Water-Soluble Dendrigrafts Bearing Saccharidic Moieties: Elaboration

to load: https://cdn.mathjax.org/mathjax/contrib/a11y/accessibility-menu.js .... and Application to Enzyme Linked OligoSorbent Assay (ELOSA) Diagn...
0 downloads 0 Views 697KB Size
Bioconjugate Chem. 2006, 17, 6−14

6

ARTICLES Water-Soluble Dendrigrafts Bearing Saccharidic Moieties: Elaboration and Application to Enzyme Linked OligoSorbent Assay (ELOSA) Diagnostic Tests Julien Bernard,† Michel Schappacher,† Alain Deffieux,*,† Pascal Viville,‡ Robert Lazzaroni,‡ Marie-He´le`ne Charles,§ Marie-The´re`se Charreyre,*,§ and Thierry Delair§ Laboratoire de Chimie des Polyme`res Organiques, UMR 5629 CNRS-ENSCPB-Universite´ Bordeaux I, 16 Avenue Pey Berland, 33607 Pessac, France, Unite´ Mixte CNRS-bioMe´rieux, UMR-2714, ENS-Lyon, 46, Alle´e d’Italie 69364 Lyon Cedex 07, France, and Service de Chimie des Mate´riaux Nouveaux, Universite´ de Mons-Hainault, Mons, Belgium. Received December 7, 2004; Revised Manuscript Received September 15, 2005

The synthesis of a series of water-soluble galactopyranose-functionalized polystyrene-polyvinyl ether dendrigrafts and their characterization (in solution and thin solid deposits) have been achieved. The presence of external galactopyranose groups on dendritic polymers has been exploited to prepare dendrigraft-oligonucleotide conjugates using a simple one-step coupling procedure with amino-ended oligonucleotides (ODNs). Several parameters such as the peripherical density of hydrophilic branches, the polymerization degree of polystyrene or poly(hydroxyethyl vinyl ether) blocks, and the number of galactopyranose groups were tuned. A capture test with short labeled complementary ODNs (25 bases) confirmed the presence of covalently bound ODNs on various kinds of dendrigrafts. The ability of the dendritic polymers to enhance the sensitivity of enzyme-linked oligosorbent assay (ELOSA) diagnostic tests (detection of hepatitis B virus, DNA target of 2400 bases) was then evaluated, especially the influence of the macromolecular architecture and the impact of the structural parameters. The dendrigraftODN conjugate with the lower saccharide external density was found to lead to a very significant amplification of the fluorescence signal, corresponding to a limit of sensitivity of 109 DNA copies per milliliter (instead of 1011 DNA copies per milliliter without using dendrigrafts). Conversely, the dendrigrafts exhibiting a very high number of branches and galactopyranose groups at their periphery were not able to induce a better sensitivity due to steric hindrance generated by the peripheral congestion on these polymers.

INTRODUCTION ELISA1

Contrary to the traditional diagnostic tests, ELOSA tests are based on the capture and detection of nucleic acids and provide a direct and specific detection of genetic or infectious diseases. The principle of these tests lies, first, in the capture of a target material (single-stranded DNA sequences) by oligonucleotide probes anchored onto a solid support and, second, in the detection and quantification of the captured DNA †

Universite´ Bordeaux I. Unite´ Mixte CNRS-bioMe´rieux. § Universite´ de Mons-Hainault. 1 List of abbreviations: DLS, dynamic light scattering; DMF, dimethyl formamide; DMSO, dimethyl sulfoxide; EGP, 1,2:3,4-di-Oisopropylidene-6-O-epoxypropyl-D-galactopyranose; ELISA, enzymelinked immunosorbent assay; ELOSA, enzyme-linked oligosorbent assay; DiGVE, 1,2:3,4-di-O-isopropylidene-6-O-(2-vinyloxy ethyl)-Dgalactopyranose; Gal, galactose; HBV, hepatitis B virus; NMP, N-methylpyrrolidone; ODN, oligonucleotide; PCEVE, poly(chloroethyl vinyl ether); PDiPGVE, poly(1,2:3,4-di-O-isopropylidene-6-O-(2-vinyloxy ethyl)-D-galactopyranose); PEO, poly(ethylene oxide); PEGP, poly(1,2: 3,4-di-O-isopropylidene-6-O-epoxypropyl-D-galactopyranose); PGVE, poly(6-O-(2-vinyloxy ethyl)-D-galactopyranose); PHEVE, poly(hydroxyethyl vinyl ether); PS, polystyrene; PSiVE, poly(tert-butyldimethylsilyloxyethyl vinyl ether); PVE, poly(vinyl ether); SEC, size exclusion chromatography; SiVE, 5,10,15,20-((tert-butyldimethylsilyl)oxy)ethyl vinyl ether; SLS, static light scattering; THF, tetrahydrofuran; TMAFM, tapping mode atomic force microscopy; VPO, vapor pressure osmometry ‡

target (colorimetric or fluorescent signal). As the quantities of target material are typically very low, this strategy generally includes either amplification of the target sequence (1) or signal amplification, to enhance the sensitivity of the test. Since the target amplification strategy gives rise to numerous drawbacks such as cross-contamination or poor reproducibility, attention has then been paid to the second approach. Several signal amplification strategies have been proposed, either based on the introduction of multiple reporter groups such as biotin (2), fluorophores (3), or 2,4-dinitrophenyl (4) in the ODN used for the detection or focusing on the use of branched oligonucleotides (5, 6) in combination with detection probes labeled with an enzyme. However, the increase of sensitivity obtained by these routes is still limited. An alternative way consists of the preparation of (co)polymer-ODN amplification structures (conjugates). With this aim, a number of new polymer supports, consisting of functional (co)polymers bearing different reactive groups (activated ester, anhydride, aldehyde), have been developed to immobilize ODNs (7-17). The conjugates were typically prepared either by binding amino-ended single-stranded oligonucleotide probes (synthesized via automated synthesis with an -NH2 group at the 5′ position) onto reactive (co)polymers via amination or amidation reaction (7-15) or by direct synthesis of the ODNs from the polymer backbone (16-17). Both kinds of conjugates enable the capture of complementary DNA targets through hybridization and lead to an increase in the sensitivity of the ELOSA test.

10.1021/bc049708j CCC: $33.50 © 2006 American Chemical Society Published on Web 12/24/2005

Water-Soluble Dendrigrafts Bearing Saccharidic Moieties

Bioconjugate Chem., Vol. 17, No. 1, 2006 7

Scheme 1. Reaction Pathways for the Elaboration of Hydrophobic Hyperbranched Precursors [PCEVE-g-(PS-b-PCEVE) or PCEVE-g-(PS-b-(PCEVE-g-PS))] and Amphiphilic Dendrigrafts [PCEVE-g-(PS-b-(PCEVE-g-(PEO-b-PEGP))) or PCEVE-g-(PS-b-(PCEVE-g-(PS-b-PHEVE-b-PGVE)))]

Concerning polymer-ODN conjugate properties desirable for an effective enhancement of the signal, it was shown that accessibility of the immobilized ODNs (rather than their number) was a major issue. Among the above-mentioned functional (co)polymers, various architectures have been designed such as (i) linear (co)polymers, namely poly(maleic anhydride-co-methyl vinyl ether) (8), poly(N-vinylpyrrolidoneco-N-acryloxysuccinimide) (7, 10), poly(N-(2,2-dimethoxyethyl)-N-methyl acrylamide) (11), and synthetic glycopolymers (9, 12, 13), (ii) graft (co)polymers (14, 15), and (iii) latex particles (18, 19). However, to our knowledge, no study describing the use of dendrigrafts as supports for ODN immobilization has been previously reported. In a recent article, we described the synthesis of hydrosoluble amphiphilic dendrigrafts bearing saccharide groups located at the periphery (20). These polymers were prepared from functional arborescent PS precursors (21) by taking advantage of the presence of terminal acetal functions on every PS branch to grow poly(vinyl ether) blocks of poly(tert-butyldimethylsilyloxyethyl vinyl ether) and poly(1,2:3,4-di-O-isopropylidene6-O-(2-vinyloxyethyl)-D-galactopyranose) via living cationic polymerization. After deprotection, the dense hydrophilic PVE shell (PHEVE and PGVE) surrounding the PS core ensured the solubilization of the dendrigrafts in water. The presence of both PHEVE blocks (acting as hydrophilic spacers) and PGVE external blocks (bearing galactopyranose groups with the potential ability to covalently immobilize amino-ended oligonucleotides (13, 15)) on these dendritic polymers encouraged us to evaluate their efficiency as supports for ELOSA diagnostic tests. In addition, we also developed a second route to galactopyranose-functionalized water-soluble dendrigrafts based on the coupling of living polyoxirane chains onto PCEVE-g(PS-b-PCEVE) precursors during this work. In this article, we report the synthesis and the characterization of a series of galactopyranose-functionalized dendrigrafts. Several parameters such as the peripheral density of branches, the polymerization degree of PS or PHEVE blocks, and the

number of Gal groups were tuned. The ability of these polymers to enhance the sensitivity of ELOSA tests was then evaluated, especially the influence of the macromolecular architecture and the impact of the structural parameters on the results of the test.

EXPERIMENTAL PROCEDURES Materials. Toluene (99.5%, J. T. Baker, Deventer, The Netherlands) was purified by distillation over calcium hydride and stored over polystyryllithium seeds. Tetrahydrofuran (99%, J. T. Baker, Deventer, The Netherlands) was purified by distillation over calcium hydride and stored over potassium/ benzophenone. Ethylene oxide (99.5%, Sigma Aldrich Chimie, St Quentin Fallavier, France) was purified and stored over calcium hydride. Dimethyl sulfoxide (99.9%, Sigma Aldrich Chimie, St Quentin Fallavier, France), dimethyl formamide (99.9%, Sigma Aldrich Chimie, St Quentin Fallavier, France), N-methylpyrrolidone (99.9%, Sigma Aldrich Chimie, St Quentin Fallavier, France), epichlorhydrine (99%, Sigma Aldrich Chimie, St Quentin Fallavier, France), 2-hydroxyethyl vinyl ether, 1,2: 3,4-di-O-isopropylidene-R-D-galactopyranose, and potassium (98%, Sigma Aldrich Chimie, St Quentin Fallavier, France) were used as received. SiVE was prepared as described previously by Fukui et al. (22) from 2-hydroxyethyl vinyl ether and distilled over calcium hydride twice before use. DiGVE was prepared as recently described by D’Agosto et al. (13). EGP was synthesized from epichlorhydrine and 1,2:3,4-di-O-isopropylidene-R-D-galactopyranose according to the procedure reported by Koell et al. (23). Synthesis of the PCEVE/PS Dendrigraft Precursors. The dendrigraft precursors 200/50/50/A150 (A) (DPnPCEVE1/DPnPS1/ DPnPCEVE2/DPnA1-PS2, A1 ) R-acetal group), 200/50/50/A15 (B), and 200/120/40 (C) used in this work were elaborated according to a previously reported procedure (21, 24) (Scheme 1). PCEVE/PS dendrigraft precursor properties. A: Mw ) 3.8 × 107g‚mol-1; PDI ) 1.05; Rg ) 34.5 nm; Rh ) 34 nm. B:

8 Bioconjugate Chem., Vol. 17, No. 1, 2006

Mw ) 0.6 × 107g‚mol-1; PDI ) 1.18; Rg ) 22.5 nm; Rh ) 25 nm. C: Mw ) 0.3 × 107g‚mol-1; PDI ) 1.13. Synthesis of Amphiphilic Dendrigrafts (I, II, III) with a Hydrophilic PVE Shell. The dendrigrafts were prepared as previously described (20) via a “grafting from” procedure: initiation of living cationic polymerization of SiVE and DiGVE from the diethylacetal ends of PS branches of the PCEVE/PS hyperbranched cores, using trimethylsilyl iodide and zinc dichloride as a catalytic system. Typically, the chain extension was carried out under nitrogen in toluene at -30 °C in a 250 mL baked flask with a double-way stopcock equipped with a magnetic stirring bar. The PS precursor (200/50/50/A150, 5 g, 1 mmol acetal termini) was dissolved in anhydrous toluene and vacuum-dried to remove traces of moisture. The polymer and SiVE (19 g, 93 mmol) were subsequently dissolved in anhydrous toluene (100 mL) under nitrogen. Trimethylsilyl iodide (0.17 mL, 1.2 mmol) was then added to the solution and vigorously stirred for 1 h. A ZnCl2 solution (1 mL, 3 mmol‚L-1 in diethyl ether) was added to initiate the polymerization of SiVE. After complete conversion of SiVE (12 h), DiGVE (5 g, 15 mmol) was added to the living polymer solution to grow a third block and generate external PS-b-PSiVE-b-PDiGVE branches. After 10 h, the polymerization was stopped by adding a methanol/ lutidine mixture. The polymer was washed several times with potassium thiosulfate aqueous solution (2 wt %) and then with water. The solvent was evaporated, and the polymer was precipitated in methanol. The deprotection of the saccharidic moieties was achieved by treatment with a trifluoroacetic acid/ water mixture (v/v, 5/1, 3 h, room temperature), and the hydroxyl functions on PSiVE were regenerated via a treatment with NBu4F in THF (3 h, room temperature). After deprotection of the hydroxyl groups of SiVE and DiGVE, the dendrigrafts were purified by dialysis against pure water (I, II) or methanol (III) for 3 days (membrane Spectra/Por 7). On the basis of the previously measured PS precursor characteristics (Mw, DPnPS2, number of external branches) and assuming quantitative initiation from the precursor acetal ends, PSiVE and PDiGVE degrees of polymerization were evaluated by 1H NMR, from the relative integration of PS aromatic (5H), tert-butyldimethylsilyl group (15H), and saccharidic protons (1H, H1 or H2). The molecular weight of the dendrigrafts was calculated with the following equation: MnRMN ) Mn(precursor) + (DPnPSiVE (RMN) × mSiVE + DPnPDiGVE (RMN) × mDiGVE), with n equal to the number of branches in the precursor. The number of galactopyranose groups per dendrigraft was determined from the number of PS2 external branches and the degree of polymerization of PDiGVE (Nber of Gal ) Nber of external branches × DPnPDiGVE). 1H NMR (CDCl ) δ (ppm): -0.1-0.2 (-Si-CH ), 0.7-1.1 3 3 (-Si-C-CH3), 1-2.2 (PSiVE and PDiGVE backbones: -CH2-CH-O-; PS backbone: -CH2-CH-; PDiGVE: CH3C-O), 3-3.8 (PSiVE and PDiGVE backbones: -CH2-CHO-; PSiVE and PDiGVE pendent groups: O-CH2-CH2-O; PDiGVE pendent groups: O-CH2-CH-), 3.8-4 (PDiGVE, H5), 4.1-4.3 (PDiGVE, H3 and H4), 4.5-4.7 (PDiGVE, H2), 5.4-5.6 (PDiGVE, H1), 6.2-7.2 (PS, Ph). Synthesis of Amphiphilic Dendrigrafts (IV) with a Hydrophilic Polyoxirane Shell. (a) Preparation of the PEGP-bPEO-K+ Diblocks. The synthesis of PEGP-b-PEO chains was carried out by living anionic polymerization. The potassium alcoholate salt of 1,2:3,4-di-O-isopropylidene-D-galactopyranose was first prepared under vacuum by addition of 1 g of 1,2:3,4di-O-isopropylidene-R-D-galactopyranose (3.8 mmol) dissolved in 150 mL of distilled THF in a flask containing an excess of potassium (0.3 g, 7.7 mmol, ∼2 equiv). After reacting for 24 h at room temperature, the slightly yellow solution of alcoholate

Bernard et al.

was filtered under vacuum to eliminate residual potassium. Polymerization of EGP was carried out in THF at 50 °C in a 500 mL flask equipped with a magnetic stir bar after addition to the potassium alcoholate solution of 6 g of EGP (19 mmol) dissolved in 50 mL of dry THF. After reacting for 24 h, an aliquot of the solution was collected and analyzed by size exclusion chromatography in THF, confirming the complete polymerization of EGP. Ethylene oxide (15 g, 0.34 mol) was then added to the “living” polymer solution at room temperature in order to grow the second block. After 24 h of reaction, an aliquot of PEGP-b-PEO- K+ block copolymer solution was deactivated by addition of degassed methanol and then analyzed by SEC in THF, 1H NMR in CDCl3, and VPO in toluene. Copolymer characteristics: DPn(PEGP) ) 5; DPn(PEO) ) 90 (calculated from 1H NMR spectrum and VPO measurement of PEGP-b-PEO); MwGPC ) 4200 g‚mol-1; MwVPO ) 5880 g‚mol-1; PDI ) 1.14. (b) Synthesis of Amphiphilic Dendrigrafts IV by the Grafting onto Procedure. The synthesis of the amphiphilic dendrigrafts bearing a polyoxirane shell was performed by grafting the living PEGP-b-PEO- K+ chains onto the PCEVE2 blocks of (PCEVE1g-(PS1-b-PCEVE2)) comblike precursors. Typically, the comb copolymer (C, 1.1 g, 3.2 mmol of chloroethyl functions) and NaI (20 mg, 0.13 mmol) were dissolved in 25 mL of anhydrous THF in a 500 mL flask equipped with a double-way stopcock and a magnetic stir bar. To remove traces of moisture, the solvent was evaporated under vacuum and the polymer was vacuum dried for 1 h at 60 °C. This operation was repeated twice. The polymer was then redissolved in 20 mL of anhydrous THF, and the solution containing the “living” oxanionic polyoxirane chains was then slowly added. After 3 weeks, the reaction was deactivated by addition of degassed methanol and 250 mL of CH2Cl2 was introduced. The polymer solution was washed several times with potassium thiosulfate/water (2 wt %), to eliminate iodine, and then with water. The organic solution was then dried with magnesium sulfate, filtered, and precipitated in diethyl ether and redissolved in acetone. The dendrigraft was isolated from unreacted polyoxirane chains by selective precipitation in an acetone/diethyl ether mixture. The pure arborescent polymer was then characterized by SEC (THF), SLS (THF), DLS (THF), and 1H NMR (CDCl ). Deprotection of the saccharidic groups (1 g 3 of polymer in 10 mL of THF) was achieved by acidic treatment with a 5:1, v:v CF3COOH/H2O mixture (10 mL) for 3 h. After neutralization of the solution by addition of aqueous NaOH (1 M), the polymer was purified by dialysis against pure water (membrane Spectra/Por7, molecular weight cutoff ca. 1000) for a week. 1H NMR (CDCl ) δ (ppm): 1-2.2 (PCEVE backbone: 3 -CH2-CH-O-; PS backbone: -CH2-CH-Ph; PEGP: CH3C-O), 3.4-3.8 (PCEVE backbone: -CH2-CH-O-; PCEVE: -O-CH2-CH2-Cl, -O-CH2-CH2-Cl; PEO backbone: -CH2O-; PEGP backbone: O-CH2-CH-O-, O-CH2CH-O-), 3.9-4.1 (PEGP, H5), 4.2-4.4 (PEGP, H3 and H4), 4.5-4.7 (PEGP, H2), 5.4-5.6 (PEGP, H1), 6.2-7.2 (PS, Ph). Binding of ODN onto Galactopyranose-Functionalized Dendrigrafts. NH2-terminated oligonucleotides were synthesized by standard DNA cyanoethyl N,N-di-isopropylamino phosphoramidite chemistry. They were modified at their 5′ positions with an aminohexyl spacer arm for coupling to the polymer. The following sequence was used (given from the 5′ to 3′ position): 5′-TCA-ATC-TCG-GGA-ATC-TCA-ATGTTA-G-3′. To optimize the conditions for ODN coupling onto the dendrigrafts, several organic solvents (DMF, DMSO, and NMP), different aqueous buffer to organic solvent ratios, and various pH and ionic strength values of the aqueous buffer were tested.

Bioconjugate Chem., Vol. 17, No. 1, 2006 9

Water-Soluble Dendrigrafts Bearing Saccharidic Moieties

Typically, 10.2 µL of an aqueous solution of oligonucleotides (5 nmol) and 27 µL of borate buffer (20 mM, pH ) 7.2) were introduced in an Eppendorf tube. After removing water under reduced pressure (speed-vac), the pellet (ODNs and salts) was redissolved in 21.7 µL of pure water. Next, 5.3 µL of an aqueous solution of dendrigrafts IV (100 nmol of galactopyranose moieties) and 243 µL of DMF was introduced in the tube (run 11, total volume 270 µL, 10/90 v/v aqueous buffer/DMF). The tube was finally placed under stirring at 50 °C for 5 days. Capture Test: Dendrigraft-ODN Conjugate Characterization. The binding of ODNs onto the dendrigraft was indirectly confirmed by evaluating the ability of the dendrigraftODN conjugates to capture short complementary ODN sequences labeled with an enzyme (alkaline phosphatase). These capture tests were performed on bioMe´rieux’s Vidas immunoanalysis system. The dendrigraft-ODN conjugates (300 µL, 15 nmol) were first physically adsorbed onto the Vidas solidphase receptacle. After a washing step, the capture step was performed over a period of 1 h with complementary labeled ODNs (300 µL, 10 nmol). After another wash and addition of the enzyme substrate (methyl umbelliferyl phosphate), the emitted signal was measured and expressed by the Vidas detector in relative fluorescence units (RFU). Each series of tests included one control sample without conjugate (labeled ODNs only). ELOSA Tests. Complete ELOSA tests were performed as previously described (9, 10) with the conjugates providing a significant signal in the capture tests. The dendrigraft-ODN conjugates (300 µL, 15 nmol) were first physically adsorbed onto the Vidas solid phase (coating step). After a washing step, the capture step was carried out with increasing concentrations of HBV DNA target (2339 base pairs) (0, 107, 108, 109, 1010, and 1011 DNA copies per mL, 200 µL). Intermediate detection probes (called Mix) ensured the link between the captured DNA targets and the polymer-ODN detection conjugate (300 µL, 15 nmol). This detection conjugate was based on a saccharidic graft copolymer (AF 112) (15). The fluorescent signal (RFU) was finally produced by labeled ODNs hybridized on the polymer-ODN detection conjugate. Polymer Characterization. 1H NMR spectra were recorded in CDCl3 or in D2O on a Bruker AC 200 FT apparatus. Size exclusion chromatography (SEC) measurements of PVE amphiphilic dendrigrafts (I, II, III) before deprotection were performed in tetrahydrofuran at 25 °C (flow rate 0.7 mL/min) on a Varian apparatus equipped with refractive index (Varian) and laser light scattering (Wyatt Technology) dual detection, and fitted with four TSK gel HXL columns (250, 1500, 104, and 105 Å). For the light scattering measurements, the dn/dc values of the different PVE and polyoxirane dendrigrafts were determined in THF at 25 °C with a laser source operating at 633 nm. Size exclusion chromatography analyses of PEO-b-PEGP blocks, dendrigrafts IV, and deprotected PVE dendrigrafts I, II, and III were performed at 25 °C in a water/methanol mixture (v/v 80/20) containing NaNO3 (1.27 g/L) on a Varian apparatus equipped with a refractive index detector and four TSK PWXL columns (300 × 7.7 mm; 250, 1500, 104, and 105 Å), at a flow rate of 1 mL/mn. Dynamic light scattering measurements were performed at 25 °C in THF and in water on a Malvern apparatus (Zetasizer 3000HS) equipped with a laser source (633 nm). Correlation functions were analyzed by the Contin method. Latex particles were used as calibration standards. VPO measurements were performed on a Gonotec Osmomat 090 in toluene. Samples for AFM of the dendrigrafts and their precursors were prepared by solvent casting at ambient conditions on substrates, starting from either MeOH solutions of

Table 1. Dimensions and Solution Characteristics of Amphiphilic PVE-Based Dendrigrafts I, II, and III

a

DPnPS2 DPnPSiVEa DPnPDiGVEa MwRMN × 10-7 a Mw × 10-7 b Mw/Mcc Rgb Rh (THF)d Rh after deprotection (water)d (Nb of Gal units/dendrigraft)e

I

II

A

B

III B

50 90 12 14.4 14.8 1.03 56 55 (34f) 56 96 000

5 110 4 20.3 16.3 1.16 51 44 (25f) 58 33 400

5 11 3 3.1 2.8 1.35 37 36 (25f) --25 000

a Determined by 1H NMR in CDCl . b Measured by SLS in THF at 25 3 °C. c Measured by SEC in THF. d Measured by DLS at 25 °C. e Nb of Gal units/dendrigraft ) (experimental number of branches of A (I) or B (II and III) × DPnPDiGVE). f Rh of PCEVE1-g-(PS-b-(PCEVE2-g-A1PS)) precursors measured by DLS in THF at 25 °C.

dendrigrafts or THF solutions for the precursors. Typically, 20 µL of a dilute solution (0.002-0.1 wt %) was cast on a 1 × 1 cm2 freshly cleaved mica substrate. Samples were analyzed after complete evaporation of the solvent at room temperature. All AFM images were recorded in air with a Nanoscope IIIa microscope operated in tapping mode. The probes were commercially available silicon tips with a spring constant of 2452 N/m, a resonance frequency lying in the 264-339 kHz range, and a typical radius of curvature in the 10-15 nm range. Both the topography and the phase signal images were recorded with the highest sampling resolution available, i.e., 512 × 512 data points.

RESULTS AND DISCUSSION The hydrosoluble dendritic polymers were synthesized from functional PCEVE/PS dendrigraft (A, B) or comb (C) precursors. As previously reported, these monodispersed polymers were prepared by covalently assembling elementary macromolecular blocks by repeated grafting of R-end-functionalized polystyryllithium onto chloroethyl vinyl ether units of linear or comblike chains (obtained via living cationic polymerization) as described in Scheme 1, steps 1, 2, and 3. Two strategies have been developed to prepare hydrosoluble dendritic polymers suitable for DNA immobilization, from these hydrophobic precursors. Synthesis of the Water-Soluble Dendrigrafts I, II, and III. The first route (20) was based on the use of acetal-terminated PCEVE1-g-(PS1-b-(PCEVE2-g-A1PS2)) (precursors A and B) as multifunctional initiators to grow a surrounding hydrophilic shell with galactopyranose units available at the periphery. As previously described, this was achieved by initiationsfrom the terminal acetal groupssof the living cationic polymerization of two types of poly(vinyl ether) blocks (see Scheme 1, step 4): the PSiVE block that ensured the solubilization of the dendrigrafts in aqueous solution after deprotection of hydroxyl groups (PHEVE) and the PDiGVE block that enabled the binding of NH2-ended oligonucleotides after deprotection of the saccharidic moieties. To evaluate the influence of parameters such as the length of the hydrophilic spacer (PHEVE) or the number of the GVE units on the immobilization of ODNs and on the ability of the resulting conjugates to provide an amplification enhancement, we prepared three different dendrigrafts tuning the degree of polymerization of the external PS (DPnPS2 ) 5 or 50), PSiVE, and PDiGVE blocks. The dimensions and the main characteristics of dendrigrafts I, II, and III are indicated in Table 1. The growth of the PVE blocks was confirmed by 1H NMR. The increase of the hydrodynamic

10 Bioconjugate Chem., Vol. 17, No. 1, 2006

Bernard et al. Table 2. Dimensional and Solution Characteristics of the Amphiphilic Polyoxirane-Based Dendrigraft IV Synthesized from Precursor C

Figure 1. SEC chromatograms of dendrigraft IV (SEC water/methanol 80/20 v/v): (A) crude product; (B) dendrigraft IV after fractionation and elimination of unreacted PEGP-b-PEO.

ref

IV

PEGP-b-PEO grafting yield (%) MnRMN × 10-6 (g‚mol-1)a MwLS × 10-6 (g‚mol-1) (THF)a PDI Rhb (THF) (nm) Rhc (water) (nm) Nb of Gal/dendrigraft

15 8.8 10.6 1.16 30 39 6120

a Protected dendrigraft. b Measured by DLS at 25 °C. c Measured by DLS at 25 °C on the unprotected dendrigraft.

Scheme 2. Synthesis of PEGP-b-PEO Living Chains via Anionic Polymerization

Figure 2. Hydrodynamic radius of deprotected dendrigraft IV in water measured by dynamic laser light scattering.

radii (in THF) from 34 nm (A) to 55 nm (I), and from 25 nm (B) to 44 (II) and 36 nm (III) corroborated the successful chain extension from the acetal termini. Consistent with a living polymerization process, the resulting polymers exhibited unimodal and narrow molar mass distributions. The molecular weights of the dendrigrafts (evaluated by 1H NMR and SLS) were ranging from 2.8 to 16.3 × 107 g‚mol-1, and the number of galactopyranose groups introduced on each dendrigraft scaled from 25 × 103 (III) to 96 × 103 (I). It is worth noting that the number of Gal per macromolecule is significantly higher than the values reported with linear or graft saccharidic copolymers (typically less than 2000 Gal/macromolecule) (14, 15). After regeneration of the hydroxyl functions, I and II were soluble in water while III could only be dissolved in MeOH, likely due to the presence of short PHEVE and PGVE blocks. Dynamic light scattering analysis confirmed that the unprotected dendrigrafts behave as unimeric macromolecules in water (I and II). Synthesis of the Water-Soluble Dendrigrafts IV. The second approach aimed at the preparation of water-soluble dendrigrafts with PEO-b-PEGP external branches, considering that the PEGP blocks bearing the reactive saccharidic groups would enable the binding of NH2-terminated oligonucleotides, while the PEO blocks would ensure the water solubility of the dendrigraft and improve, as spacers, the accessibility of the Gal units. The synthesis of these amphiphilic dendrigrafts was achieved by grafting living polyoxirane chains onto the activated chloroethyl functions of PCEVE external blocks of PCEVE1g-(PS1-b-PCEVE2) precursors (C: 200/120/40). Living PEGP-b-PEO- K+ diblock copolymers were first obtained by anionic polymerization (Scheme 2). SEC analysis clearly established the preparation of narrow dispersed diblock chains (PDI ) 1.14). The exact composition and molar mass of the copolymer were determined by 1H NMR and VPO analysis of PEGP-b-PEO: DPnPEGP ) 5; DPnPEO ) 90. Amphiphilic dendrigrafts IV were subsequently prepared by grafting living PEGP-b-PEO- K+ oxanionic species onto PCEVE1-g-(PS1-b-PCEVE2) precursors (C) in the presence of

NaI as activator (see Scheme 1, step 3′). After reaction, the hyperbranched polymer was separated from free PEGP-b-PEO chains by selective precipitation. The SEC analysis of the dendrigrafts (Figure 1) confirmed the unimodal and narrow molar mass distribution of IV. The main dimensional characteristics of IV are collected in Table 2. The previous characterization of C enabled us to calculate, by 1H NMR, the yield of PEGP-b-PEO grafting, the final molar mass of dendrigraft IV, and the number of Gal units per dendrigraft. The poor reactivity of the oxanionic species resulted in low efficiency of grafting (15%), giving birth, after deprotection of the hydroxyl groups, to water-soluble dendrigrafts displaying a low peripheral density and bearing about 6100 Gal units each. In contrast to the cases of dendrigrafts I and II, the dynamic light scattering analysis revealed the presence of a few aggregates (constituted of two to three dendrigrafts) in aqueous solution (see Figure 2). Characterization by Tapping-Mode Atomic Force Microscopy (TMAFM). TMAFM characterization of PHEVE- and PGVE-functionalized dendrigrafts (especially I) with a high grafting density of PVE hydrophilic blocks has been described in a previous paper (20). In this case, individual molecules can be easily observed as unique olivelike shaped objects with a diameter of about 100 nm (Figure 3, left). In this work, we mainly focused on the molecular characteristics of dendrigraft IV with PEO90-b-PEGP5 external hydrophilic blocks, which exhibits a lower peripheral grafting density (15%). The typical morphology of submonolayer deposits obtained by drop casting of 0.02 wt % MeOH solutions on mica is shown in Figure 3, right. Deposits of IV exhibit both globular and wormlike structures. The globular structures, more clearly seen in the topographic image inset (c) of Figure 3, correspond to unimeric dendrigrafts exhibiting a narrow size distribution (diameter ) 22 nm ( 2 nm), in agreement with the low polydispersities measured in solution by SEC. The much reduced size of IV measured here by AFM, compared to the DLS data of Table 2 (diameter 22 nm as solid deposit vs 60 nm in THF), most probably arises from the fact that AFM measurements are performed in the dry state while the DLS values in a good

Water-Soluble Dendrigrafts Bearing Saccharidic Moieties

Bioconjugate Chem., Vol. 17, No. 1, 2006 11

Figure 3. Tapping mode AFM (800 × 800 nm2) phase image of a monolayer of the dendrigraft IV obtained from a 0.002 wt % methanol solution deposited on mica (b). The inset (c) shows a higher resolution (400 × 400 nm2) topographic image. The left image (a) shows, as a comparison, the morphology of the dendrigraft I deposited from a 0.01 wt % THF solution on mica. Scheme 3. Preparation of Dendrigraft-ODN Capture Conjugates Using a Simple One-Step Coupling Procedure (Schiff Base) with Amino-Ended ODNs

solvent correspond to swollen molecules. This is in contrast with the close diameter observed for I in the dry state and in solution (100 nm as solid vs 110 nm in THF). The presence of wormlike structures for IV, with a length up to 300 nm and a constant width corresponding to the diameter of unimeric dendrigrafts, clearly results from unidirectional assemblies of a large number of dendrigrafts. This organization also strongly contrasts to the one observed for dendrigraft I. Such an intermolecular organization and the size reduction of IV could be interpreted as the consequence of the quite low density of peripheral branches, which allows hydrophobic (PCEVE, PS) and hydrophilic domains of several dendrigrafts to interpenetrate, whereas the much higher lateral grafting density in I makes it much more compact, avoiding swelling and interpenetration phenomena. Preparation of Dendrigraft-ODN Conjugates and ELOSA Tests. The covalent binding of DNA sequences on the dendrigraft surface was performed in aprotic polar solvent (DMSO, DMF, or NMP)/(aqueous buffer) mixtures, by reaction of the primary amine at the 5′-end of the ODNs with the aldehyde function at the anomeric position of the galactopyranose residues (formation of a Schiff base) as shown in Scheme 3. Various parameters such as the Gal/ODN molar ratio, the (organic solvent)/(aqueous buffer) volume ratio, the nature of the solvent, the ionic strength, and the pH of the aqueous buffer were tuned

to optimize the coupling reaction. Due to the size of the dendrigraft-ODN conjugates and their tendency to form aggregates in organic/aqueous mixtures, it was not possible to follow the coupling reaction by SEC as previously reported (7, 9, 10) by comparing the relative peak areas of the bound ODNs and the unreacted ODNs. Formation of aggregates during the coupling reaction had already been reported for polymer-ODN conjugates (7) and had been attributed to possible side reactions from heterocyclic base amino groups and/or hydrogen bondings between the polymer and the ODNs. Finally, the presence of covalently bound ODNs on the dendrigrafts was indirectly evaluated by a capture test. The ability of the prepared dendrigraft-ODN conjugates to capture short complementary ODNs (25 bases) through hybridization was studied on the Vidas instrument. As these complementary ODNs were labeled with an alkaline phosphatase enzyme, addition of methylumbelliferyl phosphate substrate led to the appearance of a fluorescence signal (Scheme 4A). The main results of the capture tests performed with I, II, III, and IV dendrigraft-ODN conjugates are reported in Table 3. Except III, the dendrigrafts were all suitable for the elaboration of dendrigraft-ODN capture conjugates, as revealed by the high fluorescence signal obtained under specific coupling conditions. The nature of the organic solvent, the (organic solvent)/(aqueous buffer) volume ratio, and the Gal/ODN molar ratio appeared to be the main parameters of the coupling reaction. In contrast, no influence of the buffer ionic strength or pH was observed. To estimate the proportion of ODNs immobilized on the dendrigrafts by physical adsorption, a run was carried out using an ODN without an amino spacer arm (run 6 to be compared with run 4). This test indicated that the amount of adsorbed ODNs was low (the signal was of the same order of magnitude as that of the control run performed without dendrigrafts). Although partially due to adsorption phenomena, the fluorescence signals obtained with I-ODN, II-ODN, or IV-ODN capture conjugates clearly confirmed the presence of ODNs

12 Bioconjugate Chem., Vol. 17, No. 1, 2006

Bernard et al.

Table 3. Preparation Conditions of the Dendrigraft-ODN Conjugates and Performance of the Capture Tests

a

run

dendrigraft

solvent

org/aq v/v ratio

pH of borate buffer

Gal/ODN ratio (nmol)

optimal signal (RFU) capture test

1 2 3 4 5 6 7 8 9 10 11 12b

I I II II II II III III IV IV IV

NMP DMF DMF DMF DMF DMF DMSO DMSO DMF DMF DMF

97/3 97/3 50/50 50/50 50/50 50/50 96/4 96/4 90/10 90/10 90/10

9.2 9.2 7.2 7.2 7.2 7.2 7.2 7.2 9.2 9.2 9.2

∼60 ∼60 20 100 300 100 100 100 4 10 20

1100 9400 2200 6800 8700 1900a 2600 2900a 10500 10500 10500 2600

“Coupling reaction” performed using ODNs without NH2-ended spacer arm. b Control run without dendrigrafts (labeled ODNs only).

Scheme 4. Principles of the Capture Test (A, top) and of the Complete Sandwich Hybridization Assay (ELOSA Diagnostic Test) with Capture and Detection Amplification (B, bottom)

Figure 4. Complete sandwich hybridization assays (ELOSA tests) with capture and detection amplification (I-ODN, II-ODN, and IV-ODN, runs 2, 5, and 11).

bound at the periphery of the dendrigrafts and their availibility for hybridization. In contrast, concerning III-ODN conjugates (run 7), the detected signal was attributed only to the presence of some adsorbed ODNs, since the run performed using ODNs without an amino spacer arm produced a similar fluorescence signal (run 8). This result clearly underlines the influence of the hydrophilic spacer block (PHEVE) on the ODN coupling reaction and the necessity to grow long PHEVE blocks (III: DPnPHEVE ) 11; I and II: DPnPHEVE ) 90 and 110, respectively) to improve the accessibility of the Gal moieties and consequently to favor the binding of ODNs. The conjugates synthesized from I and IV provided high fluorescence signals with moderate Gal/ODN ratios (4-60) (see runs 2, 9, 10, and 11) whereas in the case of II it was necessary to perform the coupling reactions with high Gal/ODN ratios (∼100-300) to obtain a significant signal (runs 4 and 5). Although it remains difficult to draw conclusions, this unexpected trend could be correlated to the short saccharidic block (DPn ) 4) at the periphery of dendrigraft II. In addition, in opposition to the case of I or IV, the ODN coupling reaction with II was more efficient at a higher water content (50% v/v instead of 10% v/v).

Finally, IV appeared to have the best structure to favor the binding of ODNs, since a Gal/ODN ratio of 4 was enough to give a saturation signal (run 9). In comparison with I, IV is smaller (Rh ) 30 instead of 55 nm in THF) and thus has a much higher specific area (1086 instead of 155 m2 per gram). In addition, each dendrigraft IV bears an average of 6100 Gal residues (instead of 96 000 for dendrigraft I) and thus a lower saccharidic density at the surface which favors accessibility (0.32 and 2.52 Gal/nm2 for IV and I, respectively). Consequently, the more favorable structure for the binding of ODNs seems to correspond to dendrigrafts of 60 nm diameter bearing a moderate number of Gal residues located at the end of long hydrophilic spacer arms. Complete sandwich hybridization assays (ELOSA tests) with capture and detection amplification were further performed with a HBV DNA target (∼2400 bases) (Scheme 4B). Each dendrigraft-ODN capture conjugate which had provided a high fluorescence signal in the capture test was evaluated (I-ODN, II-ODN, and IV-ODN, runs 2, 5, and 11, respectively). The different dendrigraft-ODN capture conjugates were independently compared in the presence of a detection conjugate prepared from a saccharidic graft copolymer (AF112) (14, 15). The results of the different tests are presented in Figure 4. Despite the presence of ODNs on each conjugate, the results clearly showed that the conjugates made from IV (the dendrigraft of lowest external branch density) were the only ones leading to an increase in sensitivity. I- and II-based conjugates did not provide a significant increase in sensitivity since the signal/noise ratio remained very low (e2) even at high

Bioconjugate Chem., Vol. 17, No. 1, 2006 13

Water-Soluble Dendrigrafts Bearing Saccharidic Moieties Table 4. Amplification Performances of IV-ODN Conjugates in Complete Sandwich Hybridization Assays (ELOSA Diagnostic Tests) with Capture and Detection Amplification polymer-ODN capture conjugate

signal/noise ratio (1011 DNA copies/mL)

signal/noise ratio (109 DNA copies/mL)

AF112-ODNa IV-ODN (run 9) IV-ODN (run 11) capture ODNb

13.5 7.6 10 1.4

2.9 1.7 2.5 1

a Polymer-ODN capture conjugate made from a saccharidic graft copolymer (AF112) (15) bearing an average of 650 Gal residues per copolymer. b Control run without dendrigraft-ODN capture conjugate (ODNs alone).

(1011

concentrations of DNA copies/mL). The difference of behavior between I-ODN, II-ODN, and IV-ODN conjugates was correlated with the relative steric hindrance induced by the external structure of conjugates I and II. In comparison with the case of IV-ODN, the high external density of saccharidic residues could affect the availability of the immobilized capture ODNs and/or prevent either the approach or the hybridization of the target of 2400 nucleic bases and/or prevent the approach or the hybridization of the detection conjugate. It was not possible to determine which of these steps was finally responsible for the low resulting signal. In the case of IV-ODN, the signal/noise ratio was equal to 8-10 at 1011 DNA copies per mL and was still superior to 2 (which is the detection limit) at 109 DNA copies per mL (Table 4). In comparison, the same test without dendrigrafts (with capture ODNs alone) reached the same detection limit for 1011 DNA copies per mL and thus was 100 times less sensitive. Moreover, the background signal provided by the IV-ODN conjugate was relatively low (1000 RFU, a value similar to that of linear copolymers) (9). Finally, it was interesting to compare these dendrigraft-ODN conjugates (IV-ODN) with a capture conjugate based on a saccharidic graft copolymer, AF112-ODN, bearing on average 650 ODNs per copolymer chain. In both cases, the sensitivity of the test reached 109 DNA copies per mL (Table 4), confirming the interest of the less crowded dendrigrafts bearing a moderate number of external branches and Gal residues for such diagnostic tests. In summary, two families of amphiphilic water-soluble dendrigrafts exhibiting well controlled dimensions and bearing saccharidic moieties have been synthesized. The presence of external galactopyranose groups on these polymers has been exploited to prepare dendrigraft-oligonucleotide capture conjugates via a simple one-step coupling procedure (Schiff base) using amino-ended ODNs. A capture test with short labeled complementary ODNs (25 bases) confirmed the presence of covalently bound ODNs on various kinds of dendrigrafts (I, II, and IV) except for the dendrigraft (III) displaying short hydrophilic spacer blocks, which led exclusively to some ODN adsorption. The conjugates made from I, II, and IV were finally evaluated in a nucleic acid detection test with a real DNA target from hepatitis B virus, HBV DNA with 2400 nucleic bases. The IVODN conjugate which possesses a low saccharidic external density (6120 Gal per dendrigraft) was the only capture conjugate leading to a significant amplification of the fluorescence signal, corresponding to a limit of sensitivity of 109 DNA copies per mL (instead of 1011 DNA copies per mL without using dendrigrafts) and a low background signal. Conversely, despite the large number of galactopyranose groups located at their periphery (respectively 96 000 and 33 000), I and II were not able to induce a better sensitivity of the test, probably due to the steric hindrance generated by the peripheral congestion of these polymers.

ACKNOWLEDGMENT The authors would like to thank Arnaud Favier (Unite´ Mixte CNRS-bioMe´rieux) for the sample of the saccharidic graft copolymer AF112 and Aude Dassonville (Unite´ Mixte CNRSbioMe´rieux) for some help with the ELOSA tests. We would also like to acknowledge the cooperation agreement (Tournesol) between France (LCPO) and the Communaute´ Franc¸ aise de Belgique (Mons) for financial support.

LITERATURE CITED (1) Abramson, R., and Myers, T. W. (1993) Nucleic acid amplification technologies. Curr. Opin. Biotechnol. 4, 41-47. (2) Nelson, P. S., Sherman-Gold, R., and Leon, R. (1989) A new and versatile reagent for incorporating multiple primary aliphatic amines into synthetic oligonucleotides. Nucleic Acids Res. 17, 7179-7186. (3) Haralambidis, J., Duncan, L., Angres, K., and Tregear, G. W. (1990) The synthesis of polyamide-oligonucleotide conjugate molecules. Nucleic Acids Res. 18, 493-499. (4) Grzybowski, J., Will, D. W., Randall, R. E., Smith, C. A., and Brown, T. (1993) Synthesis and antibody-mediated detection of oligonucleotides containing multiple 2,4-dinitrophenyl reporter groups. Nucleic Acids Res. 21, 1705-1712. (5) Horn, T., Werner, B. D., Running, J. A., Downing, K., Clyde, J., and Urdea, M. S. (1989) The synthesis of branched oligonucleotides as signal amplication multimers for use in nucleic acid assays. Nucleosides Nucleotides 8, 875-877. (6) Chang, C. A., Horn, T., Ahle, D., and Urdea, M. S. (1991) Improved methods for the synthesis of branched DNAb(DNA) for use as amplification multimers in bioassays. Nucleosides Nucleotides 10, 389-392. (7) Erout, M. N., Troesch, A., Pichot, C., and Cros, P. (1996) Preparation of conjugates between oligonucleotides and N-vinylpyrrolidone/N-acryloxysuccinimide copolymers and applications in nucleic acid assays to improve sensitivity. Bioconjugate Chem. 7, 568-575. (8) Ladaviere, C., Veron, L., Delair, T., Domard, A., Pichot, C., and Mandrand, B. (1997) Reactive polymers in diagnostics: synthesis and characterizations of nucleic acid probes and maleic anhydrideco-methyl vinyl ether polymers. J. Appl. Polym. Sci. 65, 25672577. (9) Delair, T., Badey, B., Charles, M. H., Laayoun, A., Domard, A., Pichot, C., and Mandrand, B. (1997) Conjugation of nucleic acid probes to 6-aminoglucose-based homo and copolymers. II. Application to diagnostics. Polym. AdV. Technol. 8, 545-555. (10) Delair, T., Charles, M. H., Cros, P., Laayoun, A., Mandrand, B., and Pichot, C. (1998) Synthesis and characterization of nucleic acid(co)polymer conjugates: application to diagnostics. Polym. AdV. Technol. 9, 349-361. (11) Ve´ron, L., de Bignicourt, M. C., Delair, T., Pichot, C., and Mandrand, B. (1996) Synthesis of poly[N-(2,2-dimethoxyethyl)-Nmethacrylamide] for the immobilization of oligonucleotides. J. Appl. Polym. Sci. 60, 235-244. (12) Badey, B., Boullanger, P., Domard, A., Cros, P., Delair, T., and Pichot, C. (1996) Radically initiated polymerization of a methacryloylamido-terminated saccharide. Part 1. Monomer synthesis, homopolymerization and characterizations. Macromol. Chem. Phys. 197, 3711-3728. (13) D’Agosto, F., Charreyre, M. T., Pichot, C., and Mandrand, B. (2002) Polymer of controlled chain length carrying hydrophilic galactose moieties for immobilization of DNA probes. Macromol. Chem. Phys. 203, 146-154. (14) D’Agosto, F., Charreyre, M. T., Melis, F., Pichot, C., and Mandrand, B. (2003) High molecular weight hydrophilic functional copolymers by free radical copolymerization of acrylamide and of N-acryloylmorpholine with N-acryloxysuccinimide: Applications to the synthesis of a graft copolymer. J. Appl. Polym. Sci. 88, 18081816. (15) Favier, A. (2002) Ph.D. Thesis, Claude Bernard University, Lyon I.

14 Bioconjugate Chem., Vol. 17, No. 1, 2006 (16) Minard-Basquin, C., Chaix, C., Pichot, C., and Mandrand, B. (2000) Oligonucleotide-polymer conjugates: Effect of the method of synthesis on their structure and performance in diagnostic assays. Bioconjugate Chem. 11, 795-804. (17) Minard-Basquin, C., Chaix, C., D’Agosto, F., Charreyre, M. T., and Pichot, C. (2004) Oligonucleotide synthesis onto poly(Nacryloylmorpholine-co-N-acryloxysuccinimide): Assessement of the resulting conjugates in a DNA sandwich hybridization. J. Appl. Polym. Sci. 92, 3784-3795. (18) Charreyre, M. T., Hiver, A., Delair, T., Cros, P., Pichot, C., Mandrand, B., Tcherkasskaya, O., and Winnik, M. A. (1997) Fluorescence energy transfer study of the conformation of oligonucleotides covalently bound to polystyrene latex particles. Langmuir 13, 3103-3110. (19) Elaissari, A., Chevalier, Y., Ganachaud, F., Delair, T., and Pichot, C. (2000) Structure of adsorbed and grafted single-stranded DNA fragments on aminated latex particles. Langmuir, 16, 1261-1269. (20) Schappacher, M., Deffieux, A., Putaux, J. L., Viville, P., and Lazzaroni, R. (2003) Synthesis and characterization of water-soluble amphiphilic polystyrene-based dendrigrafts. Macromolecules 36, 5776-5783.

Bernard et al. (21) Muchtar, Z., Schappacher, M., and Deffieux, A. (2001) Hyperbranched nanomolecules: Regular polystyrene dendrigrafts. Macromolecules 34, 7595-7600. (22) Fukui, H., Yoshihashi, S., Sawamoto, M., and Higashimura, T. (1996) Multifunctional coupling agents for living cationic polymerization. 7. Synthesis of amphiphilic tetraarmed star block polymers with R-methylstyrene and 2-hydroxyethyl vinyl ether segments by coupling reactions with tetrafunctional silyl enol ether. Macromolecules 29, 1862-1866. (23) Koell, P., Saak, W., Pohl, S., Steiner, B., and Koos, M. (1994) Preparation and crystal and molecular structure of 6-O-[(2S)-2,3epoxypropyl]-1,2:3,4-di-O-isopropylidene-galactopyranose. Pyranoid ring conformation in 1,2:3,4-di-O-isopropylidene-galactopyranose and related systems. Carbohydr. Res. 265, 237-248. (24) Schappacher, M., Bernard, J., and Deffieux, A. (2003) Functional polystyrene combs and dendrigrafts, 1- Selective introduction of dihydroxymethyl functions at their periphery. Macromol. Chem. Phys. 204, 762-769. BC049708J