Flexible-substratum technique for viewing cells from the side: some in

vivo properties of primary (9+0) cilia in cultured kidney epithelia

KAREN E. ROTH1-2, CONLY L. RIEDER1'3'* and SAMUEL S. BOWSER1

'Wadsu-arlh Center for Laboratories and Research, Umpire Stale Plaza, IJO Box 509, Albanv, AT 12201, USA^Department of Biological Sciences, State University ofXeic )'ork, Albany, AT 12222, USA^School of Public Health Sciences, State University of Sen- York, Albany, AT 12222, USA

•Author for correspondence

Summary

Cells cultured on thin plastic (e.g. Formvar,Teflon, polycarbonate) membranes can be clearlyimaged from the side in vivo by video mi-croscopy. We have used this flexible-substratumtechnique to examine the behaviour and proper-ties of primary cilia in confluent cultures of thekidney epithelial cell lines PtK,, PtK2, LLC-PKj,MDCK and BSC-40. In these cells primary ciliaappear as rigid rods, up to 55 fim long, -whichproject at various angles from the dorsal cellsurface. The length distribution of primary ciliain confluent cultures is a distinct characteristic ofeach established kidney cell line examined, withLLC-PK, exhibiting three distinct length popu-lations. Primary cilia of kidney cell lines bend

passively in response to flow but do not displaypropagated bending or vortical motions. Up to26 % of the cilia in the cell types examined possessone or more conspicuous swellings along theciliary shaft. Treatment with 0*05% trypsin,which is sufficient to cause cell rounding, does notinduce the resorption or shedding of the cilium.These direct observations demonstrate that kid-ney epithelial-cell primary cilia are non-motileand longer than previously thought, and suggestthat their length represents a phenotypic markerfor each cell line.

Key words: primary cilium, kidney epithelium, flexiblesubstratum, cell profiles, video microscopy.

Introduction

Side views of cells are often useful in studies of cellshape and motility, especially in those cases wheremovements away from or towards the substratum arebeing investigated. For example, Dellinger (1906) usedside views to examine the extension of pseudopods inamoebae. More recently, Ingram (1969) and Harris(1969) studied the ruffling membrane in moving fibro-blasts by viewing single cells from the side. Theseruffles often extended upwards perpendicular to thesubstratum (i.e. parallel to the optical axis), makingthem difficult to study when viewed from above.

Several methods have been developed to obtain sideviews of cells. Most of them involve growing cells on anarrow horizontal substratum (e.g. glass edges, glassfibres, plasma clots, gelatin strips), which is thenturned vertically for viewing (Dellinger, 1906; Ingram,1969; A. Harris, 1969; J. K. Harris, 1978; Hlinka &

Journal of Cell Science 89, 457-466 (1988)Printed in Great Britain © T h e Company of Biologists Limited 1988

Sanders, 1972; Sanders & Prasad, 1979). Alternatively,optical components can be introduced into culturechambers to permit the observation of specimens inboth top and side views (Boocock el al. 1985). Ingeneral all of these approaches can, under the appropri-ate conditions, provide good views of single cells at restor moving across a substratum.

Primary cilia, which extend from the dorsal cellsurface, are difficult to observe when viewed fromabove. As a result little is known concerning theirbehaviour and properties in vivo. We reasoned that wecould generate high-resolution side views of primarycilia in living cells by some of the above methods.However, these approaches were found to be of limitedvalue, since the incidence of primary cilia within a cellculture is positively correlated with the degree ofconfluency (Archer & Wheatley, 1971; Jensen et al.1979; Mori et al. 1979; Tuckers al. 1979) and none ofthese methods generates or permits the examination of

457

confluent cell monolayers. To overcome this limitationwe have grown cells to confluency on flexible plasticsubstrata, which are then folded to form an edge. Byexamining this folded edge we can routinely image,with good resolution, large numbers of cell profiles andprimary cilia. The details of this technique, along withour initial observations on primary cilia in living kidneyepithelial cell lines derived from various vertebrates,are described in this paper.

Materials and methods

Cell cultureStock cultures of PtK|, PtK2and LLC-PK, cells were grownat 37°C in T-30 flasks in L-1S medium buffered to pH7-2 with 10 mM-A'-2-hydroxyethyl-piperazme-A"-2-ethane-sulphonic acid (Hepes) and fortified with 10% foetal calfserum and antibiotics (Jensen el al. 1979; Riedere/ al. 1979).Stock cultures of MDCK and BSC-40 cells were similarlymaintained in bicarbonate-buffered Dulbecco's MEM forti-fied with 5% foetal calf serum and antibiotics.

Flexible growth substrataDuring the course of this investigation we grew cells on anumber of flexible plastic substrata including Aclar 22A(Allied Chemical, Morristown, NJ), polyvinylidene chloride(Saran wrap; Dow Chemical, Indianapolis, IN), Butvar(Monsanto, Springfield, MA), Formvar (E. Fullam, Schen-cctady, NY), porous polycarbonate (PVP-free, tissue-cul-turc-trcatcd; Nucleopore, Pleasanton, CA), and Teflon FEP(Dupont, Willmington, DE). Cells readily grew on all ofthese substrata except Teflon, which needed to be first coatedwith the bio-adhesion polymer Cell-Tak (Biopolymers Inc.,Farmington, CT).

We found Formvar and porous polycarbonate to be theeasiest to use on a routine basis. The data presented in thisstudy were obtained from cells grown on Formvar by thefollowing procedure: clean glass microscope slides werecoated with a 0-5(±0-l),um thick film (as determined byinterference microscopy) by dipping them into a 2% solutionof Formvar in ethylene dichloride. The slides were thensterilized by ultraviolet irradiation. Cells were removed froma stock flask by trypsin treatment, seeded at medium densityonto the Formvar-coated slides, and grown at 37°C in ahumid chamber.

Once the cell cultures on a coated slide had reached thedesired stage of confluency, a rectangular section of Formvarwas outlined by cutting it with a scalpel. While still sub-merged in medium, the outlined Formvar film was strippedfrom the slide using fine forceps. The long edge of thisrectangular strip was folded over on itself, cell side out. Athin (5mm) piece of Teflon tape (type TV350; CHRIndustries, New Haven, CT), the length of the slide, wasplaced just behind and parallel to one of the long edges of asterile 24 mm X 60 mm coverslip. This coverslip was sub-merged in medium adjacent to the folded cell-containingFormvar strip. The Formvar strip was teased onto this slideso that its long edge was behind and parallel to the edge of thetape (see Fig. 1). It was anchored in this position by placing a

10 mm X 60 mm coverslip behind the edge containing thecells. After these manipulations, the preparation wasremoved from the medium and covered with a 24 mm X40 mm coverslip. The cell chamber was filled with mediumand sealed with VALAP. The preparations were maintained,during all observations, at 37°C with an air-curtain incubator(Arenberg Sage Inc., Boston, MA). When prepared in thismanner at least 92% of the cells were viable after 8h, asjudged by Trypan Blue exclusion (Absoloin, 1986).

Light microscopy

Cell preparations were examined with either a Nikon Opti-phot microscope equipped wtih Hoffman modulation-con-trast and phase-constrast objectives or a Zeiss Photomicro-scope II equipped with Nomarski differential interference-contrast (DIC) objectives. Video recordings were made on aPanasonic 8050 l/2-inch time-lapse video recorder using aDage MIT-65 camera, which was coupled to a NipponAvionics digital image processor and a time/date generator.Single frames were photographed from the video screen usingKodak Plus-X film with a 35 mm camera mounted in a fixedposition. Modulation- and phase-contrast photomicrographswere also taken directly from the microscope, using KodakTechnical Pan film, which was push-processed to ASA 400with Perfection XR1 developer (Perfection PhotographicProducts, Sun Valley, CA). For differential interference-contrast photomicrographs, Kodak technical pan film wasexposed at ASA 80 developed with D-19 developer.

For fluorescence microscopy, cell cultures were grown onFormvar and processed for the indirect immunofluorescentlocalization of microtubules by standard procedures (Rieder& Bowser, 1986; Rupp et al. 1985) using a monoclonal /5-tubulin antibody (Tu27B) kindly supplied by Dr L. I.Binder (University of Alabama, Birmingham). Control prep-arations lacking the primary antibody showed no fluor-escence. Epifluorescence photomicrographs were recordedon Kodak Tri-X film, which was push-processed to ASA 3200with Perfection XR1 developer.

MicropeifusionPreparations were perfused with media or experimentalsolutions by leaving the ends of the chambers unsealed. Aliberal volume of solution was placed at one side of thechamber and pulled through it by removing the solution fromthe other side with absorbent paper (e.g. see Cande, 1982).For trypsin treatment a 0"05 % solution of trypsin in Ca -,Mg +-free saline was stored frozen and thawed just beforeuse.

StatisticsCells were grown to confluency on Formvar-coated slides andprepared for side viewing as described above. Lengthmeasurements were made, to the nearest micrometre, di-rectly from the video screen, which was calibrated to within±0-5 jt/m using a stage micrometer. Each cilium wasmeasured from the dorsal cell surface to the tip of the cilium.For PtK| cells, 100 cilia were measured from each of twopreparations, for a total of 200 measurements. For theremaining cell types, 75 cilia were measured from each ofthree preparations, for a total of 225 measurements for eachcell type. In addition to length, all cilia were scored for the

458 K. E. Roth

FOLDED FORMVAR WITHCELL MONOLAYER

UPPER 24 x 40 mmCOVERSLIP

TAPE ATTACHINGLOWER 24 x 60 mmCOVERSLIP TOSTEEL COVERSLIPHOLDER

GLASS COVERSLIP ORTEFLON TAPE SPACER

FOLDED FORMVARWITH CELLS

TEFLON TAPE ORCOVERSLIP SPACER

UPPER 24 x 40 mmCOVERSLIP

STEEL COVERSLIPHOLDER

VALAP

COVERSLIP SPACER

LOWER 24 x 60 mmCOVERSLIP

Fig. 1. Schematic diagram illustrating theflexible-substratum method for viewingcells from the side. A. Top view; 13, side

presence or absence of swellings along their shaft. A statisticalanalysis was conducted to determine whether the lengthdistributions of cilia were significantly different between celllines. Length histograms indicated that the size distributionsof cilia in each cell line were skewed to the right, and anormalizing transform was therefore sought. Several trans-formations were attempted and the square root was finallyaccepted as giving an adequate approximation to a Gaussiandistribution.

The relationship between cilium length and the occurrenceof a swelling along the shaft in PtK| cells was analysed using alogistic regression model (Cox, 1970).

Results

Primary cilia viewed from the sideWhen monolayers of kidney epithelial cells (PtK|,PtK2, MDCK, BSC-40, LLC-PK,) were viewedthrough-focus from above, primary cilia were barelydetectable as dots above the dorsal cell surface(Fig. 2A). However, when these cells were viewedfrom the side using our flexible-substratum techniquethe entire primary cilium was clearly visible by differ-ential interference contrast microscopy (Fig. 2B-F) orHoffman modulation-contrast (Fig. 2G-I) optics. Pri-mary cilia were also visible with phase-contrast optics(Fig. 2J) but the image was less than optimum, due tothe strong phase halos associated with the foldedFormvar edge. The fact that these structures wereprimary cilia was confirmed by the observation thateach fluoresces after processing for the indirect immu-nofluorescent localization of microtubules (Fig. 3; seealso Jensen et al. 1979). The great majority of ciliated

cells in each cell line possessed a single cilium, but afew possessed two or more (Fig. 2K). When 100X ob-jectives were used, two or more ciliated cells were fre-quently visible in the same field of view (Figs 21-J, 3);the entire folded edge often contained over 600 cells.Profiles of primary cilia were also seen where the topcoverslip flattened them against the dorsal cell surface(Fig. 2L).

When viewed from the side, primary cilia appearedas conspicuous thin rods projecting from the dorsal cellsurface, usually in the vicinity of the nucleus(Fig. 2C,K). The projection angle of these cilia relativeto the growth substratum varied from near perpendicu-lar (Fig. 2E,G) to acute (Fig. 2B,C,H). However, overtime the projection angle of a cilium could slowlychange while cilia of adjacent cells remained stationary(data not shown). At no time did any cilia exhibitpropagated bending or vortical movements. However,all cilia over a few micrometres in length did bend inpassive response to solutions flowing through theviewing chamber, but returned to their original pos-ition upon cessation of flow (Fig. 4A-D).

On rare occasions, cilia from the same cell (Fig. 2K)or adjacent cells (not shown) appeared to adhere to oneanother. We also observed cilia that arched to touchneighbouring cells. One cilium was encountered thathad looped to touch its own base. These unusual ciliaryconfigurations may have been caused by the mechanicalmanipulation of the Formvar substratum during speci-men preparation.

Some of the primary cilia possessed one or moreswellings along the length of the shaft (Fig. 2F,11) or at

Primarv cilia in vivo 459

Fig. 2. Primary cilia in living kidney epithelial cells as viewed from above (A) and from the side using the flexible-substratum method (B-K). When viewed from above, cilia appear as dots (A, arrows), but are clearly imaged as thin rodswhen viewed from the side (B-L). In L the cilia (arrows) have been flattened against the cell surface by the coverslip.Swellings were frequently seen associated with the cilium shaft (F,H, arrows) or tip ( D - G . I J ) . I,J, the same fieldobserved by Hoffman modulation-contrast and phase-contrast optics, respectively. PtK, (A,I,J); PtK? (B—F,K,L); MDCK(G); BSC-40 (II). Hoffman modulation-contrast (G,I); phase-contrast (A,J); DIC (B-F,II,K,L). Bars: A,B,L, 20jum;C-K, lOjimi.

460 K. li. Roll,

Fig. 3. DIC (A) and fluorescence (B) micrographs of two primary cilia after fixation and staining for the indirectimmunofluorescent localization of microtubulcs. Bar, lOjUm.

Fig. 4. Selected photographs from a Hoffman modulation-contrast video recording of a primary cilium bending in responseto flow. A. Prior to flow; B, bending in response to flow; C, straightening as flow ceases; D, after cessation of flow. Bar,10 ;im.

the tip (Fig. 2D,E,I). Thes varicosities were sym-metrical about the shaft and varied in size from barelydiscernible thickenings (Fig. 2F,H) to approximately1 /im in diameter (Fig. 2D,E). They were found on theshortest (Fig. 2D,E) and longest (Fig. 2G-I) cilia inall cell types examined. Indeed, a logistic regressionanalysis (not shown) indicated that there was nocorrelation between the presence of these varicositiesand the length of a cilium in PtK| cells. Our im-pression, based on through-focus observations onmany cilia using DIC optics, was that the swellingswere bounded by the ciliary membrane and were notadhering debris. This conclusion was strengthened bythe fact that the percentage of cilia exhibiting thesevaricosities differed significantly for each cell line(PtK,, 14%; PtK2, 26%; LLC-PK,, 7%; BSC-40,12%).

Cilium length in confluent culturesThe primary cilia within a single preparation of a givencell line varied considerably in length (Fig. 5). How-ever, certain kidney cell lines consistently possessed

longer cilia than other lines. LLC-PK | cells had boththe greatest mean length (17-4jum) and the longestindividual cilium (55jim). Testing for the equality ofthe variances of PtK,, PtF<2 and BSC-40 cells showedthat these populations were statistically different(Fig. 6). LLC-PKi cells, however, retained a sig-moidal shape under this transformation, indicating thatthese cells possessed three poorly separated popu-lations of cilia.

Response of primary cilia to trypsin treatmentThe frequency of ciliation in freshly plated epithelialcells is low relative to that in confluent cultures(Wheatley, 1972; Rieder et al. 1979; Tucker el al.1979). This fact suggests that trypsin treatment duringsiibculturing induces the resorption or detachment ofcilia. To test this hypothesis, ciliated PtKi cells weretreated by perfusion with 0-05 % trypsin and examinedby video microscopy. As the cell rounds in response tothe trypsin its plasma membrane detaches from neigh-bouring cells and the substratum (Fig. 7A-F).

Pnmarv cilia in vivo 461

40-

30-

20-

30 40 50 60

40-

30-

20-

10-

0- 1

PtK,

1I10 20 30 40 50 60

BSC-40

20 30 40Length (/im)

50 60

Fig. 5. Histograms of primary cilia lengths in confluentcultures. The mean cilium length (± standard deviation)for each cell type was as follows: LLC-PK,,17-4± ll-lj<m; PtK2, 12-5 ± 6-2jUm; PtK,, 11-7 ± 4-8J(mi;BSC-40, 10-6 ± 6-5jUm.

Throughout this process, which takes approximately5 min, the cilium base remains positioned near thenucleus. As a result, the base of the cilium becomes

buried within the cell (Fig. 7D-F) . This findingdemonstrates that trypsin treatment does not leaddirectly to the loss of cilia during subculturing.

Discussion

Other than its simplicity, a major advantage of ourflexible-substratum technique is that many cells can beobserved from the side, and screened for primary cilia,in a single preparation. Although this approach workedwell for viewing primary cilia it does have certainlimitations. For example, optical interference pro-duced at the Formvar-cell and cell-medium interfacestends to obscure some detail, especially at the ventralcell surface. The refractive index (RI) of Teflon FEP(1-338) is similar to that of tissue-culture medium andthus provided significantly improved phase-contrastimages, but its birefringence limited its usefulness withpolarization optics. Furthermore, Teflon must betreated (e.g. with the bio-adhesive Cell-Tak) to pro-mote cell attachment and, in our hands, the productionof confluent monolayers was inconsistent. Porous sub-strata (e.g. polycarbonate) had the major advantage ofpromoting the partial differentiation of epithelial cells(reviewed by Steele et al. 1986). Porous polycarbonatemembranes also withstood the rigours of fixation andembedding for correlative light- and electron-micro-scopic studies, but their high refractive index of 1-586and greater thickness (lOjum) proved to be limitingfactors. Finally, Butvar (RI = 1-490) was also a goodgrowth substratum but did not adhere well to glasswhen submerged in growth medium. Despite its highRI (1-502), Formvar proved to be the most convenientand economical growth substratum for studying pri-mary cilia.

To our knowledge, previous descriptions concerningthe motile and behavioural properties of primary ciliain living cells are limited to three studies. Two of thesemake only single-sentence statements within the con-text of larger studies. In an early report Stubblefield &Brinkley (1966) noted, in a figure legend, that theprimary cilia of Chinese hamster fibroblasts "beaterratically in some cells". By contrast, Valentich (1981)mentioned that the primary cilia in MDCK cells arenon-motile. In a more thorough study, Odor & Blan-dau (1985) observed that primary cilia in rabbit oviductepithelial cells exhibited either the usual "to-and-fromotility" or vortical movements. However, it is notclear whether the 2/7m long 'solitary' cilia described inthis study were primary cilia, because the fimbrialepithelial cells employed were differentiating into cellscontaining numerous secondary (i.e. motile 9+2) cilia.Furthermore, their ultrastructural analysis of random80 nm sections did not allow them to determine withcertainty the axonemal symmetry of the cilia in ques-tion. Our results clearly demonstrate that the primary

462 K. E. Roth

74-2 p-

64-9

55-7

46-4oX̂•a>

IO 27-8

18-6

9-3

0-0L L

-30-0 -23-3 -16-7 -10-0 -3-3 3-3Expected (xlO1)

10-0 16-7 23-3 30-0

Fig. 6. Normalprobability plot ofthe square roottransformation ofcilia lengths inLLC-PK, (A),PtK2(B), PtK,(C) and BSC-40(D). The slope ofthe line is directlyproportional tothe variance andnon-normality inthe data appear asdeviations fromlinearity.

Fig. 7. Selected photographs from a Hoffman modulation-contrast video recording of a PtK| cell rounding in response totrypsin (0-05 %) treatment. In A two primary cilia (arrowheads), arising from the same cell, are shown 48 s after treatment.As the cell rounds, the cilia remain attached and the close association between the ciliary bases and the nucleus becomesclear (D-F) . Bar, lOjUm.

Piimarv cilia in vivo 463

cilia of kidney epithelial cells from various organismsare non-motile. This finding is not surprising, sinceprimary cilia are known to lack the dynein armsresponsible for the motility of secondary cilia. How-ever, striated rootlets are associated with the centriole/primary eilium complex of many cells (e.g. PtK2,Baron & Salisbury, 1987; 3T3, Albrecht-Buehler &Bushnell, 1980; pig kidney embryo, Vorobjev & Chent-sov, 1982; connective tissue, Pooler/ al. 1985; retinalneurones and pigment epithelia, Allen, 1965) and theseare thought to be contractile (Salisbury et al. 1984). Itis therefore possible that the primary eilium of somecells could undergo erratic motions, as those describedby Stubblefield & Brinkley (1965) and Odor & Blandau(1985), or change its projection angle, as reported here,without necessarily having a motile axoneme.

Previous electron-microscopic and immunofluor-escence studies have shown that the primary eilium ofmany cell types is short (^6jitm) and often containedwithin a deep invagination of the cell surface (reviewedby Wheatley, 1982). As a result the eilium may notprotrude, or may protrude only a short distance, fromthe cell. By contrast, we have found that the primarycilia of kidney epithelial cells can project up to 55jUmfrom the dorsal cell surface. This characteristic allowedus to compare the lengths of primary cilia of kidneyepithelia dervied from different organisms. The factthat primary cilia in PtK,, PtK2 and BSC-40 cells havestatistically different length populations indicates thatprimary eilium length is a distinct phenotypical charac-teristic of these cells. The primary cilia of LLC-PKicells do not show a normal length distribution andtherefore cannot be compared directly with the otherkidney cell lines. Rather, LLC-PK, cilia appear to havethree length populations, which might indicate that,like some other kidney cell lines (e.g. MDCK; Valen-tich, 1981), our LLC-PKi culture is heterogeneous(see also Gstraunthaler & Handler, 1987).

Primary cilia are common to most cells of thevertebrate kidney, including parietal and visceral layersof Bowman's capsule, proximal tubule, thin loop ofI Ienle, distal tubule, collecting tubules and occasion-ally in cells of the papillary duct; they appear to beabsent from the glomerular endothelium, calyx epi-thelium and the dark cells of the distal convoluted andcollecting tubules (Andrews & Porter, 1974; Andrews,1975; Ilucker & Frenzel, 1975). The distribution andmorphology of primary cilia within the kidney ledAndrews & Porter (1974) to suggest that they serve asensory function. Our results demonstrate that primarycilia of kidney epithelial cells bend passively in re-sponse to flow and return to their original positions atthe cessation of flow. Consistent with the hypothesisthat primary cilia act as mechanoreceptors (reviewedby Poole et al. 1985), the long primary cilia in kidney

epithelial cells may be an integral part of the biofeed-back system used to monitor changes in fluid motionwithin the nephron. Finally, primary cilia within thekidney may act as cellular osmometers to regulate waterand/or ion transport. Indeed, the swellings associatedwith the ciliary shaft may represent "circumscribedregions of the ciliary membrane which are sensitive tochanges in the environmental osmotic pressure" (seeDalen, 1981).

Ultrastructural studies have described similar swell-ings in primary and secondary cilia of vertebrate(Albrecht-Beuhler & Bushnell, 1980; Dalen, 1981;Eloffssone^ al. 1984; Menco & Farbman, 1985; Jensenet al. 1987) and invertebrate cells (Tamarin et al. 1974;Bergquist et al. 1977; Dilly, I977a,b; Ehlers & Ehlers,1978; Storch & Alberti, 1978; Arnold & Williams-Arnold, 1980; Boneet al. 1982; Matera & Davis, 1982;Hoverd, 1985). Relevant to our study is the fact thatsuch swellings are seen by scanning electron mi-croscopy (SEM) to be associated with the primary ciliaof rat collecting and human distal convoluted tubuleepithelia fixed in situ by perfusion (Andrews & Porter,1974; Hucker & Frenzel, 1975). Some have suggestedthat these swellings are artifacts of chemical fixation(e.g. see Ehlers & Ehlers, 1978; Dalen, 1981). How-ever, since they are seen hi vivo in both primary (ourstudy) and secondary cilia (Arnold & Williams-Arnold,1980) they are not, at least in these examples, fixationartifacts.

Primary cilia in supporting cells of developing ratnasal epithelium possess swollen or 'bulbed' tips at day14, but not at day 17, of development (Menco &Farbman, 1985). The formation and loss of bulbs inthese cells is thought to be a cellular response tochanging environmental or physiological conditionsduring development. Similarly, the paddle-like tips onthe motile 9+2 'discocilia' of Pleurobranchaea, whichare also found in many other marine invertebrate cells,are lost in hypertonic but not hypotonic sea water(Matera & Davis, 1982). This observation, and thosefrom additional studies, led Matera & Davis (1982) toconclude "that dilations of ciliary membranes representa common morphological specialization subservingchemoreception".

Recently, Jensen et al. (1987) suggested that theswellings associated with primary cilia are formed atthe tips of growing cilia as a result of differingelongation rates between the axonemal microtubulesand the ciliary membrane. Our observation that vari-cosities are present at various points along the ciliaryshaft, regardless of eilium length, is not consistent withthis hypothesis. However, Jensen et al. also noted thatlarge bulbs, containing an electron-opaque material,form in association with ciliating centrioles that hadfailed to generate axonemal microtubules in the pres-ence of colcemid. This observation suggests that ciliary

464 K. E. Roth

components are actively transported from the cyto-plasm into the cilium during ciliogenesis and that in theabsence of elongating axonemal microtubules thesecomponents 'stack up' to form a large bulb. Dilly(1977a) has similarly suggested that the swollen regionsseen along the shafts of discocilia are the result ofmaterial transport within the cilia. A prediction of thishypothesis is that these swellings should move alongthe ciliary shaft in the absence of ciliary growth. We arecurrently using time-lapse video microscopy to evalu-ate this prediction and to study other properties ofprimary cilia.

We thank Dr Andrew Reilly for help with the statisticalanalysis, Dr Colin Izzard for his assistance with DICmicroscopy, and Dr L. I. Binder for supplying the mono-clonal /3-tubulin antibody used in this study. We also thankMr Edwin Davison for assisting in the initial development ofthis project, and Mr Gerald Rupp, Mr Richard Cole, MsNadine Myers and Ms Elizabeth Mandeville for assistance inthe laboratory. Portions of this work were submitted by K. E.Roth in partial fulfilment of the requirements for the Masterof Science degree at the State University of New York atAlbany. This work was supported by NIH BiotechnologicalResource Related grant RR-021S7 awarded to C. L. Rieder.S. S. Bowser is an American Heart Association, New YorkState Affiliate Research Fellow.

References

ABSOLOM, D. R. (1986). Phagocytosis: Generalmethodology. Meth. Enzyw. 132, 85-180.

ALBRECHT-BEUHLER, G. & BUSHNELL, A. (1980). The

ultrastructure of primary cilia in quiescent 3T3 cells.lixpl Cell Res. 126, 427-437.

ALLEN, R. A. (1965). Isolated cilia in inner retinal neuronsand in retinal pigment epithelium. J. Ullrastnict. Res.12, 730-747.

ANDREWS, P. M. (1975). Scanning electron microscopy ofhuman and Rhesus monkey kidneys. Lab. Invest. 32,610-618.

ANDREWS, P. M. & PORTER, K. R. (1974). A scanningelectron microscopic study of the nephron. Am. J. Anat.140,81-116.

ARCHER, P. L. & WHEATLEY, D. N. (1971). Cilia in cell-cultured fibroblasts. II. Incidence in mitotic and post-mitotic BHK21/C13 fibroblasts. J. Anal. 108, 223-237.

ARNOLD, J. M. & WILLIAMS-ARNOLD, L. D. (1980).

Development of the ciliature pattern on the embryo ofthe squid Longio pealei: A scanning electron microscopestudy. Biol. Bull. mar. biol. Lab., Woods Hole 159,102-116.

BERGQUIST, P. R., GREEN, C. R., SINCLAIR, M. E. &

ROBERTS, II. S. (1977). The morphology of cilia insponge larvae. Tissue & Cell 9, 179-184.

BARON, A. T. & SALISBURY, J. L. (1987). Identificationand localization of a novel cytoskeletal centrosome-associated protein, "centrin", in PtK2 cells. J. Cell Biol.105, 205a.

BONE, Q., RYAN, K. P. & PULSFORD, A. (1982). The

nature of complex discocilia in the endostyle of Ciona(Tunicata: Ascidiacae). Mikivskopie 39, 149-153.

BOOCOCK, C. A., BROWN, A. F. & DUNN, G. A. (1985). A

simple chamber for observing microscope specimens inboth top and side views. J . Micmsc. 137, 29-34.

CANDE, W. Z. (1982). Permeabilized cell models forstudying the chromosome movements in dividing PtK|cells. Meth. Cell Biol. 25, 57-68.

Cox, D. R. (1970). The Analysis of 'Binary Data. London:Methuen Co. Ltd.

DALEN, H. (1981). An ultrastructural study of primarycilia, abnormal cilia and ciliary knobs from the ciliatedcells of the guinea pig trachea. Cell Tiss. Res. 220,685-697.

DELLINGER, O. P. (1906). Locomotion of amoebae andallied forms. J . exp. Zoot. 3, 337-358.

DILLY, P. N. (1977«). Material transport within specializedciliarv shafts on Rtiabdopleura zooids. Cell I'iss. Res.180/367-381.

DILLY, P. N. (19776). Further observations of transportwithin paddle cilia. Cell Tiss. Res. 185, 105-113.

EHLERS, U. & EHLERS, B. (1978). Paddle cilia anddiscocilia - genuine structures? Cell Tiss. Res. 192,489-501.

ELOFFSSON, R., ANDERSSON, A., FALCK, B. & SJOBORG, S.

(1984). The ciliated human kcratinocyte.,7- Ullrastnict.Res. 87, 212-220.

GSTRAUNTHALER, G. & HANDLER, J. S. (1987). Isolation,growth and characterization of a gluconeogenic strain ofrenal cells. Am. J. Physiol. 252, C232-C238.

HARRIS, A. (1969). Initiation and propagation of the rufflein fibroblast locomotion. J. Cell Biol. 43, 165a-166a.

HARRIS, J. K. (1978). A photo elastic substrate techniquefor dynamic measurements of forces exerted by movingorganisms. J . Micmsc. 114, 219-228.

HLINKA, J. & SANDERS, F. K. (1972). Real and reflectedimages of cells in profile. I. A method for the study ofcell movement and adhesion..°J. Cell Sci. 11, 221-231.

HOVERD, W. A. (1985). Histological and ultrastructuralobservations of the lophophore and larvae of thebrachiopod, Notosaria nigricans (Sowerby 1846). .7. nat.Hist. 19, 831-850.

HOCKER, H. & FRENZEL, H. (1975). Scanning electronmicroscopy of the distal nephron and calyx of the humankidney. Yirchoics Arch. B, Zellpath. 18, 157-164.

INGRAM, V. M. (1969). A side view of moving fibroblasts.Nature, Land. 222, 641-644.

JENSEN, C. G., DAVISON, E. A., BOWSER, S. S. & RIEDER,

C. L. (1987). Primary cilia in PtK, cells: Effects ofcolcemid and taxol on cilia formation and resorption.Cell Motil. Cytoskel. 7, 187-197.

JENSEN, C. G., JENSEN, L. C. W. & RIEDER, C. L. (1979).

The occurrence and structure of primary cilia in asubline of Potorous tridactylus. lixpl Cell Res. 123,444-449.

MATERA, E. M. & DAVIS, W. J. (1982). Paddle cilia(discocilia) in chemosensitive structures of the gastropodmollusk Pleurobranchaea califomica. Cell Tiss. Res. 222,25-40.

Primary cilia in vivo 465

MENCO, B. M. & FARBMAN, A. I. (1985). Genesis of ciliaand niicrovilli of rat nasal epithelia during development.I. Olfactory epithelium, qualitative studies. J. Cell Sci.78, 283-310.

MORI, Y., AKEDO, H., TANIGAKI, Y., TANAKA, K. &

OKADA, M. (1979). Ciliogenesis in tissue-cultured cellsby the increased density of cell population. Expl CellRes. 121, 435-439.

ODOR, D. L. & BLANDAU, R. J. (1985). Observations on

the solitary cilium of rabbit oviductal epithelium: Itsmotility and ultrastructure. Ani.J. Anat. 174, 437-453.

POOLE, C. A., FLINT, M. H. & BEAUMONT, B. W. (1985).

Analysis of the morphology and function of primary ciliain connective tissues: A cellular cybernetic probe? CellMotil. 5, 175-193.

RIEDER, C. L. & BOWSER, S. S. (1985). Correlative

iinmunofluorescence and electron microscopy on thesame section of Epon-embedded material. J . Hislochem.Cylochem. 33, 165-171.

RIEDER, C. L., JENSEN, L. & JENSEN, C. (1979). The

resorption of primary cilia during mitosis in a vertebrate(PtK,) cell line. J . Ultrastmct. Res. 68, 173-185.

RUPP, G., BOWSER, S. S., MANNELLA, C. A. & RIEDER, C.

L. (1986). Naturally occurring tubulin-containingparacrystals in Allogromicr. Immunocytochemicalidentification and functional significance. Cell Motil.Cytoskel. 6, 363-375.

SALISBURY, J. L., BARON, A., SUREK, B. & MELKONIAN,

M. (1984). Striated flagellar roots: Isolation and partialcharacterization of a calcium-modulated contractileorganelle. J . Cell Biol. 99, 962-970.

SANDERS, E. J. & PRASAD, S. (1979). Observation ofcultured embryonic epithelial cells in side view. J. CellSci. 38, 305-314.

STEELE, R. E., PRESTON, A. S., JOHNSON, J. P. &

HANDLER, J. S. (1986). Porous-bottom dishes for cultureof polarized cells. Am. J. Physiol. 251, C136-C139.

STORCH, V. & ALBERTI, G. (1978). Ultrastructural

observations on the gills of polychaetes. Helgolanderwiss. Meeresunters 31, 169-179.

STUBBLEFIELD, E. & BRINKLEY, B. R. (1966). Cilia

formation in Chinese hamster fibroblasts in vitro as aresponse to colcemid. J. Cell Biol. 30, 645-652.

TAMARIN, A., LEWIS, P. & ASKEY, J. (1974). Specialized

cilia of the byssus attachment plaque forming region inMytilus califontiamts. J. Morph. 142, 321-328.

TUCKER, R. W., PARDEE, A. B. & FUJIWARA, K. (1979).

Centriole ciliation is related to quiescence and DNAsynthesis in 3T3 cells. Cell 17, 527-535.

VALENTICH, J. D. (1981). Morphological similaritiesbetween the dog kidney cell line MDCK and themammalian cortical collecting tubule. Ann. N.Y. Acacl.Sci. 372, 384-405.

VOROBJEV, I. A. & CHENTSOV, Y. S. (1982). Centrioles inthe cell cycle. 1. Epithelial cells. J . Cell Biol. 93,938-949.

WHEATLEY, D. N. (1972). Cilia in cell-cultured fibroblasts.IV. Variation within the mouse 3T6 fibroblastic line. J'.Anat. 113, 83-93.

WHEATLEY, D. N. (1982). The Centriole: A CentralEnigma of Cell Biology, pp. 147-184. New York:Elsevier Biomedical Press.

(Received 11 September 1987 - Accepted 17 December1987)

466 K. E. Roth