INTRODUCTION

The major challenge in understanding ciliary and flagellarmotility is to determine the mechanisms that locally control themultiple dynein motors (Satir, 1985; Brokaw, 1994; Kamiya,1995). Structural and biochemical analysis of Chlamydomonasmutants has revealed that the central pair apparatus and radialspokes are central regulators of dynein arm activity (Smith andLefebvre, 1997a). For example, mutations that disrupt eitherthe central pair or the radial spokes result in paralysis by amechanism involving the inactivation of flagellar dyneinactivity (Huang et al., 1982; Piperno et al., 1992; Porter et al.,1992). Mutant, paralyzed axonemes can still undergo dynein-driven microtubule sliding in vitro (Witman et al., 1978), butat greatly reduced rates of microtubule sliding comparedto wild-type axonemes (Smith and Sale, 1992). In vitroreconstitution and functional assays demonstrated the radialspokes are required for wild-type dynein activity and that thevelocity of dynein-driven microtubule sliding is mediated byposttranslational modification of the inner dynein arms (Smithand Sale, 1992). The results of the in vitro experiments areconsistent with other evidence indicating the central pair/radialspoke complex controls dynein activity (Huang et al., 1982;Porter et al., 1992; Piperno et al., 1992, 1994; Gardner et al.,1994; Rupp et al., 1996; Yoshimura and Shingyoji, 1999).

Several studies have revealed that ciliary and flagellaractivity is controlled by phosphorylation (Brokaw, 1987; Satiret al., 1993; Walzak and Nelson, 1994; Tash and Bracho, 1994).Furthermore, in vitro analysis has indicated the protein kinasesand phosphatases responsible for regulation are anchored in theaxoneme (Hasagawa et al., 1987; Hamasaki et al., 1989; SanAgustin and Witman, 1994; Chaudhry et al., 1995). Forexample, based on the studies of Hasegawa et al. (1987), wepredicted Chlamydomonas flagellar dyneins are controlled bykinases and phosphatases anchored in the axoneme. To testthis, a pharmacological approach was taken with the predictionthat inhibitors of kinases and/or phosphatases could rescuewild-type dynein activity in isolated, paralyzed axonemes(Howard et al., 1994). Using this approach, specific inhibitorsof cAMP dependent kinase (PKA) rescue wild-type dyneinactivity in isolated axonemes lacking the radial spokes or thecentral pair apparatus (Howard et al., 1994). Evidently in theabsence of these axonemal structures, an axonemal PKAblocks dynein-driven microtubule sliding. In the same study, invitro reconstitution experiments also demonstrated that one ofthe target proteins is located in the inner arm dynein fraction.Thus, it was concluded that PKA is anchored in the axoneme,possibly in position to control inner arm dynein driven motility.Furthermore, since the kinase inhibitors resulted in rescue ofwild-type dynein activity, it was postulated the axoneme must

91Journal of Cell Science 113, 91-102 (2000)Printed in Great Britain © The Company of Biologists Limited 2000JCS1026

We postulated that microcystin-sensitive proteinphosphatases are integral components of theChlamydomonas flagellar axoneme, positioned to regulateinner arm dynein activity. To test this, we took a directbiochemical approach. Microcystin-Sepharose affinitypurification revealed a prominent 35-kDa axonemalprotein, predicted to be the catalytic subunit of type-1protein phosphatase (PP1c). We cloned theChlamydomonas PP1c and produced specific polyclonalpeptide antibodies. Based on western blot analysis, the 35-kDa PP1c is anchored in the axoneme. Moreover, analysisof flagella and axonemes from mutant strains revealed thatPP1c is primarily, but not exclusively, anchored in thecentral pair apparatus, associated with the C1 microtubule.Thus, PP1 is part of the central pair mechanism that

controls flagellar motility. Two additional axonemalproteins of 62 and 37 kDa were also isolated usingmicrocystin-Sepharose affinity. Based on direct peptidesequence and western blots, these proteins are the A- andC-subunits of type 2A protein phosphatase (PP2A). Theaxonemal PP2A is not one of the previously identifiedcomponents of the central pair apparatus, outer armdynein, inner arm dynein, dynein regulatory complex orthe radial spokes. We postulate PP2A is anchored on thedoublet microtubules, possibly in position to directlycontrol inner arm dynein activity.

Key words: Dynein, Cilium, Flagellum, Phosphatase, PP1, PP2A,Chlamydomonas, Cell motility

SUMMARY

Protein phosphatases PP1 and PP2A are located in distinct positions in the

Chlamydomonas flagellar axoneme

Pinfen Yang1, Laura Fox1, Roger J. Colbran2 and Winfield S. Sale1,*1Department of Cell Biology, Emory University School of Medicine, 1648 Pierce Drive, Atlanta, GA 30322, USA2Department of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, TN 37232, USA*Author for correspondence (e-mail: win@cellbio.emory.edu)

Accepted 2 November; published on WWW 9 December 1999

92

also bear protein phosphatases anchored in position to controlinner arm dynein activity.

Based on a pharmacological approach using axonemesisolated from wild-type and mutant cells, an axonemal proteinphosphatase type 1 (PP1) was discovered that either directly orindirectly acts to control inner arm dynein activity(Habermacher and Sale, 1996). This conclusion was based onspecific inhibitors of PP1, but the data did not exclude thepresence and involvement of additional axonemal proteinphosphatases in the activation of dynein activity. Moreover,genetic and biochemical analyses revealed that one targetsubstrate is the 138-kDa intermediate chain (IC138) of innerarm dynein I1, and that an axonemal microcystin-sensitiveprotein phosphatase, such as PP1, is either directly or indirectlyresponsible for de-phosphorylation of IC138 and rescue ofwild-type dynein driven motility (Habermacher and Sale,1997). The simplest model is that a microcystin-sensitiveprotein phosphatase, or phosphatases, is anchored in theaxoneme in position to regulate inner arm dynein activity.Alternatively, the microcystin sensitive axonemal phosphatasesmay act indirectly (King and Dutcher, 1997).

To test this model, a direct biochemical and molecularapproach was selected to define the presence and location ofmicrocystin-sensitive protein phosphatases in isolatedaxonemes from both wild-type and mutant Chlamydomonasflagella. As predicted from the physiological andpharmacological studies, we determined the axonemecontains two microcystin-sensitive phosphatases, proteinphosphatase type 1 (PP1) and protein phosphatase type 2A(PP2A). The major fraction of axonemal PP1 is located in thecentral pair apparatus associated with the C1 microtubule(Dutcher et al., 1984; Smith and Lefebvre, 1996, 1997a,b;Mitchell and Sale, 1999). The axonemal PP2A is likelyanchored on the outer doublet microtubules, possibly inposition to directly control phosphorylation of inner armdynein. This system of control, which involves the centralpair, radial spokes, and inner arm dynein I1, is likely toregulate flagellar waveform (Brokaw et al., 1982; Brokaw andKamiya, 1987; King and Dutcher, 1997). The discovery thatPP1 and PP2A are anchored in the wild-type 9+2 axoneme,but not anchored in specific Chlamydomonas mutants,provides a new opportunity to define subunits that anchorphosphatases in the cell.

MATERIALS AND METHODS

Cell strainsCell strains unless otherwise specified were provided byChlamydomonas Genetics Center (E. H. Harris, Duke University).Wild-type CC-125 was used for preparation of DNA and RNA. Drs E.F. Smith and P. Lefebvre (University of Minnesota and DartmouthCollege) provided pf16 and pf20 (Smith and Lefebvre, 1996, 1997a,b).All cells were grown in L-medium with aeration over a 14/10 light/darkcycle (Witman, 1986).

Molecular characterization of PP1cFor RT-PCR and northern analysis, total RNA was prepared from wild-type cells and/or at timed intervals following deflagellation as describedpreviously (Yang and Sale, 1998). Genomic DNA was prepared asdescribed (Wilkerson et al., 1995). For cloning PP1c, primers for RT-PCR were designed based on published peptide sequences (Song et al.,

1993). These sequences included KIKYPENF and EDGYEFF. Thedegenerate primer pairs for RT-PCR included Ps1 and Pa1: Ps1[CGCGGAATTCAARATIAARTAYCCIGAGAAYTTY];Pa1[CGCGCTCGAGRAARAACTCRTAICCRTCCT];

The additional primer pair, Ps2 and Pa2, was designed based on thegenomic sequence and used for RT-PCR to identify the 5′ end of thePP1c cDNA sequence:Ps2[CAAGGAGGGACCACTCTGACCTGCC];Pa2[GAAGAACTCGTACCCCTCCTCCAC].

The RT-PCR reactions were carried out as described previously(Yang and Sale, 1998). The 480 bp product from primers Ps1 and Pa1,was cloned into the EcoRI and XhoI site of pBluescript II KS(−)(Strategene, La Jolla, CA) and designated pRT1. The 840-bp productfrom primers Ps2 and Pa2 was cloned into pGemEasy (Promega,Madison, Wi.) and designated pRT2. The insert from pRT1 was usedas a probe for library screening and for Southern and northern blots.

Library screeningcDNA and genomic libraries were screened as described previouslyusing the insert of the pRT1, the RT-PCR clone, as a probe (Yang andSale, 1998). A λzapII cDNA library was kindly provided by C. G.Wilkerson and G. Witman (University of Massachusetts; Wilkerson etal., 1995) and a λfixII genomic library was kindly provided by E. F.Smith and P. Lefebvre (University of Minnesota; Smith and Lefebvre,1997b). The cDNA clone pC1 was excised from purified plaques withhelper phage (Wilkerson et al., 1995). The inserts from genomic cloneswere released from the phage vector by digestion with XbaI, SacI orNotI in separate reactions. The products of restriction digestion weresubcloned into pBluescript KS II(−).

BiochemistryThe flagella and axonemes from various strains were prepared asdescribed (Smith and Sale, 1992). Briefly, cells were de-flagellated withdibucaine, and axonemes were isolated by suspension of flagella inBuffer A (30 mM NaCl, 10 mM Hepes, pH 7.4, 5 mM MgSO4, 1 mMDTT, 0.5 mM EDTA, PMSF and aprotinin) containing 0.5% Nonidet-P 40 (NP-40, Calbiochem, LaJolla, CA). For SDS-PAGE, flagella andaxonemes from duplicate aliquots were resuspended with Buffer A toa final axoneme concentration of 5 mg/ml. Protein was determinedusing the Coomassie binding assay (Bio-Rad, Hercules, CA) using BSAas a standard. For SDS-PAGE, protein samples were fixed in theappropriate volume of 5×-sample buffer. For peptide sequencing,proteins were separated by SDS-PAGE, transferred to PVDF membraneand band purified as described (Yang and Sale, 1998). Dr John Leszyk(Protein Chemistry Core, University of Massachusetts) performedtryptic digestion, peptide separation by HPLC and microsequencing.

Cross-linking of axonemes with EDC (1-ethyl-3-[3-dimethylaminopropyl]-carbodiimide hydrochloride, Pierce, Rockford,IL) was carried out as described (Yang and Sale, 1998). Briefly,axonemes (suspended at 2 mg/ml in buffer A without DTT) were treatedfor 1 hour at room temperature with EDC freshly dissolved in water.The reaction was by terminated with a 20-fold molar excess of β-mercaptoethanol. Western blot analysis was carried out as described(Yang and Sale, 1998). Unless stated otherwise, an 8% mini-gel wasused for SDS-PAGE.

AntibodiesPolyclonal antibodies were raised in rabbits (Spring Valley laboratories,Sykesville, MD) against a synthetic peptide containing cysteine,tyrosine and the C-terminal 11 amino acids residues from theChlamydomonas PP1c (CYKPAETKAKGRR). For immuno-fluorescence localization, antibodies to PP1c were affinity purified tothe peptide conjugated to Affi-Gel 10 following directions supplied bythe manufacturer (Bio-Rad, Hercules, CA). Dr D. Pallas (EmoryUniversity) provided the monoclonal antibodies against the catalyticsubunit of PP2A (1D6, 4G7; Orgis et al., 1999). Drs E. F. Smith and P.Lefebvre (Dartmouth College and the University of Minnesota) kindly

P. Yang and others

93Phosphatases anchored in flagellar axonemes

provided the antibodies to the pf16 and pf20 gene products (Smith andLefebvre, 1996, 1997b).

Immunofluorescence microscopyFor immuno-localization, axonemes were prepared by two differentprotocols with identical results. For both procedures, cells suspended in10 mM HEPES, were attached to the glass coverslips. In the firstmethod, cells were then permeabilized with 0.5% Nonidet in amicrotubule stabilizing buffer (MTSB; Johnson and Rosenbaum, 1992)for 10 minutes at room temperature and then fixed in 4% para-formaldehyde/MTSB for 30 minutes at room temperature. In the secondmethod, cells were fixed in 4% para-formaldehyde/MTSB for30 minutes at room temperature and then immersed in 0.5%Nonidet/MTSB/4% para-formaldehyde for 20 minutes at roomtemperature. In each case, coverslips were rinsed in MTSB, immersedin –20°C methanol and briefly air-dried. Cells were re-hydrated in 3changes of PBS, blocked with 3% BSA in PBS, and incubated inprimary antibodies in PBS/3% BSA overnight at 4°C. Followingextensive rinses with PBS, cells were incubated with FITC-conjugatedgoat anti-rabbit antibodies (1:400 dilution in PBS/10% goat serum,ICN, Costa Mesa, CA) or TRITC-conjugated goat anti-mouseantibodies (1:100 dilution in PBS/10% goat serum, Zymed Labs, Inc.,South San Francisco, CA) for 2 hours at room temperature. Afterextensive washing in PBS, coverslips were mounted in Citifluor (TedPella, Inc., Redding CA). Data were collected using a 100× Zeiss plan-neofluar lens and air-cooled CCD camera (DAGE MTI, Michigan City,IN City) using Scion Image software (Scion Corporation, Fredrick,MD) to capture images. Digital images were analyzed and prepared forpublication using Photoshop (Adobe Systems, Inc., San Jose, CA).

Microcystin-Sepharose affinityTwo methods were used to prepare extracts for microcystin-Sepharoseaffinity. First a salt extract from isolated axonemes was prepared.Isolated axonemes were suspended at 8 mg/ml in Buffer Asupplemented with 0.57 M NaCl, 10 mM NaF, 100 µM ortho-vanadate,and 0.1% β-mercaptoethanol (replacing 1 mM DTT). This buffer, whencontaining 0.12 M NaCl, is referred to as Buffer B. The resultant saltextract was separated from the extracted axonemes by centrifugationand dialyzed in Buffer B. Second, in some cases, an extract from wholeflagella was prepared by a novel procedure as follows,. Isolated flagellawere suspended at 10 mg/ml in Buffer B, drop-frozen in liquid nitrogencontained in a porcelain mortar. The frozen samples were then groundinto powder with a pestle. The frozen flagellar powder was thawed anddiluted 5× with Buffer B. The whole flagellar extract was then separatedfrom insoluble material by centrifugation in a microfuge.

For microcystin affinity, 0.5 ml extract from isolated axonemes orwhole flagellar extract (~2-4 mg/ml) was either pretreated with 50 µMmicrocystin-LR (Calbiochem, La Jolla, CA) or with buffer alone fortwo hours at 4°C. The samples were then incubated with ~20 µlmicrocystin-Sepharose (15 µM, final microcystin concentration,Upstate Biotechnology, Lake Placid, NY) overnight at 4°C. Sepharosebeads were then washed extensively with buffer B and then re-suspended to 1/5 the original volume in 1× SDS-PAGE sample buffer.For microcystin-Sepharose precipitation of the EDC cross-linkedproteins, the EDC treated axonemes were extracted with 0.6 M NaCland the extract prepared for microcystin-Sepharose affinity as describedabove.

RESULTS

Isolation of axonemal protein phosphatases bymicrocystin-Sepharose affinityBased on pharmacological analysis of dynein-drivenmicrotubule sliding, we predicted isolated axonemesmust contain microcystin-sensitive protein phosphatases

(Habermacher and Sale, 1996, 1997). To directly test thishypothesis, we performed the microcystin affinity-precipitation on salt extracts from isolated axonemes. Thisapproach has been used to identify PP1, PP2A, and associatedproteins, in several cell types (Moorhead et al., 1995; Camposet al., 1996; Colbran et al., 1997; Damer et al., 1998). Threeaxonemal proteins, with apparent masses of 35, 37 and 62 kDa(lane 2, Fig. 1), specifically bound to the microcystin-Sepharose matrix compared to the control treated with excess,soluble microcystin-LR (lane 1, Fig. 1). The same threeproteins specifically bound microcystin-Sepharose usingextracts from whole flagella prepared by the freeze-thawmethod described in Materials and Methods. Based onmolecular size and microcystin affinity, we speculated theprominent 35-kDa protein is the catalytic subunit of PP1(PP1c). To test this, we cloned the Chlamydomonas PP1c, andproduced polyclonal, peptide antibodies specific to a uniquesequence for localization studies.

Molecular characterization of Chlamydomonas PP1For cloning, we took advantage of the highly conserved proteinsequence of PP1c. A pair of degenerate oligonucleotideprimers was designed from two regions of PP1c containingresidues completelyconserved in PP1c, but notfound in the catalyticsubunit of PP2A (Materialsand Methods). RT-PCRwith this primer pair(designated Ps1 and Pa1,Fig. 2) resulted in a 480-bpproduct designated thepRT1 clone (Fig. 2A).Sequence analysis indicatedthat the predicted proteinsequence of pRT1 is nearlyidentical to that of otherknown PP1cs. Thus thepRT1 clone was used as aprobe for screening thecDNA and genomiclibraries as well as forSouthern and northern blotanalysis.

Three cDNA clones andsix distinct genomic cloneswere recovered. The largest

Fig. 1. Microcystin-Sepharoseaffinity of salt extractedaxonemal proteins. Silverstained SDS-PAGE revealedthat 35, 37 and 62 kDaaxonemal proteins specificallybound to microcystin-Sepharose (arrowheads, lane2). In control experiments, inwhich the extract was pre-incubated with solublemicrocystin, these proteinsfailed to bind to microcystin-Sepharose (lane 1).

94 P. Yang and others

Fig. 2. Cloning of the Chlamydomonas PP1c gene. (A) Stepwise strategy for cloning Chlamydomonas PP1c gene and message. Key: arrowsdesignate primers used for RT-PCR; solid bar designates the exons; and the open box designates the un-translated region. (B) Genomic and thepredicted protein sequence for Chlamydomonas PP1c. Arrows designate primer positions; bold italicized amino acids designate peptides usedto design the degenerate primers; bases in capitals designate the exons; bold bases designate potential polyadenylation signal; the asterisk (*)indicates the predicted stop codon; the arrowhead designates the 5′ end of the pC1 cDNA clone; and the boxed letters designate the syntheticpeptide used to raise anti-PP1 peptide antibody. These sequence data are available from GenBank under accession number (AF156101).

95Phosphatases anchored in flagellar axonemes

cDNA clone was 1.3 kb (pC1, Fig. 2A), which contained theC-terminal half of the coding sequence and the complete 3′ un-translated region, including a polyadenylation signal (TGTAA)located 19 bp upstream of a poly(A) tail (Fig. 2B). Mappingof six genomic clones indicated that the PP1c gene is locatedwithin two SacI subclones, a 5′, 5.5-kb fragment (designatedpG1 in Fig. 2A) and a 3′, 1.9-kb fragment (designated pG2 inFig. 2A). The pC1, pG1 and pG2 clones were sequenced. Torecover the 5′ end of the cDNA, a second RT-PCR wasperformed using the primer pair Ps2 and Pa2. An 820-bpproduct, designated pRT2 (Fig. 2A), was recovered. An openreading frame of 912 bp was identified in the pRT2 and pC1clones, and the first ATG is preceded by an in-frame stopcodon-TAG 21-bp upstream. Thus, it is likely that the completecoding sequence has been recovered (Fig. 2B).

The Chlamydomonas PP1c gene consists of six exons,which, when combined, predict a 1.8 kb message. The 912 bpcoding sequence predicts a 35-kDa protein with pI 4.8.Sequence analysis predicts >90% of the amino acid residuesare identical to the PP1 from other species (Fig. 3). TheChlamydomonas PP1c is most similar to the PP1c fromAcetabularis, and among mammalian isoforms most similar toPP1γ. Greatest sequence diversity is found at the N- and Cterminus as described for PP1c from earlier studies. Asdescribed below, we took advantage of unique sequence toproduce a peptide antibody that specifically reacts withChlamydomonas PP1c. Southern blot analysis of genomicDNA, digested with either HindIII and PstI, revealed only oneband suggesting that a single gene encodes PP1c inChlamydomonas (Fig. 4A). Based on RFLP analysis using pC1

as a probe, the gene is located on the right arm of linkage groupVI (Silflow, 1998). Northern blot analysis revealed thepredicted 1.8-kb message (Fig. 4B). Moreover, the transcript ismore abundant following experimental deflagellation, acharacteristic common to the mRNA of many axonemalproteins (Lefebvre et al., 1980).

Chlamydomonas PP1c is anchored in the axonemeA rabbit polyclonal antibody was raised against a syntheticpeptide containing the unique C-terminal 11 amino acids(boxed sequence in Fig. 2B), with cysteine and tyrosine addedat the N terminus for conjugation and phosphorylationrespectively. Western blot analysis demonstrated that theantibody bound to a 35-kDa protein in isolated flagella (arrow,Fig. 5A), and, as predicted, the major fraction of the 35-kDaprotein remained bound to isolated axonemes (lane 3, Fig. 5).The 35-kDa protein is extractable with 0.6 M NaCl buffer,permitting functional analysis and purification as a solubleprotein. Thus, using microcystin-Sepharose affinity, we wereable to confirm that the 35-kDa immuno-reactive protein isPP1c (Fig. 5B).

Immuno-fluorescence localization using affinity purifiedPP1c antibodies revealed that PP1c is located along the lengthof both flagellar axonemes, as well as being abundant in thecell body (Fig. 5C). Pre-immune serum did not stain theaxonemes. Notably, compared to many other axonemalproteins including PP2A (see below) the staining of PP1cappeared punctate irrespective of mode of fixation,permeabilization or irrespective of the source of secondaryantibody. One explanation for the punctate staining of PP1c is

Fig. 3. Chlamydomonas PP1 is highly conserved. Amino acid sequence comparison of PP1c from various species was performed with theprogram PileUp using default parameters. Shaded regions indicate identical amino acid residues. The accession numbers for the listedsequences include, from top to bottom: p48481, p37139, p36874, p48727, p48461, p37140 and p08128.

96

that in the axoneme, PP1c is not easily accessible to theantibodies. Consistent with this interpretation, we have notproduced a strong signal using the PP1c antibody and electronmicroscopic immuno-gold localization with either pre- or post-embedding methods. This failure with electron microscopiclocalization is despite successful positive controls usingantibodies to other axonemal proteins. Thus, direct localizationwill require new reagents and/or approaches or discovery ofpreviously characterized axonemal proteins that interact withand anchor PP1c.

Chlamydomonas axonemal PP1 is located in thecentral pair apparatusTo further determine the location of PP1c in the axoneme,western blot analysis was performed using axonemes isolatedfrom flagellar mutants. We postulated that axonemal PP1c isanchored to a distinct structure, and that failure to assemble theanchor structure would result in failure of PP1c to co-purifywith the axoneme, despite the presence of PP1c in flagella. Totest this, we analyzed and compared PP1c fin both flagella andaxonemes from various mutants using the PP1c antibody.Among the mutants analyzed, only the axonemes from mutantsthat fail to assemble the central pair structures lack a significantfraction of the axonemal PP1c (Fig. 6). Axonemes lackingPP1c were derived from pf15, pf16, pf18, pf19 and pf20 (Fig.6A). Each of these mutants fails to assemble all or part of thecentral pair apparatus, and, in particular, in pf16 the ‘C1’microtubule and associated structures are unstable (Dutcher et

P. Yang and others

Fig. 4. (A) Southern blot analysis of genomic DNA (10 µg, derivedfrom wild type or motility mutants), digested as indicated, revealedsingle band suggesting Chlamydomonas PP1c is encoded by a singlegene. (B) Northern blot analysis of total RNA isolated from wild-type cells revealed a single 1.8 kb message for PP1c, and themessage was amplified following experimental deflagellation (toppanel). Ethidium bromide staining of the same gel illustrated sampleloading (bottom panel).

Fig. 5. Western analysis of flagellar fractions revealed PP1c isanchored in the axoneme. (A) The PP1C peptide antibodyrecognized a 35 kDa protein (arrow) in isolated flagella, the detergentextract (NP-40 ext.), isolated axonemes, the high salt extract (HSE)derived from axonemes, and the resultant extracted axonemes(Ext.ed Axo). Fractions were prepared from equal amounts offlagella. (B) The 35-kDa antigen in the extract specifically bound tomicrocystin-Sepharose (lane 2). Microcystin-Sepharose affinity wasblocked when extract was pre-incubated with soluble microcystin,insuring that the 35-kDa antigen is PP1c. The difference in migrationof PP1C is thought to be a result of covalent association pf PP1c withsoluble microcystin (MacKintosh et al., 1995).(C) Immunofluorescence analysis with the affinity-purified PP1cantibody revealed a punctate staining pattern along the length of bothflagellar axonemes and the cell body. Bar, 5 µm. Pre-immune serumfailed to stain the axonemes (not shown).

kb kb

kb

97Phosphatases anchored in flagellar axonemes

al., 1984; Smith and Lefebvre, 1996; Mitchell and Sale, 1999).In contrast, axonemes from mutants lacking the outer or innerdynein arms, radial spokes, dynein regulatory complex, or the‘mbo’ gene products all contain the wild-type amount of PP1c.The isolated flagella and axonemes from ‘mia’ mutants, whichare defective in phosphorylation of IC138 (King and Dutcher,1997), also contain wild-type levels of PP1c. The simplestinterpretation is that a major fraction of axonemal PP1c isanchored in the central pair apparatus, associated with the C1microtubule. Moreover, axonemes from pf6, which lack oneprojection from the C1 microtubule (Smith and Lefebvre,1996; Mitchell and Sale, 1999), retain the wild-type levels ofPP1c. Thus, PP1 is not likely to be anchored to that projection.

Despite the lack of PP1c in isolated axonemes from centralpair mutants, PP1c is found in the whole flagella of thesemutants at a similar concentration found in wild-type flagella(top row, Fig. 6A). This indicates that none of the central pairmutants analyzed are defective in PP1c expression, and thatPP1c is correctly targeted and transported to the flagellarcompartment. In contrast, blots probed with antibodies to twoother central pair proteins, the pf16 and pf20 gene products(Smith and Lefebvre, 1996; 1997A), revealed these proteins areabsent in both flagella and axonemes of mutants that fail toassemble the central pair complex (compare axonemes andflagella for pf18 and pf19, Fig. 6A). This experiment indicatedthat PP1c is transported to the flagella and axoneme differentlyfrom the pf16 and pf20 proteins.

To determine whether the axonemal PP1c is solely locatedin the central pair apparatus, we used larger protein loads andcompared the proteins from isolated flagella, Nonidet extract,and axonemes from wild-type and pf18 (Fig. 6B). Westernblots revealed the axonemes from pf18, which lack the centralpair structure, retain a small amount of PP1c (compareaxonemes from pf18 with wild type, Fig. 6B). The simplestinterpretation is that PP1c is anchored in more than oneaxonemal location. Of potential relevance, an outer arm dyneinlight chain with sequence homology to the PP1 binding proteinSDS22 has been reported (Patel-King et al., 1997).

PP1c forms a molecular complex with a ~60 kDaaxonemal proteinTo further confirm that PP1 is docked in the axoneme andidentify PP1 interacting proteins, western blots were used toanalyze axonemes treated with the ‘zero-length’ cross-linker,EDC, which has proven useful in identification of interactionsamong protein subunits in dynein and the radial spokes (Kinget al., 1991; Diener et al., 1993; Yang and Sale, 1998).Following exposure of isolated axonemes to 1.5 mM EDC, anew band with a mass of ~97 kDa appeared in blots probedwith the PP1c antibody (closed circle, Fig. 7A). As a furthercontrol, the cross-linked 97 kDa axonemal product wassolublized with 0.6 M NaCl, and following dialysis, wasspecifically precipitated by microcystin-Sepharose (Fig. 7B).This result confirmed the 97 kDa cross-linked product is acomplex containing axonemal PP1c tightly associated with anaxonemal protein of about 60 kDa. Furthermore, the distinctivecross-liked product produced, rather than multiple, spuriouscross-linked products or lack of product, supports the modelthat PP1c is anchored in specific positions in the axoneme. The97-kDa cross-link product did not react on western blots usingantibodies to tubulin or dynein intermediate chain subunits.

However, based on the preponderance of PP1c in the centralpair complex, it is likely the ~60 kDa protein is the central paircomponent that secures PP1c in the structure. The 97 kDacross-linked product failed to appear when EDC cross-linking

Fig. 6. The major axonemal fraction of PP1c is located in the centralpair apparatus, associated with the C1 microtubule. (A) Top pair ofpanels: western blot analyses revealed that PP1c is dramaticallyreduced in the axonemes isolated from mutants that fail to assemblecentral pair apparatus. In contrast, PP1c is present in the flagella ofall strains examined. Middle and lower pairs of panels: in controlexperiments, two other central pair proteins, the pf16 and pf20 geneproducts, were found in similar quantities in both flagella andaxonemes from the various cell types illustrated. Phenotypes: pf15,pf18 and pf19 axonemes each lack the entire central pair apparatus;pf6 axonemes lack a projection of the C1 central microtubule; pf16axonemes lack the C1 microtubule and associated projections; andpf20 axonemes predominantly lack the central pair components.(B) PP1c is not exclusively located to the central pair apparatus.PP1c distribution was compared with higher protein loads betweenflagella, axonemes and detergent extract (NP-40 Ext.) from wild typeand mutants lacking the central pair (pf18). Notably, in the absenceof the central pair, a small amount of PP1c remains in the axoneme(compare lanes 5 and 6).

98

was performed in the salt extract, following 0.6 M NaClextraction. This suggested that without cross-linking, PP1c andthe 60-kDa axonemal protein become dissociated followingexposure to the 0.6 M NaCl extraction buffer.

PP2A is also anchored to the axonemeIn addition to PP1, microcystin-Sepharose affinity purificationrevealed two other axonemal proteins with masses of 37 and62 kDa (Fig. 1). Based on molecular weight and microcystinaffinity, we postulated the proteins are the C- and A-subunitsof PP2A. To test this prediction, the microcystin bindingproteins were blotted to the PVDF membrane and the 62-kDaprotein was band purified for peptide sequencing. One of thepeptides recovered was nearly identical to the A-subunit ofPP2A (Fig. 8A), strongly suggesting the 62 kDa axonemalprotein is the A-subunit. To test whether the 37 kDa protein isthe catalytic subunit of PP2A, two different monoclonalantibodies to the C-subunit were used for western blot analysis(Orgis et al., 1999). In both cases, the antibodies to the C-subunit specifically bound to the 37-kDa, axonemal protein(Fig. 8B). The identical blot re-probed with anti-PP1 antibodyshowed that the 37-kDa protein is slightly larger than PP1 (Fig.8B and Fig. 1). Moreover, the PP1c antibody selectivelyprecipitated the 35 kDa PP1c: the 37 and 62-kDa microcystinbinding proteins were not found in such immune precipitates

(data not shown). Immuno-fluorescence microscopy using themonoclonal antibody to the C-subunit revealed PP2A is foundalong the length of both flagellar axonemes (Fig. 8C). Usingthese antibodies, we have not succeeded in localization ofPP2A with electron microscopic immuno-gold techniques.

PP2A is localized to the outer doublet microtubules To further determine the location of PP2A in the axoneme,western blot analysis was performed using the flagella andaxonemes isolated from flagellar mutants. As described abovefor axonemal PP1, we postulated that axonemal PP2A isanchored to distinct structures, and that failure to assemble theanchor structure would result in failure of PP2A to co-purifywith the axoneme. To test this, we analyzed PP2A from bothflagella and axonemes from each mutant using the monoclonalantibody to the C-subunit of PP2A. Based on western blots,isolated flagella and axonemes from mutants lacking outer or

P. Yang and others

Fig. 8. The axoneme also contains PP2A. (A) Peptide sequence fromthe 62 kDa, microcystin-binding protein (see Fig. 1). The bandpurified 62-kDa axonemal protein contains a sequence highlyhomologous to the A-subunit of PP2A (amino acids #148-157), butnot found in sequences from subunits of other phosphatases.(B) Based on western blots with two monoclonal antibodies to the C-subunit of PP2A, the 37 kDa axonemal, microcystin binding protein(see Fig. 1) is the C-subunit of PP2A (left panel). As furtherdemonstration, the PP2A 37-kDa antigen was specificallyprecipitated by microcystin-Sepharose (middle panel) and migratesmore slowly than the 35 kDa PP1c (right panel).(C) Immunofluorescence analysis using anti-PP2Ac revealed an evendistribution of PP2A along the length of both flagellar axonemes andbright staining of the cell body. Bar, 5 µm.

Fig. 7. Western blot analyses of EDC-treated axonemes revealed thatPP1c is tightly associated with a ~ 60 kDa axonemal protein.(A) Axoneme treated with the EDC of concentrations indicated. Notethe occurrence of a 97-kDa band (dot) in addition to PP1c (arrow)after exposure to EDC. (B) The 97-kDa cross-linked complex, alongwith the 35 kDa PP1c, were salt extractable and precipitated bymicrocystin-Sepharose, confirming the presence of PP1 in the cross-linked complex.

kDa

99Phosphatases anchored in flagellar axonemes

inner arm dyneins, radial spokes, and the dynein regulatorycomplex all contained wild-type levels of PP2A. In contrast toPP1, PP2A is found in both the flagella and axonemes ofmutants that fail to assemble the central pair apparatus. This isbest illustrated in Fig. 9A, comparing the staining of the PP1to the staining of PP2A for flagella, Nonidet extract, andaxonemes from pf19 and wild-type cells. Notably, PP2A isdistributed in an equivalent manner in wild type and pf19,indicating the axonemal PP2A is not located in the central pairapparatus. The isolated flagella and axonemes from ‘mia’mutants, which are defective in phosphorylation of IC138(King and Dutcher, 1997), also contain wild-type levels ofPP2A. Among all the mutants analyzed, only flagella andaxonemes from pf15 contained reduced amount of PP2A (Fig.9B). Although pf15 lacks the central pair, it is not likely that

the axonemal PP2A is located in the central pair structure sinceother central pair mutants, including pf6, pf16, pf18, pf19, andpf20, all contain wild-type concentration of PP2A (Fig. 9B).Moreover, other studies have shown that lack of the central paircomplex in pf15 is likely to be a secondary consequence of thedefect in pf15 (Smith and Lefebvre, 1998). Since the mutantslacking dyneins, central pair, radial spoke, dynein regulatorycomplex all contain a similar concentration of PP2A, weconclude that PP2A is anchored to the nine outer microtubuledoublets.

DISCUSSION

We determined that two protein phosphatases, PP1 and PP2A,are transported to the flagellar compartment and anchored inthe axoneme. This conclusion is founded on microcystin-affinity of extracts from flagella and isolated axonemes. Theconclusion is also based on western blot analysis of flagellarfractions including the membrane-matrix and isolatedaxonemes and based on light microscopic immuno-localization. Localization of the phosphatases to the axonemeis consistent with previously published pharmacologicalanalysis revealing PP1 and possibly PP2A are anchored to themicrotubules of the axoneme in positions to control thephosphorylation and activity of inner arm dynein(Habermacher and Sale, 1996, 1997). The results are alsoconsistent with immuno-localization analysis of PP1 in cilia(Momayezi et al., 1996). Not surprisingly, additional fractionsof flagellar PP1c and PP2A are found in the detergent soluble,membrane-matrix fraction where they may play additionalroles.

The axonemal fraction of PP1c is predominantly localizedto the central pair apparatus and is associated with the C1microtubule. This conclusion is founded on analysis of bothflagella and axonemes from mutant cells that fail to assemblespecific axonemal structures (Fig. 6). In particular, localizationof PP1c to the C1 microtubule of the central pair apparatus isbased on analysis of axonemes from pf16, which fails toassemble the C1 central pair structure (Dutcher et al., 1984;Smith and Lefebvre, 1996, 1997b; Mitchell and Sale, 1999).Importantly, in each central pair mutant examined the isolatedflagella retain the wild-type concentration of PP1c despite thelack of PP1c in the isolated axonemes. The simplestinterpretation is that the majority of axonemal PP1c isanchored in the central pair apparatus, and failure to assemblethe key anchor structures results in failure of PP1c to sedimentwith the axoneme. The significance of the central pair toaxonemal function is discussed below.

The axonemal fraction of PP2A is localized to the outerdoublet microtubules. This conclusion is also founded onwestern blot analysis of both flagella and axonemes frommutant cells that fail to assemble specific structures. Based onthis analysis, PP2A is found at identical concentration inaxonemes from wild-type cells and cells lacking the centralpair apparatus, radial spokes, outer or inner arm dyneins, or thedynein regulatory complex. By elimination, PP2A must beassociated with the outer doublet microtubules. Thus, onehypothesis is that PP2A is anchored in position to directlycontrol phosphorylation of IC138 in inner arm dynein I1 (seeHabermacher and Sale, 1997). The minor fraction of axonemal

Fig. 9. Axonemal PP2A is not located in the central pair apparatus.(A) Comparison of the distribution of PP2A in isolated flagella,detergent extract and axonemes from central pair mutant pf19 andwild-type cells. PP2A is distributed in an identical manner in pf19and wild-type fractions. In contrast to PP1, PP2A remains anchoredin the axoneme despite lack of the central pair structure. (B) PP2A isdramatically reduced in the pf15 flagella and axonemes.

100

PP1c, not located in the central pair, may also be anchored inposition to control dynein activity, but this may be unlikelysince only a very small amount of axonemal PP1c is foundoutside of the central pair. To test these models, it will benecessary to develop new reagents or approaches to directlylocalize PP1 and PP2A in the axoneme. It will also benecessary to determine the relative stoichiometry anddetermine the anchor proteins for the axonemal fractions ofPP1 and PP2A.

Diverse evidence has revealed that regulation of dynein isimposed in part by activities of the central pair apparatus, theradial spokes and dynein regulatory complex (Huang et al.,1982; Smith and Sale, 1992; Piperno et al., 1992, 1994; Porteret al., 1992, 1994; Yoshimura and Shingyoji, 1999). Themechanism is not understood, but at least one of theseinteractions involves phosphorylation of inner dynein arms(Howard et al., 1994; Habermacher and Sale, 1996, 1997; Kingand Dutcher, 1997). The mechanism appears to involve controlof flagellar waveform (Brokaw et al., 1982; Hosokawa andMiki-Noumura, 1987; Brokaw and Kamiya, 1987; King andDutcher, 1997), however, other axonemal components are alsoinvolved in this control (Frey et al., 1997; Wakabayashi et al.,1997). In other examples, the interactions involve outer armdynein (Huang et al., 1982; Porter et al., 1994) and control ofbeat frequency (Mitchell and Sale, 1999). The challenge ineach case is to determine how signals in the central pairapparatus become transmitted through the radial spokes anddynein regulatory complex, to alter dynein activity.

Based on localization of PP1c in the central pair apparatus,the central pair apparatus may operate, in part, throughchanges in phosphorylation of central pair proteins,consequently leading to change in the structural interactionwith the radial spokes. As described above, this model isconsistent with genetic studies that revealed the centralpair/radial spoke system controls dynein activity. The modelis also consistent with structural studies showing a transientinteraction between the spokes and central pair projections(Warner and Satir, 1974; Goodenough and Heuser, 1985), andwith discovery that in many cases the central pair rotatesrelative to the radial spokes (Omoto et al., 1999). In vivoanalysis has revealed that at least eight of twenty-three centralpair proteins and at least five of seventeen radial spokeproteins are phosphoproteins (Piperno et al., 1981; Huang etal., 1981; Adams et al., 1981). Therefore, predictably, thecentral pair complex and radial spokes contain protein kinases.Mammalian sperm tail axonemes bear a unique catalyticsubunit for PKA with structure that may play a role inanchoring PKA and restricting mobility (San Agustin et al.,1998). In preliminary studies, AKAPs (A-kinase anchorproteins, Dell’Acqua and Scott, 1997) have been localized tothe central pair apparatus and in the radial spoke (Roush andSale, 1998). Thus, as predicted, the central pair/radial spokeapparatus may anchor both kinases and phosphatases. Theseenzymes are likely to control flagellar dynein activity and/orthe mechanism that drives the rotation of the central paircomplex in cilia and flagella of several organisms (see Omotoet al., 1999). Further tests of such models will require novel,direct biophysical measurement of structural changesassociated with changes in phosphorylation of keycomponents. Tests will also require continued definition ofphysical organization of each protein in the spoke and the

central pair apparatus, including understanding the proteinsthat anchor the kinases and phosphatases in the structure.

The axoneme offers a new opportunity to definemechanisms that anchor protein phosphatases in the cell.Protein phosphatases 1 and 2A dephosphorylate serine andthreonine residues on proteins that control a wide range ofcellular functions (Cohen, 1989; Wera and Hemmings, 1995).Thus, the question is how diverse substrates, within a cell, areregulated independent of one another. In many cases substratespecificity of PP1c appears to be conferred by localization ofthe catalytic subunit through regulatory or targeting subunits(Hubbard and Cohen, 1993; Brautigan, 1997; Campos et al.,1996; Egloff et al., 1997; Colbran et al., 1997; Damer et al.,1998; Liao et al., 1998). However, it is likely that we are justbeginning to identify such anchor mechanisms.Chlamydomonas mutants that fail to properly anchor PP1cwill contribute new members of the class of anchor proteins.One candidate is the 60-kDa axonemal protein identified bychemical cross-linking (Fig. 6). Evidence is accumulating thata B-type subunit acts, in part, to target PP2A in position tointeract with specific substrates (e.g. Mumby and Walter,1993; Kamibayashi et al., 1994; Sontag et al., 1995;Zolnierowicz et al., 1996; Csortos et al., 1996; Zhao et al.,1997; Seeling et al., 1999). We have not yet identified the B-type subunit (s) for the axonemal PP2A. We postulate that theaxonemal B-subunit is not extracted in 0.6 M salt, a conditionthat solubilizes the A- and C-PP2A heterodimer. We thereforepredict that identification of the axonemal B-subunit will helpdefine the location of PP2A in the axoneme, define the relativestoichiometry of PP2A and test our hypothesis that PP2A isanchored in position to control inner arm dynein.

We are grateful to Drs D. Pallas, E. Smith, P. Lefebvre, C.Wilkerson, and G. Witman for antibodies and libraries, S. Dutcher formutant cells and C. Silflow for RFLP mapping of PP1c inChlamydomonas. We also gratefully acknowledge helpful discussionwith Drs D. Pallas, L. Quarmby, and K. Saxe and R. Finst and A.Roush. The work was supported in part by Research Grant # FY97-0407 from the March of Dimes Birth Defects Foundation and fromthe NIH (GM 51173 and GM 17666). R.J.S. is an EstablishedInvestigator of the American Heart Association and is supported byNIH grant NS37508.

REFERENCES

Adams, G. M. W., Huang, B., Piperno, G. and Luck, D. J. L. (1981).Central-pair microtubule complex of Chlamydomonas flagella: polypeptidecomposition as revealed by analysis of mutants. J. Cell Biol. 91, 69-76.

Brautigan, D. (1997). Phosphatases as partners in signaling networks. Advan.Sec. Mess. Phospho-Protein Res. 31, 113-124.

Brokaw, C. J., Luck, D. J. L., and Huang, B. (1982). Analysis of themovement of Chlamydomonas flagella: The function of the radial spokesystem is revealed by comparison of wild type and mutant flagella. J. CellBiol. 92, 722-732.

Brokaw, C. J. (1987). Regulation of sperm flagellar motility by calcium andcAMP-dependent phosphorylation. J. Cell. Biochem. 35, 175-184.

Brokaw, C. J. and Kamiya, R. (1987). Bending patterns on Chlamydomonasflagella: IV. Mutants with defects in inner and outer arms indicatedifferences in dynein arm function. Cell Motil. Cytoskel. 8, 68-75.

Brokaw, C. J. (1994). Control of flagellar bending: a new agenda based ondynein diversity. Cell Motil. Cytoskel. 28, 199-204.

Campos, M., Fadden, B., Alms, G., Qian, Z. and Haystead, T. (1996).Identification of PP1 binding proteins by microcystin-biotin affinitychromatography. J. Biol. Chem. 271, 28478-28484.

Chaudhry, P., Creagh, S., Yu, N. and Brokaw, C. J. (1995). Multiple protein

P. Yang and others

101Phosphatases anchored in flagellar axonemes

kinase activities required for activation of sperm flagellar motility. CellMotil. Cytoskel. 32, 65-79.

Cohen, P. (1989). The structure and regulation of protein phosphatases. Annu.Rev. Biochem. 58, 453-508.

Colbran, R., Bass, M., Blair, R., Bollen, M., Zhao, S., Wadzinski, B. andStrack, S. (1997). Association of brain PP1 with cytoskeletaltargeting/regulatory subunits. J. Neurochemistry 69, 920-929.

Csortos, C., Zolnierowicz, S., Bako, E., Durbin, S. and DePaoli-Roach, A.(1996). High complexity in the expression of the B subunit of PP2A. J. Biol.Chem. 271, 2578-2588.

Damer, C., Partridge, J., Pearson, W. and Haystead, T. A. J. (1998). Rapididentification of PP1-binding proteins by mixed peptide sequencing and database searching. J. Biol. Chem. 273, 24396-24405.

Diener, D., Ang, L. and Rosenbaum, J. (1993). Assembly of flagellar radialspoke proteins in Chlamydomonas: identification of axonemal bindingdomains in RSP3. J. Cell Biol. 123, 183-190.

Dell’Acqua, P. and Scott, J. (1997). Protein kinase A anchoring. J. Biol.Chem. 272, 12881-12884.

Dutcher, S. K., Huang, B. and Luck, D. J. L. (1984). Genetic dissection ofthe central pair microtubules of the flagella of Chlamydomonas reinhardtii.J. Cell Biol. 98, 229-236.

Egloff, M., Johnson, D., Moorhead, G., Cohen, P. T., Cohen, P. andBarford, D. (1997). Structural basis for the recognition of regulatorysubunits by the catalytic subunit of PP1. EMBO J. 16, 1876-1887.

Frey, E., Brokaw, C. J. and Omoto, C. K. (1997). Reactivation at low ATPdistinguishes among classes of paralyzed flagella mutants. Cell Motil.Cytoskel. 38, 91-99.

Gardner, L. C., O’Toole, E., Perrone, C. A., Giddings, T. and Porter, M.E. (1994). Components of a ‘Dynein Regulatory Complex’ are located atthe junction between the radial spokes and the dynein arms inChlamydomonas flagella. J. Cell Biol. 127, 1311-1325.

Goodenough, U. W. and Heuser, J. E. (1985). Substructure of inner dyneinarms, radial spokes, and the central pair/projection complex. J. Cell Biol.100, 2008-2018.

Habermacher, G., and Sale, W. S. (1996). Regulation of flagellar dynein byan axonemal type 1 phosphatase in Chlamydomonas. J. Cell Sci. 109, 1899-1807.

Habermacher, G. and Sale, W. S. (1997). Regulation of flagellar dynein byphosphorylation of a 138-kD inner arm dynein intermediate chain. J. CellBiol. 136, 167-176.

Hamasaki, T., Murtaugh, T., Satir, B. and Satir, P. (1989). In vitrophosphorylation of Paramecium axonemes and permeabilized cells. CellMotil. Cytoskel. 12, 1-11.

Hasagawa, E., Hayashi, H., Asukura, S. and Kamiya, R. (1987).Stimulation of in vitro motility of Chlamydomonas axonemes by inhibitionof cAMP dependent phosphorylation. Cell Motil. Cytoskel. 8, 302-311.

Hosokawa, Y. and Miki-Noumura, T. (1987). Bending motion ofChlamydomonas axonemes after extrusion of the central pair microtubules.J. Cell Biol. 105, 1297-1301.

Howard, D., Habermacher, G., Glass, D., Smith, E. F. and Sale, W. S.(1994). Regulation of Chlamydomonas flagellar dynein by an axonemalprotein kinase. J. Cell Biol. 127, 1683-1692.

Hubbard, M. J. and Cohen, P. (1993). On target with a new mechanism forregulation of protein phosphatases. Trends Biochem. Sci. 18, 172-177

Huang, B., Piperno, G., Ramanis, Z. and Luck, D. J. L. (1981). Radialspokes of Chlamydomonas flagella: genetic analysis of assembly andfunction. J. Cell Biol. 88, 80-88.

Huang, B., Ramanis, Z. and Luck, D. J. L. (1982). Suppressor mutations inChlamydomonas reveal a regulatory mechanism for flagellar function. Cell28, 115-124.

Johnson, K. and Rosenbaum, J. (1992). Polarity of flagellar assembly inChlamydomonas. J. Cell Biol. 119, 1605-1611.

Kamibayashi, C., Yang, S., Craft, S. and Mumby, M. (1994). Comparisonof heterotrimeric PP2A containing different B subunits. J. Biol. Chem. 269,20139-20148.

Kamiya, R. (1995). Exploring the function of inner and outer dynein armswith Chlamydomonas mutants. Cell Motil. Cytoskel. 32, 98-102.

King, S., Wilkerson, C. and Witman, G. B. (1991). The Mr 78,000intermediate chain of Chlamydomonas outer arm dynein interacts with α-tubulin in situ. J. Biol. Chem. 266, 8401-8407.

King, S. and Dutcher, S. (1997). Phosphorylation of an inner dynein armcomplex in Chlamydomonas reinhardtii is altered in phototactic mutantstrains. J. Cell Biol. 136, 177-191.

Lefebvre, P., Silflow, C., Weiber, E. and Rosenbaum, J. (1980). Increased

levels of mRNA for tubulin and other flagellar proteins after amputation orshortening of Chlamydomonas flagella. Cell 20, 469-470.

Liao, H., Li, Y., Brautigan, D. and Gundersen, G. (1998). PP1 is targetedto microtubules by the Map tau. J. Biol. Chem. 273, 21901-21908.

MacKintosh, R., Dalby, K., Campbell, D., Cohen, P. T. W., Cohen, P. andMacKintosh, C. (1995). The cyanobacterial toxin microcystin bindscovalently to cysteine-273 on protein phosphatase1. FEBS Lett. 371, 236-240.

Mitchell, D. and Sale, W. S. (1999). Characterization of a Chlamydomonasinsertional mutant that disrupts flagellar central pair microtubule-associatedstructures. J. Cell Biol. 144, 293-304.

Momayezi, M., Wloga, D., Kissmehl, K., Plattner, H., Jung, G., Klumpp,S. and Schultz, J. (1996). Immunolocalization of protein phosphatase type1 in Paramecium cells using antibodies against recombinant protein andpeptides. J. Histochem. Cytochem. 44, 891-905.

Moorhead, G., MacKintosh, C., Morrice, N. and Cohen, P. (1995).Purification of the glycogen associated form of PP1 by microcystin –sepharose affinity chromatography. FEBS Lett. 362, 101-105.

Mumby, M. and Walter, G. (1993). Protein serine/threonine phosphatases:structure, regulation, and functions in cell growth. Physiol. Rev. 73, 673-699.

Orgis, E., Du, X., Nelson, K., Mak, E., Yu, X., Lane, W. and Pallas, D.(1999). A novel phosphatase methylesterase is one of several novel proteinsstably associating with two inactive mutants of PP2A. J. Biol. Chem. 274,14382-14391.

Omoto, C., Gibbons, I. R., Kamiya, R., Shingyoji, C., Takahashi, K. andWitman, G. B. (1999). Rotation of the central pair microtubules ineukaryotic flagella. Mol. Biol. Cell 10, 1-4.

Patel-King, R., Benashski, S. and King, S. (1997). A leucine-rich repeatprotein related to the SDS22 modulator of protein type-1 phosphatase fromthe Chlamydomonas outer dynein arm. Mol. Biol. Cell 8, 163a.

Piperno, G., Huang, B. and Luck, D. J. L. (1981). Radial spokes ofChlamydomonas flagella: composition and phosphorylation of stalkcomponents. J. Cell Biol. 88, 73-79.

Piperno, G., Mead, K. and Shestak, W. (1992). The inner dynein arms I2interact with a ‘dynein regulatory complex’ in Chlamydomonas flagella. J.Cell Biol. 118, 1455-1464.

Piperno, G., Mead, K., LeDizet, M. and Moscatelli, A. (1994). Mutationsin the “Dynein regulatory complex” alter the ATP-insensitive binding sitesfor inner arm dyneins in Chlamydomonas axonemes. J. Cell Biol. 125, 1109-1117.

Porter, M. E., Power, J. and Dutcher, S. K. (1992). Extragenic suppressorsof paralyzed flagellar mutations in Chlamydomonas reinhardtii identify locithat alter the inner dynein arms. J. Cell Biol. 118, 1163-1176.

Porter, M. E., Knott, J., Gardner, L., Mitchell, D. and Dutcher, S. (1994).Mutations in the SUP-PF-1 locus of Chlamydomonas reinhardtii identify aregulatory domain in the beta dynein heavy chain. J. Cell Biol. 126, 1495-1507.

Roush, A. M. and Sale, W. S. (1998). A-kinase anchor proteins (AKAPs) arelocalized to the central pair and radial spokes in flagellar axonemes. Mol.Biol. Cell 9, 397a.

Rupp, G., O’Toole, E., Gardner, L., Mitchell, B. and Porter, M. (1996).The sup-pf-2 mutations of Chlamydomonas alter the activity of the outerdynein arms by modification of the gamma heavy chain. J. Cell Biol. 135,1853-1865.

San Agustin, J. T. and Witman, G. B. (1994). Role of cAMP in thereactivation of ram sperm. Cell Motil. Cytoskel. 27, 206-218.

San Agustin, J. T., Leszyk, J., Nuwaysir, L. and Witman, G. B. (1998). Thecatalytic subunit of the cAMP dependent protein kinase of ovine spermflagella has a unique amino terminal sequence. J. Biol. Chem. 273, 24874-24993.

Satir, P. (1985). Switching mechanisms in the control of ciliary motility. InModern Cell Biology (ed. P. Satir), pp. 1-46. Alan R. Liss, Inc., NY.

Satir, P., Barkalow, K. and Hamasaki, T. (1993). The control of ciliary beatfrequency. Trends Cell Biol. 3, 409-412.

Seeling, J., Miller, J., Gil, R., Moon, R., White, R. and Virshup, D. (1999).Regulation of B-catenin signaling by the B56 subunit of PP2A. Science 283,2089-2091.

Silflow, C. (1998). Organization of the nuclear genome. In The MolecularBiology of Chloroplasts and Mitochondria in Chlamydomonas. (ed. J.Rochais, M. Goldschmidt-Clemont, and S. Merchant), pp. 25-40. KluwerAcademic Publications. The Netherlands.

Smith, E. F. and Sale, W. S. (1992). Regulation of dynein-driven microtubulesliding by the radial spokes in flagella. Science 257, 1557-1559.

102

Smith, E. F. and Lefebvre, P. A. (1996). PF16 encodes a protein witharmadillo repeats and localizes to a single microtubule of the centralapparatus in Chlamydomonas flagella. J. Cell Biol. 132, 359-370.

Smith, E. F. and Lefebvre, P. A. (1997a). The role of central apparatuscomponents in flagellar motility and microtubule assembly. Cell Motil.Cytoskel. 38, 1-8.

Smith, E. F. and Lefebvre, P. A. (1997b). PF20 Gene Product contains WDrepeats and localizes to the intermicrotubule bridges in Chlamydomonasflagella. Mol. Biol. Cell 8, 455-467.

Smith, E. F. and Lefebvre, P. A. (1998). Pf15 is required for centralmicrotubule assembly and is homologous to katanin P80. Mol. Biol. Cell 9,278a.

Sontag, E., Nunbhakdi-Craig, V., Bloom, G. and Mumby, M. (1995). Anovel pool of PP2A is associated with microtubules and is regulated duringthe cell cycle. J. Cell Biol. 128, 1131-1144.

Song, Q., Khanna, K., Lu, H. and Lavin, M. (1993). Cloning andcharacterization of a human protein phosphatase-1 cDNA. Gene 129, 291-295.

Tam, L.-W. and Lefebvre, P. A. (1993). Cloning of flagellar genes inChlamydomonas reinhardtii by DNA insertional mutagenesis. Genetics 135,375-384.

Tash, J. S. and Bracho, G. (1994). Regulation of sperm motility: Emergingevidence for a major role for protein phosphatases. J. Androl. 15, 505-509.

Wakabayashi, K., Yagi, T. and Kamiya, R. (1997). Ca2+-dependentwaveform conversion in the flagellar axoneme of Chlamydomonas mutantslacking the central-pair radial spoke system. Cell Motil. Cytoskel. 38, 22-28.

Walczak, C. E., and Nelson, D. L. (1994). Regulation of dynein-drivenmotility in cilia and flagella. Cell Motil. Cytoskel. 27, 101-107.

Warner, F. D. and Satir, P. (1974). The structural basis of ciliary bendformation. Radial spoke positional changes accompanying microtubulesliding. J. Cell Biol. 63, 35-63.

Wera, S. and Hemmings, B. (1995). Serine/threonine protein phosphatases.Biochem J. 311, 17-29.

Wilkerson., C., King, S., Pazour, G. and Witman, G. B. (1995). The 78 kDaintermediate chain of Chlamydomonas outer arm dynein is a WD repeatprotein required for arm assembly. J. Cell Biol. 129, 169-178.

Witman, G. B. (1986). Isolation of Chlamydomonas flagella and flagellaraxonemes. Meth. Enzymol. 134, 280-290.

Witman, G. B. Plummer, J., and Sander, G. (1978). Chlamydomonasflagellar mutants lacking radial spokes and central tubules. J. Cell Biol. 76,729-747.

Yang, P. and Sale, W. S. (1998). The 140, 000 Mr intermediate chain ofChlamydomonas inner arm dynein is a WD-repeat protein implicated indynein anchoring. Mol. Biol. Cell 9, 3335-3349.

Yoshimura, M. and Shingyoji, C. (1999). Effects of the central pair apparatuson microtubule sliding velocity in sea urchin sperm flagella. Cell Struct.Funct. 24, 43-54.

Zhao, Y., Bogulawski, G., Zitomer, R. and DePaoli-Roach, A. (1997). S.cerevisiae homologs of mammalian B and B′ subunits of PP2A directdistinct functions. J. Biol. Chem. 272, 8256-8262.

Zolnierowicz, S., Van Hoof, C., Stevens, P., Stevens, I., Merlevede, W.,Goris, J. and Hemmings, B. (1996). The variable subunit associated withphosphatase 2A defines a novel multimember family of regulatory subunits.Biochem. J. 317, 187-194.

P. Yang and others