Chlamydomonas Axoneme Isolation and Dynein Purification

Flagellar axonemes were prepared from Chlamydomonas reinhardtii using standard protocols (Witman, 1986; King, 1995). Dynein was extracted with 0.6 M NaCl and purified by centrifugation in a 5–20% sucrose density gradient (King et al., 1986). Gradient fractions were concentrated in a Centricon 30 unit (Amicon Corp., Danvers, MA). Nonspecific protein binding was minimized by preincubating the unit for 48 h with 5% Tween-20.

For the initial identification of the 14,000-M _r protein, the wild-type strain cc124- was used. A mutant lacking the outer arm, oda9, was used for all subsequent purifications. Axonemes from mutants lacking inner arm components (ida1, ida2, ida3, and ida4), radial spokes (pf14), and the central pair microtubule complex (pf18) were used to localize the 14,000-M _r protein.

Peptide Sequencing

Sucrose gradient fractions containing the 14,000-M _r protein were concentrated, separated by electrophoresis in a 5–15% acrylamide gradient gel, and blotted to polyvinylidene difluoride (PVDF) membrane (Immobilin P^sq; Millipore Corp., Bedford, MA) in 10 mM NaHCO₃, 3 mM Na₂CO₃, 0.01% SDS, and 20% methanol. The blot was stained with 0.2% Ponceau S, and a thin strip was probed for the 14,000-M _r protein. The immunoreactive band was then identified on the unprobed strip, excised, and treated with trypsin. Peptides eluting from the membrane were purified by reverse-phase chromatography on a C₈ column. Two peptides were sequenced (model 492A sequencer; Applied Biosystems Inc., Foster City, CA) at the Protein Chemistry Facility at the Worcester Foundation for Biomedical Research (Shrewsbury, MA). For one peptide, a 12-h cycle of treatment with TFA vapor at 60°C was required to remove a blocked NH₂ terminus via an Asp-Pro cleavage. Peptide masses were determined by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry using a Linear Voyager Biospectrometry workstation (PerSeptive Biosystems, Framingham, MA).

Electrophoresis and Immunoblotting

All Chlamydomonas samples were electrophoresed in 5–15% polyacrylamide gradient gels and either stained with Coomassie brilliant blue or blotted to nitrocellulose (0.2 μm pore size; Schleicher & Schuell, Keene, NH) in 10 mM NaHCO₃, 3 mM Na₂CO₃, 0.01% SDS, and 20% methanol. For immunoblotting, the nitrocellulose was blocked in 5% dried milk, and 0.1% Tween-20 in TBS before being probed with blot-purified primary antibody followed by a peroxidase-conjugated secondary antibody (King et al., 1996a ). After washing, the antibody signal was visualized using an enhanced chemiluminescent system (ECL; Amersham Corp., Arlington Heights, IL) and Rx film (Fuji Photo Film Co. Ltd., Tokyo, Japan). Total protein was subsequently detected using 0.2% Ponceau S. Antibodies used were R4058 vs. the Chlamydomonas 8,000-M _r LC (King and Patel-King, 1995b ), R5205 vs. human Tctex1 (King et al., 1996b ), and R5391 vs. the Chlamydomonas Tctex2 homologue LC2 (Patel-King et al., 1997).

Quantitation of Coomassie blue–stained gels was performed using an IS1000 digital imaging system (Alpha Innotech, San Leandro, CA).

Molecular Cloning

A gene-specific primer (5′-GCGCGAATTCCTSCAGAACCAGCAGTAC- 3′) was designed based on the peptide sequence LQNQQY and incorporated the Chlamydomonas codon bias (Harris, 1989). The primer also incorporated an EcoRI site and a GC clamp at the 5′ end. The reverse primer was the standard oligo (dT) adapter primer (5′-GCGCGTCGACTCGAGT₂₀V-3′) that contains SalI and XhoI sites at the 5′ end. 100-μl PCRs were performed in 10 mM Tris-Cl, pH 8.85, 25 mM KCl, 5 mM (NH₄)₂SO₄, 2 mM Mg SO₄, and 0.2 mM dNTP, and they contained 1 μg of each primer. The template used was first-strand cDNA made from mRNA isolated from cells actively regenerating their flagella. Samples were heated at 95°C for 5 min, cooled on ice, and 2.5 U of Pfu DNA polymerase (Stratagene, La Jolla, CA) were added. Samples were subjected to 35 rounds of the following program: 96°C for 1 min, 35°C for 1 min, and 75°C for 2 min. Products were subcloned into pBluescript II SK⁻ (Stratagene) and were sequenced using Sequenase v2.0 and a 7-deaza-dGTP sequencing kit (U.S. Biochemical Co., Cleveland, OH).

The PCR product corresponding to the 14,000-M _r protein was used to screen a λZapII cDNA library made from mRNA derived from cells actively regenerating flagella (Wilkerson et al., 1995). Phagemids were rescued using helper phage, and the longest clone was sequenced on both strands using a double-stranded DNA template. Northern and Southern blots were prepared by standard methods (Sambrook et al., 1989) and probed using the conditions described in King and Patel-King (1995b). The probes used were the 14,000-M _r protein full-length cDNA, a full-length cDNA for LC4 (King and Patel-King, 1995a ), and a full-length calmodulin cDNA that had been isolated during a screen of the λZapII library for the calmodulin-like LC4 protein.

Tctex1 Is a Component of Inner Arm I1

To further define the intraflagellar associations of the Chlamydomonas Tctex1-like protein, axoneme samples were prepared from mutants lacking various axonemal structures, including the outer arm (oda9), inner arm I1 (ida1, ida2, and ida3), inner arms I2/I3 (ida4), the radial spokes (pf14), and the central pair complex (pf18). When these samples were probed with an antibody (R5391) against LC2 of the outer arm, which is a homologue of the putative t complex distorter Tctex2, this protein was found to be missing only in axonemes prepared from the oda9 mutant, as described previously (Patel-King et al., 1997). Interestingly, the minor 15,000-M _r protein recognized by this antibody was present in all axoneme samples, indicating that it is not a component of the dynein arms, radial spokes, or central pair complex. In contrast, analysis of the same samples with the R5205 antibody revealed that the Tctex1-like protein was missing specifically in those strains (ida1–ida3) unable to assemble inner arm I1 (Fig. ​(Fig.3).3). This result strongly suggests that in Chlamydomonas, the Tctex1 protein is an integral component of the inner arm I1 complex.

Open in a separate window

Figure 3

Intra-axonemal localization of Tctex1. Axonemes were prepared from wild-type Chlamydomonas (WT) and from strains lacking the outer arm (oda9), inner arm I1 (ida1, ida2, and ida3), inner arms I2/I3 (ida4), radial spokes (pf14), and central pair complex (pf18). 100 μg of each preparation was electrophoresed in a 5–15% acrylamide gradient gel and either stained with Coomassie blue (upper panel) or blotted to nitrocellulose (lower panels) and probed with R5205 and R5391 antisera to detect Tctex1 and LC2 (Tctex2) of outer arm dynein, respectively. LC2 is missing only in the oda9 mutant, which lacks the outer arm (the smaller immunoreactive band is present in all strains tested as described previously [Patel-King et al., 1997]). Tctex1 is specifically absent in those strains that fail to assemble inner arm I1, suggesting that it may be a component of that complex.

To further confirm the association deduced from genetic dissection, axonemes were prepared from the oda9 mutant and subjected to high salt extraction. The extract (which contained no outer arm components) was then sedimented through a 5–20% sucrose gradient. Electrophoretic and immunological analysis of the resulting fractions revealed that the Tctex1-like protein sedimented at ~18 S and indeed precisely comigrated with bona fide inner arm I1 components (Fig. ​(Fig.4).4). This result confirmed that Tctex1 is an inner arm I1 polypeptide. The small membrane pool of Tctex1 was missing from ida1 flagella, suggesting that it represents inner arm I1 that either dissociated during detergent extraction or had not yet been assembled into the axonemal superstructure (not shown).

Open in a separate window

Figure 4

Tctex1 is a component of inner arm dynein I1. Proteins extracted from oda9 axonemes (which completely lack the outer arm) by high salt were separated by sucrose density gradient centrifugation. Equal volumes of each fraction were electrophoresed in 5–15% gradient gels. One gel was stained with Coomassie blue (upper panel), while the other was blotted to nitrocellulose and probed with R5205 antibody (lower panel). The bottom of the gradient is at the left. After immunoblotting, the nitrocellulose was restained with Ponceau S to reveal the location of individual lanes and the relative molecular mass markers. Tctex1 comigrates with inner arm I1 components (IC140 and IC138) at ~18 S, but it is clearly distinct from arms I2/3, which contain actin, p28, and centrin. (inset, upper panel) The I1 dynein–containing fractions from two gradients were pooled and concentrated in a Centricon-30 ultrafiltration unit and electrophoresed in a 5–15% acrylamide gel. The LC region of the Coomassie blue–stained gel is shown. Three bands are evident. The upper band was recognized by the R5205 antibody and is therefore Tctex1. The lower band was detected by the R4058 antibody, indicating that it represents an additional pool of the 8,000-M _r LC (King and Patel-King, 1995b ). A third LC band of unknown origin is also present.

To further analyze the LC complement of this dynein, the I1-containing fractions from several gradients were pooled and concentrated in a Centricon 30 ultrafiltration unit. After electrophoresis and staining with Coomassie blue, three distinct bands were observed in the low molecular weight region (Fig. ​(Fig.4,4, inset). Immunological analysis revealed that the slowest migrating band was Tctex1. The fastest migrating band was recognized by antibody R4058 (King and Patel-King, 1995b ) and, thus, is likely to be identical to the 8,000-M _r LC previously found in both outer arm and cytoplasmic dyneins, as well as in several other enzyme systems (Benashski et al., 1997). Intriguingly, a third band (~12,000-M _r) was also observed in arm I1; at the present time, no further information is available concerning the identity of this protein.

Quantitative densitometry of Coomassie blue–stained gels was used to determine the stoichiometry of the Tctex1 and 8,000-M _r LCs within the I1 dynein (Table ​(TableI).I). This analysis revealed that there are two copies of Tctex1 and suggested the presence of 8–12 copies of the 8,000-M _r LC per I1 arm (this latter value is likely an overestimate; see Discussion). Furthermore, based on the presence of two HCs (Iα and Iβ), the data indicate that the I1 particle contains one to two copies of IC140, a single IC138, and one to two copies of IC97. A 34,000-M _r protein has also been suggested to be part of the I1 complex (King and Dutcher, 1997). Although a band of that relative molecular mass is present in the appropriate region of the sucrose gradients, the 34,000-M _r polypeptide peak is offset from that of I1 dynein by one fraction and, instead, it appears to comigrate with 100,000- and 106,000-M _r proteins.

Table I

Stoichiometry of Components within Inner Dynein Arm I1^*

Component‡		Relative stoichiometry§		Copies per dynein arm
HC		1.00		2‖
IC140		0.71 0.53 0.63 (0.5–1.0)		1–2
IC138		0.53 (0.5)		1
IC97		0.95 0.67 0.56 (0.5–1.0)		1–2
Tctex 1		1.28 1.12 0.71 (1.0)		2
8,000-M r LC		5.50 4.00 (4.0–5.5)		8–12 ¶

Open in a separate window

^*Determined by quantitative densitometry of Coomassie blue–stained gels.  

^‡The ~12,000-M _r protein observed in I1 samples is present at a stoichiometry of one per particle. This polypeptide, however, has not yet been formally demonstrated to be a dynein LC.  

^§The values are quoted relative to the HCs. The most likely stoichiometry is in parentheses.  

^‖Previously published results indicate that there are two HCs (Iα and Iβ) per particle (Piperno et al., 1990).  

^{¶ 35}S-Labeling of Chlamydomonas proteins in vivo suggests this value may be an overestimate by approximately twofold (Wilkerson, C.G., and G.B. Witman, personal communication).  

Molecular Analysis of Chlamydomonas Tctex1

To conclusively identify the 14,000-M _r protein as the Chlamydomonas homologue of Tctex1, it was essential to compare the sequences of the algal and mammalian proteins. Therefore, the sucrose gradient–purified inner arm I1 from the oda9 mutant was concentrated in a Centricon 30, and the components were separated by electrophoresis and blotted to a PVDF membrane. The 14,000-M _r band was identified using blot-purified R5205 antibody, excised from the PVDF blot, and digested in situ with trypsin. Eluted peptides were purified by reverse-phase chromatography (Fig. ​(Fig.5,5, a). Mass spectrometry revealed that peak (i) contained a single peptide with a mass (M+H⁺) of 1,825 D (Fig. ​(Fig.5,5, b). This peptide subsequently yielded the sequence ESIDAVLQNQQYSEAK, which has a calculated molecular weight of 1,823 D. Analysis of peak (Fig. ​(Fig.55 a, ii) revealed a major peptide of 2,463 D and a minor one of 2,479 D (Fig. ​(Fig.5,5, c). As the mass difference between these two molecules is 16 D, the latter very likely represents a methionine oxidation product of the former. Initially, Edman degradation of the peak (Fig. ​(Fig.55 a, ii) peptide yielded no products, indicating that the NH₂ terminus was blocked. However, after prolonged treatment with TFA, which cleaves Asp-Pro bonds, the sequence PAVEEAAFVADDVSNII was obtained.

Open in a separate window

Figure 5

Peptide purification from Tctex1. The I1-containing fractions from seven gradients were pooled and concentrated in a Centricon 30 ultrafiltration unit. The concentrate was electrophoresed in a 5–15% acrylamide gradient gel and blotted to polyvinylidene difluoride. A strip was excised and probed with R5205 to localize the Tctex1 band. The remaining Tctex1 band was then excised from the unprobed PVDF and digested in situ with trypsin. Peptides eluting from the membrane were purified by reverse-phase chromatography on a C₈ column (a). Peaks (i) and (ii) were sequenced. Mass spectrometry (b and c) revealed a single peptide of 1,825 D present in peak (i). Peak (ii) contained a single major peptide of 2,463 D and a minor amount of its methionine oxidation product at 2,479 D.

A gene-specific primer based on the sequence LQNQQY and an oligo (dT) adaptor primer were then used in the PCR using first-strand cDNA derived from mRNA enriched for flagella sequences and generated a product of ~1 kb. Sequencing of the 5′ end of this product revealed that the predicted residues were encoded immediately after the primer. Therefore, this product was used to screen a λZapII cDNA library made from RNA derived from Chlamydomonas actively regenerating their flagella (Wilkerson et al., 1995). Multiple clones were obtained, and the longest was sequenced on both strands.

The largest clone isolated was 1,190 bp in length and contained a single open reading frame of 342 bp encoding a protein of 114 residues with a calculated molecular weight of 12,668 D and a predicted isoelectric point of 5.17 (Fig. ​(Fig.6).6). The reading frame is preceded by a 115-bp 5′ untranslated region that contains three in-frame stop codons before the first Met residue. The reading frame terminates with a single stop codon followed by a 707-bp 3′ untranslated region before the poly A tail. Both peptide sequences obtained from the authentic molecule were found within the predicted coding sequence (33/33 residues correct), confirming that the clone indeed encodes Chlamydomonas Tctex1.

Open in a separate window

Figure 6

Sequence of Chlamydomonas Tctex1. Nucleotide and predicted amino acid sequence of the cDNA clone encoding Tctex1 are shown. The residues shown in bold type were obtained from peptide sequencing (33/33 residues correct) and confirm the identity of the clone. The 5′ untranslated region contains three in-frame stop codons. These sequence data are available from GenBank/EMBL/DDBJ under accession number AF039437.

The peak (i) peptide was found to be preceded by a Lys residue, as required for a tryptic fragment. The peptide from peak (ii) represents the NH₂ terminus of the protein. It contains the predicted acid-sensitive Asp-Pro bond immediately before the peptide sequence obtained. Intriguingly, the actual mass of the blocked NH₂-terminal peptide was 42-D greater than calculated from the encoded sequence. This mass difference strongly suggests that the modification resulting in NH₂-terminal blockage was caused by acetylation of the initial methionine residue.

Southern blot analysis of Chlamydomonas genomic DNA revealed a single band after digestion with BamHI, PstI, PvuII, and SmaI (Fig. ​(Fig.77 a), suggesting that a single gene for this protein exists within Chlamydomonas. Two messages for this protein, however, were detected on Northern blots (Fig. ​(Fig.77 b). The larger message of ~1.35 kb was observed in total RNA samples from nondeflagellated cells, but it was not detected in samples from cells that were undergoing flagellar regeneration. In contrast, the smaller ~1.30-kb mRNA was greatly upregulated in cells actively regrowing flagella. This differential regulation of both message levels was very apparent when compared with the levels of the single messages for two other proteins (Fig. ​(Fig.77 c). The mRNA for LC4 of the outer arm was essentially undetectable in nondeflagellated cells, but it was highly upregulated during regeneration (King and Patel-King, 1995a ). Calmodulin mRNA exhibited a different pattern and was readily detectable in both samples because the same protein is used constitutively in the cytoplasm but is also upregulated and required for flagellar assembly and/or function (Gitelman and Witman, 1980; Zimmer et al., 1988).

Open in a separate window

Figure 7

Southern and Northern blot analyses. (a) Southern blot of 10 μg genomic DNA from Chlamydomonas strain S1D2 digested with BamHI, PstI, PvuII, and SmaI. Single bands were observed in all lanes, suggesting that there is only a single gene for this LC within Chlamydomonas. Standards are indicated on the left in kilobases. (b) Northern blot of 20 μg total RNA obtained from nondeflagellated cells (NDF) and from cells that had been deflagellated and allowed to regenerate flagella for 30 min (30′ post DF). Standards are shown on the left in kilobases. Two messages were observed. The message at ~1.35 kb was detected only in the NDF sample, while a smaller message (~1.30 kb) was greatly upregulated after deflagellation. (c) Enlargement of Fig. ​Fig.77 b (upper panel) and comparison with both a deflagellation-regulated message (LC4 of outer arm dynein) and one that is present constitutively but is also upregulated (calmodulin).

The secondary structure for Chlamydomonas Tctex1 was predicted using PHD (Fig. ​(Fig.88 a; Rost and Sander, 1993). The NH₂-terminal half of the molecule has two segments that have a high probability of being helical. Significant portions of both segments (Fig. ​(Fig.88 a, i and ii) are predicted to be highly amphiphilic and may therefore be involved in protein–protein interactions. The COOH-terminal region consists of a series of extended sheet structures.

Open in a separate window

Figure 8

Sequence analysis of Chlamydomonas flagellar Tctex1. (a) The secondary structure of Tctex1 was predicted using PHD (Rost and Sander, 1993). E, extended sheet; H, helix. Helical stretches (i) and (ii) are amphiphilic and displayed using HELICALWHEEL. Hydrophobic and hydrophilic residues cluster to opposite sides of the helix. (b) Sequence comparison between Chlamydomonas flagellar Tctex1 and human Tctex1 (D50663). The alignment was generated by GAP using the default parameters. These proteins share 62% identity (68% similarity) with the smallest Poisson probability P _(n) = 6.3 × 10⁻⁴⁵ (calculated by BLAST). (c) Phylogenetic analysis of the members of the Tctex1 protein family. The relationship was calculated with DISTANCES and plotted with GROWTREE (UPGMA option). Chlamydomonas flagellar Tctex1 is more closely related to mammalian and insect Tctex1 than it is to human rp3, which is an additional cytoplasmic dynein LC sharing 55% identity with Tctex1 (King, S.M., E. Barbarese, J.F. Dillman III, S.E. Benashski, K.T. Do, R.S. Patel-King, and K.K. Pfister. 1997. Mol. Biol. Cell. 8:163a).

The Chlamydomonas 14,000-M_r Protein Is Closely Related to Murine Tctex1

The Tctex1 protein family consists of two major branches exemplified by Tctex1 (shown in Fig. ​Fig.88 c) and by murine Tctex2 and its Chlamydomonas homologue LC2 (Patel-King et al., 1997). Molecular cloning of the Chlamydomonas 14,000-M _r protein described here revealed that it shares 62% identity with mammalian Tctex1. The closest known mammalian homologue of Tctex1 is the rp3 polypeptide that was originally cloned by Roux et al. (1994) and is ~55% identical with Tctex1. The rp3 protein has recently been demonstrated to be an additional cytoplasmic dynein LC that is most prevalent in cells and tissues distinct from those expressing Tctex1 (King, S.M., E. Barbarese, J.F. Dillman III, S.E. Benashski, K.T. Do, R.S. Patel-King, and K.K. Pfister. 1997. Mol. Biol. Cell. 8:163a). Importantly then, the dendrogram clearly distinguishes rp3 from the Chlamydomonas, insect, and mammalian Tctex1 proteins. Thus, phylogenetic analysis strongly supports the identification of the 14,000-M _r protein as the Chlamydomonas version of murine Tctex1. Furthermore, there are several partial and complete sequences of additional mammalian, nematode, and trypanosome Tctex1 homologues present in the Expressed Sequence Tag database (not shown). All of these proteins are significantly less related to Tctex1 than is the Chlamydomonas LC and, indeed, several of them group with the much more diverse Tctex2 branch of this family.

Implications of a Flagellar Form of Tctex1 for the Mechanism of Meiotic Drive

Meiotic drive or transmission ratio distortion is thought to derive from defects in spermiogenesis, where sperm bearing the t haplotype–containing chromosome are normal, but those with the wild-type t complex are dysfunctional and unable to fertilize an oocyte (Olds-Clarke and Peitz, 1985). This phenotype could be caused by aberrant flagellar motility because sperm from t/+ males exhibit subtle defects in motility that have been shown to affect their ability to reach the oocytes both in vivo and in vitro (see Olds-Clarke [1997] for recent review). Genetic analysis suggests that meiotic drive derives from the interaction of three to four distorter proteins with a responder that is expressed after meiosis (Lyon, 1984, 1986). To achieve distortion, the responder must be kept in close association with the nucleus that encoded it. We have recently suggested that distortion might be achieved if the responder acts as the gatekeeper or sorting mechanism for flagella assembly located at or near the basal body (Patel-King et al., 1997). The wild-type responder is thought to interact with both wild-type and t mutant distorter proteins, whereas the mutant responder may offer a protective effect by interacting poorly with mutant distorters and thereby incorporating mainly the wild-type versions (Lyon, 1984, 1986; Cebra-Thomas et al., 1991). The consequences of this model for incorporation of Tctex1, Tctex2, and the Tcd2 distorter into + and t sperm are illustrated in Fig. ​Fig.9.9. Simultaneous analysis of the speed and path shape of sperm populations from t/+ males shows two distinct peaks, one of which resembles that of wild-type sperm and one that is similar to sperm from mice carrying two t haplotypes (Olds-Clarke and Johnson, 1993). These data support the idea that t/+ males produce two subpopulations of sperm that differ in their motility characteristics.

Open in a separate window

Figure 9

A mechanism for flagellar dynein–mediated meiotic drive. In this model, the responder is hypothesized to be a “gatekeeper” or sorting mechanism that determines what may enter the growing flagellum during spermiogenesis. Because t mutations in both putative distorters Tctex1 and Tctex2 could lead to flagellar dynein dysfunction, their incorporation into + sperm by the wild-type responder (Tcr ⁺) might result in defective motility. In contrast, the t mutant responder (Tcr^t) is thought to protect the t sperm with which it associates by being unable to interact with (or having a lower affinity for) the mutant distorters. In the heterozygous case, this would lead to incorporation of only wild-type dyneins into t haplotype–bearing sperm and thus to normal motility. The molecular identities of all the distorters are unknown at the present time. In this model, it is hypothesized that Tctex1 and Tctex2 are the Tcd1 and Tcd3 distorters, respectively.

Tctex1 and Tctex2 are located in the appropriate regions of the t haplotype to be candidates for distorter factors Tcd1 and Tcd3, respectively (Lader et al., 1989; Rappold et al., 1987; Huw et al., 1995). The t haplotype forms of both proteins contain mutations that are likely to affect dynein function. In Tctex1, the mutation Q41H disrupts a tripeptide sequence that has been completely conserved in all Tctex1 proteins (except for Drosophila, which contains Asn at the equivalent position) and in the closely related rp3. Similarly, the t form of Tctex2 contains a three-residue deletion and a proline insertion within a predicted helical segment. As Tctex2 is an essential outer arm component, it is relatively easy to see how its dysfunction might affect sperm motility. However, identification of Tctex1 as a cytoplasmic dynein LC was more problematic, because this protein is present in many tissues but heterozygous t haplotype–bearing mice show an abnormal phenotype only in male germ cells. If Tctex1 is a component of an inner arm in mouse sperm flagella, this would provide a mechanism by which the t haplotype allele could lead to alterations in sperm activity.

The action of the distorter proteins is cumulative such that the degree of ratio distortion is directly related to the particular alleles present in a given haplotype (Lyon, 1984, 1986). These allele-specific effects can be readily understood through increasingly severe consequences for flagellar dynein function and, thus, for the motile properties of sperm. In our model, distortion arises because the t responder interacts only weakly with the t distorters and, therefore, preferentially incorporates the wild-type forms into the growing sperm tail. Sperm from t^x/t^y mice³ are known to contain both Tctex1 and Tctex2 (Huw et al., 1995; O'Neill and Artzt, 1995; this study). All sperm from these mice, however, exhibit defective motility and are completely nonprogressive (Olds-Clarke and Johnson, 1993). This observation is also predicted by our model (Table ​(TableII).II). The t^x/t^y homozygous animals represent the sole occasion on which the t forms of both the distorters and responder are the only versions present within the cell. We suggest that in the absence of the competing “high affinity” + distorters, the “low affinity” t forms would interact with the t responder and become incorporated by default. Thus, in the absence of competition with the high affinity distorters, low affinity interactions could suffice for the insertion of the t mutant distorters. The predicted variation in motile properties between + and t sperm from heterozygotes may then reflect a qualitative difference in the type of distorter protein (+ or t) incorporated into the two sperm classes.

Tctex1 and Tctex2 are candidates for the proximal Tcd1 and Tcd3 distorters. Analysis of rare partial haplotypes has revealed, however, that the strongest distorter (Tcd2) is located in the distal portion of the t complex (see Fig. ​Fig.11 in Pilder et al., 1993). The molecular identity of this protein, which may be encoded at the Hybrid Sterility-6 locus (Pilder et al., 1993), is unknown at the present time. Sperm from the Hst6 mutant exhibit an abnormal flagellar curvature as well as poor motility. This phenotype is only seen in motile gametes, raising the possibility that it derives from the misregulation of flagellar beating, perhaps through direct effects on the dynein arm (and thus potentially on Tctex1 and/or Tctex2) or radial spoke systems. Intriguingly, a presumptive axonemal dynein HC has recently been mapped to the distal region of the t complex (Dnahc8; Vaughan et al., 1996). Although the precise location of the Dnahc8 gene relative to Hst6 is not yet clear, this does raise the possibility that the distal distorter could be a dynein component. Such a scenario might then explain how Tcd2 ^t greatly enhances the effects due to the other distorters.

A Generic Requirement for Tctex1 Family LCs in Dyneins Containing Multiple HCs

Tctex1 is part of a diverse protein family, members of which have now been identified in the major cytoplasmic dynein isoform (Tctex1 [King et al., 1996b ] and rp3 [King, S.M., E. Barbarese, J.F. Dillman III, S.E. Benashski, K.T. Do, R.S. Patel-King, and K.K. Pfister. 1997. Mol. Biol. Cell. 8:163a]), outer arm dynein (Tctex2; Patel-King et al., 1997), and inner arm I1 (Tctex1; this study). At least in the outer arm, the Tctex2 LC is an essential component and null mutants fail to assemble the entire structure (Pazour, G.J., A. Koutoulis, H. Sheng, R.S. Patel-King, S.M. King, and G.B. Witman, unpublished data). This LC family, however, is not apparently represented in the inner arm I2/3 complexes, all the components of which are now known (LeDizet and Piperno, 1995). Thus, Tctex1 family– containing dyneins are distinct in that they all include two or more HCs and at least one IC that is a member of the WD repeat protein family. In outer arm and cytoplasmic dynein, the Tctex1 family LCs are known to interact with the ICs and form part of the basal IC/LC complex (Mitchell and Rosenbaum, 1986; King and Witman, 1990; for review see Witman et al., 1991; King, S.M., E. Barbarese, J.F. Dillman III, S.E. Benashski, K.T. Do, R.S. Patel-King, and K.K. Pfister. 1997. Mol. Biol. Cell. 8:163a). By analogy, a similar IC-associated location is likely in inner arm I1 and suggests that the Tctex1 homologues play a generic and essential role in maintaining the integrity of these multi-HC complexes.

We thank S.E. Benashski and R.S. Patel-King for their assistance, and Dr. J. Leszyk (Worcester Foundation for Biomedical Research, Shrewsbury, MA) for peptide sequencing.

This study was supported by a New Investigator award from the Patrick and Catherine Weldon Donaghue Medical Research Foundation (to S.M. King) and by grants GM 51293 (to S.M. King) and HD 15045 (to P. Olds-Clarke) awarded by the National Institutes of Health.