Atomic force microscopy of histological sections using a chemical etching method
7
Atomic force microscopy of histological sections using a new electron beam etching method T. OSADA,* H. ARAKAWA,* M. ICHIKAWA² & A. IKAI* *Department of Biological Sciences, Faculty of Bioscience and Biotechnology, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 227, Japan ²Anatomy and Embryology, Tokyo Metropolitan Institute for Neuroscience, Fuchu, Tokyo 183, Japan Key words. Atomic force microscope (AFM), histological section, transmission electron microscope (TEM), vomeronasal organ. Summary In order to examine histological sections of the rat vomeronasal epithelium with the atomic force microscope (AFM), we developed an electron beam etching method that improves the resolution of AFM images. This method results in AFM images comparable to those obtained with the transmission electron microscope (TEM). Ultrathin tissue sections embedded in epoxy resin were observed before and after the treatment with electron beam radiation. Before electron beam treatment, epithelial structures such as the microvilli surface, dendritic processes, the supporting cell layers and the neuronal cell layers were all visible using the AFM. However, only a few subcellular structures could also be resolved. The AFM images were not as clear as those obtained with the TEM. After electron beam treatment, however, the resolution of AFM images was greatly improved. Most of the subcellular structures observed in TEM images, including the inner membrane of mitochon- dria, ciliary-structure precursor body, junctional complexes between the neurons and supporting cells, and individual microvilli were now visible in the AFM images. The electron beam treatment appeared to melt the embedding resin, bringing subcellular structures into high relief. The result of this study suggests that electron beam etching of histolo- gical samples may provide a new method for the study of subcellular structure using the AFM. Introduction The atomic force microscope (AFM) is a type of scanning probe microscope (SPM) which gives three-dimensional images by recording interactions between the scanning probe and sample surface (Binnig et al., 1982, 1986). The AFM is expected to become a promising new tool for the study of biological materials because it can be used for both conductive and insulating materials in liquid as well as in air (Lal & John, 1994). The resolution of the AFM depends primarily on the condition of the sample surface and the sharpness of the AFM probe tip. The AFM is typically used to resolve atomic details on the surface of sample speci- mens. Its use in examining histological sections has been limited. Recently AFM images using biological materials have been reported. For example, the typical circular structure of plasmid DNAs and complexes between DNA and proteins on the substrate have been resolved with AFM (Bustamante et al., 1992; Vesenka et al., 1992; Hansma & Hoh, 1994). Other biological materials such as proteins (Weisenhorn et al., 1990; Hallett et al., 1995), viruses (Ikai et al., 1993; Imai et al., 1993) and cell surfaces (Paul et al., 1994; Eppell et al., 1995) have been studied extensively with the AFM. One advantage of the AFM is its ability to obtain real-time three-dimensional images and monitor biological processes such as the clotting process of fibrinogen (Drake et al., 1989) and cell growth of the yeast cells (Gad & Ikai, 1995). The AFM can also be used to measure forces between the probe tip and the sample surface. As the AFM tip makes contact with the sample surface, physical properties of the sample surface as well as its topography can be obtained (Mitsui et al., 1996; Ikai, 1996). Thus, the binding forces between avidin–biotin (Lee et al., 1994), ligand–receptor (Florin et al., 1994; Moy et al., 1994) and antigen–antibody (Dammer et al., 1996) have been measured using the AFM. In this report we describe an electron beam etching method which allows subcellular structures to be imaged by the AFM. AFM images of subcellular structures in the vomeronasal epithelium were examined using this method. Vomeronasal epithelium was used here because of its structural simplicity. The sensory epithelium of the vomer- onasal organ (Wysocki, 1979; Halpern, 1987) is similar in structure to the olfactory epithelium, consisting of only three cell types: supporting cells, receptor neurons and Journal of Microscopy, Vol. 189, Pt 1, January 1998, pp. 43–49. Received 28 April 1997; accepted 10 July 1997 43 q 1998 The Royal Microscopical Society
Transcript of Atomic force microscopy of histological sections using a chemical etching method
Atomic force microscopy of histological sections using anew electron beam etching method
T. OSADA,* H. ARAKAWA,* M. ICHIKAWA† & A. IKAI**Department of Biological Sciences, Faculty of Bioscience and Biotechnology, Tokyo Institute ofTechnology, 4259 Nagatsuta, Midori-ku, Yokohama 227, Japan†Anatomy and Embryology, Tokyo Metropolitan Institute for Neuroscience, Fuchu, Tokyo 183,Japan
Key words. Atomic force microscope (AFM), histological section, transmissionelectron microscope (TEM), vomeronasal organ.
Summary
In order to examine histological sections of the ratvomeronasal epithelium with the atomic force microscope(AFM), we developed an electron beam etching method thatimproves the resolution of AFM images. This method resultsin AFM images comparable to those obtained with thetransmission electron microscope (TEM). Ultrathin tissuesections embedded in epoxy resin were observed before andafter the treatment with electron beam radiation. Beforeelectron beam treatment, epithelial structures such as themicrovilli surface, dendritic processes, the supporting celllayers and the neuronal cell layers were all visible using theAFM. However, only a few subcellular structures could alsobe resolved. The AFM images were not as clear as thoseobtained with the TEM. After electron beam treatment,however, the resolution of AFM images was greatlyimproved. Most of the subcellular structures observed inTEM images, including the inner membrane of mitochon-dria, ciliary-structure precursor body, junctional complexesbetween the neurons and supporting cells, and individualmicrovilli were now visible in the AFM images. The electronbeam treatment appeared to melt the embedding resin,bringing subcellular structures into high relief. The result ofthis study suggests that electron beam etching of histolo-gical samples may provide a new method for the study ofsubcellular structure using the AFM.
Introduction
The atomic force microscope (AFM) is a type of scanningprobe microscope (SPM) which gives three-dimensionalimages by recording interactions between the scanningprobe and sample surface (Binnig et al., 1982, 1986). TheAFM is expected to become a promising new tool for thestudy of biological materials because it can be used for bothconductive and insulating materials in liquid as well as in
air (Lal & John, 1994). The resolution of the AFM dependsprimarily on the condition of the sample surface and thesharpness of the AFM probe tip. The AFM is typically usedto resolve atomic details on the surface of sample speci-mens. Its use in examining histological sections has beenlimited.
Recently AFM images using biological materials havebeen reported. For example, the typical circular structure ofplasmid DNAs and complexes between DNA and proteins onthe substrate have been resolved with AFM (Bustamante etal., 1992; Vesenka et al., 1992; Hansma & Hoh, 1994).Other biological materials such as proteins (Weisenhorn etal., 1990; Hallett et al., 1995), viruses (Ikai et al., 1993;Imai et al., 1993) and cell surfaces (Paul et al., 1994; Eppellet al., 1995) have been studied extensively with the AFM.One advantage of the AFM is its ability to obtain real-timethree-dimensional images and monitor biological processessuch as the clotting process of fibrinogen (Drake et al.,1989) and cell growth of the yeast cells (Gad & Ikai, 1995).The AFM can also be used to measure forces between theprobe tip and the sample surface. As the AFM tip makescontact with the sample surface, physical properties of thesample surface as well as its topography can be obtained(Mitsui et al., 1996; Ikai, 1996). Thus, the binding forcesbetween avidin–biotin (Lee et al., 1994), ligand–receptor(Florin et al., 1994; Moy et al., 1994) and antigen–antibody(Dammer et al., 1996) have been measured using the AFM.
In this report we describe an electron beam etchingmethod which allows subcellular structures to be imaged bythe AFM. AFM images of subcellular structures in thevomeronasal epithelium were examined using this method.Vomeronasal epithelium was used here because of itsstructural simplicity. The sensory epithelium of the vomer-onasal organ (Wysocki, 1979; Halpern, 1987) is similar instructure to the olfactory epithelium, consisting of onlythree cell types: supporting cells, receptor neurons and
Journal of Microscopy, Vol. 189, Pt 1, January 1998, pp. 43–49.Received 28 April 1997; accepted 10 July 1997
43q 1998 The Royal Microscopical Society
basal or precursor cells. Chemosensory neurons in thevomeronasal epithelium are typically bipolar with a sensorydendrite that extends to the surface of the epithelium andan axon process that reaches the olfactory bulb. Thevomeronasal epithelium and subcellular structures providean excellent specimen for testing the resolution of this newAFM method in histological sections. A dose–responserelationship between electron beam treatment and imageresolution was determined. This new method of AFMimaging may be applicable to a variety of histologicalsamples and preparations. By varying the electron beamexposure time, optimum resolution can be obtained for theobservation of structures that normally would requiretransmission electron microscopy.
Materials and methods
Adult Sprague–Dawley rats were anaesthetized withsodium pentobarbital and perfused with physiological salinefollowed by 2% glutaraldehyde, 2% paraformaldehyde in0.1 M phosphate buffer. Coronal sections were cut throughthe vomeronasal organ and postfixed with 1% osmiumtetroxide in 0.1 M phosphate buffer. The fixed slices weredehydrated and embedded in epoxy resin (Quetol 812).Ultrathin sections with a silver grey interference colourwere cut and mounted on formvar-coated one-hole coppergrids. Sections of the vomeronasal epithelium were firstexamined with the AFM and then stained with uranylacetate and lead citrate for examination with the electronmicroscope (JEOL JEM 1200 EXII). Electron beam etchedsamples for the AFM were prepared by radiating thevomeronasal epithelium sections in the same way aswhen prepared for TEM observations, using the 80 kVelectron beam from an electron microscope. The currentdensity typically used here was 50 6 10 pA cm¹2. Thistreatment was identical to that used for TEM observationsexcept that the time of radiation exposure and the positionof the radiation spot were carefully controlled.
AFM imaging was performed in tapping mode in air witha Nanoscope III (Digital Instruments, Santa Barbara, CA)mounted with the J scanner. Micro Cantilevers for AC mode(OMCL-AC120TS) were purchased from Olympus OpticalCo., Ltd (Tokyo, Japan). The cantilever oscillation was set ata rate oscillated at slightly below its resonance frequency(typically 300 kHz) with the drive amplitude set around 200mV. This resulted in a detected amplitude of cantileveroscillation of 4.0 V. The set-point was adjusted to 95% orslightly below the amplitude of free oscillation after theprobe was positioned on the sample. The quality of theobserved images was not dependent on the set-point.Samples were scanned at a rate of 1 Hz with appropriatefeedback gain (typically, integral gain ¼ 0.5, proportionalgain ¼ 1.0).
Results
A comparison of TEM and AFM images of the vomeronasalepithelium sections was first obtained from samples
Fig. 1. Comparison of AFM image and TEM image of the samevomeronasal epithelium. (a) AFM image of histological thin sectionof the vomeronasal epithelium. The image is a composite of over-lapping areas spanning the full depth of the epithelium. AFMimages were obtained using the tapping mode in air. Height iscoded on an intensity scale. (b) TEM image of the same histologicalvomeronasal epithelium section. The tissue section was stainedwith uranyl acetate and lead citrate prior to TEM scanning. Struc-tures visible in both images include the microvilli surface (MS),supporting cell layer (SCL), neuronal cell layer (NCL) and basementmembrane (BM). Arrows indicate dendritic processes. Scalebar ¼ 10 mm.
44 T. OSADA ET AL .
q 1998 The Royal Microscopical Society, Journal of Microscopy, 189, 43–49
prepared according to standard TEM procedures. Figure1(a) shows a montage of AFM images of the vomeronasalepithelium. The microvilliar surface, dendritic processes ofsensory neurons, and supporting cell layer can be identifiedin this image. This same histological section was thenstained with uranyl acetate and lead citrate, and examinedwith TEM. Figure 1(b) shows a TEM image of the left half ofthe same section shown in Fig. 1(a). Identical structureswere observed with both AFM and TEM, which demon-strates that the AFM can be used to examine ultrathinsections prepared for TEM observation on copper grids.Figure 2 shows images at higher magnification for bothAFM and TEM. Subcellular structures such as mitochon-dria, ciliary-structure precursor bodies, microtubles andsmooth endoplasmic reticulum on dendritic processes werenot as clear in the AFM images as they were in the TEMimages. The apical ends of dendritic processes and microvilliwere distinguishable in TEM but while not in AFM images.
In the course of our study, we discovered that AFMimages were improved dramatically if the images were takenafter they had been examined with the TEM. Figure 3(a)shows an AFM image of the circular area exposed to theTEM electron beam. The epithelial surface and dendriticprocess were observed clearly within this circular spot.Figure 3(b) shows this region at a higher magnification. It islikely that the tissue space filled with resin was alteredfollowing exposure to the electron beam, creating anincrease in surface relief and enhanced resolution ofsubcellular structures. This led to a significant improvementin image contrast compared with the pretreated sample. Athigher magnification the AFM image reveals dendriticprocesses (Fig. 3c), mitochondria and the ciliary-structureprecursor body (Fig. 3c). Junctional complexes could also beseen between neurons and supporting cells and individualmicrovilli were now visible on the distal end of the dendrites.
Figure 4 shows high-magnification AFM images of the
q 1998 The Royal Microscopical Society, Journal of Microscopy, 189, 43–49
Fig. 2. High-magnification images of thedendritic region of the vomeronasal epithe-lium. AFM images (a and c) are not asgood as TEM images (b and d). The dendri-tic process can be seen in both a and b.Scale bar ¼ 2 mm. At higher magnification(c and d) subcellular structures can beseen. Scale bar ¼ 7 mm.
ELECTRON BEAM ETCHING FOR AFM 45
subcellular structure of dendritic processes. Figure 4(a)shows the distal end of a dendritic process and manymicrovilli were observed. A fibrous structure which waslikely to be a microtubule can be seen in Fig. 4(b). Figure4(c) shows cistae within mitochondrial structures. Figure4(d) shows a subcellular structure known as the ciliary-structure precursor body (Taniguchi & Mikami, 1985) or
granular dark body (Kratzing et al., 1971). The veins of aleaf-like structure were also observed in AFM images. AFMgave similar or better resolution images when comparedwith TEM images of the same sample.
In order to determine the effect of electron beamtreatment, sections were radiated for 30, 15, 5 and 0 swith 80 kV (Fig. 5a–d). Prior to radiation, indistinguishable
Fig. 3. AFM images of the epithelium after exposure to electron beam treatment. (a) Area exposed to electron beam observed using wide areascan AFM image. Circular area appears to be depressed or shrunken by the electron beam exposure. Scale bar ¼ 10 mm. (b) At higher mag-nification the AFM image reveals dendritic structures. Scale bar ¼ 4 mm. (c) With additional magnification, the AFM images reveal moredetails. Scale bar ¼ 1 mm. (c) A composite of overlapping regions scanned at high resolution. When compared with images in Fig. 2, thecontrast of the AFM image after electron beam exposure is much better and is similar to that obtained with TEM. Arrowheads denote junc-tional complexes between neurons and supporting cells. Small and large arrows indicate ciliary-structure precursor body and mitochondria,respectively.
46 T. OSADA ET AL .
q 1998 The Royal Microscopical Society, Journal of Microscopy, 189, 43–49
subcellular structures including mitochondria and ciliary-structure precursor bodies were difficult to observe andappeared concave against other parts of dendritic processes(Fig. 5d). With increased electron beam exposure time, thesubcellular structures became clearer, and they stood out insharper relief against other structures. The topographicalimages of subcellular structures improved in resolution andcontrast with electron beam radiation. Individual microvilliwere distinguishable. The contrast of AFM images seemed tobe reversed at around 5 s of exposure.
Discussion
It is critical in AFM studies to prepare sample surfacesproperly. They should be fairly smooth with gentlyundulations corresponding to the structure of the surface.Although Epon resin sections used for TEM have a smoothsurface, they do not have sharp relief. There is a need to
develop new methods to improve the resolution of surfacestructures. The electron beam radiation method we havedeveloped engraves the epoxy resin deeply, resulting inimproved resolution of resolving the subcellular structures.
If the surface is flat a topological image cannot beobtained with AFM. Interestingly, tissue sections cut with adiamond knife have a relatively flat surface, yet can beresolved with AFM, although the images are not very clear.It is not known why topographic images can be obtainedfrom the histological sections with flat surfaces. The surfaceof Epon thin sections of bacterial cells have also been imagedwith AFM when used in the contact mode (Amako et al.,1993), but the authors they did not explain how theseimages were obtained from a flat surface. It is possible thatregional differences in hardness and softness of the surfacemay result in a topographical image. The treatment of thesections with electron staining reagents alone did notimprove AFM images (data not shown).
q 1998 The Royal Microscopical Society, Journal of Microscopy, 189, 43–49
Fig. 4. AFM images of subcellular structures. (a) Distal end of dendritic processes. Individual microvilli can be seen after the electron beamtreatment. (b) Fibrous structures likely to be microtubules (arrows). (c) Mitochondria and their cistae (arrows). (d) Structure referred to asthe ciliary-structure precursor body or granular dark body. Scale bar ¼ 0.1 mm.
ELECTRON BEAM ETCHING FOR AFM 47
The electron beam radiation typically used in TEM bringsthe subcellular structures into sharp and high relief. Undersufficient exposure of the electron beam, the subcellularstructures become quite visible and details of the subcellularstructures can be obtained with high-resolution AFMimages. We did not investigate any other etching methodsfor AFM. Ushiki et al. (1994) examined embedment-freesections of the biological materials using AFM. Theyintroduced this method because Epon-embedded sectionsgave them little biological information. Our method hasovercome this problem by engraving the embedmentembedding material with electron beam radiation. Ushikiet al. (1994) also examined biological materials with
different types of AFM. Additional studies are needed todetermine which method is optimal for AFM samplepreparation.
When the sections were exposed to electron beams for along time, or to a stronger electron beam, we observeddeformation of the sample (data not shown). Randomradiation of the electron beam onto the section bycontinuous observation with TEM results in irregulardeformation of the surface. It is difficult to scan a largearea of the deformed section with AFM, but the detail of thesubcellular structure was still observed within the smallscanning area. We did not investigate the details ofdeformation of sections by the electron beam, but our
Fig. 5. AFM images of the vomeronasal sections after electron beam exposure of (a) 30 s, (b) 15 s, (c) 5 s, (d) 0 s. Scale bar ¼ 3 mm. Optimumresolution seems to occur after 30 s of exposure.
48 T. OSADA ET AL .
q 1998 The Royal Microscopical Society, Journal of Microscopy, 189, 43–49
initial results suggest that optimum conditions of electronbeam radiation must be determined for different histologicalsections. So far we have not seen any deformation ofsections during AFM observation. Our results indicate thatAFM observation causes less damage to the Epon samplethan TEM observation. Our new AFM method may prove tobe better than TEM in some cases since it does less damageand has good resolution of subcellular structures. It couldprovide a new approach to the study of epoxy resinembedded histological tissue.
Acknowledgments
We thank Dr R. Costanzo for a critical review of themanuscript and for helpful discussion. We also thank N.Iwasaki for her technical assistance.
References
Amako, K., Takade, A., Umeda, A. & Yoshida, M. (1993) Imagingof the surface structures of Epon thin sections created with aglass knife and a diamond knife by the atomic force microscope.J. Electron Microsc. 42, 121–123.
Binnig, G., Quate, C.F., Gerber, C.H. & Weibel, E. (1982) Surfacestudies by scanning tunneling microscopy. Phys. Rev. Lett. 49,57–61.
Binnig, G., Rohrer, H. & Gerber, C.H. (1986) Atomic forcemicroscopy. Phys. Rev. Lett. 56, 930–933.
Bustamante, C.J., Vesenka, J., Tang, C.L., Rees, W., Guthold, M. &Keller, R. (1992) Circular DNA molecules imaged in air byscanning force microscopy. Biochemistry, 31, 22–26.
Dammer, U., Hegner, M., Anselmetti, D., Wagner, P., Dreier, M.,Huber, W. & Guntherodt, H.J. (1996) Specific antigen/antibodyinteraction measured by force microscopy. Biophys. J. 70, 2437–2441.
Drake, B., Orater, C.B., Weisenhorn, A.L., Gould, S.A.C., Albrecht,T.R., Quate, C.F., Cannell, D.S., Hansma, H.G. & Hansma, P.K.(1989) Imaging crystals, polymers and processes in water withthe atomic force microscope. Science, 243, 1586–1589.
Eppell, S.J., Simmons, S.R., Albrecht, R.M. & Marchant, R.E. (1995)Cell surface receptors and proteins of platelet membranes imagedby scanning force microscopy using immunogold contrastenhancement. Biophys. J. 68, 671–680.
Florin, E.L., Moy, V.T. & Gaub, H.E. (1994) Adhesion forces betweenindividual ligand pairs. Science, 264, 415–417.
Gad, M. & Ikai, A. (1995) Method for immobilizing microbial cellson gel surface for dynamic AFM studies. Biophys. J. 69, 2226–2233.
Hallett, P., Offer, G. & Miles, M.T. (1995) Atomic force microscopyof the myosin molecule. Biophys. J. 68, 1604–1606.
Halpern, M. (1987) The organization and function of thevomeronasal system. Ann. Rev. Neurosci. 10, 325–362.
Hansma, H.G. & Hoh, J. (1994) Biomolecular imaging with theatomic force microscope. Ann. Rev. Biophys. Biomol. Struct. 23,115–139.
Ikai, A. (1996) STM and AFM of bio/organic molecules andstructure. Surf. Sci. Rep. 26, 261–332.
Ikai, A., Yoshimura, K., Arisaka, F., Ritani, A. & Imai, K. (1993)Atomic force microscopy of bacteriophage T4 and its tube-baseplate complex. FEBS Lett. 326, 39–41.
Imai, K., Yoshimura, K., Tomitori, M., Nishikawa, O., Kokawa, R.,Yamamoto, M., Kobayashi, M. & Ikai., A. (1993) Scanningtunneling and atomic force microscopy of T4 bacteriophage andtobacco mosaic viruses. Jpn. J. Appl. Phys. 32, 2962–2964.
Kratzing, J. (1971) The structure of the vomeronasal organ in thesheep. J. Anat. 108, 247–260.
Lar, R. & John, S.A. (1994) Biological applications of atomic forcemicroscopy. Am. J. Physiol. 266, C1–C21.
Lee, G.U., Kidwell, D.A. & Colton, R.C. (1994) Sensing discretestreptavidin–biotin interactions with atomic force microscopy.Langmuir, 10, 354–357.
Mitsui, K., Hara, M. & Ikai, A. (1996) Mechanical unfolding ofalpha-2-macroglobulin molecules with atomic force microscope.FEBS Lett. 385, 29–33.
Moy, V.T., Florin, E.L. & Gaub, H.E. (1994) Intermolecular forcesand energies between ligands and receptors. Science, 266, 257–259.
Paul, J.K., Nettikadan., S.R., Ganjeizadeh, M., Yamaguchi, M. &Takeyasu, K. (1994) Molecular imaging of Naþ, Kþ-ATPase inpurified kidney membranes. FEBS Lett. 346, 289–294.
Taniguchi, K. & Mikami, S. (1985) Fine structure of the epithelia ofthe vomeronasal organ of horse and cattle. A comparative study.Cell Tissue Res. 240, 41–48.
Ushiki, T., Shigeno, M. & Abe, K. (1994) Atomic force microscopyof embedment-free sections of cells and tissues. Arch. Histol.Cytol. 57, 427–432.
Vesenka, J., Guthold, M., Tang, C.L., Keller, D., Delaine, E. &Bustamante, C. (1992) A substrate preparation for reliableimaging of DNA molecules with the scanning force microscope.Ultramicroscopy, 42/44, 1243–1249.
Weisenhorn, A.L., Drake, B., Prater, C.B., Gould, S.A.C., Hansma,P.K., Ohnesorge, F., Egger, M., Heyn, S.P. & Gaub., H.E. (1990)Immobilized proteins in buffer imaged at molecular resolution byatomic force microscopy. Biophys. J. 58, 1251–1258.
Wysocki, C.J. (1979) Neurobehavioral evidence for the involve-ment of the vomeronasal system in mammalian reproduction.Neurosci Biobehav. Rev. 3, 301–341.
q 1998 The Royal Microscopical Society, Journal of Microscopy, 189, 43–49
ELECTRON BEAM ETCHING FOR AFM 49
