Fluorescent microparticles with controllable enhanced stokes shift

Description Submit all comments and votes

FIELD OF THE INVENTION

The invention relates to polymeric materials incorporating multiple fluorescent dyes to allow for controlled enhancement of the Stokes shift. In particular, the invention describes microparticles incorporating a series of two or more fluorescent compounds having overlapping excitation and emission spectra, resulting in fluorescent microparticles with a desired effective Stokes shift. The novel fluorescent microparticles are useful in applications such as the detection and analysis of biomolecules, such as DNA and RNA, that require a very high sensitivity and in flow cytometric and microscopy analytical techniques.

BACKGROUND OF THE INVENTION

Microparticles labeled with fluorescent dyes have found use in a wide variety of applications. Microparticles are generally considered to be spherical or irregular in shape, and to be less than about 50 micrometers in diameter. They may be prepared by several practical methods from a variety of polymerizable monomers, including styrenes, acrylates and unsaturated chlorides, esters, acetates, amides and alcohols. Microparticles can be further modified by coating with one or more secondary polymers to alter the surface properties of the particles.

Fluorescent microparticles are most commonly used in applications that can benefit from use of monodisperse, chemically inert, biocompatible particles that emit detectable fluorescence and that can bind to a particular substance in the environment. For example, fluorescent particles to which biological molecules have been attached have been used for immunoassays (U.S. Pat. No. 4,808,524 (1989)), for nucleic acid detection and sequencing (Vener, et al. ANALYT. BIOCHEM. 198, 308 (1991); Kremsky, et al., NUCLEIC ACIDS RES. 15, 2891 (1987); Wolf, et al., NUCLEIC ACIDS RES. 15, 2911 (1987)), as labels for cell surface antigens, FLOW CYTOMETRY AND SORTING, ch. 20 (2.sup.nd ed. (1990)), and as tracer to study cellular metabolic processes (J. LEUCOCYTE BIOL. 45, 277 (1989)). The high surface area of microparticles provides an excellent matrix for attaching biological molecules while the fluorescent properties of these particles enable them to be detected with high sensitivity. They can be quantitated by their fluorescence either in aqueous suspension or when captured on membranes.

Fluorescent microparticles can be visualized with a variety of imaging techniques, including ordinary light or fluorescence microscopy and laser scanning confocal microscopy. Three-dimensional imaging resolution techniques in confocal microscopy utilize knowledge of the microscope's point spread function (image of a point source) to place out-of-focus light in its proper perspective. Small, uniform, fluorescently labeled polystyrene microspheres have been employed as point sources for these microscopes (Confocal Microscopy Handbook p. 154 (rev. ed. 1990)).

Many luminescent compounds are known to be suitable for imparting bright and visually attractive colors to various cast or molded plastics such as polystyrene and polymethyl methacrylate. Uniform fluorescent latex microspheres have been described in patents (U.S. Pat. No. 2,994,697, 1961; U.S. Pat. No. 3,096,333, 1963; Brit. U.s. Pat. No. 1,434,743, 1976) and in research literature (Molday, et al., J. CELL BIOL. 64, 75 (1975); Margel, et al., J. CELL SCI. 56, 157 (1982)). A recent patent application of the inventor (Brinkley, et al., Ser. No. 07/629,466, filed 12/18/90) describes derivatives of the dipyrrometheneboron difluoride family of compounds (derivatives of 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene) as useful dyes for preparing fluorescent microparticles. This family of dyes possesses advantageous spectral data and other properties that result in superior fluorescent microparticles.

Although dipyrrometheneboron difluoride labeled materials are highly fluorescent and photochemically stable, a disadvantage of these fluorescent materials is their relatively small Stokes shift (the difference between the peak excitation and peak emission wavelengths) when only one dye is used. Because the optimum wavelength of the exciting light is close to the peak emission light, fluorescent particles with small Stokes shifts require precise excitation and emission filters to eliminate or reduce interference. The customary use of excitation filters blocks part of the excitation and emission light that would otherwise increase the efficiency of the fluorescence and reduces the intensity of the fluorescent signal. Fluorescent materials that incorporate bright fluorescent dyes with increased Stokes shifts would permit maximum utilization of the available excitation and emission light, resulting in a greater fluorescent signal (see e.g. FIGS. 3A and 3B).

Another advantage of fluorescent materials with large Stokes shifts is that they can be more easily detected in the presence of other fluorescent materials. Immunoassays are typically carried out in body fluids which contain many endogenous fluorescent molecules, such as bilins, flavins and drugs. Since the vast majority of interfering fluorescent materials have relatively short Stokes shifts, the use of a fluorescent label that emits at a wavelength far greater than its excitation wavelength makes the label easier to distinguish from background fluorescence, since its fluorescent signal is emitted at a wavelength at which most background fluorescence is minimal.

A third advantage of fluorescent materials with large Stokes shift is their usefulness in detecting multiple analytes in a single sample using a single excitation wavelength. Using two or more different fluorescent labels, each of which can be excited at a particular wavelength (e.g. the 488 nm argon laser principal emission), the emission peaks of the different labels are detected at different wavelengths, where each emission spectrum is characteristic of a single analyte. In order to successfully accomplish this, the emission peaks of the fluorescent labels must be well-separated from each other so the correction factors between the various dyes are minimized. High photostability of the label is also beneficial. Fluorescent materials with a large Stokes shift can be used in combination with fluorescent materials with a smaller Stokes shift where both materials excite at the same wavelength, but emit at different wavelengths, giving multiple signals that can be resolved using optical filters or monochromators.

Unfortunately, fluorescent compounds useful as labeling reagents that have Stokes shifts of 50-100 nm as well as high fluorescence efficiency and emission wavelengths of greater than 500 nm required for detectability are relatively rare. (Haugland, Fluorescein Substitutes for Microscopy and Imaging, OPTICAL MICROSCOPY FOR BIOLOGY pp. 143-57 (1990). The magnitude of the Stokes shift in fluorescent dyes has been found to be generally inversely proportional to the high absorbance needed to ensure a strong signal. Fluorescent dyes in use as labeling reagents for biological molecules, such as xanthenes, dipyrrometheneboron difluorides, rhodamines and carbocyanines commonly have Stokes shifts of less than about 30 nm.

The lack of suitable fluorescent dyes with large Stokes shifts has led to the development and use of protein-based fluorophores known as phycobiliproteins as labels (e.g. U.S. Pat. Nos. 4,520,110 and 4,542,104 both to Stryer, et al. (1985)). Like other fluorophores, they have been covalently attached to beads and macromolecules. See, e.g., Oi, et al., J. CELL BIO. 93,981 (1982). These large bilin-containing molecules have the desirable characteristics of very high extinction coefficients and they use internal energy transfer between unlike, covalently-linked fluorophores to accomplish a relatively large Stokes shift. They have the disadvantage of poor chemical stability, instability to photobleaching, limited long wavelength emission capability, bulky molecular size (MW >100,000 Daltons) and relatively high cost. Furthermore, only a few proteins of this type are known and one cannot select or appreciably adjust their spectral properties. In an effort to improve the fluorescent emission efficiency of phycobiliproteins without significantly increasing their molecular size, phycobiliproteins have been covalently coupled to the fluorescent dye Azure A (U.S. Pat. No. 4,666,862 to Chan (1987)).

It is known that covalent coupling of a pair of fluorophores results in a fluorescent dye with a larger Stokes shift than either of the individual dyes (e.g. Gorelenko, et al., Photonics of Bichromophores Based on Laser Dyes in Solutions and Polymers, EXPERIMENTELLE TECHNIK DER PHYSIK 37, 343 (1989)). This approach, although reportedly effective in increasing the Stokes shift, requires complex synthetic procedures to chemically couple the two dyes together and are limited by the number and location of available reactive sites. The process of carrying out the necessary synthetic procedures to attach three, four, or more dyes sufficiently close together and in the proper configuration to undergo substantial energy transfer would be exceedingly difficult, if not impossible. Furthermore, covalently linked molecules typically have sufficient freedom of movement that significant collisional deactivation occurs, leading to loss of energy by vibrational relaxation rather than by fluorescence. There is a need for a way of combining the spectral properties of dyes by methods other than complex covalent coupling to provide useful fluorescent labels with an enhanced effective Stokes shift.

In studies of energy transfer between pairs of covalently linked dyes, it has been shown that the efficiency of energy transfer between two fluorescent dyes is inversely proportional to the sixth power of the distance between the two interacting molecules, consistent with Forster's theory (Stryer & Haugland, Energy Transfer: A Spectroscopic Ruler, PROC. NAT'L ACAD. SCI. USA 58, 719 (1967)). The reference suggests that the percentage of measurable energy transfer can be used to measure the distance separating the covalently linked fluorophores in the 10 to 60 .ANG. range. A subsequent paper, Haugland, Yguerabide, & Stryer, Dependence of the Kinetics of Singlet-Singlet Energy Transfer on Spectral Overlap, PNAS 63, 23 (1969), reported that intramolecular singlet energy transfer depends on the magnitude of spectral overlap integral.

Energy transfer has been demonstrated between dyes that have been coupled to macromolecules to study intramolecular distances and conformation in biomolecules, e.g., Julien & Garel, BIOCHEM. 22, 3829 (1983); Wooley, et al., BIOPHYS. CHEM. 26, 367 (1987); and in polymer chains and networks, e.g. Ohmine, et al., MACROMOLECULES 10, 862 (1977); Drake, et al., SCIENCE 251, 1574 (1991). Energy transfer with resultant wavelength shifting has also been described for mixtures of dyes in lasing solutions, e.g. Saito, et al., APPL. PHYS. LETT. 56, 811 (1990). Energy transfer has been demonstrated between monomolecular layers of dyes and other organized molecular assemblies, e.g. Kuhn, Production of Simple Organized Systems of Molecules, PURE APPL. CHEM. 11, 345 (1966), abstracted in CHEM. ABSTRACTS 66, 671 (1967); Yamazaki, et al., J. PHYS. CHEM. 94, 516 (1990). Energy transfer between paired donor and acceptor dyes has also been demonstrated in polymer films as a way of studying the energy transfer dynamics, e.g. Mataga, et al., J. PHYS. CHEM. 73, 370 (1969); Bennett, J. CHEM. PHYSICS 41, 3037 (1964). Although the conformity of research results to Forster's theoretical formulation have been widely reported, utilitarian applications of the theory have been limited. The cited references neither anticipate nor suggest fluorescent microparticles incorporating a series of dyes to be used as labeling reagents with an enhanced effective Stokes shift.

It is an object of the invention to provide a more simple method than complex covalent coupling for combining the spectral properties of multiple dyes, while minimizing collisional deactivation, allowing efficient energy transfer and increasing the effective Stokes shift for the purpose of providing more useful fluorescent labeling reagents. It is a further object of the invention to provide materials that have not only an increased effective Stokes shift but materials for which the effective Stokes shift can be selectively controlled by the selection of appropriate dyes with overlapping spectral properties.

Immobilizing fluorescent dyes randomly in a polymeric matrix according to the subject invention provides just such a simple method of providing novel fluorescent materials with controllable, enhanced effective Stokes shifts. Certain fluorescent dyes, such as dipyrrometheneboron difluoride dyes, coumarin dyes and polyolefin dyes, have high fluorescence efficiency when incorporated into polymeric materials and are available in a large number of derivatives with a wide range of excitation and emission maxima, These characteristics allow the wavelength of excitation and the magnitude of the increase in the effective Stokes shift to be easily controlled by carefully selecting dyes with the appropriate spectral overlap for incorporation into the microparticles.

DESCRIPTION OF THE DRAWINGS

FIG. 1 is a graph of the emission spectra of 0.093 micrometer polystyrene latex particles containing various combinations of dipyrrometheneboron difluoride dyes. Spectrum A represents 16 .mu.Mol/g-latex of Compound 1 (see Table 2 below), exited at 490 nm (a preferred excitation wavelength); Spectrum B results from particles containing a mixture of 16 .mu.Mol/g-latex of Compound 1 (donor) and 9.6 .mu.Mol/g of Compound 3 (acceptor), also excited at 490 nm; spectrum C is the emission of particles containing 9.6 .mu.Mol/g-latex of Compound 3, excited at 540 nm.

FIG. 2 shows the emission spectrum of 0.093 micrometer polystyrene latex particles containing three dipyrrometheneboron difluoride dyes; Compound 1 (donor dye), Compound 3, (transfer dye) and Compound 4 (acceptor) dye. The molar ratio of the dyes used to prepare the fluorescent latex microparticles was: 1 donor: 0.6 transfer: 0.9 acceptor.

FIGS. 3A and B shows in graphic form the increase influorescent signal that is attainable by increasing the Stokes shift from about 10 nm to 70 nm in 0.093 micrometer latex microparticles. The microparticles in Spectrum A (FIG. 3A) contain 16 .mu.Mol/g-latex of Compound 1, while the microparticles in Spectrum B (FIG. 3B) contain 16 .mu.Mol/g-latex of Compound 1 (donor) and 9.6 .mu.Mol/g-latex of Compound 3 (acceptor). The relative shaded areas show the optimum filter bandwidths that can be used in each of these microparticle preparations, demonstrating the increasing signal that is obtainable from the microparticles containing the donor-acceptor dye pair.

SUMMARY OF THE INVENTION

The invention relates to polymeric microparticles incorporating multiple fluorescent dyes to allow for controlled enhancement of the effective Stokes shift. The effective Stokes shift is the Stokes shift of the microparticle, i.e. the difference between the peak excitation wavelength of the initial donor dye and the peak emission wavelength of the ultimate acceptor dye after incorporation in the microparticle. In particular, the invention describes microparticles incorporating a series of two or more fluorescent compounds having overlapping excitation and emission spectra. Efficient energy transfer from the excitation wavelength of the first dye in the series which is re-emitted at the emission wavelength of last dye in the series results in a large and readily controllable effective Stokes shift. Selection of appropriate dyes results in fluorescent probes with desired excitation and/or emission wavelengths and preselected and increased effective Stokes shift.

Selection of Dyes

A series of fluorescent dyes is selected for incorporation into the microparticles based on their excitation and emission spectral properties. The dyes included in the series form a cascade of excitation energy transferred from high energy (short wavelength) to low energy (long wavelength) resulting in enhanced optical luminescence from the final dye in the series, regardless of the sequence of their incorporation or their random physical location in the microparticles.

The spectral properties for the series of fluorescent dyes should be determined in the polymeric materials in which they will be used. Although certain 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene and 4,4-difluoro-4-bora-3a,4a,8-triaza-s-indacene dyes have been found to have spectral properties in polymer materials that are comparable to their spectral properties in solution, most dyes show significant and unpredictable spectral shifts depending on the media in which they are measured. Generally, oil soluble dyes and neutral dyes combine more readily with the polymeric materials. In addition to the desired excitation and emission peaks as described below, dyes useful for the invention also generally have a quantum yield of greater than about 0.2, preferably greater than about 0.5, as well as an extinction coefficient of greater than about 25,000 cm.sup.-1 M.sup.-1, preferably greater than about 50,000 cm.sup.-1 M.sup.-1.

The excitation and emission peaks and other spectral properties of candidate dyes are easily determined by conventional means. The excitation peak(s) of a dye can be approximately determined by running an absorption spectrum on any absorption spectrophotometer or, exactly, by running a fluorescent excitation spectrum using a scanning fluorescence spectrophotometer. The emission peak of the dye is also determined by using a fluorescence spectrophotometer to get an emission spectrum using a scanning fluorometer. Quantum yield is typically determined by measuring with a fluorometer the total fluorescence emission in the polymer of both the unknown dye and a reference dye with known absorbances at the excitation wavelength. The extinction coefficient is typically determined for a free dye in solution by using a spectrophotometer to measure absorbance of a solution with a gravimetrically determined concentration and calculating the extinction coefficient based on the Beer-Lambert law. After determining the spectral characteristics of likely dyes, dyes having the desired spectral characteristics are selected.

TABLE 1 ______________________________________ CONVENTIONAL FLUORESCENCE EXCITATION SOURCES.sup.a Useful Wavelengths Principal Lines Sources (nm) (nm) ______________________________________ Mercury Arc 250-600 254,366,436,546 Xenon Arc 250-1000 467, several > 800 Tungsten Filament 350-1000 None He--Cd Laser 325,442 442 Ar Laser 350,360,458,476, 488,514 488,496,514 He--Ne Laser 543,594,633 633 Kr Laser 530,568,647,676 647 Diode Laser >650 850 (GaAlAs).sup.b ______________________________________ .sup.a Only primary excitation sources capable for continuous operation (CW) have been considered. Several other laser sources are available that either provide pulsed output (e.g. N.sub.2 laser) or require pumping by C ion lasers (e.g. dye lasers, Ti:Sapphire lasers). .sup.b Material dependent, multiple types available.

The series of fluorescent dyes contains an initial donor dye with a desired excitation peak. The initial donor dye receives external excitation energy, such as from photons, x-rays, or decay of radioactive materials (e.g. .beta.-emitters). In one embodiment of the invention, the initial donor dye receives excitation energy from incandescent or laser-based excitation sources to avoid the hazards of radioactive materials. Lasers, including argon ion, krypton ion, He-Cd, He-Ne, excimer, diode, metal vapor, neodymiumYAG, nitrogen and others, produce from one to several discrete lines that contain sufficient power for fluorescence excitation. Laser sources are available to provide many excitation lines over the spectrum from the UV to the infrared, to excite a wide range of fluorescent dyes.

The initial donor dye in the series has an excitation peak that overlaps the emission wavelength of energy from a preferred excitation source. For example, the most widely used beam-scanning confocal microscopes currently use air cooled Ar ion lasers, or more recently Ar-Kr gas mixes, allowing for fluorescent dye excitation between about 450 and 650 nm. Preferably, the excitation peak is optimal for absorption of energy from the excitation source, i.e. the excitation peak significantly overlaps the output of the excitation source. Table 1 lists the wavelengths for a number of conventional excitation sources.

The series of fluorescent dyes used for the invention also has an ultimate acceptor dye with a desired emission peak. Generally, the desired emission peak is based on the magnitude of effective Stokes shift desired. The intensity of the fluorescence emission with respect to background "noise" (the signal-to-noise ratio), is fundamental to sensitivity and image contrast. The signal-to-noise quality of fluorescence data may be severely compromised by background signals at wavelengths different than the emitted fluorescence of interest. The background signals may result from light scatter or may be due to fluorescence intrinsically present in many biological systems or generated from an analytically useful second emitting species. It is possible, for example, to avoid cellular autofluorescence or to provide for a range of materials that can be excited by common wavelength but detected at different wavelengths by selection of a dye series that gives the desired effective Stokes shift.

The desired magnitude of effective Stokes shift may require that additional or "transfer" dyes be included in the series of fluorescent dyes to act as both intermediate donor and acceptor dyes. Individual dyes having a relatively narrow Stokes shift are preferred as transfer dyes for ease of exactly tailoring the desired excitation and emission peaks. Each transfer dye receives energy from the preceding dye in the series (acting as an acceptor dye with respect to the preceding dye) and substantially transfers the received energy to the next dye in the series (acting as a donor dye with respect to the next dye). Each dye in the series of fluorescent dyes has spectral overlap with the other dyes in the series, i.e. the emission peak of each donor or transfer dye in the series overlaps the excitation peak of the following acceptor or transfer dye in the series (Ex. 4/FIG. 2). The excitation and emission peaks overlap sufficiently so that, upon excitation, significant (i.e., greater than about 50%) transfer of the excitation energy from each dye to the next is achieved. The efficiency of energy transfer is measured by the loss in fluorescence intensity of the emission peak of the donor dye measured at the same degree of dye loading as polymer loaded with only the donor dye (FIG. 1). Typically, substantially complete energy transfer is achieved (i.e. greater than about 90%). Preferably, greater than about 95% energy transfer is achieved. More preferably, greater than about 98% energy transfer is achieved. Addition of a transfer dye is most appropriate when the efficiency of energy transfer falls below about 90%. The function of the transfer dye is to accept the excited state energy from the initially excited donor dye and to facilitate transfer of this energy to the ultimately detected acceptor dye.

Typically, the desired energy transfer is achieved when sufficient amounts of dyes are loaded into the polymeric microparticles so that the average intermolecular distance between donor and transfer and/or acceptor dyes is between about 40 .ANG. and about 25 .ANG.. Intermolecular distances between the dyes of greater than about 40 .ANG. generally result in less efficient energy transfer, while intermolecular distances between the dyes of less than about 25 .ANG. generally result in non-radiative energy loss that reduces the fluorescence emission of the dyes.

A sufficient number of donor, acceptor and transfer dyes are included in the series so that efficient energy transfer from the initial dye in the series to the ultimate dye in the series occurs. Although the dyes may be included in equal amounts, too much total dye may result in deterioration in properties of the polymer and in suboptimal fluorescence of the materials because of non-radiative energy loss. The average random distance between the fluorophores compatible with significant energy transfer generally results when the total dye concentration is less than about 10% (w/w); preferably the total dye concentration is between about 0.5 and 2% (w/w); more preferably between about 0.8 and 1.2% (w/w). Less (fewer molar equivalents) transfer dye than donor and acceptor dyes may sometimes be used to achieve the desired energy transfer. Less (fewer molar equivalents) ultimate acceptor dye than initial donor dye may also be effective in achieving the desired energy transfer. Increasing the amount of ultimate acceptor in proportion to the amount of initial donor generally has little effect on improving the ultimate fluorescent signal, whereas increasing the proportion of initial donor improves the effective extinction coefficient of and lead to greater emission from the ultimate acceptor dye. While not wishing to be bound by theory, it appears that the donor dye functions as a radiation collection mechanism for funneling energy output to the acceptor dye such that more initial donor dye results in more efficient absorption at excitation which in turn yields a more intense fluorescence signal. A workable ratio of initial donor dye to ultimate acceptor dye is between about 1:5 and about 10:1; preferably between about 1:1 and about 8:1; more preferably between about 4:1 and about 6:1.

In one embodiment of the invention, novel fluorescent materials are prepared from two or more polyazaindacene dyes (i.e. derivatives of 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene or 4,4-difluoro-4-bora-3a,4a,8-triaza-s-indacene). Polyazaindacene derivatives suitable for preparation of fluorescent polymer microparticles according to this invention have the general structure of formula (I): ##STR1## wherein R.sub.1 -R.sub.6, which may be the same or different, are hydrogen, halogen, or alkyl, alkoxy, alkenyl, cycloalkyl, arylalkyl, acyl, or aryl, heteroaryl, alone or in combination; and R.sub.7 is nitrogen, methine, or halogen-, alkyl-, alkoxy-, alkenyl-, cycloalkyl-, arylalkyl-, acyl-, aryl- or heteroaryl-substituted methine.

The alkyl, cycloalkyl, arylalkyl, acyl, alkoxy, and alkenyl substituents of the polyazaindacene derivatives generally each have independently fewer than about 20 carbon atoms, preferably fewer than about 10 carbon atoms. The term alkenyl includes ethenyl or conjugated dienyl or trienyl, which may be further substituted by hydrogen, halogen, alkyl, cycloalkyl, arylalkyl, acyl, (the alkyl portions of which each contain fewer than about 20 carbons), cyano, carboxylate ester, carboxamide, aryl or heteroaryl.

A heteroaryl group is a heterocyclic aromatic group that contains at least one heteroatom (a non-carbon atom forming the ring structure). The heteroaryl group can be a single ring structure or a fused two- or three-ring structure. Each ring can be a 5- or 6-member ring. The heteroaryl group can contain one or more heteroatoms. The term heteroaryl includes its alkyl-, aryl-, arylalkyl- or heteroaryl-substituted derivatives. For example, the heteroaryl substituent is pyrrole, thiophene, or furan (single ring, single heteroatom), or oxazole, isoxazole, oxadiazole, or imidazole (single ring, multiple heteroatoms). Alternatively, the heteroaryl group is a multi-ring structure containing one or more heteroatoms, for example, the heteroaryl substituent is benzoxazole, benzothiazole, or benzimidazole, (multi-ring, multiple heteroatoms), or benzofuran or indole (multi-ring, single heteroatom).

In general, 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene derivative dyes are prepared from suitable pyrrole precursors, according to methods known in the art (e.g. U.S. Pat. No. 4,916,711 to Boyer, et al. (1990) and U.S. Pat. No. 4,774,339 to Haugland, et al. (1988) each of which are incorporated by reference). Typically, approximately stoichiometric proportions of pyrrole precursors, one of which contains an aldehyde or ketone function in the 2-position, are condensed in a reaction mediated by a suitable acid, such as hydrogen bromide, to yield an intermediate pyrromethene salt. Cyclization of the heterocyclic ring formation is completed by addition of boron trifluoride in combination with a strong base such as trimethylamine. Derivatives of 4,4-difluoro-4-bora-3a,4a,8-triaza-s-indacene are synthesized by the cyclization of azapyrromethenes with boron trifluoride in the presence of a base such as N,N-diisopropylethylamine. The azapyrromethene intermediates are prepared by the acid catalyzed condensation of 2-nitrosopyrrole derivatives with suitable pyrrole precursors having a hydrogen on the 2-position.

A representative sample of polyazaindacene dyes suitable for preparing fluorescent microparticles with controlled effective Stokes shifts and a summary of their spectral properties is included in Table 2. Other useful dipyrrometheneboron difluoride dyes are described in pending patents ser. no. 07/704,287 to Kang, et al., filed 5/22/91 and ser. no. 07/629,596 to Haugland, et al., filed 12/18/90 (each of which is incorporated by reference).

Appropriate selection of polyazaindacene derivatives, when incorporated together into a polymeric microparticle, have desired excitation and emission wavelengths that overlap sufficiently so that efficient transfer of energy from donor to acceptor is achieved, considering 1) the wavelength of the excitation and emission of the donor, transfer and/or acceptor dyes; 2) the overlap of the donor dye emission with the transfer and/or acceptor excitation; 3) relative concentrations of donor, transfer and/or acceptor dyes; and 4) the average distance between the dye molecules.

TABLE 2 ______________________________________ Examples* of Polyazaindacene