SEARCH

SEARCH BY CITATION

Keywords:

  • warfarin;
  • fluorescence;
  • fluorescence lifetimes;
  • organic solvents;
  • molecular recognition;
  • molecular probe;
  • molecularly imprinted polymer;
  • human serum albumin;
  • HSA;
  • colloidal silica;
  • TCSPC

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

The complex nature of the structure of the anticoagulant warfarin is reflected in the diversity of binding modes observed in warfarin–protein recognition systems. A series of theoretical, 1H-NMR and steady state and time resolved fluorescence spectroscopic studies, have been used to establish correlations between the molecular environment provided by various solvent systems and the relative concentrations of the various members of warfarin's ensemble of isomers. A consequence of these observations is that the judicious choice of solvent system or molecular environment of warfarin allows for manipulation of the position of the equilibrium between isomeric structures such as the hemiacetal and open phenol-keto forms, the latter even possible in a deprotonated form, where in each case unique spectroscopic properties are exhibited by the respective structures. Collectively, warfarin's capacity to adapt its structure as a function of environment in conjunction with the fluorescence behaviours of the various isomers together provide an environment-dependent molecular switch with reporter properties, which allows for the simultaneous detection of warfarin in different states with lifetimes spanning the range < 0.10–5.5 ns. These characteristics are here used to examine warfarin binding domains in a series of materials (solvents, protein, inorganic matrix and synthetic polymer). Moreover, these studies demonstrate the potential for using warfarin, or other switchable analogues thereof, as a tool for studying molecular level characteristics, for example local dielectricity. Copyright © 2010 John Wiley & Sons, Ltd.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

The coumarin derivative 3-(α-acetonylbenzyl)-4-hydroxycoumarin, most commonly known as warfarin, has for over 50 years been extensively used as an anticoagulant in the prevention of cardiovascular disease states that can lead to life threatening conditions, for example thrombosis and myocardial infarction (Landefeld and Beyth, 1993). Warfarin elicits its anticoagulant activity by inhibiting the active site of vitamin K dependent epoxide reductase (VKOR) (Li et al., 2004; Rost et al., 2004). Despite the long and frequent use of the drug and the complications that can arise due to the drug's narrow therapeutic window, the mechanisms of action underlying the inhibition process are not yet fully elucidated. A possible reason for the general difficulties associated with the monitoring of this substance and correlation with patient physiological status is the molecular-level complexity of warfarin. Some aspects of the isomeric distribution of warfarin were demonstrated by Valente et al. (1975, 1977, 1978) during the 1970s. Using theoretical treatments, 1H-NMR spectropscopy and a series of fluorescence spectroscopic studies (steady state, time resolved and anisotropy), we have recently mapped the relative concentrations of the members of the warfarin isomer ensemble as a function of solvent or molecular environment, Figure 1 (Karlsson et al., 2007, 2009). The traditional difficulties associated with correlating warfarin concentration with biological activity are reflected in the results of studies examining its binding to proteins involve in the bioavailabilty and even the direct action of warfarin. Human serum albumin (HSA) (Ha et al., 2000; Il'ichev et al., 2002; Petersen et al., 2002) and cytochrome P2C9 (CYP2C9) (He et al., 1999) and even binding site models such as the cyclodextrins (Ishiwata and Kamiya, 1997) have been examined. In these cases, the binding sites demonstrate preferences for different isomers, Table 1. As the bioavailability of the drug is determined by a range of factors including adsorption, transport, affinity for site of action, metabolism, etc., the structural variability demonstrated by warfarin makes the study of its structure-function relationships difficult as each involves quite unique molecular recognition events and processes. An illustration of this is found in the fact that although VKOR is the site of action, at any given time 99% of the drug present in a patient is bound to HSA (Yacobi et al., 1976). On account of this it can be expected that any factor affecting warfarin's interaction with HSA will have dramatic effects on its drug action to VKOR.

thumbnail image

Figure 1. Structures of warfarin reported in the literature. OCO, open side chain coumarin; OCH, open side chain chromone; CCH, cyclic chromone; DCO, deprotonated open side chain coumarin; HCO, H-bonded open side chain coumarin; CCO, cyclic coumarin; DCH, deprotonated open side chain chromone.

Download figure to PowerPoint

Table 1. Preferred isomers for warfarin – receptor binding
Binding sitePrefered isomerRefs.
HSADeprotonated open side chainPetitpas et al

. (2001)

Cytochrome P450 2C9 (CYP2C9)Cyclic hemiketal

Heimark and Trager (1984)

, He et al

. (1999)

VKORDeprotonated open side chainGebauer (2007)
ß-cyclodextrinDeprotonated open side chain

Ishiwata and Kamiya (1997)

, Vasquez et al

. (2009)

Warfarin's coumarin moiety provides a valuable potential tool to aid in our investigation of drug–protein binding events. Although coumarin itself provides a fairly low fluorescence quantum yield at room temperature in most organic solvents (Song and Gordon, 1970; Lakowicz, 1999), the presence of different electron withdrawing or accepting substituents on the coumarin core structure can greatly influence the fluorescence properties (Mantulin and Song, 1973; de Melo et al., 1994; Reichardt, 1994; Gao et al., 2000; de Melo and Fernandes, 2001; Ammar et al., 2003; Sharma et al., 2003). The nature and position of substituents is important for the fluorescence behaviour of these compounds, and many spectroscopic and theoretical studies have been undertaken to characterize the photophysical behaviour of these substances (Machado et al., 2001, 2003; Ionescu and Hillebrand, 2003). Here we exploit the unique chameleon-like fluorescence behaviour of warfarin isomers in various solvent systems in conjunction with structural studies of warfarin–protein binding systems to deduce the nature of the nano-scale recognition domains. These insights suggest using warfarin as a probe for examining the molecular nature of environments associated with warfarin's in vivo function, or even as a general environment sensitive molecular probe.

EXPERIMENTAL SECTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Chemicals

Racemic warfarin (3-(α-acetonylbenzyl)-4-hydroxycoumarin, min 98%) and human serum albumin (HSA, essentially fatty acid free, 99%) were obtained from Sigma–Aldrich (St. Louis, MO, USA). Ethanol, EtOH (99.5%) was purchased from Solveco Chemicals AB (Täby, Sweden); acetonitrile, ACN (≥99.8%); dichloromethane, DCM (≥99.9%) was from Aldrich, chloroform, CHCl3 (≥99%); 2-propanol, 2PrOH (≥99.9%), ethyl acetate, EtOAc (≥99.8%), n-heptane, HEP (≥99.3%) and tetrahydrofuran, THF (≥99.9%) were all obtained from Merck KGaA (Darmstadt, Germany). Acetic acid, AcOH (>98%) was purchased from VWR (Stockholm, Sweden). All chemicals were used as received. Molecularly imprinted and the corresponding reference polymers were synthesised as reported by Karlsson et al. (submitted).

Fluorescence spectroscopy

All fluorescence steady state emission spectra were obtained using 14 µM solutions of warfarin. Spectra were corrected for wavelength dependent optical components at 294 ± 1 K, with the excitation and emission monochromator spectral band widths at 1 and 2 nm, unless otherwise stated.

Steady-state studies

The dependency of fluorescence emission on solvent polarity was investigated by absorption spectroscopy in different solvents using either 300 nm (CHCl3, AcOH, 2PrOH, EtOH and ACN) or 305 nm (HEP, EtOAc, THF and DCM) as the wavelength of excitation. Absorption was kept below 0.2 in order to prevent self-absorption of the emission and inner filter effects. At the excitation wavelength 305 nm the spectral band widths were placed at 2 and 4 nm.

Time-resolved

The time-resolved measurements, with resolution down to the sub-nanosecond time regime, were performed with the time-correlated single-photon counting (TCSPC) technique. Fluorescence decay times of warfarin in ACN were measured at different emission wavelengths (345, 360, 380, 389, 390, 400, 408, 420 and 450 nm) set by a single grating monochromator with spectral band widths of 8 or 16 nm. The HSA-warfarin binding studies were performed in PBS buffer using 10 µM of warfarin and 0.0065–0.65 mg mL−1 of HSA, using 420 nm as the fluorescence emission wavelength set using a single grating monochromator with spectral band widths of 12 or 16 nm.

The excitation source was either an IBH NanoLed (IBH) producing ∼295 nm (warfarin solvent experiments) or 334 nm (HSA-warfarin binding experiments) excitation pulses at 1.0 MHz repetition rates. The photon-counting rate was always maintained at lower than 2% of the excitation source repetition rate to avoid photon pile-up effects. A typical instrumental pulse using an IBH TBX-04 detector at a time calibration of 13 ps per channel and a scattering sample of LUDOX® was about 600 ps at full width half maximum. To remove the additional excitation pulse from the NanoLed at around 430 nm, a cut-off filter (UG-11) was kept in front of the excitation source. Before every sample measurement, the instrument response function was measured using LUDOX® solution (excitation and emission wavelengths 295 and 334 nm respectively). This solution was also used in warfarin-recognition characterisation. The fluorescence decays were collected over 4096 channels and fluorescence lifetimes were calculated using either a double or a triple exponential method within the software package Analysis Software (version 6.1.51, IBH). Typically, curve-fits were accepted when χ2 ≤ 1.2.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

The 4-hydroxycoumarin derivative warfarin has for decades maintained a position as the anticoagulant of choice for treatment of patients suffering from various chronic cardiovascular conditions. However, several patient dependent factors have been demonstrated to have impact upon warfarin's bioavailablity, for example diet, genetic disposition, age, pharmaceutical regime (Holbrook et al., 2005). An issue intimately related to the complexity of warfarin's bioavailability is its chameleon-like behaviour, whereby the relative concentrations of the ensemble of warfarin isomers present can be profoundly influenced by molecular environment (Karlsson et al., 2007, 2009). Furthermore, environment, and in particular in solvent, has been implicated in mediating the reaction pathways involved in the interconversion of the dominant isomeric forms of this drug as investigated through a series of DFT studies (Henschel et al., submitted).

An examination of warfarin's steady state fluorescence spectroscopic behaviour in a broad range of solvent environments was undertaken to examine correlations with its isomeric state(s). The adsorption and fluorescence emission spectra of warfarin in a series of solvents ranging from low to high polarity were studied. Data were then used to construct a Lippert–Mataga plot, Figure 2. The plot provides an association between the S0[RIGHTWARDS ARROW]S1 absorption bands and the corresponding S1[RIGHTWARDS ARROW]S0 emission bands. Straight lines are obtained for (ideal) situations were no specific interaction is present between the fluorophore and solvent, the slope of which affords the change in the dipole moment upon excitation. The clear non-linearity observed, with a break at Δf ≈ 0.2, implicates a distinct solvent mediated influence on the fluorescence behaviour, which corresponds to the presence of the open chain deprotonated forms DCO and DCH (see Figure 1) in the more polar solvent regimes. In nonpolar solvents, where Δf < 0.2, the isomeric ensemble is effectively only populated by cyclic hemiketal, as confirmed through comparison with NMR studies (Valente et al., 1977; Karlsson et al., 2007). The vast differences in warfarin's spectroscopic behaviour suggested the possibility for using this solvent dependent behaviour to provide simple approximation of the environment provided by a binding site. As the inherent complexity of spectra arising from the presence of multiple isomers a strategy based upon time resolved measurements was suggested.

thumbnail image

Figure 2. Lippert–Mataga plot demonstrating the warfarin fluorescence behaviour in (1) n-heptane, (2) chloroform, (3) ethyl acetate, (4) acetic acid, (5) tetrahydrofuran, (6) dichloromethane, (7) 2-propanol, (8) ethanol, and (9) acetonitrile.

Download figure to PowerPoint

Time resolved measurements from a selection of the solvent systems studied provided an even more detailed perspective, Table 2. Warfarin displays two fluorescence lifetimes in polar organic solvents, which are dependent on solvent character and show different contributions to the emission spectra with τ1 < 0.1 ns and τ2 = 0.5–1.6 ns. Moreover, warfarin excitation in non-polar organic solvents results in fluorescence associated with a fast decay of τ < 0.1 ns. The short lifetime (τ1), observed in both non-polar and polar organic solvents, is ascribed to the neutral isomeric forms of warfarin. In the case of the longer lifetime (τ2), this was assigned to the deprotonated open side chain form, which only exists in polar environments. The distinct distribution of the lifetimes implied that differentiation of, for example, ligand-protein binding modes may be achievable.

Table 2. The relative contributions to the emission spectrum and corresponding time resolved fluorescence lifetimes for warfarin in various molecular environments
SolventaRelative amplitudeb at different λem. (nm)Fluorescence lifetime (ns)c
345360380389390400408420450τ1τ2τ3
  • a

    HSA, MIP and REF were all in PBS buffer at pH 7.3 and the silica particles (LUDOX®) were suspended in deionized water.

  • b

    The amplitude presented describes the relative contribution (percentage) of the lifetime τ1 to the total fluorescence spectrum at the given emission wavelength. Relative amplitudes (%) for τ2 and τ3 for binding systems, respectively: HSA 46 and 29, LUDOX 1 and 1, MIP 30 and 56, REF 35 and 24.

  • c

    Fluorescence lifetimes are presented as mean ± standard deviation.

Chloroform99 99  99   <0.10  
Acetic acid99 100  99   <0.10  
2-Propanol907245    22 <0.100.53 ± 0.014 
Ethanol2960 75   72 0.20 ± 0.110.45 ± 0.062 
Acetonitrile73 6   6 40.14 ± 0.0731.6 ± 0.11 
HSA       25 0.22 ± 0.00501.3 ± 0.0243.7 ± 0.018
LUDOX®    97    <0.101.4 ± 0.0865.5 ± 0.15
MIP       15 0.18 ± 0.00711.4 ± 0.0594.3 ± 0.026
REF       41 0.21 ± 0.00421.8 ± 0.0595.3 ± 0.052

The nature of the fluorescence properties upon warfarin-recognition to each of three distinct macromolecular recognition systems (protein, synthetic polymer and inorganic material) was then examined to test the possibility of using warfarin as an environment selective molcular probe. In the first case, the recognition of warfarin by HSA was investigated, where an observed τ2 = 1.3 ns and the longer lifetime, τ3, were assigned to alternative anion binding modes. The value of τ2 suggested the presence of a binding domain with characteristics comparable to those provided by acetonitrile (ε = 36.4) (Karlsson et al., 2007). The amplitudes reflect the relative concentration of the species present. The environment provided by a colloidal silica suspension (LUDOX®), which affords no inherent binding for the anionic forms of warfarin was examined. Here the observed distribution of fluorescence lifetime corresponding to the unbound warfarin, τ1 < 0.1 ns, in its deprotonated open sidechain form, and a small population bound to the silica particles. The longer lifetimes of τ2 and τ3 are indicative of those obtained for solvents of low dielectricity, which in turn reflect the relative hydrophobicity of the environment provided by the silica particles.

The mixed-mode binding offered by a methacrylic acid-ethylene glycol co-polymer in aqueus media (Karlsson et al., 2001, 2004; Rosengren et al., 2005) was also probed. Polymers of this general type have recently been used in conjunction with molecular imprinting protocols (Alexander et al., 2006) for generating materials with warfarin-selective binding (Karlsson et al., submitted). Time resolved fluorescence spectroscopic studies of warfarin binding in an imprinted polymer (MIP, Table 2) revealed a population of lifetimes reflecting the dielectricity of acetonitrile (due to similarity in the τ2 and τ3 lifetimes). Perhaps more interesting is the similarity in this behaviour to that observed for warfarin binding to the HSA Sudlow II binding site. That the corresponding reference polymer (REF, identical chemical composition, though without imprinted sites) does not demonstrate similar spectroscopic evidence of HSA-like recognition of warfarin (due to higher amplitude for the unbound state of warfarin) further illustrates the potential for using warfarin as an environment sensitive molecular probe.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

We have employed a series of fluorescence spectroscopic studies (steady state and time resolved) of warfarin behaviour in various solvents and thereafter of binding to various macromolecular systems: biological (protein), inorganic and synthetic polymer. Collectively, warfarin's environment-dependent structural variation and the fluorescence behaviour of the various isomers provide an environment-dependent molecular switch with reporter properties. We suggest that challenging molecular environments with warfarin and monitoring fluorescence behaviour can provide molecular-level insights concerning the warfarin–material interaction domain, for example estimates of dielectricity at binding domains. Such data may ultimately lead to a better understanding of warfarin's bioavailability, for example through a better understanding of the interaction of this anticoagulant with various relevant protein systems or other structures such as membranes (Karlsson et al., submitted). In a broader context, the use of warfarin in conjunction with time resolved fluorescence methods such as TCSPC as a highly sensitive general tool for probing nano-scale molecular environments offers much promise.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

The financial support of the Swedish Research Council (VR), the Swedish Knowledge Foundation (KKS), Carl-Trygger's Foundation and Linnaeus University is most gratefully acknowledged.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL SECTION
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES