Correlative high-resolution morphologic analysis of the three-dimensional organization of human chromosomes

SCANNING VOL. 22, 273–281 (2000) Received: February 10, 2000© FAMS, Inc. Accepted with revision: May 30, 2000

Correlative High-Resolution Morphologic Analysis of the Three-Dimensional Organization of Human Chromosomes

PIETRO GOBBI, STEFAN THALHAMMER,* MIRELLA FALCONI,† ROBERT W. STARK,* WOLFGANG M. HECKL,* GIOVANNI MAZZOTTI

Department of Human Anatomy, University of Bologna, Bologna, Italy; *Institute of Crystallography, University of Munich,Munich, Germany; †Institute of Cytomorphology, CNR, Bologna, Italy

Summary: A correlative morphologic analysis was carriedout on isolated metaphase chromosomes by means of fieldemission in-lens scanning electron microscopy (FEISEM)and atomic force microscopy (AFM). Whereas FEISEMprovides ultra-high resolution power and allows the surfaceanalysis of biological structures free of any conductivecoating, the AFM allows imaging of biological specimensin ambient as well as in physiologic conditions. The analy-sis of the same samples was made possible by the use ofelectrical conductive and light transparent ITO glass asspecimen holder. Further preparation of the specimen spe-cific for the instrumentation was not required. Both tech-niques show a high correlation of the respective morpho-logic information, improving their reciprocal biologicalsignificance. In particular, the biological coat represents abarrier for surface morphologic analysis of chromosomespreads and it is sensitive to protease treatment. The chem-ical removal of this layer permits high-resolution imagingof the chromatid fibers but at the same time alters the chro-mosomal dimension after rehydration. The high-resolutionlevel, necessary to obtain a precise physical mapping of thegenome that the new instruments such as FEISEM andAFM could offer, requires homogeneously cleaned sampleswith a high grade of reproducibility. A correlative micro-scopical approach that utilizes completely different phys-ical probes provides complementary useful informationfor the understanding of the biological, chemical, andphysical characteristics of the samples and can be appliedto optimize the chromosome preparations for furtherimprovement of the knowledge about spatial genome orga-nization.

Key words: atomic force microscopy, field emission in-lens scanning electron microscopy, chromosomal structure,correlative microscopy

PACS: 87.16.Sr / 87.64.-t / 87.64.Dz / 87.64.Ee

Introduction

The detailed understanding of the nuclear cell functionsrequires an accurate knowledge of the spatial organizationof the nuclear structures. For many years, the study ofhuman metaphase chromosomes was carried out with lightmicroscopy following staining protocols disturbing thenative chromosomal structure. The approach by scanningelectron microscopy (SEM) provides higher resolutioncompared with that in light microscopy and permits surfaceanalysis of the chromosomal structure which cannot be ade-quately obtained from transmission electron microscopy(TEM). Nevertheless, in order to obtain high resolution inSEM observations, the use of a high electron acceleratingvoltage (up to 30 kV) is required (Sanchez-Sweatman et al.1993, Sumner 1996, Sumner and Ross 1989). With theseexperimental conditions, the sputtercoating or a conductivestaining of the samples is generally required (Sumner et al.1994, Wanner and Formanek 1995). Both procedures allowelectron charging dispersion from the sample but mayobscure fine details and produce sample alterations (Her-mann and Müller 1992).

Today, only few techniques are available for high-reso-lution imaging of chromosomal material with reduced arti-facts. In the present study, the field emission in-lens scan-ning electron microscope (FEISEM) and the atomic forcemicroscope (AFM) were utilized. The FEISEM representsa special kind of SEM, fitted with a cold cathode field emis-sion electron gun (Nagatani et al. 1987, Pawley 1997), thatcan operate at low accelerating voltage with reduced elec-tron charging of the sample. In fact, the low voltage and lowcurrent electron beam of the FEISEM, together with a liq-uid nitrogen anticontamination device corresponding tothe specimen area and an “in-lens” assembly of the elec-tron-optic column, allows for high-resolution imaging ofthe biological sample without any conductive staining ormetal coating. Nevertheless, contamination of the specimen

The experiments were partially supported by the Vigoni Italy-Germany CRUI-DAAD research exchange program.

Address for reprints:

Pietro GobbiIst. Anatomia Umana NormaleUniversità di Bolognav. Irnerio 4840126 Bologna, ItalyE-mail: [email protected]

is highly reduced compared with that in a conventionalSEM (Nagatani et al. 1987).

The sample location between the objective polepiecelimits the dispersion of the secondary electrons collectedby the magnetic field of the lens. Taken together, these char-acteristics allow the observation of uncoated biologicalsamples with a higher resolution compared with that in theconventional SEM (Gobbi et al. 1999, Lattanzi et al. 1998,Rizzi et al. 1995, Rizzoli et al. 1994).

Since its invention in 1986, the use of the AFM hasbecome a standard technique in biology (Binnig et al.1986, Mariani et al. 1994, Ushiki et al. 1996), not requir-ing any particular treatment of the sample, especially inconstant force mode analysis. The AFM allows imagingof DNA in ambient as well as in physiologic conditions,but with lower resolution compared with electron micro-scopy approaches (Fritzsche and Henderson 1997a, Leubaet al. 1998). Recently, the possible use of the AFM as ananomanipulator of genetic material such as DNA extrac-tion from human metaphase chromosome with higher pre-cision than standard microdissection techniques has beendemonstrated (Heckl 1998, Stark et al. 1998, Thalhammeret al. 1997).

In the present study, FEISEM and AFM were combinedon the same metaphase chromosome samples obtainedfrom standard cytogenetic preparations of human HL 60cells, after cleaning the metaphase spreads with differentprocedures (Rizzoli et al. 1994). The analysis of the samesamples was facilitated by the use of conductive glass(indium tin oxide [ITO] glass) for the chromosome mappreparation. These two different technical approaches showa high correlation of the respective morphologic informa-tion, both in normal and treated samples. The high-resolu-tion potential of the FEISEM, together with the possibilityto observe hydrated samples and/or to nanomanipulate thespecimen with the AFM, confirm morphologic data andoffer enhanced information on their biological significance.

Materials and Methods

Chromosome Preparation

Human HL 60 leukemia cells were cultivated in RPMI1640 medium supplemented with 10% fetal calf serumand 100 units/ml penicillin and streptomycin. The cellswere arrested in metaphase after addition of 0.01 µg/ml col-cemid (Sigma Chemical Co., St. Louis, Missouri, USA) for3 h, harvested, and incubated in 75 mM KCl for 15 min at37°C. After washing three times in 3:1 cold methanol:aceticacid, chromosome spreads were made by dropping thesuspension onto the conductive surface of perfectly cleanedand degreased 3 × 6 mm ITO glasses. Metaphases werethen air dried, dehydrated with an ethanol series (70, 90,100%), air dried, and stored in a dry chamber until use. Sub-sequently, one of two different cleaning solutions wereemployed. The first treatment was a mix of 1ml 3M Na-

acetate, 20 µl 10 mg/ml protease K, and 20 µl 20% sodiumdodecyl sulfate (SDS) for 2 min at 50°C (also called hotmix solution); the second was a 0.1 mg/ml protease K inH2O solution applied for 5 min at 50°C. After the clean-ing procedure, the ITO glasses with the metaphase spreadswere washed 2 min in distilled water, dehydrated in anethanol series (25, 50, 70, 90, 100%), and air dried.

Field Emission in-Lens Scanning Electron MicroscopyAnalysis

Cleaned and uncleaned metaphase spreads on ITOglasses were mounted onto the microscope specimen hold-ers and observed without any conductive coating on aJEOL JSM-890 FEISEM (Jeol Ltd., Tokyo, Japan) at 7 kVaccelerating voltage, 1 × 10–11 A probe current, and 0° to45° tilt angle.

Atomic Force Microscopy Analysis

The same ITO glasses utilized first for FEISEM obser-vations were then employed to obtain AFM data. A micro-scope with 130 µm xy-scan range and 10 µm z-scannerAFM (Topometrix Explorer by Thermo Microscopes,Sunnyvale, Calif., USA) was utilized. The AFM wasmounted on top of an inverted microscope (Axiovert 135,Carl Zeiss Mikroskopie, Göttingen, Germany) in order toselect the metaphase spreads both in ambient and wet con-ditions. Two different types of cantilevers were used formeasurements: observations of the human chromosomes inambient conditions were carried out by means of stiff can-tilevers in constant force mode (pointprobe CONT, springconstant c = 0.3 N/m, nominal tip radius r < 10 nm, Dr. OlafWolter GmbH, Wetzlar-Blankenfeld, Germany). For obser-vations of the same specimens immersed in physiologicsaline solution, soft cantilevers were utilized in constantforce mode (sharpened microlevers, c = 0.02 N/m, r < 10nm, ThermoMicroscopes, USA). The loading forces dur-ing AFM measurements were 10–20 nN in ambient con-ditions and < 5 nN in liquids.

All height measurements were obtained by the analysisof multiple AFM images with the Topometrix line mea-surement software after baseline correction. About 10 mea-surements were performed in similar chromosomaldomains of different sample preparations to confirm thedata. The maximum error of the measurements was ± 5%.

Results

Field Emission in-Lens Scanning Electron MicroscopyAnalysis

The low magnification FEISEM detection of uncoatedchromosomes is made easier by the contrast between thebiological specimen and the conductive surface of the ITOglass. The chromosomal surface shows a flattened net-

274 Scanning Vol. 22, 5 (2000)

work of fibers, which are relaxed organized in two chro-matids (Fig. 1a). At higher magnification, the chromoso-mal network appears to be poorly relieved from the ITOglass surface and to be constituted by fibers covered by ahomogeneous phase that does not allow a clear detectionof the fine arrangement of the fibers and that covers also aspread area around each single chromatid. Only globular

structures of 30–50 nm are protruding from this coat andare distinguishable (Fig. 1b). By the use of Na-acetate, pro-tease K, and SDS solution, the structure of the chromoso-mal fibers becomes more evident. The covering homoge-neous phase looks localized in limited areas of thechromosomal surface, often assuming a round shapeinscribed in a spoke-like structure or sometimes resulting

P. Gobbi et al.: AFM and FEISEM imaging of human chromosomes 275

FIG. 1 Field emission in-lens scanning electron microscopy images of metaphase chromosomes. (a) Three short untreated chromosomes appearflattened on the ITO surface (scale bar = 2 µm); (b) at higher magnification, the chromosome surface appears to be constituted of a networkcovered by a homogeneous phase. Only some nodal points of 30–50 nm are clearly detectable (arrows) (bar = 300 nm); (c) metaphase mapstreated with Na-acetate, protease K, and SDS. The surface of the samples appears to be partially covered by a homogeneous phase. A typicalspoke-like arrangement of the coat and the chromatid fibers is evident (asterisk) (bar = 1 µm); (d) the high magnification of the same surfaceshows the 10 nm size chromosomal fiber as the main constituent of a network. The larger nodal points are still evident (arrows) and some of thenetwork meshes are filled by the homogeneous phase (bar = 100 nm).

1a 1b

1c 1d

in a simple dark area on the network. When the absence ofthe covering phase discloses the underlying structure, thefibers of the network appear to form round meshes of reg-ular diameter. Chromatin fibers are also detectable scatteredaround the chromatid (Fig. 1c). At very high magnification,the chromosomal network is constituted mainly by a 10 nmfiber surrounding meshes with a diameter of 30–80 nm.Some nodal points are more irregular in size, ranging

between 5 and 50 nm. It is also clearly evident that thehomogeneous phase covers some of the chromatid’s roundmeshes (Fig. 1d). After treatment with protease K, at lowmagnification the chromosomes appear deprived of anycoat in the central part of the chromatid, and an area morein contrast with the surface of the ITO glass around thechromatid axis is detectable (Fig. 2a). By increasing mag-nification, a network constituting the central part of each

276 Scanning Vol. 22, 5 (2000)

FIG. 2 Field emission in-lens scanning electron microscopy analysis of protease K-treated samples. (a) The centromeric region (arrows) andthe chromatids are well recognizable. A dark halo surrounds the entire chromosome (bar = 1 µm); (b) increasing the magnification, the chro-mosomal surface appears to be constituted of a network. Some fibrillar structures parallel to the axis of the chromatid are well detectable (arrows).The halo around the chromatid appears to be formed by a mix of fibers and homogeneous phase (asterisks) (bar = 400 nm); (c) the mean diam-eter of the network meshes appears enlarged with respect to the controls. The homogeneous phase partially covers the periphery of the chro-matid (bar = 150 nm); (d) the diameter of the fibers constituting the network appears increased with respect to the controls (bar = 50 nm).

2a 2b

2c 2d

chromatid can be observed. Major fibrillar structures areevident parallel to the longitudinal axis of the chromosome.The area surrounding the center of the chromatid is formedby a flattened mixture of fibers and homogeneous phasecontinuous with the chromatid network (Fig. 2b). Theaspect of the periphery of this structure, characterized bythe presence of the homogeneous phase, is very similar tothe surface of untreated samples while the mean diameterof the round shaped meshes appears larger with respect toother experimental conditions (Fig. 2c). At the highestmagnifications, the visible network is formed by fibers of

increased diameter (15–20 nm) if compared with the con-trols or after treatment with Na-acetate, protease K, andSDS solution (Fig. 2d).

Atomic Force Microscopy Analysis of DryChromosomes

Despite the completely different construction of theAFM compared with the FEISEM, the features obtainedfrom the dry samples are precisely superimposable at thesame magnification level. The analysis of untreated sam-

P. Gobbi et al.: AFM and FEISEM imaging of human chromosomes 277

3a 3b

3c 3d

FIG. 3 Atomic force microscopy images of the same samples as in Figures 1 and 2. (a) A flat and short untreated chromosome is well identi-fiable on the ITO glass; (b) the surface appears to be constituted of protruding large nodal points; (c) after detergent and protease treatment, thechromosomal surface presents characteristic spoke-like structures (asterisks), whereas the pure protease K-treated samples (d) present a moredefined network structure. Scale bar = 1 µm.

ples reveals the same flattened network structure formingthe previously described chromatids (Fig. 3a). At the high-est magnification obtainable with the same cantilever andthe xy – scan piezo device, it is possible to recognize thenodal points of the underlying network in a not well definedarrangement (Fig. 3b). After treatment with Na-acetate,protease K, and SDS solution, the AFM analysis reveals thepresence of flat and round-shaped structures on the chro-mosome that partially cover the chromatid surface. Aroundthis structure, a spoke-like arrangement of fibers is clearlydetectable (Fig. 3c). The treatment of the metaphase chro-mosomes with protease K discloses a well identifiable net-work structure, where the central area of each chromatidshows a structural axis formed by large fibers. The pres-ence of chromatin material spread around the chromo-some is detectable (Fig. 3d).

Atomic Force Microscopy Analysis of RehydratedChromosomes

The rehydration of the human chromosomes causesmorphologic rearrangement of their surface appearance.Both the untreated chromosomes and those cleaned byNa-acetate, protease K, and SDS solution appear swollenin large clumps, still maintaining the two chromatid struc-tures (Fig. 4a,b). In particular, the clump arrangement ofthe chromosomes treated with Na-acetate, protease K, andSDS solution strongly resembles a G-banding longitudinalpattern (Fig. 4b). The aspect of the rehydrated protease K-cleaned maps is flattened on the ITO glass, and the pres-ence of small clumps characterizing the chromatid surfacestill remains. There is no evidence of a fibrous axis on thechromosome in this sample (Fig. 4c).

Analysis of the Chromosomal Height after and beforeRehydration

The measurements of the chromosome profiles, obtainedby AFM and indicating the mean height of the chromatids,show that the different treatments induce a modification oflittle significance in the mean height of the dried chromo-somes (Table I). Differences are more relevant after therehydration of the chromosomes. The rehydration ofuntreated chromosomes causes an about three-fold increaseof the mean height compared with the corresponding driedchromosomes. The rehydration of chromosomes treatedwith the hot mix solution and the protease K solution pro-duces a mean 1.5-fold increase of the height with respectto the dry condition (Table I).

Discussion

The different treatments performed on chromosomesmodify their surface structure and are well detectable withboth technical approaches. In spite of the differences interms of technology (electron interaction vs. nanome-

278 Scanning Vol. 22, 5 (2000)

FIG. 4 Atomic force microscopy images of rehydrated chromo-somes. Untreated samples (a) present a swollen surface characterizedby small clumps. The chromatids remain well identifiable. The sur-face of detergent-plus protease-treated maps (b) is characterized bya conformation that resembles the longitudinal G-banding. The pro-tease K only-treated chromosomes (c) appear more flattened on theITO glass, with some large clumps along the chromatids. Bar = 1 µm.

4a

4b

4c

chanical contact), the images are completely superimpos-able when the same specimen was observed at a similarmagnification by either FEISEM or AFM. This kind ofrelationship is made possible because we could observe thesame sample with the two different microscopes withoutany specific treatment related to the instrumentation.

The use of ITO conductive glass represents a funda-mental technical step for this work. The conductive surfaceof the glass allows efficient discharging of non-secondaryelectrons from the sample. In addition, the transparency ofthe ITO glass permits the identification of the region ofinterest in conventional phase contrast microscopy beforeFEISEM or AFM analysis.

By considering the two techniques, they both confirm thepresence of a homogeneous biological coat over undi-gested chromosomes after a standard cytogenetic prepara-tion protocol. The presence of this chromosome coating iswell known to the morphologists. This layer can mask theimaging of the chromosomal surface in standard conditions,after in situ hybridization labeling protocols, or duringnanomanipulations of the sample (Heckl 1995, Rizzi et al.1995, Rizzoli et al. 1994, Sumner 1996, Sumner and Ross1989).

Several authors suggested that the coating layer is madeof residues of cell membranes, RNA, and cytoplasmiccomponents (Tamayo et al. 1999); the detection of com-pletely coat-free chromosomes in standard cytogeneticpreparation is an occasional finding, is time consuming, andcannot be used for routine ultrastructural analysis. There-fore, different techniques have been proposed to removethis coat by proteolytic and/or detergent solutions (Milleret al. 1988, Sumner 1996, Sumner and Ross 1989) or byfreezing of the metaphase spreads (Martin et al. 1994).However, the results were inconsistent and seem to bestrictly dependent on specific biological models.

Considering the need for a conductive staining or ametallic sputter of the samples before observation in aconventional SEM, it is possible to understand the differ-ences in chromosomal fibers diameters previouslydescribed by other authors and probably related to theinterference of this biological coat on the samples pro-cessed for observation (Martin et al. 1994, Sanchez-Sweat-man et al. 1993, Sumner et al. 1994).

This is consistent with the FEISEM high magnifica-tions of the chromosomal surface that show different fiberdiameters and organization depending upon the treatment.In fact, the globular structures of 30–50 nm protruding fromthe untreated chromatid surface, observed also by AFM byother authors (Ali Ergun et al. 1999), are otherwise poorlycomparable with the 10 nm fibers framework organizationpreviously described on selected coat-free chromosomes(Rizzi et al. 1995, Rizzoli et al. 1994). In our experimen-tal conditions, these 10 nm fibers can be detected exclu-sively after sample digestion.

Both FEISEM and AFM reveal the sensitivity of thechromosomal coat to cleaning solutions, thus demonstrat-ing subsequent modifications of the chromatid morphology.After incubation with the detergent solution, the chromo-somal surface appears not to be completely free of thehomogeneous coat. This coat appears as aggregates local-ized in well-distinguishable areas of the chromosome. Asimilar behavior could be due to the protein precipitatingcapabilities of the detergent solution (Miller et al. 1988).Moreover, comparable features have also been detectedduring different steps of the Giemsa chromosome stainingby AFM (Musio et al. 1997). In spite of these large roundaggregates, the 10 nm fibers chromatin network, detectablein clean areas of the samples by high-magnificationFEISEM analysis, is completely comparable with the net-work structure that can be seen accurately on untreatedchromosomes selected without coating (Rizzoli et al.1994). For this reason, this kind of cleaning treatment canbe considered to preserve wide areas of the native chro-mosomal fiber architecture.

After pure protease K treatment, the homogeneous phaseappears to be scattered at the periphery border of eachchromatid. Both AFM and low-magnification FEISEMimages show the presence of a kind of framework along themajor axis of each chromatid. The origin of this framework,better observed by high-magnification FEISEM imagesand consisting mainly of enlarged fibers of about 15–20nm, cannot be explained completely. The rarified circum-scribing network suggests that this kind of fiber could bethe result of either the protease K exposition of more con-densed fibers in the inner part of the chromosome or achemically induced collapse of the 10 nm fibers.

However, the randomly contemporary presence of boththe 10 nm fibers and the enlargement of the meshes of thenetwork, together with the evidence of the strong action ofthe treatment on the chromosomal coat, lead us to believethat the second hypothesis is more likely. In this case, thefine native structure of the chromosomes and the recipro-cal spatial arrangement of the DNA fibers are substan-tially respected at lower magnification observations, but notat the higher resolution level allowed by the FEISEM.

The behavior of the coating layer under the effects of thecleaning solutions we utilized suggests the presence of a rel-evant protein component. Moreover, the finest ultrastruc-tural spatial arrangement of the 10 nm DNA fibers appearsto be independent of the chemical removal of the proteins.

P. Gobbi et al.: AFM and FEISEM imaging of human chromosomes 279

TABLE I Measurements of chromosome height in dry and wet con-dition and their variation related to the hydration status

Height increase

Mean height Mean height after dry wet rehydration

(nm) (nm) (%)

Untreated chromosomes 85 239 281 Chromosomes after

hot mix solution 97 166 171 Chromosomes after

protease K 100 165 165

Further information about the role of the chromosomecoat has been obtained by the constant force mode AFManalysis that allows to observe samples at different degreesof hydration (Fritzsche and Henderson 1997b, Stark et al.1998). Our results on dehydrated and rehydrated chromo-somes show differences in the mean increase of the chro-mosomal size after the rehydration. This increase of heightand volume of each chromatid is related to the treatmentand therefore to the amount of the coating layer, indicat-ing that this mantle on the chromosomes can modify theirmechanical and electrostatic characteristics related to thehydration status (Fritzsche and Henderson 1997b, Stark etal. 1998). Since this coat is frequently, but not always,observed on standard untreated chromosome preparations(Rizzi et al. 1995, Rizzoli et al. 1994), it is commonlythought that its origin is independent of chromosomalstructure. This coat seems to be applied, probably duringthe preparative procedure, to the chromosomal surfacewith an inconstant strength of bond, as also suggested bythe critical analysis of the physical cleaning procedureproposed by Martin et al. (1994).

The effects of rehydration on chromosomes, treated withpure proteolytic solutions or with detergent plus protease,are quite similar. In fact, the increase of the untreated chro-mosome height after rehydration is of about three-fold com-pared with the dry condition, while after the treatment fortotal or partial coat removal, this increase is about 1.5-fold.The reason for this evident difference due to the presenceof the coat during the rehydration can be related to thebehavior of the “biological colloids” in watery solutionthat modify in a nonlinear mode their volumes in the tran-sition throughout different hydration status (Kellenberger1987). In our experimental conditions, this coat can alter thehydration shell of the DNA and the effect of the water sur-face tension on the whole sample. Moreover, the behaviorof the rehydrated treated samples described by AFM rulesout a strong anchorage between the surface of the chromatidand the ITO glass mediated by the proteins of the coat, assuggested by the flattened FEISEM images of dry undi-gested samples. On the contrary, this protein layer con-tributes to increasing the thickness of the chromatids dur-ing the rehydration (239 vs. 165 nm) (Fritzsche andHenderson 1997b, Kellenberger 1987, Stark et al. 1998).

It is well known that many banding methods can beused to identify individual chromosomes (i.e., the G-band-ing) but, although these techniques are used routinely forconventional cytogenetic investigations, many of the mech-anisms involved are not completely understood (Musio etal. 1997). A longitudinal pattern similar to the G-bandingcan be clearly detected by AFM on the chromosome sur-face after adequate aging of the maps, or during the treat-ments required to perform the G-banding that involves aproteolytic digestion of the sample (Ali Ergun et al. 1999;Mariani et al. 1994; Musio et al. 1994, 1997). Under ourconditions, samples were dehydrated and dried shortlyafter the spontaneous air drying of the methanol-aceticacid fixative: a G-banding-like pattern is not detectable,

except for the chromosomes treated with the solution withNa-acetate, protease K, and SDS and rehydrated; how-ever, these chromosomes are not completely coat-free.

This finding strongly suggests a fundamental role ofthe coat and its interaction with the fibers of specific chro-mosomal region in the modeling of the G-banding-likestructure previously described. Moreover, our results cancontribute to explaining the necessity for a long aging andgentle digestion to obtain the typical and repetitive band-ing of the chromosomes after the rehydration of the sam-ple by means of the Giemsa watery solution. We feel thatthis treatment results both in a degradation and a partialremoval of the coat.

In the rehydrated samples, the chromatids of treatedmaps show a cylindrical shape with a diameter of about 160nm. This volume is smaller than similar measurementsobtained with conventional SEM observations of digestedand undigested chromosome preparations that result in achromatid diameter of about 350 nm (Harrison et al. 1981,Sumner 1991). Nevertheless, it has been demonstrated thateach preparative step for conventional SEM increases thechromosomal dimension (Rizzoli et al. 1994). Moreover,the present data are in agreement with both previousFEISEM observations (Rizzi et al. 1995) and similar AFMmeasurements recently reported by Musio et al. 1997 andAli Ergun et al. 1999. For these reasons we suggest that theestimate of the real dimension of chromatids, evaluatedwith these different techniques and under these conditions,requires a critical revision.

Therefore, the presence of the biological coating on themetaphase spreads has to be considered not only for theirsurface aspect but also for their mechanical and electrostaticcharacteristics. Consequently, any possible treatment hasto be carefully analyzed for the interaction with structuralor external molecules and for their role in chromosomalresponse to watery solution. Different final volumes of thechromosomes can, in fact, alter the reciprocal position ofspecific domains with respect to the native chromosomalstructure. This could have marked differences when, forexample, multiple fluorescent in situ hybridization proce-dures are performed.

Conclusion

The data of the two microscopes are completely super-imposable and complementary. The FEISEM offers thepossibility to observe directly the chromosome fibers atnanometric resolution and the AFM is able to operate in liq-uid environment. In this way, it was possible to describe theeffects of different chemical treatments utilized to removethe protein layer that normally covers the chromosomesurface and to evaluate the role of this coat in the inter-pretation of the biological results. In any case, the high-res-olution level, necessary to obtain a precise physical map-ping of the genome that new instruments such as FEISEMand AFM could offer, requires homogeneously cleaned

280 Scanning Vol. 22, 5 (2000)

samples with a high grade of reproducibility. A correlativemicroscopic approach that utilizes completely differentphysical probes provides complementary useful informa-tion for the understanding of the biological, chemical, andphysical characteristics of the samples and can be appliedto optimize the chromosome preparations for furtherimprovement in the knowledge about the spatial organiza-tion of the genome.

Acknowledgments

The authors are grateful to Dr. A. Galanzi and Ms. T.Drobeck for skillful technical assistance and suggestions.Mr. Marcello Maselli is also thanked for the photographicwork.

References

Ali Ergun M, Tan E, Sahin FI, Menevse A: Numerical chromosomalabnormalities detected by atomic force microscopy. Scanning21, 182–186 (1999)

Binnig G, Quate CF, Gerber C: Atomic force microscopy. Phys RevLett 56, 930–933 (1986)

Fritzsche W, Henderson E: Chicken erythrocyte nucleosomes havea defined orientation along the linker DNA—a scanning forcemicroscopy study. Scanning 19, 42–47 (1997a)

Fritzsche W, Henderson E: Mapping elasticity of rehydrated meta-phase chromosomes by scanning force microscopy. Ultrami-croscopy 69, 191–200 (1997b)

Gobbi P, Falconi M, Vitale M, Galanzi A, Artico M, Martelli AM,Mazzotti G: Scanning electron microscopic detection of nuclearstructures involved in DNA replication. Arch Histol Cytol 62,317–326 (1999)

Harrison CJ, Britch M, Allen TD, Harris R: Scanning electron micro-scopy of the g-banded human karyotype. Exp Cell Res 134,141–153 (1981)

Heckl WM: The combination of AFM nanodissection with PCR. Anew tool for the generation of genetic probes. Bioforum Intl 9,2–5 (1998)

Hermann R, Müller M: Towards high resolution SEM of biologicalobjects. Arch Histol Cytol 55, 17–25 (1992)

Kellenberger E: The response of biological macromolecules andsupramolecular structures to the physics of specimen cry-opreparation. In Cryotechniques in Biological Electron Micro-scopy (Eds. Steinbrecht RA, Zierold K). Springer-Verlag, BerlinHeidelberg (1987) 35–63

Lattanzi G, Galanzi A, Gobbi P, Falconi M, Matteucci A, Breschi L,Vitale M, Mazzotti G: Ultrastructural aspects of the DNA poly-merase α distribution during the cell cycle. J HistochemCytochem 46, 1435–1442 (1998)

Leuba SH, Bustamante C, van Holde K, Zlatanova J: Linker histonetails and N-tails of histone H3 are redundant: Scanning forcemicroscopy studies of reconstituted fibers. Biophys J 74,2830–2839 (1998)

Mariani T, Musio A, Frediani C, Sbrana I, Ascoli C: An atomic forcemicroscope for cytological and histological investigations. JMicrosc 176, 121–131 (1994)

Martin R, Busch W, Herrmann RG, Wanner G: Efficient preparationof plant chromosomes for high-resolution scanning electronmicroscopy. Chromosome Res 2, 411–415 (1994)

Miller SA, Dykes DD, Polesky HF: A simple salting out procedurefor extracting DNA from human nucleated cells. Nucleic AcidsRes 16, 1215 (1988)

Musio A, Mariani T, Frediani C, Sbrana I, Ascoli C: Longitudinal pat-terns similar to G-banding in untreated human chromosomes:Evidence from atomic force microscopy. Chromosoma 103,225–229 (1994)

Musio A, Mariani T, Frediani C, Ascoli C, Sbrana I: Atomic forceimaging of chromosome structure during G-banding treatments.Genome 40, 127–131 (1997)

Nagatani T, Saito S, Sato M, Yamada M: Development of an ultra highresolution scanning electron microscope by means of a fieldemission source and in-lens system. Scan Microsc 1, 901–909(1987)

Pawley J: The development of field-emission scanning electronmicroscopy for imaging biological surfaces. Scanning 19,324–336 (1997)

Rizzi E, Falconi M, Baratta B, Manzoli L, Galanzi A, Lattanzi G,Mazzotti G: High-resolution FEISEM detection of DNA cen-tromeric probes in HeLa metaphase chromosomes. J HistochemCytochem 43, 413–419 (1995)

Rizzoli R, Rizzi E, Falconi M, Galanzi A, Baratta B, Lattanzi G, VitaleM, Manzoli L, Mazzotti G: High resolution detection ofuncoated metaphase chromosome by means of field emissionscanning electron microscopy. Chromosoma 103, 393–400(1994)

Sanchez-Sweatman OH, de-Harven EP, Dube ID: Human chromo-somes: Evaluation of processing techniques for scanning elec-tron microscopy. Scanning Microsc 7, 97–104 (1993)

Stark RW, Thalhammer S, Wienberg J, Heckl WM: The AFM as atool for chromosomal dissection—the influence of physicalparameters. Appl Phys A 66, S579–S584 (1998)

Sumner AT: Scanning electron microscopy of mammalian chromo-somes from prophase to telophase. Chromosoma 100, 410–418(1991)

Sumner AT: Problems in preparation of chromosomes for scanningelectron microscopy to reveal morphology and to permitimmunocytochemistry of sensitive antigens. Scan Microsc 10,165–176 (1996)

Sumner AT, Ross A: Factors affecting preparation of chromosomesfor scanning electron microscopy using osmium impregnation.Scan Microsc (suppl) 3, 87–97 (1989)

Sumner AT, Ross AR, Graham E: Preparation of chromosomes forscanning electron microscopy. Methods Mol Biol 29, 41–50(1994)

Tamayo J, Miles M, Thein A, Soothill P: Selective cleaning of the celldebris in human chromosome preparations studied by scanningforce microscopy. J Struct Biol 128, 200–210 (1999)

Thalhammer S, Stark RW, Müller S, Wienberg J, Heckl WM: Theatomic force microscope as a new microdissecting tool for thegeneration of genetic probes. J Struct Biol 119, 232–237 (1997)

Ushiki T, Hitomi J, Ogura S, Umemoto T, Shigeno M: Atomic forcemicroscopy in histology and cytology. Arch Histol Cytol 50,421–431 (1996)

Wanner G, Formanek H: Imaging of DNA in human and plant chro-mosomes by high-resolution scanning electron microscopy.Chromosome Res 3, 368–374 (1995)

P. Gobbi et al.: AFM and FEISEM imaging of human chromosomes 281