PolysaccharideâPolynucleotide Complexes. 2. Complementary

Biomacromolecules 2001, 2, 641-650

641

Polysaccharide-Polynucleotide Complexes. 2. Complementary Polynucleotide Mimic Behavior of the Natural Polysaccharide Schizophyllan in the Macromolecular Complex with Single-Stranded RNA and DNA Kazuo Sakurai,*,† Masami Mizu,‡ and Seiji Shinkai*,‡ “Organization and Function”, PRESTO and Chemotransfiguration Project, Japan Science and Technology Corporation (JST), Kurume Research Center Building, 2432 Aikawa, Kurume, Fukuoka 839-0861, Japan Received November 8, 2000; Revised Manuscript Received March 7, 2001

Schizophyllan is an extracellular polysaccharide consisting of a β-1,3-D-glucan main chain and exists as a triple helix in water and as a single chain in dimethyl sulfoxide (DMSO). When the single chain of schizophyllan (s-SPG) was mixed with poly(C), poly(A), poly(dA), or poly(dT), they form a macromolecular complex. On the other hand, poly(G), poly(U), poly(I), poly(dG), and poly(dC) do not. This nucleotide specificity evidences that the hydrogen bonds are essential to form the complex, because the former nucleotides have an unoccupied hydrogen-bonding site and the latter ones use the hydrogen-bonding sites in the intramolecular aggregation (i.e., such as the G quartet for poly(G) and poly(dG) and the U hairpin for poly(U)). The hypochromic effect and the increment in the circular dichroism (CD) intensity are observed in accordance with the complex formation. These facts indicate that the base stacking is enhanced in the complex. The solvent-composition (DMSO/water) dependence demonstrates that the hydrophobic interaction is important to form the complex as well as the hydrogen-bonding interaction. With increasing temperature the complex dissociates cooperatively and the melting curve enables the thermodynamic parameters to be evaluated (∆H ) -60 to 70 kcal mol-1 and ∆S ) -150 to 200 cal mol-1 K-1). These values are comparable with those for double helix DNA. Namely, the complex can be characterized by enhancement of the base stacking, cooperative dissociation, the similar thermodynamic parameters to DNA, and combination of the hydrogen-bonding and hydrophobic interactions to form the higher-order structure. These facts surprisingly coincide with characters of the double helix of DNA. In other words, the s-SPG molecule behaves as if it were a complementary polynucleotide chain for the corresponding polynucleotide. Furthermore, stoichiometric study suggested that the complex structure is a triple helix consisting of two s-SPG and one poly(C) or poly(A) chains. Introduction Saccharides act as an identification marker in the cellrecognition process and enable the biological system to curry out a particular and specific reaction.1 Hydrogen bonds between the saccharides and other biomolecules are essential in the recognition;2 however, the one formed in aqueous solutions such as the biological system is hardly considered to provide specific affinity to a particular reaction. Nowadays, it is widely accepted that the specificity is achieved by the hybrid or synergistic effect, which occurs only when the hydrogen-bonding interaction is combined with the spatial selectivity,1,3 the hydrophobic interaction,1 or the electrostatic interaction4 (hybrid hydrogen bond). Furthermore, the specific affinity of the hybrid hydrogen bond can be intensified in the cooperative phenomenon5 in the macromolecular system, which is sometimes called “the cluster effect” or “the allosteric effect”. * To whom correspondence should be addressed. E-mail address: [email protected]. † PRESTO, Japan Science and Technology Corporation (JST). ‡ Chemotransfiguration Project, Japan Science and Technology Corporation (JST). URL: http://www.jst.ktarn.or.jp/.

The combination of the hybrid hydrogen bonds and the cooperative phenomenon is also characteristic of polynucleotides, where the complementary bases form a specific pair through the stereoselective hydrogen bonds; once the pairs form, the hydrophobic interaction induces the base-pair stacking, then the stacking occurs cooperatively to construct the double-helix of DNA.6 Since the similar mechanism (i.e., the hybrid hydrogen bonds) governs both the saccharide recognition and the DNA double helix formation, one may consider that some saccharide can specifically interact with polynucleotides. Lee et al.7 showed that a family of antibiotics, calicheamicins, utilizes the interaction between the oligosaccharide and polynucleotides to recognize the C-G pair in DNA. However, the detailed mechanism of the sequence recognition is still under investigation. As far as we know, the oligosaccharide in calicheamicins7 and its related materials8 are the only case to show a novel interaction between polynucleotides and saccharides. Understanding the specific interaction between polynucleotides and other materials and developing such functional compounds are of great importance in the gene technology. A copolymer consisting of imidazole and pyrrole moieties

10.1021/bm000121r CCC: $20.00 © 2001 American Chemical Society Published on Web 06/12/2001

642

Biomacromolecules, Vol. 2, No. 3, 2001

Figure 1. Repeating unit of schizophyllan (a) and its triple helical structure (b). The circles in (b) represent the glucose residues. The helix pitch and diameter are 1.8 and 2.6 nm as indicated, respectively.

can form a 1:1 complex with the double helix DNA.9 Their work clearly demonstrates the importance of both the acidity of the hydrogen bonding site and the distance between the binding sites (stereoselectivity) in order to form the stoichiometric complex. Nielsen et al.10 synthesized a peptide nucleic acid (PNA) which consists of the (2-aminoalkylene) glycine backbone and a nucleic base attached to the amino group and showed that PNA can also form a 1:1 or 1:2 complex with the complementary polynucleotide single chain. Since PNA has no anionic charge to generate the repulsive force between the anion of the pair polynucleotide chain, the complex is more stable than the corresponding polynucleotide/polynucleotide complex.10 This advantage can be utilized in designing an antisense DNA drug11 and controlling the telomerase activity.12 These pioneering studies have broadened our horizons in the gene science and technology; however, polysaccharides have never been reported to show a specific interaction with polynucleotides. Schizophyllan is a cell wall polysaccharide produced by the fungus Schizophyllan commune and the main chain consists of β-1,3-D-glucans and the one β-(1f6)-D-glucosyl side chain links to the main chain at every three glucose residues,13 as shown in Figure 1. Yanaki et al.14 and Young and Jacobs15 showed that schizophyllan exhibits antitumor activity against Sarcoma 180 as well as lectin activity against Limulus amebocyte lysate. The mechanism of such biological activity is not well-understood in the molecular level and still lies as a challenging issue in the field of saccharide biochemistry. Norisuye et al.16 extensively studied the dilute solution properties of schizophyllan and clarified as follows: (1) schizophyllan adopts a triple helix conformation17

Sakurai et al.

(see part b in Figure 1) in water and a random coil in dimethyl sulfoxide (DMSO);16a,b (2) the schizophyllan triple helix (t-SPG) is stabilized by the interchain hydrogen bonds between the OH groups attached to the C-2 carbons of the main chain glucoses;16a (3) when DMSO is added to the schizophyllan aqueous solution (denaturation), the helix remains intact up to the water fraction in the DMSO-water mixture (Vw) ) 0.14;16b (4) however, when water is added to the DMSO solution (renaturation), the single chain of schizophyllan (s-SPG) collapses at Vw ) 0.14 owing to the hydrophobic interaction; (5) Vw > 0.2 s-SPG aggregates due to both the intra- and intermolecular hydrogen bonds.16d We supposed that when schizophyllan was mixed with other bioactive materials, the hydrophobic nature of β-1,3D-glucose might create the hybrid hydrogen bond effect and the spatially arranged glucose residues in schizophyllan might exert the cooperative phenomenon. This was our primary motive to launch this work. In fact, we have found that s-SPG can interact with poly(C) or poly(A) and form a novel macromolecular complex.18 In this subsequent reports “Polysaccharide-Polynucleotide Complexes. 2”, we explore the characteristic nature of this newly found interaction, first examining the nucleotide specificity, then focusing on the poly(C) and poly(A) systems. On the basis of the results, we have made an attempt to propose a molecular model for the complex. Experimental Section Materials and Solution Preparation. Table 1 summarizes the materials used in this work, their codes, and some molecular characters. The triple helix of schizophyllan (tSPG) was kindly supplied from Taito Co. in Japan. The intrinsic viscosities ([η]) of t-SPG and s-SPG in water and DMSO at 25 °C were evaluated to be 6.1 and 0.92 dL g-1, respectively, and from those [η] values, the molecular weights were calculated to be 4.5 × 105 and 1.5 × 105 (690 and 230 in the degree of polymerization), respectively. All homo-polynucleotides were purchased from Amersham Phamacia and the degrees of polymerization, calculated from the reported sedimentation velocities,19 are listed in Table 1. Dextran, pullulan, amylose, poly(diallydimethylammonium chloride) (PDDA), guanosine 5′-monophosphate (GMP), urea, and poly(ethylene glycol) were purchased from Wako Chemical and used without further purification. A RNase free, deionized, and distilled water and DMSO with a spectroscopic grade were purchased from Wako Chemical and used for all measurements. A polynucleotide sample was mixed with s-SPG by adding a s-SPG/DMSO solution to a polynucleotide/water solution. Hereinafter, we denote the mixture as poly(X) + s-SPG, where X is the base symbol. The concentrations of the s-SPG/ DMSO and the polynucleotide/water solutions were chosen so as to give the desired s-SPG (Ms-SPG) and poly(X) (Mpoly(X)) molar concentrations after mixing. The apparent pH (pH*) of the DMSO/water solutions was measured by a Horiba D-22 pH meter at room temperature and found to be pH* ) 8.0-9.0 for all the solutions with Vw ) 0.97-0.7. All the samples thus prepared were kept at 5 °C for at least 3 days before the measurement.

Biomacromolecules, Vol. 2, No. 3, 2001 643

Polysaccharide-Polynucleotide Complexes Table 1. Materials Used in This Study and Their Codes and Some Characters name single chain of schizophyllan triple helix of schizophyllan dextran pullulan amylose single-stranded homo poly(adenosine monophosphate) single-stranded homo poly(cytidine monophosphate) single-stranded homo poly(guanosine monophosphate) single-stranded homo poly(uridine monophosphate) single-stranded homo poly(inosine monophosphate) single-stranded homo poly(deoxyadenosine monophosphate) single-stranded homo poly(deoxycytidine monophosphate) single-stranded homo poly(deoxyguanosine monophosphate) single-stranded homo poly(deoxythymidine monophosphate) poly(diallydimethylammonium chloride) poly(ethylene glycol) urea guanosine 5′-monophosphate a

Mw

code s-SPG t-SPG dexrtan pullulan amylose poly(A) poly(C) poly(G) poly(U) poly(I) poly(dA) poly(dC) poly(dG) poly(dT) PDDA PEG6000

1.5 × 4.5 × 105 4.0 × 104 105

1.6 × 105

notes ) 230 DP ) 690 1,2 R-1f 6-(R-D-1,4 glucose)3 R-D-1,4 glucose NBb ) 320 NB ) 570 NB ) 300 NB ) 400 NB ) 401 NB ) 300 NB ) 250 NB ) 300 NB ) 260 DPa

ca. 4.0 × 105 6000

GMP

b

Degree of polymerization. Base length estimated from the reported sedimentation velocity.

Circular Dichroism, UV Spectroscopy, and SmallAngle X-ray Scattering. Circular dichroism (CD) in the 200-320 nm region was measured on a Jasco J-720WI spectropolarimeter with a 1.0 or 0.1 cm cell. Under each condition described below, four types of measurements were carried out: (1) one to examine which polynucleotide induces the spectral change when mixed with s-SPG, (2) the temperature (T) dependence, (3) the DMSO/water composition dependence, and (4) one to determine the stoichiometric number. (1) The CD spectrum was obtained for poly(X) + s-SPG at 5 oC with Ms-SPG/ Mpoly(X) ) 4.5. Mpoly(X) was chosen so as to provide the lowest S/N spectrum for each sample. (2) The T dependence for the poly(C) system was examined within T ) 0-70 °C under the condition that Mpoly(C) ) 2.5 × 10-4 M and Ms-SPG/Mpoly(C) ) 4.5 at the three compositions: Vw ) 0.93, 0.85, 0.75. (3) The Vw dependence was measured at 10 °C for the poly(C) system and at 5 °C for the poly(A) system with Mpoly(X) ) 2.0 × 10-4 M and Ms-SPG/ Mpoly(X) ) 4.5 fixed. (4) To determine the stoichiometric number, the molar ratio [fms ) Ms-SPG/(Mpoly(X) + Ms-SPG )] dependence was measured at 10 °C for poly(C) + s-SPG under two different conditions: Mpoly(C) was kept constant (2.5 × 10-4 M) while fms was varied from 0 to 0.16, and changing fms from 0 to 0.6 with fixing Ms-SPG + Mpoly(C) ) 2.0 × 10-4 M. The same type of stoichiometric measurement was carried out for poly(A) + s-SPG at 5 °C by changing fms from 0 to 0.6, fixing Ms-SPG + Mpoly(A) ) 4.0 × 10-4 M. The molecular ellipticity ([θ]) was evaluated in a conventional method and the [θ] value at the peak top positions (λmax) was denoted as [θ]max. For comparison, after an excess amount of the materials listed in Table 1 (except for s-SPG and polynucleotides) was added to the poly(C) or poly(A) solution, the change in the CD spectrum was examined. The additive ratios were 4-5 times in mole for the polymeric materials and about 10 times for GMP and urea. Since GMP itself was found to have a CD band, the differential spectrum was used in discussion. A Jasco V-570 UV/vis/NIR spectrometer was used to measure the ultraviolet absorbance (UV) and the molar

Figure 2. Unusual change in the CD spectra occurred only when s-SPG is added to poly(C), with comparing addition of dextran, PEG6000, pullulan, amylose, PDDA, and t-SPG. The measurements were carried out at 10 oC with Mpoly(C) ) 2.5 × 10-4 M.

extinction coefficient () was obtained. Small-angle X-ray scattering (SAXS) from poly(C) and poly(C) + s-SPG solutions with Cpoly(C) ) 1.52 × 10-3 and Cs-SPG ) 7.3 × 10-3 g dL-1 were measured at room temperature with the BL45XU synchrotron X-ray biophysics beam line22 at Spring 8 in Japan. The instrumental detail was descried elsewhere21 and from the scattering data the cross-sectional diameter of the molecules was estimated by the Guinier plot.21 NuSieve agarose gel (FMC Bioproducts) prepared at 3.5 wt % concentration were electrophoresed for 3 h at 3 V cm-1 at 10 °C, and the migrants were stained with Gel Star nucleic acid stainer (FMC Bioproducts). Results and Discussion Changes in the CD and UV Spectra of Poly(C) and Poly(A) by Adding Schizophyllan. Figure 2 compares the CD spectra between poly(C) and its seven mixtures at Vw ) 0.93. The poly(C) spectrum remains intact when excess t-SPG, dextran, pullulan, or PEG6000 is added; on the other hand, PDDA decreases it by almost half. The striking feature

644


Sakurai et al.

Table 2. Results of the UV and CD Measurements and Comparison with the Reported Valuesa sample code poly(C) in water poly(C) in Vw ) 0.93 poly(C) + s-SPG in Vw ) 0.93 poly(A) in Vw ) 0.93 poly(A) + s-SPG in Vw ) 0.93 poly(C) in water

λmax in UV, nm 270 270 269 257 255 268

× 10-3, L mol cm-1 6.51 6.41 5.68 7.98 7.22 6.32

(Cp)6C (CMP 7 mer)

268.5

7.58

(Cp)3C (CMP tetramer)

268

7.96

CMP

272

9.15

poly(C) + poly(G) poly(A) in water

258

9.2

poly(A) + poly(U)

256

6.2

λmax in CD, nm 276 276 274 263 261

[θ] × 10-3, deg cm2 dmol-1

ref no.

68.2 65.0 90.9 50.9 117.7

277 (pH ) 7.5) 290 (pH ) 4.0)

75.0 ca. 56

290 (pH ) 4.0)

ca. 38.6

290 (pH ) 4.0)

ca. 14.2

290 (pH ) 4.0) 265

ca. 7.0 ca. 30

267

ca. 72

264

ca. 40

24a 24a 24b 24a 24b 24a 24b 24a 24b 24d 24c 24c 24c 24c

a The measurements were carried out at 10 °C for poly(C) and 5 °C for poly(A). : molar absorption coefficient. [θ]: molar ellipticity. V ) water/(water w + DMSO) in volume.

is that poly(C) + s-SPG shows a positive band at 274 nm increased by 70% and a new peak arising around 245 nm. This result confirms our previous data.18 Although the data are not shown, urea and GMP slightly increase the intensity around 274 nm. The values of λmax of CD, and [θ ]max for poly(C) and poly(C) + s-SPG are summarized in Table 2. The CD band of poly(C) is ascribed to the exciton coupling between the cytosines involved in the helical (or stacking) conformation.22 Therefore, the CD intensity is in the liner relation to the molar ratio of the base involved in stacking (Rs).23 The decrease by PDDA is explained by decrease in Rs, which is caused by destruction of the helical conformation due to formation of an ionic pair.24 On the other hand, the increment of the positive band in poly(C) + s-SPG can be ascribed to increment in Rs. This means that more cytosines are stacking in poly(C) + s-SPG than in poly(C). Figure 3 presents the UV spectra of poly(C) and poly(C) + s-SPG as well as their CD spectra at Vw ) 0.93. The values of λmax of UV and max were read off and summarized in Table 2, along with the literature ones. The addition of s-SPG to poly(C) decreases the absorbance of cytosine by 12% (hypochromism) and shifts λmax slightly to shorter wavelength (blue shift) in accordance with the change in the CD spectrum. According to the previous work25,26 which is presented as the literature values in Table 2, the hypochromism, the blue shift, and the increment of [θ]max take place when Rc increases (that happens with increasing the degree of polymerization of poly(C),25,26 as demonstrated in Table 2, from 1 to 4, 7, and then >200, i.e., poly(C)). In analogy with those, we can confirm that the addition of s-SPG induces the stacking between the cytosine bases. To enhance the base stacking, polar atmosphere or/and hydrogen-bonding interactions are necessary.6 Since the addition of s-SPG to an aqueous solution should not increase the polarity of the solvent, hydrogen-bonding interactions are the only possible explanation for the base stacking. If

Figure 3. Comparison of the UV and CD spectra between poly(C) and poly(C) + s-SPG measured at 10 oC for Cpoly(C) ) 7.16 × 10-3 g dL-1 (2.22 × 10-4 M) with a 1 cm cell for both the CD and UV spectroscopies.

we accept the formation of the hydrogen bonds, it means that poly(C) and s-SPG form a macromolecular complex. As mentioned in the Introduction, the OH groups of the C-2 carbon in schizophyllan forms the hydrogen bonds in water. Hence, the same OH group might be a candidate to interact with the proton acceptors in the cytosine. In Figure 2, the increment in Rc can rationalize the intensified 274 nm band, however, it cannot explain the origin of the newly arising band at 245 nm. Furthermore, when we estimated the value of [θ]max at Rc ) 1 of poly(C) by using van’t Hoff equation,27 the result ranged over (7.08.0) × 104 deg cm3 dmol-1, which is smaller than the observed value of poly(C) + s-SPG (9.1 × 104 deg cm3 dmol-1). These two facts indicate that the conformation of poly(C) in the mixture is considerably different from that of poly(C) itself. According to the exciton-coupling band


Polysaccharide-Polynucleotide Complexes

Table 3. Nucleotide Specificity in the Complex Formation in Neutral and Nonsalt Aqueous Solution

RNA

DNA

poly(C) poly(A) poly(U) poly(G) poly(I) poly(dC) poly(dA) poly(dT) poly(dG)

complex formation

conformation

yes yes no no no no yes yes no

single chain single chain intramolecular H-bond (hairpin like) 4G wire (intramolecular H-bond) intramolecular H-bond intramolecular H-bond single chain single chain 4G wire (intramolecular H-bond)

Figure 4. Comparison of the CD spectra for poly(A) + s-SPG at Vw ) 0.923, poly(A) + t-SPG in water, and poly(A) in six different Vw solutions with Mpoly(A) ) 2.2 x 10-4 M.

mechanism,28 the new band can be associated with the change in the tilting angle between the cytosines. Furthermore, the larger [θ]max than expected from Rc ) 1 can be related to the shorter distance between the bases. On the other hand, comparison between the UV and CD spectra in Figure 3 suggests that the poly(C) chain takes the A form conformation29 in poly(C) + s-SPG as well as in poly(C). Therefore, we conclude that poly(C)’s conformation is considerately altered in the mixture (different tilting angle and shorter base distance), although it is still classified to the A form. Figure 4 presents the CD spectra for the poly(A) system and some of the data are presented in Table 2. Poly(A) and poly(A) + t-SPG show the exactly same spectra, and the poly(A) spectrum in the mixed solvents shows decreasing intensity with decreasing Vw. Although the results were not shown, neither dextran, pullulan, nor amylose induces any change in the CD spectrum. We see again the unusual behavior for poly(A) + s-SPG, where the both positive and negative bands are intensified by more than two times stronger than those of poly(A). From Figure 4 and Table 2, we can confirm that the addition of s-SPG enhances the base stacking according to the following three facts: the hypochromism, the blue shift in UV, and the increment of [θ] in CD. When we estimated the value of [θ]max at Rc ) 1 in poly(A) in the same way as in poly(C), again, the resultant value was smaller than that for poly(A) + s-SPG. This means that the original A form conformation of poly(A) is also deformed by the complexation. All the results of the CD and UV measurements consistently indicate that s-SPG specifically interacts with poly(A) and poly(C) to enhance the base stacking and to create a new conformation of the nucleotide. It is worthy to emphasize that this feature is remarkably distinctive and peculiar as the nature of this s-SPG/polynucleotide interaction. To the best of our knowledge, no one has ever found such an interaction between polysaccharides and singlestranded polynucleotides. Complex Formation and Hydrogen-Bonding Interactions. Table 3 summarizes which homo-polynucleotide can induce the CD spectral change when mixed with s-SPG with comparing their conformations reported previously. None of

Figure 5. Spectral change induced by mixing poly(dT) and s-SPG (upper) and poly(dA) and s-SPG (lower).

poly(G), poly(U), poly(I), poly(dG), and poly(dC) shows any change in CD; however, we found that poly(dA) and poly(dT) interact with s-SPG as well as poly(C) and poly(A). The spectral changes for these single stranded DNAs are presented in Figure 5. According to the previous work,30 the guanines in poly(G) and poly(dG) form a tertamer, the uraciles and cytosines in poly(U) and poly(dC) form a dimer, and the inosines in poly(I) form a tetramer or trimer. For all the cases the hydrogen-bonding sites in the bases are used to form the intramolecular interaction. On the other hand, poly(C), poly(A), poly(dA), and poly(dT) do not form such an intramolecular aggression; therefore, their hydrogenbonding sites are unoccupied. There is clear correspondence between the ability of the interaction and the presence of the free hydrogen-bonding

646


Sakurai et al.

Figure 7. Composition dependence of the relative intensity of CD for poly(C), poly(C) + s-SPG, poly(A), and poly(A) + s-SPG with Mpoly(X) ) 2.0 × 10-4 M and Ms-SPG/Mpoly(X) ) 4.5. Figure 6. Comparison of the electrophoresis migration length among three polynucleotides and their complexes.

sites. Namely, only when the hydrogen bonding-site is available, the polynucleotide can interact with s-SPG. This correspondence evidences that the hydrogen-bonding interactions are essential to induce the interaction. This conclusion leads that s-SPG and polynucleotides form a macromolecular complex. Hereinafter, we denote the complex as the poly(X)/s-SPG complex, and it should be distinguished from the symbol of the mixture such as poly(X) + s-SPG. Figure 6 shows the results of electrophoresis for poly(C), poly(C) + s-SPG, poly(A), poly(A) + s-SPG, poly(dA), and poly(dA) + s-SPG. The addition of s-SPG drastically suppresses the migration length for all cases. This result indicates that the migrants in the mixture show much higher molecular weight than in poly(C), poly(A), or poly(dA) itself, respectively. This is another evidence that these polynucleotides form a macromolecular complex with s-SPG. Influence of Hydrophobic Interaction on the Complex Formation. Figure 7 plots the relative intensities of CD against Vw for both poly(A) and poly(C) systems. Here, the relative intensity is defined as the normalized value of [θ]max by that of poly(A) or poly(C) at Vw ) 1. In the low Vw region, the individual polynucleotide, i.e., poly(A) or poly(C), exhibits the same values as expected for each mononucleotide. This means no base stacking (i.e., Rs ) 0). At Vw ) 0.5 for poly(A) and Vw ) 0.6 for poly(C), the intensity begins to increase. These increments indicate that polarity of the solvent becomes significant enough to induce the hydrophilic interaction between the corresponding bases. With increasing Vw, the intensity increases essentially linearly with Vw. On the other hand, the mixtures present completely different behavior from those of each individual polynucleotide. In the low Vw region (ca., 0-0.45) the CD intensity stays at the same value as that of the mononucleotide. At Vw ) 0.58 for poly(A) + s-SPG and Vw ) 0.6 for poly(C) + s-SPG, the intensities drastically increase and they level off at Vw > 0.7. The saturation of the CD intensity indicates that the

complex formation is completed and most bases are stacking (Rs ) 1) in Vw > 0.7. The onset composition for the complex formation is close to where the stacking begins in the poly(A) and poly(C) solutions, respectively. This coincidence indicates that the hydrophobic interaction between bases is also a key to induce the complex formation. Careful examination of the difference in the onset composition between poly(A) + s-SPG and poly(C) + s-SPG reveals that the composition of the poly(A) system is lower than that of the poly(C) system. This can be ascribed to the fact that the adenine moiety is more hydrophobic than the cytosine moiety due to the difference in the number of the aromatic rings. This fact is consistent with our conclusion for importance of the hydrophobic interaction between the bases. At the onset composition s-SPG is already capable to use both the hydrophobic and hydrogen-bonding interactions. Once the base and s-SPG unit come close by the hydrophobic interaction, they can form the hydrogen bonds. Consequently, we can reasonably suppose that the complexation is triggered by combination of the hydrophobic and hydrogen-bonding interactions. Probably, they act synergistically to create this polysaccharide/polynucleotide complex. The synergistic effect is usually intensified in the macromolecular system; therefore, this is the reason for the abrupt formation of the complex at the onset composition. “Melting” Behavior of the Complex. Figure 8 compares the T dependence of the CD spectra between poly(C) and poly(C) + s-SPG at Vw ) 0.93. With increasing T, the maximum position shifts from 274 to 278 nm and the intensity decreases. These changes occur gradually for poly(C) but, on the other hand, discontinuously for poly(C) + s-SPG. This feature is more clearly demonstrated in Figure 9, where the [θ]274 values are plotted against T at Vw ) 0.75, 0.85, 0.93 and poly(C) in water. In the poly(C) solutions, [θ]274 decreases linearly with increasing T and the slope becomes smaller at the lower Vw. In poly(C) + s-SPG, however, the [θ]274 value is independent of Vw at T ) 0-40 °C; at about T ) 50 °C, it drastically decreases, and at T >



Figure 8. Comparison of the temperature dependence of the CD spectra between poly(C) and poly(C) + s-SPG.

Figure 9. Comparison of [θ]274 vs T plots between poly(C) and poly(C) + s-SPG.

60 °C, it merges into that of poly(C) solution with each Vw. This T dependence of poly(C) + s-SPG evidently demonstrates the autoaccelerate cleavage of the complex. It is interesting that there is remarkable similarity between the dissociation of the complex and the melting behavior of double DNA helixes. This similarity is also demonstrated by the T dependence of at λ ) 267 nm shown in Figure 10, where max increases discontinuously for poly(C) + s-SPG whereas it gradually increases for poly(C) with increasing T. In the double DNA helixes, the base staking is stabilized by the hydrogen-bonding interaction with the complementary base. This is the reason that the double helix dissociates in a cooperative manner. According to Causley et al.,27 the T dependence of [θ]max during the melting of DNA can be expressed by eq 1. [θ]max ) [θ]un + ([θ]st - [θ]un)/[1 + exp(∆H/RT-∆S/R)] (1) Here, [θ]un and [θ]st are the unstacked and stacked ellipticities, respectively, and ∆H and ∆S are the enthalpy and entropy changes, respectively. We assumed eq 1 applicable to the dissociation in Figures 9 and 10 and obtained ∆H and ∆S to be -60 to -70 kcal mol-1 and -150 to -200 cal mol-1 K-1, respectively.31 The resultant ∆H values are in the same order as those for double helix DNA.32

Figure 10. Meting behavior observed by UV spectroscopy for poly(C) + s-SPG.

As we already reported, the poly(A)/s-SPG complex melts in a cooperative manner similar to that of the poly(C)/sSPG complex.18 However, it melts at the lower temperature than the poly(C) complex, i.e., 32-35 °C.18 As shown in Figures 9 and 10, poly(C)/s-SPG melts at 55 °C. This difference is similar to the fact that the poly(C)/poly(G) complex melts at higher temperature (ca. 110 °C) than that of the poly(A)/poly(U) complex (ca. 65 °C).6 The difference in the DNA helix melting temperature can be ascribed to the difference in the number of the hydrogen bonds (cytosine has three hydrogen-bonding sites whereas adenosine has the two hydrogen-bonding sites). The same explanation should be applicable to the difference in the melting temperature of the s-SPG/polynucleotide complexes. Determination of Stoichiometry for the Complex. The values of [θ]max are plotted against for the poly(A) and poly(C) systems in Figure 11. The composition dependence of [θ]max is linear at small fms, turns to be upward convex, and finally levels off at a certain value. For both cases, the cross section of the plateau and increment regions can be read off without ambiguity and the stoichiometric ratio (n ) Ms-SPG/ (Ms-SPG + Mpoly(X))) is evaluated to be 0.39-0.41 for poly(C)/s-SPG and 0.41-0.45 for poly(A)/s-SPG. Although the

648


Sakurai et al.

Figure 12. Four possible structures for the poly(C)/s-SPG complex. G and C denote the glucose and cytosine in the chains, respectively. In model D1, every glucose residue in the main chain interacts with the cytosine. Models D2 and D3 improve the ratio of the glucosecytosine interaction to agree with the stoichiometric number. Model T supposes that the complex adopts a triple helix similar to the original t-SPG. The numbers under the model name indicate the stoichiometric numbers for each model.

Figure 11. Composition dependence of [θ]274 for poly(C) + s-SPG (upper) and [θ]261 for poly(A) + s-SPG (lower). In the poly(C) measurements, (9) Mpoly(C) was kept constant at 2.5 x 10-4 M/L while fms was varied from 0 to 0.16, and (b) fms was changed from 0 to 0.6 while fixing Ms-SPG + Mpoly(C) ) 2.0 x 10-4 M. For the poly(A) system, all the data were obtained while changing fms from 0 to 0.6 and fixing Ms-SPG + Mpoly(A) ) 4.0 x 10-4 M.

data are not shown, the upward convex region became larger with decreasing Vw. This feature can be explained by the lower complex formation constant at lower Vw. With decreasing Vw, it became difficult to determine n accurately, however, the n value did not depend on Vw. Proposing a Possible Molecular Model for the s-SPG/ Poly(C) Complex. It may be still premature to discuss the molecular structure of the complexes based on the results obtained so far. Nevertheless, we can speculate a few possible structures on the assumption that (1) since the C-2 carbon’s OH groups of the main chain, which are located inside of the triple helix column, generate the hydrogen-bonding interaction,16,17 the same OH groups should be involved in the complex formation, (2) the side glucose chain only plays a role to provide solubility and is not involved in the complex formation, (3) the CD and UV data for the poly(C) complex support the formation of a helical structure, which means the distance between the stacked bases should be around 3.6 Å,6 (4) in order to explain the cooperative dissociation, we also have to assume that the complex adopts the helical or regular structure, and (5) since the local polymer conformation in solutions is sometimes related to the crystalline

structure, the complex structure may be similar to either the double helix of DNA or the triple helix of SPG. Figure 12 illustrates the four possible structures for the complex. In model D1, every glucose residue in the main chain interacts with the cytosine in a DNA-like double helix. Since the distance between the neighboring C-2 carbon’s OH is about 4 Å, this model is reasonable from the stereochemistry. However, the stoichiometric number (n ) 0.20) does not agree with the experimental result. Models D2 and D3 improve the ratio of the glucose to the cytosine to agree with the stoichiometric number: in model D2 two of three glucose residues periodically interact with two cytosines, whereas they interact randomly in model D3. In order to form the double strands, however, the ribose or phosphate moieties have to be stretched unfavorably, thus these models are not acceptable in the stereochemistry. In model T, we suppose that the complex takes a triple helix similar to the original t-SPG. When we construct the helix from two s-SPG chains and one poly(C) chain, this model seems to be consistent from the viewpoints of both the stoichiometric number (n ) 0.40) and the stereochemistry. At this moment, therefore, model T is the most acceptable and reasonable structure to explain the experimental data. Needless to say, extensive X-ray crystallography is necessary to prove this model, and we are in the process of doing it. Evaluation of the Complex Size from Small-Angle X-ray Scattering (SAXS). Figure 13 shows the plots of ln qI vs q2 for poly(C) and poly(C) + s-SPG, where q and I are the magnitude of the scattering vector and the scattering intensity, respectively. Since the scattering from the s-SPG solution was weaker by about 1 order of magnitude than those of poly(C) and poly(C) + s-SPG solutions, the difference in the scattering profiles between poly(C) and poly(C) + s-SPG can be ascribed to the conformational difference. According to Guinier,33 the q dependence of I for rod like molecules is expressed by I(q) ∝ exp

( )

Rg2 2 q /q 2

(2)

where Rg is the cross-sectional radius of gyration of the molecules. From the initial slopes indicated by (1) and (2) in the figure, Rg is evaluated and the resultant values are



Committee (No. 2000B0227-NL -np and 2000B0228-NOL -np). List of Abbreviations and Symbols Used in This Paper

Figure 13. Comparison of SAXS between poly(C) and poly(C) + s-SPG. From the initial slopes, Rg was calculated with eq 2 and the results are presented in the figure. Because of ambiguity in taking the initial slope for poly(C), the possible error range is indicated by the two lines.

presented in the figure. In poly(C), both helical and random coil conformations coexist at room temperature,6,25 therefore, there is ambiguity to determine Rg form the initial slope of this plot. On the other hand, poly(C) + s-SPG shows a scattering from a typical rod molecules and the Rg is evaluated more correctly than poly(C) to be 1 ( 0.2 nm. When we suppose the solid cylinder model, Rg is related to the cylinder diameter (r) by r ) x8 Rg The resultant value of 2.8 nm is close to that for the triple helix of SPG (2.6 nm)16 and slightly larger than that of DNA double helix (2.3 nm).16 This result is quite consistent with the triple helical model for the complex. Concluding Remarks We demonstrated that s-SPG interacts with poly(C), poly(A), poly(dA), and poly(dT) through the hydrogen-bonding interactions to form a peculiar macromolecular complex. With focusing on the poly(C) and poly(A) systems, we revealed that (1) the hydrophobic interaction is important as well as the hydrogen-bonding interaction, (2) the complex is dissociated in the cooperative manner with increasing temperature, as being similar to the melting behavior of DNA double helix, (3) the order of the free energy of the complex is the same magnitude as that of DNA double helix, and (4) the stoichiometric number of the complex is 0.4. To explain these results, we propose that the complex structure may be consistent with a triple helix consisting of two s-SPG and one poly(C) chains. These features imply that the s-SPG chain can behave like a poly(G) in the double strand poly(C)/poly(G). The complementary mimic nature we reported here is surprisingly unusual with noticing the fact that there is no base or specific interaction site in the s-SPG chain and may open a door to new application fields of polysaccharides in the gene technology, such as the antisense DNA drugs, nonvirus vectors, and affinity chromatography. Acknowledgment. We thank Taito Co. for kindly providing the schizophyllan sample and Professor I. Hamachi at Kyushu University for helpful discussions. The SAXS part is performed under the approval of the SPring 8 Advisory

DMSO s-SPG t-SPG CD GMP PDDA PNA SAXS UV n [η] [θ] [θ]max [θ]st [θ]un max fms I(q) λmax Mpoly(X) Ms-SPG pH* q Rg Rs Vw

dimethyl sulfoxide single chain of schizophyllan triple helix of schizophyllan circular dichroism guanosine 5′-monophosphate poly(diallydimethylammonium chloride) peptide nucleic acid small-angle X-ray scattering ultraviolet stoichiometric ratio of schizophyllan in the complex defined by n ) Ms-SPG/(Ms-SPG + Mpoly(X)) intrinsic viscosity molecular ellipticity molecular ellipticity at the top of the positive band molecular ellipticity in the staked state molecular ellipticity in the unstacked state extinction coefficient at λmax molar ratio defined by Ms-SPG/(Ms-SPG + Mpoly(X)) scattering intensity as a function of q wavelength to give the maximum absorbance base molar concentration of polynucleotides molar concentration of s-SPG apparent pH magnitude of the scattering vector cross-sectional radius of gyration molar ratio of the base involved in stacking water vol fraction in water/DMSO mixture

References and Notes (1) (a) Lemieux, R. U. Chem. Soc. ReV. 1989, 19, 347. (b) Davis, A. P.; Wareham, R. S. Angew. Chem., Int. Ed. 1999, 38, 2978. (2) (a) Rebek, Jr., J.; Marshall, L.; Wolak, L.; Parris, K.; Killoran, M.; Askew, B.; Nemeth, D. J. Am. Chem. Soc. 1985, 107, 7476. (b) Rebek, Jr., J.; Nemeth, D. J. Am. Chem. Soc. 1986, 108, 5637. (c) Rebek, Jr., J.; Askew, B.; Killoran, M.; Nemeth, D.; Lin, F. T. J. Am. Chem. Soc. 1987, 109, 2426. (d) Rebek, Jr., J. Angew. Chem., Int. Ed. Engl. 1990, 29, 245. (e) Hamilton, A. D.; van Engen, D. J. Am. Chem. Soc. 1987, 109, 5035. (f) Chang, S. K.; Hamilton, A. D. J. Am. Chem. Soc. 1988, 110, 1318. (g) Hamilton, A. D.; Pant, N. J. J. Chem. Soc., Chem. Commun. 1988, 765. (h) Goswam, S.; Hamilton, A. D. J. Am. Chem. Soc. 1989, 111, 3425. (i) Kelly, T. R.; Maguire, P. M. J. Am. Chem. Soc. 1987, 109, 6549. (j) Kelly, T. R.; Zhao, C.; Bridger, G. J. J. Am. Chem. Soc. 1989, 111, 3744. (k) Etter, M. C.; Paunto, T. W. J. Am. Chem. Soc. 1988, 110, 5896. (l) Etter, M. C.; Adsmond, D. A. J. Am. Chem. Soc. 1990, 112, 4549. (m) Bell, T. W.; Liu, J. J. Am. Chem. Soc. 1988, 110, 3673. (n) Aoyama, Y.; Tanaka, Y.; Toi, H.; Ogoshi, H. J. Am. Chem. Soc. 1988, 110, 634. (o) Tanaka, Y.; Ubukata, Y.;. Aoyama, Y. Chem. Lett. 1988, 1905. (p) Kano, K.; Yoshiyasu, K.; Hashimoto, S. J. Chem. Soc., Chem. Commun. 1988, 801. (3) (a) Kikuchi, Y.; Kobayashi, K.; Aoyama, Y. J. Am. Chem. Soc. 1992, 114, 1351. (b) Kikuchi, Y.; Tanaka, Y.; Sutarto, S.; Kobayashi, K.; Toi, H.; Aoyama, Y. J. Am. Chem. Soc. 1992, 114, 10302. (c) Kobayashi, K.; Asakawa, Y.; Kato, Y.; Aoyama, Y. J. Am. Chem. Soc. 1992, 114, 10307. (4) (a)Lumb, K. J.; Kim, P. S. Science 1995, 268, 436. (b)Ponomarenko, E. A.; Tirrell, D. A.; MacKnight, W. J. Macromolecules 1998, 31, 1584. (5) For reviews, see: (a) Neoglycoconjugates: Preparation and Application; Lee, Y. C.; Lee, R. C., Eds.; Academic Press: San Diego, CA, 1994. (b) Lee, Y. C.; Lee, R. C. Acc. Chem. Res. 1995, 28, 321.

650


(6) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New York, 1984; Chapter 6. (7) Lee, M. D.; Dunne, T. S.; Chang, C. C.; Siegel, M. M.; Morton, G. O.; Elleatad, G. A. J. Am. Chem. Soc. 1992, 114, 985. Calicheamicin binds to the miner groove of DNA using its aglycone phenyl group to interact with the guanine in DNA. Calicheamicin’s saccharide moieties form the hydrogen bonding with DNA; however, the major factor is the hydrophobic interaction between the aromatic groups. Therefore, its interaction mechanism is different from the present case. (8) For example, Davis, S. Oligosaccharide as ligands to DNA: The Molecular and Supermolecular Chemistry of Carbohydratet; Oxford University Press: Oxford, England, 1997; pp 300-309. (9) Dervan, P. B.; Bu¨rli, R W. Curr. Opinion Cxhem. Biol. 1999, 3, 688 and references therein. (10) (a) Egholm, M.; Buchardt, O.; Nielsen, P. E.; Berg, R. H. J. Am. Chem. Soc. 1992, 114, 1895. (b) Lagriffoule, P.; Wittung, P.; Eriksson, M.; Kilsa, K. J.; Norden, B.; Buchardt, O.; Nielsen, P. E. Chem.sEur. J. 1997. (c) Hyrup, B.; Egholm, M.; Nielsen, P. E.; Wittung, P.; Norden, B.; Buchardt, O. J. Am. Chem. Soc. 1994, 116, 7964. (11) Larsen, H. J.; Nielsen, P. E. Nucleic Acids Res. 1996, 24, 458. (12) Norton, J. C.; Piatyszek, M. A.; Wright, W. E.; Shay, J. W.; Corey, D. R. Nature Biotechnol. 1996, 14, 615. (13) Tabata, K.; Ito, W.; Kojima, T.; Kawabata, S.; Misaki, A. Carbohydr. Res. 1981, 89 121. (14) Yanaki, T.; Ito, W.; Tabata, K.; Kojima, T.; Norisuye, T.; Takano, N.; Fujita, H. Biophys. Chem. 1983, 17, 337. (15) Young, S.; Jacobs, R. R. Carbohydr. Res. 1998, 310, 91. (16) (a) Norisuye, T.; Yanaki, T.; Fujita, H. J. Polym. Sci., Polym. Phys. Ed. 1980, 18, 547. (b) Norisuye, T.; Yanaki, T. Polym. J. 1982, 15, 389. (c) Yanaki, K.; Norisuye, T.; Fujita, H. Macromolecules 1980, 13, 1462. (d) Sato, T.; Sakurai, K.; Norisuye, T.; Fujita, H. Polym. J. 1983, 16, 559. (17) Itou, T,; Teramoto, A.; Matuo, M.; Suga, H. Carbohydr. Res. 1987, 160, 243.

Sakurai et al. (18) (a) Sakurai, K.; Shinkai, S. J. Am. Chem. Soc. 2000, 122, 4520 (b) Sakurai, K.; Shinkai, S.; Tabata, K. Japanese Patent, file number 11319470, 1999. (19) Technical data report in Amersham Phamacia Catalog, 1999. (20) The 2nd Spring 8 meeting; Japan Atomic Energy Research Institute, Eds.; 1998; p 48. (21) Fujisawa, T. J. Jpn. Soc. Synchrotron Radiat. Res. 1999, 12, 194. (22) Johnson, W. C., Jr.; Tinoco, I., Jr. Biopolymers 1969, 7, 727. (23) Since the CD intensity of poly(C) is 10 times as large as that of cytidine-momophosphate, the proportionality between Rs and the CD intensity can be approximately valid. (24) For example, see: Tuchida, E. Kagakusousetu 1997, 31, 182. (25) (a) Brahms, J.; Marurizot, J. C.; Michelson, A. M. J. Mol. Biol. 1967, 25, 465. (b) Marurizot, J. C.; Bricharski, J.; Brahms, J.; Biopolymers 1971, 10, 1429. (c) Warner, R. C. J. Biol. Chem. 1957, 229, 711. (26) Baba, Y.; Fujioka, K.; Kagemoto, A. Polym. J. 1978, 10, 241. (27) Causley, G.; Staskus, P. W.; Johnson, Jr. W. C. Biopolymers 1983, 22, 946. (28) Nakajima, K.; Brrova, N. Circular Dichroism: Principles and Application; VCH: New York, 1994. (29) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New York, 1984; Chapter 17. (30) (a) Saenger, W. Principles of Nucleic Acid Structure; SpringerVerlag: New York, 1984; Chapter 13. (b) Howard, F. B.; Miles, H. T. Biochemistry, 1982, 21, 6736. (31) The method used here may not be accurate to give a further detailed discussion, but the order of the magnitude of the obtained ∆H and ∆S should be correct (see ref 27) and our method is good enough to lead the conclusion in this section. (32) Sugimoto, N.; Nakano, S.; Yoneyama, M.; Honda, K. Nucleic Acid Res. 1996, 24, 4501. (33) Glatter, O.; Kratky, O. Small-Angle X-ray Scattering; Academic Press: London, 1982.

BM000121R

PolysaccharideâPolynucleotide Complexes. 2. Complementary

Recommend Documents

PolysaccharideâPolynucleotide Complexes. 2. Complementary