Ultramicroscopy 34 (1990) 141-147 North-Holland
141
Scanning tunneling microscopy of a liquid crystalline phase of poly((dA-dT) • (dA-dT)) induced by a histone H 1 peptide R. Coratger, A. Chahboun, F. Ajustron, J. Beauvillain CEMES-LOE/CNRS, 29 rue J. Maroi& BP 4347, 31055 Toulouse, France
M. Erard and F. Amalric CRBGC/CNRS, 118 route de Narbonne, 31062 Toulouse, France Received 6 June 1990
In this report, we present the first observations of uncoated poly((dA-dT)-(dA-dT)) molecules organi7ed in a liquid crystalline phase induced by the binding of a histone H 1 peptide. The effect of the peptide on the polymer condensation is clearly illustrated on the large-scale STM images which reveal a well defined spacing between parallel DNA helices. High resolution images of rare isolated molecules of poly((dA-dT)-(dA-dT)) exhibit two sets of helical pitch values, 6 and 7.5 nm. While the lower value can be correlated with the pitch of poly((dA-dT).(dA-dT)), the larger one may arise from peptide binding in the polymer minor groove.
1. Introduction Because of its ability to produce three-dimensional images of conducting [1] and semi-conducting [2] surfaces with ultra-high resolution, the scanning tunneling microscope (STM) [3] has been used in a large number of fundamental and technological fields. One of the most exciting developments has been the promise of visualising small biological objects [4,5] and molecules [6] deposited on flat surfaces with atomic resolution. Several groups have therefore used the STM to study biological structures and macromolecuies. But STM imaging in biology has proven difficult, mainly due to the mechanism of tunneling through biological materials and to the sample preparation. In this field, two approaches have been used: either the biological molecules were coated with thin conducting films [7] (for example P t - I r - C ) or these macromolecules were directly observed [8] in the hope to achieve atomic resolution.
Since DNA in vivo is generally present in a compact form, as in the cell nucleus, transition towards condensed phases have long been the focus of interest. In fact, a variety of reagents or conditions can lead to a liquid crystalline phase of DNA as assessed by circular dichroism measurements (for a review, see ref. [9]). Ip an attempt to visualise this phase by scanning tunneling microscopy, we have chosen complexes of poly((dAdT)- (dA-dT)) and a histone H 1 peptide as a model system. Indeed, this peptide was shown to efficiently condense DNA as a consequence of binding in its minor groove [10]. It also has a preferential affinity for A - T base pairs. We report here the first observations by STM of uncoated poly((dA-dT). (dA-dT)) molecules deposited on flat HOPG graphite surfaces and organized in a liquid crystalline phase induced by the binding of a histone H I peptide. We have thus been able to visualize both DNA condensation and peptide binding phenomena.
0304-3991/90/$03.50 © 1990 - Elsevier Science Publishers B.V. (North-Holland)
142
R. Coratger et al. / STM of poly((dA-dT). (dA-dT))
2. Experiment The pocket-sized STM used in this work has been largely described in the literature (see for example ref. [11] and references therein). The tips are made b y etching 0.1 m m tungsten wires in N a O H solutions and produce atomic resolution on clean graphite surfaces with high reliability [121. The typical tunneling conditions are a samplepositive polarization of about 500 mV and a tunnelling current of 0.2 hA. Taking into account the " u n c l e a r " m e c h a n i s m of electronic transfer through the molecules and the different values given in the literature (for example about DNA), we think that such great tunneling resistances keep the tip at a large distance from the sample and
m a y avoid contacts with the observed structures. For the same reasons, the scanning speed is usually chosen so as to keep the tip velocity on the sample below 300 n m / s . These tunneling conditions are supposed to reproduce the topography rather than the electronic structure of the sample [13] and prevent deformations of the graphite surface due to shorter t i p - s a m p l e distances [14]. The substrate is always a freshly cleaved H O P G surface that only presents well known cleavage steps of a height in the range of a few nanometers. A solution of poly((dA-dT). (dA-dT)) (Pharmacia Biochemicals) at a concentration of 35 # g / m l in a buffer containing 150mM NaC1, 10mM N a phosphate, p H 7.4 was mixed with an equal volume of a solution of a histone H 1 peptide ( K T P K K A K K P ) 2 at a concentration of 7 0 / ~ g / m l
Fig. 1. TEM micrograph of Pt-Ir coated poly((dA-dT).(dA-dT))/peptide complexes deposited on a carbon film and observed at 100 kV. The arrow at (1) indicates large bundles of molecules while the arrow at (2)points to smaller groups the size of which is very close to that of isolated molecules.
R. Coratger et al. / S T M of poly((dA-dT) . (dA-dT))
in the same buffer. Then 5 /xl droplets of the complex were deposited on chips of freshly cleaved HOPG and allowed to evaporate. For conventional transmission electron microscopy, 5/~1 droplets of the same poly((dA-dT) • (dA-dT))/peptide complexes were deposited on carbon-coated gl'ids and allowed to evaporate. The samples were coated with a thin Pt-Ir film and examined under a JEOL 200 EX microscope at 100 kV.
3. Results
Fig. 1 presents a micrograph of poly((dA-dT) (dA-dT))/peptide complexes as they appear in the transmission electron microscope (TEM) after deposition on a. carbon film and coating with a thin Pt-Ir film. The Pt-Ir film.~ a r e k n o w n t o m i n i m i z e the grain size so that they are often used in the preparation of shadowed samples for TEM. This image clearly shows the typical condensation in-
143
duced by the peptide since the poly((dA-dT) • (dAdT)) molecules t e n d to associate, the bundles of molecules having a thickness ranging from a few nanometers to several ten nanometers (the arrow at (1) in fig. 1)(with a maximum size of 50 nm). In addition, the view presents an arborescent structure which may suggest that the molecules do not tangle on the substrate but are sideways associated. On the othe r hand, the size of the smaller observed structures (2) implies that individual molecules may be localiTed a t the end of some arborescences. Fig. 2 is a large-scale STM image of uncoated complexes deposited on a HOPG graphite surface. The scanned area is 75 nm × 140 nm with a maximum corrugation of about 4 nm. The upper-right part of this image shows two steps which join and end by three molecules running to the lower-left part of the image. We have often observed such arrays of closely packed molecules. As we had checked by circular dichroism that the complexes in solution displayed
Fig. 2. STM image of poly((dA-dT).(dA-dT))/peptide complexes deposited on freshly cleaved HOPG graphite. The scanned area is 75 x 140 nm2 with a maximum corrugation of about 4 nm (scan speed 1 Hz, polarization 0.9 V, I T = 0.2 nA). This image shows a bidimensional array of polymers ending by three molecules lying side by side with an interhelical spacing of about 6.3 nm.
144
R. Coratger et al. / S T M of polt,((dA-dT) . (dA-dT))
Fig. 3. STM image of a 250 x 200 n m 2 area showing poly((dA-dT). (dA-dT))/peptide complexes deposited on the graphite. The larger group of molecules is 40 n m large a n d about 2.5 n m high while the second group exhibits a thickness of 20 nm. This type of organization and the size of the bundles of molecules are in perfect agreement with the T E M observations (same tunneling conditions).
Fig. 4. STM image showing the displacement of the molecules by the tip during the scan. The total area and the corrugation are respectively 63 × 92 n m 2' and 2.5 nm. The group of molecules has been displaced from 16.5 n m against the step indicated by an arrow (scan speed 1 Hz, polarization 0.6 V, I T = 0.2 hA).
R. Coratger et al. / S T M of poly((dA-dT). (dA-dT))
145
Fig. 5. STM image of an isolated poly((dA-dT). (dA-dT)) molecule adsorbed on a graphite defect of the H O P G substrate. The total area is 56 x 56 n m 2 with a m a x i m u m corrugation of 2 n m (same tunneling conditions). Two helical pitch values, 6 and 7.5 nm, can be seen.
the typical spectrum of a liquid crystalline phase, our STM results prove that a certain degree of order is preserved during deposition and observation of the samples. The lateral spacing between the parallel strands is about 6.3 nm. This rather large value can be ascribed to the peptide binding, as discussed further on. Another example of this ordering is shown in Fig. 3. The image comes from another sample processed in the same way. The total area and the corrugation are respectively 250 × 250 nm z and about 2.5 nm. A group of molecules (20 nm thick) emerges from another group (upper part of the image), the thickness of which is about 40 nm. The striking feature of this image is its perfect agreement with the arborescent structure, as observed by TEM, leading to virtually rectilinear bundles. As the steps or defects on the graphite surface have proven to be more reactive than the basal
plane (0001) [15,16], we often observe bundles of macromolecules embedded at or near their edges but we cannot detect periodic corrugations along them. The movement of the biological samples under the tip mechanical disturbance appears to be one of the most important obstacles to the observation of isolated molecules [16]. To illustrate this phenomenon, we present a STM image (fig. 4) in which the tip has displaced a group of molecules during a scan despite the great tunneling resistances used.. The molecules have been displaced from 16.5 n m against a cleavage step visible on the upper part of the image and indicated by an alTOW.
Fig. 5 shows an isolated poly((dA-dT)-(dAdT)) molecule which has been adsorbed on the graphite surface. This image quite clearly depicts a fight-handed helix, but both the helical pitch and
146
R, Coratger et al. / S T M of poly((dA-dT) . (dA-dT))
the width of the molecule appear enlarged in different places which probably correspond to the peptide binding sites.
4. Discussion
Regular arrangements of DNA molecules have already been visualized by STM [15,17]. Furthermore, as recently summarized by Keller et al. [18], it now appears that images of uncoated DNA molecules have most often been obtained when these molecules were either adsorbed to step edges or tightly packed in ordered arrays. In both cases, the molecules ran less risk of being swept aside by the tip. Our present report gives full support to these observations. The TEM micrographs allowed us to compare the number of molecules deposited with the number of molecules really observed by STM per unit area. As the STM observations always revealed lower densities, the displacement of the molecules weakly bound to the substrate during the scans (fig. 4) may explain why the localization of single molecules is a time-consuming procedure. In addition, this phenomenon occurs even for large tip-sample distances (large polarizations, low tunneling currents), and this again underlines the important role of the sample preparation in such experiments. Our aim was to visualize a liquid crystalline phase of DNA. There are different ways of inducing such a phase transition in DNA solutions [9], in particular the use of very high DNA and salt concentrations [19]. In this case, however, STM experiments are made rather difficult by the formation of areas in which the tunneling current becomes very noisy, giving rise to large corrugations. In contrast, our protocol requires relatively low concentrations in DNA at a physiological ionic strength. The peptide actually binds to poly((dA-dT). (dA-dT)), taking the place of the hydration layer in its minor groove [10]. It is the disturbance of the hydration network around the polymer which leads to its condensation by solvent exclusion. STM has then enabled us to visualize both the condensation and the peptide binding phenomena.
Circular dichroism signals such as the one we get from the poly((dA-dT)-(dA-dT))/peptide complexes in solution are correlated with the presence of a cholesteric liquid crystalline phase [9]. This phase consists of layers of parallel double helices stacked together with a slight twist between layers. We have never visualiT~l more than one layer, as assessed by the corrugation values ranging from 2.5 nm (figs. 3 and 4) to a maximum of 4 nm (fig. 2). As we can exclude that the tip would pierce through the uppermost layers to image only the layer in contact with the graphite, we have to conclude either that spreading conditions led to monolayers or that stacked layers gave unstable tunneling conditions. We were intrigued to observe no periodicity along the closely packed molecules in contrast with the situation for isolated ones. We then suggest that the stronger interaction with the graphite of the condensed molecules somehow alters the polymer structure. As for the periodicity of the isolated molecules (fig. 5), we have found two types of helical pitch values, 6 and 7.5 nm. We think that 6 nm can be assigned to the pitch of poly((dA-dT). (dA-dT)). Indeed, on the one hand, this polymer was shown to have a larger periodicity than standard B-form DNA (4.2 nm instead of 3.4 nm according to ref. [20]), and on the other hand, large periodicities up to 6.3 nm, as depicted by STM, have already been reported by Beebe et al. [21]. The other pitch value of 7.5 nm probably corresponds to places where the rather bulky peptide of 18 amino acids is bound. Another consequence of the peptide binding is the irregularity in the apparent width of the isolated molecules shown in fig. 5. This thickening of the molecule at the peptide binding sites might also explain the rather large interhelical spacing (6.3 nm) of the ordered array presented in fig. 2.
5. Conclusion
Our results again illustrate the potential of scanning tunneling microscopy to resolve the ultrastructure of biomolecules. The liquid crystalline phase of poly((dA-dT). (dA-dT)) that this tech-
R. Coratger et al. / S T M of poly((dA-dT) . (dA-dT))
nique has enabled us to observe is undoubtedly a well suited system. Indeed, the hquid crystalline organization of the molecules is almost a prerequisite to their adsorption on the graphite substrate, according to a yet unknown mechanism, as already stated in other works. It is worth remarking that we have been able to visualize not only how the nucleic acid is organized but also how its parameters have been modified by the binding of the peptide used as a condensation factor. STM technique should then be useful in the study of interactions between nucleic acids and proteins.
References [1] G. Binnig and H. Rohrer, Surf. Sci. 126 (1983) 237. [2] G. Binnig, H. Rohrer, Ch. Gerber and E. Weibel, Phys. Rev. Lett. 50 (1983) 120. [3] G. Binnig, H. Rohrer, Ch. Gerber and E. Weibel, Phys. Rev. Lett. 49 (1982); Appl. Phys. Lett. 40 (1982) 178. [4] A. Baro, R. Miranda, J. Alaman, N. Garcia, G. Binnig, H. Rohrer, Ch. Gerber and J.L. Carrasco, Nature 315 (1985) 253. [5] B.L. Blackford, M.O. Watanabe, D.C. Dahn, M.H. Jericho, G. Southham and T.J. Beveridge, Ultramicroscopy 27 (1989) 427. [6] T. Sleator and R. Tycko, Phys. Rev. Lett. 60 (1988) 1418.
147
[7] M. Amrein, A. Stasiak, H. Gross, E. Stoll and G. Travaglini, Science 240 (1988) 514. [8] M. Amrein, R. Diirr, A. Stasiak, H. Gross and G. Travaglini, Science 243 (1989) 1708. [9] I. Tinoeo, C. Bustamante and M. Maestre, Annu. Rev. Biophys. Bioeng. 9 (1980) 107. [10] M. Erard, F. Lakhdar-Ghazal and F. Amalric, Eur. J. Biochem. 191 (1990) 19. [11] R. Coratger, F. Ajustron and J. Beauvillain, Spectra 2000 No. 147 (1990) 34. [12] R. Coratger, A. Claverie, F. Ajustron and J. Beauvillain, Surf. Sci. 227 (1990) 7. [13] J. Tersoff and D.R. Hamann, Phys. Rev. B 31 (1985) 805. [14] H.J. Mamin, E. Ganz, D.W. Abraham, R.E. Thomson and J. Clarke, Phys. Rev. B 34 (1986) 9015. [15] D. Keller, C. Bustamante and R.W. Keller, Proc. Natl. Acad. Sci. USA 86 (1989) 5356. [16] M. Salmeron, T. Beebe, J. Odriozola, T. Wilson, D. Ogletree and W. Sieklaus, J. Vac. Sci. Technol. A 8 (1990) 635. [17] G. Lee, P.G. Arscott, V.A. Bloomfield and D.F. Evans, Science 244 (1989) 475. [18] R.W. Keller, D.D. Dunlap, C. Bustamante, D.J. Keller, R.G. Garcia, C. Gray and M.F. Maestre, J. Vac. Sci. Technol. A 8 (1990) 706. [19] R.L. Rill, Proc. Natl. Acad. Sci. USA 83 (1986) 342. [20] S. Brain, Nature New Biol. 232 (1971) 174. [21] T.P. Beebe, T.E. Wilson, D.F. Ogletree, J.E. Katz, R. Balhorn, M.B. Salmeron and W.J. Sieklaus, Science 243 (1989) 370.
141
Scanning tunneling microscopy of a liquid crystalline phase of poly((dA-dT) • (dA-dT)) induced by a histone H 1 peptide R. Coratger, A. Chahboun, F. Ajustron, J. Beauvillain CEMES-LOE/CNRS, 29 rue J. Maroi& BP 4347, 31055 Toulouse, France
M. Erard and F. Amalric CRBGC/CNRS, 118 route de Narbonne, 31062 Toulouse, France Received 6 June 1990
In this report, we present the first observations of uncoated poly((dA-dT)-(dA-dT)) molecules organi7ed in a liquid crystalline phase induced by the binding of a histone H 1 peptide. The effect of the peptide on the polymer condensation is clearly illustrated on the large-scale STM images which reveal a well defined spacing between parallel DNA helices. High resolution images of rare isolated molecules of poly((dA-dT)-(dA-dT)) exhibit two sets of helical pitch values, 6 and 7.5 nm. While the lower value can be correlated with the pitch of poly((dA-dT).(dA-dT)), the larger one may arise from peptide binding in the polymer minor groove.
1. Introduction Because of its ability to produce three-dimensional images of conducting [1] and semi-conducting [2] surfaces with ultra-high resolution, the scanning tunneling microscope (STM) [3] has been used in a large number of fundamental and technological fields. One of the most exciting developments has been the promise of visualising small biological objects [4,5] and molecules [6] deposited on flat surfaces with atomic resolution. Several groups have therefore used the STM to study biological structures and macromolecuies. But STM imaging in biology has proven difficult, mainly due to the mechanism of tunneling through biological materials and to the sample preparation. In this field, two approaches have been used: either the biological molecules were coated with thin conducting films [7] (for example P t - I r - C ) or these macromolecules were directly observed [8] in the hope to achieve atomic resolution.
Since DNA in vivo is generally present in a compact form, as in the cell nucleus, transition towards condensed phases have long been the focus of interest. In fact, a variety of reagents or conditions can lead to a liquid crystalline phase of DNA as assessed by circular dichroism measurements (for a review, see ref. [9]). Ip an attempt to visualise this phase by scanning tunneling microscopy, we have chosen complexes of poly((dAdT)- (dA-dT)) and a histone H 1 peptide as a model system. Indeed, this peptide was shown to efficiently condense DNA as a consequence of binding in its minor groove [10]. It also has a preferential affinity for A - T base pairs. We report here the first observations by STM of uncoated poly((dA-dT). (dA-dT)) molecules deposited on flat HOPG graphite surfaces and organized in a liquid crystalline phase induced by the binding of a histone H I peptide. We have thus been able to visualize both DNA condensation and peptide binding phenomena.
0304-3991/90/$03.50 © 1990 - Elsevier Science Publishers B.V. (North-Holland)
142
R. Coratger et al. / STM of poly((dA-dT). (dA-dT))
2. Experiment The pocket-sized STM used in this work has been largely described in the literature (see for example ref. [11] and references therein). The tips are made b y etching 0.1 m m tungsten wires in N a O H solutions and produce atomic resolution on clean graphite surfaces with high reliability [121. The typical tunneling conditions are a samplepositive polarization of about 500 mV and a tunnelling current of 0.2 hA. Taking into account the " u n c l e a r " m e c h a n i s m of electronic transfer through the molecules and the different values given in the literature (for example about DNA), we think that such great tunneling resistances keep the tip at a large distance from the sample and
m a y avoid contacts with the observed structures. For the same reasons, the scanning speed is usually chosen so as to keep the tip velocity on the sample below 300 n m / s . These tunneling conditions are supposed to reproduce the topography rather than the electronic structure of the sample [13] and prevent deformations of the graphite surface due to shorter t i p - s a m p l e distances [14]. The substrate is always a freshly cleaved H O P G surface that only presents well known cleavage steps of a height in the range of a few nanometers. A solution of poly((dA-dT). (dA-dT)) (Pharmacia Biochemicals) at a concentration of 35 # g / m l in a buffer containing 150mM NaC1, 10mM N a phosphate, p H 7.4 was mixed with an equal volume of a solution of a histone H 1 peptide ( K T P K K A K K P ) 2 at a concentration of 7 0 / ~ g / m l
Fig. 1. TEM micrograph of Pt-Ir coated poly((dA-dT).(dA-dT))/peptide complexes deposited on a carbon film and observed at 100 kV. The arrow at (1) indicates large bundles of molecules while the arrow at (2)points to smaller groups the size of which is very close to that of isolated molecules.
R. Coratger et al. / S T M of poly((dA-dT) . (dA-dT))
in the same buffer. Then 5 /xl droplets of the complex were deposited on chips of freshly cleaved HOPG and allowed to evaporate. For conventional transmission electron microscopy, 5/~1 droplets of the same poly((dA-dT) • (dA-dT))/peptide complexes were deposited on carbon-coated gl'ids and allowed to evaporate. The samples were coated with a thin Pt-Ir film and examined under a JEOL 200 EX microscope at 100 kV.
3. Results
Fig. 1 presents a micrograph of poly((dA-dT) (dA-dT))/peptide complexes as they appear in the transmission electron microscope (TEM) after deposition on a. carbon film and coating with a thin Pt-Ir film. The Pt-Ir film.~ a r e k n o w n t o m i n i m i z e the grain size so that they are often used in the preparation of shadowed samples for TEM. This image clearly shows the typical condensation in-
143
duced by the peptide since the poly((dA-dT) • (dAdT)) molecules t e n d to associate, the bundles of molecules having a thickness ranging from a few nanometers to several ten nanometers (the arrow at (1) in fig. 1)(with a maximum size of 50 nm). In addition, the view presents an arborescent structure which may suggest that the molecules do not tangle on the substrate but are sideways associated. On the othe r hand, the size of the smaller observed structures (2) implies that individual molecules may be localiTed a t the end of some arborescences. Fig. 2 is a large-scale STM image of uncoated complexes deposited on a HOPG graphite surface. The scanned area is 75 nm × 140 nm with a maximum corrugation of about 4 nm. The upper-right part of this image shows two steps which join and end by three molecules running to the lower-left part of the image. We have often observed such arrays of closely packed molecules. As we had checked by circular dichroism that the complexes in solution displayed
Fig. 2. STM image of poly((dA-dT).(dA-dT))/peptide complexes deposited on freshly cleaved HOPG graphite. The scanned area is 75 x 140 nm2 with a maximum corrugation of about 4 nm (scan speed 1 Hz, polarization 0.9 V, I T = 0.2 nA). This image shows a bidimensional array of polymers ending by three molecules lying side by side with an interhelical spacing of about 6.3 nm.
144
R. Coratger et al. / S T M of polt,((dA-dT) . (dA-dT))
Fig. 3. STM image of a 250 x 200 n m 2 area showing poly((dA-dT). (dA-dT))/peptide complexes deposited on the graphite. The larger group of molecules is 40 n m large a n d about 2.5 n m high while the second group exhibits a thickness of 20 nm. This type of organization and the size of the bundles of molecules are in perfect agreement with the T E M observations (same tunneling conditions).
Fig. 4. STM image showing the displacement of the molecules by the tip during the scan. The total area and the corrugation are respectively 63 × 92 n m 2' and 2.5 nm. The group of molecules has been displaced from 16.5 n m against the step indicated by an arrow (scan speed 1 Hz, polarization 0.6 V, I T = 0.2 hA).
R. Coratger et al. / S T M of poly((dA-dT). (dA-dT))
145
Fig. 5. STM image of an isolated poly((dA-dT). (dA-dT)) molecule adsorbed on a graphite defect of the H O P G substrate. The total area is 56 x 56 n m 2 with a m a x i m u m corrugation of 2 n m (same tunneling conditions). Two helical pitch values, 6 and 7.5 nm, can be seen.
the typical spectrum of a liquid crystalline phase, our STM results prove that a certain degree of order is preserved during deposition and observation of the samples. The lateral spacing between the parallel strands is about 6.3 nm. This rather large value can be ascribed to the peptide binding, as discussed further on. Another example of this ordering is shown in Fig. 3. The image comes from another sample processed in the same way. The total area and the corrugation are respectively 250 × 250 nm z and about 2.5 nm. A group of molecules (20 nm thick) emerges from another group (upper part of the image), the thickness of which is about 40 nm. The striking feature of this image is its perfect agreement with the arborescent structure, as observed by TEM, leading to virtually rectilinear bundles. As the steps or defects on the graphite surface have proven to be more reactive than the basal
plane (0001) [15,16], we often observe bundles of macromolecules embedded at or near their edges but we cannot detect periodic corrugations along them. The movement of the biological samples under the tip mechanical disturbance appears to be one of the most important obstacles to the observation of isolated molecules [16]. To illustrate this phenomenon, we present a STM image (fig. 4) in which the tip has displaced a group of molecules during a scan despite the great tunneling resistances used.. The molecules have been displaced from 16.5 n m against a cleavage step visible on the upper part of the image and indicated by an alTOW.
Fig. 5 shows an isolated poly((dA-dT)-(dAdT)) molecule which has been adsorbed on the graphite surface. This image quite clearly depicts a fight-handed helix, but both the helical pitch and
146
R, Coratger et al. / S T M of poly((dA-dT) . (dA-dT))
the width of the molecule appear enlarged in different places which probably correspond to the peptide binding sites.
4. Discussion
Regular arrangements of DNA molecules have already been visualized by STM [15,17]. Furthermore, as recently summarized by Keller et al. [18], it now appears that images of uncoated DNA molecules have most often been obtained when these molecules were either adsorbed to step edges or tightly packed in ordered arrays. In both cases, the molecules ran less risk of being swept aside by the tip. Our present report gives full support to these observations. The TEM micrographs allowed us to compare the number of molecules deposited with the number of molecules really observed by STM per unit area. As the STM observations always revealed lower densities, the displacement of the molecules weakly bound to the substrate during the scans (fig. 4) may explain why the localization of single molecules is a time-consuming procedure. In addition, this phenomenon occurs even for large tip-sample distances (large polarizations, low tunneling currents), and this again underlines the important role of the sample preparation in such experiments. Our aim was to visualize a liquid crystalline phase of DNA. There are different ways of inducing such a phase transition in DNA solutions [9], in particular the use of very high DNA and salt concentrations [19]. In this case, however, STM experiments are made rather difficult by the formation of areas in which the tunneling current becomes very noisy, giving rise to large corrugations. In contrast, our protocol requires relatively low concentrations in DNA at a physiological ionic strength. The peptide actually binds to poly((dA-dT). (dA-dT)), taking the place of the hydration layer in its minor groove [10]. It is the disturbance of the hydration network around the polymer which leads to its condensation by solvent exclusion. STM has then enabled us to visualize both the condensation and the peptide binding phenomena.
Circular dichroism signals such as the one we get from the poly((dA-dT)-(dA-dT))/peptide complexes in solution are correlated with the presence of a cholesteric liquid crystalline phase [9]. This phase consists of layers of parallel double helices stacked together with a slight twist between layers. We have never visualiT~l more than one layer, as assessed by the corrugation values ranging from 2.5 nm (figs. 3 and 4) to a maximum of 4 nm (fig. 2). As we can exclude that the tip would pierce through the uppermost layers to image only the layer in contact with the graphite, we have to conclude either that spreading conditions led to monolayers or that stacked layers gave unstable tunneling conditions. We were intrigued to observe no periodicity along the closely packed molecules in contrast with the situation for isolated ones. We then suggest that the stronger interaction with the graphite of the condensed molecules somehow alters the polymer structure. As for the periodicity of the isolated molecules (fig. 5), we have found two types of helical pitch values, 6 and 7.5 nm. We think that 6 nm can be assigned to the pitch of poly((dA-dT). (dA-dT)). Indeed, on the one hand, this polymer was shown to have a larger periodicity than standard B-form DNA (4.2 nm instead of 3.4 nm according to ref. [20]), and on the other hand, large periodicities up to 6.3 nm, as depicted by STM, have already been reported by Beebe et al. [21]. The other pitch value of 7.5 nm probably corresponds to places where the rather bulky peptide of 18 amino acids is bound. Another consequence of the peptide binding is the irregularity in the apparent width of the isolated molecules shown in fig. 5. This thickening of the molecule at the peptide binding sites might also explain the rather large interhelical spacing (6.3 nm) of the ordered array presented in fig. 2.
5. Conclusion
Our results again illustrate the potential of scanning tunneling microscopy to resolve the ultrastructure of biomolecules. The liquid crystalline phase of poly((dA-dT). (dA-dT)) that this tech-
R. Coratger et al. / S T M of poly((dA-dT) . (dA-dT))
nique has enabled us to observe is undoubtedly a well suited system. Indeed, the hquid crystalline organization of the molecules is almost a prerequisite to their adsorption on the graphite substrate, according to a yet unknown mechanism, as already stated in other works. It is worth remarking that we have been able to visualize not only how the nucleic acid is organized but also how its parameters have been modified by the binding of the peptide used as a condensation factor. STM technique should then be useful in the study of interactions between nucleic acids and proteins.
References [1] G. Binnig and H. Rohrer, Surf. Sci. 126 (1983) 237. [2] G. Binnig, H. Rohrer, Ch. Gerber and E. Weibel, Phys. Rev. Lett. 50 (1983) 120. [3] G. Binnig, H. Rohrer, Ch. Gerber and E. Weibel, Phys. Rev. Lett. 49 (1982); Appl. Phys. Lett. 40 (1982) 178. [4] A. Baro, R. Miranda, J. Alaman, N. Garcia, G. Binnig, H. Rohrer, Ch. Gerber and J.L. Carrasco, Nature 315 (1985) 253. [5] B.L. Blackford, M.O. Watanabe, D.C. Dahn, M.H. Jericho, G. Southham and T.J. Beveridge, Ultramicroscopy 27 (1989) 427. [6] T. Sleator and R. Tycko, Phys. Rev. Lett. 60 (1988) 1418.
147
[7] M. Amrein, A. Stasiak, H. Gross, E. Stoll and G. Travaglini, Science 240 (1988) 514. [8] M. Amrein, R. Diirr, A. Stasiak, H. Gross and G. Travaglini, Science 243 (1989) 1708. [9] I. Tinoeo, C. Bustamante and M. Maestre, Annu. Rev. Biophys. Bioeng. 9 (1980) 107. [10] M. Erard, F. Lakhdar-Ghazal and F. Amalric, Eur. J. Biochem. 191 (1990) 19. [11] R. Coratger, F. Ajustron and J. Beauvillain, Spectra 2000 No. 147 (1990) 34. [12] R. Coratger, A. Claverie, F. Ajustron and J. Beauvillain, Surf. Sci. 227 (1990) 7. [13] J. Tersoff and D.R. Hamann, Phys. Rev. B 31 (1985) 805. [14] H.J. Mamin, E. Ganz, D.W. Abraham, R.E. Thomson and J. Clarke, Phys. Rev. B 34 (1986) 9015. [15] D. Keller, C. Bustamante and R.W. Keller, Proc. Natl. Acad. Sci. USA 86 (1989) 5356. [16] M. Salmeron, T. Beebe, J. Odriozola, T. Wilson, D. Ogletree and W. Sieklaus, J. Vac. Sci. Technol. A 8 (1990) 635. [17] G. Lee, P.G. Arscott, V.A. Bloomfield and D.F. Evans, Science 244 (1989) 475. [18] R.W. Keller, D.D. Dunlap, C. Bustamante, D.J. Keller, R.G. Garcia, C. Gray and M.F. Maestre, J. Vac. Sci. Technol. A 8 (1990) 706. [19] R.L. Rill, Proc. Natl. Acad. Sci. USA 83 (1986) 342. [20] S. Brain, Nature New Biol. 232 (1971) 174. [21] T.P. Beebe, T.E. Wilson, D.F. Ogletree, J.E. Katz, R. Balhorn, M.B. Salmeron and W.J. Sieklaus, Science 243 (1989) 370.
