|Home | About | Journals | Submit | Contact Us | Français|
We have combined our fluorogen activating peptide with a new tandem dye molecule to develop a biosensor that labels a cell-surface protein and displays an easily detectable pH dependent emission color change by efficient intramolecular Förster resonant energy transfer. This probe has demonstrated pH variations in β2-adrenergic receptor trafficking and revealed a process of surface to endosome inter-cellular transfer in dendritic cells with potential significance in antigen transfer.
The fate and localization of proteins through the endocytic pathway is of interest across a broad range of biological investigations because inappropriate trafficking of surface proteins can lead to various diseases[2-4]. Bulk methods for pH measurement where pits are loaded with two dyes, one pH dependent (such as fluorescein) and the other pH independent (such as rhodamine), provide the aggregate endolysosomal pH rather than trafficking-related changes associated with a single protein. Genetically tagged fluorescent proteins (pHluorins) exhibit activation (ecliptic) or excitation ratio (ratiometric) signatures in response to pH changes; but these reporters are not selective for surface proteins, requiring image-based segmentation to select the subset of cell surface proteins. Typical “surface selection” is accomplished using total internal reflection fluorescence (TIRF), imaging up to ~100 nm depth, far larger than the 5-10 nm thickness of the plasma membrane. This, however, does not truly select for events occurring at the surface of the cell, nor does it allow tracking of these events further into the cellular volume than the TIRF field.
For truly cell surface selected imaging, antibodies are conjugated to the pH dependent CypHer5 dye. Antibodies bind to cell surface proteins that are accessible without permeabilization (i.e. not in collared pits). The fluorescence increase at low pH allows determination of relative acidity of the internalized vesicles, but determination of the precise pH value is difficult. Typical ratiometric dyes such as SNARF are not available as conjugates. Ligands coupled to ratiometric DNA nanosensors were recently used to track ligand-associated pH changes through the endocytic pathway. Despite these advances, imaging of the cell surface subset of receptor proteins and their fate in live cells remains a considerable challenge
We have previously demonstrated that exposing an expressed fluorogen activating peptide (FAP)[1,13] to a cell impermeant fluorogenic dye labeled the subset of proteins present at the cell surface. This approach chemically discriminates between the proteins at the plasma membrane and those within secretory or endocytic compartments, and allows selective imaging of the transport of these proteins. To convert this label into a physiological indicator, we have coupled a pH dependent Cy5 analog (II) to a fluorogen donor derived from thiazole orange (TO1), at a distance to allow efficient Förster Resonant Energy Transfer (FRET). This provides a cell-excluded reagent that is activated on binding to an expressed tag only when it is exposed at the plasma membrane. The bound dye displays a pH dependent dual-band emission spectrum, a result of the protonation of the Cy5 derivative (II), producing increased spectral overlap and FRET efficiency at low pH.
Figure 1 illustrates the tandem dye and the pH dependent emission spectra of the HL1.0.1-TO1 FAP-tandem dye complex in citrate/phosphate buffer from pH 4.0 to pH 8.0 (Figure 1b). A systematic variation in spectral overlap and Förster radius between the fluorogen TO1 (thiazole orange derivative) and the pH dependent Cy5 derivative (II) leads to a significant change in the ratio of red to green emission (Supplementary Discussion, Table S1 Figure S1and Figure S2). The FAP-tandem dye complex gives the same ratiometric trend both in vitro and on the mammalian cell surface demonstrating a standardizable readout of this pH-biosensor (Figure 1c, Figure S3). The dye shows significant fluorogenic activation (~10-fold) in response to binding to the FAP. The tandem dye binds with its cognate FAP with Kd= 55 nM and with a fluorescence quantum yield of 0.06 (at pH 4). The pKa of the probe determined by the fluorescence ratio is 6.4 +/-0.05.
Such a reagent in effect combines the properties of the ecliptic and ratiometric pHluorins—chemoselective labeling of cell surface proteins that allows subsequent measurement of pH changes in response to trafficking of the labeled molecules (Scheme 1). This type of measurement is not yet available for studies in living cells. The ß2— Adrenergic Receptor (β2AR) is a canonical G-protein coupled receptor with well characterized trafficking. Stimulation of the receptor with an agonist results in rapid clathrin-mediated endocytosis followed by actin-dependent recycling to the plasma membrane[15-18]. This is an important catecholamine receptor that is expressed in brain, heart, lung, and a number of other tissues as well as a therapeutic target for asthma and cardiovascular diseases. Retroviral transduction of 3T3 cells (NIH) (Supporting Information) produced N-terminal labeled β2AR, with the FAP presented to the extracellular milieu. Because of the low KD of the FAP-dye interaction, the concentration and activation of dye by FAPs on the cell surface results in substantially higher signal-to-background activation than the 10-fold seen in solution measurements (100-fold or higher as measured with line-scans— data not shown). The neutral pH of the media evokes minimal fluorescent response from II but bright fluorescence from the FAP-bound TO1 (Figure 2a). Upon agonist activation (isoproterenol, 10 μM) the adrenergic receptor internalizes and rapidly sequesters away from the plasma membrane into the endosomal pathway (Figure S4). On acidification of the vesicles, II is activated by protonation, enhancing the FRET-sensitized red emission (Figure 2b, Movie S1). Drug treatments demonstrated the biological relevance and specificity of this method. Acidification of vesicles was blocked with chloroquine (a general cell permeant base) and bafilomycin A (an ATPase inhibitor)[21,22] 1 and trafficking was altered using Latrunculin A to block actin-dependent recycling of the β2AR[23,24]. Alteration in the biosensor fluorescent ratio was recorded under the influence of these drugs (Figure 2 and Figure S5) and the mean vesicle pH was estimated relative to the nigericin calibration data (Figure S6). Distinct pH phenotypes were seen after manipulation of endosomal acidity or receptor trafficking.
Dendritic cells (DCs) of the immune system are responsible for priming T cell responses to protein antigens to generate adaptive immunity to pathogens. Uptake of antigens is a critical first step in the process and DCs have multiple mechanisms for internalizing antigens, including phagocytosis, macropinocytosis, and receptor mediated endocytosis. In addition, DCs that display antigens internalized by neighbouring cells, a phenomenon termed “cross-dressing“, have been proposed as an amplification mechanism in generating effective immune responses. The precise mechanism by which these cells achieve cross-dressing has not been determined. One proposed mechanism is transfer of cellular components between DCs, where a recipient cell engulfs considerable quantities of plasma membrane of donor cells through intercellular nanotubules and other intercellular contacts, a process called trogocytosis[28-30].
Critical questions in understanding the cross-dressing of DCs are whether whole protein antigens or peptide fragments are transferred between cells, and whether transfer occurs into acidified endocytic compartments within recipient cells, where further processing can occur, or through simple mixing of membrane proteins during cell-cell contact. To address these questions, we used the surface-displayed biosensor as a surrogate protein antigen that was transfected into human monocyte-derived DCs. The interaction between biosensor-expressing DCs and naïve DCs added to the culture was then visualized. As shown in Figure 3 and Movie S2, when pH-sensor labeled dendritic cells (green signal) interact with a naïve cell, biosensor transfer occurs almost exclusively into acidified endocytic compartments, resulting in enhanced red signal from II. This demonstrates that an intact membrane protein (dye-bound biosensor) can be internalized directly into an endocytic compartment of adjacent cells, and that transfer does not result from mixing of plasma membrane during contact. In addition, this signal could not arise from direct transfer of endosomes or other vesicular structures containing the sensor, because these structures would be topologically incapable of binding to dye in the media in transit to the recipient cell. Figure S7 and Movie S3 show transfer from FAP-transfected cells to naïve cells labeled with a cell-surface antibody, revealing transfer of vesicles to the recipient cell through nanotubules, without extension of the donor cell membrane, and not due to active endocytosis in the donor cell. These data support a model for antigen sharing between dendritic cells involving transfer of patches of intact plasma membrane from donor cells through an active endocytic process into recipient cells.
We have demonstrated a genetically targeted surface selected pH biosensor that is useful for dynamic analysis of endosomal trafficking and antigen processing. This sensor is based on a genetically targeted fluorogen linked to a pH sensitive dye that undergoes a significant change in FRET efficiency in response to environmental pH changes. Use of this biosensor to study receptor mediated endocytosis in fibroblast cells revealed typical behavior, and indicated drug-induced alterations in the pH of vesicles in living cells. Because the surface selection is achieved chemically, these measurements potentially extend to relevant anisotropic 3-d culture systems and live animals. The ability to target cellular pathways enables a number of studies with direct disease relevance.
The pH biosensor concept demonstrated here can be extended to a range of different indicators. Dyes that display a change in quantum yield, spectral properties, or extinction coefficient can be converted into targeted sensors that modulate the energy transfer efficiency in response to analytes. Dyes can be made cell permeable, and cells can be treated to allow dye incorporation. Unlike direct fluorescent labeling, the fluorogen biosensor dye does not have to be washed away to reduce the background signal. This represents a significant new class of targeted biosensors for measuring local changes in the cellular environment.
This project was supported by grants from the National Center for Research Resources (5U54RR022241-08) and the National Institute of General Medical Sciences (8 U54 GM103529-08) from the National Institutes of Health. We would like to thank Jonathan W. Jarvik for the gift of stable ADRB2 cells, Yehuda Creeger for help with flow cytometry and Haibing Teng for help with confocal microscopy. We would like to thank Lauren Ernst and Chris Szent-Gyorgyi for helpful discussions and Kristen McConnell for assistance preparing the illustration. Microscopes used for the project were acquired under NIH grants 1S10RR024716 and 1S10RR026766.
Anmol Grover, Department of Biological Sciences and Molecular Biosensor and Imaging Center, Carnegie Mellon University.
Dr. Brigitte F. Schmidt, Molecular Biosensor and Imaging Centre, Carnegie Mellon University.
Dr. Russell D. Salter, Department of Cell Biology and Physiology, University of Pittsburgh School of Medicine, Pittsburgh (PA)
Dr. Simon C. Watkins, Department of Cell Biology and Physiology, University of Pittsburgh School of Medicine, Pittsburgh (PA)
Dr. Alan S. Waggoner, Department of Biological Sciences and Molecular Biosensor and Imaging Center, Carnegie Mellon University.
Dr. Marcel P. Bruchez, Department of Chemistry, Department of Biological Sciences and Molecular Biosensor and Imaging Center, Carnegie Mellon University.