Cytometry (Communications in Clinical Cytometry) 34:159–179 (1998)
`
`ReviewArticle
`Application of Fluorescence Resonance Energy Transfer
`in the Clinical Laboratory: Routine and Research
`
`Ja´nos Szo¨llo9 si,* Sa´ndor Damjanovich, and La´szlo´ Ma´tyus
`Department of Biophysics and Cell Biology, University Medical School of Debrecen, Debrecen, Hungary
`
`Fluorescence resonance energy transfer (FRET) phenomenon has been applied to a variety of scientific
`challenges in the past. The potential utility of this biophysical tool will be revisited in the 21st century. The
`rapid digital signal processing in conjunction with personal computers and the wide use of multicolor laser
`technology in clinical flow cytometry opened an opportunity for multiplexed assay systems. The concept is
`very simple. Color-coded microspheres are used as solid-phase matrix for the detection of fluorescent labeled
`molecules. It is the homogeneous assay methodology in which solid-phase particles behave similarly to the
`dynamics of a liquid environment. This approach offers a rapid cost-effective technology that harnesses a
`wide variety of fluorochromes and lasers. With this microsphere technology, the potential applications for
`clinical flow cytometry in the future are enormous. This new approach of well-established clinically proven
`methods sets the stage to briefly review the theoretical and practical aspects of FRET technology. The review
`shows various applications of FRET in research and clinical laboratories. Combination of FRET with
`monoclonal antibodies resulted in a boom of structural analysis of proteins in solutions and also in biological
`membranes. Cell surface mapping of cluster of differentiation molecules on immunocompetent cells has
`gained more and more interest in the last decade. Several examples for biological applications are discussed
`in detail. FRET can also be used to improve the spectral characteristics of fluorescent dyes and dye
`combinations, such as the tandem dyes in flow and image cytometry and the FRET primers in DNA sequencing
`and polymerase chain reactions. The advantages and disadvantages of donor-acceptor dye combinations are
`evaluated. In addition, the sensitivity of FRET provides the basis for establishing fast, robust, and accurate
`enzyme assays and immunoassays. Benefits and limitations of FRET-based assays are thoroughly scrutinized. At the
`end of the paper we review the future of FRET methodology. Cytometry (Comm. Clin. Cytometry) 34:159–179,
`1998. r 1998 Wiley-Liss, Inc.
`
`Key terms: fluorescence resonance energy transfer; tandem dyes; enzyme assay; immunoassays; DNA
`sequencing; cell surface mapping
`
`Although the phenomenon of fluorescence resonance
`energy transfer (FRET) was observed by Perrin at the
`beginning of the century, it was Theodor Fo¨rster who
`proposed a theory describing long-range dipole-dipole
`interactions between fluorescent molecules approxi-
`mately 50 years ago (29,30). He derived an equation that
`relates the transfer rate to the interchromophore and the
`spectroscopic properties of the chromophores. The inge-
`nious discovery that a fluorescence dipole-dipole interac-
`tion, besides orientational and other spectroscopic param-
`eters, which can be kept under control, depends on the
`negative sixth power of their distance provided one of the
`most sensitive methods to measure molecular and atomic
`distance relations at the nanometer level. The utilization of
`this method in chemistry and biochemistry reached a
`pinnacle in the 1970s. Cell biological applications also
`started in the 1970s, but widespread application began
`only a decade later and is still flourishing.
`
`r 1998 Wiley-Liss, Inc.
`
`FRET is widely utilized for a variety of applications. In
`one series of studies, FRET is used as a tool for ensuring
`high sensitivity. FRET technology can be incorporated into
`chromatographic assays, electrophoresis, microscopy, and
`flow cytometry. FRET also can be used for improving
`spectral characteristics of fluorescent dyes. In another
`group of studies, FRET is used to obtain structural informa-
`tion that
`is otherwise difficult
`to obtain. The major
`advantage of applying FRET for structural studies is that
`
`Contract grant sponsor: Hungarian Academy of Sciences; Contract
`grant numbers: OTKA T019372, T023835, and 6221; Contract grant
`sponsor: Ministry of Public Health; Contract grant numbers: ETT 344/96
`and 359/96; Contract grant sponsor: Ministry of Education; Contract
`grant number: FKFP 1015/1997.
`*Correspondence to: Ja´nos Szo¨llo9si, Department of Biophysics and Cell
`Biology, University Medical School of Debrecen, P.O. Box 39, Nagyerdei
`krt. 98, H-4012 Debrecen, Hungary.
`E-mail: szollo@jaguar.dote.hu
`Received 9 March 1988; Accepted 29 May 1988
`
`Page 1 of 21
`
`Fluidigm
`Exhibit 1012
`
`

`
`160
`
`SZO¨ LLO9 SI ET AL.
`
`owing to the specific labeling the experimental object can
`be investigated in situ and/or in vivo with little or no
`interference regardless of the complexity and heterogene-
`ity of the system.
`Although numerous reviews are available on fluores-
`cence resonance energy transfer (17,18,21,28,69–71,90,94,
`98,100,106–108,127), there is paucity of information re-
`garding the clinical applications of the technology. In this
`review, an attempt has been made to summarize and
`describe recent applications of the FRET in routine and
`research clinical laboratories. The topics and publications
`discussed in this review undoubtedly reflect the interest of
`the authors, but we did our best to review relevant
`literature. First, we describe briefly the theory behind
`FRET, then the measuring techniques are introduced.
`Next, papers dealing with the structure and cell surface
`distribution of cluster of differentiation (CD) molecules are
`summarized and the analytical applications of FRET are
`described. At the end of the paper, the future prospects of
`FRET applications are discussed. The papers reviewed
`were selected because they had introduced new or im-
`proved methods for FRET measurements and analysis or
`led to a better understanding of important biological
`structures.
`Three directions with enormous clinical potential influ-
`enced the selection of reviewed topics in the future. The
`current interest in quantitative fluorescence determination
`with simultaneous multicolor immunophenotyping as it
`applies to clinical
`immunology. Secondly, the rapidly
`increasing interest in flow cytometer-based multiplexed
`immunoassays. This microsphere-based technology revis-
`its the well-characterized solid-phase immunoassays and
`bioassays that were developed in the 1970s. In the case of
`the multiplexed solid-phase technology, the epitopes are
`suspended uniformly in a liquid environment on solid
`phase (31,73). This paradox-like situation permits harness-
`ing the benefit of both liquid and solid-phase technologies.
`The novel application is based on rapid identifications of
`various microsphere populations with subtle differences
`attributed to spectral emission profiles related to various
`distinct shades of dyes embedded in their surfaces. The
`permutations of discrete microsphere types rendered by
`four-color clinical flow cytometry is staggering. Each
`color-coded population of microsphere set carries reac-
`tants for a distinct bioassay. The solid-phase compartment-
`based technology opens new doors in unrelated areas,
`such as pharmacokinetics for drug discovery studies,
`nucleic acid-based tissue typing, and competitive DNA
`hybridization, that were not readily available for rapid flow
`cytometric-based applications in the past. Finally, in the
`future, the miniaturization of flow cytometric instrumenta-
`tion will force a new approach to evaluate the relationship
`between solid phase and liquid phase in the context of
`rapid immunochemistry. There is a need to revisit the role
`of transfer rate equation that relates to the interchromo-
`phore and the spectroscopic properties of the chromo-
`phores. The ultimate exploitation of FRET will come at the
`end of this century by focusing on the interface between
`molecular pharmacology and medical chemistry for drug
`
`development. Will flow cytometry have a role in this
`fascinating scientific challenge?
`
`THEORY OF FRET
`The theory of FRET was first described by Fo¨rster in the
`late forties, its application to measure distances between
`donor and acceptor molecules came decades later
`(20,21,28–30,58,69–71,94,100,106–108). FRET is a radia-
`tionless process in which energy is transferred from an
`excited donor molecule to an acceptor molecule under
`favorable conditions. One of the most important factors is
`the distance between the donor and acceptor molecules.
`Because the rate of energy transfer is inversely propor-
`tional to the sixth power of the distance between the
`donor and acceptor, the energy transfer efficiency is
`extremely sensitive to distance changes. Energy transfer
`occurs in the 1- to 10-nm distance range with measurable
`efficiency, and these distances correlate well with macro-
`molecular dimensions.
`Consider a system with two different fluorophores in
`which the molecule with higher energy absorption is
`defined as the donor (D) and the one with lower energy
`absorption is defined as acceptor (A). If the donor is in the
`excited state, it will lose energy by internal conversion
`until it reaches the ground vibrational level of the first
`excited state. If the donor emission energies overlap with
`the acceptor absorption energies, through weak coupling,
`the following resonance can occur:
`D* ⫹ A l D ⫹ A*
`
`(1)
`
`where D and A denote the donor and the acceptor
`molecules in ground state, and D* and A* denote the first
`excited states of the fluorophores. The rate of the forward
`process is kT and the rate of the inverse process is k⫺T.
`Because vibrational relaxation converts the excited accep-
`tor into the ground vibrational level, the inverse process is
`highly unlikely to occur. As a result, the donor molecules
`become quenched, while the acceptor molecules become
`excited and, under favorable conditions, can emit fluores-
`cent light. This latter process is called sensitized emission
`(Fig. 1).
`According to the theory of Fo¨rster, the rate (kT) and
`efficiency (E) of energy transfer can be written as:
`
`kT
`
`⫽ const Jn⫺4R⫺6␬2
`
`E ⫽
`
`kT
`⫹ kF
`
`kT
`
`⫹ kD
`
`(2)
`
`(3)
`
`where kF is the rate constant of fluorescence emission of
`the donor and kD is the sum of the rate constants of all
`other deexcitation processes of
`the donor. R is the
`separation distance between the donor and acceptor
`molecules, and ␬2 is an orientation factor that is a function
`of the relative orientation of the donor’s emission dipole
`and the acceptor’s absorption dipole. Other parameters
`are n, the refractive index of the medium, and J, the
`
`Page 2 of 21
`
`Fluidigm
`Exhibit 1012
`
`

`
`FRET IN CLINICAL LABORATORY
`
`161
`
`FIG. 2. Orientation of the transition moments of donor and acceptor
`molecules. ␣ is the angle between the transition moments, ␤ is the angle
`between the transition moment of the donor and the line joining the
`fluorophores, and ␥ is the angle between the transition moment of the
`acceptor and the line joining the fluorophores.
`
`where ␣ is the angle of the transition moments of the
`donor and the acceptor, and ␤ and ␥ are the angles
`between the line joining the centers of the fluorophores
`and their transition moments (Fig. 2). Uncertainties in the
`value of ␬2 cause the greatest error in distance determina-
`tion by energy transfer. (Fortunately R depends on (␬2)1/6).
`The direct measurement of its value is impossible. From
`theoretical considerations ␬2 is in the range between 0 and
`4. Assuming random orientation of the donor and the
`acceptor ␬2 becomes 2/3. In the case of cell surface
`components this assumption is reasonable (19). It can be
`shown that:
`
`(6)
`
`(7)
`
`E ⫽
`
`R⫺6
`⫺6
`R⫺6 ⫹ R0
`
`From Equations 2 and 3 it follows that:
`
`
`
`1R0R26
`
`1 ␶
`
`⫽
`
`kT
`
`where ␶ is the donor’s lifetime in the absence of the
`acceptor, and R0 is the characteristic distance between the
`donor and the acceptor when the transfer efficiency is
`50%.
`
`R0
`
`2QDn⫺4)1/6
`⫽ const( JK
`
`(8)
`
`In this equation, QD is the quantum efficiency of the donor
`in the absence of the acceptor.
`The energy transfer efficiency, as follows from the above
`formulas, can be determined in a number of different
`ways. Since energy is transferred from the excited donor
`to the acceptor, the lifetime (␶), quantum efficiency (Q),
`and fluorescence intensity (F) of the donor decrease, if the
`acceptor is present. As a consequence, the fluorescence
`intensity of the acceptor increases if the donor is present.
`
`(9)
`
`␶
`
`DA
`
`␶
`
`D
`
`1 ⫺ E ⫽
`
`FIG. 1. Energy balance of FRET. The top part of the figure shows the
`Jablonski energy level diagram. The donor fluorophore is excited and
`rapidly drops to the lowest vibrational level of the excited state, where it
`can decay radiatively (fluorescence) or by internal conversion (heat) to the
`ground state, or transfer energy to the acceptor. Only those levels of the
`donor and acceptor with similar energies contribute significantly to the
`transfer rate. Once the acceptor is excited, rapid vibrational relaxation
`prevents back transfer. The acceptor then decays to the ground state via
`fluorescence or heat. The bottom part of the figure shows the spectral
`characteristics and changes of the donor and acceptor undergoing FRET.
`The donor intensity decreases and the acceptor increases (i.e.,
`is
`sensitized) with energy transfer. The spectral overlap that makes FRET
`possible is shown in gray. The absorbance and emission intensities are
`normalized for display purposes.
`
`spectral overlap integral, which is proportional to the
`overlap in the emission spectrum of the donor and the
`absorption spectrum of the acceptor:
`
`兰 FD(␭)⑀
`A(␭)␭⫺4 d␭
`兰 FD(␭) d(␭)
`
`J ⫽
`
`(4)
`
`where FD(␭) is the fluorescence intensity of the donor at
`wavelength ␭, ⑀A(␭) is the molar extinction coefficient of
`the acceptor.
`For dipole-dipole energy transfer it can be shown that:
`
`␬2 ⫽ (cos ␣ ⫺ 3 cos ␤ cos ␥)2
`
`(5)
`
`Page 3 of 21
`
`Fluidigm
`Exhibit 1012
`
`

`
`162
`
`SZO¨ LLO9 SI ET AL.
`
`1 ⫺ E ⫽
`
`A
`F D
`FD
`
`⫽
`
`A
`QD
`QD
`
`D
`F A
`FA
`
`⫽ 1 ⫹ 1⑀
`
`DCD
`⑀
`
`ACA2 E
`
`transfer in a microscope has the advantage of the spatial
`resolution, thus providing structural information at the
`same time. The only disadvantage is statistical accuracy,
`because only a relatively small number of cells can be
`investigated.
`
`(10)
`
`(11)
`
`In the previous formulas the lower indexes refer to the
`donor (D) or acceptor (A), whereas the upper indexes
`indicate the presence of the donor (D) or the acceptor (A)
`in the system, while CD and CA are the molar concentra-
`tions, ⑀D and ⑀A are the molar absorption coefficients of the
`donor and the acceptor. Another possibility for determin-
`ing the energy transfer efficiency is based on the more
`depolarized emission of the acceptor. Detailed evaluation
`of the energy transfer measurements can be found in
`several recent reviews (69–71,100,106,108).
`Calculation of distance relationships from energy trans-
`fer efficiencies is easy in the case of a single donor single
`acceptor system if the localization and relative orientation
`of the fluorophores is known. However, if cell membrane
`components are investigated, a two dimensional restric-
`tion applies for the labeled molecules. Analytical solutions
`for randomly distributed donor and acceptor molecules
`and numerical solutions for nonrandom distribution have
`been elaborated by different groups (24,26,27,35,36,
`92,124). In order to differentiate between random and
`nonrandom distributions, energy transfer efficiencies have
`to be determined at different acceptor concentrations.
`
`MEASURING TECHNIQUES
`Spectrofluorimetry
`For the case of spectrofluorometric measurements,
`Equations 10 and 11 are used to determine the energy
`transfer efficiency. A complete set of samples for transfer
`efficiency determination should contain at least one unla-
`beled, two single-labeled (one labeled with donor only and
`one labeled with acceptor only) and a double-labeled
`(labeled with donor and acceptor) sample. The measured
`fluorescence intensities have to be corrected intensities,
`i.e. for autofluorescence and light scattering. This can be
`achieved using the unlabeled sample. At the same time the
`fluorescence intensities should be normalized to the same
`donor (Equation 10) or acceptor (Equation 11) concentra-
`tion. For both corrections very accurate sample prepara-
`tion is required. The dye concentration should be carefully
`controlled.
`In the case of cell suspension another possible error
`source is the contribution from unbound fluorophores and
`cell debris to the specific fluorescence. These are very
`difficult,
`if not impossible to control, especially if the
`fluorescent label has a low binding constant. Multiple
`washing decreases the contribution of free fluorophores
`to the fluorescence intensity, but unavoidably increases
`the amount of cell debris. Fluorescence microscopy over-
`comes most of the problems one faces using spectrofluo-
`rometry. The distortion caused by dead cells or cell debris
`can be avoided, and uncertainties in cell concentration do
`not cause a problem either. Measurement of energy
`
`Flow Cytometry
`Flow cytometry offers a good compromise for both
`measuring energy transfer on cell surfaces, combining
`some of the advantages of the spectrofluorometric and
`microscopic methods. In the case of flow cytometry, the
`effect of light scattering on fluorescence intensities is
`practically negligible. Because the receptor densities have
`great variation in a cell population, normalization of
`fluorescence intensities on a cell-by-cell basis can not be
`done using the spectrofluorometric approach. Real single-
`cell determination of energy transfer requires the measure-
`ment of all parameters on the same cell. In flow cytometric
`measurements there are three unknown parameters: the
`unquenched donor fluorescence intensity, the nonen-
`hanced acceptor intensity, and the efficiency of the energy
`transfer. In order to determine these parameters one has to
`measure three independent signals from the same cell.
`Two of these parameters are the emission intensities
`detected from different spectral bands. The third emission
`intensity is the one resulting from a second exciting laser
`beam. To this end, a conventional flow cytometer may be
`modified by the introduction of a second excitation laser
`beam. The technical details of such a system were de-
`scribed elsewhere (22,100,102,103,109). Briefly, in such a
`system the 488-nm and the 514-nm lines of an argon ion
`laser are used for excitation. The laser beams are displaced
`by about 0.5 mm at the so-called intersection point, where
`the laser beams illuminate the cells. Spectral ranges of
`fluorescence are detected around the emission maximum
`of fluorescein (535 nm) and usually above 590 nm (emis-
`sion of rhodamine). Data collection is done in list mode,
`meaning that the corresponding light scatter and fluores-
`cence intensities from each cell are stored separately. The
`calculation of energy transfer efficiency is done by a
`computer using Equations 15 and 16. In Equations 12–15,
`the measured intensities are I1, I2, and I3 and the excitation
`and emission wavelengths are given in parentheses. IF and
`IR stand for the theoretical (unquenched and nonen-
`hanced) intensities of the donor (excited at 488 nm,
`emission detected at 535 nm) and of the acceptor (excited
`at 514 nm, detected at ⬎590 nm), respectively. Because
`the emission spectra of the fluorescein and rhodamine
`overlap, and both molecules can be excited by the use of
`both laser beams, correction factors have to be intro-
`duced. These factors are S1, S2, and S3. The definitions of
`these parameters are as follows:
`
`S1
`
`⫽ I2/I1
`
`(determined using only donor labeled cells);
`
`S2
`
`⫽ I2/I3
`
`Page 4 of 21
`
`Fluidigm
`Exhibit 1012
`
`

`
`FRET IN CLINICAL LABORATORY
`
`163
`
`detected at different regions of the spectrum, a more
`elaborate correction method is also possible. Here another
`independent parameter should be detected, a fourth
`fluorescence intensity, and this way the autofluorescence
`can be calculated on a cell-by-cell basis. Naturally high
`autofluorescence may decrease the precision of the mea-
`surements.
`
`IMAGE CYTOMETRY
`Photobleaching FRET Digital Imaging Microscopy
`Jovin and Arndt-Jovin introduced another approach to
`determine transfer efficiencies in a microscope (47,48).
`The energy transfer is calculated from the photobleaching
`kinetics of the donor in the presence and in the absence of
`the acceptor. Their method is based on the fact that the
`integrated fluorescence intensity during complete photo-
`bleaching is independent from the quantum efficiency of
`the fluorophore and therefore it is proportional to the
`donor concentration (41). It is assumed that the donor’s
`photobleaching occurs from the excited singlet state. It
`can be derived that the energy transfer efficiency can be
`calculated as follows:
`
`E ⫽ 1 ⫺
`
`␶
`ble
`
`␶
`
`A
`ble
`
`⫽ 1 ⫺
`
`A )(I0A/Iint
`
`
`(I0/Iint)
`
`(17)
`
`where ␶ble is the time constant of photobleaching, I0 is the
`is the integrated
`initial fluorescence intensity, and Iint
`fluorescence intensity upon complete photobleaching of
`the donor. The A upper indexes indicate the presence of
`an adequate acceptor. The detailed derivation of the above
`formula is described in references (47,48). In the case of
`energy transfer, because there is extra possibility for
`de-excitation, the availability of excited donors for photo-
`bleaching will decrease. As a consequence, the rate of
`photobleaching will be slower, starting from a quenched
`initial fluorescence intensity, but the integrated fluores-
`cence intensity remains unchanged.
`This method offers a greater sensitivity with an internal
`control for real donor concentration. The only drawback
`of the method is the limited number of cells that can be
`measured. This may cause that inhomogeneities in the
`sample are not revealed. However, if the same experiment
`is performed on a flow cytometer this disadvantage can be
`overcome. Since the introduction of this approach several
`successful adaptations for different systems have been
`elaborated (2,95–97).
`
`Intensity-Based Microscopy
`The first semiquantitative intensity-based microscopic
`method for measuring FRET was introduced approxi-
`mately 10 years ago (112,113). Uster and Pagano installed
`an additional filter combination for detecting the ‘‘transfer
`signal’’ by using the excitation wavelength of the donor
`and detecting the sensitized emission of the acceptor. This
`approach is suitable for proving the existence of energy
`transfer, but the accurate determination of the transfer
`
`(determined using only acceptor labeled cells); and
`
`S3
`
`⫽ I3/I1
`
`(determined using only donor labeled cells).
`The three detected intensities can be expressed as
`
`I1(488 = 535) ⫽ IF(1 ⫺ E)
`
`(12)
`
`I2(488 = ⬎590) ⫽ IF(1 ⫺ E)S1
`
`⫹ IRS2
`
`⫹ IFE␣ (13)
`
`I3(514 = ⬎590) ⫽ IF(1 ⫺ E)S3
`
`⫹ IR
`
`⫹
`
`S3
`S1
`
`IFE␣ (14)
`
`I1 is smaller than IF because the energy transfer causes
`donor quenching. I2 consists of three additive terms: 1) the
`overlapping fraction of the quenched fluorescein inten-
`sity, 2) the direct contribution of rhodamine, and 3)
`sensitized emission due to energy transfer. The proportion-
`ality factor ␣ is the ratio of I2 for a given number of
`rhodamine molecules and I1 for the same number of
`fluorescein molecules. ␣ is constant for each experimental
`setup, and has to be determined for every defined case. I3
`is a sum of: 1) a fraction of the quenched fluorescein
`intensity, 2) the rhodamine intensity, and 3) the sensitized
`emission of rhodamine due to energy transfer corrected
`for the lower molar extinction coefficient of fluorescein at
`514 nm than at 488 nm.
`From Equations 12–14, the following equation can be
`derived:
`
`(15)
`
`⫺ S2I3)
`3 (I2
`S3S2
`11 ⫺
`S1 2 I1
`
`⫺ S14
`
`1 ␣
`
`⫽
`
`E
`1 ⫺ E
`
`All the parameters in Equation 15 can be determined
`experimentally. If we substitute B in the right side of
`Equation 15, E can be expressed as follows:
`
`E ⫽
`
`B
`1 ⫹ B
`
`(16)
`
`Because fluorescein is used as a donor and rhodamine as
`an acceptor in most flow cytometric energy transfer
`experiments, the above considerations apply to them. The
`equations are valid for other donor acceptor pairs, but the
`different spectral characteristics must be considered.
`In Equations 12–15, it is assumed that the contribution
`of cellular autofluorescence to the specific fluorescence
`signals is negligible. If the autofluorescence is substantial,
`corrections should be done. In this case, however, the
`correction for autofluorescence can be done using the
`average autofluorescence intensities of the entire cell
`population. It should be noted that since in most cell types
`there is a good correlation between the autofluorescence
`
`Page 5 of 21
`
`Fluidigm
`Exhibit 1012
`
`

`
`164
`
`SZO¨ LLO9 SI ET AL.
`
`efficiency is not possible. More recently, others (83,132)
`used a similar approach with slight modifications.
`In the family of the intensity-based microscopic meth-
`ods, two relatively new versions emerged. The first (79)
`uses a set of equations similar to those described by Tro´n
`(109), whereas the other calculates corrected ratio images
`taken from the donor and the acceptor as well (54). Both
`versions are suitable for determining the transfer efficien-
`cies on pixel basis.
`
`APPLICATION OF FRET
`Research
`Cell surface distribution of hematopoietic cluster
`of differentiation (CD) molecules. A nonrandomized
`codistribution of membrane-bound proteins play an impor-
`tant role in signal transduction across the cell membrane.
`The primary target for external stimuli
`is the plasma
`membrane. In addition to the well-known biochemical
`mechanisms of ligand-receptor interaction there are numer-
`ous physical events that induce alterations of the cell
`surface in the vicinity of the receptor. Signal transduction
`is often accompanied by the dynamic rearrangement of
`the two dimensional patterns of the macromolecular
`constituents at the cell surface.
`The most commonly used methods for determining
`molecular associations in membranes include cocapping,
`co-immunoprecipitations, chemical cross-linking, modi-
`fied fluorescence recovery after photobleaching, electron
`microscopy, atomic force microscopy, and FRET measure-
`ments. Application of one method alone usually does not
`give a conclusive picture. Results of two or more methods,
`however, reinforce each other and lead to a consistent
`picture of the molecular interactions in the plasma mem-
`brane. We will focus mostly on results gained by FRET
`technique, some of the supporting data obtained by other
`biochemical and biophysical methods will also be men-
`tioned. There have been several general reviews on
`mapping cell surface elements using FRET technique
`(17,18,21,70,71,98,108,127); this chapter will focus mainly
`on the latest results concerning this field.
`Although the major histocompatibility complex (MHC)
`class I and class II molecules are structurally and function-
`ally distinct, there is evidence that MHC class I and class II
`molecules can be more intimately related than previously
`thought. It has been reported that MHC class I-specific
`antibodies can cocap MHC class II antigens on B lympho-
`cytes (77,78). These observations prompted us to perform
`studies in which a more direct approach, FRET technique,
`was applied for the investigation of the possible associa-
`tion between HLA class I and class II molecules on PGF and
`JY B lymophoblastoid cells (99). A panel of monoclonal
`antibodies specific for various class I and class II antigens
`was labeled with either FITC (donor) or TRITC (acceptor).
`Flow cytometric energy transfer measurements were made
`on cells labeled with fluoresceinated and rhodaminated
`antibodies simultaneously. FRET efficiency was calculated
`on a cell-by-cell basis and the results were displayed as
`energy transfer distribution histograms. Mean values of
`such distribution histograms were used to draw conclu-
`
`sions about the proximity relationship of cell surface
`proteins under investigation. This type of FRET study
`revealed that HLA class I and class II molecules are
`expressed mostly in a monomeric state on the cell surface.
`Class II antigens may form heteroassociations among
`themselves and there is association between class I and
`class II molecules. There is an equilibrium between free
`nonassociated HLA class I and class II molecules and the
`associated class I-class II antigen complexes. The number
`of class I-class II molecule associations formed depends
`upon the available class I and class II molecules expressed
`in the plasma membrane. These results are also in agree-
`ment with cocapping experiments demonstrating class
`I-class II interaction. Our data, however, demonstrated
`that these complexes are physically associated before
`cocapping (99). While there was no homoassociation
`between HLA class I molecules on PGF cells, we could
`detect homoassociation between them on JY cells. The
`degree of homoassociation of class I antigens highly
`depends on the culturing conditions, whether the cells
`were in log phase or in plateau phase. HLA class I
`clustering correlated with the expression level of ␤2-
`microglobulin-free HLA class I heavy chains, and the
`addition of exogenous ␤2-microglobulin greatly reduced
`the HLA class I homoassociation (9). Moreover, modula-
`tion of the composition of plasma membrane also influ-
`enced the HLA class I clustering. Addition of cholesterol
`decreased the membrane fluidity and also the degree of
`homoassociation of HLA class I molecules (9).
`We have extended these types of flow cytometric
`energy transfer measurements and molecular associations
`have been detected between intercellular adhesion mol-
`ecule 1 (ICAM-1; CD54), HLA class I heavy chain, ␤2-
`microglobulin, and HLA-DR on the cell surface of JY B
`lymphoma cells (6). Similar heteroassociations were found
`in the plasma membrane of HUT-102B2 T lymphoma cells,
`but the significantly different quantitative data suggested a
`somewhat different cell surface distribution pattern of the
`molecules. In this case, interleukin-2 (IL-2) receptor ␣
`subunit (IL-2R␣, CD25), was also included in the proximity
`studies. FRET data suggested that ICAM-1 and/or IL-2R␣
`are closer to the ␤2-microglobulin than to heavy chain of
`the HLA class I complex. In addition, a high degree of
`homoassociation of ICAM-1 molecules was observed. Het-
`eroassociations involving ICAM-1 molecules may play an
`important role in antigen presentation, T-cell recognition,
`cytotoxicity, site-directed lymphokine secretion, and other
`immunological processes (6).
`Flow cytometric energy transfer measurements have
`also been used to study the topological distribution of
`transferrin receptor (TfR; CD71) relative to the heavy and
`light chains of the HLA class I molecules, class II mol-
`ecules, interleukin receptor ␣-chain (CD25), and ICAM-1
`(CD54) molecules on HUT-102B2 T and JY B cell lines
`(72). TfR showed high degree of homoassociation, and its
`cell surface distribution depended upon the growing
`condition of cells. TfR was in close vicinity to HLA class I
`molecules on the surface of JY cells in both logarithmic
`and plateau phase, whereas it was not associated with HLA
`
`Page 6 of 21
`
`Fluidigm
`Exhibit 1012
`
`

`
`FRET IN CLINICAL LABORATORY
`
`165
`
`description of the clustering of MHC class I molecules at 2
`to 10 nm and also at µm levels (23). Using flow cytometric
`and photobleaching energy transfer measurements a non-
`random distribution pattern of HLA class I molecules was
`observed in the 2 to 10 nm range. A second, nonrandom,
`and larger-scale topological organization of the HLA class I
`molecules was detected by transmission electron micros-
`copy and atomic force microscopy using immunogold
`labeling. These data suggested that HLA class I antigens
`exhibit a hierarchical arrangement consisting of specific
`patterns of localizations but with a degree of randomness
`in the distribution. The possible function of the higher
`order clusterization is that by increasing the local density
`of adhesion molecules and thereby the multiplicity of
`intercellular interactions, clusters may serve to stabilize
`weak contacts between cells (23). A similar hierarchical
`distribution pattern was observed for HLA class II mol-
`ecules. Electron microscopy also revealed that a fraction of
`the HLA class II molecules was heteroclustered with HLA
`class I molecules at the same hierarchical level (46).
`CD7 is a 40-kDa glycoprotein that is expressed on a
`major subset of human peripheral blood T cells. Cross-
`linking of CD7 monoclonal antibody (mAb) is mitogenic
`and signals delivered via CD7 molecule stimulated integrin-
`mediated adhesion (59). Co-immunoprecipitation data
`suggested that CD7 associate with CD3 and CD45. To
`confirm this observation, flow cytometric energy transfer
`measurements were performed using FITC- and TRITC-
`labeled mAbs as donor acceptor pairs. There was signifi-
`cant increase in the sensitized emission when FITC-CD7
`and TRITC-CD45, or FITC-CD7 and TRITC-CD3 interaction
`wa