Ultramicroscopy North-Holland
33 (1990) 127-131
127
SCANNING TUNNELLING MICROSCOPY OF NUCLEIC ACIDS AND POLYNUCLEOTIDES P.G. ARSCO’IT and V.A. BLOOMFIELD Department
of Biochemistry
University of Minnesota,
St. Paul, MN 55108, USA
Received 28 February 1990; in final form 8 March 1990
We have obtained high-resolution images of DNA and RNA by scanning determine the conformation, which depends on hydration state, of individual periodicity, packing distances in aggregates, and handedness.
1. Intmduction Biological processes are generally so finely tuned that they cannot be well understood at the molecular level without a topographical map. The twists, turns and surface texture of a DNA molecule determine how it is organized, and how it is recog&ed by the proteins that initiate and regulate transcription of the genetic message. The “morphogenetic code”, if there is one, has not yet been deciphered, but there are indications that critical binding sites on the DNA may be no more than three base pairs long, a fraction of a turn of the helix. Structural details smaller than about 1 nm cannot generally be resolved by electron microscopy, and optical diffraction techniques do not aBow one to look at specific sites on individual molecules. The scanning tunnelhng microscope (STM) offers a unique advantage in allowing a close look at DNA and small ligands in the absence of stains, shadows, and labels. There is no risk of radiation damage as there is in electron microscopy, and few other problems of specimen preservation when the harsh conditions of a vacuum are avoided. The instrument can be operated in air or liquids as well as in a vacuum, which gives a great deal of leeway in designing experimental conditions. It is capable of resolving structural details in atomic, dimensions, provided 0304-3991/90/$03.50
0 1990 - Elsevia
tunnelling microscopy which allow us to molecules from measurements of helical
the specimen is sufficiently conductive and stable under the probe [l]. Relatively little is known about the electronic properties of nucleic acids. The first STM image of DNA [2] was made by Birmig and Rohrer shortly after they invented the instrument. The molecule appeared to be lying below the level of the substrate, as if it were a very poor conductor. Some periodicity was detectable down the long axis, but the two strands of the double helix were not resolved and the conformation could not be determined. Travaglini et al. [3] succeeded in getting a better image by coating the DNA with Pt/C, but resolution was effectively limited to the grain size of the metal. Lindsay and coworkers [4-61 were the first to to show uncoated DNA molecules lying on top of the substrate. Their scans were made under water, which suggests that hydration was a factor in enhancing conductivity. Periodicities were measurable, but resolution was apparently limited by the mobility of the specimens in solution. In a similar study of DNA in air by Beebe et al. [7] the specimen appeared to be stable, but molecular dimensions were more irregular than expected. In order to understand the mechanism of image formation and to be able to compare images made under different conditions, a much broader data base of molecular dimensions is needed than is presently available. We have made extensive mea-
Science Publishers B.V. (North-Holland)
128
P.G. Arscott,
V.A. Bloomfield / STM of nucleic aciak andpolynucleotides
surements of poly(rA) . poly(rU), calf thymus DNA and poly(dG-me5dC) - poly(dG-me5dC) as models of A-form RNA, B- and A-, and Z-form DNAs. We performed the scans in air so as to eliminate the problems of molecular motion encountered in an aqueous environment yet retain some degree of hydration. Our results indicate that we can reliably measure periodicities and determine the conformational form of these molecules at high resolution.
served was about 19% (2.33 f 0.22 nm; n = 69), in a sample prepared more than 10 h before scanning. Sightings of single fibers were rare. The STM images we obtained were of bundles of fibers. From the diameter of the bundles, and the number of peaks in cross-sectional profiles, we estimate that they contained 5-7 closely packed molecules. The center-to-center distance between peaks is 2.4 it 0.33 nm. The diameter of RNA by comparison is 2.13 nm [lo].
2. Materials and methods Calf thymus DNA (Sigma Chemical), poly(rA) * poly(rU) and poly(dG-me5dC) - poly(dG-me5dC) (P-L Biochemicals) were suspended in 1OmM NaCl, 1mM NaCacodylate, pH 7.5, deposited on highly oriented pyrolytic graphite, and dried at room temperature. Scans were performed in air, using a Nanoscope II STM (Digital Instruments, Santa Barbara, CA) equipped with a 9 pm scan head and a platinum-iridium probe. The microscope was operated in constant current mode, with set points varying from 1 to 2 nA, bias voltages from - 100 to + 100 mV, and scan rates from 5.8 to 8.5 Hz.
3. Poly(rA)
l
poly(rU)
Poly(rA) - poly(rU) gives the same fiber diffraction pattern as naturally occurring RNA in the double-stranded replicative form [8]. RNA is much less variable than DNA and almost always adopts the A form, which has a helical pitch of 3.09 nm. An STM profile of the vertical versus translational position of the probe along the main axis of poly(rA) - poly(rU) would therefore be expected to show evenly spaced peaks with approximately the same periodicity. Our results, based on close to 300 measurements of the fibers in 6 different images, indicate an average of 2.87 f 0.11 nm [9]. The small difference (7%) between predicted and observed values is attributed to dehydration. The samples were exposed to air for about 2.5 h during the initial drying time and while they were being scanned. The maximum shrinkage that we ob-
4. Calf thymus DNA The conformation of random sequence DNAs, such as calf thymus DNA, is related to the hydration state of the molecule. The B form, which is characterized by a pitch of 3.4 run, predominates under conditions of high humidity in vitro, and presumably in vivo. The A form, detected at about 75% relative humidity, has an average pitch of 2.82 nm [lo]. STM images of calf thymus DNA made within 2.5 h of depositing the sample on the substrate indicate a helical periodicity of 3.15 f 0.13 nm (7 images, n = 140), A sample dried for > 10 h showed a periodicity of 2.28 nm (n = 62) [9]. These results are thus consistent with our findings with poly(rA) - poly(rU). After 2.5 h of exposure to air, the STM values for DNA are 7% less than what would be expected at high humidity, and after > 10 h, 19% less than expected for the A form. Calf thymus DNA is several orders of magnitude longer than the two polynucleotides with which it is compared in these experiments and was consequently more difficult to image. To alleviate problems of tangling and clumping, we added trivalent cations to our samples to neutralize the DNA’s negative charge and promote the lateral association of fibers in solution. We observed bundles and extensive arrays of aligned fibers in the presence of spermidine3+ and hexammine cobalt (III), and only random entanglements in the absence of these cations. The center-to-center packing distance is 3.06 f 0.53 nm (n = 94) in the aggregates that had the B-form periodicity (2.5 h
P.G. Arscott. V.A. Bloomfield / STM of nucleic acids andpolynucleotides
samples) and 1.99 f 0.29 nm (n = 18) in those that had A form (> 10 h sample). These values correspond to the maximum and minimum packing distances between DNA helices in crystals [ll].
5. Poly(dGmesdC)
l
poly(dGmesdC)
Poly(dG-me’dC) - poly(dG-me5dC) is used as a model for the distinctive Z form of DNA, which has a pitch of 4.54 nm under conditions of high humidity and high ionic strength [12]. It is stable in the B form only under low-salt conditions. We made no attempt to control ionic conditions as the samples were dried for scarming. Our sample buffer contained millimolar quantities of NaCl and NaCacodylate, but we can reasonably assume that the concentration of salt on the substrate was
129
much higher than in solution. Measurements of fibers imaged within 2.5 h of sample deposition (10 images, n = 360) indicate a helical periodicity of 4.19 f 0.11 nm [13]. This is 8% less than the crvstallogranhic value for Z-DNA and only 3% less than the variant of Z designated as Z *I The variant is usually found under dehydrating conditions, in high-salt and in ethanolic solutions. Z-DNA is immediately recognized by the lefthanded twist of the helix. The appearance of the fibers in our images fully supports the conclusion that poly(dG-me’dC) . poly(dG-me5dC) was in the Z form on the substrate. The zigzag path of the helical strands could even be seen in certain orientations (see fig. 1). The fibers were aligned in parallel arrays, but were not organized into bundles or closely packed aggregates.
Fig. 1. STM image of Z-DNA showing parallel alignment of fibers.
130
P.G. Arscott,
V.A. Bloomfield / STM of nucleic acids andpolynucleotides
6. Discussion
Most of what is known about the secondary and tertiary structure of nucleic acids has been derived from X-ray crystallography and fiber diffraction studies of a small number of oligo- and polynucleotides. The physical differences that exist, even between molecules of similar size and sequence, are much more extensive than anyone would have predicted when the A and B forms were first described. The discovery of left-handed Z-DNA precipitated a concerted effort to determine the conditions under which conformational transitions occur, and there is now considerable evidence from physical, chemical, and genetic assays to suggest that short stretches of DNA can undergo a transition without involving the whole molecule [14]. The evidence is indirect and limited since the methodology does not permit examination of individuals in a heterogeneous population and does not detect small differences which are smoothed by signal averaging. The present results with poly(rA) - poly(rU), calf thymus DNA and poly(dG-me5dC) . poly(dGme5dC) indicate the feasibility of determining the conformation of nucleic acids by scanning tunnelling microscopy. The specimens we examined were sufficiently conductive to image without a metal coating, and sufficiently differentiated to obtain good contrast. Profiles of vertical versus translational movement of the probe show clearly defined peaks spaced at regular intervals along the main axes of the fibers. The distance between peaks corresponds to the helical repeat period or half period determined by optical diffraction. We give average periodicities for comparative purposes, but can identify the specific regions of the molecules represented in the standard deviation of a series of measurements. Cross-sectional profiles of poly(rA) . poly(rU) and calf thymus DNA aggregates show that the helices are packed with approximately the same center-to-center spacing as in crystals. We could not get an accurate determination of helical diameters in the aggregates, but an upper limit could be inferred from the spacing of the peaks in these profiles. The calf thymus DNA prepared more than 10 h before scanning appeared to be severely
dehydrated and was classified as A form on that basis. However, the packing distance in that sample is less than the diameter of an A-form molecule, 1.99 nm versus 2.31 nm [lo], which raises the possibility that the DNA was Z form (diameter 1.93 nm) and that the 2.28 nm axial period we measured was a half period (full period 4.56 nm). The pitch of fully hydrated Z-DNA is 4.54 nm. Although the dimensions are an excellent match, it is unlikely that the sample would be hydrated to such an extent after so long an exposure to air. Poly(dG-me5dC) - poly(dG-rne’dc), which gave clear evidence of Z form, showed essentially the same rate of shrinkage as the other samples. Unfortunately, the handedness of the calf thymus DNA was not evident by eye. The presence of bundles of fibers in poly(rA) . poly(rU) was unexpected since the method of sample preparation did not ensure any sort of orderly aggregation. Mandelkem et al. [IS] reported similar bundles, containing an average of 7 fibers, in concentrated solutions of nucleosomal DNA studied by electric dichroism, equilibrium sedimentation, and laser light scattering. Formation was favored at lower ionic strengths (< lOruM), and higher temperatures in the range 320° C. This would suggest that the bundles we observed formed in solution rather than on the substrate, since salt concentrations on the substrate were presumably high enough to inhibit aggregation. Poly(dG-me5dC) - poly(dG-me5dC) imaged under the same conditions was found to be in the high salt Z* form instead of the B form which prevails in low-salt solutions. Those fibers were not bundled but were aligned in parallel array. We can therefore not rule out the possibility that in both cases the polynucleotides were brought into close proximity and aligned by the rastering motion of the probe. It is rather striking that in numerous scans we have not found disentangled fibers in a random coil configuration. Our STM images invariably show straight, stiff rods, lying at about a 60” angle on the substrate. It is unusual to get a good image of a single isolated fiber, and most of our best images are of groups of 5-10 molecules. The effects of probe and contact with the surface on conformation need further investigation.
P.G. Arscott, V.A. Bloomfield / STM of nucleic acid andpolynucleotides
Acknowledgements
We gratefully acknowledgethe collaborative efforts of G. Lee and D.F. Evans at the Center for Interfacial Engineering, University of Minnesota, Minneapolis, and grant support from the National Institutes of Health and National Science Foundation.
References
111N.D. Lang, Phys. Rev. Lett. 56 (1986) 1164. PI G. Binnig and H. Rohrer, in: Trends in Physics, Eds. J. Janta and J. Pantofhcek (European Physical Society, The Hague, 1984) p. 38. [31 G. Travaglini, H. Rohrer, M. Amrein and H. Gross, Surf. Sci. 181 (1987) 380. 141 B. Ban-is, U. Knipping, SM. Lindsay, L. Nagahara and T. T’hundat, Biopolymers 27 (1988) 1691.
131
[5] S.M. Lindsay and B. Barris, J. Vat. Sci. Technol. A 6 (1988) 544. [6] SM. Lindsay, T. Thundat and L. Nagahara, in: Biological and Artificial Intelligence Systems, Eds. E. Clementi and S. Chin (ESCOM Science, Leiden, 1988) p. 125. [7] T.P. Beebe, Jr., T.E. Wilson, F. Ogletree, J.E. Katz, R. BaIhom, M.B. SaImeron and W.J. Siekhaus, Science 243 (1989) 370. (81 S. Amott, in: Proc. 1st Cleveland Symp. on Macromolecules, Ed. A.G. Walton (Elsevier, Amsterdam, 1977) p. 87. [9] G. Lee, P.G. Arscott, V.A. Bloomfield and D.F. Evans, Science 244 (1989) 475. [lo] W. Saenger, Principles of Nucleic Acid Structure (Springer, New York, 1984). [ll] J.A. Schelhnan and N. Parthasarathy, J. Mol. Biol. 175 (1984) 313. [12] S. Fujii, A.H.-J. Wang, G. van der Mare& J.H. van Boom and A. Rich, Nucleic Acids Res. 10 (1982) 7879. [13] P.G. Arscott, G. Lee, V.A. Bloomfield and D.F. Evans, Nature 319 (1989) 484. [14] J.K. Barton, Chem. Eng. News (September 1988) 30. [15] M. Mandelkem, N. Dattagupta and D.M. Crothers, Proc. Natl. Acad. Sci. USA 78 (1981) 4294.
33 (1990) 127-131
127
SCANNING TUNNELLING MICROSCOPY OF NUCLEIC ACIDS AND POLYNUCLEOTIDES P.G. ARSCO’IT and V.A. BLOOMFIELD Department
of Biochemistry
University of Minnesota,
St. Paul, MN 55108, USA
Received 28 February 1990; in final form 8 March 1990
We have obtained high-resolution images of DNA and RNA by scanning determine the conformation, which depends on hydration state, of individual periodicity, packing distances in aggregates, and handedness.
1. Intmduction Biological processes are generally so finely tuned that they cannot be well understood at the molecular level without a topographical map. The twists, turns and surface texture of a DNA molecule determine how it is organized, and how it is recog&ed by the proteins that initiate and regulate transcription of the genetic message. The “morphogenetic code”, if there is one, has not yet been deciphered, but there are indications that critical binding sites on the DNA may be no more than three base pairs long, a fraction of a turn of the helix. Structural details smaller than about 1 nm cannot generally be resolved by electron microscopy, and optical diffraction techniques do not aBow one to look at specific sites on individual molecules. The scanning tunnelhng microscope (STM) offers a unique advantage in allowing a close look at DNA and small ligands in the absence of stains, shadows, and labels. There is no risk of radiation damage as there is in electron microscopy, and few other problems of specimen preservation when the harsh conditions of a vacuum are avoided. The instrument can be operated in air or liquids as well as in a vacuum, which gives a great deal of leeway in designing experimental conditions. It is capable of resolving structural details in atomic, dimensions, provided 0304-3991/90/$03.50
0 1990 - Elsevia
tunnelling microscopy which allow us to molecules from measurements of helical
the specimen is sufficiently conductive and stable under the probe [l]. Relatively little is known about the electronic properties of nucleic acids. The first STM image of DNA [2] was made by Birmig and Rohrer shortly after they invented the instrument. The molecule appeared to be lying below the level of the substrate, as if it were a very poor conductor. Some periodicity was detectable down the long axis, but the two strands of the double helix were not resolved and the conformation could not be determined. Travaglini et al. [3] succeeded in getting a better image by coating the DNA with Pt/C, but resolution was effectively limited to the grain size of the metal. Lindsay and coworkers [4-61 were the first to to show uncoated DNA molecules lying on top of the substrate. Their scans were made under water, which suggests that hydration was a factor in enhancing conductivity. Periodicities were measurable, but resolution was apparently limited by the mobility of the specimens in solution. In a similar study of DNA in air by Beebe et al. [7] the specimen appeared to be stable, but molecular dimensions were more irregular than expected. In order to understand the mechanism of image formation and to be able to compare images made under different conditions, a much broader data base of molecular dimensions is needed than is presently available. We have made extensive mea-
Science Publishers B.V. (North-Holland)
128
P.G. Arscott,
V.A. Bloomfield / STM of nucleic aciak andpolynucleotides
surements of poly(rA) . poly(rU), calf thymus DNA and poly(dG-me5dC) - poly(dG-me5dC) as models of A-form RNA, B- and A-, and Z-form DNAs. We performed the scans in air so as to eliminate the problems of molecular motion encountered in an aqueous environment yet retain some degree of hydration. Our results indicate that we can reliably measure periodicities and determine the conformational form of these molecules at high resolution.
served was about 19% (2.33 f 0.22 nm; n = 69), in a sample prepared more than 10 h before scanning. Sightings of single fibers were rare. The STM images we obtained were of bundles of fibers. From the diameter of the bundles, and the number of peaks in cross-sectional profiles, we estimate that they contained 5-7 closely packed molecules. The center-to-center distance between peaks is 2.4 it 0.33 nm. The diameter of RNA by comparison is 2.13 nm [lo].
2. Materials and methods Calf thymus DNA (Sigma Chemical), poly(rA) * poly(rU) and poly(dG-me5dC) - poly(dG-me5dC) (P-L Biochemicals) were suspended in 1OmM NaCl, 1mM NaCacodylate, pH 7.5, deposited on highly oriented pyrolytic graphite, and dried at room temperature. Scans were performed in air, using a Nanoscope II STM (Digital Instruments, Santa Barbara, CA) equipped with a 9 pm scan head and a platinum-iridium probe. The microscope was operated in constant current mode, with set points varying from 1 to 2 nA, bias voltages from - 100 to + 100 mV, and scan rates from 5.8 to 8.5 Hz.
3. Poly(rA)
l
poly(rU)
Poly(rA) - poly(rU) gives the same fiber diffraction pattern as naturally occurring RNA in the double-stranded replicative form [8]. RNA is much less variable than DNA and almost always adopts the A form, which has a helical pitch of 3.09 nm. An STM profile of the vertical versus translational position of the probe along the main axis of poly(rA) - poly(rU) would therefore be expected to show evenly spaced peaks with approximately the same periodicity. Our results, based on close to 300 measurements of the fibers in 6 different images, indicate an average of 2.87 f 0.11 nm [9]. The small difference (7%) between predicted and observed values is attributed to dehydration. The samples were exposed to air for about 2.5 h during the initial drying time and while they were being scanned. The maximum shrinkage that we ob-
4. Calf thymus DNA The conformation of random sequence DNAs, such as calf thymus DNA, is related to the hydration state of the molecule. The B form, which is characterized by a pitch of 3.4 run, predominates under conditions of high humidity in vitro, and presumably in vivo. The A form, detected at about 75% relative humidity, has an average pitch of 2.82 nm [lo]. STM images of calf thymus DNA made within 2.5 h of depositing the sample on the substrate indicate a helical periodicity of 3.15 f 0.13 nm (7 images, n = 140), A sample dried for > 10 h showed a periodicity of 2.28 nm (n = 62) [9]. These results are thus consistent with our findings with poly(rA) - poly(rU). After 2.5 h of exposure to air, the STM values for DNA are 7% less than what would be expected at high humidity, and after > 10 h, 19% less than expected for the A form. Calf thymus DNA is several orders of magnitude longer than the two polynucleotides with which it is compared in these experiments and was consequently more difficult to image. To alleviate problems of tangling and clumping, we added trivalent cations to our samples to neutralize the DNA’s negative charge and promote the lateral association of fibers in solution. We observed bundles and extensive arrays of aligned fibers in the presence of spermidine3+ and hexammine cobalt (III), and only random entanglements in the absence of these cations. The center-to-center packing distance is 3.06 f 0.53 nm (n = 94) in the aggregates that had the B-form periodicity (2.5 h
P.G. Arscott. V.A. Bloomfield / STM of nucleic acids andpolynucleotides
samples) and 1.99 f 0.29 nm (n = 18) in those that had A form (> 10 h sample). These values correspond to the maximum and minimum packing distances between DNA helices in crystals [ll].
5. Poly(dGmesdC)
l
poly(dGmesdC)
Poly(dG-me’dC) - poly(dG-me5dC) is used as a model for the distinctive Z form of DNA, which has a pitch of 4.54 nm under conditions of high humidity and high ionic strength [12]. It is stable in the B form only under low-salt conditions. We made no attempt to control ionic conditions as the samples were dried for scarming. Our sample buffer contained millimolar quantities of NaCl and NaCacodylate, but we can reasonably assume that the concentration of salt on the substrate was
129
much higher than in solution. Measurements of fibers imaged within 2.5 h of sample deposition (10 images, n = 360) indicate a helical periodicity of 4.19 f 0.11 nm [13]. This is 8% less than the crvstallogranhic value for Z-DNA and only 3% less than the variant of Z designated as Z *I The variant is usually found under dehydrating conditions, in high-salt and in ethanolic solutions. Z-DNA is immediately recognized by the lefthanded twist of the helix. The appearance of the fibers in our images fully supports the conclusion that poly(dG-me’dC) . poly(dG-me5dC) was in the Z form on the substrate. The zigzag path of the helical strands could even be seen in certain orientations (see fig. 1). The fibers were aligned in parallel arrays, but were not organized into bundles or closely packed aggregates.
Fig. 1. STM image of Z-DNA showing parallel alignment of fibers.
130
P.G. Arscott,
V.A. Bloomfield / STM of nucleic acids andpolynucleotides
6. Discussion
Most of what is known about the secondary and tertiary structure of nucleic acids has been derived from X-ray crystallography and fiber diffraction studies of a small number of oligo- and polynucleotides. The physical differences that exist, even between molecules of similar size and sequence, are much more extensive than anyone would have predicted when the A and B forms were first described. The discovery of left-handed Z-DNA precipitated a concerted effort to determine the conditions under which conformational transitions occur, and there is now considerable evidence from physical, chemical, and genetic assays to suggest that short stretches of DNA can undergo a transition without involving the whole molecule [14]. The evidence is indirect and limited since the methodology does not permit examination of individuals in a heterogeneous population and does not detect small differences which are smoothed by signal averaging. The present results with poly(rA) - poly(rU), calf thymus DNA and poly(dG-me5dC) . poly(dGme5dC) indicate the feasibility of determining the conformation of nucleic acids by scanning tunnelling microscopy. The specimens we examined were sufficiently conductive to image without a metal coating, and sufficiently differentiated to obtain good contrast. Profiles of vertical versus translational movement of the probe show clearly defined peaks spaced at regular intervals along the main axes of the fibers. The distance between peaks corresponds to the helical repeat period or half period determined by optical diffraction. We give average periodicities for comparative purposes, but can identify the specific regions of the molecules represented in the standard deviation of a series of measurements. Cross-sectional profiles of poly(rA) . poly(rU) and calf thymus DNA aggregates show that the helices are packed with approximately the same center-to-center spacing as in crystals. We could not get an accurate determination of helical diameters in the aggregates, but an upper limit could be inferred from the spacing of the peaks in these profiles. The calf thymus DNA prepared more than 10 h before scanning appeared to be severely
dehydrated and was classified as A form on that basis. However, the packing distance in that sample is less than the diameter of an A-form molecule, 1.99 nm versus 2.31 nm [lo], which raises the possibility that the DNA was Z form (diameter 1.93 nm) and that the 2.28 nm axial period we measured was a half period (full period 4.56 nm). The pitch of fully hydrated Z-DNA is 4.54 nm. Although the dimensions are an excellent match, it is unlikely that the sample would be hydrated to such an extent after so long an exposure to air. Poly(dG-me5dC) - poly(dG-rne’dc), which gave clear evidence of Z form, showed essentially the same rate of shrinkage as the other samples. Unfortunately, the handedness of the calf thymus DNA was not evident by eye. The presence of bundles of fibers in poly(rA) . poly(rU) was unexpected since the method of sample preparation did not ensure any sort of orderly aggregation. Mandelkem et al. [IS] reported similar bundles, containing an average of 7 fibers, in concentrated solutions of nucleosomal DNA studied by electric dichroism, equilibrium sedimentation, and laser light scattering. Formation was favored at lower ionic strengths (< lOruM), and higher temperatures in the range 320° C. This would suggest that the bundles we observed formed in solution rather than on the substrate, since salt concentrations on the substrate were presumably high enough to inhibit aggregation. Poly(dG-me5dC) - poly(dG-me5dC) imaged under the same conditions was found to be in the high salt Z* form instead of the B form which prevails in low-salt solutions. Those fibers were not bundled but were aligned in parallel array. We can therefore not rule out the possibility that in both cases the polynucleotides were brought into close proximity and aligned by the rastering motion of the probe. It is rather striking that in numerous scans we have not found disentangled fibers in a random coil configuration. Our STM images invariably show straight, stiff rods, lying at about a 60” angle on the substrate. It is unusual to get a good image of a single isolated fiber, and most of our best images are of groups of 5-10 molecules. The effects of probe and contact with the surface on conformation need further investigation.
P.G. Arscott, V.A. Bloomfield / STM of nucleic acid andpolynucleotides
Acknowledgements
We gratefully acknowledgethe collaborative efforts of G. Lee and D.F. Evans at the Center for Interfacial Engineering, University of Minnesota, Minneapolis, and grant support from the National Institutes of Health and National Science Foundation.
References
111N.D. Lang, Phys. Rev. Lett. 56 (1986) 1164. PI G. Binnig and H. Rohrer, in: Trends in Physics, Eds. J. Janta and J. Pantofhcek (European Physical Society, The Hague, 1984) p. 38. [31 G. Travaglini, H. Rohrer, M. Amrein and H. Gross, Surf. Sci. 181 (1987) 380. 141 B. Ban-is, U. Knipping, SM. Lindsay, L. Nagahara and T. T’hundat, Biopolymers 27 (1988) 1691.
131
[5] S.M. Lindsay and B. Barris, J. Vat. Sci. Technol. A 6 (1988) 544. [6] SM. Lindsay, T. Thundat and L. Nagahara, in: Biological and Artificial Intelligence Systems, Eds. E. Clementi and S. Chin (ESCOM Science, Leiden, 1988) p. 125. [7] T.P. Beebe, Jr., T.E. Wilson, F. Ogletree, J.E. Katz, R. BaIhom, M.B. SaImeron and W.J. Siekhaus, Science 243 (1989) 370. (81 S. Amott, in: Proc. 1st Cleveland Symp. on Macromolecules, Ed. A.G. Walton (Elsevier, Amsterdam, 1977) p. 87. [9] G. Lee, P.G. Arscott, V.A. Bloomfield and D.F. Evans, Science 244 (1989) 475. [lo] W. Saenger, Principles of Nucleic Acid Structure (Springer, New York, 1984). [ll] J.A. Schelhnan and N. Parthasarathy, J. Mol. Biol. 175 (1984) 313. [12] S. Fujii, A.H.-J. Wang, G. van der Mare& J.H. van Boom and A. Rich, Nucleic Acids Res. 10 (1982) 7879. [13] P.G. Arscott, G. Lee, V.A. Bloomfield and D.F. Evans, Nature 319 (1989) 484. [14] J.K. Barton, Chem. Eng. News (September 1988) 30. [15] M. Mandelkem, N. Dattagupta and D.M. Crothers, Proc. Natl. Acad. Sci. USA 78 (1981) 4294.
