Peering into the Birth Canal during Ion Channel Parturition

Neuron 214

nowski, 1991). However, a spindle defect is likely the primary cause since Feng and Walsh show that interference with Nde1 function in 293T and COS-7 cell lines impairs spindle orientation, possibly through a defect in the integrity of the centrosomes which anchor the mitotic spindle through microtubule connections to the cell cortex. Centrosomes also figure prominently in the study of the Nde1 homolog Ndel1 by Shu et al. In contrast to Nde1, Ndel1 is highly expressed by migrating neurons. Although there have been studies on the role of dynein in the orientation of the centrosome as fibroblasts migrate in vitro, most studies have focused on the microtubule connections from the centrosome to the leading edge of the cell (Etienne-Manneville and Hall, 2001; Palazzo et al., 2001). A migrating neuron translocates its nucleus/ cell soma after its leading process has extended in a manner similar to an axonal growth cone. Very little is understood about how the nucleus of a migrating neuron is translocated. The haploinsuffiency of LIS1 in a human neuronal migration defect indicates that dynein and its associated proteins must play a crucial role at some stage of somal translocation. Shu et al. first connect Lis1 and Ndel1 biochemically by examining the protein complexes that are formed in the presence and absence (via RNAi) of Ndel1 in tissue culture cells. Depletion of Ndel1 causes a slight decrease in the association between Lis1 and cytoplasmic dynein. Since Lis1 is postulated to be an activator of dynein, a deficiency in Ndel1 should have a similar effect as a Lis1 deficiency. Shu et al. find that individually inactivating Ndel1, Lis1, and the heavy chain of cytoplasmic dynein (DHC) using RNAi produces similar effects: migration is perturbed. Close examination of the transfected cells reveals a common anomaly, that the distance between the nucleus and the centrosome is increased. Based on these observations, Shu et al. propose a model that places Ndel1, Lis1, and DHC function at the nucleus, where the centrosome-directed DHC motor activity is anchored on the nuclear envelope and maintains a microtubule-based link between the nucleus and the centrosome. The centrosome acts as a hub for microtubules that are directed toward the nucleus and dynein motors its nuclear cargo along microtubules toward the centrosome. When taken together, the findings reported by Feng and Walsh and Shu et al. are remarkable in that inactivation of two separate but highly homologous genes disrupts two distinct phases of the life of a cortical neuron: cell division and migration. This is likely due to the complementary expression patterns of Nde1 and Ndel1. By studying dynein regulators that are differentially expressed, we are able to gain insight to specific functions in different milieus. A bountiful future awaits researchers at this border of cell biology and developmental neurobiology with further questions: How does Nde1 regulate spindle orientation, and what developmental cues trigger this regulation? What anchors dynein to the nucleus, enabling dynein to move the nucleus toward the centrosome? Finally, what is the mechanism for dynein’s “activation” by LIS1? Dynein is a highly cosmopolitan mechanoenzyme, appearing in diverse cellular functions, and LIS1 is proving to be a Rosetta stone for understanding

the cellular roles of dynein that underlie the development of the cerebral cortex. Bruce T. Schaar Department of Biological Sciences Stanford University Stanford, California 94305 Selected Reading Caviness, V.S., Jr., Takahashi, T., and Nowakowski, R.S. (1995). Trends Neurosci. 18, 379–383. Chenn, A., and McConnell, S.K. (1995). Cell 82, 631–641. Doe, C.Q., and Bowerman, B. (2001). Curr. Opin. Cell Biol. 13, 68–75. Efimov, V.P. (2003). Mol. Biol. Cell 14, 871–878. Etienne-Manneville, S., and Hall, A. (2001). Cell 106, 489–498. Feng, Y., Olson, E.C., Stukenberg, P.T., Flanagan, L.A., Kirschner, M.W., and Walsh, C.A. (2000). Neuron 28, 665–679. Feng, Y., and Walsh, C.A. (2004). Neuron 44, this issue, 279–293. Gupta, A., Tsai, L.H., and Wynshaw-Boris, A. (2002). Nat. Rev. Genet. 3, 342–355. Haydar, T.F., Ang, E., Jr., and Rakic, P. (2003). Proc. Natl. Acad. Sci. USA 100, 2890–2895. McConnell, S.K., and Kaznowski, C.E. (1991). Science 254, 282–285. Niethammer, M., Smith, D.S., Ayala, R., Peng, J., Ko, J., Lee, M.S., Morabito, M., and Tsai, L.H. (2000). Neuron 28, 697–711. Noctor, S.C., Martinez-Cerdeno, V., Ivic, L., and Kriegstein, A.R. (2004). Nat. Neurosci. 7, 136–144. Olson, E.C., and Walsh, C.A. (2002). Curr. Opin. Genet. Dev. 12, 320–327. Palazzo, A.F., Joseph, H.L., Chen, Y.J., Dujardin, D.L., Alberts, A.S., Pfister, K.K., Vallee, R.B., and Gundersen, G.G. (2001). Curr. Biol. 11, 1536–1541. Sasaki, S., Shionoya, A., Ishida, M., Gambello, M.J., Yingling, J., Wynshaw-Boris, A., and Hirotsune, S. (2000). Neuron 28, 681–696. Shu, T., Ayala, R., Nguyen, M.-D., Xie, Z., Gleeson, J.G., and Tsai, L.-H. (2004). Neuron 44, this issue, 263–277. Xiang, X., Osmani, A.H., Osmani, S.A., Xin, M., and Morris, N.R. (1995). Mol. Biol. Cell 6, 297–310.

Peering into the Birth Canal during Ion Channel Parturition Recent studies have provided detailed structures of the N-terminal T1 domain of Kv channel ␣ subunits that mediates contranslational subunit assembly. In this issue of Neuron, Kosolapov et al. probe T1 domain structure within the ribosomal tunnel. They find that the T1 domain forms secondary structure within the tunnel, in preparation for its immediate role in governing channel assembly upon exit. Voltage-gated potassium or Kv channels are key determinants of intrinsic electrical activity in excitable cells, and of crucial signaling events (insulin release, response to antigen, etc.) in a variety of nonexcitable cells. Kv channels are oligomeric membrane proteins, composed of four polytopic transmembrane ␣ subunits, which are the voltage-sensing and pore-forming constituents of

Previews 215

the channel complex, and up to four cytoplasmic auxiliary ␤ subunits. The precise ␣/␤ subunit composition directs diverse aspects of channel function, including but not restricted to the channel’s biophysical and pharmacological properties, and its intracellular trafficking and localization. The T1 domain present at the cytoplasmic N terminus of the ␣ subunits dictates the Kv ␣ and ␤ subunit composition of these oligomeric membrane protein complexes (Deutsch, 2003). In spite of overall structural similarity between T1 domains of different Kv subfamilies (Nanao et al., 2003), this domain serves as the recognition determinant that restricts ␣ subunit assembly to within a Kv channel subfamily (i.e., Shaker or Kv1 subfamily ␣ subunits can only assemble with one another, and not with Shab/Kv2, Shaw/Kv3, or Shal/Kv4 ␣ subunits, and vice versa [Deutsch, 2003]). The T1 domains of the Kv1 and Kv4 ␣ subunits also direct efficient subfamily-specific assembly with cytoplasmic Kv␤ and KChIP subunits, respectively (Gulbis et al., 2000; Scannevin et al., 2004). The fact that these ␣-␣ and ␣-␤ subunit assembly events occur cotranslationally (Deal et al., 1994; Shi et al., 1996), while the nascent ␣ subunits are still attached to the polyribosome complex, as well as the fact that these oligomeric complexes are thereafter irreversibly assembled, underlies the importance of having the T1 domain of the ␣ subunits achieve a correct structure as soon as possible after emergence from the ribosome tunnel. In this issue of Neuron, Carol Deutsch and her colleagues describe a series of elegant experiments investigating the nature of T1 domain structure within the ribosomal tunnel (Kosolapov et al., 2004), which can be thought of as a protein’s birth canal. As with any birth, this terminal transit event is the culmination of a long series of preceding events that together represent an essential and highly evolved developmental program. For membrane proteins such as ion channels and receptors, these events include reliable gene replication without mutation, flawless transcription, correct processing of nascent mRNAs, efficient docking of mRNA to ribosome and ribosome to ER membrane, and accurate translation of the primary structure. Once translation ensues, not only is the successful passage of the nascent polypeptide down the ribosomal tunnel toward the exit port minimally required for the successful termination of the prepartum portion of early life, but also the events that occur during the passage may have a longterm impact on the overall functional characteristics of the progeny. In this issue of Neuron, Kosolapov et al. have provided dramatic insights into structural changes that occur within the T1 domain as it transits the ribosome tunnel. These innovative studies provide novel insights into the folding of a critical domain of an ion channel while it is inside the ribosome complex. Deutsch and her colleagues cleverly recognized that the ribosomal tunnel and the pore domains of ion channels are “conceptual cousins.” They turned the tables on the ribosomal tunnel and its included T1 domain by successfully applying techniques developed to study the transmembrane pores of mature ion channels. This allowed them to probe the pore-like structure of the ribosomal tunnel and changes in the channel T1 domain while it is within the tunnel itself. This approach builds upon the use of

a cysteine-specific protein modification reagent, PDM, as a precise label for engineered cysteines on otherwise cysteine-less T1 domains, which can then be used to systematically examine the accessibility of these sites. PDM enters through the normal exit port of the ribosomal tunnel and because of steric considerations has a limited penetration into the 100 A˚ long/10–20 A˚ wide tunnel (Kramer et al., 2001). As such, accessibility of these labeling agents to target sites on the T1 domain is dependent on the distance of these sites from the exit port of the tunnel. The nature of the polypeptide chain (extended or compact) can then be inferred from the degree of accessibility of the labeling reagent. These investigators generated a series of T1 domain cDNAs of different length, all lacking termination codons such that when mRNAs transcribed from these cDNAs are translated in vitro, translation arrests at the last (nontermination) codon, and the nascent polypeptide chain remains attached to the ribosome at that fixed position. The polypeptide-ribosome complex can then be subjected to labeling with the cysteine modification reagents to assess whether it exists in an extended or compact structure. The T1 domain was the first portion of mammalian ion channel for which high-resolution structure was determined using X-ray crystallographic techniques (reviewed in Roosild et al., 2004). As such, Kosolapov et al. could target specific sites with known locations within the structure of the folded T1 domain and address where these sites were located relative to the tunnel exit port. They found convincing evidence for acquisition of T1 secondary structure within the tunnel itself. As such, the T1 domain is already prepared to undergo tertiary folding events upon emergence from the tunnel exit pore. This precocious folding within the tunnel may be necessary to achieve the rapid and precise tertiary structure necessary for accurate contranslational discrimination of different ␣ and ␤ subunits. Accurate assembly ensures that the subunit composition of the resultant channels reflects the relative abundance of available nascent polypeptides, and by extension the relative levels of gene expression. Misregulation of channel assembly could have profound consequences for cell excitability, as the precise subunit composition of these oligomeric Kv channels has profound effects on diverse aspects of function, intracellular trafficking, and localization (Deutsch, 2003). The recent explosion in detailed information from high-resolution crystal structures of both prokaryotic and eukaryotic ribosomes has provided significant insights into ribosomal function. Moreover, a growing body of literature has examined the dynamic events that are associated with folding of nascent polypeptide chains within the ribosome tunnel (reviewed in Kramer et al., 2001). It is immediately apparent that there exist differences in the details of these previous studies and the findings of Kosolapov et al. However, the results of these other studies are somewhat variable based on the details of the ribosomal preparation employed, the method of assaying folding (e.g., the acquisition of enzymatic activity, fluorescence resonance energy transfer [FRET], and covalent labeling). However, when comparing the results of Kosolapov et al. to these previous studies it is perhaps more important to remember that Kv channel

Neuron 216

␣ subunits (as well as other voltage-gated cation channels and many transport proteins) are somewhat odd beasts in the spectrum of membrane protein structure. The extended N-terminal (i.e., initially translated) domain of these polypeptides is cytoplasmic and lacks the very N-terminal canonical signal sequence that confers immediate association of classical type I membrane proteins with signal recognition particles and the endoplasmic reticulum translocon apparatus. Thus, it is likely that the translation of the Kv channel ␣ subunit N-terminal domain occurs on free ribosomes in the cytoplasm. Only when the first transmembrane or S1 domain, which contains the initial start transfer sequence (Sato et al., 2002), emerges from the ribosome tunnel port can the ribosome become associated with the signal recognition particle, the cytoplasmic face of the endoplasmic reticulum, and the translocon apparatus. Thus, the events that surround translation of the initial domains of Kv channel ␣ subunits represent a spatial and temporal hybrid between those events that occur for both bona fide cytoplasmic proteins and signal sequence-containing transmembrane type I polypeptides. Mutations in ion channel genes are the basis for a number of inherited neurological disorders, termed channelopathies (Ptacek, 2002). At least some of these mutations, for example, those in the Kv1.1 ␣ subunit that lead to episodic ataxia type I, involve protein misfolding (Manganas et al., 2001). While to date no EA-1 or other channelopathic mutations have been localized to the Kv channel T1 domains, mutations in putative C-terminal assembly domain of the related KCNQ ␣ subunits are the most common channelopathic KCNQ mutations (Singh et al., 2003). The role of the T1 domain in mediating the subunit assembly events that irrevocably define the subunit composition of Kv channels is crucial in defining diverse aspects of channel function, cell surface abundance, and subcellular localization. The paper by Kosolapov et al. provides the first insights into the folding events that shape T1 domain structure in the final moments before it exits the ribosome and is called upon to correctly fulfill its critical role. Future studies employing real-time analyses of protein folding, using FRET, for example, will lead to further insights into the dynamic aspects of these events and how they lead to the final structure of this interesting and important membrane protein functional domain. James S. Trimmer Department of Pharmacology School of Medicine University of California, Davis Davis, California 95616 Selected Reading Deal, K.K., Lovinger, D.M., and Tamkun, M.M. (1994). J. Neurosci. 14, 1666–1676. Deutsch, C. (2003). Neuron 40, 265–276. Gulbis, J.M., Zhou, M., Mann, S., and MacKinnon, R. (2000). Science 289, 123–127. Kosolapov, A., Tu, L., Wang, J., and Deutsch, C. (2004). Neuron 44, this issue, 295–307. Kramer, G., Ramachandiran, V., and Hardesty, B. (2001). Int. J. Biochem. Cell Biol. 33, 541–553.

Manganas, L.N., Akhtar, S., Antonucci, D.E., Campomanes, C.R., Dolly, J.O., and Trimmer, J.S. (2001). J. Biol. Chem. 276, 49427– 49434. Nanao, M.H., Zhou, W., Pfaffinger, P.J., and Choe, S. (2003). Proc. Natl. Acad. Sci. USA 100, 8670–8675. Ptacek, L. (2002). Novartis Found Symp. 241, 87–104; discussion 104–108, 226–132. Roosild, T.P., Le, K.T., and Choe, S. (2004). Trends Biochem. Sci. 29, 39–45. Sato, Y., Sakaguchi, M., Goshima, S., Nakamura, T., and Uozumi, N. (2002). Proc. Natl. Acad. Sci. USA 99, 60–65. Scannevin, R.H., Wang, K., Jow, F., Megules, J., Kopsco, D.C., Edris, W., Carroll, K.C., Lu¨, Q., Xu, W., Xu, Z., et al. (2004). Neuron 41, 587–598. Shi, G., Nakahira, K., Hammond, S., Rhodes, K.J., Schechter, L.E., and Trimmer, J.S. (1996). Neuron 16, 843–852. Singh, N.A., Westenskow, P., Charlier, C., Pappas, C., Leslie, J., Dillon, J., Anderson, V.E., Sanguinetti, M.C., and Leppert, M.F. (2003). Brain 126, 2726–2737.

Connecting the Dots: From Actin Polymerization to Synapse Formation

Protrusive behavior of dendritic spines on developing neurons has been previously suggested to mediate the formation of new axodendritic synaptic contacts. A study by Zito et al. in this issue of Neuron links actin polymerization in dendritic spines with the motility that the spines exhibit and the synapses that they form. Understanding the mechanisms that regulate the formation of synaptic contacts is of interest in both normal and pathological contexts. Dendritic spines, small protrusions that stud dendrites, bear the majority of excitatory input in the central nervous system. Advances in microscopy now allow the visualization of these synaptic structures and the subcellular signaling in them. Such studies indicate that, during postnatal development, dendritic spines exhibit highly protrusive motility (Dailey and Smith, 1996; Lendvai et al., 2000; Grutzendler et al., 2002). This developmentally regulated spine behavior was proposed to mediate the formation of new contacts between dendrites and axons (Ziv and Smith, 1996). The cellular mechanism that powers the motility of dendritic spines has been shown to be the polymerization of actin filaments (Fischer et al., 1998; Dunaevsky et al., 1999). These discrete findings suggest that polymerization of dendritic spines leads to spine motility, which in turn leads to the formation of synaptic contacts. How does one causally link spine motility with synaptogenesis? One way to address this problem would be to alter spine motility in a subset of developing neurons and test the effect on the formation of synaptic contacts. A paper in this issue of Neuron from the lab of Karel Svoboda (Zito et al., 2004) describes exactly such an experiment. The basic approach involved a manipulation of actin polymerization in neurons followed by the analysis of spine and synapse morphogenesis and func-