- Email: [email protected]
SPR analysis of the total reduction of protein adsorption to surfaces coated with mixtures of long- and short-chain polyethylene oxide block copolymers
Biomaterials 20 (1999) 885 — 890
SPR analysis of the total reduction of protein adsorption to surfaces coated with mixtures of long- and short-chain polyethylene oxide block copolymers Karl D. Pavey *, Cedric J. Olliff Novel Methods Group, CASS, SmithKline Beecham Pharmaceuticals, NFSP (North), Third Avenue, Harlow, Essex, CM19 5AW, UK School of Pharmacy and Biomoloecular Sciences, University of Brighton, Lewes Road, Brighton, Sussex, BN2 4GJ, UK Received 24 August 1998; accepted 28 November 1998
Abstract PEO/PPO/PEO triblock copolymers have previously been shown to reduce the binding of proteins to a variety of surfaces. In this study, mixtures of long- and short-chain copolymers have been shown to adhere to gold substrate surface plasmon resonance slides. The mixtures have been shown to significantly reduce the binding of BSA to gold surfaces, compared to the more commonly used long chain PEO copolymers. These mixtures have been shown to be more effective, than either short, or long-chain copolymers used individually, complementing a published theoretical treatise of PEO surfactant behaviour towards protein interaction with surfaces. 1999 Elsevier Science Ltd. All rights reserved Keywords: SPR; PEO/PPO triblock mixtures; Protein adsorption
1. Introduction The adsorption of poly(ethylene oxide) —poly(propyL lene oxide) —poly(ethylene oxide) triblock copolymers K L onto solid surfaces forms a surfactant layer capable of resisting nonspecific protein adsorption, cell adhesion [1] and bacterial attachment [2]. It has been established that the hydrophobic interaction of the PPO chains with surfaces provides a solid support for the polymer, whilst the PEO chains extend into solution, forming either a ‘loop and chain’ [3] or brush-like architecture [4]. It has been shown that adsorbed layers of a single poloxamer can reduce the non-specific adsorption of proteins such as albumin and fibrinogen by up to 95%, when compared to untreated surfaces [1]. The hydrated PEO layers are flexible and decrease surface interaction via a process of steric repulsion between the hydrophilic PEO blocks in solution and the protein molecules [5]. The efficiency of this system is highly dependent upon the chain length of the PEO chains.
* Corresponding author. Tel.:#44 (0)1279 627895; fax: #44 (0)1279 627896; e-mail: karl—d—[email protected]
The reduction in protein deposition, to surfaces treated with a variety of poloxamers has been monitored using a wide range of techniques including radiolabelling [1] contact angle [6], photon correlation spectroscopy [7], sedimentation [7] and ultraviolet spectroscopy [8]. A more recent introduction to the field has been the application of surface plasmon resonance [4]. This is an evanescent wave technique, which allows the real-time monitoring of interactions with surfaces. The fundamental details of SPR operation are discussed in an excellent review by Hutchinson [9]. SPR is based upon total internal reflection (TIR). TIR occurs when light beyond the critical angle and incident on an interface of higher to lower refractive index is totally reflected. An electromagnetic component, the evanescent wave is created at the junction between incident laser light and the underside of a glass prism. This moves away from the interface in to the medium of lower refractive index. If the evanescent wave is allowed to contact a thin transition metal film as is the case in SPR, free electrons in the metal film couple with the wave and resonate with certain angles of incident light. The evanescent wave penetrates the lower index medium decaying
0142-9612/99/$ — see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 2 4 5 - 2
886
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
exponentially to a distance of around 600 nm. The electron clouds (plasmons) have a collective wave parallel to the interface. The plasmons absorb energy from the incident light and a characteristic drop in reflected light intensity is observed. The evanescent wave profile is dependent upon the thickness and refractive index of the layer above the metal film. Changes in this layer requires a new angle of incident light to set-up resonance and it is the change in the surface plasmon resonance angle with time that is measured using a photo diode detector, along with the depth and shape of the ‘dip angle’. SPR has the advantage over many other methods [10], in that it is a real-time technique capable of monitoring not only the adsorption of compounds to transition metal surfaces [11], but also subsequent binding to these layers [12]. It is therefore an ideal technique for monitoring the adsorption of PEO surfactants from solution and screening with protein solutions to assess the effectiveness of the surfactant layer to reduce protein adsorption. It has been suggested that the single PEO chain length copolymers adopted thus far may not offer the most effective surface required to block protein interaction and that a surface comprising of a mixture of both long and short PEO chains would produce greater steric repulsion properties [13]. The objective of this preliminary study was to investigate the use of mixtures of PEO/PPO pluronic and tetronic clock copolymers in reducing non-specific protein adhesion on gold surfaces. A range of both long chain, short chain and mixtures of solutions of the PEO surfactants were passed over gold SPR slides via a flow cell. Surfaces were then flushed with buffer solution to remove unbound material and followed by a solution of bovine serum albumin.
2. Materials and methods Surface plasmon resonance studies were carried out on a system utilising the Kretchmann configuration (Johnson and Johnson Orthoclinical diagnostics Ltd, Chal-
font St. Giles, UK), the details and operation of this instrument have previously been described elsewhere [10, 14]. Gold SPR slides (48 nm), were obtained from the same source, being optically matched to the prism involved using Resolve, a low viscosity, microscope matching oil, (Stephens Scientific, Riverdale, NJ, USA). Three flow channels were formed on top of the slide using a clamped, thermally regulated block with a custom butyl ‘O’-ringtype arrangement. Liquid flow was achieved using an on-board syringe driver holding 1.0 ml disposable plastic syringes and driving 0.95 ml of fluid at rates between 1 and 95 ll s\. All experiments related to this work were carried out at least in triplicate, with a solution flow rate of 4 ll s\ and a maintained temperature of 30$0.1°C. Copolymers as described in Table 1 were obtained from BASF, (Parsippany, New Jersey, USA) and used without further purification. 4%w/v solutions of single polymers and 50 : 50 mixtures, to give 4%w/v, were prepared in 0.2 lm filtered phosphate buffer, (10 mM, pH 7.2). Solutions of bovine serum albumin, (Sigma, Poole, Dorset, UK), were prepared at a concentration of 1 mg ml\ using the same buffer. When in use all solutions were kept at 30°C in a water bath and immediately transferred to the SPR instrument to avoid refractive index shifts due to temperature change. Gold surfaces were treated first with buffer, then with either samples of triblock copolymers of known structure, or equal mixtures of the long and short chains, to a total concentration of 4%w/v [13]. The slides received a wash with buffer at the same flow rate to remove any excess, non-adsorbed material, followed by solutions of bovine serum albumin at a fixed concentration of 1 mg ml\ and a further buffer wash. Changes in SPR angle, expressed as milli-degree angle change (MDA), were recorded over 5 min periods with a gate time of 1 s, for both the protein adsorption and final wash steps. Custom written WINspr V1.1 software, (Johnson and Johnson Clinical Orthodiagnostics, Chalfont St. Giles, UK) was employed for the data acquisition step. Data analysis was carried out using WBplot software.
Table 1 Properties of block copolymers [15] Block copolymer
Av. mol. wt
EO
Amt molecular blocks (mol) PO
EO
Cloud point (1% aqueous soln.) (°C)
Cloud point (10% aqueous soln.) (°C)
L61 L64 F68 F127 Tetronic T908
2000 2900 8400 12 600 26 600
3 13 76 98 122
30 30 30 67 22
3 13 76 98 122
'24 58 '100 '100 '100
17 60 '100 '100 '100
34 10.5 !26.2 6.9 7.8 6.1 5.83 0.8 2.59 !7 !4.4 (!3.5) 2 1.5 34.4 36.5 2.1 (!1.9) 0.3 0.9 11.8 !11.7 0.1 0.07
Negative residual shifts for L64#F68 and L64#T908 may indicate a small amount of surfactant desorption. N"number of experiments, SD"standard deviation of data set, mda"milli degree angle.
4.6 1.2 48.5 !12.3 36.2 28.5 5.4 1.2 52.3 !11 41.3 32.5 6.2 1 51 !11.4 39.6 31.2 7 2.9 48 !11.6 36.4 28.6 2.4 1.1
SD N"5 N"3 SD N"3
SD
N"3
SD
N"5
SD
N"5
SD
N"5
SD
N"5
SD
N"5
SD
L64#L61 L64#T908 L64#F68 L64#F127 L64 F68 F127 T908#F127
49.9 !14.5 35.4 27.8 Protein (mda) Protein adsorption Protein desorption Residual protein Residual protein as % of protein to gold shift
mda—milli degree angle (1°"1000 mda). SD—standard deviation. N—number of experiments.
887
T908
SPR results are most commonly examined as timeresolved traces, however, further information about the system can be obtained directly from the reflectivity vs. angle curves. Both methods of analysis are employed here. The 1 mg ml\ solution of bovine serum albumin, (BSA), in pH 7.2 buffer absorbs strongly to a plain gold SPR slide surface, with a 127 mda, (2.9 SD, N"6), increase in angle and an initial slope of 3.2 mda s\, (0.3 SD, N"4). An immediate, subsequent, buffer wash led to a small, desorption at !0.6 mda s\, (0.1 SD, N"6), of 12.4 mda, (1.1 SD, N"6), resulting in an overall shift due to BSA adsorption of 115 mda, (4.2 SD, N"6). These values are within the range observed by other groups [14]. In the case of the surfactants solutions the adsorption profiles gave readings ranging between 370 and 420 mda with F68 giving 520 mda, except for the solution containing L64#L61, which gave increasing values in excess of 1900 mda. These shifts between 370 and 420 mda are of the order expected from the refractive indices of the solutions, compared with changes of mda previously derived from methanol and ethanol mixtures having an 8.6;10 mda/unit change with refractive index. The value for L64#L61 mixture does not correspond with its refractive index as measured by an Abbe´ refractometer i.e.1.3383. It does agree however, with the results obtained on a polystyrene coated silver slide by Green et al. (1997) and could be accounted for by the fact that the solution was turbid, due to the experimental temperature being close to the cloud point of the solution. The refractive indices of L61 and L64 are 1.4518 and 1.4570, respectively, these values indicate that a thick layer of polymer is adsorbed to the surface rather than just a monolayer. All other surfactant solutions were clear and at the experimental temperatures, well below the recorded cloud points (Table 1). Flushing of the surface with buffer removed the surfactants giving an overall shift due to adsorption of between 5 and 40 mda. These surfaces were stable with respect to buffer flow at 4 ll s\ for several hours after which time recording was stopped. The changes in resonance angle due to protein adsorption to the surfactant treated slides are given in Table 2. It can be seen that on all surfaces treated with either a single surfactant or a mixture, there was a degree of protein adsorption evident, observed as increases in resonance angle of the order of 24—32.5 mda, except for the
Surfactant C
3. Results and discussion
Table 2 Surface plasmon resonance results for poloxamer adsorption and protein interaction. Conditions: 30°C, 4 ll s\, 4% wt\ total surfactant, mixtures"1 : 1 ratio in pH 7.2, 0.2 lm filtered PBS buffer, 1 mg ml\ BSA
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
888
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
surfaces modified with the mixtures L64#F127, L64# F68 and L64#T908, all recording much lower shifts. As the mda adsorption shift is due to two factors, that is the difference in refractive index between the solution of BSA and buffer, plus changes in refractive index at the surface due to adsorption of BSA, then the results (Table 2), indicate insignificant adsorption of protein for the L64#F127, L64#F68 and L64#T908 treated gold slides and a reduced adsorption for all the other copolymer solutions used compared with the untreated gold slides (Fig. 1). All samples recorded drop in angle following a buffer wash, however the most significant feature was the final mda value after the wash. All long chain-surfactants and the L64, recorded a residual protein deposit on the surface at values between 30.1% and 36% of the residual shift of the response due to protein adsorbing to an untreated gold surface. However, for all the mixtures of long and short-chain surfactants used, the residual protein values have been significantly reduced even further (P"0.05). It is interesting to note that although results for L64#F68 and L64#L61 both indicate a certain level of adsorption of protein before the buffer wash, it cannot be strongly bound as it is removed readily on buffer washing, there being no significant difference at the 95% confidence level for the adsorption and desorption readings. However for all the others the decrease on buffer washing is 12$2.5 mda, the same as that for BSA
adsorption on to untreated gold slides, this value is most likely due to the difference in refractive index between buffer and BSA solution. The increased reduction of protein adhesion caused by the use of a mixture of long and short-chain surfactants is of both practical and theoretical significance. These data support the suggestion by Stenkamp and Berg [13] that effective steric stabilisation should be achieved by a polydisperse mixture. Thus, according to this theory, the mix of long and short chains at the surface should produce more effective steric hindrance to the protein, as the longer chains would be less inhibited in their movement and the shorter chain PEOs would be interdigitated close to the surface. Factors such as the ratio of the two polymers of different chain lengths, which may not be adsorbing in the same ratio as the solution composition, will be the subject of further investigation. Figures 2 and 3 show the SPR curves, as the initial and final scans during the adsorption of a 1 mg ml\ BSA solution, for both an untreated gold surface and one treated with a mixture of L64#F127. It can be seen that a large shift in minima position occurs with the application of BSA to an untreated gold surface. When the gold surface has been previously exposed to a mixture of long and short chain PEOs, (L64#F127), there is no discernible shift in the minimum or broadening of the SPR curve upon exposure of the surface to the protein solution,
Fig. 1. SPR angle change vs. time for the cumulative addition of a 50 : 50 mixture of L64 : 127 and a subsequent buffer wash, followed by a 1 mg ml\ BSA solution and further buffer wash (flow rate 4 ll s\, pH 7.2, ¹"30°C).
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
889
Fig. 2. Plot of average initial (pre BSA) and final (post BSA) scans for adsorption of 1 mg ml\ BSA to an untreated gold SPR slide (flow rate"4 ll s\, pH 7.2, ¹"30°C, N"3).
Fig. 3. Plot of average initial (pre BSA) and final (post BSA) scans for adsorption of 1 mg ml\ BSA to a gold SPR slide previously treated with a 4%w/v 50 : 50 mixture of L64 :127 slide (flow rate"4 ll s\, pH 7.2, ¹"30°C, N"3).
890
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
supporting the evidence that little if any protein is adsorbed. 4. Conclusions It has been shown that bovine serum albumin adsorbs, quickly and strongly, to untreated gold surfaces. The self-adsorption of PEO block copolymers of a range of molecular weight reduces protein adsorption to gold surfaces. Furthermore, if mixtures of long- and shortchain polaxamers are employed, protein adsorption can be further reduced. The specific mixtures of poloxamers L64#F127, L64#T908, L64#F68 were shown to be capable of reducing or inhibiting the binding of BSA to a gold surface to levels not discernible in this study. The mixture L64#L61 was shown to be equally good at resisting protein adsorption, despite the adsorption values being affected by the turbidity of the solution at the experimental temperature.
References [1] Amiji M, Park K. Prevention of protein adsorption and platelet adhesion on surfaces by PEO/PPO/PEO triblock copolymers. Biomater 1992;13(10):682—92. [2] Portoles M, Refojo MF, Leong FL. Poloxamer 407 as a bacterial adhesive for hydrogel contact lenses. J Biomed Mater Res 1994;28(3):303—9. [3] Lee JH, Andrade JD. Surface properties of aqueous PEO/PEO block copolymer surfactants. In: Andrade JD, editor. Polymer surface dynamics. New York: Plenum Press, 1988:119.
[4] Green RJ, Tasker S, Davies J, Davies MC, Roberts CJ, Tendler SJB. Adsorption of PEO—PPO—PEO triblock copolymers at the solid/liquid interface: a surface plasmon resonance study. Langmiur 1997;13:6510—5. [5] Lee J, Martic PA, Tan JS. Protein adsorption on pluronic copolymer-coated polystyrene particles. J Colloid Interfacial Sci 1989;131(1):252—66. [6] Freij-Larsson C, Nylander T, Jannasch P, Wesslen B. Adsorption behaviour of amphiphilic polymers at hydrophobic surfaces: effects on protein adsorption. Biomater 1996;17(22):2199—207. [7] Wu DT, Yokoyama A. An experimental study on the effect of adsorbing and non-adsorbing block sizes on diblock copolymer adsorption. Polym J 1991;23(5):709—14. [8] Carthew D, Buckton G, Parsons GE, Poole S. The influence of a phase transition in poly(oxyethylene)/poly(oxypropylene)/ poly(oxyethylene) surfactants on adsorption behaviour from dilute aqueous solution. Pharm Sci 1995;1:3—5. [9] Hutchinson AM. Evanescent wave biosensors: real time analysis of biomolecular interactions. Mol Biotech 1995;3:47—54. [10] Davies J. In: Davies J, editor. Chemistry and physics of surfaces and interfaces. Surface analytical techniques for probing biomaterial processes. New York: CRC Press, 1996:67. [11] DeBono RF, Loucks GD, Della Manna D, Krull UJ. Self-assembly of short and long chain-alkyl thiols onto gold surfaces: a real time study using surface plasmon resonance techniques. Can J Chem 1996;74:667—88. [12] Liedberg B, Nylander C, Lundstro¨m I. Biosensing with surface plasmon resonance: how it all started? Biosens Bioelectron 1995;10(8):1—6. [13] Stenkamp VS, Berg JC. The role of long tails in steric stabilization and hydrodynamic layer thickness. Langmuir 1997;13:3827—32. [14] Green RJ, Davies J, Davies MC. Roberts GJ, Tendler SJB. Surface plasmon resonance for real time in situ analysis of protein adsorption to polymer surfaces. Biomater 1997;18(5):405—13. [15] BASF performance chemicals technical literature. BASF Corporation, 1998 (http://www.basf.com/businesses/chemicals/performance/html/tetronic.html).
SPR analysis of the total reduction of protein adsorption to surfaces coated with mixtures of long- and short-chain polyethylene oxide block copolymers Karl D. Pavey *, Cedric J. Olliff Novel Methods Group, CASS, SmithKline Beecham Pharmaceuticals, NFSP (North), Third Avenue, Harlow, Essex, CM19 5AW, UK School of Pharmacy and Biomoloecular Sciences, University of Brighton, Lewes Road, Brighton, Sussex, BN2 4GJ, UK Received 24 August 1998; accepted 28 November 1998
Abstract PEO/PPO/PEO triblock copolymers have previously been shown to reduce the binding of proteins to a variety of surfaces. In this study, mixtures of long- and short-chain copolymers have been shown to adhere to gold substrate surface plasmon resonance slides. The mixtures have been shown to significantly reduce the binding of BSA to gold surfaces, compared to the more commonly used long chain PEO copolymers. These mixtures have been shown to be more effective, than either short, or long-chain copolymers used individually, complementing a published theoretical treatise of PEO surfactant behaviour towards protein interaction with surfaces. 1999 Elsevier Science Ltd. All rights reserved Keywords: SPR; PEO/PPO triblock mixtures; Protein adsorption
1. Introduction The adsorption of poly(ethylene oxide) —poly(propyL lene oxide) —poly(ethylene oxide) triblock copolymers K L onto solid surfaces forms a surfactant layer capable of resisting nonspecific protein adsorption, cell adhesion [1] and bacterial attachment [2]. It has been established that the hydrophobic interaction of the PPO chains with surfaces provides a solid support for the polymer, whilst the PEO chains extend into solution, forming either a ‘loop and chain’ [3] or brush-like architecture [4]. It has been shown that adsorbed layers of a single poloxamer can reduce the non-specific adsorption of proteins such as albumin and fibrinogen by up to 95%, when compared to untreated surfaces [1]. The hydrated PEO layers are flexible and decrease surface interaction via a process of steric repulsion between the hydrophilic PEO blocks in solution and the protein molecules [5]. The efficiency of this system is highly dependent upon the chain length of the PEO chains.
* Corresponding author. Tel.:#44 (0)1279 627895; fax: #44 (0)1279 627896; e-mail: karl—d—[email protected]
The reduction in protein deposition, to surfaces treated with a variety of poloxamers has been monitored using a wide range of techniques including radiolabelling [1] contact angle [6], photon correlation spectroscopy [7], sedimentation [7] and ultraviolet spectroscopy [8]. A more recent introduction to the field has been the application of surface plasmon resonance [4]. This is an evanescent wave technique, which allows the real-time monitoring of interactions with surfaces. The fundamental details of SPR operation are discussed in an excellent review by Hutchinson [9]. SPR is based upon total internal reflection (TIR). TIR occurs when light beyond the critical angle and incident on an interface of higher to lower refractive index is totally reflected. An electromagnetic component, the evanescent wave is created at the junction between incident laser light and the underside of a glass prism. This moves away from the interface in to the medium of lower refractive index. If the evanescent wave is allowed to contact a thin transition metal film as is the case in SPR, free electrons in the metal film couple with the wave and resonate with certain angles of incident light. The evanescent wave penetrates the lower index medium decaying
0142-9612/99/$ — see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 2 4 5 - 2
886
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
exponentially to a distance of around 600 nm. The electron clouds (plasmons) have a collective wave parallel to the interface. The plasmons absorb energy from the incident light and a characteristic drop in reflected light intensity is observed. The evanescent wave profile is dependent upon the thickness and refractive index of the layer above the metal film. Changes in this layer requires a new angle of incident light to set-up resonance and it is the change in the surface plasmon resonance angle with time that is measured using a photo diode detector, along with the depth and shape of the ‘dip angle’. SPR has the advantage over many other methods [10], in that it is a real-time technique capable of monitoring not only the adsorption of compounds to transition metal surfaces [11], but also subsequent binding to these layers [12]. It is therefore an ideal technique for monitoring the adsorption of PEO surfactants from solution and screening with protein solutions to assess the effectiveness of the surfactant layer to reduce protein adsorption. It has been suggested that the single PEO chain length copolymers adopted thus far may not offer the most effective surface required to block protein interaction and that a surface comprising of a mixture of both long and short PEO chains would produce greater steric repulsion properties [13]. The objective of this preliminary study was to investigate the use of mixtures of PEO/PPO pluronic and tetronic clock copolymers in reducing non-specific protein adhesion on gold surfaces. A range of both long chain, short chain and mixtures of solutions of the PEO surfactants were passed over gold SPR slides via a flow cell. Surfaces were then flushed with buffer solution to remove unbound material and followed by a solution of bovine serum albumin.
2. Materials and methods Surface plasmon resonance studies were carried out on a system utilising the Kretchmann configuration (Johnson and Johnson Orthoclinical diagnostics Ltd, Chal-
font St. Giles, UK), the details and operation of this instrument have previously been described elsewhere [10, 14]. Gold SPR slides (48 nm), were obtained from the same source, being optically matched to the prism involved using Resolve, a low viscosity, microscope matching oil, (Stephens Scientific, Riverdale, NJ, USA). Three flow channels were formed on top of the slide using a clamped, thermally regulated block with a custom butyl ‘O’-ringtype arrangement. Liquid flow was achieved using an on-board syringe driver holding 1.0 ml disposable plastic syringes and driving 0.95 ml of fluid at rates between 1 and 95 ll s\. All experiments related to this work were carried out at least in triplicate, with a solution flow rate of 4 ll s\ and a maintained temperature of 30$0.1°C. Copolymers as described in Table 1 were obtained from BASF, (Parsippany, New Jersey, USA) and used without further purification. 4%w/v solutions of single polymers and 50 : 50 mixtures, to give 4%w/v, were prepared in 0.2 lm filtered phosphate buffer, (10 mM, pH 7.2). Solutions of bovine serum albumin, (Sigma, Poole, Dorset, UK), were prepared at a concentration of 1 mg ml\ using the same buffer. When in use all solutions were kept at 30°C in a water bath and immediately transferred to the SPR instrument to avoid refractive index shifts due to temperature change. Gold surfaces were treated first with buffer, then with either samples of triblock copolymers of known structure, or equal mixtures of the long and short chains, to a total concentration of 4%w/v [13]. The slides received a wash with buffer at the same flow rate to remove any excess, non-adsorbed material, followed by solutions of bovine serum albumin at a fixed concentration of 1 mg ml\ and a further buffer wash. Changes in SPR angle, expressed as milli-degree angle change (MDA), were recorded over 5 min periods with a gate time of 1 s, for both the protein adsorption and final wash steps. Custom written WINspr V1.1 software, (Johnson and Johnson Clinical Orthodiagnostics, Chalfont St. Giles, UK) was employed for the data acquisition step. Data analysis was carried out using WBplot software.
Table 1 Properties of block copolymers [15] Block copolymer
Av. mol. wt
EO
Amt molecular blocks (mol) PO
EO
Cloud point (1% aqueous soln.) (°C)
Cloud point (10% aqueous soln.) (°C)
L61 L64 F68 F127 Tetronic T908
2000 2900 8400 12 600 26 600
3 13 76 98 122
30 30 30 67 22
3 13 76 98 122
'24 58 '100 '100 '100
17 60 '100 '100 '100
34 10.5 !26.2 6.9 7.8 6.1 5.83 0.8 2.59 !7 !4.4 (!3.5) 2 1.5 34.4 36.5 2.1 (!1.9) 0.3 0.9 11.8 !11.7 0.1 0.07
Negative residual shifts for L64#F68 and L64#T908 may indicate a small amount of surfactant desorption. N"number of experiments, SD"standard deviation of data set, mda"milli degree angle.
4.6 1.2 48.5 !12.3 36.2 28.5 5.4 1.2 52.3 !11 41.3 32.5 6.2 1 51 !11.4 39.6 31.2 7 2.9 48 !11.6 36.4 28.6 2.4 1.1
SD N"5 N"3 SD N"3
SD
N"3
SD
N"5
SD
N"5
SD
N"5
SD
N"5
SD
N"5
SD
L64#L61 L64#T908 L64#F68 L64#F127 L64 F68 F127 T908#F127
49.9 !14.5 35.4 27.8 Protein (mda) Protein adsorption Protein desorption Residual protein Residual protein as % of protein to gold shift
mda—milli degree angle (1°"1000 mda). SD—standard deviation. N—number of experiments.
887
T908
SPR results are most commonly examined as timeresolved traces, however, further information about the system can be obtained directly from the reflectivity vs. angle curves. Both methods of analysis are employed here. The 1 mg ml\ solution of bovine serum albumin, (BSA), in pH 7.2 buffer absorbs strongly to a plain gold SPR slide surface, with a 127 mda, (2.9 SD, N"6), increase in angle and an initial slope of 3.2 mda s\, (0.3 SD, N"4). An immediate, subsequent, buffer wash led to a small, desorption at !0.6 mda s\, (0.1 SD, N"6), of 12.4 mda, (1.1 SD, N"6), resulting in an overall shift due to BSA adsorption of 115 mda, (4.2 SD, N"6). These values are within the range observed by other groups [14]. In the case of the surfactants solutions the adsorption profiles gave readings ranging between 370 and 420 mda with F68 giving 520 mda, except for the solution containing L64#L61, which gave increasing values in excess of 1900 mda. These shifts between 370 and 420 mda are of the order expected from the refractive indices of the solutions, compared with changes of mda previously derived from methanol and ethanol mixtures having an 8.6;10 mda/unit change with refractive index. The value for L64#L61 mixture does not correspond with its refractive index as measured by an Abbe´ refractometer i.e.1.3383. It does agree however, with the results obtained on a polystyrene coated silver slide by Green et al. (1997) and could be accounted for by the fact that the solution was turbid, due to the experimental temperature being close to the cloud point of the solution. The refractive indices of L61 and L64 are 1.4518 and 1.4570, respectively, these values indicate that a thick layer of polymer is adsorbed to the surface rather than just a monolayer. All other surfactant solutions were clear and at the experimental temperatures, well below the recorded cloud points (Table 1). Flushing of the surface with buffer removed the surfactants giving an overall shift due to adsorption of between 5 and 40 mda. These surfaces were stable with respect to buffer flow at 4 ll s\ for several hours after which time recording was stopped. The changes in resonance angle due to protein adsorption to the surfactant treated slides are given in Table 2. It can be seen that on all surfaces treated with either a single surfactant or a mixture, there was a degree of protein adsorption evident, observed as increases in resonance angle of the order of 24—32.5 mda, except for the
Surfactant C
3. Results and discussion
Table 2 Surface plasmon resonance results for poloxamer adsorption and protein interaction. Conditions: 30°C, 4 ll s\, 4% wt\ total surfactant, mixtures"1 : 1 ratio in pH 7.2, 0.2 lm filtered PBS buffer, 1 mg ml\ BSA
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
888
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
surfaces modified with the mixtures L64#F127, L64# F68 and L64#T908, all recording much lower shifts. As the mda adsorption shift is due to two factors, that is the difference in refractive index between the solution of BSA and buffer, plus changes in refractive index at the surface due to adsorption of BSA, then the results (Table 2), indicate insignificant adsorption of protein for the L64#F127, L64#F68 and L64#T908 treated gold slides and a reduced adsorption for all the other copolymer solutions used compared with the untreated gold slides (Fig. 1). All samples recorded drop in angle following a buffer wash, however the most significant feature was the final mda value after the wash. All long chain-surfactants and the L64, recorded a residual protein deposit on the surface at values between 30.1% and 36% of the residual shift of the response due to protein adsorbing to an untreated gold surface. However, for all the mixtures of long and short-chain surfactants used, the residual protein values have been significantly reduced even further (P"0.05). It is interesting to note that although results for L64#F68 and L64#L61 both indicate a certain level of adsorption of protein before the buffer wash, it cannot be strongly bound as it is removed readily on buffer washing, there being no significant difference at the 95% confidence level for the adsorption and desorption readings. However for all the others the decrease on buffer washing is 12$2.5 mda, the same as that for BSA
adsorption on to untreated gold slides, this value is most likely due to the difference in refractive index between buffer and BSA solution. The increased reduction of protein adhesion caused by the use of a mixture of long and short-chain surfactants is of both practical and theoretical significance. These data support the suggestion by Stenkamp and Berg [13] that effective steric stabilisation should be achieved by a polydisperse mixture. Thus, according to this theory, the mix of long and short chains at the surface should produce more effective steric hindrance to the protein, as the longer chains would be less inhibited in their movement and the shorter chain PEOs would be interdigitated close to the surface. Factors such as the ratio of the two polymers of different chain lengths, which may not be adsorbing in the same ratio as the solution composition, will be the subject of further investigation. Figures 2 and 3 show the SPR curves, as the initial and final scans during the adsorption of a 1 mg ml\ BSA solution, for both an untreated gold surface and one treated with a mixture of L64#F127. It can be seen that a large shift in minima position occurs with the application of BSA to an untreated gold surface. When the gold surface has been previously exposed to a mixture of long and short chain PEOs, (L64#F127), there is no discernible shift in the minimum or broadening of the SPR curve upon exposure of the surface to the protein solution,
Fig. 1. SPR angle change vs. time for the cumulative addition of a 50 : 50 mixture of L64 : 127 and a subsequent buffer wash, followed by a 1 mg ml\ BSA solution and further buffer wash (flow rate 4 ll s\, pH 7.2, ¹"30°C).
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
889
Fig. 2. Plot of average initial (pre BSA) and final (post BSA) scans for adsorption of 1 mg ml\ BSA to an untreated gold SPR slide (flow rate"4 ll s\, pH 7.2, ¹"30°C, N"3).
Fig. 3. Plot of average initial (pre BSA) and final (post BSA) scans for adsorption of 1 mg ml\ BSA to a gold SPR slide previously treated with a 4%w/v 50 : 50 mixture of L64 :127 slide (flow rate"4 ll s\, pH 7.2, ¹"30°C, N"3).
890
K.D. Pavey, C.J. Olliff / Biomaterials 20 (1999) 885—890
supporting the evidence that little if any protein is adsorbed. 4. Conclusions It has been shown that bovine serum albumin adsorbs, quickly and strongly, to untreated gold surfaces. The self-adsorption of PEO block copolymers of a range of molecular weight reduces protein adsorption to gold surfaces. Furthermore, if mixtures of long- and shortchain polaxamers are employed, protein adsorption can be further reduced. The specific mixtures of poloxamers L64#F127, L64#T908, L64#F68 were shown to be capable of reducing or inhibiting the binding of BSA to a gold surface to levels not discernible in this study. The mixture L64#L61 was shown to be equally good at resisting protein adsorption, despite the adsorption values being affected by the turbidity of the solution at the experimental temperature.
References [1] Amiji M, Park K. Prevention of protein adsorption and platelet adhesion on surfaces by PEO/PPO/PEO triblock copolymers. Biomater 1992;13(10):682—92. [2] Portoles M, Refojo MF, Leong FL. Poloxamer 407 as a bacterial adhesive for hydrogel contact lenses. J Biomed Mater Res 1994;28(3):303—9. [3] Lee JH, Andrade JD. Surface properties of aqueous PEO/PEO block copolymer surfactants. In: Andrade JD, editor. Polymer surface dynamics. New York: Plenum Press, 1988:119.
[4] Green RJ, Tasker S, Davies J, Davies MC, Roberts CJ, Tendler SJB. Adsorption of PEO—PPO—PEO triblock copolymers at the solid/liquid interface: a surface plasmon resonance study. Langmiur 1997;13:6510—5. [5] Lee J, Martic PA, Tan JS. Protein adsorption on pluronic copolymer-coated polystyrene particles. J Colloid Interfacial Sci 1989;131(1):252—66. [6] Freij-Larsson C, Nylander T, Jannasch P, Wesslen B. Adsorption behaviour of amphiphilic polymers at hydrophobic surfaces: effects on protein adsorption. Biomater 1996;17(22):2199—207. [7] Wu DT, Yokoyama A. An experimental study on the effect of adsorbing and non-adsorbing block sizes on diblock copolymer adsorption. Polym J 1991;23(5):709—14. [8] Carthew D, Buckton G, Parsons GE, Poole S. The influence of a phase transition in poly(oxyethylene)/poly(oxypropylene)/ poly(oxyethylene) surfactants on adsorption behaviour from dilute aqueous solution. Pharm Sci 1995;1:3—5. [9] Hutchinson AM. Evanescent wave biosensors: real time analysis of biomolecular interactions. Mol Biotech 1995;3:47—54. [10] Davies J. In: Davies J, editor. Chemistry and physics of surfaces and interfaces. Surface analytical techniques for probing biomaterial processes. New York: CRC Press, 1996:67. [11] DeBono RF, Loucks GD, Della Manna D, Krull UJ. Self-assembly of short and long chain-alkyl thiols onto gold surfaces: a real time study using surface plasmon resonance techniques. Can J Chem 1996;74:667—88. [12] Liedberg B, Nylander C, Lundstro¨m I. Biosensing with surface plasmon resonance: how it all started? Biosens Bioelectron 1995;10(8):1—6. [13] Stenkamp VS, Berg JC. The role of long tails in steric stabilization and hydrodynamic layer thickness. Langmuir 1997;13:3827—32. [14] Green RJ, Davies J, Davies MC. Roberts GJ, Tendler SJB. Surface plasmon resonance for real time in situ analysis of protein adsorption to polymer surfaces. Biomater 1997;18(5):405—13. [15] BASF performance chemicals technical literature. BASF Corporation, 1998 (http://www.basf.com/businesses/chemicals/performance/html/tetronic.html).