The Dimerization Mechanism of LIS1 and its Implication for Proteins Containing the LisH Motif

doi:10.1016/j.jmb.2006.01.002

J. Mol. Biol. (2006) 357, 621–631

The Dimerization Mechanism of LIS1 and its Implication for Proteins Containing the LisH Motif Agnieszka Mateja1, Tomasz Cierpicki1,2, Marcin Paduch1 Zygmunt S. Derewenda2 and Jacek Otlewski1,2* 1

Laboratory of Protein Engineering, Institute of Biochemistry and Molecular Biology, University of Wroclaw Tamka 2, 50-137 Wroclaw Poland 2

Department of Molecular Physiology and Biological Physics, University of Virginia Charlottesville, VA 22908 USA

Miller-Dieker lissencephaly, or “smooth-brain” is a debilitating genetic developmental syndrome of the cerebral cortex, and is linked to mutations in the Lis1 gene. The LIS1 protein contains a so-called LisH motif at the N terminus, followed by a coiled-coil region and a seven WD-40 repeat forming b-propeller structure. In vivo and in vitro, LIS1 is a dimer, and the dimerization is mediated by the N-terminal fragment and is essential for the protein’s biological function. The recently determined crystal structure of the murine LIS1 N-terminal fragment encompassing residues 1–86 (NLIS1) revealed that the LisH motif forms a tightly associated homodimer with a four-helix antiparallel bundle core, while the parallel coiled-coil situated downstream is stabilized by three canonical heptad repeats. This homodimer is uniquely asymmetric because of a distinct kink in one of the helices. Because the LisH motif is widespread among many proteins, some of which are implicated in human diseases, we investigated in detail the mechanism of N-LIS1 dimerization. We found that dimerization is dependent on both the LisH motif and the residues downstream of it, including the first few turns of the helix. We also have found that the coiledcoil does not contribute to dimerization, but instead is very labile and can adopt both supercoiled and helical conformations. These observations suggest that the presence of the LisH motif alone is not sufficient for highaffinity homodimerization and that other structural elements are likely to play an important role in this large family of proteins. The observed lability of the coiled-coil fragment in LIS1 is most likely of functional importance. q 2006 Elsevier Ltd. All rights reserved.

*Corresponding author

Keywords: homodimerization; protein structure; neuronal migration; lissencephaly; protein stability

Introduction The Miller-Dieker lissencephaly, or “smoothbrain”, is a developmental syndrome caused by mutations in an autosomal gene Lis1.1 The gene codes for a protein which contains an N-terminal motif (residues 8–37) found in numerous eukaryotic proteins and denoted LisH (LIS1-homology motif),2 followed by a coiled-coil region and a seven-blade b-propeller domain found in a number of signaling proteins of the WD-40 family.3 LIS1 forms homodimers, and is implicated in interactions with other Abbreviations used: PAF-AH, PAF-acetylhydrolase; a.a., amino acid residue(s); GdmCl, guanidinium chloride; TFE, trifluoroethanol; HSQC, heteronuclear single quantum coherence. E-mail address of the corresponding author: [email protected]

proteins including the catalytic homo- and heterodimers of the brain PAF-acetylhydrolase (PAFAH).4 The homodimerization of LIS1 is essential for its biological function. Heterozygous mice lacking exon 1 (residues 1–63) show a typical lissencephaly phenotype, and the mutant protein no longer interacts with the PAF-AH catalytic subunits.5 Recently, two crystallographic studies provided the first insight into the molecular architecture of LIS1. The first described a high-resolution structure of a homodimer of the N-terminal fragment of murine LIS1 (referred to hereinafter as N-LIS1), encompassing residues 1–86 (PDB, 1UUJ),6 and the second described the complex of LIS1 with the catalytic a2-homodimer of the PAF-AH (PDB, 1VYH).7 The homodimeric structure of N-LIS1 is surprisingly asymmetric, because the two helices destined to form a coiled-coil at the C-terminal end,

0022-2836/$ - see front matter q 2006 Elsevier Ltd. All rights reserved.

622 originate with their axes set at an angle of w558, so that one needs to form a distinct kink to allow for a parallel alignment downstream. The coiled-coil fragment, visualized by the crystal structure, had been originally predicted on the basis of the presence of three canonical heptad repeats within the amino acid sequence (abcdefg)3, where a and d are hydrophobic amino acids, while e and g carry opposite charges.8 These heptads indeed form a hydrophobic “zipper”, which is vital for the ˚ integrity of the coiled-coil. Interestingly, the 3.4 A resolution structure of the full-length LIS1 in complex with the PAF-AH shows no interpretable electron density for the N-LIS1 fragment, 7 suggesting that this fragment is disordered in the crystals. The N termini of the two b-propeller ˚ apart, implying that domains of LIS1 are w60 A parallel supercoiling upstream of these domains is virtually impossible. This suggests a possibility that the N-LIS1 domain may alternate between closed (i.e. coiled-coil) and open conformations, and that this flexibility has a functional role. The LisH (LIS1-homology) sequence motif (residues 8–37) is found in numerous proteins in all eukaryotic genomes, including human. At least three proteins containing the LisH motif are implicated in genetic syndromes: the transducin b-like 1X (TBL1), which causes ocular albinism with late-onset sensorineural deafness;9 the OFD1, involved in oral-facial-digital syndrome type 1;10 and TCOF1, implicated in the TreacherCollins-Franeschetti syndrome.11 TBL1 contains a LisH domain (a.a. 54–87), an F-box, and seven WD-40 repeats in its C-terminal region. Mutations in the fly ortholog, Ebi, affect multiple processes including neuronal differentiation through the epidermal growth factor receptor pathway.12,13 The OFD1 syndrome is an X-linked dominant disease that is lethal in males and in females it is characterized by malformations of the face, oral cavity, and digits, and by a highly variable presentation. The phenotype may also include mental retardation or polycystic kidney disease.10 The gene product mutated in this disease, OFD1, contains both a LisH motif (a.a. 69–102) and several coiled-coil domains. Considering the ubiquitous nature of the LisH motif and its potential biological significance, we investigated in detail the LisH-dependent mechanism of homodimerization of N-LIS1. Using NMR, spectroscopic techniques and chemical denaturation, we show that the LisH motif accounts for only a part of the free energy of the homodimerization of N-LIS1, while the other part originates primarily from a single residue downstream of the LisH motif, Trp55. By contrast, the coiled-coil fragment is very labile, and has the ability to alternate between “closed” and “open” conformations depending on the ionic strength of the solution and possible other factors in vivo. These results illustrate the complexity of the intermolecular interactions involving the LisH-containing proteins.

Dimerization of LIS1

Results N-LIS1 dimer The crystal structure of the N-LIS1 fragment6 is shown in Figure 1(a). Briefly, the LisH motif, built of two short helices, forms a dimer of tightly packed four-helix bundles with a clearly defined hydrophobic core. A comparison of various LIS1 sequences shows that the amino acid conservation pattern corresponds to the location of buried, hydrophobic residues (Figure 1(a)), suggesting that the overall tertiary structure of the homodimer is preserved in other proteins similar to LIS1 (Figure 1(b)). Downstream of residue Leu37 the structure forms a solvent-exposed loop with a labile conformation, and with some helical content between residues Asn40 and Leu50. The sequence, which initiates the coiled-coil, i.e. LEKKWT, is the most conserved hexapeptide in all known LIS1 homologues from fungi to man, with only one conservative change (MEKKWT) in Ciona intestinalis. This oligopeptide adopts an a-helical conformation upstream of the three heptad repeats and the axes of the two helices form an angle of w558 (open scissors). The helices run across the homodimer interface so that the side-chains of Trp55 actually cross over to adjacent monomers. Downstream of Trp55 one of the helices forms a dramatic kink, so that Thr56 and Ser57 adopt non-helical

Figure 1. (a) Overall structure of the N-LIS1 fragment represented as a “worm” ribbon colored by hydrophobic scale (red to blue) and depicted by varying worm thickness (thick hydrophobic to thin hydrophilic). (b) Alignment-based phylogenetic conservation of residues in the LIS1 family mapped onto the LIS1 structure (green conserved, yellow variable). Conservation is illustrated by variability in worm thickness, where thin means conserved and thick denotes variable. The hydrophobic residues coincide with conserved regions in LIS1. High residue conservation could be observed at the beginning of the coiled-coil structure covering the Trp55 residue. Structure used to generate (b) was prepared on the ConSurf server based on the PSI-BLAST similarity search with the E-value inclusion threshold of 2!10K3 (http:// consurf.tau.ac.il).32 Both Figures were generated with Chimera software using the Render by Attribute extension (http://www.cgl.ucsf.edu/chimera).35

623

Dimerization of LIS1

Table 1. Thermodynamic parameters for the GdmCl-induced unfolding of N-LIS1 fragments and mutants 0.5 mM Construct/mutant N-Lis1(1–79) WT [GdmCl]1/2 (M) m (kcal/mol/M) H2 O (kcal/mol) DGU N-Lis1(1–57) WT [GdmCl]1/2 (M) m (kcal/mol/M) H2 O (kcal/mol) DGU N-Lis1(1–48) WT [GdmCl]1/2 (M) m (kcal/mol/M) H2 O (kcal/mol) DGU N-Lis1(1–79) T56A S57A [GdmCl]1/2 (M) m (kcal/mol/M) H2 O kcal/mol) DGU N-Lis1(1–79) W55F [GdmCl]1/2 (M) m (kcal/mol/M) H2 O (kcal/mol) DGU N-Lis1(1–79) W55A [GdmCl]1/2 (M) m (kcal/mol/M) H2 O (kcal/mol) DGU

2 mM

5 mM

FL at 341 nm

CD at 222 nm

FL at 341 nm

CD at 222 nm

FL at 341 nm

CD at 222 nm

2.81 3.42 18.21

2.87 3.85 19.63

3.05 4.84 22.54

3.06 4.66 22.06

3.18 4.80 22.56

3.2 4.49 21.61

2.82 3.62 19.00

2.87 3.71 19.41

3.05 3.98 19.93

3.09 4.00 20.11

3.25 3.81 19.81

3.14 3.82 19.27

n.d. n.d. n.d.

0.94 1.91 10.42

n.d. n.d. n.d.

1.39 2.17 10.75

n.d. n.d. n.d.

1.81 2.34 11.46

3.22 3.81 20.85

3.25 3.99 21.58

3.36 4.35 22.46

3.35 4.22 21.98

3.49 4.01 21.30

3.49 4.01 21.29

n.d. n.d. n.d.

2.13 4.23 17.55

n.d. n.d. n.d.

2.31 3.10 14.92

n.d. n.d. n.d.

2.45 2.68 13.81

n.d. n.d. n.d.

1.15 2.00 11.01

n.d. n.d. n.d.

1.74 2.12 11.18

n.d. n.d. n.d.

2.21 2.47 12.63

n.d. not determined; WT, wild-type.

secondary structures. In the second monomer these residues are a part of a contiguous helix. As a result, a coiled-coil structure is formed with one blade of the open scissors broken and aligned along the other. Previously, we showed that the N-LIS1 dimer is thermodynamically very stable and unfolds according to N2 (native dimer) 42 D (two denatured H2 O of denaturation monomers) mechanism with DGU 6 w22 kcal/mol. This value represents the Gibbs energy changes for monomers’ unfolding and dimer dissociation reactions.14 In the dimer the ˚ 2 of the buried LisH motifs account for w1900 A surface, while the coiled-coil, including Trp55, ˚ 2. To dissect the role played accounts for w1500 A by these two segments in dimerization, we studied them separately. We used a slightly shorter construct than the one used for crystallization (i.e. residues 1 to 79) as the crystal structure shows that the C termini beyond Phe79 are disordered. As expected, the stability of the 1-79 fragment is nearly identical to the longer one (Table 1; Figure 2). The stabilities of all the different N-LIS1 fragments and mutants used in this study were analyzed according to the N242D model.14 We compared chemical denaturation curves monitored by CD and fluorescence signal at different protein concentrations. In general, the two methods provided very similar denaturation parameters (Table 1). With an increase of protein concentration we always observed a shift of [GdmCl1/2] towards higher values. Finally, analytical size exclusion chromatography showed a constant molecular mass, as expected for the dimer, regardless of

protein concentration (0.5 to 200 mM range) (data not shown). The only exception was the isolated LisH fragment, which was dimeric up to 5 mM and at higher concentration gradually shifted towards tetramer. Isolated LisH To analyze the dimerization of the LisH motif alone, we expressed the fragment containing residues 1–48. This fragment excludes the

Figure 2. Normalized GdmCl-induced unfolding transitions of N-LIS1 (1–79) wild-type (WT) at three concentrations: 0.5 mM (diamonds), 2 mM (circles) and 5 mM (squares). The unfolding transition was monitored by fluorescence at 341 nm (open symbols) and far-UV (filled symbols). All curves were fitted to a two-state dimer denaturation equation, as described in Materials and Methods.

624

Dimerization of LIS1

Figure 3. (a) Far UV CD spectra of N-LIS1 fragments: N-LIS1(1–48) (filled triangles), N-LIS1(1–57) (open squares) and N-LIS1(1–79) (filled circles) in 20 mM sodium phosphate containing 150 mM NaCl (pH 7.4). (b) Normalized GdmClinduced unfolding transitions of N-LIS1 (1–48) (triangles), N-LIS1 (1–57) (squares) and N-LIS1 (1–79) WT (circles), at 2 mM concentration. The unfolding transition was monitored by fluorescence at 341 nm (open symbols) and far-UV CD (filled symbols). Unfolding of N-LIS1 (1–48) was monitored only by far-UV CD (due to lack of Trp55). All curves were fitted to a two-state dimer denaturation equation, as described in Materials and Methods.

N-terminal portion of the coiled-coil helix, and specifically eliminates the interactions mediated by Trp55, but includes the “linker” region between LisH and the coiled-coil. The circular dichroism (CD) spectrum of this fragment showed w70% helical content, in a good agreement with the crystal structure (Figure 3(a)). To assess the thermodynamic stability of the LisH dimer, we used chemical (guanidinium chloride) denaturation. We found H2 O value for LisH fragment is that the DGU w10.5 kcal/mol (Figure 3(b)). This left the possibility that either a portion of, or the whole coiledcoil segment, contributes significantly to stability. Coiled-coil derived peptides Interestingly, the CD spectrum of an isolated putative coiled-coil fragment (residues 58–79) showed primarily random coil properties (Figure 4(a)). Since the peptide contains three canonical heptads, we were surprised that it does not form a coiled-coil in solution. To check if solvent could affect the peptide’s conformation, we dissolved the peptide in 1.7 M ammonium sulfate, 100 mM sodium acetate (pH 4.5), i.e. the solvent similar to that used to crystallize N-LIS1 (1.7 M ammonium sulfate, 100 mM sodium citrate (pH 4.5)). Under those conditions, the peptide showed CD spectrum characteristic for the coiledcoil conformation, with the ratio of molar ellipticities at 222 nm and 208 nm ([q222/q208]) equal to 1.001 (Figure 4(a)).15–17 A more detailed assessment of the dependence of molar ellipticity on ammonium sulfate concentration (Figure 4(b) and (c)) shows that the conversion to a fully superhelical conformation occurs above 2 M ammonium sulfate. In addition, the presence of an isodichroic point at 203 nm suggests that this transition is a two-state process, i.e. involving random coil and coiled-coil species only.18 In contrast, 50% (v/v) trifluoroethanol (TFE) induced O90% a-helical con-

tent ([q 222 /q 208]Z0.84) but did not induce supercoiling, as expected from similar studies (Figure 4 (a)).19,20 Solvent-dependent behavior of coiled-coil fragment within N-LIS1 Since the isolated fragment encompassing residues 58–79 is unable to form a coiled-coil under low salt conditions, we asked if it could form a coiledcoil in the context of the intact N-LIS1 fragment. We subtracted the CD spectrum of N-LIS1 (1–57) from that of N-LIS1 (1–79), and we found that in 150 mM NaCl the difference CD spectrum was consistent with the presence of an identifiable fraction of coiled-coil ([q222/q208]Z0.94) (Figure 5). However, when analogous spectra were recorded and subtracted from samples dissolved in 1.7 M ammonium sulfate, a difference CD spectrum indicated the presence of primarily supercoiled species ([q222/q208]Z1.065)21,22 (Figure 5). Thus, within the context of the N-LIS1 fragment, the C-terminal fragment is helical, but supercoiling is induced only in the presence of concentrated NaCl or ammonium sulfate. This transition seems to be dependent on the sequence of the “kink”, since mutation of the critical Thr56 and Ser57 residues to Ala impedes the supercoiling (see below). In order to gain independent insight into the conformational equilibrium of N-LIS1 in solution we recorded a set of 1H–15N heteronuclear single quantum coherence (HSQC) NMR spectra at various ionic strengths. A salt (NaCl)-dependent perturbation of chemical shifts was clearly seen in the experimental data. For the salt-free sample (50 mM Tris buffer), we observed 65 out of 86 possible signals corresponding to backbone protons, although not all resonances could be resolved due to overlap and inadequate dispersion of chemical shifts (Figure 6(a)). Aside from backbone protons, we could identify signals corresponding to

Dimerization of LIS1

625

Figure 5. Difference spectra obtained after subtraction of the molar CD spectrum of N-LIS1 (1–57) from the molar CD spectrum of N-LIS1 (1–79) and normalized for a number of residues in 20 mM sodium phosphate (pH 7.4) containing 150 mM NaCl (open triangles) and additionally in 1.7 M ammonium sulfate (open circles). Difference spectra obtained after subtraction of the molar CD spectrum of N-LIS1 (1–57) from the molar CD spectrum of N-LIS1 (1–79) mutant Thr56Ala, Ser57Ala and normalized for a number of residues in 20 mM sodium phosphate (pH 7.4) containing 150 mM NaCl (filled circles) and additionally 1.7 M ammonium sulfate (filled triangles).

Figure 4. (a) The open squares line corresponds to the synthetic peptide (58–79) in buffer (20 mM sodium phosphate (pH 7.4), 1.7 M ammonium sulfate), whereas the open circles line represents peptide in 50% TFE opposed to 2.6 M ammonium sulfate (filled circles line). The top filled triangle line is for the peptide in 20 mM sodium phosphate (pH 7.4) alone. (b) Far UV CD spectra of synthetic peptide (58–79) in 20 mM sodium phosphate (pH 7.4) and 150 mM NaCl in variable concentrations of ammonium sulfate. The upper curve corresponds to the absence of ammonium sulfate (open circles), whereas the lower curve (filled circles) corresponds to 2.6 M ammonium sulfate. Each successive curve represents the peptide (58–79) spectrum in the presence of the following ammonium sulfate concentrations: 0.2, 0.6, 1.0, 1.2, 1.4, 1.6, 1.7, 2.0, 2.2, 2.4, 2.6 M. (c) The dependence of synthetic peptide helicity on ammonium sulfate concentration.

six out of seven possible side-chain amides as well as a single peak for the H3 proton of Trp55, which could be assigned with high confidence, serving as a convenient probe in our experiments. This set of resonances is consistent with the presence of a single conformer, which we believe to be the “open”, symmetric dimer (Figure 6(b)). The sample in 50 mM Tris and 50 mM NaCl shows an additional set of signals with lower intensity, consistent with the appearance of a minor fraction of a second conformer in slow exchange with the former. This second set is most likely to originate from the closed, i.e. asymmetric dimer. At higher salt concentration (i.e. 150 mM NaCl), the two conformations reach equilibrium so that the two sets of signals are of equal intensity (Figure 6(c) and (d)). A further increase of salt concentration (500 mM NaCl) results in the disappearance of the original set of signals, while the second one remains, including 71 peaks for backbone amides, one Trp55 H3 proton signal, and six for side-chain amides. Because the crystals of the asymmetric dimer of N-Lis1 were obtained from solutions of 1.7 M (NH4)2SO4, we also measured the 1H–15N HSQC spectrum under those conditions, in addition to 150 mM NaCl and the Tris buffer (Figure 6). The spectrum shows the same single set of signals as those observed for high NaCl concentration. Although we believe that at higher ionic strengths the N-LIS1 fragment forms an asymmetric dimer as seen in the crystal structure, this dimer undergoes fast exchange (i.e. the A and B molecules switch conformations) on the NMR

626

Dimerization of LIS1

Figure 6. Conformational equilibrium of N-LIS1 monitored by NMR spectroscopy. (a) Superposition of 1H–15N HSQC spectra of N-LIS1 in 50 mM Tris and in the absence of salt (blue), in 150 mM NaCl (black) and 150 mM NaCl with 1.7 M (NH4)2SO4 (red). (b) Model of conformational equilibrium of N-LIS1 consistent with NMR data; the open conformation is stable in the absence of salt while the closed conformation, involving an asymmetric homodimer exchanging alternate conformations on a fast time-scale, is stabilized by high ionic strength. (c) Examples of isolated peaks consistent with conformational equilibrium; the color coding is identical to that in (a). (d) Enlarged section of the 1H–15N HSQC spectra (box in (a)) showing salt-dependent chemical shifts of the indole proton of Trp55; the color coding is identical as in (a) with addition of spectrum measured in 50 mM Tris and 50 mM NaCl shown in green; a single backbone amide that does not experience conformational equilibrium is shown for comparison.

timescale, resulting from a single averaged set of resonances (Figure 6(b)). Trp55 mutants Since the region corresponding to the coiled-coil in the crystal structure is not super-coiled in low ionic-strength solution, the question arises why the thermodynamic stability of N-LIS1 is twice that of isolated LisH. The structure suggests that Trp55 ˚ 2 of its monomer plays a critical role. It buries 215 A accessible surface upon dimer formation. When Trp55 was mutated to Phe and Ala, a large drop in stability to 15 and 11 kcal/mol, respectively, could be observed (Table 1; Figure 7(a)). Consequently, the stability of N-LIS1 truncated after Trp55 (1–57 construct) is similar to that of the entire N-LIS1 (1–79) (Table 1). Thr56Ala/Ser57Ala mutant As we have shown earlier6 the parallel orientation of helices in coiled-coil is possible

due to the presence of a distinct kink, in the highly conserved Thr56-Ser57 sequence in one of the helices, introducing asymmetry in the LIS1 dimer. This kink seems to result from rather poor helix-forming propensities of these two amino acid residues.23 However, we showed above that the coiled-coil is formed only in the high concentration of ammonium sulfate and in low salt buffer the symmetric a-helices are formed. To probe the role played by the kink sequence, we prepared a double mutant Thr56Ala/Ser57Ala of N-LIS1 (1–79). The mutant is more stable than the wild-type and shows a slightly shifted value for [GdmCl]1/2, w0.3 M, in agreement with a higher helical propensity of alanine (Figure 7(b)). When the CD spectrum of N-LIS1 (1–57) was subtracted from that of N-LIS1 (1–79) the resulting difference CD spectrum was consistent with a-helical structure both in low salt buffer and in 1.7 M ammonium sulfate, in contrast to the situation observed for the wild-type where a difference CD spectrum in 1.7 M ammonium sulfate indicated the presence of a coiled-coil (Figure 5).

Dimerization of LIS1

627

Figure 7. Chemical-induced denaturations of N-LIS1 mutants. (a) Normalized GdmCl-induced unfolding transitions of N-LIS1 (1–79) WT (circles), N-LIS1 (1–79) mutant Trp55Ala (open squares) and mutant Trp55Phe (filled triangles) at 2 mM concentration. Unfolding of N-LIS1 (1–79) WT was monitored by fluorescence at 341 nm (open circles) and far-UV (filled circles). Unfolding of N-LIS1 (1–79) Trp mutants was monitored only by far-UV. (b) Normalized GdmCl-induced unfolding transitions of N-LIS1 (1–79) WT (circles) and N-LIS1 (1–79) mutant Thr56Ala, Ser57Ala (triangles) at 0.5 mM concentration monitored by fluorescence at 341 nm (open symbols) and far-UV at 222 nm (filled symbols). All curves were fitted to a two-state dimer denaturation equation, as described in Materials and Methods.

The Phe31Ser patient mutation There are five missense mutations in the Lis1 gene that have been identified in lissencephaly patients.24–26 Four are found in the b-propeller domain and one (Phe31Ser) in the N-LIS1 fragment. Phe31, together with Ile15, Tyr18, Leu19 and Tyr24, provides a hydrophobic surface that is responsible for a tight dimerization of LisH motif. To rationalize the impact of this pathogenic mutation on N-LIS1 structure and stability, we studied stability and spectral properties of Phe31Ser mutant. The substitution strongly affects both the circular dichroism and fluorescence emission spectra, suggesting a severe distortion of protein conformation. Compared with the wild-type, the fluorescence emission spectrum decreased by 34% and its maximum is red-shifted by 2 nm to 343 nm (Figure 8(a)). The CD spectrum is also much less pronounced, with the ellipticity at 222 nm less negative by 28% (Figure 8(b)). This suggests that the mutant is only partially folded at room temperature. The unfolding transition begins at very low denaturant concentration and, therefore, H2 O value of the we can only estimate that the DGU mutant is very low (Figure 8(c)), consistent with a predominantly monomeric, denatured structure. This result confirms the crucial role played by the LisH motif in defining N-LIS1 dimer stability and reveals the molecular mechanism of the Phe31Ser mutant’s pathogenesis.

Discussion Studies of the mice heterozygous for a shorter version of LIS1 (missing residues 1–63), showed that this fragment, solely responsible for the

dimerization of the protein, is vital for the biological function.5 While the crystal structure of the N-LIS1 fragment revealed the molecular basis of dimerization and the role of the LisH motif,6 the crystal structure of the full-length LIS1 in complex with a2PAF-AH suggested that dynamic equilibrium between two alternative conformations, i.e. open and closed, made possible by a labile coiled-coil segment, is also of vital importance.7 Here, we document NMR and CD spectroscopic experiments that provide evidence that N-LIS1 indeed exists in an open conformation as predicted before, and that it is stable at very low salt concentrations. Interestingly, under physiological conditions the open and closed conformations exist in equilibrium with possible functional implications. LIS1 is involved in the dynein/dynactin complex,27–31 required for the movement of the dynein molecular motor along microtubules, and the flexibility of the LIS1 dimer may be necessary to accommodate the dynamic properties of the complex. The evolutionary conservation of the amino acid sequence in the N-LIS1 fragment underscores the functional and structural complexity (Figure 1(b)). The integrity of the dimer is ensured by the tight association of the LisH motifs, assisted by Trp55. The latter amino acid residue is also vital for the mutual disposition of the N termini of the parallel coiled-coil segment (558 angle) that defines the open conformation. Finally Thr56 and Ser57 provide the flexible switch in the coiled-coil that allows for the formation of the closed (supercoiled) conformations made possible by the three canonical coiled-coil heptads. In our experiments, the conformational exchange between the two states is triggered by a high concentration of ammonium sulfate, but we cannot exclude the possibility that other factors, such as LIS1 binding proteins, may exert a similar effect.

628

Dimerization of LIS1

LisH motif and colored by hydrophobic group property (Figure 9(b)). A highly non-polar surface groove composed of conserved residues 6, 9, 10 (LisH motif numbering) is clearly seen (Figure 9(c)). The presence of LisH motif in a multi-domain protein might depend on additional structured elements that would bury this large non-polar patch. In case of LIS-1 this groove is almost entirely covered by the two Trp55 residues leaving only a small internal cavity (Figure 9(a)). According to our database search, the majority of LisH motifs are indeed accompanied on their C-terminal side by short putative helical regions, coiled-coils or CTLH motifs, which might fulfill the same role as the coiled-coil from LIS1. Thus, the LisH-containing proteins may be as complex in their detailed architecture as LIS1.

Materials and Methods Protein expression and peptide synthesis The N-LIS1 (1–48) and N-LIS1 (1–57) constructs were prepared by substitution of Gly49 and Val58, respectively, with a stop codon in the mouse N-LIS1 fragment (residues 1–86), subcloned into the NcoI and XhoI sites of the pGSTUni1 expression vector.6 The mutation was confirmed by sequencing and the protein was expressed in Escherichia coli BL21(DE3)RIL strain and purified by combination of glutathione affinity chromatography and gel filtration Superdex 75 (Amersham Biosciences) after tag removal with rTEV protease. The purified protein was additionally examined by mass spectrometry. The coiledcoil region (amino acid sequence 58–79) and its dimeric version extended on the N terminus by Cys-Gly-Gly sequence were purchased from Pepscan Systems B.V. (The Netherlands). The peptide contained one substitution, F79Y, to help concentration measurements. Circular dichroism

Figure 8. Characterization of the Phe31Ser pathogenic mutant. (a) Fluorescence emission spectra of N-LIS1 (1–79) wild-type (continuous line) and its pathogenic mutant (broken line) in 20 mM sodium phosphate containing 150 mM NaCl (pH 7.4). (b) Far-UV CD spectra of N-LIS1 (1–79) wild-type (open diamonds) and its pathogenic mutant (filled circles) in 20 mM sodium phosphate containing 150 mM NaCl (pH 7.4). (c) Denaturation curves of N-LIS1 wild-type (open diamonds) and its pathogenic, Phe31Ser mutant (filled circles). Since the mutant is partially unfolded in the absence of denaturant, only raw ellipticity data are shown.

Based on the alignment of the homologous LisH motifs (as found in the non-redundant NCBI database), as well as structural information inferred from the N-LIS1 study, we propose a consensus sequence that is visualized on the surface of the

CD spectra were collected with a J-175 spectropolarimeter (Jasco) using a 2 nm band width and response time of 4 s. All the experiments were carried out at 20 8C using cuvettes of various path-lengths (0.1 cm and 1 cm). Ellipticity was reported as mean residue ellipticity [q] in deg cm2 dmolK1 using the following equation: [q]Z (qobs!MRW)/(10lc), where q is the observed ellipticity in degrees, MRW is the mean residue weight, l is the optical path length of cell in cm, and c is the peptide concentration in mg/ml. The helical content (%) was calculated from the ratio of the observed [q]222 value divided by the predicted molar ellipticity!100%. The predicted molar ellipticity was calculated from the equation [q]222Z40!103!(1K4.6/n) for the chain length dependence of an a-helix,40 where n is the number of residues in the peptide. Difference CD spectra were obtained after computer subtraction of direct recorded spectra and expressed in molar ellipticity values. Chemical denaturation Equilibrium GdmCl denaturation were monitored by steady-state fluorescence and circular dichroism signal at

629

Dimerization of LIS1

Figure 9. Surface representations of LisH motif from (a) N-LIS1 and (b) consensus LisH sequence mapped onto the NLIS1 structure with the non-polar parts of the surface colored red. In the N-LIS1 structure (PDB, 1UUJ) the conserved hydrophobic cavity is covered by two symmetric Trp55 residues located next to the coiled-coil as shown in the upper section of (a) where the coiled coil structure obscuring the view was removed for clarity. The conserved residues building the hydrophobic cavity are marked on the “sequence logo” on (b). Three PSI-BLAST36 cycles were used with an E-value inclusion threshold of 2!10K3 to detect homologous sequences in the non-redundant (nr) database (ftp://ftp.ncbi.nlm. nih.gov/blast/db). Obtained sequences were aligned in ClustaW (http://www.ebi.ac.uk/clustalw)37 and converted to the sequence logo (http://weblogo.berkeley.edu).38 Figures were prepared in spdbv 3.739 and rendered in POV-Ray 3.6 (http://www.povray.org).

222 nm. Stock solutions of GdmCl were prepared in 20 mM sodium phosphate (pH 7.4), 150 mM NaCl and mixed to final concentrations indicated and measured with a refractometer (Carl Zeiss). Protein samples were equilibrated at increasing GdmCl concentrations ranging from 0 M to 6 M for 12 h at 20 8C. Fluorescence emission spectra were carried out in 1 cm cuvettes using a FP-750 spectrofluorimeter equipped with an ETC-272 Peltier temperature controller (Jasco). The samples were excited at 280 nm and emission was measured at 341 nm. The fluorescence signal from each sample was averaged from five measurements (5 nm excitation and emission band widths with the response time 8 s). CD denaturation followed the signal at 222 nm.

The profiles of denaturation of N-LIS1 (1–79), N-LIS1 (1–57), LisH (1–48) together with N-LIS1 (1–79) Trp55Phe, Trp55Ala and Thr56Ala/Ser57Ala mutants determined by CD and fluorescence were normalized and analyzed according to the two-state dimer denaturation model: N242D, with N2 describing native dimer and D, denatured monomer. The equilibrium constant of unfolding (KU) was calculated according to the equation which describes the dependence of KU on the fraction of denatured protein and on the total protein concentrations (Pt):32 KU Z

½D2 2Pt fU2 Z ½N2  ð1KfU Þ

630

Dimerization of LIS1

The standard denaturation parameters, [GdmCl]1/2, m, H2 O and DGU were determined by fitting individual data sets to an equation given by Mateau33 with the program GraFit (Erithacus Software Ltd).

8.

NMR spectroscopy

9.

NMR spectra were recorded using Varian Inova 500 spectrometer at 25 8C. A set of 1H–15N HSQC experiments was measured for 0.5 mM to 0.8 mM solutions of 15Nlabeled N-LIS1. In order to study the effect of salt we prepared six samples: (1) 50 mM Tris (pH 7.5); (2) 50 mM Tris (pH 7.5), 50 mM NaCl; (3) 50 mM Tris (pH 7.5), 150 mM NaCl; (4) 50 mM Tris (pH 7.5), 500 mM NaCl; (5) 50 mM Tris (pH 7.5), 150 mM NaCl and 0.75 M Na2SO4; (6) 50 mM Tris (pH 7.5), 150 mM NaCl and 1.5 M Na2SO4. To verify the position of the Trp indole proton we recorded a 2D 15N-edited NOESY experiment. All spectra have been processed and analyzed using NMRpipe34 and Sparky programs (SPARKY 3; T. D. Goddard and D. G. Kneller)†.

10.

11.

12. 13.

Acknowledgements

14.

Supported by NIH grant NS36267 (to Z.S.D.). This collaborative research is also supported by the NATO-Link grant (to Z.S.D. and J.O.). J.O. thanks the Howard Hughes Medical Institute for generous support. The TEV protease expression plasmid was kindly provided by J. A. Doudna.

16.

References

17.

1. Reiner, O., Carrozzo, R., Shen, Y., Whenert, M., Faustinella, F., Dobyns, W. B. et al. (1993). Isolation of a Miller-Dieker lissencephaly gene containing G protein ß-subunit-like repeats. Nature, 364, 717–721. 2. Emes, R. D. & Ponting, C. P. (2001). A new sequence motif linking lissencephaly, Treacher Collins and oralfacial-digital type 1 syndromes, microtubule dynamics and cell migration. Hum. Mol. Genet. 10, 2813–2820. 3. Neer, E. J., Schmidt, C. J., Nambudripad, R. & Smith, T. F. (1994). The ancient regulatory-protein family of WD-repeat proteins. Nature, 371, 297–300. 4. Reiner, O. (2000). LIS1: Let’s Interact Sometimes.(part 1). Neuron, 28, 633–636. 5. Cahana, A., Escamez, T., Nowakowski, R. S., Hayes, N. L., Giacobini, M., von Holst, A. et al. (2001). Targeted mutagenesis of Lis1 disrupts cortical development and LIS1 homodimerization. Proc. Natl Acad. Sci. USA, 98, 6429–6434. 6. Kim, M. H., Cooper, D. R., Oleksy, A., Devedjiev, Y., Derewenda, U., Reiner, O. et al. (2004). The structure of the N-terminal domain of the product of the lissencephaly gene Lis1 and its functional implications. Structure, 12, 987–998. 7. Tarricone, C., Perrina, F., Monzani, S., Massimiliano, L., Kim, M. H., Derewenda, Z. S. et al.

15.

18.

19. 20.

21.

22.

23.

24. † http://www.cgl.ucsf.edu/home/sparky/

(2004). Coupling PAF signaling to dynein regulation: structure of LIS1 in complex with PAF-acetylhydrolase. Neuron, 44, 809–821. Gruber, M. & Lupas, A. N. (2003). Historical review: another 50th anniversary—new periodicities in coiled coils. Trends. Biochem. Sci. 28, 679–685. Bassi, M. T., Ramesar, R. S., Caciotti, B., Winship, I. M., De Grandi, A., Riboni, M. et al. (1999). X-linked lateonset sensorineural deafness caused by a deletion involving OA1 and a novel gene containing WD-40 repeats. Am. J. Hum. Genet. 64, 1604–1616. Ferrante, M. I., Giorgio, G., Feather, S. A., Bulfone, A., Wright, V., Ghiani, M. et al. (2001). Identification of the gene for oral-facial-digital type I syndrome. Am. J. Hum. Genet. 68, 569–576. Shoo, B. A., McPherson, E. & Jabs, E. W. (2004). Mosaicism of a TCOF1 mutation in an individual clinically unaffected with Treacher Collins syndrome. Am. J. Med. Genet. ser. A, 126, 84–88. Boulton, S. J., Brook, A., Staehling-Hampton, K., Heitzler, P. & Dyson, N. (2000). A role for Ebi in neuronal cell cycle control. EMBO J. 19, 5376–5386. Dong, X., Tsuda, L., Zavitz, K. H., Lin, M., Li, S., Carthew, R. W. & Zipursky, S. L. (1999). Ebi regulates epidermal growth factor receptor signaling pathways in Drosophila. Genes Dev. 13, 954–965. Neet, K. E. & Timm, D. E. (1994). Conformational stability of dimeric proteins: quantitative studies by equilibrium denaturation. Protein Sci. 3, 2167–2174. Lau, S. Y., Taneja, A. K. & Hodges, R. S. (1984). Synthesis of a model protein of defined secondary and quaternary structure. Effect of chain length on the stabilization and formation of two-stranded alphahelical coiled-coils. J. Biol. Chem. 259, 13253–13261. Cooper, T. M. & Woody, R. W. (1990). The effect of conformation on the CD of interacting helices: a theoretical study of tropomyosin. Biopolymers, 30, 657–676. Zhou, N. E., Kay, C. M. & Hodges, R. S. (1992). Synthetic model proteins. Positional effects of interchain hydrophobic interactions on stability of twostranded alpha-helical coiled-coils. J. Biol. Chem. 267, 2664–2670. Krylov, D., Mikhailenko, I. & Vinson, C. (1994). A thermodynamic scale for leucine zipper stability and dimerization specificity: e and g interhelical interactions. EMBO J. 13, 2849–2861. Jasanoff, A. & Fersht, A. R. (1994). Quantitative determination of helical propensities from trifluoroethanol titration curves. Biochemistry, 33, 2129–2135. Hicks, M. R., Holberton, D. V., Kowalczyk, C. & Woolfson, D. N. (1997). Coiled-coil assembly by peptides with non-heptad sequence motifs. Fold Des. 2, 149–158. Choy, N., Raussens, V. & Narayanaswami, V. (2003). Inter-molecular coiled-coil formation in human apolipoprotein E C-terminal domain. J. Mol. Biol. 334, 527–539. Dubin, M. J., Stokes, P. H., Sum, E. Y., Williams, R. S., Valova, V. A., Robinson, P. J. et al. (2004). Dimerization of CtIP, a BRCA1- and CtBP-interacting protein, is mediated by an N-terminal coiled-coil motif. J. Biol. Chem. 279, 26932–26938. Fersht, A. R. (1999). Energetics of formation of structure. In Structure and Mechanism in Protein Science: a Guide to Enzyme Catalysis and Protein Folding (Fersht, A. R., ed.), pp. 522–529, Freeman, New York. Lo Nigro, C., Chong, C. S., Smith, A. C., Dobyns, W. B., Carrozzo, R. & Ledbetter, D. H.

631

Dimerization of LIS1

25.

26.

27.

28. 29.

30. 31.

32.

(1997). Point mutations and an intragenic deletion in LIS1, the lissencephaly causative gene in isolated lissencephaly sequence and Miller-Dieker syndrome. Hum. Mol. Genet. 6, 157–164. Pilz, D. T., Matsumoto, N., Minnerath, S. R., Mills, P., Gleeson, J. G., Allen, D. H. et al. (1998). LIS1 and XLIS (DCX) cause most classical lissencephaly, but different patterns of malformation. Hum. Mol. Genet. 7, 2029–2037. Cardoso, C., Leventer, R. J., Matsumoto, N., Kuc, J. A., Ramocki, M. B., Mewborn, S. K. et al. (2000). The location and type of mutation predict malformation severity in isolated lissencephaly caused by abnormalities within the LIS1 gene. Hum. Mol. Genet. 9, 3019–3028. Payne, C., St John, J. C., Ramalho-Santos, J. & Schatten, G. (2003). LIS1 association with dynactin is required for nuclear motility and genomic union in the fertilized mammalian oocyte. Cell. Motil. Cytoskel. 56, 245–251. Morris, R. (2000). A rough guide to a smooth brain. Nature Cell. Biol. 2, E201–E202. Faulkner, N. E., Dujardin, D. L., Tai, C. Y., Vaughan, K. T., O’Connell, C. B., Wang, Y. & Vallee, R. B. (2000). A role for the lissencephaly gene LIS1 in mitosis and cytoplasmic dynein function. Nature Cell. Biol. 2, 784–791. Liu, Z., Steward, R. & Luo, L. (2000). Drosophila Lis1 is required for neuroblast proliferation, dendritic elaboration and axonal transport. Nature Cell. Biol. 2, 776–783. Smith, D. S., Niethammer, M., Ayala, R., Zhou, Y., Gambello, M. J., Wynshaw-Boris, A. & Tsai, L. H. (2000). Regulation of cytoplasmic dynein behaviour and microtubule organization by mammalian Lis1. Nature Cell. Biol. 2, 767–775. Glaser, F., Pupko, T., Paz, I., Bell, R. E., BechorShental, D., Martz, E. & Ben-Tal, N. (2003). ConSurf:

33.

34.

35.

36.

37.

38. 39. 40.

identification of functional regions in proteins by surface-mapping of phylogenetic information. Bioinformatics, 19, 163–164. Mateu, M. G. (2002). Conformational stability of dimeric and monomeric forms of the C-terminal domain of human immunodeficiency virus-1 capsid protein. J. Mol. Biol. 318, 519–531. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J. & Bax, A. (1995). NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR, 6, 277–293. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C. & Ferrin, T. (2004). UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612. Schaffer, A. A., Aravind, L., Madden, T. L., Shavirin, S., Spouge, J. L., Wolf, Y. I. et al. (2001). Improving the accuracy of PSI-BLAST protein database searches with composition-based statistics and other refinements. Nucl. Acids Res. 29, 2994–3005. Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T. J., Higgins, D. G. & Thompson, J. D. (2003). Multiple sequence alignment with the Clustal series of programs. Nucl. Acids Res. 31, 3497–3500. Crooks, G. E., Hon, G., Chandonia, J. M. & Brenner, S. E. (2004). WebLogo: a sequence logo generator. Genome Res. 14, 1188–1190. Guex, N. & Peitsch, M. C. (1997). SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis, 18, 2714–2723. Gans, P. J., Lyu, P. C., Manning, M. C., Woody, R. W. & Kallenbach, N. R. (1991). The helix-coil transition in heterogeneous peptides with specific side chain interactions: theory and comparison with CD spectral data. Biopolymers, 31, 1605–1614.

Edited by P. Wright (Received 7 September 2005; received in revised form 2 December 2005; accepted 3 January 2006) Available online 19 January 2006