One of the most widely used techniques in drug discovery and basic biology is fluorescence resonance energy transfer (FRET), which utilizes the transfer of energy between two fluorophores, a donor and an acceptor. When two partners, each coupled with a fluorescent label, come in close proximity, excitation of the donor by an energy source triggers energy transfer toward the acceptor, which in turn emits specific fluorescence at a given wavelength.
Buffers, proteins, and even compounds of chemical libraries with intrinsic fluorescence commonly hamper FRET chemistries (i.e., GFP, AlexaFluors, BODIPY dyes, and others). If FRET fluorophores have a fluorescent lifetime with orders of magnitude longer than background fluorescence, then the energy transfer could be measured after the interfering signal has completely decayed.
Time-resolved FRET (TR-FRET) has emerged as a method that utilizes long-lived fluorophores to enable the measurements to be delayed by 50–150 microseconds. This time delay allows the signal to be cleared of most nonspecific short-lived emissions.
Cisbio (www.htrf.com) has developed HTRF®, a TR-FRET technology based on the long-lived fluorescence of europium cryptate. This rare-earth complex consists of a macrocycle cage with an embedded Eu3+ion. The company has developed several acceptors for Eu-cryptate, including XL665, a phycobilliprotein pigment, and d2, a small synthetic motif.
Applications for the novel technology were recently discussed at a Cisbio HTRF user group meeting in Sonoma, CA.
“Our FRET pairs exhibit several unique characteristics that make them indispensable for the high-throughput screening stage of drug development,” says Francois Degorce, head of HTRF marketing. “Stability of chemistries is a critical parameter affecting performance of an assay.”
“Not only does Eucryptate form stable conjugates,” he notes, “but it can be easily coupled to a variety of biomolecules.” It is not subject to photobleaching, a change that affects many previously known fluorophores. “Our donor-acceptor pairs consistently demonstrate high signal-to-background ratios. Moreover, HTRF-based assays perform equally efficiently from 96- to 1,536-well formats and in volumes as low as 4 µL.”
Cisbio now owns a portfolio of more than 100 reagents and assay kits based on this technology, including assay platforms for GPCR and kinase screening, such as the KinEASE™ platform developed in collaboration with Millipore.
Homogeneous Assays for Kinase Inhibitors
In the last decade, protein kinases have become hot targets for drug discovery. Kinases play an integral role in intracellular signaling cascades. Dysregulation of signaling pathways has been implicated in a host of diseases including cancer, diabetes, hypertension, cardiopulmonary disorders, and many others.
Human genome sequencing has identified 518 putative kinases (known as the kinome), roughly divided into eight main phylogenic clusters. Over 200 kinases have been cloned and are now available as screens for kinase inhibitors. Primary high- throughput screening aims to select a few hits from a compound library against a desired target, whereas selectivity assays create a profile of a few selected hits against the kinome. And yet, despite an avalanche of available information about structure and function of kinases, only 128 kinase inhibitors are making their way through clinical trials.
After 20 years of work on genetically well-established targets, only a handful of approved compounds have emerged. “This statistic underlies the challenges in effective lead discovery,” comments Ulf Bömer, Ph.D., head of technology development and HTS infrastructure at Bayer Schering Pharma (BSP; www.schering.de). “Different types of kinase assays have different needs, thus requiring a specific technology for each assay. We could, however, achieve greater synergies between various projects if only the assay formats were standardized, so that over 90% of data is in a single format.”
BSP began its standardization process by decoupling compound preparation from the actual assays. Dr. Bömer’s group prepared multiple replicas of 1,536-well compound plates (50 nanoliters/well) that were frozen and ready-to-use in all ongoing assays. “This step ensured comparability of results and increased productivity,” adds Dr. Bömer. “Now, we only need one person to prepare 300,000 samples per day. In addition, we needed a single assay technology that could deliver cost efficiency, sensitivity, and robustness in our miniaturized assay format.”
After a thorough comparative assessment, HTRF was chosen as a core kinase screening technology by the BSP. In a typical assay, a specific biotinylated peptide (kinase substrate) becomes phosphorylated by a kinase. The addition of the phosphate is detected by FRET between a phospho-specific antibody coupled with Eu-cryptate and a streptavidine-XL655 conjugate. While tyrosine-phosphate recognition is independent of the peptide sequence, detection of a phosphorylated serine or threonine requires an antibody specific for each substrate.
“We used a biotinylated peptide library from Jerini Peptide Technologies (www.jerini.de) and a collection of 250 Ser/Thr phosphospecific antibodies. “Normally, we have a specific peptide substrate-antibody pair for a new kinase within a week,” says Dr. Bömer. “HTRF assays use 10–100 times less enzymes than scintillation proximity assays (SPA). When using a generic secondary antibody conjugated with Eu-cryptate, the cost per well is a fraction of the cost of SPA.”
Screening of Natural Products
Over the last 25 years, nearly 50% of all new chemical entities have originated from natural products. Australia’s Eskitis Institute for Cell and Molecular Therapies (www.eskitis.org.au) focuses on the development of drugs from natural products against a range of cell-signaling targets. It holds a library of 42,000 archived biota samples originating from Queensland, China, Tasmania, and Papua New Guinea, and has been undertaking a natural product discovery collaboration with AstraZeneca (www.astrazeneca.com).
“Screening natural product extracts is far from trivial,” comments Michael Jobling, Ph.D., a research fellow in the lead discovery biology group. “Because solubility is an issue, we have to repeatedly re-optimize experimental conditions. At the same time, we have to work with limited samples. Thus, for our screening purposes, the detection system has to be extremely sensitive and able to perform under homogeneous reaction conditions.”
One of the projects studies heparanase (Hpa), an enzyme that cleaves to heparan sulphate proteoglycan (HSPG), a major component of extracellular matrix, vital for its structural integrity. Degradation of HSPGs by heparanase facilitates entry of tumor cells into the tissues. Inhibition of Hpa expression in emerging tumor cells may prevent tumor invasion and metastasis.
“The classical assay for Hpa activity measures the release of the fluorescent label from immobilized HSPG substrate,” continues Dr. Jobling. “This is a multistep procedure that takes up to 72 hours to complete.” It also requires large quantities of natural extracts for screening. “We switched to an HTRF-based assay because it is a one-tube, one-hour procedure, and it requires less material.”
The HTRF heparanase assay is based on HSPG substrate labeled with biotin on one end and Eu-cryptate on the other end. Streptavidin-XL665 binds to the substrate, causing energy transfer and a fluorescent signal. When an active enzyme cleaves to the substrate, the biotinylated end is released, and the emission is reduced. Dr. Jobling adds, “The only caveat is that sometimes the cleavage occurs on the wrong end of Eu-cryptate, resulting in incomplete inhibition of the signal.”
HTRF assay works in a homogeneous format, when all components of the assay remain in the well at the time of readout. The acceptor fluorescence could be affected by media variability in each well. A correction method accounting for well-to-well variations is based on measuring the ratio of fluorescent signals of the acceptor (665 nm) to Eu-cryptate (620 nm). Since the emission of Eu-cryptate is affected in the same proportion as the acceptor, the ratio of fluorescent signals is reflective of the interaction studied.
“We found, however, that natural libraries are full of nuisance compounds affecting 620 nm fluorescence, and resulting in false-positives,” says Dr. Jobling. “To remove the artefacts we simply looked at the average 620 baseline across the whole screening campaign. Any sample that showed a 620 reading outside of 80 percent to 150 percent of the average was discounted from the final hit list.”
“Dysregulation of CD4+ T-cell cytokine production directly contributes to chronic inflammation, autoimmune disease, and asthma,” says Jennifer Brogdon, Ph.D., research investigator, Novartis Institute for Biomedical Research (www.nibr.novartis.com). Many existing chemical entities that target T cells are general immunosuppressants and thus have broad effects on the immune system.
“These mechanisms of action do not get to the heart of the problem. One of the goals of our group at the developmental and molecular pathways department is to find selective inhibitors of Th1 and Th2 pathways, without a priori knowledge of their target.”
The lack of appropriate cell lines that mimic T-cell cytokine regulation only complicates the task. “We chose to use primary cells selectively differentiated into Th1 and Th2 types.”
Because of the limited source of primary cells and the need to keep the screening assay suitable for automation, an assay with an “add-add-add” format was highly desirable. “After considering several different platforms, we chose Cisbio’s HTRF since it had been shown to work in a cell-based format and could be miniaturized,” continues Dr. Brogdon. “The company not only offered several cytokine assays, including IFN-a, but also provided custom antibody labeling and assay optimization services.”
T cells were plated on compounds, stimulated with CD3/28+ beads, and incubated with two antibodies recognizing IFN-a epitopes. TR-FRET reaction between Eu-cryptate and XL-665 conjugated antibodies demonstrated a 7–10 fold signal-to-noise ratio in a fully automated screening format.
Cisbio further assisted Novartis (www.novartis.com) in the development of a custom kit for detecting IL-5. Cisbio identified the best antibody pair based on binding affinities and conjugation with the fluorophores, and validated the assay on supernatants provided by Novartis. “After miniaturization to a 1,536- well format with only 1,500 to 2,000 cells per well, the assays were screened against 400,000 compounds. Several clear hits are now being pursued,” adds Dr. Brogdon.
Confirming GPCR Dimerization
The GPCR superfamily includes approximately 450 genes found in every cell of the human body, as well as in other organisms including insects. GPCRs play a critical role in numerous downstream signaling events and, thus, are a target of intense drug development efforts. There are many indications that GPCRs may exist as homodimers and heterodimers, and could even form an oligomeric assembly.
“If two GPCRs came together and formed a new entity, information flow within the cell could be significantly altered,” notes Jean-Philippe Pin, Ph.D., Institut de Génomique Fonctionnelle (www.igf.cnrs.fr). “If dimerization correlates with a disease state, then these dimers could represent specific targets for drug development. But for a few instances, there is no clear evidence if dimers actually form in native cells, and if they are stable.”
Both FRET and BRET (bioluminescent resonance energy transfer) are commonly used to study GPCR interactions, but both have several significant limitations. In these systems, recombinant GPCRs are fused with C-terminal tags, where GFP or luciferase serves as an energy donor and YFP as an energy acceptor. The recombinant GPCRs are expressed in the cell and then transported via endoplasmic reticulum and Golgi to the cell surface. Neither FRET nor BRET is able to distinguish dimers on the cell surface from those in the intracellular compartments.
“Another issue with FRET, using green or yellow fluorescent proteins, is that the energy transfer is dependent on the spatial orientation of the fluorophores. If they are not aligned at a particular angle, FRET does not occur, even though the proteins may be in close proximity,” adds Dr. Pin. “We chose HTRF because it is not subject to this orientation constraint, it possesses excellent spectral properties, and we can selectively detect the events on the cell surface.”
The group modified metabotropic glutamate (mGlu) GPCRs with short HA or myc tags on N-termini. The constructs were also modified on C-termini with two distinct affinity tags (c1 and c2) for COP I protein. Only dimers between c1 and c2 would pass through the endoplasmic reticulum (ER) onto the cell surface, whereas c1-c1 or c2-c2 dimers are retained in the ER by COP I. The dimers on the cell surface were detected by TR-FRET between an anti-HA antibody coupled with Eu-cryptate and anti-myc-antibody coupled with d2.
“By manipulating transfection conditions, we established that HTRF was sensitive at physiological densities of GPCRs at the cell surface,” adds Dr. Pin. “We found no evidence for mGlu dimers of dimers, but we detected dimers of heterodimers of GABAB, a receptor for the major inhibitory neurotransmitter in the mammalian CNS.”
Dr. Pin’s group adapted the assay for 96- and 384-well plates and is currently working on a system where antibodies are replaced with an SNAP technology. Not only will this approach allow confirmation of the oligomeric state of GPCRs but it would also be useful for the analysis of the interactions between any cell surface proteins in an HTS-compatible format.