Report

Protamine-Induced Condensation and Decondensation of the Same DNA Molecule

See allHide authors and affiliations

Science  01 Oct 1999:
Vol. 286, Issue 5437, pp. 120-123
DOI: 10.1126/science.286.5437.120

Abstract

The DNA in sperm and certain viruses is condensed by arginine-rich proteins into toroidal subunits, a form of packaging that inactivates their entire genome. Individual DNA molecules were manipulated with an optical trap to examine the kinetics of torus formation induced by the binding of protamine and a subset of its DNA binding domain, Arg6. Condensation and decondensation experiments with λ-phage DNA show that toroid formation and stability are influenced by the number of arginine-rich anchoring domains in protamine. The results explain why protamines contain so much arginine and suggest that these proteins must be actively removed from sperm chromatin after fertilization.

Protamine and other polycations have been shown to coil DNA into toroidal structures containing up to 60 kb of DNA (1–3). Individual bacteriophage appear to contain a single toroid folded inside the protein capsid (3), whereas a sperm cell contains as many as 50,000 toroids packed inside its nucleus (1). The protamines responsible for inducing torus formation and packaging DNA in maturing spermatids contain a series of arginine-rich anchoring domains (4) that bind to the phosphodiester backbone of DNA in a base sequence–independent fashion (5). One protamine molecule is bound to each turn [∼11 base pairs (bp)] of DNA (5, 6), and adjacent arginines in the anchoring domains interlock both strands of the helix. Arginine-rich sequences are also present in the proteins that package DNA in several viruses (7), but the viral proteins contain fewer anchoring domains per molecule.

In vitro studies using light scattering (8, 9), electron and atomic force microscopy (1, 2, 10), fluorescence microscopy (11, 12), and DNA elasticity measurements (13) have examined how protamine and other polycations induce torus formation. The interpretation of light-scattering experiments has been complicated by DNA aggregation, whereas electron and atomic force microscopy studies characterized only the structure of the final product. Toroid formation and the kinetics of the condensation process could not be observed by fluorescence microscopy because the molecules were not sufficiently extended. To examine toroid formation under conditions that preclude aggregation and precipitation and allow a detailed analysis of kinetics, we used an optical trap to isolate individual DNA molecules and fluorescence microscopy to monitor the formation of toroids in real time as they are induced by protamine (or Arg6) binding.

λ-phage DNA concatemers (20 to 80 μm long) were tagged at one end with a biotinylated oligonucleotide attached to a 1-μm streptavidin-coated polystyrene bead and stained with the intercalating dye YOYO-1 (14). These molecules were introduced through one port of a “bifurcated flow cell” (Fig. 1A) and the condensing agent protamine (or Arg6) through another port so that the two solutions flowed side by side with minimal mixing. An infrared optical trap (15) (Fig. 1B) was used to move an individual DNA molecule, via its attached bead, from the sample (DNA) side to the condensing agent (protein) side of the flow cell. The molecule was extended by the force of the flowing buffer, and its entire length became visible because of the fluorescence of the intercalated dye. Toroid formation (condensation) (Fig. 1C) was monitored in real time by measuring the change in length of the molecule as a function of time after moving it into the buffer stream containing protein (16).

Figure 1

(A) Top view of the flow cell (25) showing how the DNA molecules (attached to beads) and protamine enter the cell and form an interface (−−−) with little or no mixing. (B) Side view of the system showing the optical trap (orange) holding a bead attached to a single DNA molecule. The cell is illuminated from beneath by a 1-mW argon-ion laser, λ = 488 nm, to excite the YOYO-1 dye bound to the DNA. (C) Model of a DNA molecule condensing in protamine (protamine molecules not shown). The upper molecule shows the initiation of coiling and the lower molecule depicts the progression of coiling to form the torus.

Images documenting the progression of condensation for a 50-μm DNA concatemer in protamine (17) are shown in Fig. 2. The toroid, which appeared as a bright spot at the end of the DNA molecule, increased in brightness as it moved toward the bead. In 1.1 μM protamine, the condensation process was completed in ∼19 s. The change in length versus time for four different DNA molecules as they condensed in different concentrations of protamine is shown in Fig. 3A. As long as the DNA was extended by flow, torus formation initiated at the free end of the molecule, and its length decreased linearly with time, as predicted by Ostrovsky and Bar-Yam (18). The movement of the toroid often exhibited a jerky, start-and-stop motion, but when the process was repeated with the same DNA molecule (Fig. 4) this motion was not reproduced. Although these sporadic fluctuations in condensation rate cannot be DNA sequence or conformation dependent, they may relate to the cooperative nature of protamine binding and a nonrandom, incomplete coverage of the DNA molecule that develops as the protamines bind. Condensation (toroid movement) should decrease or stop if the toroid encountered regions that were not completely covered by protamine.

Figure 2

Progression of condensation of a 50-μm concatemer of λ-phage DNA attached to a 1-μm bead. The bead was held stationary by an optical trap and the DNA molecule is extended by the flow of the incoming buffer containing 1.1 μM protamine (v = 17.6 μm/s). Each time frame was captured after initiation of condensation as shown.

Figure 3

(A) Change in length versus time measured for four different DNA molecules condensed in a different protamine concentration (flow speed, v = 50 μm/s): (a) 3.1 μM (diamonds); (b) 1.6 μM (squares); (c) 1.2 μM (triangles); (d) 0.93 μM (circles). Lines are least-squares fits to the data points. (B) Condensation rates were determined by collecting data for about 200 individual DNA molecules condensed by protamine.

Figure 4

Condensation (blue dots) and decondensation (red dots) of the same DNA molecule in 107 μM Arg6 (flow speed, v = 28 μm/s) performed four successive times (A to D).

Experiments conducted at different protamine concentrations showed that the rate of condensation was limited by the rate of protamine binding to the DNA molecule. The change in rate (Fig. 3B) was linear, with a slope of 2.6 ± 0.47 μm/μM-s. This corresponds to a rate of protamine binding to DNA of 600 ± 110 molecules/μM-s. The rate of condensation was measured at two different concentrations of YOYO-1 (0.1 and 0.02 μM) to determine whether intercalated YOYO-1 molecules affect the condensation rate. No statistically significant difference in the rates was observed. The potential effect of buffer flow and its frictional force on toroid movement was also assessed by calculating the force that would be exerted on a sphere of equivalent volume to the toroid and by performing an experiment to test the effect of the force directly. The force was calculated by Stokes' law to be 0.43 pN (19). The effect of the force was determined experimentally by comparing the condensation rates measured for 30 DNA molecules condensed in 1.5 μM protamine over a range of buffer flow rates (20 to 70 μm/s). The slope of the fitted line through the data points (0.00 ± 0.038) indicated that the condensation rate was not affected by the velocity of the buffer flow.

To estimate the off-rate of protamine, we measured the decondensation rates of individual condensed DNA molecules after moving them back to the DNA side of the flow cell. We monitored decondensation, which occurs only when protamine dissociates from the DNA, for periods up to 25 min and used the lengthening (uncoiling) of the molecule to estimate the rate of protamine dissociation. The measured rate of increase in DNA length, 3.1 ± 1.3 nm/s, corresponds to a protamine off-rate of 0.71 molecule per second (20). At this rate, the complete dissociation of protamine from the sperm genome (1.5 × 109 bp) would require at least 6 years. Because sperm chromatin takes only 5 to 10 min to decondense after fertilization, these results support the hypothesis that protamine must be actively removed from DNA (21) once the chromatin enters the egg's cytoplasm. The protamine dissociation constant derived from these measurements, 1.17 ± 0.53 nM, is similar to values obtained by others for herring protamine (1.25 nM and 1.15 nM) with bulk DNA (9, 22).

Similar studies conducted with Ac-RRRRRR-amide (Arg6), one of three Arg6 anchoring domains present in protamine, revealed that a 68-fold higher concentration of Arg6 (160 μM) was required to achieve a condensation rate comparable to that measured for protamine (4.2 ± 1.1 μm/s). This suggests that either Arg6 has a much lower binding affinity for DNA, or the consecutive binding of three individual Arg6 molecules is statistically much less likely than the binding of a single protamine molecule. In contrast to the protamine complexes, DNA molecules condensed by Arg6 decondensed rapidly when pulled back across the interface and out of the buffer containing the peptide. Four successive condensation-decondensation measurements performed on the same molecule in 107 μM Arg6 are shown in Fig. 4. The mean condensation rate for these four measurements was 9.1 ± 2.5 μm/s, and the mean decondensation rate was 21.25 ± 2.9 μm/s. Because the Arg6 molecule binds to only 3 bp of DNA, these data indicate that the off-rate for Arg6 is 18 × 103 molecules per second—four orders of magnitude higher than the off-rate of protamine (23). The dissociation constant, 0.25 ± 0.08 mM, has not been measured previously.

Experiments conducted with concatemers of different lengths also indicated that there may be a limit to the amount of DNA that can be coiled into a toroid. Only one toroid was observed during the condensation of single λ-phage DNA molecules, whereas multiple toroids (Fig. 5) were observed when concatemers containing two to four molecules were condensed at a low flow rate (v < 10 μm/s). Although additional experiments must be conducted to verify the maximum length of sequence that can be coiled into a torus, the present studies suggest that one torus is formed for each length of λ-phage DNA (48.5 kb). Although it is not clear why such a limit should exist, this estimate (∼50 kb) is close to the 60 kb derived from toroid volumes measured by electron microscopy (10).

Figure 5

Condensation of a 45-μm concatemer of λ-phage DNA attached to a 1-μm bead held stationary by an optical trap in 1.1 μM protamine (flow speed, v = 6.1 μm/s). Three toroids were observed to form, sequentially starting at the free end of the DNA molecule.

By combining the use of an optical trap, a dual-port flow cell, and single molecule imaging, we have examined the condensation kinetics of individual DNA molecules, obtained estimates of the binding and dissociation rates for protamine and Arg6 in the absence of competing aggregation reactions, and monitored the formation and motions of toroids in real time to obtain results that only single molecule studies can provide. In addition to providing new insight into the mechanism of DNA condensation by protamine, this work presents an example of how experiments with single DNA molecules can be used to study the kinetics and biophysics of protein-DNA interactions.

REFERENCES AND NOTES

View Abstract

Stay Connected to Science

Navigate This Article