atomic force microscopy: a new way to look at chromatin

Michael J. Allen Digital Instruments Inc., Santa Barbara, CA

fter several decades of modem re- A search approaches to chromatin structure, there are many important details that we do not yet know. Most notably, there are gaps in our knowledge of the dynamic interplay between chromatin structural states and the functional states of genes. For example, while a clear high-resolution picture of the nucleosome structure has been established, it remains unclear how nucleosomes, which are packaged into higher-order structures, interact with transcription factors and are read through by RNA polymerases. Clearly, a technique that could address questions regarding the structural dynamics of chromatin would provide clues as to how certain important biological processes work in the cell, such as the regulation of genes.

This article deals with the application of a new form of microscopy, atomic force microscopy or AFM [l], to the study of chromatin structure. The main advantage of AFM is that it can operate at sub- nanometer resolution in aqueous fluids [2, 3). It has been applied successfully to study of the structural dynamics of cells [4] and molecules [S, 61. However, thus far most of the AFM work on chromatin has been limited to studies performed in ambient air environments. Even when working in ambient air, AFM offers a number of important advantages over other microscopic techniques: (1) sample preparation is sim- ple, e.g., no stains or metal coatings need to be applied to the specimen; (2) without coatings or stains present, AFM measurements can be made directly on the natural surface of the specimen; (3) imaging can be performed in humid environments where biological specimens remain hydrated with bound water molecules; and (4) the spatial resolution of AFM is suffi- cient to allow clear visualization of individual nucleosomes and linker DNA.

Somatic Chromatin: Nucleohistone Structure

The histones and the nucleosomal structures somatic chromatin form when

complexed with DNA have been well characterized using a .variety of methods, including x-ray diffraction [7], neutron scattering [8,9], electron microscopy [lo, 111, as well as by biochemical means [12]. Since the early 1970s, the nucleosome model has been established as the elemental subunit structure of somatic cell chromatin in eukaryotes [13]. The nucleosome core particle consists of 146 base pairs (bp) of DNA wrapped 1.75 times around a disc- shaped inner protein core. The inner protein core is an octameric structure composed of two molecules of each of the four core histones (H2A, H2B, H3 and H4). The octameric structure is formed by combination of one H3-H4 tetramer ([H32-H42]) and one H2A-H2B heterodi- mer ([H2A-H2B]2). The nucleosome core particle has an overall diameter of 11 nm and thickness of 5.5 nm [7-91. Nucleosome cores are interconnected by linker DNA, typically 49 bp in length; therefore, the average total DNA repeat length for each nucleosomal unit is 195 bp [14]. A fifth protein, histone H1 or histone H5, binds the nucleosome core and linker DNA forming the chromatosome structure, which contains 168 bp DNA.

Nucleosomes arranged linearly into the classical 11 nm wide “beads-on-a- string” structure can assemble zn ~z t ro into a30nmwidesolenoidfiber [15,16].Mod- eling of such 30 nm fibers indicates a left- handed helix of 6-7 nucleosomeslturn coiled tightly to give a helical pitch of 11 nm. Biochemical and electron microscope evidence suggests histone H1 is a mediator in the higher order folding of chromatin, including formation of the 30 nm fiber [17,18]. The influence of the ionic strengths of mono- and di-valent cations have also been shown to positively influence higher-order chromatin folding and 30 nm fiber formation [19]. The 30 nm fiber is hypothetically the chromatin structure forming 20-85 kbp loop domains in vivo [20]. The local structure of the chromatin within loop domains appears to be

34 IEEE ENGINEERING IN MEDICINE AND BIOLOGY 0739-51 75/97/.$10.0001997 Marth/Apri l l997

dependent on the transcriptional activity of the DNA [17]. Transcriptionally active chromatin appears unfolded, to the extent that it is nearly fully extended and resem- bles the “beads-on-a-string’’ conforma- tion [21].

Sperm Chromatin: Nucleoprotamine Structure

Protamine is the predominate DNA packaging protein in sperm. Human protamines and a small amount of their pre- sursors comprise about 80% of the nucleoprotein in human sperm. Protamine amino acid sequences have been publish- ed [22, 231. Mammalian protamines are small proteins, about 50 amino acids in

length, consisting of a long central arginine-rich region flanked by a short highly conserved cysteine-containing amino terminal domain and a short car- boxyl terminal domain also containing cysteine. Modeling suggests that protamine 1 binds tightly into one of the grooves of DNA through arginine-phos- phate salt bridges [24]. Zinc ions may also play a role in protamine binding to DNA and stabilization of higher-order structures [25]. Charge neutralization of the DNA backbone and inter di-sulfide bond formation between neighboring protamine molecules likely induce DNA condensation. It has been estimated that the DNA of mammalian sperm is six-fold

1. (A-C) AFM images of three reconstituted chromatin fibers with increasing histone octamer loadings: (A) 7; (B) 12; and (C) 18 nucleosome cores. The chromatin fiber contour lengths were: (A) 1164 nm; (B) 852 nm; and (C) 693 nm. The arrows in (B) point to the six unfilled nucleosome positions. (D) shows the effect of histone octamer loading on chromatin fiber contour length, as determined by AFM. Circles represent AFM measurements showing the compaction of the linear DNA sequence induced by increasing levels of histone octamer loading. The solid line represents the expected compaction of the DNA sequence (3780 bp) induced by the packaging of 146 bp DNA into nucleosome cores.

more compacted than the DNA in mitotic chromosomes [261. Consequently, such dense DNA packaging can not be ex- plained by a nucleosomal-like organization similar to that observed for somatic chromatin. However, the fundamental structure and higher order packaging in sperm chromatin have not been clearly elucidated.

Atomic Force Microscopy AFM analyzes surfaces with a micro-

fabricated silicon probe-tip that makes soft local contact with the specimen. The AFM probe is end-mounted to a flexible cantilever-arm, and raster-scanned across the sample under piezo-electric control. Very small forces, in the pN range, acting on the probe-tip are measured with the AFM [5,27].

Contact scanning is the traditional AFM imaging mode. Here, the AFM probe is scanned across the sample with the cantilever held at a constant deflection. An ultrasensitive laser system moni- tors cantilever deflection, while digital feedback control maintains probe-sample forces at a constant value and prevents appreciable loss of probe contact with the surface. For very smooth surfaces, such as sheets of bacterial membrane proteins, lateral resolutions beneath 1 nm and vertical resolutions of approximately 0.1 nm have been achieved [2 , 31.

A new oscillation imaging mode de- veloped recently (TappingModeTM, Digi- tal Instruments, Santa Barbara, CA) can be operated in both air and fluids [28,29]. In this tapping mode, the probe and inte- grated cantilever are oscillated up and down at resonance by a special frequency synthesizer. For air imaging, the frequency of such oscillations is usually 300 kHz and amplitudes are controlled typically between 10-100 nm. In fluid, the cantilever oscillation frequencies and amplitudes used are approximately 10-fold lower than those used in air. Changes in the amplitude of the cantilever oscillation due to the scanning probe-tip’s intermit- tent interaction with the sample surface are monitored and kept constant using an electronic feedback circuit. Accurate topographic images are generated using the amplitude change as the feedback sig- nal [30]. The key advantage of tapping mode is the elimination of lateral shear forces present in traditional contact scanning, i.e., only very brief, highly con- trollable, vertical forces are applied. For many specimens, lateral shear forces

Morth/Aprill997 IEEE ENGINEERING IN MEDICINE AND BIOLOGY 35

cause excessive perturbation, movement or damage to the structure being imaged, and tapping has helped overcome this problem [3 11. As an extension to tapping mode, monitoring the phase-angle lag of the AFM cantilever oscillation allows probe-tip detection of local sample proper- ties other than topography (e.g., adhesion, viscoelasticity, magnetic fields, etc).

A F M A New Approach to Chromatin Structure

AFM is emerging as a useful tool for investigating chromatin structure, and some of the questions addressed include: (1) quantitative measurements of the extent of DNA packaging into nucleosomes [32], (2) DNA condensate structures in native [33, 341 and reconstituted [35] mammalian sperm chromatin, (3) higher- order structuring and the influences of ionic strength variations on native fibers [36, 371, (4) rigorous testing of AFM results against existing models for regular solenoidalpackaging [38], (5) in situ studies [39-421, (6) nanodissection of DNA segments along individual chromosomes [43, 441, and (7) measurements of chromatin volume variations in mammalian sperm [45, 461 and metaphase chromosomes [41]. Several of these applications are presented here in more detail below.

Chromatin structure investigations to be addressed by AFM in the future include: (1) protein/chromatin interactions (e.g., binding and effect of transcription factors, polymerases, etc), (2) effects of histone chemical modifications (e.g., acetylation), (3) effects of ionic strength var- ation by mono and divalent metal cations, (4) real-time dynamics in fluid, and (5) technical aspects, (Le., improved sample adsorption techniques and increased use of high resolution fluid imaging).

AFM of Reconstituted Chromatin Fibers

As seen in Fig. 1, AFM measurements made on reconstituted nucleosome cores show a compaction of DNA by histone octamers, consistent with 146 bp of DNA wrapped 1.75 tums about the histone octamer to form the 11 nm nucleosome core [32]. The chromatin was reconstituted using isolated histone octamers and a 3.78 kilo base pair (kbp) DNA fragment containing 18 tandem-repeats of a 208 bp nucleosome- positioning sequence. The chromatin was imaged by AFM in an ambient air environment following adsorption from a 10 mM NaCl solution onto a cover glass substrate.

36

2. (A) AFM image of a 2.5 kilo base pair linearized plasmid DNA sequence adsorbed to mica. (B) A diluted solution containing an 18 kDa histone H1-like bacterial protein was co-incubated with the same DNA as in (A), following pre-adsorption of the DNA to mica. The AFM image in (b) shows the complex formed by the binding of three protein molecules to the DNA. The DNA is bent and coiled slightly in the region of protein binding. Loosely binding DNA to mica prior to addition of DNA binding protein prevents sample aggregation and allows the imaging of individual complexes. Scale: bars =lo0 nm.

3. (A, B) AFM scans of H1+ native chromatin fibers. The partially open higher-order structures were held at 10 mM NaCl during fixation prior to imaging.The nucleosome conforms to a “zig-zag” type array along the fiber axis. @’) 3D computer modeling of the AFM data-image in B reveals the average edge-to-edge distance between adjacent nucleosomes to be equivalent to the long axis of the core particle (11 nm).

IEEE ENGINEERING IN MEDICINE AND BIOLOGY Morth/Aprill997

With increasing loading of histone octamers on the 208-18 DNA sequence, the lengths of the chromatin fibers decreased at an incremental rate consistent with 146 bp of DNA packaged into nucleosome cores (Fig. 1D). Figure 1B displays an

AFM image of a 208-18 chromatin fiber in which 12 ofthe 18 nucleosome-locating positions are filled by nucleosome cores. The remaining six unfilled positions can be identified by their spacings along the chromatin fiber (see arrows).

4. (A-F) In situ AFM imaging of chromatin spread from lysed chicken erythrocyte nuclei. (A) Single erythrocyte after lysis and treatment with 100 mM NaCI. Mem- brane (mem), location of the nuclear region (nuc), and released chromatin (chr) are indicated (Bar = 5 pm), Z range = 65 nm). (B) Zoom of the chromatin fibers (cfl in a subregion from A (arrow) (Bar = 500 nm, Z range =lo nm). (C) Erythrocytes after detergent treatment. The central region of cell debris is surrounded by a dense chromatin layer (ch) (Bar = 5 pm, Z range = 80 nm). (D) A zoom of the chromatin layer in (C) reveals a granular substructure. A single supranucleosomal chain (cf) is detected (Bar = 1 pm, Z range = 25 nm). (E) Supranucleosomal fiber morphology at higher resolution (Bar = 200 nm, Z range = 10 nm). (F) extended nucleosomal chains seen following different lysis procedure (Bar = 200 nm, Z range = 10nm). Re- printed by permission (from, Fritzsche, et al., (1994) Probing chromatin with the scanning force microscope. Chromosoma 103,231-236).

Center-to-center measurements between adjacent nucleosome cores along the chromatin fiber axis are not affected by probe-sample shape convolutions. For those cores directly adjacent along a chromatin fiber, center-to-center measurements of inter-core spacing averaged 37.2 nm (n=36, +/- 8.8 nm). The accuracy of such center-to-center measurements is limited mainly by the AFM piezo-electric scanner, usually calibrated to the atomic spacings of graphite and mica. Conse- quently, AFM studies involving the effects of various proteins and their modifications (e.g., histone acetylation) on inter-core spacing, using reconstituted or native chromatin samples, can be con- ducted with detection limits on the sub- nanometer scale.

In other AFM experiments, a bacterial histone H 1 -like protein isolated from Chlamydia trachomatis was bound to linearized plasmid DNA following immo- bilization of the DNA onto mica (Fig. 2). This method of adding protein to mica-im- mobilized DNA is particularly useful when studying such hydrophobic protein- DNA complexes [35]. Large, dense aggre- gates are formed when the DNA and protein are mixed in free solution, thereby preventing the study of individual complexes by microscopy or by other techniques requiring soluble complexes (e.g., gel electrophoresis). Bustamante and co- workers have successfully imaged a number of isolated protein:DNA complexes using AFM, including heat-shock transcription factor 2: plasmid DNA structures [47].

AFM of Native Chromatin Fibers Higher-ordered H1+ native chromatin

fibers (Fig. 3) were imaged by AFM under the same conditions as the reconstituted H1- nucleosome core fibers (fixation at 10 mM NaC1, adsorption to cover glass, and ambient air imaging) [36]. Clearly, the nucleosomes seen in AFM images of the native chromatin fibers are three- dimensional higher-order structures re- sembling a “zig-zag” type array (Fig. 3). Such “zig-zag” arrangements have been observed by EM in the low-salt chromatin structures generated following cell lysis [ 181. AFM measurements of 50 center-to-center nucleosome spacings for the H1+ native chromatin fibers averaged 22 nm. H1 -depleted native chromatin ap- peared less structured and spacings of the nucleosomes increased to 32 nm (not shown). Not only do the AFM images

Marth/April l997 IEEE ENGINEERING IN MEDICINE AND BIOLOGY 37

confirm the existence of the “zig-zag” structures, they also show that 3D infor- mation about the structures can be ac- cessed by AFM. From the images of the native chromatin held at 10 mM NaCl during fixation, we can assign, with good accuracy, the relative positions of nucleosomes within individual fibers. Computer modeling of a 10 mM NaCl higher-order chromatin fiberreveals an 11 nm edge-to-edge nucleosome spacing (Fig. 3B’).

Other AFM studies have revealed the existence of low-salt chromatin fibers ex- hibiting “irregular three dimensional ar- rays’’ of nucleosomes, not “zig-zag” structures [37, 381. The irregular structures were encountered far more often in the samples than the “zig-zag” structures. As suggested by Leuba, et al. [37], these results have provided good cause for the re-thinking of canonical models showing regular nucleosomal packaging into highly repetitive structures.

Unlike the AFM images of reconstituted chromatin, the AFM data for the DNA linker regions in the native fibers are poorly resolved and only a “smudge” can be detected in these regions (Fig. 3). We have only been able to resolve clearly the linker DNA on extended reconstituted chromatin fibers (Fig. 1) and for reconstituted mononucleosomes where the DNA extends out away from the nucleosome core (data not shown). For isolated fixed chromatin held at higher ionic strengths (25-100 mM NaCl or 0.5 mM MgC12) the fibers appear to coil into large “super-bead”- like structures, and if extended, the fibers exhibit very little detectable substructure (i.e., nucleosomes; data not shown). Another problem encountered in these studies is that of sample adsorption. Specifically, nucleosomes within a fiber appear to adsorb, preferentially, “face- down” or lengthwise onto the AFM substrate (see Figs. 1A-C, 3B’). Sample preparation and adsorption techniques that minimize such effects and preserve the 3D solution structure will be very useful.

In Sifu AFM Studies of Chromatin and Chromosomes

Figure 4 displays a number of AFRL images of chromatin fibers spread from lysed chicken erythrocytes and imaged in air [40]. These results clearly demonstrated the ability of AFM to: (1) distin- guish the chromatin material from other

cell structures (e.g., the nucleus and membrane), and (2) characterize the supranucleosomal fiber structures at relatively high resolution in ambient air. Although AFM resolution was adversely affected by the motion of the chromatin material in the fluid environment, Fritzsche, et al. [41], have also reported AFM images of rehy-

drated chromatin spread from B lympho- cytes.

Chromosome imaging and dissection using a single AFM probe is demonstrated in Fig. 5 [43]. Two parallel 150 nm wide incisions through the chromosome can be seen in Fig. 5B. The incisions were made in a single-pass scan of the AFM probe

5. (A,B) AFM images of a salivary chromosome region (A) before and (B) after dissection using the AFM probe. (A) AFM image of the tip of the X chromosome prior to dissection. (B) A subregion was selected and precisely scored with the AFM tip, and chromatin material removed by “scooping” onto the AFM probe. The parallel score lines are each approximately 150 nm wide and are separated by 840 nm. The material “scooped” represents approximately 70 kilo base pairs of DNA. Reprinted by permission (from, Vesenka, et al., (1995) Combining optical and atomic force microscopy for life sciences research. BioTechniques 19,240-253).

Position Sensitive Detector

Cells or Molecules

Sample or Specimen Support

6. Schematic of AFM fluid cell chamber.

38 IEEE ENGINEERING I N MEDICINE AND BIOLOGY Marth/Aprill997

7.3D rendering of a fluid tapping-mode AFM image of a demembranated bull sperm nucleus in water. The characteristic size and “paddle” shape of the nucleus are displayed. Nuclear thickness increases at the posterior end (far left) where the tail normally attaches.

and by increasing the force load of AFM probe against the chromosome during the

matin can then be recovered from the AFM probe and biochemically analyzed.

incision. Half of the 840 nm wide dissected “strip” was removed by a “scooping” method, which results in the chromatin

AFM Volume Measurements of Sperm Chromatin

Previous volume measurements made on mouse sperm nuclei using serial-sec-

material being collected onto the AFM probe. The DNA component of the chro-

tion electron microscopy [48] suggested that sperm DNA is packed so densely inside the nucleus that there would be no room for extensive hydration of the chromatin. Because of its ability to measure specimen thickness in fluid (Fig. 6), we used AFM to measure the nuclear volumes of intact bull and mouse sperm heads and demembranated sperm nuclei, both in the fully hydrated (Fig. 7) and dehydrated states [46]. Data were ob- tained by analyzing a small population of cells/nuclei (Table 1) or by performing repeated measurements on single cells imaged following the addition of increasing concentrations of propanol (Fig. 8). As an example, Fig. 7 shows a tapping-mode AFM image of a demembranated bull sperm nucleus in fluid. The bull sperm nucleus exhibits a characteristic “paddle” shape, with an increase in thickness seen at the posterior end where the tail normally is attached. The results show that the volume of fully hydrated, intact sperm heads and demembranated sperm chromatin particles are at least twice the volume of their air-dried counterparts. Dehydra- tion occurs rapidly in air, and the reduc- tion in volume of chromatin induced by water loss appears to be completely re- versible. These studies demonstrate that sperm chromatin is extensively hydrated in the native state and is not as compact as previous studies have suggested. Looser packing of the chromatin in sperm would

Morth/Aprill997 IEEE ENGINEERING IN MEDICINE AND BIOLOGY 39

Instruments. Thanks to Bobbie Offen, Drs. Eric Henderson and Wolfgang Fritzsche for their contributions.

8. Effect of dehydration on sperm nuclear volume as measured by AFM. Upper pair of horizontal lines marks the range of volume determined for air-dried mouse sperm nuclei. The lower pair show the range in volumes determined for mouse sperm nuclei by serial section EM.

allow ions and small proteins to access sites buried within the nucleus. Access to these internal sites might be critical to the proper unpackaging and disassembly of the sperm chromatin, or to gene activation events in the fertilized egg.

Conclu~ing Remarks High-resolution physical techniques

such as x-ray crystallography have rle- vealed the nucleosome’s structure on the nm scale [7]. However, much less is known about the dynamic nature of nucleosomes and assembled structures in the cell-particularly chromatin structural changes influenced by ionic strength, histone chemical modifications, cell cycle proteins, transcription events, and DNA- binding drugs. Regarding mammalian sperm chromatin, the elemental subunit and higher order packaging structures are not known. Determining the structure of the chromatin in sperm is important to o w understanding of the mechanisms iii- volved in DNA condensation, infertility, genomic imprinting, and early events im

the egg following fertilization. AFM is a research tool that provides a

new way to image and measure uncoated

chromatin structures at nanometer-level resolution in ambient air and fluid environments. As shown here, AFM has been successfully applied to studying reconstituted [32] and native chromatin fibers [36- 421, chromosomes [43, 441, and sperm chromatin [34, 35,451. More clinical and biomedical applications of A M , such as chromosome karyotyping [49], chromosome mapping [43,44], and human sperm head shape classification and abnormalities [SO] have met preliminary success. Therefore, although it is new, AFM has already proved useful in chromatin research. In the future it is anticipated that AFM images acquired under aqueous fluid will become more common and provide exciting new applications and infor- mation regarding chromatin structure and function.

Acknowledgments 1 gratefully acknowledge Drs. Rod

Balhorn and E. Morton Bradbury for their guidance and support. Major portions of this work were supported by DOE grant no. W-7405-ENG-48 to R. Balhom and .DEFG88ER60673, and NIH grant no. GM-26901 to E.M. Bradbury, and Digital

Michael J. Allen earned a B.Sc. in Microbiology from the University of Illinois, Urbanain 1987. He was then employed for s ix years as Biomedical Scientist at the Lawrence Liver- more National Labora-

tory in Livermore, California, where he became involved with developing scanning tunneling (STM) and atomic force (AFM) techniques for biological samples. Dr. Allen continued working with AFM during doctoral studies at the University of California, Davis, where he earned a Ph.D. in Cell Biology in 1995. Recently, Dr. Allen joined Digital Instruments lo- cated in Santa Barbara, California, as Bio- logical Applications Scientist.

Address for Correspondence: Dr. Mi- chael J. Allen, Digital Instruments, 520 E. Montecito St., Santa Barbara, CA 93103. E-mail: [email protected]

References 1. Binnig G, Quate CF & Gerber C (1986): Atomic force microscope. Phys. Rev. Lett. 56:

2. Schabert FA, Henn C & Engel A (1995): Native Escherichia coli OmpF porin surfaces probed by atomic force microscopy. Science 268; 92-94. 3. Mou J, Czajkowsky DM, Sheng S, Ho R & Shao Z (1996): High resolution surface structure of E. coli GroES oligomer by atomic force microscopy. FEBS Lett. 381,161-164. 4. Henderson E, Haydon PG, Sakaguchi DS (1992): Actin filament dynamics in living glial cells,imaged by atomic force microscopy. Science 257: 1944.1946. 5. Muller DJ, Buldt G & Engel A (1995): Force- induced conformational change of bacteriorho- dopsin. J. Mol. Biol. 249: 239-243. 6. Radmacher M, Fritz M, Hansma HG & Hansma PK (1994): Direct observation of en- zyme activity with the atomic force microscope. Science 265: 1577-1579. 7. Finch JT, Lutter LC, Rhodes D, Brown AS, Rushton B, et al. (1977): Structure of nucleosome core particles of chromatin. Nature 269,29-36. 8. Bradbury EM, Baldwin JP, Carpenter BG, Hjelm RP, Hancock R & Ibel K (1975): Brook- haven Symp. in Biol. 27, (1V) 97. 9. Suau, P, Kneale GG, Braddock GW, Bald- win, JP & Bradbury EM (1977): A low resolution model for the chromatin core particle by neutron scattering. Nucl. Acids Res. 4,3769-3786. 10. Olins AL & Olins DE (1974): Spheroid chromatin units (nu bodies). Science. 183,330-332.

930-933.

40 IEEE ENGINEERING IN MEDICINE AND BIOLOGY MarchlAprill997

mailto:[email protected]

11. Woodcock CLF (1973): Ultrastructure of in- active chromatin. J. Cel/ Biol. 59,368~1. 12. Hewish DR & Burgoyne LA (1973): Chro- matin substructure. The digestion of chromatin DNA at regularly spaced sites by a nuclear de- oxyribonuclease. Bioch. Biophys. Res. Comm. 52,

13. Kornberg R (1974): Chromatin structure: A repeating unit of histones and DNA. Science 184,

14. Felsenfeld G (1978): Chromatin. Nature 271,

15. Carpenter BG, Baldwin JP, Bradbury EM & Ibel K: (1976) Organization of subunits in chromatin. Nud. Acids Res. 3, 1739-1746. 16. Finch JT & Klug A (1976): Solenoid model for superstructure in chromatin. Proc. Nutl. Acad. Sci. USA 73,1897-1901. 17. Felsenfeld G (1992): Chromatin as an essen- tial part of the transcriptional mechanism. Nature 3 5 5 2 19-224. 18. Thoma F, Koller T & Klug A (1979): In- volvement of histone H1 in the organization of the nucleosome and of the salt-dependent superstruc- tures of chromatin. J. Cell Biol. 83,403-427. 19. Christiansen G & Griffith J (1977): Salt and divalent cations affect the flexible nature of the natural beaded chromatin structure. Nucl. Acids Res. 4, 1837-1851. 20. Benyajati C & Worcel A (1976): Isolation, characterization, and structure of the folded inter- phase genome of Drosophila melanogaster. Cell 9,393-407. 21. Bjorkroth B, Ericson C, Lamb, MM & Daneholt B (1988): Structure of the chromatin axis during transcription. Chromosoma 96, 333- 340. 22. McKay DJ, Renaux BS & Dixon GH (1985): The amino acid sequence of human protamine PI. Biosci. Rep. 5,383-391. 23. McKay DJ, Renanx BS & Dixon GH (1986): Human sperm protamines: amino acid sequences of two forms of protamine P2. Eur. J. Biochem. 156,5-8. 24. Balhoru R (1982): A model for the structure of chromatin in mammalian sperm. J. Cell Biol.

25. Kvist U (1980): Sperm nuclear chromatin decondensation ability. An in vitro study on ejacu- lated human spermatozoa. Acta Physiol. Scand.

504-510.

868-871.

115-122.

93,298-305,

(Suppl) 486, 1-24.

26. Ward W. (1993): Deoxyribonucleic acid loop domain tertiary structure in mammalian spermatozoa. Biol. Reprod. 48,1193-1201. 27. Florin EL, Moy VT & Gauh (1994): Adhe- sion forces between individual ligand-receptor pairs. Science 264: 415-417. 28. Elings V & Gurley J (1993): U.S. Patent 5,266,80 1. 29. Hansma PK, Cleveland JP, RadmacherM, Walters DA, Hillner PE, et al. (1994): Tapping mode atomic force microscopy in liquids. Appl. Phys. Lett. 64: 1738-1740. 30. Schabert FA & Rabe JP (1996): Vertical dimension of hydrated biological samples in tapping-mode scanning force microscopy. Biophys.

31. Yip C & Ward MD (1996): Atomic force microscopy of insulin single crystals: direct visualization of molecules and crystal growth. Bio- phys. J. 71, 1071-1078. 32. Allen MJ, Dong XF, O’Neill TE, Yan P, Kowalczykowski SC, et al. (1993): Atomic force microscope measurements of nucleosome cores assembled along defined DNA sequences. Bio- chemistry 32,8390-8396 33. Allen MJ, Bradbury EM & Balhorn R (1996): The chromatin structure of well-spread demembranated human sperm nuclei revealed by AFM. Scanning Microsc. (In press). 34. Allen MJ, Lee C, Lee JD, Pogany GC, Balooch M, et al. (1993): Atomic force microscopy of mammalian sperm chromatin. Chro- mosoma. 102,623-630. 35. Allen MJ, Kaul R, Bradbury EM & Bal- horn R (1996): Reactivity of mica-sequestered DNA and two DNA binding proteins. Proc. ofthe European Congress on Microsc. (In press). 36. Allen MJ, Yau P, Bradbury EM (1994): Application of Atomic Force microscopy to the study of higher order chromatin structure. Molec. Biol. of the Cell 5,211~1. 37. Leuba SH, Yang G, Robert C, Samori B, van Holde K, et al. (1994): Three-dimensional structure of extended chromatin fibers as revealed by tapping-mode scanning force microscopy. PNAS 91: 11621-11625. 38. Yang G, Leuba SH, Bustamante C, Zla- tanova J & van Holde K (1994): Role of linker histones in extended chromatin fibre structure. Nature: Structural Biology 1: 761-763. 39. Martin LD, Vesenka JP, Henderson E,

J. 70, 1514-1520.

Dobbs DL (1995): Visualization of nucleosomal substructure in native chromatin by atomic force microscopy. Biochemistry 34: 4610-4616. 40. Fritzsche W, Schaper A & Jovin TM (1995): Scanning force microscopy of chromatin fibers in air and liquid. Scanning 17,148-155. 41. Fritzsche W, Schaper A & Jovin TM (1994): Probing chromatin with the scanning force microscope. Chromosoma 103,23 1-236. 42. Fritzsche W & Henderson E (1996): Chicken erythrocyte nucleosomes have a defined orientation along the linker DNA- a scanning force microscope study. Scanning (In press). 43. Vesenka J, Mosher C, Schaus S, Ambrosio L & Henderson E (1995): Combining optical and atomic force microscopy for life sciences research. BioTechniques 19,240-253. 44. Rasch P, Wiedemaun U, Wienberg J & Heck1 WM (1993): Analysis of banded human chromosomes and in situ hybridization pattems by scanning force microscopy. Proc. Natl. Acad. Sci. 90,2509-2512. 45. Allen MJ, Hud NV, Lee C, Pogany G, Siek- haus, WJ & Balhorn R (1991): Analysis of sperm nuclear volumes and extent of chromatin compaction by atomic force microscopy. J. Cell Biol. 115,50a. 46. Allen MJ, Lee JD, Lee C & Balhorn R (1996): Extent of sperm chromatin hydration determined by atomic force microscopy. Molec. Re- prod. & Devel. 45,87-92. 47. Wyman C, Grotkopp E, Bustamante C & Nelson CM (1995): Determination of heat-shock transcription factor 2 stoichiometry at looped DNA complexes using scanning force microscopy. EMBO J. 14: 117.123. 48. Wyrobek AJ, Meistrich ML, Furrer R & Bruce WR (1976): Physical characteristics of mouse sperm nuclei. Biophys. J. 16,811-825. 49. McMaster TJ, Winfield MO, Baker AA, Karp A & Miles MJ (1996): Chromosome classification by atomic force microscopy volume measurement. J. Vac. Sci. Technol. B 14, 1438- 1443. 50. Lee JD, Allen MJ & Balhorn R (1996): AFM analysis of chromatin volumes in human sperm with head shape abnormalities. Biol. of Reproduc- tion. (In Press, November issue).

Morth/April1997 IEEE ENGINEERING I N MEDICINE AND BIOLOGY 41

atomic force microscopy: a new way to look at chromatin

Documents