proteins and processes are all required for cytokinesis (1, 2), but their precise roles in timing and spatial organization have yet to be defined. References and Notes
1. A. F. Straight, C. M. Field, Curr. Biol. 10, R760 (2000). 2. M. Glotzer, Annu. Rev. Cell Dev. Biol. 17, 351 (2001). 3. R. Rappaport, Cytokinesis in Animal Cells (Developmental and Cell Biology Series 32, Cambridge Univ. Press, Cambridge, UK, 1996). 4. D. McCollum, K. L. Gould, Trends Cell Biol. 11, 89 (2001). 5. A. J. Bardin, A. Amon, Nature Rev. Mol. Cell Biol. 2, 815 (2001). 6. S. P. Wheatley, Y. Wang, J. Cell Biol. 135, 981 (1996). 7. J. R. Peterson, T. J. Mitchison, Chem. Biol. 9, 1 (2002). 8. S. N. Martineau, P. R. Andreassen, R. L. Margolis, J. Cell Biol. 131, 191 (1995). 9. J. C. Canman, D. B. Hoffman, E. D. Salmon, Curr. Biol. 10, 611 (2000). 10. I. Mabuchi, M. Okuno, J. Cell Biol. 74, 251 (1977). 11. A. De Lozanne, J. A. Spudich, Science 236, 1086 (1987). 12. A. Cheung et al., Nature Cell Biol. 4, 83 (2002). 13. Materials and methods are available as supporting material on Science Online. 14. K. Fujiwara, T. D. Pollard, J. Cell Biol. 71, 848 (1976). 15. C. M. Field, B. M. Alberts, J. Cell Biol. 131, 165 (1995). 16. A. Hirano, T. Kurimura, Exp. Cell Res. 89, 111 (1974). 17. P. R. Andreassen, O. D. Lohez, F. B. Lacroix, R. L. Margolis, Mol. Biol. Cell 12, 1315 (2001). 18. C. B. Shuster, D. R. Burgess, Curr. Biol. 12, 854 (2002). 19. R. Wasch, F. R. Cross, Nature 418, 556 (2002). 20. J. M. Peters, Mol. Cell 9, 931 (2002). 21. A. F. Kisselev, A. L. Goldberg, Chem. Biol. 8, 739 (2001). 22. M. Gatti, M. G. Giansanti, S. Bonaccorsi, Microsc. Res. Tech. 49, 202 (2000). 23. D. Cimini, D. Fioravanti, C. Tanzarella, F. Degrassi, Chromosoma 107, 479 (1998).
24. U. T. Ruegg, G. M. Burgess, Trends Pharmacol. Sci. 10, 218 (1989). 25. R. M. Golsteyn, K. E. Mundt, A. M. Fry, E. A. Nigg, J. Cell Biol. 129, 1617 (1995). 26. Y. Terada et al., EMBO J. 17, 667 (1998). 27. M. Murata-Hori, Y. L. Wang, J. Cell Biol. 159, 45 (2002). 28. D. Fabbro et al., Pharmacol. Ther. 82, 293 (1999). 29. J. F. Henri, B. A. George, K. N. John, M. A. Austen, World Intellectual Property Organization (WPO) Patent #WO0121596. (2001). 30. J. F. Henri, M. A. Austen, WPO Patent #WO0155116 (2001). 31. K. Oegema, M. S. Savoian, T. J. Mitchison, C. M. Field, J. Cell Biol. 150, 539 (2000). 32. M. Uehata et al., Nature 389, 990 (1997). 33. H. Kosako et al., Oncogene 19, 6059 (2000). 34. We thank E. M. Ostap for purified myosin Ib; J. Dantzig, A. Shaw, and Y. M. Goldman for rabbit skeletal myosin SI; L. Flanagan and T. Stossel for M2
cells; K. Pierce and M. M.-C. Lo for assistance with chiral chromatography; S. Miller and T. Kapoor for advice on chemical synthesis; T. Kapoor for assistance with aurora kinase inhibitors; and A. Farrell, W. Brieher, and R. Ward for stimulating discussion and critical review of the manuscript. This work was supported by grants from the NIH (GM62566, GM23928) to T.J.M. and from Merck & Co. and E. Merck. A.F.S. was supported by the Cancer Research Fund of the Damon Runyon–Walter Winchell Foundation. Supporting Online Material www.sciencemag.org/cgi/content/full/299/5613/1743/ DC1 Materials and Methods Figs. S1 to S6 References and Notes Movies S1 to S4 11 December 2002; accepted 4 February 2003
Protein Insertion into the Mitochondrial Inner Membrane by a Twin-Pore Translocase Peter Rehling,1 Kirstin Model,2 Katrin Brandner,1,3 Peter Kovermann,4 Albert Sickmann,5 Helmut E. Meyer,5 Werner Ku¨hlbrandt,2 Richard Wagner,4 Kaye N. Truscott,1 Nikolaus Pfanner1* The mitochondrial inner membrane imports numerous proteins that span it multiple times using the membrane potential ⌬ as the only external energy source. We purified the protein insertion complex (TIM22 complex), a twin-pore translocase that mediated the insertion of precursor proteins in a three-step process. After the precursor is tethered to the translocase without losing energy from the ⌬, two energy-requiring steps were needed. First, ⌬ acted on the precursor protein and promoted its docking in the translocase complex. Then, ⌬ and an internal signal peptide together induced rapid gating transitions in one pore and closing of the other pore and drove membrane insertion to completion. Thus, protein insertion was driven by the coordinated action of a twin-pore complex in two voltage-dependent steps. The mitochondrial inner membrane contains two translocases that are responsible for the specific import of hundreds of different proteins (1–4). The presequence translocase (TIM23 complex) typically transports hydrophilic preproteins with amino-terminal presequences by using the inner membrane potential ⌬ for initiation of translocation and the ATP-driven Hsp70 motor for completion of transport. The protein insertion complex (TIM22 complex, carrier translocase) carries 1 Institut fu¨r Biochemie und Molekularbiologie, Universita¨t Freiburg, Hermann-Herder-Strasse 7, D-79104 Freiburg, Germany. 2Max-Planck Institut fu¨r Biophysik, Abteilung Strukturbiologie, Heinrich-Hoffmann-Strasse 7, D-60528 Frankfurt am Main, Germany. 3Fakulta¨t fu¨r Biologie, Universita¨t Freiburg, D-79104 Freiburg, Germany. 4Biophysik, Universita¨t Osnabru¨ck, FB Biologie/Chemie, D-49034 Osnabru¨ck, Germany. 5Medizinisches Proteom-Center, Geba¨ude ZKF E/143, Ruhr-Universita¨t Bochum, Universita¨tsstrasse 150, D-44780 Bochum, Germany.
*To whom correspondence should be addressed. Email:
[email protected]
out the insertion of numerous abundant multispanning inner membrane proteins, e.g., metabolite carriers, that contain internal targeting signals. The TIM22 complex of ⬃300 kD contains six subunits, the integral membrane proteins Tim54, Tim22, and Tim18 and the peripheral proteins Tim12, Tim10, and Tim9 (1–4). Its function depends on only one external energy source, the membrane potential ⌬. Purified Tim22 forms a single pore that is voltage-activated and responds to a synthetic signal peptide (5). How the TIM22 complex inserts membrane proteins that contain multiple transmembrane segments is unclear, because the entire complex has never been purified in a functional form. Also, how ⌬ alone can drive the import and insertion of multispanning inner membrane proteins remains unclear. We developed a strategy to isolate the yeast mitochondrial TIM22 complex. When we tagged Tim18 with ProtA, which contains two immunoglobulin G (IgG)– binding do-
www.sciencemag.org SCIENCE VOL 299 14 MARCH 2003
1747
REPORTS mains, the growth of yeast cells and the function of mitochondria were not impaired. Mitochondria were isolated and lysed with digitonin. After purification on IgG-Sepharose, proteins in the eluate were separated by SDS– polyacrylamide electrophoresis (SDS-PAGE) and detected by immunodecoration. Only the subunits of the TIM22 complex were found in the eluate (Fig. 1A); none of the other mitochondrial translocase complexes were present. After selective proteolytic removal of the ProtA-tag, the mobility of the purified TIM22 complex on blue native gels (BNPAGE) was indistinguishable from that of the TIM22 complex in a detergent lysate of whole wild-type mitochondria (Fig. 1B). Protein staining did not reveal any other complex (Fig. 1C). All major components of the purified complex were identified by electrospray ionization mass spectrometry. Thus, the entire TIM22 complex was purified. We reconstituted the purified TIM22 complex into small unilamellar liposomes and fused them with a planar lipid bilayer. Current recordings revealed a channel activity that exhibited the basic characteristics as observed for expressed Tim22 alone (5) [fig. S1, A and B (6)]. However, although Tim22 alone mainly acts as a single pore with multiple subconductance states (5), all recordings with the TIM22 complex showed the characteristics of two coupled pores with a lower occurrence of subconductance states within a particular pore. Direct transitions between the two open-channel amplitudes revealed that the pore activities were tightly coupled (Fig. 2, A and B). Among several internal segments of the phosphate carrier shown to interact specifically with Tom receptors (7), only the 13-amino acid peptide named P2 interacts with purified Tim22, inducing a flickering of the recombinant Tim22 channel at a membrane potential above 140 mV (5). Amino acid stretches related to P2 are conserved in a number of other carrier proteins (6). When P2 was added to the TIM22 complex, it also induced a rapid gating activity, but two major differences were observed. First, P2 induced flickering of only one of the two pores; the other pore was mainly closed (Fig. 2, A and B). Second, the TIM22 complex required a much lower threshold voltage to induce the increased frequency of channel gating than did Tim22 alone. A membrane potential of 75 mV induced rapid gating of the complex (Fig. 2, A and C). Moreover, the frequency of gating transitions was significantly higher for the TIM22 complex than for Tim22 alone (Fig. 2C). When application of a membrane potential was prolonged in the absence of P2, the Tim22 channels closed (Fig. 2, A and B, and fig. S1C), which suggests that, in the absence of precursor proteins, the Tim22 channels of energized mitochondria were mainly closed and thus a major
1748
Fig. 1. Purification of the TIM22 complex from yeast mitochondria. (A) Specificity of the isolation procedure. A tobacco etch virus ( TEV )– ProtA encoding cassette was integrated into the Saccharomyces cerevisiae chromosome downstream of the TIM18 gene ( Tim18ProtA). Mitochondria were isolated from wild-type and Tim18ProtA yeast cells, lysed in buffer containing 1% digitonin, and incubated with humanIgG–Sepharose. After extensive washing, adsorbed proteins were eluted at low pH, resolved by SDS-PAGE, and probed for associated proteins via immunoblotting. For comparison, 1% of the extract (lanes 1 and 4), 1% of the unbound (lanes 2 and 5), and 20% of the eluate (lanes 3 and 6) were loaded. (B) Mobility of the isolated TIM22 complex and the TIM22 complex in mitochondrial lysate on blue native gels (BN-PAGE). Lane 1, To isolate the native TIM22 complex, proteins bound to IgG-Sepharose were eluted by cleavage with TEV protease. Lane 2 shows mitochondria lysed with digitonin. The samples were separated by BN-PAGE and subjected to immunoblotting with antibodies against Tim22. (C) Integrity of the TIM22 complex. TEV protease eluates of a control (wild-type mitochondria; lane 1) and Tim18ProtA isolation (lane 2) were separated by BN-PAGE, and protein complexes were visualized by Coomassie Brilliant blue staining.
leakage of ions was prevented. When P2 peptide was present, however, the TIM22 complex was preferentially in the rapid gating state, which was also true after prolonged voltage application (Fig. 2, A and B, and fig. S1C). Thus, the assembly of Tim22 channels into a multisubunit complex led to a coordinated regulation of the channel activities: only one out of two pores underwent rapid gating; at the same time, the sensitivity to the signal peptide was increased because the threshold voltage for induction of rapid gating was reduced by half. The structure of the purified TIM22 complex was investigated by single-particle electron microscopy. Assessment of the micrographs revealed different projection views (fig. S2A). The analysis of the dataset was performed by both hierarchical ascendant classification (fig. S2B) and the method of self-organizing maps (fig. S2, C and D). Both methods yielded a substantial proportion of particles with two stain-filled centers (fig. S2, B to D). Other particles showed connections between stain-filled centers or extensions of one center to the exterior of the particle (fig. S2, B and C). The processed images reveal-
ing two stain-filled centers were in excellent agreement with the electrophysiological observation of a twin-pore translocase. Singleparticle analysis of the translocase of the outer mitochondrial membrane (TOM) has indicated the presence of two or three pore-like structures (8–10). We compared the image density profiles of the TIM22 complex with the two-pore TOM core complex (of ⬃400 kD) (Fig. 3). The TIM22 complex was smaller (⬃110 Å in overall length) and exhibited a smaller pore diameter (⬃16 Å) than the TOM core complex (⬃25 Å), consistent with the electrophysiological results (5, 8, 9, 11). We asked how the activation of the purified TIM22 complex by membrane voltage and a carrier signal peptide corresponds to the actual transport of complete precursor proteins in organello. The import of the carrier precursors into mitochondria has been divided into five sequential stages (fig. S3): cytosolic transport form (stage I); binding to surface receptor (stage II); translocation across outer membrane and association with the small, soluble Tim proteins (Tim9-Tim10) in the intermembrane space (stage III); insertion into the inner membrane (stage IV); and assembly to the function-
14 MARCH 2003 VOL 299 SCIENCE www.sciencemag.org
REPORTS Fig. 2. The purified TIM22 complex contains two coupled pores. The TIM22 complex was isolated from mitochondria by means of IgG chromatography, eluted by TEV protease treatment, and subjected to BN-PAGE. After elution in buffer containing 1% N-heptyl-thioglucopyranoside, the complex was reconstituted into liposomes that were subsequently fused with a planar lipid bilayer. (A and B) Two coupled pores formed by the TIM22 complex responded differentially to an internal signal peptide of the phosphate carrier [P2: TSTTLLNLLSGLT (26)]. Current recordings were performed at the indicated membrane potentials in the absence or presence of 100 nM P2 peptide under symmetric buffer conditions. Gray boxes represent time scale– expanded traces of the above current recordings. The right panels of (B) represent mean variance plots of the current recordings, and they disclose gating events within single complexes in the presence or absence of P2 peptide at 150 mV. The plots were calculated from the complete corresponding current traces. (C) Quantification of gating frequency transitions per channel at different membrane potentials in the presence or absence of 100 nM P2 peptide (averages of 1-min recordings). For comparison, Tim22 that was expressed in Escherichia coli, purified, and renatured (Tim22expr) (5), was analyzed under similar conditions.
al, mature dimer (stage V) (12–15). Although stages I, II, III, and V have been analyzed in detail, the actual insertion into the inner membrane (stage IV) has remained elusive, because neither detailed import kinetics nor import at lower temperature led to an efficient accumulation of precursor proteins at stage IV. We gradually lowered the mitochondrial membrane potential by adding increasing concentrations of the protonophore CCCP (6) [in the presence of oligomycin (16)] and analyzed the import of the dicarboxylate carrier by BN-PAGE. To analyze the late stages of carrier import (from III to V), mitochondria were treated with proteinase K to remove surface-exposed precursors. At a high ⌬ (no CCCP), the carrier was mainly found in the mature dimeric form (Fig. 4A), whereas in the absence of a ⌬ (high concentration of CCCP), the lower-molecular-mass stage III
Fig. 3. Single-particle electron microscopy analysis of the purified TIM22 complex. Electron micrographs of the isolated TIM22 complex and the TOM complex (10) from yeast mitochondria were collected in 2% uranyl acetate. Upper panel, comparison of the two-pore class average images (for each, n ⫽ 252) of the TIM22 complex (derived from fig. S2B, class 8) and the TOM core complex. Scale bar, 10 nm. Lower panel, Image density profiles of the classes showing the two-pore density projection of the TOM core complex and the TIM22 complex.
www.sciencemag.org SCIENCE VOL 299 14 MARCH 2003
1749
REPORTS form was observed as expected (Fig. 4A) (13, 17). At intermediate levels of the membrane potential, however, a high molecular mass form of the radiolabeled carrier was observed (Fig. 4, A and B). The size of this putative stage IV– intermediate corresponded to the size of the TIM22 complex plus a carrier precursor. To obtain direct evidence for an association of the carrier with the TIM22 complex, we accumulated the precursor in mitochondria car-
rying ProtA-tagged Tim18 and purified the complex. The isolated TIM22 complex exclusively carried the stage IV intermediate and no other import stage of the carrier (Fig. 4B). Although, in fully energized mitochondria with a membrane potential of 150 mV or more, the carrier precursors were transported through the TIM22 complex (12, 18), the reduction of ⌬ permitted the observation of an intermediate at stage IV. The
membrane potential had to be below 60 mV for an efficient arrest of the carrier precursor in the TIM22 complex. The observed stage IV–arrested carrier protein was efficiently chased into its fully assembled, dimeric form upon restoration of the membrane potential, shown for both the dicarboxylate and phosphate carriers (Fig. 4C). Thus, the stage IV–arrested protein represented a productive translocation intermediate.
Fig. 4. Membrane potential (⌬)⫺dependent in organello accumulation of a translocation intermediate in the TIM22 complex. (A) Reduction of ⌬ leads to accumulation of a carrier precursor at defined transport stages. 35S-labeled dicarboxylate carrier was imported into isolated yeast mitochondria (50 g protein) in the presence of 1% bovine serum albumin and 20 M oligomycin and in the absence (lane 10) or presence (lanes 1 to 9) of decreasing concentrations of the protonophore (CCCP) (6). Mitochondria were treated with proteinase K and subjected to BN-PAGE and digital autoradiography (6). (B) Accumulation of the stage IV transport intermediate in the TIM22 complex. 35 S-labeled carrier was imported into wildtype and Tim18ProtA mitochondria. Valinomycin (1 M Val) or CCCP were present as indicated. After the import reaction, mitochondrial protein complexes were directly assessed by BN-PAGE (lanes 1 to 4, wild-type mitochondria) or a purification was performed (lanes 5 to 8, mock isolation from wild-type mitochondria; lanes 9 to 12, TIM22 complex isolated from Tim18ProtA mitochondria). (C) Carrier arrested at stage IV can be chased into stage V. 35S-labeled dicarboxylate carrier (lanes 1 to 6) or phosphate carrier (lanes 7 and 8) were accumulated at low membrane potential [40 M carbonyl cyanide p-trifluoromethoxyphenyl hydrazone (FCCP) (6)] in the presence of 1% BSA and 20 M oligomycin. Mitochondria were reisolated, resuspended in fresh import buffer containing NADH and ATP in the presence (no chase) or absence of FCCP (⫹ chase), and import was continued for the indicated times at 25°C. After import, samples were analyzed by BN-PAGE. (D) Specificity of carrier association with the TIM22 translocase. 35S-labeled carrier was imported into wild-type ( WT ) and Tim18ProtA ( Tag) mitochondria as described for lanes 5 to 12 of B. In addition, a mock import reaction without 35S-labeled carrier was performed. Before solubilization, wild-type and Tim18ProtA mitochondria were mixed as indicated. Asterisk, nonspecific radiolabeled protein frequently observed in BN-PAGE. (E) Tethered and docked precursor show differential salt resistance. 35S-labeled carrier was arrested in Tim18ProtA mitochondria in the presence of 30 M CCCP or valinomycin, solubilized in NaCl-containing buffer, and the TIM22 complex was isolated. Eluates were analyzed by BN-PAGE, and the amount of the stage IV intermediate was quantified by digital autoradiography.
1750
14 MARCH 2003 VOL 299 SCIENCE www.sciencemag.org
REPORTS A small, yet significant, fraction of the carrier was found in association with the TIM22 complex even in the complete absence of a membrane potential (addition of valinomycin) (Fig. 4B). To exclude a post-lysis association of the carrier with the TIM22 complex, we performed a mixing experiment. 35S-labeled carrier was imported into either wild-type mitochondria or mitochondria carrying ProtA-tagged Tim18. After the import reaction, the mitochondria were mixed and lysed with digitonin (Fig. 4D). The purified TIM22 complex carried the stage IV– carrier intermediate only when the radiolabeled precursor was imported into Tim18ProtA mitochondria (Fig. 4D), which demonstrates that the association of carrier with the TIM22 complex had occurred in intact mitochondria. We asked if the fraction of carrier associated with the TIM22 complex in the absence of a membrane potential (“tethered form”) was bound in a different way to the stage IV intermediate at low membrane potential (“docked form”). Perhaps in the absence of a ⌬, the carrier preferentially associates with the small Tim proteins of the TIM22 complex that interact with the hydrophobic segments of carrier proteins (19), whereas in the presence of a low ⌬, the carrier comes into contact with the pore where ionic interactions could also take place. The carrier can contact the small Tim proteins at the TIM22 complex (including Tim12) in the absence of a ⌬ (15, 20), whereas contact to Tim22 requires the membrane potential (21). Indeed, when the carrier was arrested at the different steps the tethered precursor was significantly more resistant to increasing ionic strength of the buffer than the docked precursor (Fig. 4E). Thus, tethered and docked forms of the carrier reside in different molecular environments at the TIM22 complex, the tethered protein probably being in a more hydrophobic environment. We conclude that the interaction of carrier proteins with the TIM22 complex occurred in three steps. (i) Tethering of the precursor to the translocase occurred in a ⌬-independent manner. (ii) The first of two voltage-dependent steps resulted in docking of the precursor in the translocase. A ⌬ below 60 mV, which did not influence the activity of the channel itself, was sufficient for the docking step, indicating that ⌬ was acting on the precursor protein. Because the matrix-exposed loops of the carrier proteins have a net positive charge (22, 23), an electrophoretic effect of the ⌬ on the translocation of the loops is conceivable (18, 24, 25). (iii) In the presence of a carrier signal peptide, a membrane potential above 70 mV then activated channel gating. The peptide P2 is uncharged; this excludes a direct effect of ⌬ on the peptide but indicates an influence of ⌬ on the channel itself. The P2 peptide is composed of part of an intermembrane space loop and the beginning of the fifth transmembrane segment of the carrier (7). Because the carrier module consisting of the fifth and sixth transmembrane segments with
the connecting matrix loop probably inserts first into the inner membrane (14), P2 will come into contact with Tim22 when most of this module has already been inserted. It is unlikely that P2 blocks the Tim22 pore by itself, but rather induces an activity of the TIM22 complex that leads to the next steps of insertion, i.e., the closing of one channel pore, possibly to keep the inserted transmembrane segments tightly in the translocase, and the concomitant rapid gating activity of the other channel pore to promote the insertion of additional transmembrane segments. Finally, the carrier protein is laterally released into the lipid phase of the inner membrane. The single energy source, ⌬, thus plays a dual role in protein insertion by acting on both the precursor protein and the channel. References and Notes
1. C. M. Koehler et al., Trends Biochem. Sci. 24, 428 (1999). 2. M. F. Bauer et al., Trends Cell. Biol. 10, 25 (2000). 3. R. E. Jensen, A. E. Johnson, Nature Struct. Biol. 8, 1008 (2001). 4. N. Pfanner, A. Geissler, Nature Rev. Mol. Cell Biol. 2, 339 (2001). 5. P. Kovermann et al., Mol. Cell 9, 363 (2002). 6. Materials and methods are available as supporting material on Science Online. 7. J. Brix et al., J. Biol. Chem. 274, 16522 (1999). 8. K. P. Ku¨nkele et al., Cell 93, 1009 (1998). 9. U. Ahting et al., J. Cell Biol. 147, 959 (1999). 10. K. Model et al., J. Mol. Biol. 316, 657 (2002).
11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
K. Hill et al., Nature 395, 516 (1998). N. Pfanner et al., Cell 49, 815 (1987). M. T. Ryan et al., J. Biol. Chem. 274, 20619 (1999). M. Endres et al., EMBO J. 18, 3214 (1999). C. M. Koehler et al., Science, 279, 369 (1998). A. Geissler et al., Mol. Biol. Cell. 11, 3977 (2000). V. Zara et al., J. Mol. Biol. 310, 965 (2001). N. Pfanner, W. Neupert, EMBO J. 4, 2819 (1985). S. P. Curran et al., EMBO J. 21, 942 (2002). C. Sirrenberg et al., Nature 391, 912 (1998). C. Sirrenberg et al., Nature 384, 582 (1996) B. El Moualij et al., Yeast 13, 573 (1997). D. R. Nelson et al., J. Mol. Biol. 277, 285 (1998). A. J. Davis et al., Mol. Biol. Cell 9, 2577 (1998). K. Ka ´ldi, M. F. Bauer, C. Sirrenberg, W. Neupert, M. Brunner, EMBO J. 17, 1569 (1998). 26. Single-letter abbreviations for the amino acid residues are as follows: G, Gly; L, Leu; N, Asn; S, Ser; and T, Thr. 27. We are grateful to M. Radermacher for advice and critical comments on the manuscript, to I. Perschil and C. Joppich for technical assistance, to E. Schiebel for PCR templates, and to C. Meisinger and T. Prinz for providing purified TOM complex. Supported by the Deutsche Forschungsgemeinschaft, the Sonderforschungsbereich 388, the Sonderforschungsbereich 431, the Fonds der Chemischen Industrie and BMBF. Supporting Online Material www.sciencemag.org/cgi/content/full/299/5613/1747/ DC1 Materials and Methods SOM Text Figs. S1 to S3 28 November 2002; accepted 24 January 2003
Asymmetric Inheritance of Oxidatively Damaged Proteins During Cytokinesis Hugo Aguilaniu,1,2,3 Lena Gustafsson,2 Michel Rigoulet,3 Thomas Nystro¨m1 Carbonylated proteins were visualized in single cells of the budding yeast Saccharomyces cerevisiae, revealing that they accumulate with replicative age. Furthermore, carbonylated proteins were not inherited by daughter cells during cytokinesis. Mother cells of a yeast strain lacking the sir2 gene, a life-span determinant, failed to retain oxidatively damaged proteins during cytokinesis. These findings suggest that a genetically determined, Sir2p-dependent asymmetric inheritance of oxidatively damaged proteins may contribute to free-radical defense and the fitness of newborn cells. In the yeast Saccharomyces cerevisiae, cell division is asymmetrical and some genetic material is unequally distributed (1). For example, the daughter cell does not inherit extrachromosomal ribosomal DNA circles (ERCs), which accumulate in mother cells during growth and have been suggested to cause replicative senes1 Department of Cell and Molecular Biology–Microbiology, Go¨teborg University, Go¨teborg, Sweden. 2Department of Molecular Biotechnology, Chalmers University of Technology, Go¨teborg, Sweden. 3Institut de Biochimie et de Ge ´ne ´tiques Cellulaires du CNRS, Universite ´ Victor Segalen, Bordeaux, France.
*To whom correspondence should be addressed. Email:
[email protected]
cence (2). Reports have shown that cells lacking the silent information regulator, Sir2p, an NAD (nicotinamide adenine dinucleotide)–dependent histone deacetylase (3), contain more ERCs than their wild-type counterparts and have a reduced replicative potential (4). It has also been suggested that the age asymmetry may depend on partition of undamaged cellular components to the progeny (5) and that old mother cells show markers of oxidative stress (6). However, no information is available on whether oxidative damage, like ERCs, accumulates as a function of replicative age. To determine whether protein carbonylation, an irreversible oxidative damage (7),
www.sciencemag.org SCIENCE VOL 299 14 MARCH 2003
1751
