Use of confocal linescan to document ciliary beat frequency

Authors

  • R. T. DOYLE,

    1. Department of Genetics, Development and Cell Biology, Roy J. Carver Laboratory of Ultrahigh Resolution Biological Microscopy, Institute for Combinatorial Discovery, Iowa State University, Ames, IA 50011, U.S.A.
      *Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, IA 52242, U.S.A.
      †Department of Biomedical Sciences, College of Veterinary Medicine, Iowa State University, Ames, IA 50011, U.S.A.
    Search for more papers by this author
  • T. MONINGER,

    1. Department of Genetics, Development and Cell Biology, Roy J. Carver Laboratory of Ultrahigh Resolution Biological Microscopy, Institute for Combinatorial Discovery, Iowa State University, Ames, IA 50011, U.S.A.
      *Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, IA 52242, U.S.A.
      †Department of Biomedical Sciences, College of Veterinary Medicine, Iowa State University, Ames, IA 50011, U.S.A.
    Search for more papers by this author
  • * N. DEBAVALYA,

    1. Department of Genetics, Development and Cell Biology, Roy J. Carver Laboratory of Ultrahigh Resolution Biological Microscopy, Institute for Combinatorial Discovery, Iowa State University, Ames, IA 50011, U.S.A.
      *Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, IA 52242, U.S.A.
      †Department of Biomedical Sciences, College of Veterinary Medicine, Iowa State University, Ames, IA 50011, U.S.A.
    Search for more papers by this author
  • W. H. HSU

    1. Department of Genetics, Development and Cell Biology, Roy J. Carver Laboratory of Ultrahigh Resolution Biological Microscopy, Institute for Combinatorial Discovery, Iowa State University, Ames, IA 50011, U.S.A.
      *Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, IA 52242, U.S.A.
      †Department of Biomedical Sciences, College of Veterinary Medicine, Iowa State University, Ames, IA 50011, U.S.A.
    Search for more papers by this author

R. T. Doyle. Tel: 515 294 6513; fax: 515 294 7134; e-mail: rtdoyle@iastate.edu

Summary

We present a method to document ciliary beat frequency with the linescan function of a scanning confocal microscope, using ciliated tracheal cells and free-swimming rotifers as examples. Depending on the clarity of the original data, the ciliary beat frequency can be determined from the confocal linescan directly or from an intensity linescan analysis of the original data. Fast Fourier transform treatment of the data can be used to verify the derived ciliary beat frequency. The linescan approach allows analysis of simple ciliary movements displayed by the ciliated tracheal cells, as well as complex movements performed by free-swimming rotifers while feeding.

Introduction

The determination of the ciliary beat frequency (CBF) of ciliated cells has been performed for many years. Collection of these data has always been closely associated either with techniques that allow rapid collection of images of the beating cilia (Salathe & Bookman, 1999; Li et al., 2000; Nlend et al., 2002; Hirst et al., 2003, 2004; Sisson et al., 2003; Wyatt et al., 2003; Zhang & Sanderson, 2003; Piatti et al., 2004; Robertson et al., 2004; Shiima-Kinoshita et al., 2004) or with optical methods to document the light intensity changes imparted by beating cilia (Dalhamn & Rylander, 1962; Naito & Kaneko, 1973; Mercke et al., 1974; Lee & Verdugo, 1976, 1977; Verdugo, 1980; Puchelle et al., 1982; Verdugo & Golborne, 1988; Devalia et al., 1990; Bayram et al., 1998; Thanou et al., 1999; Dimova et al., 2003; Barrera et al., 2004).

Image-based methods result in long and painstaking analysis of frame-by-frame observation of cilial movement to document the CBF. Photocell- and photomultiplier-based methods produce real-time light intensity records that can be saved and analysed or processed on the fly to produce fast Fourier transform (FFT) interpretation of the frequency records; this helps to position and focus the optical system in real time to optimize data collection (Salathe & Bookman, 1999; Nlend et al., 2002).

Here we report a photomultiplier-based technique to determine CBF using the linescan function of a scanning confocal microscope. Both cultured porcine ciliated tracheal cells and free-living Brachionus rotifers were imaged to test this system. We used the 512-linescan function of a Prairie Technologies (Middleton, WI, U.S.A.) scanning confocal microscope to produce a light intensity record of the active ciliated cell, and then produced a light intensity data record with the linescan analysis function of the image analysis software metamorph (Universal Imaging Corporation, Downingtown, PA). Depending on the clarity of the cilia frequency record, the data can be analysed by counting the intensity cycles in the confocal linescan record or the metamorph record and converting that to CBF. The data can also be imported into software for FFT analysis.

Materials and methods

Preparation of cultured porcine ciliated tracheal cells

Epithelial cells were isolated from healthy 3–6 month old specific-pathogen-free pigs. The methods of cell isolation were modified from those of Young et al. (2000). Mid-portions of tracheas were aseptically collected in cold phosphate buffered saline and were transported to the laboratory. Upon arrival, the connective tissues were aseptically removed from the tracheas, leaving only the membranous and cartilaginous parts. The tracheas were each rinsed and transferred into separated centrifuge tubes. Tracheal epithelial cells were dissociated by chilled enzyme solution containing 0.1% DNase and 0.005% pronase in Ca2+- and Mg2+-free minimum essential medium (5.37 mm KCl, 110 mm NaCl, 44 mm NaHCO3, 0.91 mm NaHPO4, 0.25 µm Fe(NO3)3, 0.0015% phenol red, 1 mm sodium pyruvate, 100 units of penicillin per mL, and 0.1% streptomycin), for 18–24 h.

On the following day, 10% fetal bovine serum was added to inactivate the enzymes. Cell pellets were collected at 125 g for 5 min and resuspended in 5% serum airway medium (a 1 : 1 mixture of Delbecco's minimum essential media (high glucose) and Ham's F-12 medium containing 5% fetal bovine serum, 1% of 10 mm minimum essential medium non-essential amino acids, 0.12 units of insulin per mL, 100 units of penicillin per mL, and 0.1% streptomycin; prepared and kept in a 5% CO2 incubator overnight prior to use). The cell suspension was transferred to tissue culture dishes and incubated in a 5% CO2 incubator for at least 1 h to remove attached fibroblasts. The non-attached epithelial cells were collected and gently triturated to break the cell clumps. The cells were counted and stored at −140 °C in serum airway medium containing 20% fetal bovine serum and 7% dimethylsulfoxide.

On the day of the experiment, the vial of cells was thawed in the 37 °C water bath and washed with Krebs Ringer bicarbonate buffer solution (119 mm NaCl, 1.5 mm CaCl2, 1.2 mm KH2PO4, 4.75 mm KCl, 1.2 mm MgSO4, 20 mm HEPES (4-(2-Hydroxyethyl)piperazine-l-ethanesulfonic acid), 0.1% bovine serum albumin and 11 mm glucose, pH 7.3).

Preparation of free-living protozoa

Approximately 900 mL of water, floating plants and sediment were collected from a small freshwater lake (Lake Laverne) on the campus of Iowa State University. Approximately 100 g of soiled litter from the equine waste collection system at Iowa State University College of Veterinary Medicine was added to the lake water in a clear plastic container. The container was illuminated 24 h per day with a 100 W tungsten light bulb. The culture was examined daily for type and abundance of ciliated protozoa. A large population of Brachionus quadridentatus was present in the culture by day 5. These protozoa were used as subjects to test the CBF detection system.

Imaging setup

The system is composed of an inverted Nikon Eclipse TE200 (Nikon, Melville, NY, U.S.A.) microscope with DIC (differential interference contrast) optics attached to a Prairie Technologies laser scanning confocal microscope. The excitation lasers of the confocal microscope were not used to illuminate the ciliated cells. The cells were illuminated in transmission mode with the 12 V 100 W halogen microscope lamp. The light path was directed into the confocal microscope and passed through a 488 nm/510 nm longpass dichroic mirror, through a 150 micrometre (µm) pinhole and on to one of the photomultipliers in the confocal microscope. A 512 × 512 pixel image was taken of the ciliated cells to serve as a reference image. The position of the 512 linescan was determined by a cursor mark on the reference image, and the confocal control software was set to save each linescan to disc automatically. Where possible, cells were chosen that presented cilia ‘in-line’ with the linescan such that the moving cilia would pass the linescan and generate differences in light intensity as they moved. Free-moving rotifers were imaged once they stopped swimming and fed in place, therefore it was not possible to choose only cilia ‘in-line’ with the linescan direction.

The full cycle time of the confocal linescan function is set by the interline time. The interline time for the Prairie Technologies laser scanning confocal microscope is 5.037 ms, which yields a frame rate for the linescan of 198.6 Hz.

Experimental protocols

Ciliary beat frequency measurement in ciliated tracheal epithelial cells

The tracheal epithelial cells were loaded on to poly-L-lysine-coated cover slips and mounted in a custom-made chamber. The cells were imaged with the imaging system described above at room temperature, 23–24 °C.

Images were captured using a Nikon 60X 1.4 NA Plan Apo DIC oil immersion objective. Viable ciliated cells with good attachment and active cilia were chosen. CBFs of cilia were recorded using the linescan mode of the confocal and a white light source in transmission mode as explained above. Cells that presented active cilia that were parallel to the linescan direction were chosen. Patterns of cilial movement through time were captured in the confocal linescan images. Images were taken while the microscope was manually focused to produce the clearest images possible. Images were analysed using the linescan measurement subroutine of the metamorph offline software (Universal Imaging Corporation, Downingtown, PA) to reveal the CBF, which was recorded as beats s−1. The image intensity data were converted to an ASCII file and imported to the sigview software (sigview 32 version 1.9.5.1, http://www.sigview.com) for FFT analysis. The intensity data were pre-conditioned in sigview before FFT analysis by subtracting the mean from each data point, removing linear trends, removing any values greater than 2 standard deviations of the mean, and applying a Hanning window. The data could be analysed for the CBF at several points along the data analysis stream, depending on the clarity of the captured images. For very clear images, the beats per second could be counted directly from the original linescan image. Less clear patterns could be discerned from the metamorph linescan analysis of the confocal linescan image, and converted to CBF. The FFT software treatment was used in this study to provide an added level of analysis and to serve as a comparison with the results derived from the linescan analysis.

Ciliary beat frequency measurement of B. quadridentatus

Individual B. quadridentatus were isolated with a hand-held micropipette from cultured lake water and placed on to a glass microscope slide. A cover glass outfitted with four dental wax ‘feet’ was placed over the B. quadridentatus to provide adequate space for the rotifers to swim freely. The microscope slide was inverted and mounted on the Nikon Eclipse TE200 microscope. Images were captured using a Nikon 20X 0.75 NA 0.75 Plan Apo DIC objective.

It was necessary to follow individual rotifers until they stopped swimming and fed in place before imaging them with the linescan mode of the confocal microscope. For this reason it was not possible to choose cilia that were orientated ‘in-line’ with the direction of the linescan. This empirical approach coupled with the distinctive behaviour and morphology of the rotifers revealed additional information about the dynamics of the feeding apparatus. Re-alignment of the linescan orientation to accommodate specimen position is possible on systems with the software capability to re-orient the galvanometers.

Ciliated tracheal epithelial cells were imaged with a Prairie Technologies scanning confocal microscope in transmission mode as stated above to generate linescan records of ciliary behaviour. Figure 1(A) displays a reference image showing the ciliated cells and a line indicating the position of the linescan. The linescan image is shown in Fig. 1(B) with time as the y-axis of the image. The image is composed of 512 linescans with time starting at the top of the image and progressing to the bottom of the image. Movement of the cilia creates intensity differences that can be seen in the left-hand portion of Fig. 1(B), whereas constant features of the cell give rise to solid vertical lines.

Figure 1.

Example of a linescan reference image and resulting linescan record. A small cluster of ciliated porcine tracheal cells is imaged in panel (A) with a white line intersecting the cells indicating where the linescan image will be taken. Panel (B) is the linescan record produced from the linescan. It is composed of 512 single linescans progressing in time. Time is the y-axis, with the first acquired scan at the top and the last scan at the bottom of the record. Movement during the linescan produces intensity differences, as can be seen along the left-hand portion of the record, whereas stationary features give rise to straight lines and columns in the record.

The clarity of the linescan images was variable, depending on the cells being imaged. This led to varying amounts of uncertainty in the analysis of the CBF of the cells. Some images allowed direct computation of the CBF from the linescan image as seen in Fig. 2(A). For linescan records that were less clear, the linescan image was analysed with a linescan subroutine of the image analysis software metamorph, as shown in Fig. 2(B), to produce an optical intensity record (Fig. 2C). For further verification of the CBF, the metamorph intensity record was converted into an ASCII file and imported into a software program (sigview) for FFT analysis to determine the frequency spectrum of the intensity data (Fig. 2D). For the data shown in Fig. 2, the CBF determined from a hand count of the confocal linescan image was 6.20 Hz, whereas the CBF determined by a hand count of the metamorph intensity data was 6.57 Hz. FFT analysis of the metamorph intensity data yielded a CBF of 6.59 Hz. The improvement in CBF determination resulting from using the intensity record generated in metamorph was used to document the CBF from confocal linescans where no clear visual pattern was evident. An example of this is seen in Fig. 3, where a comparison of the metamorph linescan analyses is shown in two regions of the confocal linescan image. One region is where active cilia are located (Fig. 3, line 1) and the other is a random area away from any cells (Fig. 3, line 2). The corresponding metamorph intensity records along with the resultant FFT records display the ability of this technique to extract the CBF from visually unclear images. The FFT analysis of line 1 yields a CBF of 3.86 Hz, whereas the FFT analysis of line 2 shows no dominant frequency.

Figure 2.

Analysis of the confocal linescan record. The CBF can be determined directly from a very clear confocal linescan record, as seen in panel (A) where each cycle of cilial movement is marked with a white arrow. Another method of CBF determination is to analyse the confocal linescan record using the linescan analysis function of the image analysis software metamorph. Panel (B) shows the placement of the active analysis line on the confocal linescan, whereas panel (C) displays the intensity data generated from that line. FFT frequency analysis of the metamorph intensity data is shown in panel (D).

Figure 3.

Analysis of a visually unclear confocal linescan record. Panel (A) shows the placement of two active metamorph linescan analysis lines. Line 1 is placed on an area of active cilia, whereas line 2 is positioned in an area away from any cells. The intensity data and FFT analysis of each of these lines are shown in panels (B) and (C). Analysis of the line 1 data reveals a CBF of 4.28 Hz (visual inspection of the intensity data) or 3.86 Hz (FFT analysis). The line 2 data do not reveal any useful frequency information.

Analysis of B. quadridentatus ciliary beat frequency

It was decided also to test the system using free-living ciliates from a nearby freshwater lake on the Iowa State University campus. Lake water cultures were set up as explained above and individual rotifers (B. quadridentatus, Fig. 4B inset) were chosen as test subjects. A rotifer was allowed to swim freely until it came to rest. At this time the animal would begin to feed; a 512 × 512 reference image was taken and the point where the linescan images would be taken was marked on the reference image (Fig. 4A). The rotifer has a complex feeding apparatus composed of rings of cilia called trochal discs (Fig. 4B arrows) which are extruded from the anterior of the animal (arrow in Fig. 4A) in a rhythmic pattern. The feeding behaviour results in a complex series of movements of the individual cilia and trochal discs. This complexity can be seen in Fig. 4(C) where the feeding apparatus of the rotifer encounters the confocal linescan. During feeding the animal protracts and retracts the feeding apparatus in addition to moving the cilia on the trochal discs. In the example shown in Fig. 4, the feeding apparatus encountered the confocal linescan at an approximate 40° angle. The linescan data record now contains information about several aspects of the feeding apparatus because of the angle at which the apparatus encountered the linescan. Figure 4(C) line 1 reflects the movements of one of the trochal discs, and line 2 displays the movement of cilia. The trochal disc has two major components, one moving at a rate of 1.55 Hz and another at 2.70 Hz (Fig. 4D). The cilia are moving at a rate of 28.23 Hz (Fig. 4D).

Figure 4.

Documentation of complex movements of the feeding apparatus of B. quadridentatus. Panel (A) shows a reference image of a living B. quadridentatus with a white line indicating where confocal linescans will be acquired. An enlarged image of the anterior portion of a fixed B. quadridentatus is presented in panel (B) to show details of the trochal discs (arrows) containing cilia. The inset displays the entire organism. Panel (C) is a confocal linescan record of feeding apparatus activity; line 1 is placed where a trochal disc encountered the linescan and line 2 where cilia encountered the linescan. The intensity and FFT records for both lines are shown in panel (D).

Discussion

Documentation of the CBF of ciliated cells has been accomplished by two basic methods: high-speed photography and photocell interpretation of light intensities. The method that we present here is a variation of the photocell methods using a photomultiplier. One constraint placed upon whatever method is used to determine the CBF is the Nyquist sampling criterion (Squire et al., 2000) or Nyquist minimum (Zhang & Sanderson, 2003), which requires that the sampling rate of the imaging method be at least twice the maximum speed documented. The sampling rate of the confocal linescan used in this study is 198.6 Hz, well above the CBFs documented in the ciliated porcine cells or rotifers in this study.

Extraction of the CBF depended upon the clarity of the linescan image produced, and could be accomplished at several stages in the analysis procedure. For very clear images the cilial cycles can be counted on the confocal linescan image and converted to Hz. The entire linescan takes 2.578 s to complete; therefore, the number of cilial cycles is divided by this value to yield the CBF. For less clear confocal linescan records, the linescan analysis function of the metamorph image analysis software was used to document the cilial activity. The metamorph linescan analysis function allows the user to choose the number of adjacent rows of pixels to include in the analysis. Using one row of pixels will deliver a very sharp record, whereas including more than one row of pixels in the analysis results in a smoothed record. The user can experiment with different settings to arrive at an optimal output from the original data. The data generated by the metamorph software can then be analysed by FFT software for comparison.

Use of the confocal linescan approach to document the CBF also allows for investigation of more complex movement behaviour as shown by the rotifer data in this study. Therefore, this technique can lead to collection and analysis of live cell behaviour other than just ciliary movement.

Acknowledgements

We would like to thank Dr Elizabeth Walsh from the University of Texas at El Paso for assistance in identifying the rotifer species used in this study.

Ancillary