Lissencephaly is a developmental brain disorder characterized by a smooth cerebral surface, thickened cortex and misplaced neurons. Classical lissencephaly is caused by mutations in LIS1, which encodes a WD-repeat protein involved in cytoplasmic dynein regulation, mitosis and nuclear migration. Several proteins required for nuclear migration in Aspergillus bind directly to Lis1, including NudC. Mammalian NudC is highly expressed in ciliated epithelia, and localizes to motile cilia in various tissues. Moreover, a NudC ortholog is upregulated upon deflagellation in Chlamydomonas. We found that mammalian Lis1 localizes to motile cilia in trachea and oviduct, but is absent from non-motile primary cilia. Furthermore, we cloned a gene encoding a Lis1-like protein (CrLis1) from Chlamydomonas. CrLis1 is a ∼37 kDa protein that contains seven WD-repeat domains, similar to Lis1 proteins from other organisms. Immunoblotting using an anti-CrLis1 antibody revealed that this protein is present in the flagellum and is depleted from flagella of mutants with defective outer dynein arm assembly, including one strain that lacks only the α heavy chain/light chain 5 thioredoxin complex. Biochemical experiments confirmed that CrLis1 associates with outer dynein arm components and revealed that CrLis1 binds directly to rat NudC. Our results suggest that Lis1 and NudC are present in cilia and flagella and may regulate outer dynein arm activity.
Introduction
Lissencephaly is a developmental brain disorder characterized by a smooth cerebral surface, thickened cortex and misplaced neurons (Wynshaw-Boris and Gambello, 2001). Classical or type I lissencephaly is caused by mutations in LIS1, which encodes a WD-repeat protein initially identified as one of the three subunits of the brain platelet-activating factor acetyl-hydrolase (PAFAH)-1b complex (Hattori et al., 1994; Reiner et al., 1993). However, it is now clear that this complex plays no significant role in brain development (Koizumi et al., 2003; Yan et al., 2003). The first clue towards elucidating the molecular pathway by which Lis1 functions in brain development was provided by genetic studies in Aspergillus nidulans. Mutations of the Aspergillus nuclear distribution (Nud) genes compromise the migration of nuclei within the growing hypha. NudF, an ortholog of mammalian Lis1 is involved in this pathway (Xiang et al., 1995), as are cytoplasmic dynein heavy chain (NudA) and the NudC protein (Morris, 2000). Furthermore, mutations in NudA are able to suppress the NudF nuclear migration phenotype (Willins et al., 1997), suggesting that Lis1/NudF functions in the cytoplasmic dynein pathway. In support of this notion, vertebrate Lis1 coimmunoprecipitates with cytoplasmic dynein and dynactin components in brain extracts, and colocalizes with these protein complexes in mouse brain cortex and various cultured cell types (Faulkner et al., 2000; Smith et al., 2000). The interaction between Lis1 and the cytoplasmic dynein heavy chain (HC) is direct and occurs at multiple sites on the HC molecule, including the motor domain and cargo binding regions (Sasaki et al., 2000; Tai et al., 2002), and may directly stimulate the ATPase activity of the motor (Mesngon et al., 2006). In addition to cytoplasmic dynein and dynactin, Lis1 binds to a variety of other proteins, including NudC (see below), which in turn interact with cytoplasmic dynein and dynactin (reviewed by Vallee and Tsai, 2006). Finally, Lis1 has been found to associate with microtubules and affect microtubule dynamics (Sapir et al., 1997), but this finding is somewhat controversial (Faulkner et al., 2000).
Overexpression of Lis1 in cultured vertebrate cells leads to alterations in microtubule organization, mitotic progression, spindle orientation and chromosome attachment (Faulkner et al., 2000; Smith et al., 2000). Furthermore, microinjection of anti-Lis1 antibody into dividing cultured vertebrate cells results in delayed mitotic progression, suggesting a role for Lis1 in cell division (Faulkner et al., 2000). This is supported by the observation that homozygous Lis1 knockout mice die before birth (Hirotsune et al., 1998), and that individuals suffering from classical lissencephaly are defective in only one LIS1 allele (Wynshaw-Boris and Gambello, 2001). Consequently, functional studies of Lis1 during brain development in animals have been performed on heterozygotes and/or hypomorphs that contain reduced levels of Lis1 protein (Gambello et al., 2003; Hirotsune et al., 1998), or by in utero eletroporation of LIS1 small interference RNA and dominant negative cDNA constructs (Tsai et al., 2005). These studies have indicated that the primary effect of Lis1 depletion is to disrupt division of neural progenitor cells in the developing brain, as well as their migration (Hatten, 2005; Tsai et al., 2005). Consistent with an important function for Lis1 in brain development, this protein is more highly expressed in neurons than in non-neuronal cells (Smith et al., 2000). Nevertheless, there is still Lis1 protein present in cells from non-neuronal tissue such as liver (Smith et al., 2000), indicating that Lis1 may function in non-neuronal cells as well.
In mammals, motile cilia are involved in fluid transport at the embryonic node, and within many organs (e.g. trachea, oviduct, and brain ventricles); in addition, flagella power sperm motility (Ibanez-Tallon et al., 2003). The dynein-based movement of these microtubular organelles is known to be regulated by several distinct signaling mechanisms including cyclic adenosine monophosphate (cAMP), Ca2+, phosphorylation and redox poise (for a review, see Sakato and King, 2004). Furthermore, most cells contain a single non-motile primary cilium that lacks dynein arms and is now considered to function as a sensory organelle (for reviews, see Pazour and Witman, 2003; Christensen et al., 2007). Interestingly, the mammalian NudC protein, which binds directly to Lis1 (Morris et al., 1998), is highly expressed in ciliated epithelia, and localizes to motile cilia in various tissues (Gocke et al., 2000). In addition, the genome of Chlamydomonas reinhardtii contains a NudC ortholog, which is upregulated upon deflagellation (Stolc et al., 2005), suggesting that Chlamydomonas NudC is a flagellar protein (Lefebvre et al., 1980). However, despite the conservation of ciliary localization of NudC in algae and vertebrates, the physiological relevance of these observations is unclear.
We show here that Lis1 localizes to motile cilia in murine trachea and oviduct, but is absent from non-motile primary cilia in the ovary and in cultured growth-arrested fibroblasts. Furthermore, we cloned a gene encoding a Lis1-like protein (CrLis1) from Chlamydomonas and demonstrate that this protein is present in the flagellum and depleted from flagella of mutants with defective outer dynein arm (ODA) assembly. Finally, we provide evidence that CrLis1 associates with the ODA in Chlamydomonas flagellar extracts, and binds directly to rat NudC. These findings indicate that Lis1 (and NudC) are present in cilia and flagella, suggesting that they may regulate ODA activity.
Results
Mammalian Lis1 localizes to motile cilia
The findings that mammalian NudC localizes to motile cilia in various tissues (Gocke et al., 2000) and interacts directly with Lis1 (Morris et al., 1998) led us to investigate whether mammalian Lis1 is present in motile cilia and flagella. First, we performed immunofluorescence microscopy (IFM) analysis of sections from paraffin-embedded mouse trachea, oviduct and ovary using a goat polyclonal antibody (sc-2577) raised against human Lis1. Immunoblot analysis of lysates derived from the same tissues demonstrated that this antibody specifically crossreacts with murine Lis1 in these tissues and does not recognize other components (Fig. 1). The IFM analysis revealed that Lis1 is present in the motile cilia of trachea and oviduct, but is absent from (non-motile) primary cilia of follicular granulosa cells in the ovary (Fig. 2A,B). Further, Lis1 was absent from primary cilia of growth-arrested NIH3T3 fibroblasts, although punctate localization of Lis1 along the cytoplasmic microtubule network could be observed in these cells (Fig. 2C), consistent with a previously published report (Smith et al., 2000). We also performed IFM analysis of isolated ciliated cells from murine tracheae using a different polyclonal Lis1-specific antibody generated in rabbits (ab2607; Fig. 1). This analysis confirmed that Lis1 localizes to motile cilia (Fig. 3) and was present in multiple punctate structures along the ciliary length. Furthermore, Lis1 was also located in punctate structures throughout the cytoplasm and a strong accumulation of signal was consistently observed at the basal body region from which the cilia assemble (Fig. 3). No ciliary signal was observed if the primary antibody was omitted or if an irrelevant rabbit antibody (against the p24 component of dynactin) was used (Fig. 3). These results indicate that Lis1 is present in motile cilia and is associated with a compartment or substructure of these organelles that is absent from non-motile cilia.
Identification of a Chlamydomonas Lis1 ortholog
To study Lis1 in motile cilia, we used the biflagellate green alga Chlamydomonas as a model organism. By BLAST search analysis (Altschul et al., 1990) of the Chlamydomonas genome sequence (http://genome.jgi-psf.org/Chlre3/Chlre3.home.html) and EST databases using the human Lis1 polypeptide sequence as query we identified a gene (GenBank accession number DQ647383), hereafter referred to as CrLIS1, located on scaffold 20 (JGI version 3, nucleotides 1535128-1537336), which was predicted to encode a 347-residue (37.2 kDa) polypeptide with 45% sequence similarity and 30% identity to human Lis1 (Fig. 4). Southern blot analysis of genomic DNA isolated from Chlamydomonas cells demonstrated that there is a single copy of CrLIS1 in the Chlamydomonas genome (Fig. 5A). Northern blot analysis further revealed that this gene is expressed and is upregulated upon deflagellation in Chlamydomonas (Fig. 5B), suggesting that CrLis1 is a flagellar protein (Lefebvre et al., 1980). Analysis of the CrLis1 polypeptide sequence using the SMART algorithm (http://smart.embl-heidelberg.de/) indicated that CrLis1 contains seven WD repeats, similar to Lis1 proteins from other organisms (Reiner and Coquelle, 2005), but lacks the characteristic 33-residue LisH domain at its N-terminus (Fig. 5C). However, ClustalW alignment (Fig. 4) and phylogenetic analysis (Fig. 5D) of multiple WD-repeat proteins indicated that CrLis1 is evolutionarily related to mammalian Lis1.
CrLis1 is present in flagella
In order to characterize the CrLIS1 gene product, we cloned the full-length CrLIS1 open reading frame, expressed it as a fusion protein in Escherichia coli, and generated an anti-CrLis1 polyclonal antibody (see Materials and Methods). Immunoblot analysis of isolated flagella using affinity-purified CrLis1 antibody revealed the presence of a ∼37 kDa band that co-migrated with recombinant CrLis1 (Fig. 6A). The antibody also crossreacted with a slower migrating band (indicated with an asterisk in Fig. 6A); this band is presumably a component of the axonemal central pair complex as it is missing from flagella of the central pair mutant pf18 (data not shown). We conclude that the ∼37 kDa band detected by the CrLis1 antibody (Fig. 6A) corresponds to CrLis1 and that CrLis1 therefore is present in the flagellum. This conclusion is consistent with the northern blot analysis shown in Fig. 5B.
To determine the localization of CrLis1 within the flagellar compartment, isolated flagella were demembranated with 1% IGEPAL-630 and the resulting axonemes were further extracted with 0.6 M NaCl to remove the dynein arms (King et al., 1986). Immunoblot analysis of these flagellar fractions showed that using this fractionation procedure, CrLis1 is mostly present in the detergent-insoluble axonemal fraction although a small amount of CrLis1 was observed in the detergent-soluble membrane plus matrix fraction as well (Fig. 6B). We note, however, that under different fractionation conditions (e.g. freezing and thawing flagella, adding ATP that can extract axonemal dyneins and/or using alternative detergent types and concentrations) a significant proportion of CrLis1 was present in the soluble membrane plus matrix fraction (Fig. 7 and Fig. 8B).
CrLis1 is absent from flagella of mutants with defects in ODA assembly
To assess the interactions of CrLis1 within the flagellum, we first screened a variety of mutants with known defects in flagellar assembly and function with respect to the flagellar level of CrLis1. To this end, we isolated flagella from mutant cells and subjected them to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot analysis using the anti-CrLis1 antibody. Interestingly, the results of this analysis revealed that CrLis1 is completely absent from flagella of mutants with defects in ODA assembly (oda1, oda3, oda6), but is present in flagella of mutants lacking other axonemal structures, such as inner dynein arm I1/f (ida1), a subset of inner dynein arms I2/3 (ida4), radial spokes (pf14), and the central pair microtubule complex (pf18) (Fig. 6C, and data not shown).
The Chlamydomonas ODA contains three distinct HC motor units (α, β and γ) in addition to WD-repeat intermediate chains and multiple LC components, several of which bind directly to the HCs. To further investigate the relationship between CrLis1 and the ODAs, we next tested for the presence of CrLis1 in mutant flagella lacking specific motor subunits of the ODA complex: the entire outer arm (oda2), the outer arm α chain and light chain (LC) 5 thioredoxin (oda11), the motor domain of the outer arm β HC (oda4-s7), and the oda4-s7 oda11 double mutant that contains only an intact γ HC (Sakakibara et al., 1991; Sakakibara et al., 1993); for details of ODA structure and organization, see DiBella et al. (DiBella et al., 2004). Immunoblot analysis of these mutant flagella showed that CrLis1 is almost completely missing from oda2 and oda11 flagella, but is present in the oda4-s7 mutant (Fig. 6C). These observations suggest that CrLis1 is associated specifically with the ODA α HC or possibly the LC5 thioredoxin, and that this association is required for targeting to or retention of CrLis1 in the flagellum.
CrLis1 associates with Chlamydomonas ODA and recombinant rat NudC
Our observation that CrLis1 is depleted from flagella of mutants with defects in ODA assembly (Fig. 6C) suggested that CrLis1 might be physically associated with the ODA complex, specifically the ODA α HC or possibly the LC5 thioredoxin, and that this association is essential for CrLis1 flagellar localization. To test whether CrLis1 associates with ODA components, we performed sucrose density gradient centrifugation analysis of 0.6 M NaCl axonemal extracts as well as flagellar membrane plus matrix extract prepared from fresh flagella by extraction with 0.5% Tergitol plus 10 mM ATP (see Materials and Methods), which contains both CrLis1 and ODA components. Individual fractions were probed for CrLis1 as well as IC2 and LC1, which sediment with the ODA αβ and γ HC complexes, respectively (Pfister et al., 1982). This analysis showed that whereas IC2 (a component of the αβ HC subparticle) sedimented at ∼20 S and LC1 (part of the γ HC subunit) sedimented at 12 S as expected, CrLis1 was detected only near the top of the gradients with a peak at ∼3 S (Fig. 7). This suggests that CrLis1 is not an integral part of the ODA αβ or γ HC complexes and/or that CrLis1 dissociates from these complexes during extraction or sucrose density gradient centrifugation.
Although our sucrose density gradient analysis indicated that CrLis1 is not an integral ODA component, CrLis1 may associate transiently with the ODA αβ HC complex given the results of our ODA mutant analysis (Fig. 6C). To address this issue, we performed a glutathione-S-transferase (GST) pull-down assay. To avoid potential interference of high salt concentration with this assay, we chose to use flagellar membrane plus matrix extracted by freezing and thawing flagella in the presence of 0.05% Nonidet NP-40 and 10 mM ATP rather than using 0.6 M NaCl axonemal extract. The membrane plus matrix detergent and ATP extract was mixed with glutathione beads coated with purified recombinant GST-CrLis1 or GST alone, and bound proteins were collected by centrifugation and analyzed by SDS-PAGE and immunoblotting. The results showed that IC2, which is part of the ODA αβ HC complex (Pfister et al., 1982), cosedimented with GST-CrLis1, but not with GST alone (Fig. 8A,C). Furthermore, α-tubulin did not cosediment with GST-CrLis1 suggesting that association of the outer arm and GST-CrLis1 is specific (Fig. 8A,C).
Given that mammalian Lis1 binds directly to the nuclear movement protein NudC (Morris et al., 1998), and that Chlamydomonas contains a NudC ortholog (C_590115 in the Chlamydomonas genome database) that is upregulated by deflagellation (Stolc et al., 2005), we tested whether rat NudC associates with CrLis1 and the ODA αβ HC complex. For this purpose, flagellar membrane plus matrix detergent and ATP extract (see above) was mixed with glutathione beads coated with purified recombinant GST-rat NudC (GST-rNudC) (Morris et al., 1998) or GST alone, and bound proteins were recovered and analyzed as described above. Interestingly, this analysis showed that CrLis1, IC2 and the α HC-associated LC5 thioredoxin cosedimented with GST-rNudC but not with GST alone (Fig. 8B,C, and data not shown). By contrast, α-tubulin did not cosediment with GST-rNudC, indicating that the association of IC2, LC5 and CrLis1 with GST-rNudC is specific (Fig. 8B,C). Therefore, we conclude that CrLis1 and rNudC interact, at least indirectly, with each other and with the ODA αβ HC complex in Chlamydomonas flagella.
CrLis1 associates directly with rat NudC
Although CrLis1 contains seven WD-repeat domains similar to mammalian Lis1, CrLis1 lacks the N-terminal LisH domain characteristic of other Lis1 proteins (Fig. 5C). Therefore, it was important to determine whether CrLis1 is a functional ortholog of mammalian Lis1 or an unrelated protein with a similar domain structure. Since Lis1 from rat was shown to bind directly to rNudC (Morris et al., 1998), we tested whether CrLis1 similarly could bind directly to rNudC. First, we performed a GST pull-down experiment in which a rat brain extract was mixed with GST-CrLis1 beads and bound proteins were analyzed by SDS-PAGE and immunoblotting using an rNudC-specific antibody (Fig. 8D,E). The results showed that native rNudC specifically interacts, at least indirectly, with GST-CrLis1, but not with GST alone (Fig. 8D,E). Next, we performed a similar analysis in which purified recombinant maltose-binding protein (MBP; control) or MBP fused in-frame to CrLis1 (MBP-CrLis1) was mixed with beads coated with GST-rNudC, and bound proteins were recovered by centrifugation. SDS-PAGE and immunoblot analysis of the pellet and supernatant fractions demonstrated that MBP-CrLis1 cosedimented with GST-rNudC, whereas MBP alone did not cosediment (Fig. 8F). Since MBP-CrLis1 does not associate with GST alone (Fig. 8G), we conclude that CrLis1 interacts directly with rNudC and hence is a functional ortholog of mammalian Lis1.
Discussion
In this report, we have demonstrated that the lissencephaly protein Lis1 is specifically present in motile mammalian cilia and furthermore, we have identified a Lis1 homologue in Chlamydomonas that depends on the ODA for targeting to and/or retention in the flagellum, and which can bind directly to mammalian NudC.
The Chlamydomonas ODA has a molecular mass of ∼2 MDa and the isolated particle contains three different HCs (α, β and γ), two distinct ICs, at least ten different LCs and a trimeric docking complex (DiBella and King, 2001; DiBella et al., 2004). Our analysis of various Chlamydomonas mutants suggested that CrLis1 is associated specifically with the α HC or the LC5 thioredoxin of the ODA complex. Mammalian Lis1 binds directly to the AAA1 domain and N-terminal stem of cytoplasmic dynein HC (Sasaki et al., 2000; Tai et al., 2002), and biochemical experiments indicated that Lis1 directly stimulates ATP hydrolysis by the motor (Mesngon et al., 2006). Therefore, we predict that CrLis1 associates with the equivalent regions of the ODA α HC, but whether this association affects the ATPase activity of the ODA α HC is uncertain, because CrLis1 lacks the N-terminal LisH domain (Fig. 5C) that is required for Lis1 dimerization and stimulation of cytoplasmic dynein ATPase activity (Mesngon et al., 2006). However, as the Chlamydomonas ODA contains three different HC motors whereas cytoplasmic dynein is a HC homodimer, it is possible that this Lis1 dimerization activity is unnecessary within the confines of the axonemal superstructure. We note that the α HC exhibits a dominant regulatory effect on the ATP-sensitive microtubule-binding properties of the ODA complex (Sakato and King, 2003), and also that the α HC is phosphorylated in vivo at multiple sites (King and Witman, 1994), indicating that Lis1 potentially could be important for regulation of ODA activity. Such regulation may involve input from a variety of signaling pathways, since Chlamydomonas axonemal dyneins are controlled by cAMP, Ca2+, redox poise and phosphorylation (Bessen et al., 1980; Howard et al., 1994; Porter and Sale, 2000; Wakabayashi and King, 2006).
Our results indicate that Lis1 (and probably NudC) are associated with the ODA complex of motile cilia and may be involved in regulating its activity. These data do not dispute previous findings demonstrating an essential role for Lis1 in cytoplasmic dynein regulation, mitosis and neuronal migration during embryonic brain development (Faulkner et al., 2000; Gambello et al., 2003; Hirotsune et al., 1998; Smith et al., 2000; Tsai et al., 2005). Rather, we propose that in addition to this function, Lis1 may be involved in regulating ODA activity and ciliary beating in the trachea, oviduct and other ciliated tissues such as the brain ependyma. Interestingly, a recent report demonstrated that beating of brain ependymal cilia is required for normal cerebrospinal fluid (CSF) flow, which in turn is necessary for formation of a concentration gradient of CSF guidance molecules and directional neuronal migration in the adult brain (Sawamoto et al., 2006). Therefore, if Lis1 is required for regulating ODA activity in ependymal cilia, it may be required not only for neuronal migration in the developing embryonic brain, but also in the adult brain.
Typically, genes required for ciliary motility have been associated with a disease known as Kartagener's Syndrome or Primary Ciliary Dyskinesia (PCD) (Afzelius, 1976; Ibanez-Tallon et al., 2003). PCD is characterized by recurrent infections of the upper and lower respiratory tract, male infertility and left-right asymmetry defects or situs inversus. In rare cases, hydrocephalus internus, eye anomalies and cystic kidney disorder are also observed (Ibanez-Tallon et al., 2003). To the best of our knowledge, none of these ciliary diseases have previously been linked to mutations in LIS1. However, Lis1 is an essential protein, homozygous Lis1 knockout mutations in the mouse are embryonic lethal, and individuals suffering from classical lissencephaly contain sporadic mutations in only one LIS1 allele resulting in Lis1 haploinsufficiency (Hirotsune et al., 1998; Wynshaw-Boris and Gambello, 2001). It is possible that the amount of Lis1 protein produced by these individuals is sufficient for appropriate Lis1 activity in the cilia, but is insufficient to maintain Lis1 cytoplasmic functions. In addition, some types of cilia, such as motile ependymal cilia, are not formed until after birth (Banizs et al., 2005) and hence potential defects in these cilia would not be observed in embryos of homozygous Lis1 mutant animals, which die between embryonic day 5.5 and 9.5 (Hirotsune et al., 1998). However, in mouse models with one functional Lis1 gene higher instances of hydrocephalus were reported (Assadi et al., 2003; Hirotsune et al., 1998); a condition that is commonly observed when normal ciliary motility is disrupted (Ibanez-Tallon et al., 2004; Sapiro et al., 2002; Taulman et al., 2001). Consequently, despite the lack of direct genetic and clinical evidence of a cilia-related function for Lis1, it remains likely that Lis1 plays a role in regulating ciliary activity in the brain and other tissues through its association with the ODAs.
Materials and Methods
Strains and culture conditions
C. reinhardtii strains CC125 MT+ (wild-type), oda1, oda2, oda3, oda6, oda11, oda4-s7, ida1, ida4, pf14 and pf18 were obtained from the Chlamydomonas Genetics Center (Duke University, Durham, NC). For large-scale flagellar isolation, cells were grown in Tris-acetate-phosphate (TAP) or R medium at 22°C and synchronized with a 14-hour light and 10-hour dark or 15-hour light and 9-hour dark cycle and bubbling with air supplemented with 5% CO2. Cells were collected by centrifugation and resuspended in 10 mM HEPES, pH 7.4, prior to deflagellation.
PCR, cloning procedures and plasmids
For cloning of CrLIS1, the full-length CrLIS1 cDNA coding region (1041 bp) was PCR-amplified from C. reinhardtii wild-type (strain CC125 MT+) cDNA by using the Advantage GC cDNA PCR Kit (BD Biosciences) and primers SAE5′ (5′-GGATCCTCAGCGGAAGTGGCGACCAC-3′) and AVQ3′ (5′-AAGCTTTTGCA CCGCCCTTTCCTTCC-3′). Oligo dT-primed template cDNA was prepared as described previously (Pedersen et al., 2003). The PCR product was ligated into pCR2.1 TOPO (Invitrogen), and the ligated product was then used to transform competent E. coli DH5α cells. Plasmids were isolated from the transformants and the inserts from three different plasmids sequenced by standard procedures. The sequences were identical to that derived from alignment of 15 different overlapping ESTs (GenBank accession numbers BG859959.1, BQ824296.1, BI531759.1, BQ820892.1, BG854313.1, BU654540.1, BQ825427.1, BQ809915.1, BU651787.1, BE227932.1, BI717926.1, BQ824140.1, BE452811.1, AW758377.1 and BI530720.1), except that nucleotide 62 was a T instead of C and nucleotide 441 was a T instead of C leading to replacement of alanine at position 21 with valine. These nucleotide changes are probably as a result of strain differences because sequencing of plasmid inserts derived from different PCR reactions gave the same result. The CrLIS1 open reading frame predicted from the C. reinhardtii genome sequence version 3 (http://genome.jgi-psf.org/Chlre3/Chlre3.home.html; scaffold 20 nucleotides 1535128-1537336) lacked the first 147 nucleotides at the 5′ end, but was otherwise identical to that derived from the ESTs. The insert from one of the CrLIS1-containing plasmids, named pLP2, was excised with BamH1 and HindIII, ligated into plasmid pQE30 (Qiagen) and transformed into E. coli XL1-Blue cells. One of the transformants (strain LP4) was used for production of 6xHis-tagged CrLis1 fusion protein (see below). The CrLIS1 coding sequence from strain CC125 MT+ was submitted to GenBank (accession number DQ647383).
For GST-CrLis1 fusion protein, a ∼1.1 kb CrLIS1-containing fragment was excised from plasmid pLP2 with BamH1 and EcoR1 and ligated into plasmid pGEX-2T (Amersham-Pharmacia) to generate an in-frame fusion of GST with the 5′ end of CrLis1. After initial propagation in E. coli XL1-Blue cells, the resulting plasmid (pLP21) was transformed into E. coli BL21 (DE3) pLysE (Novagen) to generate strain LP22. For generation of rat NudC-GST fusion protein, plasmid pGEX2T containing a ∼1.3 kb rat NUDC fragment in the EcoR1 site (generously provided by Li-Yuan Yu-Lee, Baylor College of Medicine) (Morris et al., 1998) was transformed into E. coli BL21 (DE3) pLysE (Novagen) to generate strain LP23. For production of GST, E. coli XL1-Blue harboring pGEX5X-2T (Amersham-Pharmacia) without insert (strain LP36) was used.
To generate a MBP-CrLis1 fusion protein, a ∼1.1 kb CrLIS1-containing fragment was excised from plasmid pLP4 (see above) with BamH1 and Hind111, ligated into plasmid pMalc2 (New England Biolabs) and transformed in E. coli XL1-Blue cells. The resulting strain (LP45) was used for production of MBP-CrLis1 fusion protein. For production of MBP, E. coli XL1-Blue containing pMALc2 without insert (strain LP44) was used.
Southern and northern blot analysis
The CrLIS1 cDNA was used to probe a Southern blot of Chlamydomonas genomic DNA restricted with BamHI, PstI, PvuII and SmaI, and a northern blot containing total RNA from non-deflagellated cells and from cells that had been allowed to undergo flagellar regeneration for 30 minutes.
Antibodies
For preparation of anti-CrLis1 antibody, E. coli strain LP4 was grown at 37°C in Luria Bertani broth containing 100 μg/ml ampicillin and when the culture had reached an optical density at 600 nm (OD600) of 0.5, isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to 1 mM final concentration. After two hours of incubation with IPTG, the cells were harvested by centrifugation and 6xHis-tagged CrLis1 fusion protein was purified under denaturing conditions as recommended by Qiagen (QIAexpressionist, Third Edition, protocols 9 and 14). The purified fusion protein was dialyzed against PBS and used to raise polyclonal antibodies in rabbits at Pocono Rabbit Farm & Laboratory Inc. (Canadensis, PA). The resulting antibody was affinity-purified against the recombinant protein essentially as described previously (Talian et al., 1983). Mouse monoclonal antibodies against α-tubulin, acetylated α-tubulin, and IC2/IC69 of the Chlamydomonas ODA complex (King et al., 1985) were obtained from Sigma. Rabbit polyclonal antibodies against actin and MBP were obtained from Sigma and New England Biolabs, respectively, whereas Li-Yuan Yu-Lee (Baylor College of Medicine) generously provided us with rabbit polyclonal antibody specific for rat NudC (Morris et al., 1998). Rabbit polyclonal antibodies against LC1 and LC2 of the Chlamydomonas ODA complex and the p22/24 component of rat dynactin have been described previously (Benashski et al., 1999; Patel-King et al., 1997; Pfister et al., 1998). Antibodies against mammalian Lis1 were obtained from Abcam (ab2607 rabbit polyclonal) and Santa Cruz Biotechnology (sc-2577 goat polyclonal).
Flagellar isolation and fractionation
Flagella from wild-type and mutant Chlamydomonas strains were isolated as described previously (King, 1995; Witman, 1986) and demembranated with 1% IGEPAL-630 (Sigma) in HMEK buffer (10 mM HEPES, 5 mM MgSO4, 25 mM KCl, 0.5 mM EDTA, pH 7.4). The resulting axonemes were then extracted with 0.6 M NaCl to remove the dynein arms (King et al., 1986).
Sucrose density gradient centrifugation
Sucrose density gradient centrifugation was performed on a 0.6 M NaCl axonemal extract (see above) as well as a flagellar membrane plus matrix extract prepared as follows. Freshly prepared wild-type flagella (Cole et al., 1998) were extracted with HMDEK buffer [10 mM HEPES, 5 mM MgSO4, 25 mM KCl, 0.5 mM EDTA, 1 mM dithiothreitol (DTT), pH 7.2] containing 0.5% Tergitol type NP-40 (Sigma) and 10 mM ATP, and the detergent-insoluble axonemal fraction was removed by centrifugation for 10 minutes at 14,000 g. The rationale for using Tergitol type NP-40 for extraction was that Tergitol appears to be a more gentle detergent than, for example, Nonidet type NP-40 (Wirschell et al., 2004), and we therefore reasoned that Tergitol is less likely to disrupt protein-protein interactions than other detergent types. Flagellar extracts (0.5 ml) were loaded onto 10-30% sucrose density gradients (12.4 ml) prepared in HMDEK buffer, and subjected to centrifugation at 36,000 g for 18 hours at 4°C. Fractions (0.5 ml each) were collected and analyzed by SDS-PAGE and western blotting. For calibration, a similar gradient was loaded with catalase (232 kDa; 11.3 S), aldolase (158 kDa; 4.4 S) and BSA (67 kDa, 4.4 S) and run in parallel. The sedimentation standards were all obtained from Amersham-Pharmacia. As internal standards, we used IC2 and LC1, which peak at around 19.4 S and 12 S, respectively (Pfister et al., 1982).
GST pull-down assays
For GST pull-down assays, wild-type (CC125 MT+) Chlamydomonas flagella were prepared and stored at –80°C as described previously (Cole et al., 1998). Frozen flagella in HMDEK buffer were thawed, extracted with 0.05% Nonidet type NP-40 (Calbiochem) plus 10 mM ATP, and axonemes removed by two periods of centrifugation at 14,000 g for 5 minutes. Extract from freshly excised 10-week-old female rat brain frontal cortex (generously supplied by David Wells, MCDB Department, Yale University) was prepared in HMDEK buffer plus 0.05% Nonidet NP-40 using a Dounce homogenizer, and the extract was cleared by centrifugation at 14,000 g for 10 minutes.
GST fusion proteins were produced by growing E. coli strains LP22 (GST-CrLis1), LP23 (GST-rat NudC) or LP36 (GST) at 30°C in Luria Bertani broth supplemented with appropriate antibiotics until the cultures had reached an OD600 of 0.5. IPTG was then added to 1 mM final concentration and the cultures were incubated overnight at 18°C, harvested by centrifugation, rinsed in PBS and flash-frozen in liquid N2. Frozen cells from 500 ml of culture were lysed by incubation in 50 ml PBS plus 0.1% Nonidet NP-40 for 10 minutes at room temperature, the DNA sheared by sonication, and the lysates cleared by two periods of centrifugation at 12,000 g for 10 minutes. The cleared lysates were each incubated with 1 ml of glutathione-agarose (Sigma) for 1 hour at 4°C, the agarose beads collected by centrifugation and washed three times for 10 minutes with HMDEK buffer containing 0.1% Nonidet NP-40.
Growth, induction and harvest of bacteria expressing MBP-CrLis1 (LP45) or MBP (LP44) were done as described above for GST fusion proteins, and frozen bacterial pellets were lysed in PBS containing 1 mg/ml of lysozyme and 0.1% Nonidet NP-40. After sonicating the lysates to shear the DNA, followed by centrifugation to pellet cellular debris, the cleared lysates were incubated for 1 hour at 4°C with amylose resin (New England Biolabs). The resin was collected by centrifugation, washed three times for 10 minutes in PBS plus 0.1% Nonidet NP-40, and bound proteins were eluted with PBS plus 10 mM maltose. After concentrating the proteins in Centricon 10 tubes (Millipore), glycerol was added to 25% final volume and the samples stored at –80°C until use.
GST fusion protein-coated beads were incubated with either Chlamydomonas flagellar extract (∼2-3 mg/ml), rat brain extract (∼2-3 mg/ml) or purified MBP-CrLis1 or MBP (0.1 mg/ml) for 1 hour at 4°C, the beads collected by centrifugation and washed three times for 10 minutes with HMDEK buffer plus 0.05% Nonidet NP-40.
SDS-PAGE and immunoblot analysis
Isolated Chlamydomonas flagella, flagellar fractions or pellets from GST pull-down assays were solubilized in Laemmli buffer and analyzed by SDS-PAGE and immunoblotting using standard procedures. Freshly excised murine tracheae, oviduct and ovary as well as whole rat brain were solubilized in PBS containing 1% SDS, vortexed vigorously in the presence of glass beads, and the DNA sheared with a syringe. The lysates were cleared by centrifugation, mixed with Laemmli buffer and analyzed by SDS-PAGE and immunoblotting using standard procedures. Secondary antibodies were detected by chemiluminescence or colorometric methods. For immunoblotting, antibodies were diluted as follows: anti-CrLis1, 1:1000; anti-IC2, 1:4000; anti-α-tubulin, 1:1000; anti-rat NudC, 1:500; anti-actin, 1:1000; anti-LC1, 1:100; anti-LC2, 1:100; anti-MBP, 1:4000; anti-Lis1 ab2607, 1:1000; anti-Lis1 sc-7577, 1:500.
Immunofluorescence microscopy of paraffin-embedded tissue sections and cultured fibroblasts
Oviduct, ovaries and tracheae were removed from adult female mice, fixed, embedded in paraffin and cut into 8-μm-thick sections that were collected on microscope slides for IFM analysis as previously described (Teilmann and Christensen, 2005). NIH3T3 fibroblasts were cultured for 24 hours in serum-free medium to induce growth arrest, fixed and subjected to IFM analysis as described by Schneider et al. (Schneider et al., 2005). Fixed cells and tissue sections were incubated at room temperature for 2 hours with mouse monoclonal anti-acetylated α-tubulin antibody (diluted 1:50,000) and goat polyclonal Lis1 antibody (diluted 1:200). After washing, cells and tissue sections were incubated for 1 hour at room temperature with Alexa Fluor 568-conjugated rabbit anti-mouse immunoglobulin G (IgG) and Alexa Fluor 488-conjugated donkey anti-goat IgG secondary antibodies (Molecular Probes). Sections and cells were mounted in PBS containing 70% glycerol and 2% N-propylgallate and the slides sealed with nail polish. Samples were observed on an Eclipse E600 microscope (Nikon) with EPI-FL3 filters and MagnaFire cooled CCD camera (Optronics, Goleta, CA), and digital images were processed using Adobe Photoshop version 6.0.
Immunofluorescence microscopy of isolated tracheal cells
Adult mice were euthanized by CO2 asphyxiation and their tracheae removed and treated with Hanks' balanced salt solution supplemented with 25 mM HEPES, pH 7.4. Ciliated cells were obtained by scratching the epithelial surface with a scalpel or alternatively by enzymatic digestion with 0.5% type XIV bacterial protease (Sigma). Enzymatically isolated cells were kept in Dulbecco's modified Eagle's medium (DMEM) supplemented with 25 mM HEPES, pH 7.4, 10% fetal bovine serum and checked for viability by assessing ciliary beating. Cells were briefly washed in PBS pH 7.4 and left to adhere on poly-L-lysine-coated slides prior to fixation with 4% paraformadehyde for 15 minutes at room temperature. Cells were permeabilized with 1% IGEPAL CA-630 (Sigma), blocked with 5% normal goat serum and incubated with rabbit polyclonal anti-hLis1 antibody (ab2607, Abcam) or anti-p24 dynactin antibody at 4°C for 16 hours. Subsequently, samples were incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG (Molecular Probes) and DNA counterstained with DAPI. Cells were observed with an Olympus BX51 epifluorescence microscope equipped with a Magnafire CCD camera for image acquisition.
Acknowledgements
We thank Ramila Patel-King and Oksana Gorbatyuk (University of Connecticut Health Center) for their assistance with molecular and biochemical analysis of CrLis1, David Wells (Yale University) and Gert Christoffersen (University of Copenhagen) for providing us with rat brain lysates, Li-Yuan Yu-Lee (Baylor College of Medicine) for rat NudC plasmids and antibodies, and Anne Grete Byskov (Rigshospitalet, Copenhagen, Denmark) for supplying us with murine tissue sections for IFM and immunoblot analyses. Hue Tran provided expert technical assistance with Chlamydomonas culture and flagellar preparations, and Christian A. Clement assisted with IFM. This work was supported by grants from the National Institutes of Health to J.L.R. and S.M.K., from the Danish Natural Science Research Council (no. 272-05-0411) and the Novo Nordisk Foundation to L.B.P., and from the Carlsberg Foundation to S.T.C. S.M.K. is an investigator of the Patrick and Catherine Weldon Donaghue Medical Research Foundation.