Role of Disulfide Bonds on DNA Packaging Forces in Bull Sperm Chromatin

Abstract

Short arginine-rich proteins called protamines mediate the near crystalline DNA packaging in most vertebrate sperm cells. Protamines are synthesized during spermiogenesis and condense the paternal genome into a transcriptionally inactive state in late-stage spermatids. Protamines from eutherian mammals, including bulls and humans, also contain multiple cysteine residues that form intra- and interprotamine sulfur-sulfur bonds during the final stages of sperm maturation. Although the cross-linked protamine network is known to stabilize the resulting nucleoprotamine structure, little is known about the role of disulfide bonds on DNA condensation in the mammalian sperm. Using small angle x-ray scattering, we show that isolated bull nuclei achieve slightly lower DNA packing densities compared to salmon nuclei despite salmon protamine lacking cysteine residues. Surprisingly, reduction of the intermolecular sulfur-sulfur bonds of bull protamine results in tighter DNA packing. Complete reduction of the intraprotamine disulfide bonds ultimately leads to decondensation, suggesting that disulfide-mediated secondary structure is also critical for proper protamine function. Lastly, comparison of multiple bull collections showed some to have aberrant x-ray scattering profiles consistent with incorrect disulfide bond formation. Together, these observations shed light on the biological functions of disulfide linkages for in vivo DNA packaging in sperm chromatin.

Introduction

The compaction of DNA is a common theme in cell biology. The most tightly-packed DNA is found in bacteriophages and vertebrate sperm nuclei with packing densities of 400–500 mg/mL. Spermiogenesis, the formation of mature spermatozoa, is a unique multistep process resulting ultimately in the self-assembly of DNA in sperm to a final volume roughly 1/20th that of a somatic nucleus (1, 2). To facilitate this extreme compaction, a dramatic reorganization occurs in developing spermatids where the vast majority (>90%) of somatic histones are replaced by small basic proteins called protamines (3). Protamines are short, arginine-rich, nuclear proteins that condense the spermatid genome into a transcriptionally inactive state (4, 5). Various hypotheses have been given to explain why sperm exhibit this unique specialized chromatin structure. The dense, nearly crystalline packaging of DNA in sperm nuclei is generally considered essential for both efficient genetic delivery as well as paternal genome protection against damage by mutagens and reactive oxidative species (1, 6). More recently, protamine has also been proposed to be involved in gene imprinting and epigenetic regulation of the spermatozoa (4, 7, 8). Mispackaging in the sperm head is linked to protamine dysfunctions resulting in increased DNA damage (7). Higher rates of de novo mutations in sperm are associated not only with increased infertility (9) but also higher rates of miscarriages (10, 11) and higher rates of genetic disease in the offspring (12, 13).

Unlike histone-compacted DNA, the physical properties of protamine-DNA assemblies closely resemble those of DNA condensed in vitro by smaller, simpler multivalent cations (14, 15, 16, 17, 18). Reconstituted protamine-DNA forms toroids that are on the same order of size as DNA condensed with cations of charge ≥ +3 (19, 20, 21, 22, 23, 24). In vivo, protamine also forms toroids where the sperm DNA is packaged in a hexagonal close-packed lattice (25, 26, 27). Protamines have two main conserved features throughout most vertebrate life (1). Firstly, there are several arginine-rich domains essential for binding to DNA and, second, are serine and threonine residues that are initially phosphorylated after translation and subsequently dephosphorylated upon completion of successful sperm chromatin remodeling. Protamines in eutherian mammals are further enriched with multiple cysteine residues. During epididymal maturation, mammalian protamines undergo a thiol oxidation to first form intra- followed by intermolecular disulfide bonds (28). The covalent sulfur-sulfur (S-S) bonds stabilize the sperm DNA and are thought crucial to condense the mammalian sperm nucleus into its fully mature state. Although much is known about how polycations package DNA in vitro, little is known about the specific role of the intra- and interprotamine S-S bonds on DNA packaging in vivo.

In this study, we used small angle x-ray scattering (SAXS) measurements to examine the structural role of disulfide bonds on DNA packaging in bull sperm nuclei. Many mammals, including humans, have two protamines (P1 and P2). Bull sperm cells only use P1 and thus represent a model system for understanding the role of disulfide bonds on mammalian DNA packaging in sperm heads. We show that DNA in salmon sperm nuclei is more tightly packaged than native bull sperm nuclei despite the lack of cysteine residues in the fish protamine. Osmotic stress force curves of nonreduced bull and salmon sperm nuclei are remarkably similar to each other, as well as force curves measured previously for reconstituted DNA condensed by a wide variety of multivalent cations. Time-resolved studies of dithiothreitol (DTT) reduction show that, as interprotamine disulfide bonds are reduced, tighter packaging of the DNA is observed. With further but incomplete reduction of the disulfide bonds, a broad scattering peak at large DNA-DNA spacings (>40 Å) is observed. Fully reduced bull sperm nuclei cause decondensation and lead to a complete loss of Bragg scattering. Finally, in comparisons of multiple bull semen collections, we found certain bull nuclei with aberrant scattering patterns consistent with improperly formed disulfide bonds.

Methods

Materials

Frozen bull sperm was purchased from Cattle Visions and Hawkeye Breeders for a total of 21 semen collections from 17 bulls. The semen was stored at −80°C until needed. Percoll was obtained from GE Healthsciences. The Qubit Protein Assay Kit was purchased from Life Technologies. Bioultra-grade polyethylene glycol (PEG), with an average molecular weight of 8000 and Nonidet P-40 substitute (NP-40) were obtained from Fluka Chemical Company. Iodoacetamide, DTT, 3-((3-cholamidopropyl)dimethylammonio)-1-propanesulfonate detergent, and Trichloroacetic acid (TCA) were purchased from Sigma. All commercially available reagents were used without further purification.

Sperm nuclei isolation

Bull sperm nuclei were isolated using a modified protocol from Gusse et al. (29) Frozen semen was thawed and spun down at 750 × g for 15 min at 5°C. Sperm cells were then suspended with a cold buffer of 1× buffer A (150 mM NaCl, 20 mM TrisCl (pH 7.5), 2 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride) containing 0.25 M sucrose and collected after centrifugation at 5000 × g for 5 min. The resulting pellet was then resuspended and washed twice in cold 1× buffer A and pelleted at 2000 × g for 5 min. A solution of 0.5× buffer A with 50% Percoll was then used to resuspend the sperm before being sonicated on ice. Sonication was performed with an Ultrasonic Dismembrator Model 150E (Fisher Scientific) with a one-eighth inch Microtip at 35% power (6 × 10 s) with 2 min rest between sonications. Sperm heads were recovered by centrifugation at 20°C for 20 min at 14,000 × g. After centrifugation, cell and nuclear membranes were solubilized in 100 mM TrisCl (pH 7.5), 10 mM 3-((3-cholamidopropyl)dimethylammonio)-1-propanesulfonate, and 20% glycerol. After incubating at 35°C for 30 min, sperm nuclei were pelleted by centrifugation at 5000 × g for 5 min. Nuclei were then sonicated again, as described above, and repelleted. Lastly, the isolated nuclei were washed twice and stored with buffer B (150 mM NaCl, 10 mM TrisCl (pH 7.5), 0.25 sodium azide, and 0.02% NP-40) at 4°C.

Sample preparation for x-ray measurements

Approximately one straw of sperm nuclei was used per sample for SAXS experiments. Isolated nuclei were pelleted and resuspended in 50 μL of 150 mM NaCl and 10mM TrisCl (pH 7.5) at the desired DTT concentration and PEG weight percentage. Samples with PEG were given several days to reach equilibrium. After reaching equilibrium, the nuclei and bath solution were transferred to a 0.7 mm glass capillary (Charles Supper). Nuclei in the capillary were then spun down in a Centrifuge 5804R (Eppendorf) with a swing bucket rotor at 150 × g for 15 min. After centrifugation, the capillary was sealed to maintain constant bathing solution conditions. The capillary was placed in a SAXS sample holder and fixed in position for up to several days for time-resolved studies.

Small angle x-ray scattering

SAXS experiments were carried out on a XTG UltraBright Microfocus Source (Oxford) with Ni-filtered CuKα radiation. The beam was collimated with polycapillary focusing optics and four adjustable slits. The flight path between the sample and detector was helium-filled to minimize background scattering. Samples were sealed in a capillary with a bath of an equilibrating solution and mounted in a temperature-controlled holder at 20°C. Diffraction patterns were recorded by direct exposure of BAS image plates and digitized with a Fujifilm Typhoon FLA 7000 scanner. Images were analyzed using the FIT2D (European Synchrotron Radiation Facility) software program. Silver behenate was used to determine the sample-to-detector distance. The diffraction patterns were radially integrated to give the scattering intensity I(q) versus the scattering vector q = (4π/λ)sinθ, with 2θ being the scattering angle. The electron density of DNA is significantly larger than the electron density of protamine or water and thus dominates the observed scattering profiles. The maximum reflection seen in the radial integration are related to the Bragg spacing, D_Br, of the DNA. The Bragg spacing is calculated as D_Br = 2π/q_max. For a hexagonal lattice, the Bragg spacing is directly related to the DNA interhelical spacing, D_int, by D_int = 2^∗D_Br /√3. Typical exposure times were 45–90 min. There was no significant sample degradation due to x-ray exposure for a single exposure or multiple daily exposures up to 2 weeks. For different samples equilibrated under the same PEG-buffer conditions, D_int values were reproducible to within ∼0.2 Å.

The use of PEG to apply osmotic pressure to DNA assemblies has been described previously (30). In brief, the condensed phase is equilibrated against a bathing polymer solution of known osmotic pressure. The polymer, typically PEG, is too large to enter into the condensed DNA phase, whereas water and salt ions are allowed to freely exchange. The excluded PEG exerts a direct osmotic pressure on the condensate. After equilibration, the osmotic pressures in both phases are the same and can be directly measured by vapor pressure osmometry. Using SAXS, DNA-DNA spacings within the DNA condensed phase can be determined from the Bragg scattering of x rays as a function of the applied PEG osmotic pressure. We have previously used this osmotic stress method to directly determine thermodynamic forces within cation condensed DNA phases, including reconstituted salmon protamine-DNA (15, 16, 31).

Protamine extraction

Purified sperm nuclei were pelleted by light centrifugation and resuspended in 500 μL of UNGD buffer (2 M urea, 1 M NaCl, 1 M Guanidinium Cl, and 50 mM DTT) to break S-S bonds inside the sperm chromatin. Nuclei were incubated overnight at room temperature or until all nuclei dissolved and then sonicated for 10 s at 35% power. After sonication, 80 μL of cold 6 M HCl was added to extract proteins and the samples placed on ice for 10 min. Samples were then spun at 10,000 × g for 10 min at 4°C to pellet the hydrolyzed DNA. The resulting supernatant was brought up to a total volume of 2 mL with UNGD buffer and 240 μL of 4 M sodium acetate added. The solution was transferred to an Amicon Ultra-2 with an Ultracel-3 centrifuge filter (Millipore) and concentrated at 1000 × g for 30 min at room temperature. The retentate was concentrated again after being mixed with an additional 2 mL of UNGD buffer. The retentate was subsequently concentrated four more times with 2 mL of 500 mM NaCl at 1000 × g for 30 min. The final retentate was diluted to 500 μL with 1 M NaCl.

Protamine was precipitated with TCA. Next, 150 μL of cold 100% (w/v) TCA was added to the isolated protamine and placed on ice for 20 min. The solution was centrifuged for 15 min at 15,000 × g at 5°C to pellet the protamine. The protamine was washed once with 25% (w/v) TCA and once with cold acidified acetone with a 15,000 × g spin for 15 min at 5°C after each wash. The recovered pellet was dried with a desiccator pump and dissolved in 50 μL of 100 mM TrisCl (pH 8) with 9 mM DTT and stored at −20°C until needed.

Protamine alkylation

Protamine solutions were thawed, incubated for 1 h at 50°C, and placed on ice. A solution of 900 mM iodoacetamide in deionized water was made fresh and protected from exposure to light. To alkylate the reduced cysteines, 5 μL of the iodoacetamide stock was added to the protamine solutions and incubated in the dark at room temperature for 2 h. The reaction mixture was then brought up to 500 μL total volume with 1 M NaCl and then TCA precipitated as described above. After drying, the alkylated protamine was diluted to 30 μL total in 10 mM TrisCl and 0.5 mM sodium azide buffer. Protamine concentration was determined by Qubit Protein Assay Kit (Life Technologies) on a Typhoon FLA 7000 (GE Healthcare Life Sciences) with a 96-well plate. An acetic acid-urea polyacrylamide gel stained with AcquaStain (Bulldog Bio) was used to confirm protamine concentration and successful alkylation.

Results

Comparison of in vivo packaging in salmon and bull nuclei

A major difference between piscine and bull sperm chromatin packaging is the presence of the disulfide bridge network in bulls. Bull protamine, like that of most eutherian mammals, has multiple cysteine residues and is known to form inter- and intramolecular S-S bonds (24, 32, 33). These cysteines are lacking in fish protamine. Fig. 1 gives the sequences for both salmon and bull protamine as well as the positions of the S-S bonds formed during maturation of the bull sperm in the epididymis. Basic arginine residues are shown in red to emphasize the DNA-binding domains, and intermolecular and intramolecular disulfide bonds are represented by dashed and solid lines, respectively. Prior work shows that in native bull sperm chromatin, one protamine molecule is bound per ∼11 bp of DNA (34). Of the seven cysteine residues on bull protamine, four participate in the formation of two intramolecular disulfide bonds that fold the amino- and carboxy-terminal ends toward the center of the molecule. The central arginine-rich DNA-binding domain of bull P1 has a sequence and size comparable to salmon protamine. Once formed, the S-S bonds prevent protamine extraction from DNA with the reduction of disulfide bridges to free sulfhydryls required for in vitro and in vivo decondensation of mammalian sperm nuclei (27, 35).

Figure 1.

Salmon and bull protamine sequences including the positions of cysteine residues responsible for inter- and intramolecular disulfide bond formation in bull protamine. Structures as drawn are directly based on the model of bull protamine-DNA proposed by Balhorn et al. (32), with inter- and intramolecular disulfide bonds represented as dashed and solid lines, respectively. Evidence suggests bull protamine dimers are formed binding to approximately two turns of the major groove of B-form DNA with DNA arranged in a close-packed hexagonal lattice. To see this figure in color, go online.

Fig. 2 shows typical SAXS profiles for isolated salmon and bull sperm nuclei. The maximum peak for both nuclei is the helix-helix Bragg reflection. The observed DNA packing densities in both salmon and bull sperm nuclei are comparable. The Bragg peak for bull nuclei occurs at q∼0.241 nm⁻¹, whereas for salmon the Bragg peak is observed at q∼0.249 nm⁻¹. In a hexagonal lattice, the equilibrium interaxial spacing (D_int) represents the actual distance between DNA helices and is inversely proportional to the experimentally-measured scattering vector q from SAXS. Calculated D_int values from the scattering profiles correspond to 30.1 Å for bull nuclei and 29.2 Å for salmon nuclei. Both isolated nuclei samples in Fig. 2 have the same buffered bathing solution of near physiological salt (150 mM NaCl) without added reducing agent or an osmotic agent such as PEG. SAXS experiments were performed at 20°C. D_int values were reproducible to within ∼0.1 Å.

Figure 2.

SAXS profiles of salmon and bull nuclei. Plotted is the radially averaged scattering intensity as a function of scattering vector Q. Samples were irradiated in a capillary with a bathing solution of 150 mM NaCl, 10 mM TrisCl (pH 7.5), 0.5 mM Azide, and 0.02% NP-40. In sperm chromatin, DNA toroids are known to result from the condensation of DNA by protamines (inset). Within the toroid, DNA helices are packed in a hexagonal lattice with a characteristic interaxial DNA spacing (D_int). To see this figure in color, go online.

Although the overall packaging densities are similar, there are some key differences in the scattering profiles to note. At similar sample volumes, bull sperm nuclei do not scatter nearly as strongly as salmon sperm nuclei. Here, the bull nuclei were exposed for twice as long as the salmon nuclei (90 vs. 45 mins), yet the DNA-DNA correlation peak is significantly weaker. DNA-DNA correlation in-plane is inversely related to the full width at half-maximum of the observed Bragg reflections. Consistently, the bull sperm nuclei Bragg reflections were broader indicating poorer in-plane ordering of the DNA when compared to salmon. Another difference is that only salmon protamine shows a higher order reflection located at q∼0.31 nm⁻¹. This peak is consistent with a 101 reflection for a hexagonal lattice and has been previously assigned as evidence of the binding of the polycation in DNA grooves (36). The absence of such a reflection in bull is also consistent with a poorer long-range order in these samples compared to salmon.

Packing and forces in sperm nuclei

To further study the differences between DNA packaging in bull and sperm nuclei, we used osmotic stress coupled with x-ray scattering allowing for the direct determination of the intermolecular pressure-distance relationship for protamine-DNA in the nuclei. Fig. 3 shows the osmotic stress force curve, the log of the excluded PEG polymer osmotic pressure versus the intermolecular spacing between DNA helices (D_int), measured by SAXS for both bull and salmon sperm nuclei. Arrows indicate the corresponding equilibrium spacings for bull and salmon sperm nuclei in the absence of applied osmotic pressure. These equilibrium spacings represent a balance of the attractions and repulsions in the protamine-DNA systems. As we increase the osmotic pressure, we compact DNA past its equilibrium spacing to progressively smaller interhelical distances whose magnitude is dominated by the short-range repulsions in the system. Maximum applied osmotic pressures achieved here are ∼100 atm.

Figure 3.

Osmotic pressure force curves for various DNA condensates. Bull sperm nuclei and salmon sperm nuclei measurements were performed in a capillary with a PEG bathing solution of 150 mM NaCl, 10 mM TrisCl (pH 7.5), 0.5 mM sodium azide, and 0.02% NP-40. Arrows indicate equilibrium D_int values without applied osmotic pressure. Reduced bull sperm nuclei were maintained in a bathing solution of 150 mM NaCl, 10 mM TrisCl (pH 7.5), 0.5 mM sodium azide, 0.02% NP-40, and 50 mM DTT. DNA-only samples were measured in a bathing solution of 150 mM NaCl, 10 mM TrisCl (pH 7.5), and 0.5 mM sodium azide. To see this figure in color, go online.

The osmotic stress force curve for bull sperm nuclei with intact disulfide bonds is remarkably similar to that for salmon sperm nuclei. Both systems are also consistent with force curves measured for DNA condensed in vitro by a wide variety of cations including salmon protamine (15, 16, 37, 38). These force curves are well-described by a simple double exponential equation with variable preexponential factors A and R, Π = Π_r + Π_a = Re^−2D_int^/λ + Ae^−D_int^/λ, with a fixed decay length, λ, of 5 Å as described elsewhere (15, 16, 31). Although similar, the bull nuclei maintain a larger DNA-DNA spacing than salmon nuclei for the entire force curve with the difference in D_int being more pronounced at the highest pressures. This likely arises from an increase in the repulsive component of the total interaction due to the presence of the disulfide cross-linked protamine network.

Complete reduction of the disulfide bonds in bull sperm chromatin leads to significant swelling of the nuclei and eventually a complete loss of any Bragg scattering such that no reflections are observed. The presence of the disulfide bonds is thus essential for DNA condensation in bull nuclei. Here, reduction was performed using 50 mM DTT in a 150 mM NaCl salt solution buffered with Tris. We found that nuclei can still be held together and interhelical spacings measured, using an osmotic stressing agent such as PEG to constrain swelling. To prevent the samples from dissolving under our conditions, reduced bull nuclei samples were originally reduced in a 10% PEG solution before being moved to their respective final PEG concentrations and equilibrated.

Surprisingly, the fully reduced bull samples exhibit a DNA force curve that is strictly repulsive at all spacings and nearly identical to DNA alone. Under low osmotic pressure, a liquid crystalline-like DNA phase is observed in reduced bull sperm with large (>34 Å) DNA-DNA spacings. At high pressures, tight packaging consistent with the hexagonal arrays of cation condensed DNA is again observed. Na⁺-DNA force curves have been previously reported and shown to be dependent on salt concentration at low osmotic pressure (39). However, at high osmotic pressures, the forces converge to a salt-insensitive interaction. Plotted in Fig. 3 for comparison is the force curve data for Na⁺-DNA measured in the same 150 mM NaCl bathing solution as the reduced and nonreduced nuclei.

Effect of DTT on sperm chromatin packaging by time-resolved x-ray studies

To better understand the effects of inter- and intramolecular disulfide bonds on the chromatin structure, we conducted time-resolved studies in the presence of the reducing agent DTT. Fig. 4 shows x-ray scattering intensity profiles from bull nuclei reduced with a bathing solution of 1 mM DTT, 150 mM NaCl, 10 mM TrisCl (pH 7.5), and 10% PEG. Under these conditions, reduction of disulfide bonds in the nuclei in the capillary occurs on the timescale of days. One-dimensional SAXS data was obtained by azimuthally averaging the two-dimensional scattering data. Without DTT, the equilibrium Bragg reflection is observed at q ∼0.24 Å⁻¹ corresponding to a Bragg spacing D_Br of 26.2 Å. As the DNA in the bull nuclei is known to form a hexagonal array, this can easily be calculated as corresponding to an interaxial spacing D_int of 30.1 Å. DTT reduction of the nuclei shows an increase (day 3) followed by a gradual lowering in intensity of the Bragg reflection and a rise in a poorly ordered DNA phase characterized by a broad peak observed at q ∼0.165 Å⁻¹ by day 10. This broad DNA peak corresponds to a more loosely ordered structure with a Bragg spacing D_Br of ∼39.3 Å. This liquid-crystalline-like DNA phase results from the disruption of the disulfide bonds, and under 10 wt% PEG bath conditions is maintained in equilibrium with the original hexagonal DNA phase. Nuclei without DTT were tested and observed to show no signs of degradation from exposure to x rays even after daily irradiation for 14 days. DTT added to salmon sperm nuclei also had no effect on DNA-DNA spacings or scattering intensity. This strongly suggests the observed DNA-DNA spacing changes in the bull are due to disulfide bond breakage and that DTT is not significantly damaging the sperm chromatin DNA.

Figure 4.

Multiple day x-ray series for bull nuclei in a reducing environment. Bull nuclei were maintained in a bathing solution of 150 mM NaCl, 10 mM Tris (pH 7.5), 1 mM DTT, and 10% w/w PEG inside a capillary. Bull sperm nuclei samples were left in the sample holder for the duration of the experiment and stored at room temperature. To see this figure in color, go online.

To more closely examine the early time course of this rearrangement, we slowed down the disulfide bond reduction further by lowering the DTT concentration. Fig. 5 shows time-resolved x-ray series for bull nuclei being reduced in a bathing solution of 0.4 mM DTT, 150 mM NaCl, and 10 mM TrisCl (pH 7.5). Under these conditions, protamines were not fully reduced even after multiple days and PEG was not required. It is known that the interprotamine disulfides are reduced first as they are more accessible to solvent and reducing agents and presumably more strained than intramolecular disulfides. After 1 day, a clear increase in the amplitude of the Bragg reflection is observed as intermolecular disulfide bonds are first reduced. This data is consistent with an improved long-range ordering of the DNA, although maintaining a nearly constant DNA-DNA spacing compared to nonreduced bull nuclei. Interestingly, by day 3, we see not only a further increase in the Bragg reflection intensity but also a shift to higher q, or equivalently smaller DNA-DNA spacings. The equilibrium Bragg reflection is now observed at q ∼0.25 Å⁻¹ or a D_int of ∼29 Å, comparable to salmon nuclei. In addition to tighter DNA packaging, the peak width narrows and there is a slight appearance of a 101 peak observed at q ∼0.31 Å⁻¹. As shown in Fig. 2, this 101 reflection is commonly observed in piscine sperm but normally missing in bull.

Figure 5.

Multiple day x-ray series for bull nuclei in a dilute reducing environment. Nuclei were loaded in a capillary with a bathing solution of 150 mM NaCl, 10 mM Tris (pH 7.5), 0.5 mM Azide, and 0.4 mM DTT. The capillary was left in the sample holder for the duration of the experiment and stored at room temperature. To see this figure in color, go online.

Comparison of normal and aberrant bull nuclei

In comparison of the 21 bull semen collections that were purchased, we observed that a few semen collections showed aberrant x-ray scattering profiles, as shown in Fig. 6. Although there was some minor variation in the central scattering of all bull samples, only two collections had the pronounced aberrant scattering shown characterized by an increased scattering at low q (q < 0.20 Å⁻¹). This aberrant scattering data is consistent with a two-phase system where fully condensed DNA is in coexistence with a loosely ordered phase. The aberrant scattering profiles are similar to that of the isolated bull nuclei shown in Fig. 4 after exposure to a reducing environment, suggestive of incomplete disulfide bond formation.

Figure 6.

Comparison of normal and aberrant scattering from isolated nuclei from multiple bulls. The nuclei were in a bathing solution of 150 mM NaCl, 10 mM Tris (pH 7.5), 0.5mM Azide, and 0.02% NP-40. To see this figure in color, go online.

Discussion

In the late stages of spermiogenesis, chromatin remodeling in vertebrates occurs ultimately through a replacement of histones by protamines. Tight packaging of the chromatin is thought crucial to both minimize the volume of the sperm nuclei for transport and protect the genetic material from potentially damaging agents at the same time as the genetic repair mechanisms are inactive. There are several differences between mammalian and piscine protamines. Fish typically only have one protamine (P1). Salmon P1 is a 32-amino acid peptide consisting of 21 arginine and 4 serine residues that can be used as phosphorylation sites. Many mammals have two protamines, P1 and P2, both of which are longer than piscine P1. Mammalian P1 is typically 50−60 amino acids with an average arginine fraction (50−55%) significantly lower than for fish (65−70%) (1). Eutherian mammal protamines, such as humans and bulls, also contain multiple cysteine residues that must form specific disulfide bonds to ensure proper maturation of the sperm. Despite the important role of the disulfide bond formation in the function and stability of mammalian spermatids, little is known about how these S-S bonds regulate the DNA packaging in vivo.

Bulls only have P1 protamine and represent an intermediate in complexity between fish and human sperm. Bull P1 is therefore an ideal model system for understanding the role of the S-S bond network on DNA packing in vivo. The sequences of salmon and bull protamine, as well as the disulfide bond locations of bull P1, are given in Fig. 1. Bull P1 has a central arginine-rich DNA-binding domain similar in sequence and size to salmon P1. This domain is flanked by cysteine-rich domains at both ends that fold toward the center of the molecule and are held together by disulfide bonds (33). We show that despite the presence of the disulfide bond network, the packing density of DNA in bull is slightly less dense than salmon sperm chromatin (Fig. 2). DNA-DNA spacings in mature bull sperm nuclei are ∼1 Å larger than salmon sperm nuclei. The low intensity and broadness of the SAXS profiles, as well as the lack of higher order reflections, indicate poorer crystalline ordering of the bull sperm DNA compared to salmon. Although the DNA in the bull nuclei is less dense, it cannot be determined if the nuclei have more void volume than piscine nuclei due to the differences in protamine. The determination of the void volume would be helpful in comparing the likelihood of DNA damage by small reactive oxidative species or mutagens between mammalian and piscine nuclei.

The role of the disulfide bonds on the intermolecular forces was examined by coupling osmotic stress with x-ray scattering to directly determine the protamine-mediated attractive and repulsive forces in sperm nuclei. Previously, we and others have shown that the same double exponential fit can be used to describe force curves for reconstituted DNA condensed by a wide variety of cations, including protamine, as well as for salmon nuclei (15, 16, 31, 37, 38). Here, we show that mature bull protamine sperm nuclei also show similar force curves as the salmon sperm nuclei (Fig. 3). The common force characteristic for these very different systems suggests that fundamental physical principles underlie in vitro DNA condensation as well as in vivo protamine-DNA compaction. However, bull nuclei force curves did show an increased difference in DNA-DNA spacings resulting in a slightly steeper slope at the highest applied pressures. A possible explanation is that the rigid disulfide network contributes additionally to the short-range repulsive forces inside the sperm head.

Reduction of disulfide bonds has interesting consequences for DNA order. Complete reduction of the disulfide bonds with DTT leads to dramatic swelling of the nuclei (at least 10-fold in volume) and complete loss of Bragg scattering. The disulfide bonds are thus absolutely essential for dense packing of DNA. The osmotic stress force curve for fully reduced bull nuclei is nearly identical to DNA-only samples where DNA helices are strictly repulsive at all spacings. Concerned that the reduced protamine had diffused out of the protamine-DNA assembly, unsuccessful efforts to detect protamine in the supernatant were carried out (data not shown). Due to the high binding affinity of P1 to DNA (40, 41, 42) and the presence of the osmotic stressing agent, this possibility also seems unlikely.

Based on disulfide bond assignments, Balhorn et al. (32) concluded that bull protamine forms a hairpin structure with two intradisulfide bridges and three interdisulfide bridges, which connect to adjacent protamine to form the protamine-DNA assembly. In the hairpin structure, the DNA-binding section of the bull protamine maintains an arginine fraction comparable to salmon protamine at ∼70% arginine. Based on previous results with arginine oligopeptides, we showed that the increase in neutral amino acid content increases repulsions resulting in lower packaging densities in condensed DNA (16). The loss of the hairpin structure due to extensive reduction would allow the protamine to relax into a linear shape, which would presumably associate with the DNA grooves as proposed in the electrostatic zipper model (43, 44, 45). We hypothesize that as the inter- and intradisulfide bonds are reduced, the full bull protamine (∼50% arginine) is presented to the DNA. The reduced protamine would likely interact with DNA, but be insufficient to condense the DNA. This would explain why no protamine was found in the supernatant, yet the reduced protamine did not contribute to condensing/attractive forces. Another possible rationale is that the N- and C-terminal ends of bull protamine, once reduced, frustrate the adoption of the preferred highly ordered, high-density state.

Comparisons of protamine sequences shows that the location of the cysteines are highly conserved among many species. Although the hairpin structure is well-established in bull, we hypothesize that this structure is likely present in many species and that the formation of the intraprotamine disulfide bonds is critical to enable proper DNA condensation. In such a model, the secondary structure of the hairpin resulting from the intraprotamine disulfide bridges is thus required to overcome the increased repulsions resulting from the lower fraction of arginines in mammalian protamines or chain-end effects and create the tightest packaging possible. This is consistent with previous work that showed native-like packaging of reconstituted DNA resulted from condensation with folded bull P1 with the hairpin structure in place (24). We are currently extending our efforts into other mammalian species to more fully understand the role of intramolecular disulfide bonds in eutherian mammal chromatin.

The failure of fully reduced bull protamine to maintain DNA condensation led to an interest in better understanding the role of the inter- and intraprotamine disulfide bonds through time course studies. In vivo, the intramolecular disulfides form shortly after the protamines are deposited onto DNA, followed by very specific interprotamine disulfides that are thought critical for the correct formation of the highly compacted chromatin. Although crystal structures do not exist for bull protamine-DNA, it is hypothesized that the central DNA-binding domain of bull protamine is in the major groove, with the folded terminal ends protruding slightly out of the major groove facilitating the formation of intermolecular disulfide bonds (24). We show that the disulfide network appears to hold helices in a strained, nonoptimal conformation, likely due to a torque resulting from the formation of the interprotamine S-S bonds. Reduction of the intermolecular disulfide bonds relieves this torque force resulting in a relaxation of DNA to a more ordered state (Fig. 4). We hypothesize that this more tightly packaged bull sperm DNA is a result of condensation by bull P1 that maintains the hairpin structure but has cleaved the interdisulfide bridges between neighboring protamine molecules. This allows the bull sperm nuclei to achieve a packing density nearly identical to salmon sperm nuclei. With further disulfide reduction, a more loosely organized liquid-crystalline DNA phase results. Here, some but not all of the intraprotamine disulfide bonds are likely broken. Typically, this more loosely packaged phase is in equilibrium with the tightly packaged protamine-DNA phase. Ultimately, fully reduced bull protamine cannot maintain DNA in the condensed phase.

The presence of the C- and N-terminal peptide folds held by the intraprotamine S-S bonds was previously proposed as a method for refinement of sperm chromatin packaging afforded only to mammals and higher vertebrates (46). We hypothesize that the initial phosphorylation of the protamine after production, which should lower DNA affinity, combined with the inability of the linear mammalian protamine to condense DNA, may allow for controlled transport of the protamine to the nuclei. The weak affinity for the DNA would also allow saturation and optimized protamine placement on the DNA in the nuclei before toroid formation. Maturation of the bull sperm would thus require both dephosphorylation and correct S-S bond formation after protamine deposition on DNA. There is evidence that phosphorylation/dephosphorylation is involved in correct disulfide formation (8, 47, 48). In this way, developing bull spermatozoa could use the protamine chemistry and its secondary hairpin structure to effectively regulate protamine-DNA assembly and ultimately DNA packing density.

Lastly, several commercial bull specimens were screened via SAXS to determine if their frozen semen collections contained any abnormal scattering patterns. Two collections showed aberrant central scattering, which was reminiscent of the bull nuclei samples with partially reduced disulfide bridges. It is widely known in the animal reproduction community that a myriad of problems (timing, animal health, etc.) can have a large impact on semen collection quality even from fertile animals. Further studies into differences between the aberrant and normal scattering were carried out. The aberrant scattering bull nuclei did not appear to be morphological abnormal from the normal scattering collections by light microscopy. Attempts to find protamine/DNA ratio deficiency in the aberrant scattering bulls was not outside of experimental error. One possible explanation for the observed aberrant scattering would be if the bull protamines were dephosphorylated too soon to allow for correct disulfide linkage formation. Correlations between possible residual protamine phosphorylation and the observed aberrant or normal scattering nuclei and its possible correlation to fertility state is currently under investigation.

Conclusions

By combining osmotic stress with x-ray scattering, we have examined the time course and structure of sperm chromatin nuclei from salmon and bull to elucidate how inter- and intraprotamine disulfide bonds impact DNA condensation in mammalian sperm. The role of protamine in spermatozoa is likely critical for paternal genome protection and epigenetic regulation, as DNA repair mechanisms are turned off within sperm cells. We show that interprotamine disulfide bonds result in a torque force that reduces the packaging efficiency in bull sperm chromatin. The secondary structure of bull P1 resulting from intraprotamine S-S bonds appears necessary to facilitate DNA condensation. Our results provide important insight on specific ways that inter- and intraprotamine disulfide bonds are used to manipulate DNA packaging in the sperm. Bull sperm protamines are less complex than human protamines, but the insights gained by SAXS will be critical to interpreting future human experiments. Currently, experiments are underway to extend this work to mammals with P1 and P2, and further probe the interconnection between protamine mispackaging in sperm nuclei and DNA damage.

Author Contributions

D.C.R. and J.E.D. designed the research. J.M.H., D.C.R., and J.E.D. performed the experiments and analyzed the data. J.M.H. and J.E.D. wrote the article.

Editor: Jason Kahn.