Activation-dependent clustering of the erbB2 receptor tyrosine kinase detected by scanning near-field optical microscopy

ErbB2 (HER2, Neu), a member of the epidermal growth factor (EGF) receptor tyrosine kinase family, is often overexpressed in breast cancer and other malignancies. ErbB2 homodimerizes but also presents as a common auxiliary subunit of the EGF and heregulin receptors (erbB1 or EGFR; and erbB3-4, respectively), with which it heteroassociates. ErbB2 is generally regarded as an orphan (ligand-less) receptor with a very potent kinase domain activated either via its associated partners or constitutively as a consequence of discrete mutations. It follows that the extent and regulation of its cell surface interactions are of central importance. We have studied the large-scale association pattern of erbB2 in quiescent and activated cells labeled with fluorescent anti-erbB2 monoclonal antibodies using scanning near-field optical microscopy (SNOM). ErbB2 was found to be concentrated in irregular membrane patches with a mean diameter of approx. 0.5 microm in nonactivated SKBR3 and MDA453 human breast tumor cells. The average number of erbB2 proteins in a single cluster on nonactivated SKBR3 cells was about 10(3). Activation of SKBR3 cells with EGF, heregulin as well as a partially agonistic anti-erbB2 monoclonal antibody led to an increase in the mean cluster diameter to 0.6-0.9 microm, irrespective of the ligand. The EGF-induced increase in the erbB2 cluster size was inhibited by the EGFR-specific tyrosine kinase inhibitor PD153035. The average size of erbB2 clusters on the erbB2-transfected line of CHO cells (CB2) was similar to that of activated SKBR3 cells, a finding correlated with the increased base-line tyrosine phosphorylation of erbB2 in cells expressing only erbB2. We conclude that an increase in cluster size may constitute a general phenomenon in the activation of erbB2.

Fluorescence resonance energy transfer detected by scanning near-field optical microscopy

Fluorescence resonance energy transfer (FRET) between excited fluorescent donor and acceptor molecules occurs via the Förster mechanism over a range of 1-10 nm. Because of the strong (sixth power) distance dependence of the signal, FRET has been used to assess the proximity of molecules in biological systems. We used a scanning near-field optical microscope (SNOM) operated in the shared-aperture mode using uncoated glass fibre tips to detect FRET between dye molecules embedded in polyvinyl alcohol films and bound to cell surfaces. FRET was detected by selective photobleaching of donor and acceptor fluorophores. We also present preliminary results on pixel-by-pixel energy transfer efficiency measurements using SNOM.

Comparison of fixation protocols for adherent cultured cells applied to a GFP fusion protein of the epidermal growth factor receptor

BACKGROUND

The analysis of the subcellular distribution of proteins is essential for the understanding of processes such as signal transduction. In most cases, the parallel analysis of multiple components requires fixation and immunofluorescence labeling. Therefore, one has to ascertain that the fixation procedure preserves the in vivo protein distribution. Fusion proteins with the green fluorescent protein (GFP) are ideal tools for this purpose. However, one must consider specific aspects of the fluorophore formation or degradation, i.e. reactions that may interfere with the detection of GFP fusion proteins.

METHODS

Fusion proteins of the epidermal growth factor receptor (EGFR) with GFP as well as free, soluble GFP stably or transiently expressed in adherent cultured cells served as test cases for comparing the distribution in vivo with that after fixation by conventional epifluorescence and laser scanning microscopy. Indirect immunofluorescence was employed to compare the distributions of the GFP signal and of the GFP polypeptide in the fusion protein.

RESULTS

Paraformaldehyde (PFA) fixation with subsequent mounting in the antifading agent Mowiol, but not in Tris- or HEPES buffered saline, led to a partial redistribution of the EGFR from the plasma membrane to the perinuclear region. The redistribution was confirmed with the GFP and EGFR immunofluorescence. The in vivo distribution in Mowiol mounted cells was preserved if cells were treated with a combined PFA/methanol fixation procedure, which also retained the fluorescence of soluble GFP. The anti-GFP antiserum was negative for the N-terminal fusion protein.

CONCLUSIONS

The combined PFA/methanol protocol is universally applicable for the fixation of transmembrane and soluble cytoplasmic proteins and preserves the fluorescence of GFP.

Picosecond multiphoton scanning near-field optical microscopy

We have implemented simultaneous picosecond pulsed two- and three-photon excitation of near-UV and visible absorbing fluorophores in a scanning near-field optical microscope (SNOM). The 1064-nm emission from a pulsed Nd:YVO4 laser was used to excite the visible mitochondrial specific dye MitoTracker Orange CM-H2TMRos or a Cy3-labeled antibody by two-photon excitation, and the UV absorbing DNA dyes DAPI and the bisbenzimidazole BBI-342 by three-photon excitation, in a shared aperture SNOM using uncoated fiber tips. Both organelles in human breast adenocarcinoma cells (MCF 7) and specific protein bands on polytene chromosomes of Drosophila melanogaster doubly labeled with a UV and visible dye were readily imaged without photodamage to the specimens. The fluorescence intensities showed the expected nonlinear dependence on the excitation power over the range of 5-40 mW. An analysis of the dependence of fluorescence intensity on the tip-sample displacement normal to the sample surface revealed a higher-order function for the two-photon excitation compared to the one-photon mode. In addition, the sample photobleaching patterns corresponding to one- and two-photon modes revealed a greater lateral confinement of the excitation in the two-photon case. Thus, as in optical microscopy, two-photon excitation in SNOM is confined to a smaller volume.

FRET microscopy demonstrates molecular association of non-specific lipid transfer protein (nsL-TP) with fatty acid oxidation enzymes in peroxisomes

The fate of fluorescently labeled pre-nsL-TP (Cy3-pre-nsL-TP) microinjected into BALB/c 3T3 fibroblasts was investigated by confocal laser scanning microscopy. The protein exhibited a distinct punctate fluorescence pattern and colocalized to a high degree with the immunofluorescence pattern for the peroxisomal enzyme acyl-CoA oxidase. Proteolytic removal of the C-terminal leucine of the putative peroxisomal targeting sequence (AKL) resulted in a diffuse cytosolic fluorescence. These results indicate that microinjected Cy3-pre-nsL-TP is targeted to peroxisomes. The association of nsL-TP with peroxisomal enzymes was investigated in cells by measuring fluorescence resonance energy transfer (FRET) between the microinjected Cy3-pre-nsL-TP and Cy5-labeled antibodies against the peroxisomal enzymes acyl-CoA oxidase, 3-ketoacyl-CoA thiolase, bifunctional enzyme, PMP70 and catalase. The technique of photobleaching digital imaging microscopy (pbDIM), used to quantitate the FRET efficiency on a pixel-by-pixel basis, revealed a specific association of nsL-TP with acyl-CoA oxidase, 3-ketoacyl-CoA thiolase and bifunctional enzyme in the peroxisomes. These observations were corroborated by subjecting a peroxisomal matrix protein fraction to affinity chromatography on Sepharose-immobilized pre-nsL-TP. Acyl-CoA oxidase was retained. These studies provide strong evidence for a role of nsL-TP in the regulation of peroxisomal fatty acid beta-oxidation, e.g. by facilitating the presentation of substrates and/or stabilization of the enzymes.

FTIR and UV spectroscopy of parallel-stranded DNAs with mixed A*T/G*C sequences and their A*T/I*C analogues

The infrared spectra of parallel-stranded (ps) hairpin duplexes with mixed A*T/G*C composition and either isolated or sequential G*C pairs were studied in comparison with antiparallel-stranded (aps) duplexes and a corresponding set of molecules with hypoxanthine as a G base analogue lacking the exocyclic amino group. The ps duplexes showed the characteristic bands for the C2=O2 and C4=O4 stretching vibrations of thymine residues in trans-Watson-Crick A*T pairing at 1683 and 1668 cm-1. The latter band was superimposed on the stretching vibration of the free C6=O6 group of guanine. Substitution of guanine by hypoxanthine inhibited the formation of ps hairpin duplexes whatever the sequence, demonstrating that in the H-bonding between G and C the 2-NH2 group is necessary for stabilizing all of the investigated ps duplexes. This result is in agreement with a model of trans-Watson-Crick G*C base pairs with two H-bonds [N2H2(G)-N3(C) and N1H(G)-O2(C)]. However, trans-Watson-Crick A*T and G*C base pairs with two H-bonds are not isomorphous, which may explain the decreased stability of the ps, but not the aps, duplexes upon increasing the number of A*T/G*C steps. Molecular modeling studies performed on two of the ps duplexes reveal the existence of propeller twist for avoiding a clash between the N2(G) and N4(C) amino groups, and favorable stacking of sequential G*C base pairs. The optimized hairpin ps duplexes invariably incorporated G*C base pairs with two H-bonds, regardless of the initial structures adopted for the force field calculations.

Fluorescence correlation microscopy of cells in the presence of autofluorescence

Fluorescence correlation microscopy (FCM), the combination of fluorescence correlation spectroscopy (FCS) and digital microscopy (Brock and Jovin, 1998. Cell. Mol. Biol. 44:847-856), has been implemented for measuring molecular diffusion and association in living cells with explicit consideration of autocorrelations arising from autofluorescence. Autofluorescence excited at 532 nm colocalizes with mitochondria, has flavin-like spectral characteristics, exhibits relaxation times characteristic for the diffusion of high-molecular-weight proteins, and depends on the incubation conditions of the cells. These time- and location-dependent properties preclude the assignment of universal background parameters. The lower limit for detection of microinjected dextran molecules labeled with the carboxymethylindocyanine dye Cy3 was a few thousand molecules per cell, and the diffusion constant of 1.7 x 10(-7) cm2/s agreed well with values measured with other methods. Based on the fluorescence signal per molecule (fpm) and the molecule number derived from autocorrelation analysis, a new method is devised to define intracellular association states. We conclude that FCM is a powerful, noninvasive method for probing molecular interactions in femtoliter volume elements within defined subcellular locations in living cells.

Improved restoration from multiple images of a single object: application to fluorescence microscopy

We present an approach for the combined restoration of multiple different images of a single object. A linear Tikhonov filter adapted for this purpose is derived in detail. Nonlinear constrained algorithms can also be adapted, and we illustrate this possibility for an iterative constrained Tikhonov algorithm. Both the linear and the iterative constrained Tikhonov algorithms were used to analyze performance in fluorescence confocal imaging by use of simulated and experimental data. One can improve the quality of restored confocal images significantly if the signal that normally is rejected by the detection pinhole of a confocal laser scanning microscope is also recorded on a separate detector such that the two recorded signals are used together for image restoration according to the proposed algorithms.

Continuous wave two-photon scanning near-field optical microscopy

We have implemented continuous-wave two-photon excitation of near-UV absorbing fluorophores in a scanning near-field optical microscope (SNOM). The 647-nm emission of an Ar-Kr mixed gas laser was used to excite the UV-absorbing DNA dyes DAPI, the bisbenzimidazole Hoechst 33342, and ethidium bromide in a shared aperture SNOM with uncoated fiber tips. Polytene chromosomes of Drosophila melanogaster and the nuclei of 3T3 Balb/c cells labeled with these dyes were readily imaged. The fluorescence intensity showed the expected nonlinear (second order) dependence on the excitation power in the range of 8-180 mW. We measured the fluorescence intensity as a function of the tip-sample displacement in the direction normal to the sample surface in the single- and two-photon excitation modes (SPE, TPE). The fluorescence intensity decayed faster in TPE than in SPE.

Two-photon near- and far-field fluorescence microscopy with continuous-wave excitation

We report on scanning far- and near-field two-photon microscopy of cell nuclei stained with DAPI and bisbenzimidazole Hoechst 33342 (BBI-342) with the 647-nm laser line of a cw ArKr mixed-gas laser. Two-photon-excited fluorescence images are obtained for 50-200 mW of average power at the sample. A nearly quadratic dependence of fluorescence intensity on laser power confirmed the two-photon effect. The nonlinearity was further supported by evidence of three-dimensional sectioning in a scanning far-field microscope. We find that the cw two-photon irradiation sufficient for imaging within typically 5 s does not significantly impair cell cycling of BBI-342-labeled live cells. Finally, high-resolution imaging in scanning near-field microscopy with good contrast is demonstrated.

Fluorescence correlation microscopy (FCM)-fluorescence correlation spectroscopy (FCS) taken into the cell

Confocal fluorescence correlation spectroscopy (FCS) and other confocal spectroscopic techniques are ideally suited for the analysis of molecular interactions at the subcellular level. However, one requires exact positioning in three dimensions within the cell. Our instrument integrates FCS with high sensitivity digital imaging microscopy and high precision positioning. We present first measurements of intracellular FCS, with specification of the instrumental requirements and methods of data analysis. We propose the term fluorescence correlation microscopy (FCM) for this extended modality of FCS.

Cell biological applications of scanning near-field optical microscopy (SNOM)

Scanning near-field optical microscopy (SNOM) yields high-resolution topographic and optical images and is an important technique for visualizing biological systems. We summarize the literature on SNOM of biological systems and present some of our recent applications in cellular biology. These include studies of: i) the binding of fluorescently conjugated lectins to cell surface glycoproteins on 3T3 Balb/c cells, ii) molecular interactions by fluorescence resonance energy transfer using photobleaching techniques, and iii) green fluorescent protein (GFP) expressed in bacteria.

Temporally and spectrally resolved imaging microscopy of lanthanide chelates

The combination of temporal and spectral resolution in fluorescence microscopy based on long-lived luminescent labels offers a dramatic increase in contrast and probe selectivity due to the suppression of scattered light and short-lived autofluorescence. We describe various configurations of a fluorescence microscope integrating spectral and microsecond temporal resolution with conventional digital imaging based on CCD cameras. The high-power, broad spectral distribution and microsecond time resolution provided by microsecond xenon flashlamps offers increased luminosity with recently developed fluorophores with lifetimes in the submicrosecond to microsecond range. On the detection side, a gated microchannel plate intensifier provides the required time resolution and amplification of the signal. Spectral resolution is achieved with a dual grating stigmatic spectrograph and has been applied to the analysis of luminescent markers of cytochemical specimens in situ and of very small volume elements in microchambers. The additional introduction of polarization optics enables the determination of emission polarization; this parameter reflects molecular orientation and rotational mobility and, consequently, the nature of the microenvironment. The dual spectral and temporal resolution modes of acquisition complemented by a posteriori image analysis gated on the spatial, spectral, and temporal dimensions lead to a very flexible and versatile tool. We have used a newly developed lanthanide chelate, Eu-DTPA-cs124, to demonstrate these capabilities. Such compounds are good labels for time-resolved imaging microscopy and for the estimation of molecular proximity in the microscope by fluorescence (luminescence) resonance energy transfer and of molecular rotation via fluorescence depolarization. We describe the spectral distribution, polarization states, and excited-state lifetimes of the lanthanide chelate crystals imaged in the microscope.

Distamycin-stabilized antiparallel-parallel combination (APC) DNA

The formation of Antiparallel-Parallel-Combination (APC) DNA, a liner duplex with a segment of parallel-stranded (ps) helix flanked by conventional B-DNA, was tested with a number of synthetic oligonucleotides. The groove-binding ligand distamycin A (DstA) was used to stabilize the ps segment comprising five A x T base pairs. Two drug molecules bound per APC, one in each of the two equivalent grooves characteristic of ps-DNA. APC-DNA, reference molecules and their complexes with DstA were analysed by several methods: circular dichroism and absorption spectroscopy, thermal denaturation, chemical modification, and molecular modeling. The dye binding stoichiometry differed significantly due to inherent structural differences in the groove geometries of ps-DNA (trans base pairs, similar grooves) and conventional antiparallel-stranded (aps) B-DNA (cis base pairs, distinct major and minor grooves). The data support the existence of APC folding in solution.

Interferometric measurement of fluorescence excitation spectra

A Michelson interferometer has been adapted as an excitation source for fluorescence spectroscopy. A moving fringe pattern was generated by linear displacement of the movable mirror of the Michelson interferometer coupled to a xenon-arc lamp. This spectrally modulated source was monitored by a reference photomultiplier and used for exciting a Rhodamine B solution. The fluorescence emission at >645 nm was detected by a second photomultiplier. The two interferograms were acquired by a dual-channel digital oscilloscope, and their discrete Fourier transforms and corresponding power spectra were generated in a computer. The power spectrum of the emission signal represented the excitation spectrum, as was confirmed by comparison with the absorption spectrum of Rhodamine B. Thisoptical arrangement is well suited for acquiring fluorescence excitation spectra in the optical microscopy of biological specimens.

Tumor suppressor protein p53 binds preferentially to supercoiled DNA

Wild type human tumor suppressor protein p53 (expressed in insect cells) binds strongly to negatively supercoiled (sc) plasmid DNA at a native superhelix density, as evidenced by electrophoretic retardation of scDNA in agarose gels and imaging by scanning force microscopy (SFM). The binding occurs both in the presence and absence of the p53 consensus sequence. At relatively low p53/DNA ratios, binding of p53 to scDNA results in the appearance of several retarded DNA bands on the gels, similar to a conventional topoisomer ladder generated enzymatically. However, after removal of p53 by deproteination, the original mobility of the scDNA is recovered, indicating that the reduction of torsional stress accompanying p53 binding does not reflect changes in linking number. In DNA samples partially relaxed by topoisomerase I p53 binds preferentially to the scDNA molecules with the largest negative superhelix density. SFM imaging of the p53/scDNA complex reveals a partial or total relaxation of the compact scDNA, the degree of which increases with the number of bound p53 molecules. Competition assays with linear DNA reveal a preference of p53 for scDNA. In addition, scDNA induces dissociation of p53 from a preformed complex with a DNA fragment (474 bp) containing the consensus sequence. We conclude that the affinity of p53 for negatively supercoiled DNA is greater than that for the consensus sequence in linear fragments. However, thermally denatured linearized plasmid DNA is efficient in competing for the binding of p53 to scDNA, although the first retarded band (presumed to contain one bound p53 molecule) is retained in the case of the plasmid containing the consensus sequence. Thus, it appears that interactions involving both the core domain and the C-terminal domain regulate the binding of p53 to scDNA. The above results are not restricted to human p53; the wild type rat p53 protein also results in the retardation of scDNA on agarose gels. The biological implications of the novel DNA binding activities of p53 are discussed.

Striving for atomic resolution in biomolecular topography: the scanning force microscope (SFM)

The invention in 1986 of scanning force microscopy (SFM) provided a new and powerful tool for the investigation of biological structures. SFM yields a three-dimensional view at nanometer resolution of the surface topography associated with biological objects. The potential for imaging either macromolecules or biomolecules and cells under native (physiological) conditions is currently being exploited to obtain functional information at the molecular level. In addition, the forces involved in individual bimolecular interactions are being assessed under static and dynamic conditions. In this report we focus on the imaging capability of the SFM. The rather broad spectrum of applications represented is intended to orient the prospective user of biological SFM.

Actinomycin D binding to single-stranded DNA: sequence specificity and hemi-intercalation model from fluorescence and 1H NMR spectroscopy

We have studied the sequence specificity in the binding of the potent antitumor drug actinomycin D (AMD) to single-stranded DNA (ssDNA) by fluorescence and NMR spectroscopy and by molecular modeling. The significant absorption and emission changes accompanying the interaction of the fluorescent derivative 7-amino-AMD with DNAs varying in length and base composition were used to calculate affinity constants for the drug-DNA complexes. The guanine-containing trinucleotide sequences AGT, AGA, and TGT embedded within 25-base oligonucleotides, constituted favorable binding sites. In contrast, the sequence TGA did not bind the drug appreciably. Among the DNAs studied, the highest affinity was for the tetranucleotide sequence TAGT. The binding was length dependent, an oligonucleotide of at least 14 bases being required for effective complex formation (Ka > 10(4) M1=). AMD also bound to poly(d(AGT)). Gel electrophoresis confirmed that the complex was formed between the drug and individual unstructured DNA strands. The 1H NMR spectra of oligonucleotides containing the TAGT site and their complexes with AMD provided further insight into the mode(s) of interaction. A comparison of the measured chemical shifts with those estimated from ring-current calculations provided strong evidence for a hemi-intercalation of AMD between the A and G purine bases with a preference for one of two possible relative orientations. The latter were modeled as complexes with the sequence T3AGT3 and refined by force field calculations with the AMBER program. The biological implications for this novel form of interaction of AMD with single-stranded DNA are discussed.

Fluorescently labeled phosphatidylinositol transfer protein isoforms (alpha and beta), microinjected into fetal bovine heart endothelial cells, are targeted to distinct intracellular sites

Upon permeabilization of Swiss mouse 3T3 fibroblasts, an isoform of phosphatidylinositol transfer protein (PI-TP) was preferentially retained, a major part of which was associated with the perinuclear Golgi system (K. J. de Vries, A. Momchilova-Pankova, G. T. Snoek, and K. W. A. Wirtz, Exp. Cell Res. 215, 109-113, 1994). In the present study, the intracellular localization of this isoform (PI-TP beta) and the regular form (PI-TP alpha) was investigated in fetal bovine heart endothelial cells by microinjection of fluorescently labeled analogs followed by confocal laser scanning microscopy. The PI-TP alpha and PI-TP beta used were purified from bovine brain cytosol and covalently labeled with sulfoindocyanine dyes. By this novel method it was found that PI-TP beta was preferentially associated with perinuclear membrane structures whereas PI-TP alpha was predominantly present in the nucleus and in the cytoplasm. This intracellular localization was confirmed by indirect immunofluorescence indicating that the fluorescently labeled PI-TP alpha and PI-TP beta were targeted to the same sites as their endogeneous counterparts.

Imaging the intracellular trafficking and state of the AB5 quaternary structure of cholera toxin

The subcellular localization and corresponding quaternary state of fluorescent labelled cholera toxin were determined at different time points after exposure to living cells by a novel form of fluorescence confocal microscopy. The compartmentalization and locus of separation of the pentameric B subunits (CTB) from the A subunit (CTA) of the toxin were evaluated on a pixel-by-pixel (voxel-by-voxel) basis by measuring the fluorescence resonance energy transfer (FRET) between CTB labelled with the sulfoindocyanine dye Cy3 and an antibody against CTA labelled with Cy5. The FRET efficiency was determined by a new technique based on the release of quenching of the Cy3 donor after photodestruction of the Cy5 acceptor in a region of interest within the cell. The results demonstrate vesicular transport of the holotoxin from the plasma membrane to the Golgi compartment with subsequent separation of the CTA and CTB subunits. The CTA subunit is redirected to the plasma membrane by retrograde transport via the endoplasmic reticulum whereas the CTB subunit persists in the Golgi compartment.

Microspectroscopic imaging tracks the intracellular processing of a signal transduction protein: fluorescent-labeled protein kinase C beta I

We have devised a microspectroscopic strategy for assessing the intracellular (re)distribution and the integrity of the primary structure of proteins involved in signal transduction. The purified proteins are fluorescent-labeled in vitro and reintroduced into the living cell. The localization and molecular state of fluorescent-labeled protein kinase C beta I isozyme were assessed by a combination of quantitative confocal laser scanning microscopy, fluorescence lifetime imaging microscopy, and novel determinations of fluorescence resonance energy transfer based on photobleaching digital imaging microscopy. The intensity and fluorescence resonance energy transfer efficiency images demonstrate the rapid nuclear translocation and ensuing fragmentation of protein kinase C beta I in BALB/c3T3 fibroblasts upon phorbol ester stimulation, and suggest distinct, compartmentalized roles for the regulatory and catalytic fragments.

Sequence-specific labeling of superhelical DNA by triple helix formation and psoralen crosslinking

Site-specific labeling of covalently closed circular DNA was achieved by using triple helix-forming oligonucleotides 10, 11 and 27 nt in length. The sequences consisted exclusively of pyrimidines (C and T) with a reactive psoralen at the 5'-end and a biotin at the 3'-end. The probes were directed to different target sites on the plasmids pUC18 (2686 bp), pUC18/4A (2799 bp) and pUC1 8/4A-H 1 (2530 bp). After triple helix formation at acid pH the oligonucleotides were photocrosslinked to the target DNAs via the psoralen moiety, endowing the covalent adduct with unconditional stability, e.g. under conditions unfavorable for preservation of the triplex, such as neutral pH. Complex formation was monitored after polyacrylamide gel electrophoresis by streptavidin-alkaline phosphatase (SAP)-induced chemiluminescence. The yield of triple helix increased with the molar ratio of oligonucleotide to target and the length of the probe sequence (27mer > 11mer). The covalent adduct DNA were visualized by scanning force microscopy (SFM) using avidin or streptavidin as protein tags for the biotin group on the oligonucleotide probes. We discuss the versatility of triple helix DNA complexes for studying the conformation of superhelical DNA.

Determination of DNA helical handedness by fluorescence resonance energy transfer

Fluorescence resonance energy transfer (FRET) has been used to determine the helical handedness, twist and rise of different DNA conformations. The approach is based on the construction of a set of molecules consisting of two fused helical segments, one of which is in a known reference helical form. The duplexes are covalently labeled at one end with a donor and at the other with an acceptor. By systematically shifting the position of the junction while maintaining constant the total length in base-pairs, the variation in the efficiency of energy transfer can be shown to depend primarily and sensitively on the differences in helical twist and rise of the two constituent segments. If the latter have the same helical sense, one predicts a FRET signal that is a monotonic function of the junctional position. In contrast, a periodic function arises when two segments are of opposite handedness. The formalism includes explicit consideration of dye orientation (the dipole-dipole orientation factor kappa) and an implementation valid for single helix molecules, and introduces new functions of measured fluorescence signals for establishing the FRET efficiency. The method has been applied to a family of oligonucleotides forming hairpin duplexes containing an antiparallel-stranded (aps) d(m5C.G)m segment labeled at the 5' end with fluorescein (donor) and a second parallel- stranded d(A.T)N-m segment (psAt-DNA) labeled at the hairpin loop with the sulfoindocyanine dye Cy3. The segment lengths were in the range 4 to 12, but the total length N was maintained constant at 16. The d(m5C.G) sequence was chosen due to its capacity for adopting a B or a Z conformation at low and high concentrations of salt, respectively. The parallel-stranded d(A.T) sequence served as the second segment in order to determine the helical rise and twist of psAT-DNA, presumed to be right-handed from molecular modeling and a prior study of topologically constrained DNA. A Z-DNA/ps-DNA junction was created between the two segments by inducing a B-Z transition in d(m5C.G)m with MgCl2. The range of required salt concentration was established by circular dichroism measurements. FRET efficiency values of 0.38 to 0.41 were obtained for the oligonucleotides with the d(m5C.G) segment in the B conformation. In contrast, upon induction of the B-Z transition the FRET efficiency was a decreasing function of the d(m5C.G) content (0.38 to 0.28 for m = 6 to 12). Helical parameters were estimated from functional fits of the data, and were consistent with the known properties of B- and Z-DNAs and with the conclusion that psAT-DNA has a helical rise and twist close to that of B-DNA. The approach outlined here is not restricted to DNA but can be applied to other helical structures, e.g. RNA, proteins, and protein-nucleic acid complexes.

Parallel-stranded duplex DNA containing dA.dU base pairs

DNA oligonucleotides with dA and dU residues can form duplexes with trans d(A.U) base pairing and the sugar-phosphate backbone in a parallel-stranded orientation, as previously established for oligonucleotides with d(A.T) base pairs. The properties of such parallel-stranded DNA (ps-DNA) 25-mer duplexes have been characterized by absorption (uv), CD, ir, and fluorescence spectroscopy, as well as by nuclease sensitivity. Comparisons were made with duplex molecules containing (a) dT in both strands, (b) dU in one strand and dT in the second, and (c) the same base combinations in reference antiparallel-stranded (aps) structures. Thermodynamic analysis revealed that total replacement of deoxythymine by deoxyuridine was accompanied by destabilization of the ps-helix (reduction in Tm by -13 degrees C in 2 mM MgCl2, 10 mM Na-cacodylate). The U-containing ps-helix (U1.U2) also melted 14 degrees C lower than the corresponding aps-helix under the same ionic conditions; this difference was very close to that observed between ps and aps duplexes with d(A.T) base pairs. Force field minimized structures of the various ps and aps duplexes with either d(A.T) or d(A.U) base pairs ps/aps and dT/dU combinations are presented. The energy-minimized helical parameters did not differ significantly between the DNAs containing dT and dU.

Proximity relationships between the type I receptor for Fc epsilon (Fc epsilon RI) and the mast cell function-associated antigen (MAFA) studied by donor photobleaching fluorescence resonance energy transfer microscopy

Clustering of the mast cell function-associated antigen (MAFA) on the surface of rat mucosal type mast cells line 2H3 (RBL-2H3) leads to suppression of the secretory response induced by the type I Fc epsilon receptor (Fc epsilon RI). In order to establish a possible association between MAFA and Fc epsilon RI we measured fluorescence resonance energy transfer (FRET) between the MAFA-specific monoclonal antibody (mAb) G63 and Fc epsilon RI-bound ligands as well as between Fc epsilon RI-bound ligands themselves using the donor photobleaching FRET (pbFRET) technique. Average FRET efficiencies between 6 and 9% were determined after low-temperature incubation with fluorescent dye conjugated mAb G63 bound to MAFA (donor) and IgE bound to Fc epsilon RI (acceptor) on RBL-2H3 cells. Subsequent cross-linking of IgE by a polyvalent antigen caused no change in FRET efficiencies. These results suggest that the MAFA is located in the vicinity of the Fc epsilon RI on resting cells, and that clustering of the Fc epsilon RI leads to no significant change in the proximity of the two molecular species. In view of the sequence motif identified in the cytosolic tail of the MAFA and the observed changes in its phosphorylation upon antigen stimulation (Guthmann et al., Proc. Natl. Acad. Sci. USA 1995, 92: 9397-9401), the present study suggests that the secretory response inhibition by MAFA interferes with the signal transduction cascade initiated via the Fc epsilon RI. An additional finding was that clustering of the Fc epsilon RI by antigen showed a clear increase in the efficiency of FRET between Fc epsilon RI-bound IgE molecules conjugated with fluorescent donor and acceptor.

Oligomerization of epidermal growth factor receptors on A431 cells studied by time-resolved fluorescence imaging microscopy. A stereochemical model for tyrosine kinase receptor activation

The aggregation states of the epidermal growth factor receptor (EGFR) on single A431 human epidermoid carcinoma cells were assessed with two new techniques for determining fluorescence resonance energy transfer: donor photobleaching fluorescence resonance energy transfer (pbFRET) microscopy and fluorescence lifetime imaging microscopy (FLIM). Fluorescein-(donor) and rhodamine-(acceptor) labeled EGF were bound to the cells and the extent of oligomerization was monitored by the spatially resolved FRET efficiency as a function of the donor/acceptor ratio and treatment conditions. An average FRET efficiency of 5% was determined after a low temperature (4 degrees C) incubation with the fluorescent EGF analogs for 40 min. A subsequent elevation of the temperature for 5 min caused a substantial increase of the average FRET efficiency to 14% at 20 degrees C and 31% at 37 degrees C. In the context of a two-state (monomer/dimer) model for the EGFR, these FRET efficiencies were consistent with minimal average receptor dimerizations of 13, 36, and 69% at 4, 20, and 37 degrees C, respectively. A431 cells were pretreated with the monoclonal antibody mAb 2E9 that specifically blocks EGF binding to the predominant population of low affinity EGFR (15). The average FRET efficiency increased dramatically to 28% at 4 degrees C, indicative of a minimal receptor dimerization of 65% for the subpopulation of high affinity receptors. These results are in accordance with prior studies indicating that binding of EGF leads to a fast and temperature-dependent microclustering of EGFR, but suggest in addition that the high affinity functional subclass of receptors on quiescent A431 cells are present in a predimerized or oligomerized state. We propose that the transmission of the external ligand-binding signal to the cytoplasmic domain is effected by a concerted relative rotational rearrangement of the monomeric units comprising the dimeric receptor, thereby potentiating a mutual activation of the tyrosine kinase domains.

Scanning force microscopy of chromatin fibers in air and in liquid

We have adapted specimen preparation techniques of conventional electron microscopy for visualizing chromatin structures in the scanning force microscope (SFM) in air and in liquid. The beaded substructure of the nucleoprotein filament was obtained after hypotonic lysis of chicken erythrocytes and air drying, whereas supranucleosomal structures were preserved after treatment of cell nuclei with detergent. In the latter case, the nucleosomes were still distinct but appeared more condensed. A modified droplet diffusion-spreading technique of chromatin from Namalwa cells (a human B-lymphoid line) yielded a uniform filamentous morphology and similar fiber appearance. A reversible swelling of spread chromatin was observed upon exposure of air-dried samples to solutions differing in salt concentrations.

Scanning force microscopy of microtubules and polymorphic tubulin assemblies in air and in liquid

We have investigated microtubules (MTs) and polymorphic assemblies, formed in vitro from isolated microtubule protein, by scanning force microscopy (SFM) in air and in liquid. Immobilization of MTs was achieved by placing a drop of the assembly solution on a polylysine-coated coverslip. After washing with taxol and air drying, the characteristic microtubular fibrous morphology appeared in the SFM. The MTs formed a network similar to that obtained by transmission electron microscopy (TEM). A height of approximately 9.5 nm for dried MTs was computed from the surface topography. Glutaraldehyde fixation of the MTs yielded higher structures (approximately 14 nm), which swelled to approximately 20 nm after rehydration, a value close to the MT diameter of approximately 25 nm determined from TEM images of ultrathin sections. The protofilament pattern of the MTs and surface attached MT-associated proteins were not apparent from SFM, although the height along the long axis of the MTs appeared slightly modulated. In addition to MTs, various polymorphic tubulin assemblies including ribbons, hoops and double-walled MTs were visualized by SFM.

Structural hierarchy in the clustering of HLA class I molecules in the plasma membrane of human lymphoblastoid cells

Major histocompatibility complex (MHC) class I antigens in the plasma membranes of human T (HUT-102B2) and B (JY) lymphoma cells were probed by immunochemical reagents using fluorescence, transmission electron, and scanning force microscopies. Fluorescent labels were attached to monoclonal antibodies W6/32 or KE-2 directed against the heavy chain of HLA class I (A, B, C) and L368 or HB28 against the beta 2-microglobulin light chain. The topological distribution in the nanometer range was studied by photobleaching fluorescence resonance energy transfer (pbFRET) on single cells. A nonrandom codistribution pattern of MHC class I molecules was observed over distances of 2-10 nm. A second, nonrandom, and larger-scale topological organization of the MHC class I antigens was detected by indirect immunogold labeling and imaging by transmission electron microscopy (TEM) and scanning force microscopy (SFM). Although some differences in antigen distribution between the B- and T-cell lines were detected by pbFRET, both cell lines exhibited similar clustering patterns by TEM and SFM. Such defined molecular distributions on the surfaces of cells of the immune system may reflect an underlying specialization of membrane lipid domains and fulfill important functional roles in cell-cell contacts and signal transduction.

The scanning force microscopy of DNA in air and in n-propanol using new spreading agents

We present recent advances in DNA specimen preparation technique for scanning force microscopy (SFM) based on spreading on mica in the presence of cationic and non-ionic detergents. Reproducible DNA imaging in air and in n-propanol has been achieved in the presence of the non-ionic detergent 2,4,6-tris(dimethylaminomethyl) phenol (DMP-30) or the cationic detergent cetylpyridinium chloride (CP) in a microdrop containing nanograms of DNA. In an alternative procedure, a microdrop of detergent is applied to the surface just prior to the DNA. Quantitative image analysis yields as the apparent molecular dimensions of the DNA a width of approximately 7 nm and a height of approximately 0.7 nm, and delineates the problems of DNA metrology by SFM.

A three-dimensional model for the hammerhead ribozyme based on fluorescence measurements

For the understanding of the catalytic function of the RNA hammerhead ribozyme, a three-dimensional model is essential but neither a crystal nor a solution structure has been available. Fluorescence resonance energy transfer (FRET) was used to study the structure of the ribozyme in solution in order to establish the relative spatial orientation of the three constituent Watson-Crick base-paired helical segments. Synthetic constructs were labeled with the fluorescence donor (5-carboxyfluorescein) and acceptor (5-carboxytetramethylrhodamine) located at the ends of the strands constituting the ribozyme molecule. The acceptor helix in helix pairs I and III and in II and III was varied in length from 5 to 11 and 5 to 9 base pairs, respectively, and the FRET efficiencies were determined and correlated with a reference set of labeled RNA duplexes. The FRET efficiencies were predicted on the basis of vector algebra analysis, as a function of the relative helical orientations in the ribozyme constructs, and compared with experimental values. The data were consistent with a Y-shaped arrangement of the ribozyme with helices I and II in close proximity and helix III pointing away. These orientational constraints were used for molecular modeling of a three-dimensional structure of the complete ribozyme.

Imaging subcellular structures of rat mammary carcinoma cells by scanning force microscopy

Scanning force microscopy (SFM) was used for imaging subcellular structures of cultured rat mammary carcinoma cells dried in air. Identification of cellular substructures was achieved by immunofluorescence and specific fluorescence probes. Cells grown attached to a glass support exhibited submicrometer thickness in the dried state. Inside the nuclear domain the nucleoli appeared as prominent conical protrusions. Membrane extensions, microspikes and microvilli were well preserved at the cell periphery after fixation in glutaraldehyde vapor and air-drying and were distinguishable either as isolated elements or intercellular communications. The plasma membrane and soluble proteins were selectively removed with nonionic detergent in a buffer system. The mitochondria were concentrated primarily in the perinuclear space and exhibited a well defined filamentous shape. Their identity was confirmed by specific fluorescence staining with rhodamine 123. In the membrane-free system achieved by dry-cleaving of the sample surface, the cytoskeletal network was resolved as a complex mesh of actin-containing fiber bundles interwoven with a filigree arrangement of thinner filaments. The smallest fibrous substructures revealed by SFM with the scanning tips used to date were approximately 8 to 10 nm in height and 80 nm in width.

Probing specific molecular conformations with the scanning force microscope. Complexes of plasmid DNA and anti-Z-DNA antibodies

An anti-Z-DNA IgG antibody was used to probe for the left-handed Z-DNA conformation of a d(CG)11 insert in a negatively supercoiled plasmid DNA (pAN022). The complexes were spread on mica in the presence of a quaternary ammonium detergent benzyldimethylalkylammonium chloride and imaged with a scanning force microscope (SFM). The high affinity anti-Z-DNA antibody was retained even after restriction endonuclease cleavage of the DNA. The two arms in the product molecules had unequal lengths in conformity with the known location of the Z-DNA forming insert. Most complexes exhibited one IgG per DNA molecule. The bound antibodies were up to approximately 35 nm in diameter and extended approximately 2 nm from the mica surface. They were generally in a lateral orientation relative to the DNA, in accordance with prior chemical modification experimental data indicating a bipedal mode of binding for an anti-Z-DNA IgG. However, the SFM images also suggest that the DNA bends to accommodate the two Fab combining regions of the antibody. This study demonstrates the utility of the SFM for investigating conformation-dependent molecular recognition.

Parallel-stranded duplex DNA containing blocks of trans purine-purine and purine-pyrimidine base pairs

A 30 base pair parallel-stranded (ps) duplex ps-L1.L2 composed of two adjoined purine-purine and purine-pyrimidine sequence blocks has been characterized thermodynamically and spectroscopically. The 5'-terminal 15 residues in both strands ('left-half') consisted of the alternating d(GA)7G sequence that forms a ps homoduplex secondary structure stabilized by d(G.G) and d(A.A) base pairs. The 3'-terminal 15 positions of the sequence ('right-half') were combinations of A and T with complementary reverse Watson-Crick d(A.T) base pairing between the two strands. The characteristics of the full length duplex were compared to those of the constituent left and right halves in order to determine the compatibility of the two ps helical forms. The thermal denaturation curves and hyperchromicity profiles of all three duplexes determined by UV absorption spectroscopy were characteristic of ps-DNA, in accordance with previous studies. The thermodynamic properties of the 30 bp duplex corresponded within experimental error to the linear combination of the two 15-mers. Thus, the Tm and delta HvH of ps-L1.L2 in 10 mM MgCl2, derived from analyses according to a statistical mechanical formulation for the helix-coil transition, were 43 degrees C and 569 kJ mol-1, compared to 21 degrees C, 315 kJ mol-1 (ps-F5.F6) and 22 degrees C, 236 kJ mol-1 (ps-GA15). The UV absorption and CD spectra of ps-L1.L2 and the individual 15-mer ps motifs were also compared quantitatively. The sums of the two constituent native spectra (left+right halves) accurately matched that of the 30 bp duplex, with only small deviations in the 195-215 nm (CD) and 220-240 nm (absorption) regions. Based on analysis by native gel electrophoresis, the sequences studied formed duplex structures exclusively; there were no indications of higher order species. Chemical modification with diethyl pyrocarbonate showed no hyperreactivity of the junctional bases, indicating a smooth transition between the two parallel-stranded conformations. We conclude that under given salt conditions, oligonucleotides with normal primary chemical structures can readily form a parallel-stranded double helix based on blocks of very disparate non-canonical purine-purine and purine-pyrimidine base pairs and without perceptible destabilization at the junction. There are biological implications of these findings in relation to genetic structure and expression.

Synthesis and properties of fluorescent beta-adrenoceptor ligands

We describe the synthesis of bordifluoropyrromethene (BODIPY), fluorescein, and related fluorescent derivatives of the beta-adrenergic ligand CGP 12177. With these probes we screened insect (Sf9) cells stably transformed with the human beta 2-adrenoceptor gene and expressing (2-3.5) x 10(5) human beta 2-adrenoceptors per cell. Among these derivatives only BODIPY-CGP gave a receptor-specific signal sufficiently strong for measuring the on- and off-rate constants and the equilibrium dissociation constant of beta-adrenoceptor-specific binding by spectrofluorometry or photon counting. Similar KD values for BODIPY-CGP binding were obtained by kinetic measurements (approx. 250 pM) and under equilibrium conditions (400 +/- 180 pM), and these were in the same range as those obtained with [3H]CGP 12177 (200 +/- 32 pM). The cell-bound fluorescence could be quenched specifically with nonfluorescent CGP 12177 to near background levels. The disposition of the beta 2-adrenoceptors in BODIPY-CGP-stained Sf9 cells was mainly restricted to the cell surface at 4 and 30 degrees C. Hence, beta-adrenoceptor-expressing cells can be stained specifically with BODIPY-CGP, and beta-adrenoceptors on a single cell can be assessed by photon counting under the fluorescence microscope. Cells can also be scanned by fluorescence-activated flow cytometry.

Probing chromatin with the scanning force microscope

With the scanning force microscope (SFM), one can image the topography of biological material adsorbed at air-solid or liquid-solid interfaces with up to nanometer resolution. In principle, fixation, contrast enhancement, and labeling are not required. We have adapted specimen preparation techniques of conventional electron microscopy for visualizing chromatin ultrastructures in the SFM. A beaded substructure of the nucleoprotein filament was obtained after hypotonic lysis of chicken erythrocytes and air drying. The beads-on-a-string morphology of the basic nucleosomal assembly was well delineated. The nucleosomes appeared as round protrusions with an apparent height of 4-6 nm. The histogram of center-to-center distances between adjacent nucleosome cores along the filament axis had a peak at approximately 30 nm. Reversible changes in the three-dimensional structure were observed upon exposure of air-dried samples of metaphase chromosomes to solutions of different ionic strengths.

Dynamics and relative stabilities of parallel- and antiparallel-stranded DNA duplexes

The dynamics and stability of four DNA duplexes are studied by means of molecular dynamics simulations. The four molecules studied are combinations of 4, 15 bases long, single-stranded oligomers, F1, F2, F3, and F4. The sequence of these single strand oligomers are chosen such that F1-F2 and F3-F4 form parallel (ps) DNA double helices, whereas F1-F4 and F2-F3 form antiparallel-stranded (aps) DNA double helices. Simulations were done at low (100 K) and room (300 K) temperatures. At low temperatures the dynamics are quasi-harmonic and the analysis of the trajectories gives good estimates of the low frequency vibrational modes and density of states. These are used to estimate the linear (harmonic) contribution of local fluctuations to the configurational entropy of the systems. Estimates of the differences in enthalpy between ps and aps duplexes show that aps double helices are more stable than the corresponding ps duplexes, in agreement with experiments. At higher temperatures, the distribution of the fluctuations around the average structures are multimodal and estimates of the configurational entropy cannot be obtained. The multi-basin, nonlinear character of the dynamics at 300 K is established using a novel method which extracts large amplitude nonlinear motions from the molecular dynamics trajectories. Our analysis shows that both ps DNA exhibit much larger fluctuations than the two aps DNA. The large fluctuations of ps DNA are explained in terms of correlated transitions in the beta, epsilon, and zeta backbone dihedral angles.

Raman microspectroscopic study of low-pH-induced changes in DNA structure of polytene chromosomes

The effects of low-pH treatments on DNA structure in polytene chromosomes of Chironomus thummi thummi have been studied by Raman microspectroscopy. Measurements were carried out on chromosomes at low pH and on chromosomes reneutralized after a short exposure to low pH. Protonation of adenine residues and subsequent unpairing of adenine (A) and thymine (T) were found to commence already above pH 3.6 and are completed at pH 2.2. Protonation of the cytosine-guanine base pair also starts above pH 3.6. It leads to an intermediate protonated, most likely Hoogsteen-type, guanine-cytosine base pair. Unpairing of G and C residues starts between pH 2.6 and 2.2 and continues below pH 2.2. Brief treatment of chromosomes at a pH < or = 2.2, i.e., at a pH where both AT and GC base pairs are disrupted, was found to lead to irreversible changes in DNA conformation upon return to neutral pH. These were most pronounced after treatment with 45% acetic acid. More than 10% of the A and T residues was found to have adopted a non-B-DNA conformation. Evidence was found for denaturation of the B-form backbone. The amount of protein extracted from the chromosomes was strongly pH-dependent. Treatment at pH 3.6 did not cause noticeable protein extraction, while treatment with 45% acetic acid extracted more than 50% (by weight) of the chromosomal proteins.

Rotational mobility of the fibrinogen receptor glycoprotein IIb/IIIa or integrin alpha IIb beta 3 in the plasma membrane of human platelets

Integrin alpha IIb beta 3 or glycoprotein IIb/IIIa (GPIIb/IIIa, 228 kDa) is a Ca(2+)-dependent, noncovalent heterodimer of glycoproteins IIb (GPIIb or alpha IIb, 136 kDa) and IIIa (GPIIIa or beta 3, 92 kDa), which serves as the receptor for fibrinogen and other adhesive proteins at the surface of activated platelets. We have determined the microsecond-range rotational motions of alpha IIb beta 3 in resting platelets, in isolated plasma membranes, and reconstituted in 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) bilayers. The measurements were based on the time-resolved phosphorescence anisotropy [r(t)] of erythrosin-labeled F(ab) fragments [Er-F(ab)] of monoclonal antibodies bound to alpha IIb beta 3. In general, the r(t) decays were satisfactorily fitted to the sum of the two exponential terms and a constant, from which the initial anisotropy (r(in) approximately 0.05-0.11), the short (phi 1 approximately 1-14 microseconds) and the long (phi 2 approximately 15-60 microseconds) rotational correlation times, and the limiting anisotropy (r infinity approximately 0.02-0.07) were obtained. The observed values depended on the platelet preparation, temperature, Ca2+ concentration, and the antibody used. In accordance with data on the order parameter and the viscosity of the lipid bilayer of the platelet plasma membrane, phi 2 and r infinity of the alpha IIb beta 3-Er-F(ab) complexes in the three preparations decreased with the increase of temperature, the r(t) curves being fully reversible within the interval from 5 to 35 degrees C.(ABSTRACT TRUNCATED AT 250 WORDS)

Scanning force microscopy of circular and linear plasmid DNA spread on mica with a quaternary ammonium salt

Scanning force microscopy (SFM) offers the potential for subnanometer resolution in the investigation of nucleic acids, proteins, and their complexes. SFM is not bound by the requirement of classical transmission electron microscopy (TEM) for contrast enhancement through shadow casting or negative staining. A primary challenge, however, has been the reproducible fixation of samples on an atomically flat surface such as mica. We have developed a method for the routine imaging by SFM of supercoiled, relaxed, and linearized plasmid DNA, immobilized on freshly cleaved mica through the spreading action of benzyldimethylalkylammonium chloride (BAC) at micromolar concentrations. A reproducibly high yield of well-spread, dispersed molecules is obtained and background contamination is minimal. The contour lengths of the relaxed and linearized molecules imaged in air agree well with the helical rise (3.4 A/bp) of B-DNA in solution. We have also introduced the use of quantitative image analysis of SFM images to determine apparent molecular width and height over the entire molecular path.

Structure and drug interactions of parallel-stranded DNA studied by infrared spectroscopy and fluorescence

The infrared spectra of three different 25-mer parallel-stranded DNAs (ps-DNA) have been studied. We have used ps-DNAs containing either exclusively dA x dT base pairs or substitution with four dG x dC base pairs and have them compared with their antiparallel-stranded (aps) reference duplexes in a conventional B-DNA conformation. Significant differences have been found in the region of the thymine C = O stretching vibrations. The parallel-stranded duplexes showed characteristic marker bands for the C2 = O2 and C4 = O4 carbonyl stretching vibrations of thymine at 1685 cm-1 and 1668 cm-1, respectively, as compared to values of 1696 cm-1 and 1663 cm-1 for the antiparallel-stranded reference duplexes. The results confirm previous studies indicating that the secondary structure in parallel-stranded DNA is established by reversed Watson--Crick base pairing of dA x dT with hydrogen bonds between N6H...O2 and N1...HN3. The duplex structure of the ps-DNA is much more sensitive to dehydration than that of the aps-DNA. Interaction with three drugs known to bind in the minor groove of aps-DNA--netropsin, distamycin A and Hoechst 33258--induces shifts of the C = O stretching vibrations of ps-DNA even at low ratio of drug per DNA base pair. These results suggest a conformational change of the ps-DNA to optimize the DNA-drug interaction. As demonstrated by excimer fluorescence of strands labeled with pyrene at the 5'-end, the drugs induce dissociation of the ps-DNA duplex with subsequent formation of imperfectly matched aps-DNA to allow the more favorable drug binding to aps-DNA. Similarly, attempts to form a triple helix of the type d(T)n.d(A)n.d(T)n with ps-DNA failed and resulted in the dissociation of the ps-DNA duplex and reformation of a triple helix based upon an aps-DNA duplex core d(T)10.d(A)10.