Multivalent interactions between biotin–polyrotaxane conjugates and streptavidin as a model of new targeting for transporters

Journal of Controlled Release 80 (2002) 219–228 www.elsevier.com / locate / jconrel

Multivalent interactions between biotin–polyrotaxane conjugates and streptavidin as a model of new targeting for transporters Tooru Ooya, Nobuhiko Yui* School of Materials Science, Japan Advanced Institute of Science and Technology, 1 -1 Asahidai, Tatsunokuchi, Ishikawa 923 -1292, Japan Received 9 November 2001; accepted 30 January 2002

Abstract Kinetic analysis of interactions between biotin–polyrotaxane or biotin–a-cyclodextrin (biotin–a-CD) conjugates and streptavidin was carried out as a model of new targeting to transporters using the surface plasmon resonance (SPR) technique. The biotin–polyrotaxane conjugates, in which biotin-introduced a-CDs are threaded onto a poly(ethylene oxide) chain capped with bulky end-groups, are expected to increase the valency of biotin from monovalent to multivalent binding. The number of biotins conjugated with one polyrotaxane molecule varied from 11 to 78, and apparently increased the association equilibrium constant (Ka ), assuming pseudo-first-order kinetics. A detailed dissociation kinetics was analyzed and the re-binding of the biotin–polyrotaxane conjugates was observed on the streptavidin-deposited SPR surface. The magnitude of the re-binding is likely to become larger with increasing the number of biotins, suggesting multivalent interaction on the SPR surface. To quantify the effect of valency, competitive inhibition assay was performed in terms of the supramolecular structure of the polyrotaxane. The inhibitory potency of the biotin–polyrotaxane conjugate was found to be 4–5 times greater than that of the biotin–a-CD conjugate. Therefore, the biotin–polyrotaxane conjugates by supramolecular formation of the biotin–a-CD conjugate significantly switches from monovalent to multivalent bindings to the model binding protein, streptavidin.  2002 Elsevier Science B.V. All rights reserved. Keywords: Polyrotaxanes; Conjugate; Multivalent ligands; Surface plasmon resonance; Kinetics

1. Introduction Ligand–polymer or ligand–particle conjugates have been extensively studied for drug targeting [1–3], vaccines [4] and receptor-mediated gene delivery [5]. Recent advances in the conjugation of polymers with ligands made it possible to design tailor-made targeting in combination with enhanced*Corresponding author. Tel.: 181-761-511-640; fax: 181-761511-645. E-mail address: [email protected] (N. Yui).

permeation retention (EPR) effect and ligand–receptor interactions [6]. The main advantage of the ligand–polymer conjugates is that a number of ligands are possibly introduced to one macromolecule. A lot of ligand molecules can interact with surface binding proteins that widely exist on cell surfaces, resulting in polyvalent (multivalent) interactions [7]. Multivalent ligands are termed as ligands that display multiple copies of recognition elements. Omelyanenko et al. reported that several attachments of receptor binding epitopes per water-soluble macromolecule showed high binding activity to B- or

0168-3659 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved. PII: S0168-3659( 02 )00030-5

220

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

T-cells, the equilibrium association constant of which was one order of magnitude higher than a conjugate containing one epitope per macromolecule [3]. Arimoto et al. reported that a multivalent polymer of vancomycin showed enhanced antibacterial activity against vancomycin-resistant bacteria [8]. They suggested that antibiotic-introduced multivalent polymers are promising tools in the fight against multi-resistant bacteria. These facts indicate that multivalent interaction is a promising strategy for drug targeting and inhibitors. Intestinal transporters such as sodium-dependent multivitamin transporter (SMVT) and oligopeptide transporter (PEPT1) have been extensively studied for facilitating the apical transport of biotin-linked anti-human immunodeficiency virus (HIV) agents [9] and peptide-like drugs [10,11]. Several research groups have studied targeting to intestinal transporters in order to improve the oral absorption of drugs [9,12]. However, none have paid attention to drug targeting and inhibitors using multivalent polymers toward small intestines, in which a lot of transporters transport low-molecular-weight substrates such as digested proteins, peptide-mimic drugs, vitamins and saccharides. Targeting these intestinal transporters in a multivalent binding manner may be advantageous to orally delivering drugs and inhibiting the absorption of digested nutrients for patients to whom food intake is restricted. In our previous study, collaborated with Prof. Akira Tsuji of Kanazawa University, Japan, polyrotaxane–dipeptide (Val–Lys, an absorbable substrate via human peptide transporter) conjugates were synthesized and evaluated in terms of the uptake of a model dipeptide (Gly–Sar) via the human peptide transporter (hPEPT1) on HeLa cells [13]. Val–Lys groups are introduced to a-cyclodextrins (a-CDs), which are threaded onto a poly(ethylene oxide) chain capped with bulky end-groups (polyrotaxane). We demonstrated that the Gly–Sar uptake via hPEPT1 was significantly inhibited in the polyrotaxane conjugates. The inhibitory effect was not explained from the sum of interaction between hPEPT1 and a-CD–Val–Lys conjugates, suggesting an inhibitory effect via multivalent binding of Val– Lys groups with hPEPT1. The attractive characteristics of the polyrotaxanes involve the design of the bulky end-groups in the polyrotaxane as biodegradable polymers [14–17]. For example, the supramolecular structure of the

pheneylalanylglycilglycine (PheGlyGly)-terminated polyrotaxane was dissociated due to hydrolysis between Phe–Gly bonds by the action of aminopeptidase N which is expressed in intestinal and kidney brush border membranes and other mucosal surfaces [16]. Temporally-multivalent binding with the intestinal transporters using the ligand–polyrotaxane conjugates is well controlled by the dissociation to ligand–a-CD conjugates via the terminal enzymatic hydrolysis, in other words, the multivalent interaction may be switched to monovalent interaction by the dissociation, resulting in temporal binding to the transporters without altering the gastrointestinal (GI) environment (Fig. 1). Recently, Ramanathan et al. demonstrated that biotin-conjugated star-shaped poly(ethylene oxide) enhanced the uptake of biotin via SMVTs, depending on the degree of branching [18]. This report suggests that targeting SMVTs is one of the feasible approaches to non-invasively enhance the transport of large macromolecules in terms of molecular size and architecture. From the viewpoint of the biotin transporter, SMVT, biotin– polyrotaxane conjugates may be applicable to altering SMVT properties. In this study, the kinetic analysis of the interaction between ligand–polyrotaxane or ligand–a-CD conjugates and their binding proteins was performed to clarify the multivalent binding in relation to the polyrotaxane structure by means of a surface plasmon resonance (SPR) technique. Biotin and streptavidin were selected as a model ligand and a binding protein (transporter), respectively. The streptavidin-deposited SPR surface can easily be applied to the kinetic analysis, and the effect of the number of biotins per polyrotaxane molecule on the binding, dissociation and their equilibrium constants was evaluated. Competition assays were carried out to evaluate the difference in inhibition ability between the biotin–polyrotaxane and the biotin–a-CD conjugates on the interactions of streptavidin with a biotinimmobilized sensor surface.

2. Materials and methods

2.1. Materials a-CD

was

purchased

from

Bio-Research

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

221

Fig. 1. Final image of controlled-multivalent binding with the intestinal transporters using ligand–polyrotaxane conjugates for altering the transporter properties.

(Yokohama, Japan). Benzyloxycarbonyl-L-phenylalanine (Z-L-Phe), 2-aminothanol, N,N9-carbonyldiimidazole (CDI), formic acid and D-biotin were purchased from Wako (Osaka, Japan). N-Hydroxysuccinimide and 1-hydroxybenzotriazole (HOBt) were purchased from Peptide Institute (Osaka, Japan). Streptavidin from Streptomyces avidinii was purchased from Nacalai Tesque (Kyoto, Japan). Phosphate-buffered saline (pH 7.4) containing 0.05% (v / v) of Tween 20 (PBS / T) (10 mM sodium phosphate, 2.7 mM potassium chloride, 138 mM sodium chloride and 0.05% Tween 20) was prepared by the dissolution of PBS / T powder purchased from Sigma (St. Louis, MO, USA) and kept at 24 8C until use. EZ-LinkE biotin hydrazide was purchased from Pierce (Rockford, IL, USA). The biotin cuvette for the interaction analysis system (IAsys) was purchased from Affinity Sensors (Cambridge, UK). Dimethylsulfoxide (DMSO) was purchased from Wako, and distilled by the usual method. All other chemicals used were of reagent grade.

2.2. Synthesis of biotin–polyrotaxane and biotin– a -CD conjugates A polyrotaxane in which many a-CDs were threaded onto a poly(ethylene oxide) chain capped with Z-L-Phe was prepared according to our previous paper [14]. The number of a-CDs was determined to be ca. 22|30 from the 1 H-nuclear magnetic resonance (NMR) spectra. The biotin–polyrotaxane conjugates were synthesized according to our previous method [19]. Briefly, the polyrotaxane dissolved in DMSO was allowed to react with CDI to activate the hydroxyl groups of the a-CDs in the polyrotaxane. The obtained CDI-activated polyrotaxane dissolved in dry DMSO was mixed with the EZ-LinkE biotin hydrazide and HOBt under a nitrogen atmosphere, then stirred for 24 h at room temperature. 2-Aminoethanol was dropped into the reaction mixture, and stirred for additional 24 h under the same conditions. Non-reacted compounds were removed by dialysis against water (Spectra / Por  , molecular

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

222

Fig. 2. Schematic structure of the biotin–polyrotaxane conjugates.

weight cut-off 1000), and the resulting solution was lyophilized to obtain biotin–polyrotaxane conjugates (Fig. 2). The chemical structure was characterized by 750 MHz 1 H-NMR and gel permeation chromatography (GPC). Similarly, a biotin–a-CD conjugate was also synthesized. The synthetic results of the obtained conjugates are summarized in Table 1.

2.3. Binding analysis of the biotin–polyrotaxane conjugates on streptavidin-deposited surface using an SPR instrument The SPR experiments were carried out using an IAsys instrument (IAsys Auto1, Affinity Sensors) that quantifies a wide range of biomolecular interactions based on a resonant mirror biosensor [20]. The IAsys instrument temperature was set at 25 8C. The resonant layer of the biotin cuvette was washed with 40 ml of PBS / T and allowed to settle for 10 min for

equilibration. During this period, streptavidin was dissolved in PBS / T (1 mg / ml). PBS / T containing streptavidin (20 ml) was added to the PBS / T in the cuvette and allowed to settle for 10 min to facilitate binding on the biotin-immobilized surface. After washing the cuvette with 50 ml of PBS / T three times, the cuvette was left to stand for 3 min to stabilize the baseline. The density of the deposited streptavidin was calculated from the resulting sensorgram based on the IAsys calibration curve [19]. After equilibrating the streptavidin-deposited surface with 45 ml of PBS / T, 5 ml of each biotin conjugate dissolved in PBS / T (concentration: 50 nM biotin in the conjugate) was added to the PBS / T in the cuvette, then the binding was monitored for 10 min. Thereafter, the cuvette was washed with 50 ml of PBS / T, and any dissociation was monitored for additional 5 min. Finally, 1 M formic acid was added to the surface for 1 min to break the biotin–strep-

Table 1 Synthesis of the biotin conjugates for kinetic analysis Sample code a

Feed ratio ([Bio] / [Im])b

No. of biotin / mole

No. of a-CD/ mole c

No. of HEC / mole c

Total Mn

11BIO-a / E4-PHE-Z 35BIO-a / E4-PHE-Z 78BIO-a / E4-PHE-Z 1BIO-aCD

0.5 1 2 1d

11 35 78 1

20 22 22 –

104 113 188 4

33 300 43 500 60 900 1480

a

BIO-a / E4-PHE-Z and BIO-aCD denote biotin–polyrotaxane and biotin–a-CD conjugates, respectively. [Bio] and [Im] denote the concentration of the EZ-LinkE biotin hydrazide and the N-acyl imidazole groups (the activated hydroxyl groups of a-CDs in the polyrotaxane), respectively. c Calculated from 750 MHz 1 H-NMR spectra, HEC5hydroxyethylcarbamoyl groups. d The introduced N-acyl imidazole group was one per a-CD molecule. b

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

tavidin bindings, and then the cuvette was washed with PBS / T three times. The resulting sensorgrams were analyzed based on pseudo-first-order kinetics to determine the kinetic parameters (see Results and discussion).

2.4. Competition assays Competition experiments with the biotin–polyrotaxane or biotin–a-CD conjugates were carried out according to the report of Mann et al. [21]. The biotin–polyrotaxane or the biotin–a-CD conjugate (78BIO-a / E4-PHE-Z or 1BIO-aCD in Table 1, respectively) was dissolved in PBS / T to make biotin concentration of 1 mM, and then diluted to prepare the other five samples at concentrations of 0.25, 0.5, 1.0, 10, 100 mM. A 100-ml volume of streptavidincontaining PBS / T (0.1 mg / ml, 1.5 mM) was added to each sample solution (0.9 ml) and mixed well using a mixer. These solutions were incubated for 1 h at room temperature. A 5-ml volume of the resulting solution was injected into the resonant layer of the biotin cuvette that equilibrates in which 45 ml of PBS / T (10 times dilution of the sample solutions) was added in advance. Monitoring the SPR response was the same as the binding analysis using the biotin cuvette. To derive the inhibition constants (Ki ), the SPR data were analyzed with a solution competition equation using a modified rectangular hyperbolic relationship [21,22]: f 5 [I] / h[I] 1 Ki (1 1 F /Kd ) j

binding response of the streptavidin, and [SV] concentration of the streptavidin. The calculated F and Kd values were found to be 7.960.46 and 3.260.92 nM, respectively. The Ki values for 78BIO-a / E4PHE-Z and 1BIO-aCD were derived by a curve fitting of the obtained plots of f and [I] based on Eq. (1) using Microcal Origin 6.0 software.

3. Results and discussion

3.1. Synthesis of the conjugates From the results of 1 H-NMR spectroscopy and GPC, the obtained compounds were confirmed to be the biotin–polyrotaxane or biotin–a-CD conjugates, as reported in Ref. [19]. The reduced water solubility after introducing biotin, a water-soluble ligand, appeared to be due to the association of alkyl chains in biotin. To increase the water solubility, the chemical modification of a-CDs with 2-aminoethanol (hydroxyethylcarbamoylation) was also carried out, and the solubility of the polyrotaxane increased after the reaction. The number of biotin molecules per polyrotaxane molecule was varied by changing the feed ratio of the CDI-activated polyrotaxane and the EZ-LinkE biotin hydrazide (Table 1).

3.2. Effect of the number of biotin in the conjugates on binding /dissociation kinetics

(1)

where f is fractional inhibition that was calculated using equilibrium values obtained in the absence of the inhibitor (the biotin conjugates), [I] is the inhibitor concentration (biotin basis), F is the concentration of free binding sites available for the streptavidin, and Kd the dissociation constant of the streptavidin for the surface. To determine the F and Kd values, the binding response of streptavidin on the surface of the biotin cuvette (final streptavidin concentration in the cuvette: 0.1|10 mg / ml) was determined, and the response values were fitted to the following rectangular hyperbolic equation: R eq 5 R max [SV] /(Kd 1 [SV]), Kd 5 R max / 2

223

(2)

where R eq is equilibrium response, R max the maximal

In our previous study, a biotin–polyrotaxane conjugate with approximately 11 biotin molecules was found to be recognized by a streptavidin-deposited surface [19]. The streptavidin did not interact with the polyrotaxane itself (data not shown). In this study, we examined how the number of biotins in the conjugates affects the binding / dissociation kinetics, regarding the multivalency of the biotin–polyrotaxane conjugates. The SPR sensorgrams for the binding and dissociation of 11BIO-a / E4-PHE-Z, 35BIOa / E4-PHE-Z and 78BIO-a / E4-PHE-Z onto the streptavidin-deposited surface are shown in Fig. 3. The response increased when each conjugate was injected over the streptavidin-deposited surface, although there was no significant difference in the qualitative binding responses among the number of

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

224

Fig. 3. SPR binding and dissociation curves of the biotin–polyrotaxane conjugates (concentration of biotin in the conjugates: 50 nM). (? ? ?) 11BIO-a / E4-PHE-Z, (- - -) 35BIO-a / E4-PHE-Z, and (—) 78BIO-a / E4-PHE-Z. Arrow shows the time of replacing the solution with the buffer. The small graph is the dissociation curve, starting time of which is the replacement with the buffer.

biotins in the conjugates. To dissociate the biotin– polyrotaxane conjugates, the solution was replaced with a buffer (PBS / T), 4 min after the injection, and their dissociation curves were obtained (Fig. 3). The slope of the dissociation curves seemed to become smaller with increasing the number of biotins in the conjugates. These results may suggest that the number of biotins affected the dissociation rather than the binding. In order to discuss the binding and dissociation kinetics, the binding curves in Fig. 3 were analyzed in terms of pseudo-first-order kinetics, which is based on the interaction between ligands (L, in this case the biotin–polyrotaxane conjugates) and the immobilized receptor (R, in this case the streptavidin): k bind

→ LR, Ka 5 k bind /k diss L1R ← k diss

where k bind , k diss and Ka are the corresponding constants for the binding rate, the dissociation rate, and the association equilibrium, respectively. The SPR response at time t, R t , and the binding rate, dR / dt, can be used for the kinetics of the interactions:

dR / dt 5 k bind CL (R max 2 R t ) 2 k diss R t

(4a)

R t 5 R eq [1 2 exp(2k obs t)]

(4b)

k obs 5 k bind CL 1 k diss

(4c)

where CL is the concentration of the injected conjugates (in this case, the conjugated biotin bases), R max is the maximal binding response, and k obs is the pseudo-first-order rate for the binding. The k obs values of the conjugates were calculated from the binding curves that were obtained by changing the conjugate concentration. Plots of k obs as a function of the biotin concentration in the conjugate (Eq. (4c)) were well fitted to a straight line (r 2 50.987|0.998) using the linear least-square method (data not shown). Thus, k bind and k diss were calculated using Eq. (4c). Table 2 summarizes the kinetic parameters for 11BIO-a / E4-PHE-Z, 35BIO-a / E4-PHE-Z and 78BIO-a / E4-PHE-Z. The k diss values were found to drastically decrease with increasing the number of biotins in the conjugates, while the k bind values were found to be the similar order. As the results, the Ka value of 78BIO-a / E4-PHE-Z was 12|22 times larger than that of 35BIO-a / E4-PHE-Z and 11BIO-

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

225

Table 2 Apparent kinetic parameters assuming the pseudo-first-order relation of the interaction between streptavidin and biotin Ligate

k bind a (310 4 M 21 s 21 )

k diss b (310 23 s 21 )

Ka c (?10 7 M 21 )

11BIO-a / E4-PHE-Z 35BIO-a / E4-PHE-Z 78BIO-a / E4-PHE-Z

13.8 1.7 5.2

1.6 0.39 0.052

8.6 4.4 100.0

a

k bind 5the binding rate constant. k diss 5the dissociation rate constant. c Ka 5the association equilibrium constant. b

a / E4-PHE-Z. The k bind values were not proportional to the number of biotin in the polyrotaxane. This may be related to the binding thermodynamics of two multivalent entries [23]. Generally, the rate of binding is qualitatively related to thermodynamic parameters [24]. When multivalent ligands interact with receptors, the enthalpy of the multivalent interaction is unfavorable due to restricted molecular conformation [23]. Especially, if the multivalent entry is conformationally rigid and the distance between ligands became shorter, even small spatial mismatches between the ligand and its receptor will enthalpically diminish the binding. The decreased k bind values between 11BIO-a / E4-PHE-Z and 35BIO-a / E4-PHE-Z may be due to the spatial mismatches in relation to the difference in the distance between biotins in one conjugate molecule. With further increase of the biotin numbers from 35 to 78, the water solubility slightly decreased (data not shown). So, the decreased solubility might enhance hydrophobic interactions, which lead to an increase in the entropy of the surrounding water [25]. It is known that high-affinity multivalent interactions are mostly due to a decrease in the rate of dissociation of multivalent ligands from multivalent receptors [26]. Breaking energy is likely to be required when two multivalent entries dissociate. Therefore, decreasing the k diss values suggest that biotins in the conjugates bind with the deposited streptavidins in a multivalent manner. However, Kalinin et al. suggested that the pseudo-first-order kinetics is not suitable for the analysis of multivalent interactions [22]. Thus, one question arises: whether the dissociation kinetics follow the pseudo-first-order relation or not. In order to evaluate the dissociation kinetics, the

classical expression was considered for the dissociation based on the pseudo-first-order relation. Since the buffer containing the conjugates in the SPR cuvette was replaced by the buffer itself, CL in Eqs. (4a)–(4c) should be zero for the dissociation process. Thus, the equation for the dissociation kinetics can be expressed as: R t /R 0 5 exp(2k diss t)

(5a)

ln(R t /R 0 ) 5 2 k diss t

(5b)

where R 0 is the magnitude of SPR response at the beginning of the buffer injection. A linear time dependence of ln(R t /R 0 ) is expected when the dissociation kinetics follow Eqs. (5a) and (5b). Fig. 4 shows time dependence of ln(R t /R 0 ) for 11BIO-a / E4-PHE-Z, 35BIO-a / E4-PHE-Z and 78BIO-a / E4PHE-Z. The experimental plots, collected from the data of the SPR sensorgrams, did not conform to the linear form predicted by the logarithmic form of Eq. (5b). With increasing the number of biotins in the conjugates, all slopes became gradual after 0.6|0.8 min. This result leads to justifiable consideration of the re-binding the biotin–polyrotaxane conjugates on the streptavidin-deposited surface, which strongly suggests the multivalent property. The lines without any dots in Fig. 4 denoted the theoretical dependence when the k diss values in Table 2 were applied to Eq. (5b). The lines did not correspond with the experimental plots. However, the slopes of the lines seemed to conform to those of the experimental plots after the re-binding. Therefore, the calculated k diss values in Table 2 may be apparent dissociation constants that are likely to express the multivalent kinetics.

226

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

(78BIO-a / E4-PHE-Z) and biotin–a-CD (1BIOaCD) conjugates (Fig. 5). R eq values were 1000– 1200 arc seconds in the absence of the conjugates and decreased with increasing concentration of both conjugates (from 0 to 10 mM as biotin basis). At a lower concentration region (0.025|0.1 mM), R eq values for 78BIO-a / E4-PHE-Z (Fig. 5a) were relatively smaller than those for 1BIO-aCD (Fig. 5b), suggesting that the binding ability of 78BIO-a / E4PHE-Z with streptavidin in solution was superior to that of 1BIO-aCD. The Ki values for the conjugate inhibition of streptavidin binding to the biotin-immobilized surface were calculated by generating a plot of the fractional inhibition values versus the conjugate concentration (Fig. 6) using Eq. (1). The Ki values for 78BIO-a / E4-PHE-Z and 1BIO-aCD Fig. 4. First-order plots of the dissociation kinetics for the interaction of streptavidin with the biotin–polyrotaxane conjugates using the dissociation curves in Fig. 3. (j) 11BIO-a / E4-PHE-Z, (m) 35BIO-a / E4-PHE-Z and (d) 78BIO-a / E4-PHE-Z. The solid and broken lines denote the theoretical dependence (Eq. (5b)) using the k diss values in Table 2. (? ? ?) 11BIO-a / E4-PHE-Z, (- - -) 35BIO-a / E4-PHE-Z, and (—) 78BIO-a / E4-PHE-Z.

3.3. Competitive inhibition of streptavidin–biotin binding by the multivalent inhibitor ( biotin– polyrotaxane) and the monovalent inhibitor ( biotin– a -CD) Firstly, a biotin–a-CD conjugate (1BIO-aCD in Table 1) was subjected to the binding analysis onto the streptavidin-deposited surface to compare the kinetics of the biotin–polyrotaxane conjugates with the biotin–a-CD conjugate. Unfortunately, significant SPR sensorgrams of 1BIO-aCD could not be obtained due to their lower molecular weight. With current SPR technology, it is difficult to detect the interaction between the low-molecular-weight ligands (Mn , |5000) and the immobilized receptor, since increasing mass on the sensor surface is too small to change the refractive index [27]. Alternatively, we carried out the competitive inhibition assay which is a quantifying method for substrates for inhibiting the interactions of a soluble receptor with an immobilized ligand [7,21,28]. Inhibition curves were obtained by measuring the binding responses for 0.015 mM streptavidin with increasing concentration of the biotin–polyrotaxane

Fig. 5. Inhibition curves of 0.015 mM streptavidin binding to a biotin-immobilized sensor surface for 0, 0.025, 0.05, 0.1, 1 and 10 mM biotin in the conjugates. (a) 78BIO-a / E4-PHE-Z and (b) 1BIO-aCD.

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228

Fig. 6. Fractional inhibition curves for (d) 78BIO-a / E4-PHE-Z and (m) 1BIO-aCD inhibitions of streptavidin binding.

were 2.1360.25 and 9.4861.08 nM, respectively. This result suggests that the biotin–polyrotaxane conjugate is 4–5 times more active than the biotin– a-CD conjugate. Streptavidin is known to form a tetramer that has four binding sites, and its size is assumed to be 5.5 nm [29]. The length of the polyrotaxane rod was theoretically assumed to be 32 nm, since the depth of a-CD is 0.7 nm and the stoichiometric number of a-CDs threaded onto a PEO chain (Mn : 4000) is ca. 45 [14]. Considering the threading number of a-CDs in 78BIO-a / E4-PHE-Z (ca. 22), the biotin molecules of the population can span the two binding sites of the streptavidin tetramer, which may lead to noncovalent cross-linking in the streptavidin tetramer. On the other hand, 1BIOaCD cannot span the whole binding site. Therefore, the enhanced inhibitory potency of the biotin–polyrotaxane conjugate indicates a multivalent binding effect due to an array of biotin-conjugated a-CDs, the cavity of which is threaded onto the PEO chain (polyrotaxane structure).

4. Conclusion The binding / dissociation kinetics between the biotin–polyrotaxane conjugates and streptavidin using the SPR technique was performed as a model of multivalent ligand targeting to transporters. Increasing the number of biotins conjugated with one polyrotaxane molecule decreased the dissociation

227

rate constant (k diss ) rather than increasing the binding rate constant (k bind ), assuming pseudo-first-order kinetics. The dissociation process was found not to follow the pseudo-first-order kinetics, but the rebinding of the biotin–polyrotaxane conjugates on the streptavidin-deposited SPR surface was observed. From the results of competitive inhibition assay, the inhibitory potency of the biotin–polyrotaxane conjugate was greater than that of the biotin–a-CD conjugate. Therefore, the supramolecular formation of the biotin-introduced a-CD as a biotin–polyrotaxane conjugate significantly switches monovalent biotin binding to multivalent binding with streptavidin. These model conjugates can be applicable to targeting of ligand–polymer conjugates to intestinal transporters and temporal inhibiting or altering the transporter properties in conjunction with biodegradable dissociation of the polyrotaxanes, as illustrated in Fig. 1.

Acknowledgements This study was financially supported by Grants-inAid from Scientific Research on Priority Areas ‘‘Molecular Synchronization for Construction of New Materials System’’ (No. 404 / 11167238), The Ministry of Education, Science, Sports, and Culture, Japan, Mitsubishi Foundation, Japan, Micromachine Center, Japan, and the Science and Engineering Foundation, Japan.

References ´ T. Minko, Z.-R. Lu, C.M. [1] J. Kopecek, P. Kopeckova, Peterson, Water soluble polymers in tumor targeted delivery, J. Control. Release 74 (2001) 147–158. ´ J. Kopecek, Polymerizable Fab9 [2] Z.-R. Lu, P. Kopeckova, antibody fragments for targeting of anticancer drugs, Nat. Biotechnol. 17 (1999) 1101–1104. ´ R.K. Prakash, C.D. Ebert, J. [3] V. Omelyanenko, P. Kopeckova, Kopecek, Biorecognition of HPMA copolymer–adriamycin conjugates by lymphocytes mediated by synthetic receptor binding epitopes, Pharm. Res. 16 (1999) 1010–1019. [4] N. Hussain, V. Jaitley, A.T. Florence, Recent advances in the understanding of uptake of microparticulates across the gastrointestinal lymphatics, Adv. Drug Deliv. Rev. 50 (2001) 107–142. [5] D. Larocca, A. Baird, Receptor-mediated gene transfer by

228

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13]

[14]

[15]

[16]

[17]

T. Ooya, N. Yui / Journal of Controlled Release 80 (2002) 219 – 228 phage-display vectors: applications in functional genomics and gene therapy, Drug Discov. Today 6 (2001) 793–801. ´ M. Jelınkova, ´ ´ J. Strohalm, V. Ubr, D. Plocova, ´ O. B. ´Ihova, ´ D. Plundrova, ´ Y. Germano, K. Ulbrich, Hovorka, M. Novak, Polymeric drugs based on conjugates of synthetic and natural macromolecules. II. Anti-cancer activity of antibody or (Fab9) 2 -targeted conjugates and combined therapy with immunomodulators, J. Control. Release 64 (2000) 241–261. M. Mammen, S.-K. Choi, G.M. Whitesides, Polyvalent interactions in biological systems: implications for design and use of multivalent ligands and inhibitors, Angew. Chem. Int. Ed. 37 (1998) 2754–2794. H. Arimoto, K. Nishimura, T. Kinumi, I. Hayakawa, D. Uemura, Multi-valent polymer of vancomycin: enhanced antibacterial activity against VRE, Chem. Commun. 1999, 1361–1362. S. Ramanthaan, S. Pooya, A. Stein, P.D. Prasad, J. Wang, M.J. Leibowitz, V. Ganapathy, P.J. Sinko, Targeting the sodium-dependent multivitamin transporter (SMVT) for improving the oral absorption properties of a retro-inverso Tat nonapeptide, Pharm. Res. 18 (2001) 950–956. P.V. Balimane, I. Tamai, A. Guo, T. Nakanishi, H. Kitada, F.H. Leibach, A. Tsuji, P.J. Sinko, Direct evidence for a peptide transporter (PepT1)-mediated uptake of a nonapeptide prodrug valacyclovir, Biochem. Biophys. Res. Commun. 250 (1998) 246–251. P.J. Sinko, G.L. Amidon, Characterization of the oral absorption of beta-lactam antibiotics. II. Competitive absorption and peptide carrier specificity, J. Pharm. Sci. 78 (1989) 723–727. I. Tamai, A. Tsuji, Carrier-mediated approaches for oral drug delivery, Adv. Drug Deliv. Rev. 20 (1996) 5–32. N. Yui, T. Ooya, T. Kawashima, Y. Saito, I. Tamai, Y. Sai, A. Tsuji, Inhibitory effect of supramolecular polyrotaxane–dipeptide conjugates on digested peptide uptake via intestinal human peptide transporter, Bioconjugate Chem., submitted for publication. T. Ooya, N. Yui, Synthesis and characterization of biodegradable polyrotaxane as a novel supramolecular-structured drug carrier, J. Biomater. Sci. Polym. Ed. 8 (1997) 437–456. T. Ooya, K. Arizono, N. Yui, Synthesis and characterization of an oligopeptide-terminated polyrotaxane as a drug carrier, Polym. Adv. Technol. 11 (2000) 642–651. T. Ooya, M. Eguchi, N. Yui, Enhanced accessibility of peptide substrate toward a membrane-bound metalloexopeptidase by supramolecular structure of polyrotaxane, Biomacromolecules 2 (2001) 200–203. T. Ichi, J. Watanabe, T. Ooya, N. Yui, Controllable erosion time and profile in poly(ethylene glycol) hydrogels by

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

supramolecular structure of hydrolyzable polyrotaxane, Biomacromolecules 2 (2001) 204–210. S. Ramanathan, S. Stein, M. Leibowitz, P.J. Sinko, Targeting the intestinal biotin transporter to enhance the permeability of peptides and their biopolymeric conjugates, in: Proceedings of International Symposium on Controlled Release of Bioactive Materials. (Published by the Controlled Release Society), Vol. 29, 2001, No. 678. T. Ooya, T. Kawashima, N. Yui, Synthesis of polyrotaxane– biotin conjugates and surface plasmon resonance analysis of streptavidin recognition, Biotechnol. Bioprocess Eng. 6 (2001) 293–300. R.J. Green, R.A. Frazier, K.M. Shakesheff, M.C. Davies, C.J. Roberts, S.J.B. Tendler, Surface plasmon resonance analysis of dynamic biological interactions with biomaterials, Biomaterials 21 (2000) 1823–1835. D.A. Mann, M. Kanai, D.J. Maly, L.L. Kiessling, Probing low affinity and multivalent interactions with surface plasmon resonance: ligand for concanavarin A, J. Am. Chem. Soc. 120 (1998) 10575–10582. L. Kalinin, L.D. Ward, D.J. Winzor, Effects of solute multivalence on the evaluation of binding constants by biosensor technology: studies with concanavalin A and interleukin-6 as partitioning protein, Anal. Biochem. 228 (1995) 142–151. M. Mammen, S. Choi, G.M. Whitesides, Polyvalent interactions in biological systems: implications for design and use of multivalent ligands and inhibitors, Angew. Chem. Int. Ed. 37 (1998) 2754–2794. N. Agmon, R.D. Levine, Energy entropy and the reaction coordinate: thermodynamic-like relations in chemical kinetics, Chem. Phys. Lett. 52 (1977) 197–201. W. Blokzijl, J.B.F.N. Engberts, Hydrophobic effects. Opinions and facts, Angew. Chem. Int. Ed. Engl. 32 (1993) 1545–1579. F. Karush, The affinity of antibody: range, variability and the role of multivalence, in: G. Litman, R.A. Good (Eds.), Comprehensive Immunology: Immunogloblins, Vol. 5, Plenum Press, New York, 1978, pp. 85–116. R. Karlsson, Real-time competitive kinetic analysis of interactions between low-molecular-weight ligands in solution and surface-immobilized receptors, Anal. Biochem. 221 (1994) 142–151. G.B. Sigal, M. Mammen, G. Dahmann, G.M. Whitesides, Polyacrylamides bearing pendant a-sialoside groups strongly inhibit agglutination of erythrocytes by influenza virus: the strong inhibition reflects enhanced binding through cooperative polyvalent interactions, J. Am. Chem. Soc. 118 (1996) 3789–3800. N.M. Green, Avidin, Adv. Protein Chem. 29 (1975) 25–133.