Extreme conformational diversity in human telomeric DNA

Abstract

DNA with tandem repeats of guanines folds into G-quadruplexes made of a stack of G-quartets. In vitro, G-quadruplex formation inhibits telomere extension, and POT1 binding to the single-stranded telomeric DNA enhances telomerase activity by disrupting the G-quadruplex structure, highlighting the potential importance of the G-quadruplex structure in regulating telomere length in vivo. We have used single-molecule spectroscopy to probe the dynamics of human telomeric DNA. Three conformations were observed in potassium solution, one unfolded and two folded, and each conformation could be further divided into two species, long-lived and short-lived, based on lifetimes of minutes vs. seconds. Vesicle encapsulation studies suggest that the total of six states detected here is intrinsic to the DNA. Folding was severely hindered by replacing a single guanine, showing only the shortlived species. The long-lived folded states are dominant in physiologically relevant conditions and probably correspond to the parallel and antiparallel G-quadruplexes seen in high-resolution structural studies. Although rare under these conditions, the short-lived species determine the overall dynamics because they bridge the different long-lived species. We propose that these previously unobserved transient states represent the early and late intermediates toward the formation of stable G-quadruplexes. The major compaction occurs between the early and late intermediates, and it is possible that local rearrangements are sufficient in locking the late intermediates into the stably folded forms. The extremely diverse conformations of the human telomeric DNA may have mechanistic implications for the proteins and drugs that recognize G-rich sequences.

Fig. 1.

Probing potassium-dependent conformations of human telomeric DNA with FRET. (A) Schematic diagram of DNA construct. The G-quadruplex strand and the complementary stem strand were annealed to form a duplex stem. The quadruplex is depicted in the antiparallel conformation, and the rectangles represent the guanines (yellow in syn and red in anti conformations). Biotin is used to immobilize the DNA to a streptavidin-coated quartz surface, and FRET between tetramethylrhodamine (donor) and Cy5 (acceptor) was measured. (B) Single-molecule FRET histograms as a function of K⁺ concentration. (C) An expanded view of FRET histogram at 2 mM K⁺ showing three nonzero FRET peaks, named U (unfolded), F₁, and F₂ (folded). Also shown is the fit by four Gaussians. (D) Single-molecule FRET histograms as a function of K⁺ concentration for a mutant sequence as indicated above the graph. (E) Time traces (0.1-s integration time) of donor and acceptor intensities and corresponding FRET from a single molecule at 2 mM K⁺ exhibit interconversion between U, F₁, and F₂. A transition to zero FRET at 65 s is due to Cy5 blinking. Note that the transition from F₂ to F₁ occurred through U.

Fig. 2.

Extreme conformational diversity revealed by single-molecule dynamics. (A) Representative FRET time traces of single molecules at 2 mM K⁺ and room temperature (0.9-s integration time). Although some molecules show a single conformation for hundreds of seconds (bottom two traces, acceptor photobleaching events at 810 and 850 s), others show rapid fluctuations between U, F₁, and F₂ (third trace from the top). The top two traces show a mixed behavior in which switching events between the long- and short-lived species are observed. Segments of long- and short-lived species are marked by solid and dotted lines, respectively. (B) Fraction remaining in F₂ after a given time. The result is fitted by a double-exponential decay. (C) An example FRET time trace (0.1-s integration time) of a mutant sequence as indicated above the graph. The acceptor was photobleached at 115 s. (D) Bulk FRET efficiency, calculated as the ratio of the fluorescence intensity at the acceptor emission peak wavelength divided by the sum of the fluorescence intensities of the donor and the acceptor at their emission peak wavelengths, was measured as a function of time after adding 2 mM K⁺ at t = 0. The result is fitted by a double exponential.

Fig. 3.

Conformational distribution of various states. The relative populations of the folded (▪) vs. unfolded (○) conformations (Top), antiparallel (▪) vs. parallel (○) conformations (Middle), and long-lived (▪) vs. short-lived (○) species (Bottom). The results are shown for various K⁺ concentrations at room temperature (Left), and for various temperatures at 2 mM K⁺ (Center) and 100 mM K⁺ (Right).

Fig. 4.

Conformation transition rates between various states. (A and B) The rates of transition between the long- and short-lived species as a function of K⁺concentration at room temperature (A) and as a function of temperature at 2 mM K⁺ (B)(○, transition from ^SF to ^LU; □, transition from ^SU to ^LF; ⋄, transition from ^LU to ^SF; and ▵, transition from ^LF to ^SU). Superscripts S and L denote short- and long-lived respectively. (C and D) The rates of folding (▪) and unfolding (○) within the short-lived species as a function of K⁺ concentration at room temperature (C) and as a function of temperature at 2 mM K⁺ (D).

Fig. 5.

Proposed model for conformational identities and reaction pathways. The cartoon for ^LU denotes a disordered conformation of the ssDNA, and ^LF₁ and^LF₂ are depicted as the parallel and antiparallel G-quadruplex structures, respectively, holding K⁺ ions between G-quartets. The short-lived species (^SX) include^SU, ^SF, and ^{S 1} F₂. The solid arrows represent the proposed reaction paths, and the dashed arrows point toward the states that are favored upon increasing K⁺concentration or temperature.