Bioanalytical Applications of Fluorescence Spectroscopy - Analytical

Sep 15, 1988 - Markus Sauer , Andreas Schulz , Stefan Seeger , Jürgen Wolfrum , Jutta Arden-Jacob , Gerhard Deltau , Karl H. Drexhage. Berichte der ...
0 downloads 0 Views 10MB Size
REPORT

Bioanalytical Applications of Fluorescence Spectroscopy

Frank V. Bright Department of Chemistry State University of New York at Buffalo Buffalo, NY 14214

Fluorescence spectroscopy is an extremely powerful analytical tool that has been used in areas ranging from biology to physics. In this REPORT, we will present applications of fluorescence spectroscopy to bioanalytical systems, including DNA and RNA determinations, enzymatic methods, protein-ligand interactions, fluoroimmunoassays, and fiber-optic-based sensing. Space limitations preclude mention of other relevant areas (e.g., fluorescence detection in liquid chromatography). The article is divided into the following sections: historical perspective, theory, instrumentation, advantages of fluorescence, applications, and the future. Historical perspective Analytical luminescence can be traced to the initial observations of the Spanish botanist and physician Nicolas Monardes. In 1565 Monardes noted that water contained in a cup made from a specific wood (Ligirium nephiticiem) exhibited an odd blue shimmer. During the next century, eminent sci0003-2700/88/A360-1031 /$01.50/0 © 1988 American Chemical Society

entists such as Robert Boyle also observed this "unknown" phenomenon. In 1833 Sir David Brewster recognized a red emission from green leaf extracts. Although he first attributed the color to some type of dispersive-scattering phenomenon, today we know this characteristic emission as fluorescence from chlorophyll. In 1845 Herschel recorded the first "fluorescence emission spectrum" of quinine. He also recognized that the sample was excited in the blue region of the spectrum more than in the green or red regions. However, it was not until 1852 that George Stokes determined that the emission from quinine was at a longer wavelength than the excitation, a difference that bears Stokes's name. In addition to coining the term fluorescence, Stokes was the first to propose the use of fluorescence as an analytical tool. In 1867 Goppelsroder was the first to perform a fluorescence-based analysis. He developed a method for the quantitation of nonluminescent Al(III) by forming a strongly fluorescent morin: Al(III) complex. Still further evidence of the power of luminescence was demonstrated in 1877, when Adolf Baeyer used fluorescence to demonstrate a link between

the Rhine and Danube rivers. He suggested that 10 kg of fluorescein be thrown into the Danube. Three days later, the characteristic green fluorescence of fluorescein was detected in the waters of the Rhine, demonstrating that these two great rivers were linked. This brief historical perspective should give the reader some idea of the colorful (no pun intended) past of fluorescence. The main events from 1565 to 1960 are highlighted in Table I. The interested reader can also find more information in References 1-3. Theory Under standard temperatures, more than 99% of molecular species are found to occupy the ground electronic state. Furthermore, these molecules can only absorb discrete quanta of electromagnetic radiation when an incident photon has an energy coincident with the energy difference between two allowed energy states. The net result of this photon-induced absorption is the promotion of a few molecules to a higher energy state. Of course, between each of the electronic states there are vibrational levels, and within each vibrational level there are still more rotational levels. This entire set of levels makes up the electronic-vibrational-

ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988 · 1031 A

Table I. Year

Milestones in fluorescence spectroscopy Observation or invention

Researcher(s)

1565 1680 1833 1845 1852 1864

Monardes Boyle Brewster Herschel Stokes Stokes

1867 1877 1920

Goppelsroder Baeyer Desha

1925 1928 1939 1955 1960

Bayle & Fabre Jette & West Zworykin & Rajchman Bowman Schawlow

First observation of fluorescence Environmental sensitivity Chlorophyll fluorescence Spectrum for quinine Concept of fluorescence Suggested the application of fluorescence to chemical analysis First fluorescence-based analysis (morin:AI(lll)) Link between Rhine and Danube (fluorescein) Determined that intensity is proportional to concentration Drug determinations and purity First photoelectronic fluorometer Photomultiplier tube developed First commercial fluorometer

I rotational manifold for a molecular species. In Figure 1, these energy levels are designated by subscripts 0 , 1 , and 2. In addition to the electronic levels (S and T), Figure 1 shows several vibration levels. For clarity, the rotational levels within each vibrational level are not shown. The ground electronic state is designated as the lowest singlet state (So), the first excited singlet state as Si, the second excited state as S2, the first excited triplet state as ΤΊ, and the sec­ ond excited triplet state as T2. The difference between the singlet and triplet states is a consequence of a

difference in the spin (S) of the elec­ tron. In all systems, electrons have spins of either —V2 or +V2. Molecular species in the ground state generally have an even number of electrons; at least two electrons are needed for each covalent bond. In addition, the spins of these electrons are paired; one electron of the pair has a spin of +V2 and the other has a spin of —lkThe orbital angular momentum of any given state is represented by the multiplicity (M), where M = 2S + 1 and S is the total spin for the state in question. Therefore, when all electrons are paired, the multiplicity equals 1, or

2(+V2 — V2) + 1, and when the spin of a single electron is reversed, the multi­ plicity becomes 3. States in which the multiplicity is 1 or 3 are referred to as singlet (Sn) or triplet (T„) states, re­ spectively. In general, the triplet state is "generated" only upon excitation into a singlet ( S J followed by intersystem crossing (Figure 1) to the triplet state. Intersystem crossing simply in­ volves the flipping of one of the elec­ tron spins so that unpaired electron spins result. From an energized state (S„ or T„), a molecule rapidly loses its photoninduced excess energy and relaxes back to the ground state. This relaxation generally takes approximately 1 ps and occurs predominately through vibra­ tional and rotational modes and colli­ sions with solvent molecules. In Figure 1, this is depicted by the dashed lines labeled radiationless deexcitation. However, in certain mole­ cules (e.g., rigid conjugated species), the losses caused by vibration and rota­ tion may be much slower. For these species, the excited state can also lose energy by photon emission. When this photon emission is from a triplet state, phosphorescence is the result. Con­ versely, when the emission is from a singlet electronic state, it is referred to as fluorescence. The quantum yield (Q) is a physical characteristic of every molecule and is essentially the ratio of photons emitted to photons absorbed. Quantum yield values range from 0 to 1; compounds with larger quantum yields exhibit stronger luminescence. The relation­ ship between luminescent intensity (/) and concentration is given by (4): I = QIQ[1 - exp(-ebc)]

(1)

where IQ is the incident power of the exciting beam, e is the molar absorptiv­ ity of the species under study, b is the sample cell path length, and c is the molar concentration of the fluorophore. For dilute solutions, Equation 1 reduces to / = KQI0ebc

Figure 1. Abbreviated schematic of an energy-level diagram. 1032 A · ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988

(2)

where Κ is an instrumental factor. Equation 2 predicts a linear relation­ ship between luminescence intensity and concentration. However, at higher concentrations significant deviations from linearity are often observed be­ cause of concentration quenching and self-absorption. All molecules absorb photons; how­ ever, only a few fluoresce and fewer still phosphoresce at ambient tempera­ tures. Thus the rarity of luminescent materials makes fluorescence an at­ tractive technique for the resolution of trace fluorescent components in com­ plex mixtures. Furthermore, with the advent of inexpensive laser sources, subnanogram amounts of fluorescent

Figure 2. Schematic representation of a basic fluorescence instrument. Solid lines indicate electrical connections; dashed lines represent optical paths.

materials can now be routinely deter­ mined. Instrumentation

Figure 2 shows a schematic of a gener­ alized fluorescence instrument. Essen­ tial components include an excitation source (Hg lamp or laser), a means of "monochromatizing" the excitation source (filter or monochromator), a thermostatically controlled sample chamber, an emission selector (filter or monochromator), a detector (diode ar­ ray, photomultiplier tube, or chargecoupled device), detection electronics (picoammeter or photon counter), and a readout device (strip chart recorder or computer). Inspection of Equation 2 reveals sev­ eral ways to increase the fluorescence signal, including increasing the collec­ tion efficiencies, molar absorptivity, quantum yield, and intensity of excita­ tion. Of these variables, the one that is most easily increased is the excitation intensity. Most commonly, this is achieved with laser-based excitation; however, one cannot simply increase the fluorescence signal without bound. For example, for weakly emitting sam­ ples, the water Raman scatter from aqueous samples can become so intense that it actually overwhelms fluores­ cence. Fortunately, techniques have been developed to alleviate these types of problems (5). In addition, at high power densities even normally stable fluorophores begin to photodecompose. The only general recourse to this decomposition problem is to lower the incident power or employ a flowing sample stream. Advantages of fluorescence When compared with conventional absorbance-based techniques, fluores­ cence is several orders of magnitude more sensitive and more selective. Again, the increase in sensitivity arises because the emitted radiation is mea­ sured directly (Equation 2) and can be

increased simply by increasing the inci­ dent power. Moreover, fluorescence is often considered a zero background technique, whereas absorbance is a measure of the difference between inci­ dent and transmitted intensities (i.e., signals are measured as a small change in a large background). Selectivity in fluorescence-based techniques is also much higher than in absorbance-based approaches. Until several years ago, this was attributed primarily to the relatively low number of fluorescent species compared with absorbing ones and the ability to em­ ploy both excitation and emission wavelengths as selectivity parameters. Recently, quite a few other selectivity parameters have been employed suc­ cessfully; these multidimensional lu­ minescence-based approaches have been reviewed recently (6). This multi­ dimensional nature of fluorescencebased techniques makes fluorescence spectroscopy an ideal technique for bioanalytical studies. Applications

DNA and RNA. Fluorometric meth­ ods of analysis are probably most often performed in the clinical laboratory. The goal of such clinical analysis is sim­ ply to determine the amount of a given species present in a biological sample (e.g., urine, liver tissue). Although all types of biomolecules are important, nucleic acids such as ribonucleic acid (RNA) and deoxyribonucleic acid (DNA) are especially important be­ cause they store, replicate, and trans­ mit genetic information. DNA and RNA are the biomolecules that deter­ mine what a cell is and what it does. Therefore assays directed at determin­ ing DNA and RNA clearly are of criti­ cal importance. Two fluorescent reagents, Hoechst 33258 and ethidium bromide, are com­ monly used for such assays. Hoechst 33258 is a fluorescent dye that interacts physically with DNA, yielding a dra­

1034 A · ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988

matic increase in fluorescence intensity (7). The degree of fluorescence en­ hancement (excitation wavelength is 355 nm, emission wavelength is 460 nm) depends on several factors, but the overall result is that adenine/thyminerich DNA yields the most intense fluo­ rescence whereas guanine/cytosinerich DNA and single-stranded DNA give progressively less fluorescence. Using this approach, one can deter­ mine DNA in tissue homogenate. For­ tunately, RNA does not interfere be­ cause it only gives a signal 0.25% of that of the DNA-Hoechst 33258 complex. Ethidium bromide (EB) is often used for the determination of RNA or DNA (8). Both nucleic acids form a complex with EB, resulting in an en­ hancement of the EB fluorescence. The excitation is usually at 530 nm, and the emission is monitored at 620 nm. To determine only RNA, the enzyme RNAase is added to one of a pair of samples and the samples are heated at 37 °C for 30 min. During heating, the RNAase destroys all the RNA, and from the difference in intensities of the two samples (DNA + RNA and DNA alone), one can easily determine the amount of RNA present (8). These two methods are designed to facilitate determination of a class of macromolecules, but they will also re­ spond to other proteins and nucleic ac­ ids. To determine a specific target analyte, especially in a tissue extract, a more selective enzyme- or antibodybased assay is generally employed. Enzymatic methods. Glucose deter­ minations are often based on enzyme assays. Specifically, glucose is convert­ ed to glucose-6-phosphate (G6P) by the enzyme hexokinase. In this conver­ sion, adenosine triphosphate (ATP), which has an extremely reactive phosphoanhydride linkage, acts as a phos­ phate donor. G6P is then oxidized by nicotinamide adenine dinucleotide phosphate (NADP + ) employing glucose-6-phosphate dehydrogenase to yield reduced NADPH. The NADPH is strongly fluorescent and is monitored using an excitation wavelength of 340 nm and an emission wavelength of 460 nm. Detection limits for glucose are in the 0.1-nmol range (9). Sucrose and fructose can also be easily quantitated by using the enzymes invertase and hexose phosphate isomerase, respec­ tively. Kidney function is often evaluated by correlation to serum urea nitrogen. A simple schematic of an enzymatieally coupled approach is shown below: urease*

urea

ηΛΤΤΤ

+

• 2NH4

NADH + 2NHJ" + a-ketoglutarate glutamine dehydrogenase^

N A D

+

g l u t a m k

ad(j

The rate of NADH disappearance is monitored fluorometrically using an excitation wavelength of 365 nm and an emission wavelength of 460 nm, and it is proportional to the concentration of urea in the sample (10). Protein-ligand interaction. The specific biochemistry of protein func­ tionality and specificity necessarily in­ volves the association (binding) of the protein with other chemical or biomolecular species (ligands). Because this initial binding step dictates the out­ come of a protein's functionality, this simple process has received a great deal of attention. For example, an under­

standing of the stoichiometry and af­ finity of protein-binding sites is of crit­ ical importance in elucidating protein function and dynamics. Fluorescence spectroscopy is one of the premier techniques used to elucidate such pro­ tein-ligand interactions. The types of proteins and nucleic acids studied, as well as the ligands employed, comprise a huge, often overwhelming body of lit­ erature. Fortunately, a well-compiled set of references is available in this area (Π). Fluoroimmunoassays. In a clinical setting, many of the species routinely determined (e.g., drugs, proteins) ei­

Simply incredible. Incredibly simple.

HPLC/GC DATA SYSTEM

Introducing the most incredible PC-based, multi-channel, multi­ tasking chromatography data system ever designed with you in mind. Powerful, yet simple to operate. All the capabilities you want....and more, including our unique, effort-free Auto-Integration for automatic selection of integration parameters. Edit and reprocess chromatograms from storage while processing new data from one, two or three chromatographs. Control up to three separate HPLC systems. Monitor up to six chromatograms at the same time on one master screen. Get the system that's simply incredible, and incredibly simple. Call today for complete information.

Axxiom CHROMATOGRAPHY

23966 Craftsman Road Calabasas, CA 91302-1489 Toll Free: 800-423-5008 In California: 818-346-1800

CIRCLE 1 ON READER SERVICE CARD 1036 A · ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988

ther do not exhibit native lumines­ cence or they exhibit it at an inconve­ nient wavelength. Thus it is necessary to develop techniques such as immuno­ assays for indirectly quantitating these species. In its simplest form, the immunoas­ say is based on the competitive binding between a highly selective antibody (Ab), an antigen (Ag), and an appropri­ ately labeled form of the antigen (Ag*): Ag + Ag* + Ab *»· AgAb + Ag*Ab Basically, the immunoassay exploits some difference between the Ag* and Ag*Ab species to indirectly quantify the target analyte, Ag. Early immuno­ assays employed radioisotopes (e.g., 125 I) as the label. These radioimmuno­ assays (RIAs) demonstrated excellent sensitivity; however, the shelf life for the common radioactive labels is short, and governmental control of radioac­ tive materials has forced investigators to search for alternative labels. Of the myriad alternative labeling schemes employed to date, fluorescent labels have probably enjoyed the most success. Fluoroimmunoassays (FIAs) can be either homogeneous (no physi­ cal separation of Ag* and Ag*Ab) or heterogeneous (requiring physical sep­ aration of Ag* and Ag* Ab). Of the two approaches, homogeneous immunoas­ says are preferred because they are simpler and faster and they consume fewer reagents. However, they are gen­ erally less sensitive and are prone to interference problems. Although they were expected to be much more sensitive than RIAs (re­ searchers using laser-based techniques have demonstrated single-molecule de­ tection limits for proteins labeled with several fluorescent labels [12]), the ear­ liest FIAs did not contend too well with RIAs because serum components inter­ fered with the fluorescence of the label. Fortunately, many new fluorescent la­ bels have been developed that are unaf­ fected by environmental changes. The problem of background luminescence, however, still plagues FIA, and its main shortcoming remains spectral interfer­ ences from the sample matrix. Because fluorescence is inherently a multidimensional technique, one can employ several selectivity parameters in concert to help alleviate these inter­ ference problems. Lovgren and co­ workers (13) have minimized matrix interference problems by employing rare earth chelates, such as Eu(III) and Tb(III), as labels. These are unique be­ cause they have relatively long lumi­ nescence lifetimes (10-100 μβ). There­ fore, with simple time-gating they can eliminate the short-lived (~10 ns) background fluorescence from the longer lived label fluorescence. In the case of fiber-optic-based sensing (pre­ sented below), the associated back-

P=(/|,-/ ± )/(/|, + / J

Figure 3. Representation of polarization measurements. Excitation is along the χ axis and fluorescence is detected at 90°, along the y axis.

grounds can be minimized by using these rare earth chelates (14). Fluorescein (first synthesized by Baeyer in 1871) is the most common fluorescent label used in FIAs because of its high quantum yield and the avail­ ability of analogs with highly reactive functional groups. Unfortunately, the quantitation of fluorescein in serumcontaining samples is diminished by the presence of bilirubin, which has ex­ citation and emission spectra that com­ pletely overlap the fluorescein spec­ trum. Conventional wavelength selec­ tivity does not provide the necessary resolution between bilirubin and fluo­ rescein. Fortunately, however, the fluo­ rescence lifetimes of bilirubin (100 ps) and fluorescein (4.0 ns) differ signifi­ cantly. Using phase-resolved fluores­ cence spectroscopy, Bright and McGown (15, 16) were able to quantitate fluorescein at the nanomolar level in the presence of up to 7.5 μΜ biliru­ bin. Hieftje and co-workers (17) ex­ tended this approach by employing picosecond laser excitation, and they were able to routinely determine nano­ molar levels of fluorescein in the pres­ ence of up to 83 μΜ bilirubin. The first homogeneous FIA based solely on fluorescence lifetime selectiv­

ity was described by Bright and McGown in 1985 (18). A sensitive im­ munoassay for phénobarbital was developed in which free fluoresceinlabeled phénobarbital (Ag*) and antiphenobarbital-bound Ag* (Ag*Ab) exhibited a difference of 100 ps in fluorescence lifetime. Again, by employing phase-resolved fluorescence spectroscopy, they were able to quantitate phénobarbital over the entire therapeutic range ^ g / m L levels). Of all the FIAs developed to date, polarization-based fluoroimmunoassays (PFIA) have been the most successful (19). These assays are based on the fact that the rotational diffusion rate of any molecular species depends on its size; a small species will rotate faster than a larger species. Because the fluorescent label has a finite lifetime, on the order of the rotational diffusion rate (Brownian motion), it can be used to indirectly probe this dynamic rotational phenomenon. In short, the fluorescent label can be used to track rotational diffusion, which manifests itself in a larger polarization for a larger species versus a smaller species (20). Thus Ag* and Ag*Ab are easily distinguished simply by measuring the fluorescence polarization (P).

(3)

Figure 3 shows how polarization measurements are performed. Vertically polarized light is used to excite the sample, and the intensities of vertically polarized emission (7| ]) and horizontally polarized emission (I±) are used to calculate the fluorescence polarization. The polarization of the emission is selectively monitored by the rotation of a simple polarizer. FPIAs have been developed for many drugs, antibiotics, and steroids. However, larger species are not easily quantitated because the relative change in polarization is insignificant. (The polarization is not affected as much for larger antigens because the fluorescent label can rotate locally as well as globally.) In our laboratory, we are developing a new approach to homogeneous FIA based on fluorescence anisotropy selective technique (FAST). Briefly, FAST allows direct resolution of components (i.e., Ag* and Ag*Ab) based solely on the difference in their rotational dynamics (21). Although this sounds much like steady-state polarization, FAST can directly determine the individual rotational rates for a molecular species. For example, if a sample species exhibits two unique rotational correlation times (RCTs), FAST can distinguish each RCT and exploit a given RCT to quantify a given species. Conversely, steady-state polarization provides information regarding only the ensemble average of the two RCTs. Thus by employing FAST, one can zero in on a specific rotational diffusion rate and directly monitor the effects of things such as antibody binding. Recently we developed a FASTbased FIA for bovine serum albumin (BSA) (22). In this case, 10% of the observed fluorescence depolarization was from the global (total) rotation of the entire BSA molecule, and the remaining 90% was from local rotation of the fluorescent label (rhodamine B) around its covalent labeling bond. The 90% local rotation was unaffected by antiBSA binding, but the 10% global motion was extremely sensitive to the degree of association between BSA* and antiBSA. By selectively monitoring the global motion, we were able to develop an assay for BSA. Figure 4 shows our initial working curve for this assay in which we monitor only the RCT for the 10% global motion. We are currently refining FAST for other biochemical species. Fiber-optic sensors. All of the assays discussed above are necessarily restricted to the laboratory environment. A distinct advantage would be gained if these measurements could be performed in environments that are inaccessible, hot, cold, radioactive, or oth-

ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988 · 1037 A

erwise restrictive. During the past decade, the communications industry has developed inexpensive, small, high-throughput optical fibers. Because optical fibers provide a passive means of transporting light from one location to another, many conventional spectroscopic techniques can be carried out with fiber-optic sensors. Several excellent reviews on fiber-optic sensors have appeared recently (23-25). Fluorescence-based fiber-optic sensors have been described for blood pH, pC0 2 , bilirubin, pC>2, metal ions, glucose, and biomaterials (23-25). Of these, the most exciting sensors are those directed toward the determination of bioactive molecules. The first such sensor, which employed evanescent wave coupling and exhibited subnanogram detection limits, was developed for determination of human immunoglobulin G (IgG) by Hirschfeld and co-workers (26). Vo-Dinh, Sepaniak, and co-workers have also developed two very clever fluorescence immunosensors. The first employs immobilized anti-IgG and fluorescein isothiocyanate-labeled IgG for the in situ determination of IgG and has a typical detection limit of 25 fmol (27). The second fiber-optic-based immunosensor was developed in a similar fashion, but it employs antibenzo[a]pyrene antibodies for the in situ determination of benzo[a]pyrene. The detection limit for this sensor is 1 fmol (28). For the most part, fiber-optic-based sensors employ wavelength selectivity. In our laboratory, however, we are developing fiber-optic-based instrumentation and sensors employing fluorescence lifetime selectivity. Figure 5 shows a simplified schematic of a multifrequency fiber-optic-based phasemodulation fluorometer. Using this design, we have successfully resolved single, double, and triple exponential decays of fluorescence with a bifurcated fiber-optic probe up to 100 m in length (29, 30).

Figure 4. Working curve for the determination of bovine serum albumin (BSA) using the fluorescence anisotropy selective technique.

through simple fiber-optic sensors. Finally, fluorescence-based techniques should find applications in areas as diverse as solid-liquid interfaces and protein dynamics. References

(1) Harvey, E. N. A History of Luminescence from Earliest Times until 1900; American Philosophical Society: Philadelphia, 1957. (2) Chem. Eng. News, April 24, 1967, p. 72. (3) O'Haver, T. C. J. Chem. Educ. 1978,55, 423. (4) Harris, D. C. Quantitative Chemical Analysis; W. H. Freeman and Co.: San Francisco, 1982; p. 500. (5) Demas, J. N.; Keller, R. A. Anal. Chem. 1985,57, 538.

(6) Warner, I. M.; Patonay, G.; Thomas, M. P. Anal. Chem. 1985, 57, 463 A. (7) Darzynkiewicz, Z.; Traganos, F.; Kapuscinski, J.; Staiano-Coico, J.; Melamed, M. R. Cytometry 1984,5, 355. (8) Karsten, V.; Wallenberger, A. Anal. Biochem. 1972, 77,464. (9) Bradford, C. L.; Harris, D. A. Spectrophotometry and Spectrofluorimetry: A Practical Approach; IRL Press: Washington, DC, 1987; Chapter 4. (10) Guilbault, G. G. Pure Appl. Chem. 1985,57,495. (11) Haugland, R. P. Handbook of Fluorescent Probes and Research Chemicals; Molecular Probes, Inc.: Eugene, OR, 1985. (12) Nguyen, D. C; Keller, R. A. Anal. Chem. 1987,59, 2158. (13) Pettersson, K.; Siitari, H.; Hemmila, I.; Soini, E.; Lovgren, T.; Hanninen, V.; Tan-

The future

The future for fluorescence-based bioanalytical techniques is extremely promising. With greater understanding of bioanalytical systems and the steady development in instrumentation, we believe that even greater accomplishments are on the horizon. Enzymaticbased approaches should continue to grow, and fluorescence-based fiber-optic sensors should become a more intense area of study. The ultimate goal is to develop real-time in vivo sensors (chemical and immunochemical) for the trace determination of specific biomolecules and bioactive species. Within the next few years, we should see essentially every measurement that one can make in a cuvette being made

Figure 5. Schematic of the fiber-optic-based multifrequency phase-modulation fluorometer.

1038 A · ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988

ner, P.; Stenman, U. Clin. Chem. 1983,29, 60. (14) Sepaniak, M. J.; Petrea, R.; Vo-Dinh, T. Talanta 1988, 35,139. (15) Bright, F. V.; McGown, L. B. Anal. Chim. Acta 1984,162, 275. (16) McGown, L. B.; Bright, F. V. C.R.C. Crit. Rev. Anal. Chem. 1987,18, 245. (17) Bright, F. V.; Vickers, G. H.; Hieftje, G. M. Anal. Chem. 1986,58,1225. (18) Bright, F. V.; McGown, L. B. Talanta 1985,32,15. (19) Provost, Y.; Farinotti, R. J. J. Pharm. Clin. 1984,3,197. (20) Spencer, R. D.; Toledo, F. B.; Wil­ liams, B. T.; Yoss, N. L. Clin. Chem. 1973, 19 838 (21)'Bright, F. V. Appl. Spectrosc, in press. (22) Bright, F. V. Presented at the 1987 FACSS Conference, Detroit, MI. (23) Seitz, W. R. Anal. Chem. 1984, 58, 16 A. (24) Angel, S. M. Spectroscopy 1987,2, 38. (25) Seitz, W. R. C.R.C. Crit. Rev. Anal. Chem. 1988,19,135. (26) Klainer, S.; Hirschfeld, T.; Bowman, H.; Milanovich, F.; Perry, D.; Johnson, D. Technical Report No. LBL 11981, 1981; Lawrence Berkeley Laboratory, Berkeley, CA. (27) Tromberg, B. J.; Sepaniak, M. J.; VoDinh, T.; Griffin, G. D. Anal. Chem. 1987, 59 1226 (28) Vo-Dinh, T.; Tromberg, B. J.; Griffin, G. D.; Ambrose, K. R.; Sepaniak, M. J.; Gardenhire, Ε. Μ. Appl. Spectrosc. 1987, 41, 735. (29) Bright, F. V. SPIEProc, 1988,909,23. (30) Bright, F. V. Appl. Spectrosc, in press. The work from the author's laboratory described here was supported by the following grants and organizations: BRSG S07 RR 07066, awarded by the Biomedical Research Support Grant Program, Division of Resources, National Institutes of Health; the donors of the Petroleum Research Fund, administered by the American Chemical Society; a New Faculty Development Award from the New York State/United University Profes­ sions; a Nontenured Faculty Grant from 3M, Inc.; and a grant from the Health Care Instruments and Devices Institute at the State University of New York at Buffalo.

Frank V. Bright received his B.S. de­ gree in 1982 from the University of Redlands and his Ph.D. degree in 1985 from Oklahoma State University un­ der the direction of Linda B. McGown. He was a postdoctoral fellow with Gary M. Hieftje at Indiana University before joining the faculty of SUNY. His research interests include laserbased fluorescence spectroscopy, fi­ ber-optic sensors, antigen-antibody binding, polycyclic aromatic hydro­ carbons, metal-ligand complexation, and high-pressure effects.

Sigma PI at software creates publication quality charts & graphs For the scientist IBM PC and Compatibles

Create professional graphs For scientific publications, poster sessions, or lectures. Unlike business software, Sigma-Plot con­ tains essential scient/Tic Features. m Automatic error bars m Multiple graphs per page u 65,000 points per dataset

Free brochure 8 0 0 - 8 7 4 - 1 8 8 In CA 4 1 5 - 9 2 4 - 8 6 4 0 FAX: 415-924-2850 / Telex: 4931977 In Europe: R.J.A. Handers, Germany Ph #02101/666268

• Greek and math sgmbols m Polynomial curve Fitting, and more... Accepts ASCII and DIP data Files From Lotus, dBase, or your own programs. Support For Hewlett-Packard and compatible plotters, HP Laserjets, and selected dot matrix printers. Moneyback guarantee.

Jandel

WBSÊSÊm

S C I Ε Ν Τ I F I C

ι 65 Koch Road • Corte Madera, CA 94925

CIRCLE 88 ON READER SERVICE CARD

How do you get three pegs in two holes?

It's not magic; it's simply good engineering and design. The OIC Model 4440 Tandem detector com­ bines the Electrolytic Conductivity detector and the Photoionization detector into a single unit taking up only one of the GC's detector ports. This leaves the "A" port open for a third detector such as the ECD or FID. So you can do more. You'll spend less time replumbing the system and you'll be able to analyze more samples with the same ma­

chine. With the detectors in tan­ dem, simultaneous detection of halogen and aromatic compounds is possible, giving you two analy­ ses in the time it normally takes for one. The detectors may be used separately, if desired, and excellent results can be obtained with either packed or capillary columns. Call us or write O.I.Corporation today and we'll tell you more about how easily you can get three pegs in two holes.

O I CORPORATION Graham Rd. at Wellborn Rd. · P.O. Box 2980 · College Station, Texas 77841-2980 Telephone (409) 690-1711 · FAX: (409) 690-0440 · TWX NO: 510-892-7944 CIRCLE 122 ON READER SERVICE CARD

ANALYTICAL CHEMISTRY, VOL. 60, NO. 18, SEPTEMBER 15, 1988 · 1039 A