Cyclodexrtrins as Hosts for the Array-Based Detection of Persistent Organic Pollutants in Complex Media

The ability to quickly determine the nature of small-molecule toxicants after an anthropogenic event would greatly benefit first responders and medical personnel. Current detection methods, while elegant, require several separation and purification steps before the samples can be submitted for analysis, which can be a timeconsuming process. There is a crucial knowledge gap that exists as a result. Reported herein is the use of a non-toxic, commercially-available molecule, cyclodextrin, to rapidly isolate and detect the toxic components involved in a spill event that would address this knowledge gap. This cyclodextrin-based scheme would work as a compliment to established analysis procedures by establishing a rapid, highthroughput procedure that can be used to quickly scan samples to determine the nature of the compounds involved in a spill event. This would provide first responders with the information they need to develop an effective response in a timely manner, and samples would still be sent for more intense analysis using standardized procedures, such as those set by the Environmental Protection Agency (EPA) to confirm the results and quantify them. Similarly, this method can be used by medical personnel to quickly analyze samples from patients to determine if their symptoms are a result of a

PCBs were historically used as refrigerator coolants and in a variety of manufacturing products. 6 Although the use of PCBs was banned in the United States in 1979, 7 their atmospheric stability means that PCBs still persist in the environment. 8 Some examples of PCBs are shown in Figure 1; the FDA-recommended concentration limits for PCBs in food ranges from 0.2-3.0 parts per million (ppm). 9 Current methods for the detection of PAHs and PCBs generally rely on separation using chromatography, followed by detection via mass spectrometry (for PAHs 10 and PCBs 11 ) or fluorescence spectroscopy (for PAHs). The development of new methods for the detection of these compounds remains a high priority, especially if such methods have improved sensitivity and/or selectivity.
We previously reported that energy transfer occurs between anthracene and a squaraine fluorophore inside the cavity of γ-cyclodextrin, with up to 35% energy transfer observed from anthracene excitation compared to direct squaraine excitation. 12 The energy transfer efficiency is defined as: where IDA is the integrated emission of the fluorophore from PAH excitation and ID is the integrated fluorophore emission from direct excitation.
Although examples of energy transfer with covalently-modified cyclodextrins have been reported, 13 non-covalent energy transfer inside cyclodextrin cavities is much less developed, 14 even though such energy transfer is substantially easier to tune and optimize. 15

MATERIALS AND METHODS
Reported herein is the development of a widely applicable non-covalent energy transfer system between PAH and PCB energy donors and fluorophore acceptors.
These fluorophores ( Figure 2) were chosen because of their high quantum yields, 16 and established use in a variety of sensing schemes. 17 Compound 8 is commercially available, and compounds 9 and 10 were synthesized following known procedures. 18 Energy transfer from the analytes to the fluorophores in the presence of cyclodextrin was measured by mixing the analyte and fluorophore in a γ-cyclodextrin solution in phosphate-buffered saline (PBS), which should generate a ternary complex.
The complex was then excited near the absorbance maximum of the analyte and near the maximum of the fluorophore, and energy transfer efficiencies were calculated.
Control experiments were also done in which the fluorophore was excited at the analyte's excitation wavelength in the absence of any analyte, to determine whether peaks previously identified as energy transfer peaks might be due to fluorophore emission from excitation at a wavelength where it has non-zero absorbance.
The results of these experiments were quantified as ''fluorophore emission ratios,'' defined as the integrated fluorophore emission in the absence of an analyte divided by the integrated fluorophore emission in the presence of the analyte (Table   1). Defined as the ratio of fluorophore emission via indirect excitation in the absence of the analyte to fluorophore emission via indirect excitation in the presence of the analyte. Any values between 0.95 and 1.05 indicate that any presumed energy transfer is merely a result of the fluorophore absorbing a nontrivial amount of energy via the "indirect" excitation pathway.

RESULTS AND DISCUSSION
These experiments revealed that some of the analyte-fluorophore pairs that have a significant fluorophore peak from analyte excitation actually have the same fluorophore peak in the absence of analyte (fluorophore emission ratio between 0.95 and 1.05). For several cases, however, the fluorophore emission ratios were significantly higher than 1 (indicating that the analyte actually quenches fluorophore emission), and in other cases the ratio was significantly less than 1 (indicating that the desired energy transfer is observed). The maximum energy transfer efficiencies for all analyte-fluorophore combinations that demonstrate energy transfer are shown in Table   2. a Excessive overlap between the analyte and fluorophore prevented accurate integration. b No fluorophore peak was observed from analyte excitation. c Fluorophore emission ratios indicate no real energy transfer is occurring (emission ratios between 0.95 and 1.05).
Although anthracene 1 does not undergo significant energy transfer with fluorophores 8 and 9 (as measured by the fluorophore emission ratios), the highly toxic PAHs 2 and 3 demonstrate significant energy transfer. Benzo[a]pyrene 3 acted as an energy donor with fluorophores 8 and 9 (and to a limited extent with squaraine 10).
The energy transfer peaks with compounds 8 and 10 are clearly visible at 558 nm and 659 nm, respectively ( Figure 3). Control experiments also demonstrated the necessity of γ-cyclodextrin for energy transfer, as in the absence of cyclodextrin only 3% energy transfer was observed for benzo[a]pyrene with compound 8 (compared to 10% in the presence of 10mM γ-cyclodextrin). The detection of benzo[a]pyrene is particularly crucial, due to its low recommended concentration limit (0.132 ppm) and high carcinogenicity.   In order for this energy transfer to be practical for the detection of toxic analytes, it needs to be both sensitive and selective. The sensitivity of this method was determined by quantifying the limits of detection for all analyte-fluorophore combinations, 19 and the results are shown in Table 3. The limits of detection are defined as the amount of analyte necessary to observe a signal that is distinguishable from the baseline (see Supporting Information for details). 20 The limits of detection for compounds 2 and 5 are below the FDA-recommended concentration limits, thus providing a useful mechanism for the detection of these highly toxic analytes. Selectivity in the detection of toxic PAHs and PCBs can be accomplished using array-based detection. Such detection systems have also been referred to as ''chemical noses,'' and have been used successfully by a number of research groups. 21 Array-based detection generally requires exposure of an analyte to a receptor array.
Statistical analyses of the resulting array of signals then lead to the selective detection of particular analytes.
Preliminary efforts towards developing an array-based detection system have yielded promising results. Using the three different fluorophores (compounds 8-10) in combination with 10 mM γ-cyclodextrin, each analyte (PAH or PCB) displayed qualitatively different fluorescence patterns when excited at 365 nm ( Figure 5).
Qualitatively different fluorescent responses were observed even in cases where the fluorophore emission ratios indicate some degree of fluorophore quenching from introduction of the analyte. The fact that each vertical column looks different means that each analyte has a different pattern of responses with the fluorophores investigated. Efforts to translate this qualitative observation into a quantitative, selective detection system are underway.

Figure 5.
Photograph of a preliminary array using 10 mM γ-cyclodextrin (excitation at 365 nm with a hand-held TLC lamp).

CONCLUSION
In summary, reported herein is the development of highly efficient noncovalent energy transfer in γ-cyclodextrin cavities between toxic energy donors and fluorescent energy acceptors. This energy transfer has a number of advantages compared to previously-developed systems, including: (a) high sensitivity (as low as 5.9 ppm for compound 2); (b) ease of tunability; and (c) widespread applicability to two classes of highly toxic compounds. The development of a full array-based detection system, and a detailed investigation of the energy transfer mechanism, are underway and the results will be reported in due course.

MATERIALS AND METHODS
All chemicals were purchased from Sigma-Aldrich Chemical Company and used as received, unless otherwise noted. 1 H NMR spectra were obtained using a Bruker 300 MHz spectrometer. UV-Visible spectra were obtained using an Agilent 8453 spectrometer equipped with a photodiode array detector. Fluorescence spectra were obtained using a Shimadzu RF-5301PC spectrophotofluorimeter.

SYNTHESES OF FLUOROPHORES
The synthesis of BODIPY 9 was performed according to literature procedures. Procedure: Compound S5 (0.968 g, 2.07 mmol, 1.0 eq.) and compound S6 (0.27 grams, 2.36 mmol, 1.14 eq.) were dissolved in 50 mL of acetone. The reaction mixture was heated to reflux for two hours. After two hours, the reaction mixture was cooled to room temperature, acetone was removed, and the crude solid was re-dissolved in dichloromethane and washed with water. The organic extract was dried over sodium sulfate, filtered and concentrated, to yield compound S7 in 97% yield (0.932 grams).

Reaction 3:
Procedure: Compound S7 (0.932 grams, 2.01 mmol, 1.0 eq.) was dissolved in 150 mL of anhydrous ethanol that was purged with nitrogen. Potassium carbonate was added, and the reaction mixture was warmed to 30 o C. The reaction mixture was stirred under nitrogen for 4 hours at 30 o C. The contents of the flask were poured over 40 mL of aqueous saturated ammonium chloride, at which point the solution turned bright orange. The product was extracted with dichloromethane and washed several times with water. The organic layer was dried over sodium sulfate, filtered, and concentrated. The product was purified via flash chromatography (1:1 dichloromethane: hexanes) to yield compound 9 in 76% yield (674 mg).
The synthesis of squaraine 10 was performed according to literature procedures: Procedure: Compound S11 (0.877 mmol, 1.0 eq., 375 mg) was dissolved in 37 mL of ethanol. 10% palladium on carbon (516 mg) was added, followed by cyclohexene (compound S12, 102 mmol, 116 eq., 10.32 mL). The reaction mixture was heated to reflux for two hours. The reaction mixture was then cooled to room temperature, and filtered through celite to remove the palladium. The filtrate was concentrated, and purified by flash chromatography (10% ethyl acetate in hexanes) to yield compound S13 (296 mg, quantitative yield). 1  Procedure: Compound S13 (0.877 mmol, 2.0 eq, 296 mg) was dissolved in 8 mL of benzene and 8 mL of n-butanol. Compound S14 (0.439 mmol, 1.0 eq, 50 mg) was added, and the reaction mixture was equipped with a Dean-Stark trap and condenser, and heated to reflux for 24 hours. After 24 hours, the reaction mixture was cooled to room temperature and concentrated to yield compound 10 directly. 1

CONTROL EXPERIMENTS
These experiments were designed to determine the emission of the fluorophores from excitation at various wavelengths (in the absence of the analyte) and compare it to the emission of fluorophores at the same wavelengths in the presence of the analyte. This will determine whether an observed "energy transfer" peak may simply be a result of exciting the fluorophore at a wavelength where it has non-zero absorbance. These experiments were conducted as follows: (a) The fluorophore was mixed with γ-cyclodextrin and excited at the excitation wavelength of the analyte (but in the absence of any analyte); and (b) the fluorophore and analyte were both mixed in γ-cyclodextrin and excited at analyte excitation wavelength.
The fluorophore emission that resulted from excitation at the analyte wavelength in the absence of the analyte was compared to the fluorophore emission from excitation at the analyte wavelength in the presence of the analyte. The ratio of these two emissions, shown as "ratio of fluorophore emissions" in the tables below, is defined as: Fluorophore emission via low wavelength excitation in the absence of an analyte/ fluorophore emission via low wavelength excitation in the presence of the analyte.
This was used to determine what fraction of that peak was a result of legitimate energy transfer rather than simple excitation of the fluorophore at a wavelength where it has non-zero absorbance.
All of these experiments were done with 1.5 nm excitation slit width and 1.5 nm emission slit width.
Pairs that fall into this category: (c) Fluorophore emission ratios less than 1. In these cases, energy transfer from the analyte to the fluorophore occurs, resulting in amplified fluorophore emission from analyte excitation.
Pyrene (2) -Rhodamine (8) Pyrene (2) where IDA is the integrated emission of the fluorophore from analyte (PAH or PCB) excitation and ID is the integrated fluorophore emission from direct fluorophore excitation.
All fluorescence emissions were integrated using Origin 8.5, and were integrated vs.
wavenumber on the X-axis.
General procedure for energy transfer experiments:

11
The synthesis of BODIPY 11 was performed according to literature procedures. In these experiments, BODIPY 11 was excited at 360 nm in the presence and absence of analyte. These results are quantified in Table S1, where the ratio of fluorophore emission is defined as: Fluorophore emission via low wavelength excitation in the absence of an analyte/ fluorophore emission via low wavelength excitation in the presence of the analyte.
Values close to 1 indicate that the analyte does not affect the fluorophore emission, and that no energy transfer is occurring between the analyte and fluorophore.
Energy transfer percentage is defined as: where IDA is the integrated emission of the fluorophore from PAH excitation and ID is the integrated fluorophore emission from direct excitation. All experiments were done at a 1.5 nm excitation slit width and 1.5 nm emission slit width.

EXPERIMENTAL DETAILS FROR LIMIT OF DETECTION EXPERIMENTS
The limit of detection (LOD) is defined as the lowest concentration of analyte at which a signal can be detected. The limit of quantification is defined at the lowest concentration of analyte that can be accurately quantified.
To determine the limit of detection (LOD) and limit of quantification (LOQ), each fluorophore-analyte combination was examined in the following manner: 1. 2.5 mL of 10 mM γ-cyclodextrin in phosphate-buffered saline (PBS) was measured into a cuvette and 100 μL of a fluorophore solution in THF was added. The solution was excited at the analyte's excitation wavelength and the fluorescence emission spectrum was recorded. Four repeat measurements were made for the fluorescence emission spectra.
2. 20 μL of a 1 mg/mL analyte solution in THF was added to the cuvette and the solution was again excited at the analyte excitation wavelength. Four repeat measurements were taken.

3.
Step 2 was repeated for 40 μL of analyte, 60 μL of analyte, 80 μL of analyte, and 100 μL of analyte. In each case, the solution was excited at the analyte excitation wavelength and the fluorescence emission spectrum was recorded four times.
4. All fluorescence emission spectra were integrated vs. wavenumber, and we generated calibration curves with the analyte concentration on the X-axis (in mM) and the integrated fluorophore emission on the Y-axis. The curve was then fitted to a straight line and an equation for the line was determined.
5. For each case, the fluorophore with γ-cyclodextrin (before any analyte was added) was also excited at the excitation wavelength for the analyte, and the fluorescence emission spectrum was recorded (as per step 1). These measurements are referred to as the "blank." 6. The limit of the blank is defined according to the following equation: Where m is the mean of the blank integrations and SD is the standard deviation.
7. The limit of the blank was then entered into the equation determined in step 4 (for the y value), and the corresponding X value was determined. This value provided the LOD in μM, which was converted into parts per million (ppm) to better compare with FDA and EPA recommended concentration limits.
8. The limit of quantification (LOQ) was calculated in a similar way to the limit of detection. First, the limit of the blank for quantification was determined according to the following equation: This value was entered into the equation determined in step 4 (for the y value), and the corresponding X value was determined to be the limit of quantification in mM. This LOQ was then converted into parts per million (ppm).     These four anthropogenic disasters highlight the need for a sensing platform that can detect a wide variety of POPs with sensitivity, selectivity, generality, and rapidity. Such a detection scheme would fill a crucial knowledge gap for first responders, who currently need to wait for time-consuming laboratory tests to accurately classify the nature of the pollutants. It would work in conjunction with current methods, by allowing first responders to screen numerous samples to rapidly understand the nature of the pollutants involved and the extent of the event so that they can begin an effective response. Previous research in our groups has demonstrated that cyclodextrin-promoted energy transfer can be used for the detection of a wide range of aromatic toxicants, 5 and that array-based detection enables the sensitive, selective, and accurate identification of a wide variety of analytes. 6 We present herein the design, execution, and evaluation of an extremely accurate array-based detection system for aromatic POPs based on cyclodextrin-promoted energy transfer from the POPs to high quantum yield fluorophores.
γ-Cyclodextrin promoted energy transfer uses γ-cyclodextrin as a supramolecular scaffold that enforces close proximity between the aromatic analyte energy donor and high quantum yield fluorophore acceptor. 7 Once bound in close proximity, excitation of the donor results in energy transfer to and emission from the fluorophore, generating a unique highly emissive fluorophore signal ( Figure 1). Because each fluorophore-analyte combination yields a distinct signal, statistical analyses of the response patterns of multiple fluorophores in cyclodextrin to a single analyte identifies a unique "fingerprint" for each analyte of interest. Analytes 1-14 are PAH and PAH metabolites, and have been found in the blood 12 and breast milk 13 of individuals living in polluted areas, with many of them known or suspected carcinogens. PCBs (15)(16)(17)(18) cause neurotoxicity and endocrine disruption, 14 and many of them are known or suspected carcinogens.
Many aromatic pesticides (19)(20)(21)(22) are suspected carcinogens, 15 and others are designated as EPA Priority Pollutants. Compounds 23 and 24 are known carcinogens and endocrine disruptors, 16 and compound 25 is a widely used additive with suspected endocrine disrupting effects. 17 Brominated flame retardants (26 and 27) are a class of pollutants that has been investigated for possible toxicity. 18 Compound 28 is classified by the IARC as Group 1 carcinogen, has been linked to bladder and lung cancer, 19 and is an EPA Priority Pollutant. Compound 29 is an amine derivative of biphenyl and has been linked to bladder cancer. 20 Compound 30 was chosen for its structural similarity to 28, to assay the array's ability to distinguish such structural variations.  LDA was successful in classifying all 30 analytes with 100% accuracy via jackknifed classification analysis (JCA), which eliminates any potential bias in the array. 22 The array was also 96% successful in identifying unknown samples from the training set correctly (115/120 correct identifications). These results represent a substantially larger substrate scope than many literaturereported arrays, 23 and a success rate in line with or better than literature reports of analogous systems. 24 The array was divided into two sections to more clearly analyze the relationships between the analytes: (1) PAHs and PAH metabolites; and (2) PCBs, endocrine disruptors, pesticides, biphenyls and flame retardants ( Figure   3). Figure S1 demonstrates that all but five of the PAHs are clustered together. The five outliers are compounds 5, 7, 9, 10, and 13; many of these are structurally related to benzo[a]pyrene and are highly fluorescent analytes (which leads to a stronger emission signal). Figure 3A shows the remaining PAHs, and highlights other key structural relationships: Anthracene 1 and two of its metabolites, compounds 2 and 3, cluster together in the array but generate well-separated signals. Fluorene 11 and three derivatives, 12, 13, and 14 also appear in the same region, but again demonstrate good separation. Similarly, carbazole 12 and partly saturated analogue 13 are close together but still well separated. Figure 3B shows the LDA plot with biphenyl-type analytes. Structural relationships can clearly be seen, for example: chlorinated compounds with similar structures cluster together, including compounds 19 and 20, and Overall, every one of the 30 analytes generates a unique signal on the LDA plot, with analytes with structural similarities grouped in a similar area.
The array successfully identified 115 out of 120 cases of unknowns for a 96% accuracy. For those analytes that appear to have overlap in the Figure 3 plots, their successful differentiation occurs in the third score, along the Z-axis (details shown in the ESI). It is important to note that LDA identifies the axis of greatest differentiation. A low score for one of the axes does not directly translate into "small feature changes" dictating differentiation, but can instead be a reflection of particularly strong differentiation across other axes. For our studies the ellipsoids provide a better qualitative measure of the degree of differentiation.  Two critical control experiments were performed. In the first experiment, an array was generated in the absence of any analyte, using γ-cyclodextrin and the three fluorophores. The blank samples excited at 300 nm and 360 nm were correctly classified as blank samples, whereas samples excited at 250 nm and 400 nm were misclassified as PCBs or DDT, respectively. These results indicate that there is a relatively weak response between these chlorinated compounds and the sensor platform.
A second control experiment was performed where the array was generated without γ-cyclodextrin. Ten analytes (6, 8, 11, 14, 17, 18, 19, 20, 28, and 30) were used for this experiment and the results are reported in Table S11 of the Supporting Information. LDA was able to differentiate between the analytes with 53% accuracy via JCA, in stark contrast to the results achieved  Figure 5). Furthermore, the array was also able to correctly identify 55 out of 60 unknown analytes.
Notably, many of the general trends that were observed in the buffer array were also observed in urine. For example, benzo[a]pyrene 6, pyrene 5,

1-3. The fact that similar trends can be seen in both matrices clearly indicates
that the association that occurs between the γ-cyclodextrin host and guest molecules is specific for each analyte-fluorophore combination and occurs similarly in both matrices.
In conclusion, we have developed an array-based strategy to detect a wide variety of POPs in both simple (phosphate-buffered saline) and complex (urine) environments. This work has shown that individual analytes can be identified with exceptional accuracy, highlighting the ability of this detection scheme to provide specific information that will be useful for first responders. The success of this array relies on strong non-covalent interactions between a toxicant donor, fluorophore acceptor, and cyclodextrin host to achieve efficient proximityinduced energy transfer, and the cyclodextrin host is crucial to ensure association between the toxicant and fluorophore. This method is expected to be generally applicable for multiple classes of aromatic analytes in a range of complex environments. Applications of this array-based sensor for POP detection in real-world matrices is currently underway, and results of these and other investigations in our laboratories will be reported in due course.

MATERIALS AND METHODS
All chemicals were purchased from Sigma-Aldrich Chemical Company and used as received. Urine samples were provided by an anonymous donor and used without any pre-treatment. 1 H NMR spectra were obtained using a Bruker 300 MHz spectrometer.   Table S1. Excitation wavelengths used for each analyte. Figure S1. LDA score plot for all analytes.    ARRAY  Table S4. Classifications of all analytes ("Analyte ID"), including misclassifications of unknowns ("Unknown Classification"). Table S5. All integration values used for the training set ("Array Integrations") and unknowns ("Unknown Integrations").

General Procedure -Sample Preparation
Two samples were prepared for each fluorophore: one served as the sample for the training set, and the other served as the unknown. For each sample, 2.5 mL of 10 mM γ-cyclodextrin and 100 μL of fluorophore were added to a vial and vigorously shaken by hand for approximately 30 seconds. The sample remained on a rotary mixer until use to ensure thorough mixing. A 96 well microplate was used, and into each well was pipetted 100 μL of the sample solution, and each solution was repeated four times (i.e. each solution was pipetted into four separate wells) to ensure data reproducibility.  A 96W microplate was used, and into each well was pipetted 100 μL of the sample solution, and each solution was repeated four times (ie each solution was pipetted into four separate wells) to ensure data reproducibility.

General Procedure -Fluorescence Studies
A BioTek Synergy Mx Multi-Mode Microplate Reader was used to generate the fluorescence data for the array. The samples were excited at the excitation of the analyte (see Table S1). The emission of each was recorded:        Table S10. Classifications of all analytes ("Analyte ID"), including classifications of unknowns ("Unknown Classification") Jackknifed Classification Matrix  14  8  28  6  20  19  11  18  17  30  %correct  14 2 Table S11. Jackknifed classification matrix for 0 mM γ-Cyclodextrin array.   Table S13. LDA Score values for an array generated in 0 mM γ-cyclodextrin.

General Procedure -Sample Preparation
Two samples were prepared for each fluorophore: one served as the sample for the training set, and the other served as the unknown. For each sample, 1.25 mL of 10 mM γ-cyclodextrin and 1.25 mL of urine were combined and mixed in a vial. Then, 100 μL of fluorophore was added and vigorously shaken by hand for approximately 30 seconds. The sample remained on a rotary mixer until use to ensure thorough mixing.
A 96 well microplate was used, and into each well was pipetted 100 μL of the sample solution, and each solution was repeated four times (i.e. each solution was pipetted into four separate wells) to ensure data reproducibility.  Table S14.

General Procedure -Fluorescence Studies
Integration data for all analytes tested in a 1:1 v/v matrix of urine and 10 mM γ-cyclodextrin.    Table S17. LDA Score values for each analyte ("Analyte ID") and the unknown classification identities ("Unknown Classification").

INTRODUCTION
Significant oil spills in recent years 1 have highlighted a number of pressing medical 2,3 and environmental 4,5 problems associated with oil spill cleanup, 6 postincident monitoring of toxicants, 7 and the prevention of future oil spills. Such problems include the long-term environmental persistence of highly toxic polycyclic aromatic hydrocarbons (PAHs) (including the known carcinogen benzo[a]pyrene), 8,9 and the accumulation of PAHs at various points in the food chain. [10][11][12][13] Methods for removing PAHs from the environment include (a) the biodegradation of PAHs into less toxic products; 14  Reported herein is the successful implementation of a γ-cyclodextrin-based system to accomplish these two key functions: (a) extracting PAHs from complex oils and binding them with moderate to good efficiencies; and (b) promoting non-covalent, proximity-induced energy transfer from the isolated PAHs to a high quantum yield BODIPY fluorophore. The oils used in these investigations (vacuum pump oil, motor oil, vegetable oil, and cod liver oil) contain varying levels of PAH contaminants: from no known PAHs in cod liver oil, 29,30 to small amounts of PAHs in several types of vegetable oil, 31,32 and large quantities of PAHs in used motor oil. 33 These 'innate' PAH amounts were detected by measuring the energy transfer efficiencies from 'undoped' oil samples to the fluorophore. Samples were separately 'doped' with small amounts of concentrated PAH solutions, which adds to the innate PAHs found in the oils and allows for a robust PAH-to-fluorophore energy transfer signal. In addition to investigating the ability of a buffered solution of γ-cyclodextrin to extract and bind toxic PAHs, we also investigated an "oil-spill-like scenario": cyclodextrin was dissolved in Narragansett Bay seawater where it was still able to extract PAHs with moderate efficiencies from motor oil samples.
This system of extraction followed by energy transfer has a number of advantages compared to previously-reported methods for the detection of PAHs, including the ability to easily modulate the fluorescence signal generated from the energy transfer via judicious choice of fluorophore. Results reported herein used BODIPY-based fluorophore 6; however, a simple replacement of this fluorophore with other known structures will lead to a fluorescence emission signal at a different wavelength. The ability to use a variety of fluorophores with different emission maxima will allow for the facile development of an array-based detection system. 34 In such a system, each analyte will interact differently with a set of fluorophores bound in cyclodextrin. Statistical analysis of the resulting response patterns will enable the selective detection of highly toxic PAHs, which is an exciting application of the results reported herein. 35 Overall, this dual-function system has significant potential applications for the isolation and detection of carcinogenic PAHs in complex, realworld environments.

Materials and methods
Where IDA is the integration of the fluorophore emission from analyte excitation and IA is the integrated fluorophore emission from direct excitation. An illustration of such energy transfer for a generic donor-acceptor pair is shown in Figure 1. Control ratio = Ifluorophore-analyte/Ifluorophore-control (Eq 3) Where Ifluorophore-analyte is the integration of the fluorophore emission in the presence of the analyte; and Ifluorophore-control is the ratio of the fluorophore emission in the absence of the analyte. Ratios greater than 1.05 were taken to represent cases of legitimate energy transfer. Ratios close to 1 indicated that no significant energy transfer was occurring, and that the existence of a fluorophore peak via analyte excitation was merely a result of the fluorophore having a non-zero absorbance at that particular wavelength. These control ratios were measured in both the oil layer and aqueous layer (full results are reported in the ESI).

RESULTS AND DISCUSSION
The two functions of this cyclodextrin-based system (extraction and energy transfer) will be discussed individually:  a All data represents an average of at least five trials b The low quantum yield of this analyte prevented accurate identification Table 1 highlights some significant differences in the ability of the 10 mM γ- The potential contributions of each of these factors are discussed in turn: 1a. γ-Cyclodextrin binding constants. Binding affinities of analytes 1-5 are shown in Table 2. The fact that all binding constants are similar (the largest value is only 1.3 times the smallest value) indicates that the differences in binding are unlikely to be responsible for the differential behavior of the analytes in the oil extraction experiments. 332 M -1 a The binding constant of benzo[a]pyrene in γ-cyclodextrin was not reported in the literature; attempts to calculate the binding constant directly using the Benesi-Hildebrand method were unsuccessful, likely due to a complex equilibrium between binary and ternary complexes.  For vacuum pump oil, vegetable oil, and cod liver oil, the enhancement factors for all analyte-oil combinations were much closer to 1, indicating limited contributions by γ-cyclodextrin to PAH extractions. These results contrast with a recent report that showed enhanced extraction efficiencies using hydroxypropyl-β-cyclodextrin to remove PAHs from contaminated soil. 38 The difference between these reported results and the relatively modest efficiencies reported herein is likely a result of the increased binding affinities of the PAHs in hydroxypropyl-β-cyclodextrin compared to their more modest affinities in γ-cyclodextrin (Table 2).
Interestingly, the motor oil-seawater series demonstrated different behavior than the motor oil-PBS series, with lower enhancement factors for all seawater cases (and enhancement factors less than 1 for analytes 1-3). The fact that the enhancement factors for analytes 4 and 5 are greater than 1 is likely a result of their increased solubility in water compared to compounds 1-3. Reasons for this atypical behavior in motor oil-seawater extractions may be related to the particular properties of the seawater, including the presence of surfactants and the high salt content.
(a) Surfactants: Sea water is known to contain high concentrations of surfactants. 39 These surfactants can form micelles that bind the PAH donor and the BODIPY acceptor in the hydrophobic interior, 40 thereby interfering with the ability of the cyclodextrin to form the necessary ternary complexes. In addition, surfactants contain a hydrophobic tail that can bind in the cyclodextrin cavity, forming an inclusion complex with the cyclodextrin that can hinder PAH binding.
(b) High salt concentration: The high salinity of sea water can also affect the ability of the cyclodextrin to form ternary complexes and promote energy transfer. 41 This complex formation is largely driven by hydrophobic binding, which is known to depend heavily on salt concentration. 42,43 Preliminary experiments using a phosphate buffer without saline (but under otherwise identical conditions) indicated that substantially more analyte was extracted into γ-cyclodextrin dissolved in phosphate buffer (saline-free) compared to γ-cyclodextrin dissolved in sea water (for example, the analyte comparison for pyrene is 0.34 in seawater compared to 0.75 in phosphate buffer). The high salinity of sea water is thus expected to lead to a further decrease in the hydrophobic binding necessary for cyclodextrin-promoted energy transfer.
1c. Solubility of analytes in oil and aqueous layers. The solubilities of PAHs 1-5 vary widely, with compounds 4 and 5 having markedly higher aqueous solubilities compared to compounds 1-3. 44 This increased solubility had no measurable effect on the observed enhancement factors for most extraction series (motor oil, vegetable oil, pump oil, and cod liver oil). However, the seawater-motor oil extractions demonstrated greater enhancement factors for analytes 4 and 5 compared to analytes 1-3. These results demonstrate that the solubility of the analytes can facilitate the cyclodextrin-promoted extraction and binding.

Energy transfer from
PAHs to fluorophore 6. The extraction of PAHs into the aqueous layer proceeded with moderate efficiencies in most cases. Even in cases of low extraction efficiencies, many of the analytes underwent efficient energy transfer to the highly fluorescent energy acceptor 6. The results are summarized in Table 3, and the results of energy transfer from a sample analyte (compound 2) to fluorophore 6 are shown in Figure 3.
18 ± 5% b 31 ± 4% b b a All data represents an average of at least five trials b No energy transfer was observed The efficient detection of benzo[a]pyrene 2 is particularly important due to its high toxicity and known carcinogenicity. 45,46 The results summarized in Figure 3 demonstrate that benzo[a]pyrene can participate efficiently in extraction and energy transfer across a broad range of complex oils. There are a number of other aspects of this energy transfer that merit discussion.   Figure 4D has been digitally altered to remove the double harmonic peak at twice the excitation wavelength; a copy of the unaltered spectrum is shown in the Supporting Information].
2c. Innate energy transfer from the oils. In addition to measuring energy transfer efficiencies with analyte-doped samples, the direct energy transfer of the undoped oils to fluorophore 6 was measured. These experiments were conducted by adding the fluorophore to the oil-water mixture (in the absence of the analyte), followed by separating the layers. Energy transfer efficiencies were measured in the oil layers by exciting the oil at both the analyte excitation wavelength and at the fluorophore excitation wavelength.
The results of these experiments are summarized in Table 4, and indicate some degree of energy transfer for all oils investigated. This energy transfer was most efficient for motor oil, vegetable oil, and pump oil (with 360 nm excitation), and least efficient for cod liver oil. This data is consistent with literature reports of some degree of PAH contamination in motor oil, vegetable oil, and pump oil, and no PAHs in cod liver oil, [29][30][31][32][33] and supports the idea that PAHs in the actual oils participate in cyclodextrin-promoted energy transfer.   Table 5.

SUMMARY
In summary, these experiments report the use of γ-cyclodextrin for two sequential functions: extraction of carcinogenic analytes from a variety of commercially available oils to an aqueous solution, followed by energy transfer from the analytes to a high quantum yield BODIPY fluorophore. The extraction of analytes into the aqueous layer proceeded with moderate efficiencies, depending on the particular analyte and oil investigated. Even in cases where the extraction efficiency was only modest, good to excellent energy transfer was observed from the newly extracted analyte to fluorophore 6. This multi-step system of extraction followed by efficient energy transfer can have significant applications in the development of turnon detection systems for oil-spill related carcinogens. Efforts towards this goal are in progress, and results will be reported in due course.

Materials and Methods
Vacuum pump oil (Fisherbrand19 mechanical pump fluid) was obtained from Fisher Chemical Company. Crisco pure vegetable oil, Pennzoil motor oil, and CVS-brand cod liver oil were obtained from local retailers. Seawater was obtained from the Narragansett Bay in Rhode Island. All PAHs were obtained from Sigma-Aldrich chemical company (Chart 1). BODIPY fluorophore 6 was synthesized following literature-reported procedures. UV-Visible spectra were recorded on an Agilent 8453 spectrometer. Fluorescence measurements were recorded on a Shimadzu RF 5301 spectrophotometer with slit widths of 1.5 nm excitation and 1.5 nm emission slit widths. All fluorescence spectra were integrated vs. wavenumber on the X-axis, using

Oil Extraction Experimental Details
Sample preparation: Samples for vegetable oil, vacuum pump oil, and cod liver oil were added to the oil-hexane mixture in a vial. The contents were vigorously shaken by hand for approximately 1 minute. 2.5 mL of γ-cyclodextrin (10 mM in aqueous solution (either PBS or Narragansett Bay sea water)) was added to the vial and the contents were once again shaken. The sample was allowed to sit undisturbed for [16][17][18][19][20][21][22][23][24] hours to ensure that the layers were fully separated. The aqueous layer was removed via pipette and placed in a new vial for analysis.
Control sample preparation: The same procedures were followed as above, but instead of using 10 mM γ-cyclodextrin, a PBS solution without γ-cyclodextrin, or a seawater solution without cyclodextrin, was added.

Energy Transfer Experimental Details
After the extraction experiments were performed, the oil layer and aqueous layer were recombined in a vial. 100 µL of fluorophore 6 (0.1 mg/mL in THF; final concentration = 4.75 µM) was added to the oil-water mixture, and the contents of the vial were vigorously shaken by hand for approximately 1 minute to ensure thorough mixing.
The layers were separated and each layer was excited at two different wavelengths: (a) the excitation wavelength of the PAH (see Table 1); and (b) 460 nm, which is the excitation wavelength necessary to excite fluorophore 6 directly.
The fluorophore emission was integrated with respect to wavenumber on the X-axis, and the energy transfer efficiencies were calculated as in Equation 2: Energy transfer efficiency = IDA/IA x 100% (Eq 2) Where IDA is the integration of the fluorophore from analyte excitation and IA is the integrated fluorophore emission from direct excitation.
Energy transfer efficiencies from the oil itself (without doping with a PAH analyte) were also conducted by following the above procedures precisely, except for eliminating the analyte. Each oil layer was excited at the analyte's excitation wavelength (but in the absence of the analyte) to determine the innate energy donor capabilities of the oil samples.
All experiments were repeated 5-6 times, and the values reported are averages of the results.

Summary Tables of all Energy Transfer Experiments
The energy transfer from analytes 1-5 to fluorophore 6 was quantified according to Where IDA is the integration of the fluorophore from analyte excitation and IA is the integrated fluorophore emission from direct excitation.
Energy transfer was measured in both the aqueous layer and oil layer for all samples.
Energy transfer was also measured from the oil to the fluorophore, without spiking the oil layer with a particular analyte (called "Energy transfer from the oil layer" in Table   S3, below). a -no energy transfer was observed.    seawater-motor oil fish oil pump oil vegetable oil motor oil compound 10 mM γ-CD 0 mM γ-CD 10 mM γ-CD 0 mM γ-CD 10 mM γ-CD 0 mM γ-CD 10 mM γ-CD 0 mM γ-CD 10 mM γ-CD 0 mM γ-CD

Summary Tables for Control Experiments
Control experiments were also performed, wherein the fluorophore in each layer was Control ratios between 0.95 and 1.05 are defined as "non-legitimate energy transfer," meaning that the fluorophore peak is relatively equivalent in the presence and absence of the analyte.
Control ratios greater than 1.05 represent cases of legitimate energy transfer.
Control ratios less than 0.95 represent cases where the fluorophore emission was quenched in the presence of the analyte.
Control ratios were measured for both the aqueous and oil layers for each sample.        skimming or booning of the oil, 6 burning oil on the surface of the water, 7 applying chemical dispersants to facilitate oil dispersion, 8 and introducing oil-eating bacteria for environmental bioremediation. 9 Many of these methods suffer from potentially serious drawbacks, including the environmental damage from oil burning, 10 the unknown toxicity of many dispersants, 11 and the long-term disruption to the ecosystem from the introduction of non-native oil-eating bacteria. 12 In recognition of these problems, newer environmentally-friendly cleanup methods have been developed by several research groups, including the synthesis of new hydrophobic materials, including thermally reduced graphene, a sponge, and porous materials. [13][14][15] We have developed a new approach for the cleanup of oil spills in marine environments that focuses on the removal of aromatic toxicants such as polycyclic aromatic hydrocarbons (PAHs). 16 The removal of PAHs is particularly important because many of these compounds are known carcinogens or pro-carcinogens, 17 including the Class I carcinogen benzo[a]pyrene (Chart 1, compound 3). 18 This approach uses commercially available, non-toxic γ-cyclodextrin to bind PAHs and extract them from complex oils. Following the extraction, the PAHs are detected using cyclodextrin-promoted energy transfer to a high quantum yield fluorophore (compound 4); analogous energy transfer has already been established as an efficient method for toxicant detection in multiple complex environments. [19][20][21][22] Other research groups have also reported the use of cyclodextrin derivatives to extract PAHs from complex environments, including from contaminated soil 23,24 and river sediments. 25 In practice, our approach uses cyclodextrin for the tandem extraction and detection of PAHs from contaminated samples by using the cyclodextrin as a filter.
For example, a contaminated water sample would be passed through the cyclodextrin filter. The efficiency of PAH removal can then be monitored by taking random samples from the filtered water sample and monitoring its fluorescence, where decreasing fluorescence indicates successful PAH extraction. This could also be done on sediment samples in accordance with previously published preparation methods. [23][24][25] A dual-function system such as this could greatly aid environmental clean-up efforts.
Previous research in our group focused on the use of γ-cyclodextrin for the extraction and detection of PAHs from motor oil, vegetable oil, and vacuum pump oil.
Shortcomings of this method included the moderate extraction efficiencies observed using γ-cyclodextrin, as well as the use of commercially available oils rather than oils that had been collected from contaminated marine environments. Oil collected from oil spills (termed "oil spill oil") is more complex than the commercially available oils previously investigated, with a broad distribution of alkanes, aromatic compounds, and insoluble polymeric components. 26,27 These oils also contain many oxidized PAH derivatives as a result of the exposure of the oil to oxygen-rich environments. 28 Some crude oil spontaneously forms tar balls, which are oil-containing spheres formed from both oil spills as well as from naturally occurring oil sources. 29 The degradation and oxidation of toxicants in tar balls has been shown to differ from that of toxicants found in bulk oil samples. 30 Reported herein is the use of a wide variety of cyclodextrin derivatives (αcyclodextrin, β-cyclodextrin, methyl-β-cyclodextrin, 2-hydroxypropyl-β-cyclodextrin (2-HPCD), and γ-cyclodextrin) to extract and detect aromatic toxicants from motor oil, oil spill oil, and tar balls. The extraction and detection efficiencies depend both on the identity of the oil and on the cyclodextrin host. The aromatic small molecules extracted with cyclodextrin include highly toxic PAHs, polar oxidized PAH metabolites, and a variety of other toxicants that have been found in such complex matrices. 31 The ability of cyclodextrin to extract multiple classes of toxicants simultaneously provides a significant operational advantage in the environmental remediation of polluted marine environments. where IDA is the integration of the fluorophore emission from analyte excitation and IA is the integrated fluorophore emission from direct excitation.

RESULTS AND DISCUSSION
PAHs found in oil collected from environmental oil spills have undergone substantial oxidation to a variety of highly polar, oxidized products, including quinones, phenols, and other oxidized species. 33 Consistent with these reports, when the oil spill oil was mixed with an aqueous buffer solution (0 mM cyclodextrin), it demonstrated a high concentration of photophysically active compounds partitioning into the aqueous buffer solution ( Figure 1B). Water soluble photophysically active compounds extracted from oils are likely to be oxidized PAH metabolites or other water soluble aromatic moieties, a hypothesis that is supported by ample literature precedent. [34][35][36] In contrast, only a negligible concentration of photophysically active compounds partitioned from the motor oil into a cyclodextrin-free aqueous layer, reflecting the lower degree of polar fluorescent metabolites found in that oil ( Figure   1A). The oil-water partitioning of tar balls was intermediate between the oil spill oil and the motor oil, with 46% of the overall fluorescence found in the aqueous buffer layer ( Figure 1C). The differential behavior of tar balls compared to oil spill oil can be explained by the different composition of the tar balls -they are enriched in heavier components, such as asphaltenes, that are insoluble in water. 37,38 The PAHs found in the tar ball's interior are also somewhat protected from oxidation due to their limited interaction with the oxygen-rich environment, whereas the PAHs in oil spill oil are more susceptible to oxidation. 39  motor oil-buffer solutions, the addition of γ-cyclodextrin and 2-HPCD led to a substantial increase in the amount of photophysically active compounds extracted into the aqueous layer (from 24.0% in PBS to 33.6% and 34% for 2-HPCD and γcyclodextrin respectively), which is consistent with the known ability of these cyclodextrins to bind PAHs. Other cyclodextrin derivatives, including β-cyclodextrin, methyl-β-cyclodextrin, and α-cyclodextrin, have cavity sizes that are too small to bind many PAHs, and their addition had no effect on the oil-water fluorescence ratios (Table 1).   16 However, the addition of the smaller cyclodextrins also led to an increase in the percentage of fluorescence found in the aqueous layer, even though such cyclodextrins lack sufficient steric bulk to encapsulate PAHs in their hydrophobic cavities. These cyclodextrins are likely effecting fluorescence increases by binding polar PAH analytes via hydrogen bond formation; 43 this hydrogen bonding allows analytes that are too large to bind in the cyclodextrin interior to associate with the cyclodextrins, thereby enabling enhanced extraction into the aqueous layer.
Following the efficient extraction of PAHs from a variety of complex oils using cyclodextrin derivatives, the ability of the newly extracted PAHs to participate in cyclodextrin-promoted energy transfer in the aqueous layer was assayed. This energy transfer requires that fluorophore 4 partition efficiently into the aqueous layer.
The percentage of fluorophore emission in the aqueous layer was measured for all oilcyclodextrin combinations, and found to be particularly efficient for methyl-βcyclodextrin containing solutions (Figure 3). This high efficiency points to a high degree of steric and electronic compatibility between methyl-β-cyclodextrin and fluorophore 4. Notably, some degree of fluorescence emission from fluorophore 4 was found in the aqueous layer for all oil-cyclodextrin combinations, indicating the potential for efficient energy transfer in all cases.  For oil spill oil, the observed energy transfer efficiency with undoped samples in the absence of any cyclodextrin was fairly high, and the addition of β-cyclodextrin and methyl-β-cyclodextrin led to decreases in the observed energy transfer efficiencies (energy transfer efficiencies of 30% and 24% for β-cyclodextrin and methyl-βcyclodextrin, respectively, compared to 50% in the absence of any cyclodextrin) ( Table 2). The addition of larger cyclodextrins (i.e. 2-HPCD and γ-cyclodextrin) caused a substantial enhancement in the observed affinities. The large degree of cyclodextrin-free energy transfer is consistent with our previously reported results that showed cyclodextrin-free association in many complex environments. 19 In these aqueous extracts, PAH metabolites likely associate with fluorophore 4 via a combination of hydrophobic binding (between the aromatic portions of the metabolites and the aromatic moieties of the fluorophore) and hydrogen bonding (between the hydroxyl and carbonyl moieties of the metabolites and the thiol and charged portions of the fluorophore); this close association is responsible for the observed cyclodextrinfree energy transfer.
For oil collected from tar balls, a modest energy transfer efficiency in the cyclodextrin-free solution was observed in undoped samples, and this efficiency was somewhat enhanced by the addition of most cyclodextrin derivatives by 8-10 percentage points (Table 2) In aqueous extracts from motor oil, the degree of cyclodextrin-free energy transfer varied depending on the identity of the doped analyte, with analytes 2 and 3 demonstrating substantially higher degrees of cyclodextrin-free energy transfer compared to analyte 1. This is consistent with our previously reported results that demonstrated that analytes with large hydrophobic surface areas are most likely to engage in cyclodextrin-free association and cyclodextrin-independent energy transfer. 19 The energy transfer efficiencies were most improved by the addition of 2-HPCD and γ-cyclodextrin, with 73% and 74% efficiencies observed using γcyclodextrin and 2-HPCD, respectively. These results are consistent with the known ability of these cyclodextrins to form ternary complexes that promote proximityinduced energy transfer. 48 The results in Table 2 highlight the ability of cyclodextrin to remove aromatic toxicants from both oil spill oil and tar ball oil. These experiments, conducted without doping a particular PAH into the complex mixture, involve the cyclodextrins extracting a wide range of toxicants from the complex oils, including PAHs, PAH metabolites, and other aromatic moieties. Overall, the results reported herein highlight the potential of cyclodextrin derivatives to promote the efficient extraction of smallmolecule toxicants from oil spills, as well as their subsequent detection via energy transfer to a high quantum yield fluorophore. This system has a number of notable advantages, including: (1) In contrast to our previously reported results that demonstrated modest extraction efficiencies using γ-cyclodextrin to extract PAHs from motor oil, vegetable oil, and vacuum pump oil, we report herein substantially improved extraction efficiencies using a variety of cyclodextrin derivatives to extract aromatic toxicants from oil spill oil and tar ball oil, with up to 72% of the aromatic toxicants found in the cyclodextrincontaining aqueous layer, compared to our previously reported best of 34% aromatic analytes in γ-cyclodextrin-containing aqueous layer extracted from motor oil. Oil collected directly from oil spill sites and oil isolated from tar balls have different physicochemical profiles compared to motor oil, vegetable oil, and vacuum pump oil, as a result of the weathering process that promotes substantial oxidation of the aromatic toxicants. 5 Environmental remediation of oil spill oil and tar ball oil from polluted marine environments is substantially more relevant for environmental disaster efforts than the remediation of commercially available oils, and the results reported herein indicate that using a variety of cyclodextrin derivatives enables the efficient extraction of toxicants from these complex oils.
(2) The cyclodextrin-based extraction followed by detection system reported herein provides a rapid method to remove toxicants from oil spills and to confirm that photophysically active analytes were removed via fluorescence energy transfer, which is a useful tool in disaster response efforts. In many oil spill situations, the precise identification of each toxicants is less crucial than the ability to remove as many toxicants as possible as quickly as possible and confirm such removal. Using cyclodextrin derivatives to enhance the extraction of photophysically active compounds from the oil layer to the aqueous layer, as demonstrated herein, provides a practical method for such environmental detoxification, and monitoring the overall fluorescence of the extracted analytes provides a rapid method to assay the efficacy of such detoxification procedures.

CONCLUSION
In conclusion, the results reported herein demonstrate that cyclodextrin-based systems can be used for the efficient extraction and detection of aromatic toxicants from real-world oil samples collected at the sites of oil spills. The system uses a number of commercially-available, non-toxic cyclodextrin derivatives to optimize extraction and detection procedures for each oil sample investigated, and demonstrate that our previously-reported results are generally applicable for the cleanup of oilcontaminated marine environments. These results also pointed to the potential of using multiple cyclodextrins simultaneously for the cleanup of a single oil system, with the cyclodextrins that are optimal for extraction of PAHs, binding of the fluorophore, and promotion of efficient energy transfer combined into a single high-performing, multicyclodextrin system. Research in this direction is currently underway in our group, and the results to date support this idea. The full results will be reported in due course.

Funding Sources
This research was supported by a grant from the Gulf of Mexico Research Initiative. Then, ~5 mL of hexanes was added and the tar balls were mixed once more. Next, the solution was placed in a dialysis bag and placed in a beaker with ~400 mL n-octane.

Efficient Extraction and Detection of Aromatic
The sample was allowed to dialyze for 3 days until the octane turned brown in color. where Ifluorophore-analyte is the integration of the fluorophore emission in the presence of the analyte and Ifluorophore-control is the ratio of the fluorophore emission in the absence of the analyte.

Analyte comparisons
All analyte comparisons were calculated according to Equation 1. The results represent an average of at least 3 trials.

Fluorophore comparisons
All fluorophore comparisons were calculated according to the following equation: where Iaq is the integrated fluorescence emission of the fluorophore in the aqueous layer from 460 nm excitation, and Ioil is the integrated fluorescence emission of the fluorophore in the oil layer from 460 nm excitation.
All results represent an average of at least 3 trials.

Energy transfer in the aqueous layer
Energy transfer efficiencies in the aqueous extracts were quantified according to Equation 2. All results represent an average of at least 3 trials.

Control ratios for aqueous extracts
Control ratios for all aqueous extracts were calculated according to Equation 3. All results represent an average of at least 3 trials.

INTRODUCTION
The accurate detection of small-molecule organic toxicants in complex environments has significant implications for public health. Such toxicants are potentially significant contributors to human disease, 1-3 and are found in food supplies, 4-6 water supplies, 7 and in commercial products. 8 Current methods for the detection of these chemical toxicants generally require multiple steps: (a) extraction of the toxicants from the environment; 9 (b) purification of the toxicants via highperformance liquid chromatography 10 or gas chromatography; 11 and (c) detection of the toxicants by mass spectrometry 12 or fluorescence spectroscopy. 13 Such detection methods are limited in their ability to distinguish toxicants with identical molecular weights or similar fluorescence spectra.
Small-molecule toxicants can also be detected through fluorescence energy transfer-based methods. Such fluorescence energy transfer, which has been used extensively for biomolecule detection, [14][15][16] often requires significant spectral overlap between the emission spectrum of the donor and the absorption spectrum of the acceptor to achieve efficient energy transfer (i.e. a Förster-type mechanism). 17 This overlap ultimately compromises the sensitivity of the system, as even in the absence of the target analyte there is residual donor emission. 18 Efficient energy transfer that is independent of the spectral overlap (i.e. a Dexter-type mechanism) has the potential to lead to improved sensitivities in fluorescent detection schemes. 19,20 Reported herein is a highly efficient, practical approach for small-molecule detection: using the small molecules directly as energy donors in a non-covalent, macrocycle-promoted energy transfer scheme. 21 In such a scheme, both the toxicant and the fluorophore are bound in the interior of γ-cyclodextrin (Figure 1). The enforced proximity of the two molecules allows for non-covalent energy transfer to occur, with excitation of the toxicant (energy donor) resulting in energy transfer to and emission from the fluorophore (energy acceptor). The energy transfer is independent of the spectral overlap between the donor and the acceptor, and has the potential to lead to improved sensitivities in turn-on detection schemes.  We recently reported that cyclodextrin-promoted energy transfer occurred from polycyclic aromatic hydrocarbons (PAHs) (compounds 1-5, Figure 2) and polychlorinated biphenyls (PCBs) (compounds 14-19, Figure 2) to three fluorophores (two of which are shown in Figure 3). 22 Figure 3. Fluorophores investigated as energy acceptors.
Reported herein is a substantial expansion of this preliminary report to include

RESULTS AND DISCUSSION
The full chart of examined energy donors is shown in Figure 2. This chart contains several compounds that have been classified as known carcinogens (Group 1) according to the International Agency for Research on Cancer (IARC) (compounds 3, 6-10), 28 as well as a variety of other toxicants. [29][30][31][32] These structures also contain a wide variety of functional groups, steric bulk, and photophysical properties, which allows us to probe the donor features necessary for efficient energy transfer.
Energy transfer experiments were conducted by mixing the analyte and fluorophore in a 10 mM γ-cyclodextrin solution in phosphate-buffered saline (PBS), coconut water, seawater, human plasma, or human breast milk. The resulting solution was excited near the analyte's absorption maximum (defined as "analyte excitation") and near the fluorophore's absorption maximum (defined as "fluorophore excitation").
The energy transfer efficiencies were calculated according to Equation 1: where IDA is defined as the integrated fluorophore emission from indirect excitation and IA is the integrated fluorophore emission from direct excitation. A graphical depiction of IDA and IA is shown in Figure 4.

In Phosphate-Buffered Saline (PBS):
The energy transfer from analytes 7, 8, 11 and 12 to BODIPY 20 in 10 mM γcyclodextrin in PBS was exceptionally efficient, with greater than 100% efficiencies observed in all cases ( Figure 5). Control experiments with 0 mM γ-cyclodextrin in PBS showed substantially less energy transfer than the 10 mM γ-cyclodextrin solution (Table 1), highlighting the beneficial role of γ-cyclodextrin in promoting energy transfer. In coconut water: The composition of coconut water is remarkably similar to that of human plasma, and it has been used as a plasma surrogate during emergencies. 36 (Table 2), albeit with diminished efficiencies compared to energy transfer in pure PBS. Note: CD, γ-cyclodextrin a Fluorophore 20 used as the energy acceptor in all cases

In biological media
The ability to achieve cyclodextrin-promoted energy transfer in biological media can provide significant benefit for the detection of toxicants. Efficient energy transfer from compounds 7, 8, 11 and 12 to fluorophore 20 occurred in both human plasma samples and human breast milk samples that were doped with 10 mM γcyclodextrin (Table 2).

Energy transfer in seawater:
The

Comparison to Published Methods:
The ability to detect toxicants via non-covalent energy transfer has a number of advantages compared to previously-reported methods, including the ability to tune the emission signal of a single analyte throughout the spectral region through choosing a variety of fluorophores. To achieve this "tuning" ability, preliminary experiments were conducted using a third fluorophore: commercially available coumarin 6 (compound 22) as a fluorescent energy acceptor with selected analytes (10 mM γcyclodextrin, PBS solution) as energy donors. Good energy transfer efficiencies were observed for many cases (Table 3), and in most cases the energy transfer efficiencies were substantially higher in the presence of γ-cyclodextrin compared to in its absence.  Comparing the results obtained with compound 8 to those of compound 10 (which was relatively inefficient as an energy donor) highlight possible steric constraints (compound 10 is substantially larger than compound 8) and functional group requirements (compound 10 lacks the free hydroxyl moieties) that are necessary for cyclodextrin-promoted energy transfer.

CONCLUSION
In conclusion, highly efficient energy transfer from a variety of organic toxicants occurred to multiple fluorophore acceptors when bound in the cavity of γcyclodextrin. The fact that this approach is successful in many environments with a variety of analytes is very beneficial. The robust nature of this approach leaves a wide range of opportunities to expand the scope of the analytes that can be detected, as well as the environments that they can be detected in. Indeed, the only requirement is that the analyte be (at least) weakly fluorescent. Furthermore, sample preparation is simple compared to current methods, as most media simply require dilution with PBS.
The fact that γ-cyclodextrin can bind analytes within its cavity in complex environments means that it can simultaneously isolate the analytes and promote energy transfer so that the analytes can be reliably identified. This method is a significant contribution to the facile and reliable detection of toxic analytes. The ability to tune the emission signal for a particular analyte by varying the choice of fluorophore provides substantial flexibility, and can be used in the development of array-based detection schemes. The development of such an array is currently under investigation, and results of these and other experiments will be reported in due course.

EXPERIMENTAL SECTION
All chemicals were obtained from Sigma-Aldrich chemical company or Fisher Scientific and used as received. BODIPY fluorophore 20 was synthesized following literature-reported procedures. 65 Human plasma was obtained from Innovative Technologies. Human breast milk was obtained from an anonymous donor. Seawater was obtained from the Narragansett Beach in Rhode Island. Coconut water (VitaCoco 100% Pure Coconut Water) was obtained from CVS Pharmacy.
The human plasma, seawater, and coconut water were used as received. The

DETAILS FOR ENERGY TRANSFER EXPERIMENTS
All energy transfer efficiencies were calculated using Equation 1: where IDA is the integrated emission of the fluorophore from analyte excitation and ID is the integrated fluorophore emission from direct fluorophore excitation.
All fluorescence emissions were integrated using Origin 8.5, and were integrated vs.
wavenumber on the X-axis.
All analytes were dissolved at a concentration of 1 mg/mL in tetrahydrofuran (THF).
Fluorophore solutions were made as follows:  Table S1. For each combination, two fluorescence spectra were recorded: the fluorescence from excitation of the analyte and the fluorescence spectra from excitation of the fluorophore. The excitation wavelengths were chosen to be as close as possible to the maximum wavelength of absorption, without significantly truncating the emission spectrum. Excitation wavelengths are recorded below in   For each analyte, a control sample was also analyzed following the procedure outlined above, with 0 mM γ-cyclodextrin (pure PBS) in place of 10 mM γ-cyclodextrin.

EXPERIMENTAL DETAILS FOR BREAST MILK PREPARATION
Breast milk was collected from a single donor and frozen until used. The breast milk was allowed to sit in a warm water bath at 30°C until thawed. Then, the breast milk was cooled to room temperature and allowed to sit at room temperature overnight. The sample separated into a clear aqueous layer and an opaque layer with solids. The aqueous layer was carefully removed via pipette. The aqueous layer was then filtered via syringe and centrifuged for 15 minutes at 6500 rpm. The aqueous layer was then removed via pipette as some solids remained on the outside of the centrifuge tube.
For each trial, 625 μL of breast milk was added to 1.875 mL (for a total volume of 2.5 mL) of PBS (0 mM γ-cyclodextrin) or 10 mM γ-cyclodextrin, depending on the experiment.     10 mM γ-CD 0 mM γ-CD

Introduction
Cyclodextrins are widely-used supramolecular hosts, as their hydrophobic interiors and hydrophilic exteriors allow them to form inclusion complexes with a variety of small molecule guests. 1 The non-covalent interactions that promote guest: host complex formation include π -π stacking, 2 Van der Waals forces, 3 hydrophobic binding, and electrostatic interactions. 4 The binding affinities of small molecules in cyclodextrin cavities and the overall stability of the resulting inclusion complexes are determined by the electronic and steric character of the guest molecule. 5 The mechanisms that govern association complex formation are exceedingly complex and often difficult to predict and fully characterize, and numerous investigations studying complex formation in cyclodextrin hosts have been reported in the literature. 6 Due to the large scope of analytes that we have investigated (Chart 1), with a large variety of steric, electronic, and structural features, it is likely that the structures of the cyclodextrin-based host-guest complexes vary significantly. In some cases, the small molecules may associate near the cyclodextrin host rather than in the host cavity; in these cases, efficient proximity-induced energy transfer is still a likely outcome due to the enforced proximity between the donor and acceptor. 9 The mechanisms that govern the formation of inclusion complexes (wherein the small molecule is bound in the cyclodextrin cavity) or association complexes (wherein the small molecule is held near the cyclodextrin cavity) had not previously been explored, despite the fact that such mechanisms are expected to vary widely depending on the structural features of the small molecules, the nature of the host and fluorophore, and the experimental conditions (including solvent and temperature). We report herein an investigation into precisely this objective.

Experimental Section
All chemicals were obtained from Sigma-Aldrich chemical company or Fisher Scientific and used as received, including compounds 1-30 (Chart 1), 32 and 33.
Compound 31 (Chart 2) was synthesized following literature-reported procedures. 10 Fluorescence measurements were recorded on a Shimadzu RF 5301 spectrophotometer with slit widths of 1.5 nm excitation and 1.5 nm emission slit widths. All fluorescence spectra were integrated vs. wavenumber on the X-axis, using OriginPro Version 8.6.
Ultrapure water was collected an 18MΩ·cm Millipore Synergy UV. For the temperature studies, a Fisher Scientific Isotemp 6200 R20 was used to control the temperature and the spectrophotometer was equipped with a single constanttemperature cell holder. where IDA is defined as the integrated fluorophore emission from indirect excitation and IA is the integrated fluorophore emission from direct excitation.
A control experiment was also performed to ensure that the desired energy transfer was actually occurring, rather than being a result of non-zero absorbance of the fluorophore at the toxicant excitation wavelength, which would also lead to an apparent energy transfer peak. The ratio of these two emissions, shown as the "Control ratio" was calculated according to Equation 2: Control ratio = Ifluorophore-control/Ifluorophore-analyte Where Ifluorophore-analyte is the integration of the fluorophore emission in the presence of the analyte; and Ifluorophore-control is the integration of the fluorophore emission in the absence of the analyte. For ratios <0.95, legitimate energy transfer was occurring; for ratios between 0.95-1.05, the observed fluorescence response was the result of nonzero absorbance of the fluorophore at analyte excitation; and for ratios >1.05, fluorescence quenching was occurring.

Results and Discussion
The binding constants of 1-33 in gamma-cyclodextrin were determined using the Benesi-Hildebrand method 11 . Selected toxicant-fluorophore combinations were then subjected to further experimentation, which include varying the temperature of the system, varying the ionic strength of the solvent (through the addition of sodium chloride and guanidinium hydrochloride), and studying mixed aqueous-ethanol solvent systems. Each of these experiments will be discussed in turn.  Table 1, and R 2 values greater than 0.70 were considered to be reasonable linear fits. Several aspects of this data that merit discussion.
Non-covalent macrocycle complexes arise from binding affinity between the host and guest, and the contributions of hydrophobic interactions, hydrogen bonding, steric interactions, and electrostatic complementarity between the guest and the cyclodextrin host dictates the strength of this affinity. 12 In addition, high-energy water, resulting from unfavorable interactions between water and the hydrophobic cyclodextrin interior, occupy the cyclodextrin cavity. 13 Inclusion of the guest depends often depends on the capability of the guest to displace this water, which provides an important driving force for complexation. 14 In general, inclusion complexation is hindered by (1)  Lastly, using a Benesi-Hildebrand plot assumes that the fluorescence of the analyte increases with increasing cyclodextrin concentration, which is due to the decrease in radiative decay pathways available to the analyte to relax down to the ground state by greatly hindering its degrees of rotation. Therefore, it is understood that many of the analytes used in this study will not form classical inclusion complexes; rather, the analytes and fluorophores form association complexes with γcyclodextrin by using it as a scaffold 16 through hydrogen bonding and π-π stacking.
This explains the poor linear fit for some of the analytes.
The negative binding constants shown in Table 1  There are a number of potential complexes that can be formed between the small molecule analytes and gamma cyclodextrin, including 1:1, 1:2, 2:1 and 2:2 guest-host complexes, and many of these stoichiometries often occur simultaneously.
Moreover, there are a number of potentially co-occurring geometries, including ones with the analyte fully inside the gamma cyclodextrin cavity and those in which the analyte is associated outside of the cavity. As such, it is not surprising that many of the analytes do not show strong linear relationships, as the Benesi-Hildebrand plots are predicated on certain assumptions, including complete inclusion of the guest by the host, and that the concentration of the guest in the matrix is equivalent to the total guest concentration. These assumptions do not hold for all analytes, and can make accurately determining a binding constant difficult. 19 Attempts were made to fit the data to a 1:2 guest: host complex (see Table S2 in the Supporting Information), but overall the fits were stronger for 1:1 complexation.
Taken together, this data indicates that the mechanisms behind such a dynamic guest: host system is challenging to fully understand, and why definitive binding constants are difficult to obtain for a particular guest-host system.  Table 2. Control ratios confirm that these energy transfer efficiencies are a result of legitimate energy transfer rather than a result of exciting the fluorophore at a wavelength where it has non-zero absorbance (Table 3).
For each analyte, energy transfer efficiencies decreased with increasing temperature. This is likely due to hydrogen bond disruption, which decreases the stability of the complex and in turn decreases energy transfer efficiency. In general, host-guest inclusion complexes are less stable with increased temperature. 20 The results of Table 2 add further evidence to this observation. Compounds 28 and 29, which differ only in an additional amine group in 28, show the greatest decrease in energy transfer, with ~70% reduction in energy transfer efficiency between 5°C and 80°C. These compounds do not have a large hydrophobic surface area compared to other analytes (for example, compound 6); therefore, these analytes most likely rely on hydrogen bonding for complexation and therefore show the greatest sensitivity to temperature variation.
Compound 12 differs from 11 in that it has an amine group. These compounds demonstrate similar reductions in energy transfer efficiency between 5°C and 80°C (12: ~59%; 11: ~58%). Thus they share similar complexation dynamics, and the hydrogen bonding site offered by 12 hampers only the energy transfer efficiency. The complexes for both become less stable at higher temperatures, but hydrogen bonding is not a major contributor. Complexes with these compounds most likely rely on hydrophobic interactions, and previous work indicates that hydrophobic interactions are largely unhindered by an increase in temperature. 21 The fact that the stability of inclusion complexes decreases with increasing temperature explains the decreased energy transfer efficiencies.
Interestingly, the energy transfer efficiency of 11 is an order of magnitude greater than 12 (for example, 370% for 11 and 37% for 12 at 5°C), and 6 is 1-2 orders of magnitude higher than 8 (for example, 1474% for 6 and 99% for 8 at 5°C) could be due to the greater capability of 6 and 11 to π-π stack with 31, resulting in more efficient proximity-induced energy transfer. In other words, these compounds are held in closer proximity to the fluorophore than 8 and 12. This shows that 6 and 11 penetrate the cyclodextrin cavity to a greater extent (6 can be fully encapsulated), so their inclusion complexes are stable and rigid, allowing 31 to have maximum contact with these structures and resulting in higher energy transfer efficiencies. Lastly, the fact that the energy transfer efficiencies are significantly higher at 5°C than 80°C is particularly significant to our work. Many of the analytes shown in Chart 1 are weakly fluorescent, and by extension participate only weakly in noncovalent energy transfer. Detection experiments are carried out at room temperature, but by simply changing the temperature (which is a facile adjustment) the sensitivity of our method can be greatly enhanced, resulting in a wider range of analytes that can be detected by this method and improved sensitivity in detection.
Effect of salt addition. Salts have been known to influence inclusion complexation by a variety of pathways, which include hydrogen bond disruption, analyte association, and ternary complex formation. 22 They also influence the hydrophobic effect, which is the propensity of nonpolar molecules to aggregate in aqueous solution to exclude water as much as possible. Salts such as sodium chloride tend to increase the hydrophobic effect (as they make it difficult for molecules to move into the bulk water), while salts like guanidinium chloride tend to decrease the hydrophobic effect (as they make it easier for molecules to move into the bulk water), and were thus used for these studies. The results of energy transfer experiments with these salts are indicated in Table 4. Interestingly, only compound 6, and to a much less extent compound 11 and 29, showed any real difference in energy transfer efficiency, including in pure water with no salt content. The remaining analytes that showed real energy transfer (  Compound 6 showed a substantial increase in energy transfer efficiency in guanidinium chloride. This salt increases the solubility of nonpolar molecules (known as "salting in") in water by decreasing surface tension, allowing the nonpolar molecules to move into the bulk water more easily. The loss of hydrophobicity greatly increases the energy transfer efficiency from 320% in water to 514% in this salt, yet there is no difference between the control ratios (0.08 and 0.07, respectively). The fact that the energy transfer efficiency is much stronger in the presence of this salt is interesting. Previous research suggests that guanidinium chloride does not have an effect on the structure of water nor does it bind to cyclodextrin, but it does bind to the hydrophobic surface of the guest molecule, and stabilizes the analyte in water. 23 This is happening concurrently to the fluorophore 31, which could explain why the control ratios are essentially the same; in other words, the analyte and fluorophore are still forming complexes with γ-cyclodextrin. Sterics may also play a role in this; compound 6 shows strong size complementarity with the cyclodextrin cavity, and therefore has strong Van der Waals forces acting on it. This means that inclusion complex formation is highly favored for this analyte. However, the increased energy transfer efficiencies is likely a consequence of these molecules being able to better associate with one another. While they are associating similarly in space, they are able to do so such that γ-cyclodextrin holds them in much closer proximity to one another, which results in the dramatic increase of energy transfer efficiency. Effect of ethanol addition. Cyclodextrins have hydrophilic surfaces which are bonded to water, and this ordered structure can be disrupted upon addition of an alcohol. Table 6 reports the results of energy transfer experiments conducted in the absence and presence of ethanol (1:1 volume ratio with PBS), and Table 7 reports the control ratios for these experiments. The hydroxyl moiety of the alcohol form hydrogen bonds to the hydroxyl groups of the cyclodextrin cavity, and the hydrophobic portion of the alcohol enters the cavity. The result is an enhanced hydrophobic environment, and the hydrophobic effect is experienced strongly by small-molecule analytes. Furthermore, depending on the size of the analyte and the alcohol used, the alcohol may help the analyte fit more comfortably in the cyclodextrin cavity via formation of a ternary complex. 24 When looking at the control ratios in However, the efficiencies are modest. This could be due to the excess ethanol in solution: because hydrogen bonds are being disrupted, the analyte and fluorophore are being held in such a way that efficient energy transfer does not occur. Again, compound 6 is an interesting case. It decreases in efficiency from 324% without ethanol to 17% with ethanol, a substantial loss, while the control ratio is essentially unchanged (0.07 to 0.08, respectively). This result seems to support the fact that inclusion complex formation is highly favorable for this analyte, and ethanol simply disrupts efficient energy transfer while having no real effect on the inclusion complex formation.

Conclusion
Non-covalent interactions between a small-molecule guest and γ-cyclodextrin are exceedingly complex. To complement the cyclodextrin cavity, guests must possess at least some of the following characteristics: (1) favorable hydrophobic interactions; (2) attractive Van der Waals forces; (3) favorable thermodynamics for the expulsion of high-energy water; (4) favorable geometry of the guest; and (5) ability to form strong hydrogen bonds. However, even if the guest cannot be encapsulated by the cyclodextrin cavity, the guest can still form an association complex with γcyclodextrin that is capable of facilitating energy transfer or fluorescence modulation.
In this work we have explored the effect of temperature, salt addition, and ethanol addition to probe these factors. Because these analytes vary in their hydrophobic and hydrophilic structures, they are able to associate with γ-cyclodextrin by a variety of mechanisms, guided primarily by hydrogen bonding and the hydrophobic effect.
Interestingly, hydrogen bonding was found to have a leading role in complex formation over the hydrophobic effect, which is in contrast to our previous hypothesis that the hydrophobic effect would be dominant. There is much intricacy behind noncovalent interactions in general, and the interactions associated with complex formation in γ-cyclodextrin in particular. Fluorescence measurements were recorded on a Shimadzu RF 5301 spectrophotometer with slit widths of 1.5 nm excitation and 1.5 nm emission slit widths. All fluorescence spectra were integrated vs. wavenumber on the X-axis, using OriginPro Version 8.6.

Corresponding
Ultrapure water was collected an 18MΩ·cm Millipore Synergy UV. For the temperature studies, a Fisher Scientific Isotemp 6200 R20 was used to control the temperature and the spectrophotometer was equipped with a single constanttemperature cell holder.
General energy transfer procedure. All energy transfer experiments were conducted as follows: 2.5 mL of a 10 mM solution of γ-cyclodextrin dissolved in an aqueous solution (see below) was measured into a cuvette. 20 µL of the analyte (1 mg/mL) and 100 µL of the fluorophore (0.1 mg/mL) were added. After thorough mixing, the solution was excited at two wavelengths: near the analyte's absorption maximum (defined as "analyte excitation") and near the fluorophore's absorption maximum (defined as "fluorophore excitation"). See Table S1 for these values. Three repeat measurements were taken at each wavelength. The energy transfer efficiencies were calculated according to Equation 1: where IDA is defined as the integrated fluorophore emission from indirect excitation and IA is the integrated fluorophore emission from direct excitation. Control ratio = Ifluorophore-control/Ifluorophore-analyte Where Ifluorophore-analyte is the integration of the fluorophore emission in the presence of the analyte; and Ifluorophore-control is the integration of the fluorophore emission in the 3. Temperature studies. A 10 mM γ-cyclodextrin solution was prepared in PBS.
Energy transfer experiments were then conducted using the above procedures at the following temperatures: 5°C, 20°C, 35°C, 50°C, 65°C, and 80°C. The temperature control system used indicated when the desired temperature was reached, and each sample was allowed to sit in the unit for ~1 minute before the fluorescence emission spectrum was collected. This was done to ensure the sample was at the correct temperature.   Table S2. Binding constants for a 1:2 guest: host complex.

ABSTRACT
In the wake of anthropogenic releases of toxicants (for example, polycyclic aromatic hydrocarbons (PAHs)), the rapid, sensitive, and selective detection of such environmental pollutants is of great importance to first responders. Many anthropogenic events affect the local population; therefore, the detection of both the parent compound and the numerous metabolites is essential to aid medical personnel in assessing an individuals' exposure to environmental pollutants. Reported herein is the successful development of such a system using a cyclodextrin host, wherein both the toxicant of interest and a high quantum yield fluorophore are bound within the cavity of the cyclodextrin, and detection occurs via energy transfer from the toxicant to the fluorophore. In this study, samples from a non-smoker and habitual smoker were used to assess differences in analyte response. Efficient energy transfer (and thus toxicant detection) was observed in all cases.

INTRODUCTION
The occurrence, prevalence, and persistence of polycyclic aromatic hydrocarbons (PAHs) in the environment is a significant public health concern, 1,2 as many of these compounds are known or suspected carcinogens, 3 mutagens, 4 and teratogens. Medium-sized PAHs such as compounds 2, 3, 4 and 7 (Chart 1) are of particular concern due to their high toxicity 5 (i.e. benzo[a]pyrene 4 is a Group 1 carcinogen). 6 PAHs are often formed from the incomplete combustion of petroleum, and have been found in environments surrounding the sites of major oil spills, 7-9 as well as in the blood, 10 breast milk, 11 and urine 12 of inhabitants living in affected areas.
Current methods for PAH detection in complex environments generally use chromatographic separation 13 (such as gas chromatography 14 or liquid chromatography 15 ) to separate complex mixtures of compounds, followed by detection of each compound by mass spectrometry. 16,17 While such methods are highly sensitive for individual environmental toxicants, they lack broad applicability for multiple classes of toxicants in multiple complex environments. 18 The requirement for timeconsuming and often costly separation procedures prior to accurate identification further limits the broad applicability of these approaches, as well as the potential development of high-throughput assays. Because such contaminated environments are often rapidly changing (i.e. in a fast-moving stream or at the site of a rapidly spreading oil spill), 19-20 the requirement for sensitive, selective, and rapid detection methods of PAHs and other environmental toxicants is crucial.
We recently developed a fundamentally new fluorescence-based approach for the detection of aromatic toxicants, by using the aromatic toxicants directly as energy donors in combination with a variety of high-quantum yield fluorophore acceptors. [21][22][23][24][25][26][27] In this approach, energy transfer from the toxicants to the fluorophores occurs when both are bound in the cavity of a cyclodextrin host, and the fluorophore emission via toxicant excitation is used as a highly sensitive and easily tunable read-out signal. This approach has proven to be general for multiple classes of aromatic toxicants in multiple complex environments, including seawater, coconut water, and human plasma and breast milk. 21 The utility of fluorescence-based detection of toxicants would be markedly enhanced by the ability to detect environmental toxicants in human urine, as such detection would enable medical professionals to rapidly collect samples and evaluate individuals' exposure. 28 Moreover, because many aromatic toxicants undergo rapid in vivo oxidation, 29 an ideal detection strategy would be able to detect highly polar oxygenated metabolites as well as unmodified non-polar toxicants. Reported herein is the successful development of such a detection strategy: the use of cyclodextrinpromoted, proximity-induced fluorescence energy transfer to detect a wide variety of aromatic toxicants and toxicant metabolites in human urine, and the ability of such a method to distinguish urine samples collected from a habitual smoker and from a nonsmoker. Urine samples were collected from two anonymous volunteers: one of whom is a habitual smoker (ca. 25 cigarettes/day) and one who has never smoked. The Where IDA is the integration of the fluorophore emission from analyte excitation and IA is the integrated fluorophore emission from direct excitation.
Small amounts of analytes 1-15 were also added to the urine-cyclodextrinfluorophore samples, and energy transfer efficiencies in the analyte-doped samples were calculated following the same procedures.
Limits of detection for each analyte-fluorophore combination were calculated following literature-reported procedures to construct calibration curves with the analyte concentration on the X-axis and the fluorophore emission via energy transfer on the Y-axis. 31 These curves were then used to determine the minimum analyte concentration necessary to elucidate a detectable and quantifiable fluorescence response.

RESULTS AND DISCUSSION
In the absence of any added analyte or fluorophore, the urine samples   a All reported results represent the average of 4 trials b Excessive overlap between the analyte and fluorophore emission prevented accurate integration  For several of the larger sized PAHs (i. e. 2, 3, 7) we observed significant cyclodextrinfree association, which requires hydrophobic association of those compounds with the aromatic fluorophores to promote efficient energy transfer.
In contrast to these previously targeted non-polar analytes, metabolites 10-13 are oxygenated and highly polar, and are formed in vivo from cytochrome P450mediated oxidation of PAHs. 40 The extensive oxygenation decreases the hydrophobic character of the analytes; 41 nonetheless, in all cases these PAH metabolites  The sensitivity of this detection method was determined by calculating the limits of detection (minimal sample concentration that will generate a distinguishable signal) and limits of quantification (minimal sample concentration that will generate a quantifiable signal) for all analyte-fluorophore combinations in the non-smoker urine samples, and selected examples are highlighted in Table 3. Both the aromatic toxicants and the oxidized toxicant metabolites can be quantified at micromolar concentrations; current efforts are focused on lowering these detection limits even further to provide optimal sensitivity in toxicant and toxicant metabolite detection.  Where IDA is the integration of the fluorophore emission from analyte excitation and IA is the integrated fluorophore emission from direct excitation.
Control experiments were also conducted in which the 10 mM γ-cyclodextrin solution was replaced with a 0 mM solution, and the same procedure was followed.

Control experiments:
Control experiments were also conducted in which the 10 mM γ-cyclodextrin solution was replaced with a 0 mM solution, and the same procedure was followed.
All experiments were repeated 4 times, and the values reported are averages of the results.

ANALYTE COMPARISON EXPERIMENTAL DETAILS
These experiments were designed to determine the emission of the fluorophores from To determine the limit of detection (LOD) and limit of quantification (LOQ), each fluorophore-analyte combination was examined in the following manner: 1. 2.5 mL of 10 mM γ-cyclodextrin in phosphate-buffered saline (PBS) was measured into a cuvette and 100 μL of a fluorophore solution in THF was added. The solution was excited at the analyte's excitation wavelength (see table of wavelengths below) and the fluorescence emission spectrum was recorded. Six repeat measurements were made for the fluorescence emission spectra.
2. 20 μL of a 1 mg/mL analyte solution in THF was added to the cuvette and the solution was again excited at the analyte excitation wavelength. Six repeat measurements were taken.

3.
Step 2 was repeated for 40 μL of analyte, 60 μL of analyte, 80 μL of analyte, and 100 μL of analyte. In each case, the solution was excited at the analyte excitation wavelength and the fluorescence emission spectrum was recorded four times.
4. All fluorescence emission spectra were integrated vs. wavenumber, and we generated calibration curves with the analyte concentration on the X-axis (in µM) and the integrated fluorophore emission on the Y-axis. The curve was then fitted to a straight line and an equation for the line was determined.
5. For each case, the fluorophore with γ-cyclodextrin (before any analyte was added) was also excited at the excitation wavelength for the analyte, and the fluorescence emission spectrum was recorded (as per step 1). These measurements are referred to as the "blank." 6. The limit of the blank is defined according to the following equation: Where m is the mean of the blank integrations and SD is the standard deviation.
7. The limit of the blank was then entered into the equation determined in step 4 (for the y value), and the corresponding X value was determined. This value provided the LOD in µM.
8. The limit of quantification (LOQ) was calculated in a similar way to the limit of detection. First, the limit of the blank for quantification was determined according to the following equation: LoBLOQ = mblank + 10(SDblank) This value was entered into the equation determined in step 4 (for the y value), and the corresponding X value was determined to be the limit of quantification in µM.

SUMMARY TABLES
Fluorescence of urine samples was determined in the absence of any additional analyte or fluorophore. These values were determined by exciting the urine samples at a variety of excitation wavelengths, and integrating the resulting fluorescence emission vs. wavenumber on the X-axis. The ratio of the non-smoker urine fluorescence emission to smoker urine fluorescence emission was calculated, and the results are summarized in the following table: