Fluorescence modulation sensing of positively and negatively charged proteins on lipid bilayers
Detecting ligand-receptor binding on cell membrane surfaces is required to understand their function and behavior. Detection platforms can also provide an avenue for the development of medical devices and sensor biotechnology. The use of fluorescence techniques for such purposes is highly desirable as they provide high sensitivity. Herein, we describe a technique that utilizes the sensitivity of fluorescence without directly tagging the analyte of interest to monitor ligand-receptor interactions on supported lipid bilayers. The fluorescence signal is modulated according to the charge state of the target analyte. The binding event elicits protonation or deprotonation of pH-responsive reporter dyes embedded in the lipid bilayer.
Supported lipid membranes containing ortho-conjugated rhodamine B-POPE (1-hexadecanoyl-2-(9Z-octadecenoyl)-sn-glycero-3-phosphoethanolamine), which fluoresces in its protonated but not in its deprotonated form, were utilized as sensor platforms for biotin-avidin and biotin-streptavidin binding events. The membranes contained 5 mol% biotin-PE (1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(cap biotinyl) (sodium salt) as a capture ligand. Supported lipid bilayers were formed in the channels of microfluidic devices and the fluorescence intensity of the dye was monitored as protein was introduced.
The binding of avidin, which is positively charged at pH 7.2, made the bilayer surface charge more positive, which in turn deprotonated the ortho-rhodamine B dye, reducing its fluorescence. The binding of streptavidin, which is negatively charged at pH 7.2, had the opposite effect. Reducing the ionic strength of the analyte solution by removing 150 mM NaCl from the 10 mM phosphate buffered saline (PBS) solution raised the apparent pKa of the ortho-rhodamine B titration point by about 1 pH unit. This could be exploited in conjunction with bulk solution pH changes to turn the rhodamine B-POPE dye into a sensor for streptavidin involving a decrease, rather than an increase, in the fluorescence response, at pH values below streptavidin’s pI value.
This study demonstrates the ability to monitor ligand-receptor interactions on supported lipid bilayers through the protonation or deprotonation of reporter dyes for both negatively and positively charged analytes over a range of pH and ionic strength conditions. Specifically, the sensitivity and pH-operating range of this technique can be optimized by modulating the sensing conditions which are employed.
The ability to monitor ligand-receptor interactions is vital for biotechnological advances as well as for a fundamental understanding of cell biology. The single-molecule sensitivity that fluorescent labeling can provide and the relative ease of fluorescence measurements has resulted in the utilization of this technique for detecting and monitoring proteins and nucleotides. Dye labels can, however, interfere with analyte activity and be difficult to efficiently employ [1, 2]. Many sensing techniques have consequently been developed such as quartz-crystal microbalance (QCM), [3, 4, 5, 6] surface plasmon resonance (SPR), [7, 8, 9, 10] semiconductor nanowire conductivity, [11, 12] and optical microcavities that avoid the use of fluorescent labels . Although these techniques avoid the problems associated with directly attaching fluorescent tags to analytes, they can behave non-linearly in response to analyte concentration, require specialized equipment or procedures which can be difficult and/or costly to employ, or possess poorer detection limits than fluorescence-based techniques.
It is advantageous to be able to monitor ligand-receptor interactions on supported lipid bilayers  because these platforms effectively act as simplified cellular surfaces, incorporating the same lipid molecules as well as maintaining two-dimensional lipid fluidity . Indeed, the majority of current drug targets are receptor sites associated with the cell surface . This makes the ability to effectively monitor such systems key for therapeutics and drug discovery. Consequently, much work has been done to produce supported lipid bilayers that mimic the cell surface with various supports [17, 18, 19, 20] and tethers [21, 22, 23, 24, 25, 26]. By embedding these bilayers into microfluidic devices,  one can probe binding ligands rapidly while utilizing only small amounts of analyte. Heterogeneous flow-based microfluidic techniques also enable the user to easily modulate parameters such as ionic strength and buffer pH while monitoring the fluorescent response in real time. The ability to utilize glass substrates as membrane supports is also advantageous as these surfaces can readily accommodate fluid lipid bilayers via vesicle fusion [28, 29].
Recently, we have developed a sensing technique that employs the sensitivity of fluorescent measurements while avoiding the complications incurred by directly labeling the analyte of interest . This technique involves the incorporation of a pH sensitive dye, ortho-conjugated Texas Red DHPE (Texas Red 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine, triethylammonium salt), into a supported lipid bilayer. This isomer of Texas Red loses fluoresces when it is deprotonated at higher pH. On the other hand, the para-conjugated Texas Red isomer retains its fluorescence under basic conditions and therefore can serves as a reference dye . These Texas Red-containing bilayers can also accommodate binding ligands such as biotin-PE (1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(cap biotinyl) (sodium salt)). Upon the binding of a negatively charged protein, the surface potential near the binding site will be decreased. This helps partition hydronium ions to the interface, which in turn causes protonation of the ortho-Texas Red isomer and increases the fluorescence response. This sensing approach, by monitoring the change in fluorescence signal, was demonstrated using the biotin/anti-biotin binding pair . The assay shows sub-pM sensitivity and could be performed with low analyte volumes inside a microfluidic device.
2.2 Rhodamine B-POPE Conjugation and Separation
A solution of 10 mg of rhodamine B sulfonyl chloride in 1 mL of chloroform was mixed with 10 mg of POPE and 2 μL of diethylamine. The reaction was stirred for 1 h at room temperature. After the reaction, the solution was spotted onto a TLC plate (EMD, 5715-7, silica gel 60 F254) and developed in a 92:8 mixture by volume of chloroform to methanol to yield separated rhodamine-B labeled POPE isomers. The ortho-rhodamine B-POPE isomer was further purified on a second TLC plate developed with an ammonium hydroxide solution, dichloromethane, and n-propanol mixture of 5:60:35 by volume. The successful labeling of the phospholipid was confirmed with matrix-assisted laser desorption ionization mass spectrometry, which can be seen in Additional file 1: Online Resource 1.
2.3 Small Unilamellar Vesicles (SUVs)
To make SUVs, lipids of the desired molar composition were mixed in chloroform and then dried by purging with nitrogen gas. Following desiccation under vacuum for at least 1 h, 0.5 mg/mL lipid suspensions were obtained by hydration with 10 mM phosphate-buffered saline (PBS) containing 150 mM NaCl. The SUV solutions were prepared by ten freeze-thawing cycles with liquid nitrogen and warm water followed by ten extrusion cycles through two stacked 0.1-μm polycarbonate membranes (Whatman) using a Lipex extruder (Northern Lipids Inc., Vancouver, Canada). The size of the SUVs was determined to be 90 ± 10 nm in diameter by dynamic light scattering (Brookhaven Instruments 90Plus Particle Size Analyzer).
2.4 Microfluidic Device Fabrication
Microfluidic devices consisting of a PDMS (polydimethylsiloxane) mold created with a channel-patterned glass substrate were formed by soft lithographic methods . First, photoresist (Shipley 1827) was spin coated onto clean glass slides (soda-lime glass slides, Corning) which were then exposed to UV light through a Kodak technical pan film photomask consisting of the appropriate microfluidic channel pattern. These photoresist-coated glass slides were treated in developing solution and baked overnight at 120°C. The thickness of the photoresist layer was determined to be 2.3 μm to 2.4 μm using a Dektek 3 Stylus Profilometer. The patterned glass slides were then etched by immersing them into a buffered oxide etchant (BOE) solution . The BOE solution was prepared by mixing 48% HF (EMD Chemicals Inc., Germany) and aqueous NH4F (200 g/300 mL purified water, Alfa Aesar, Ward Hill, MA) in a 1:6 volume ratio. The etching process was repeated until the depth of the channels was between 15 μm and 19 μm. The slides were then washed with acetone to remove the remaining photoresist and dried under flowing nitrogen gas. Next, PDMS (Sylgard silicone elastomer-184) and its cross-linking agent (10:1 volume ratio) were mixed in a plastic beaker and degassed inside a desiccator under vacuum. The liquid PDMS mixture was poured over a patterned glass slide and allowed to cure in an oven at 70°C. After 1 h, the nascently-formed PDMS elastomer was peeled off of the glass slide. This mold and a freshly cleaned glass coverslip were placed into a 25 W oxygen plasma cleaner (PDC-32 G, Harrick) for 30 s and were then immediately brought into contact to form the PDMS/glass microfluidic device. It should be noted that both the glass slides and coverslips used in these procedures were cleaned in a boiling solution of ICN 7X detergent (Costa Mesa, CA) mixed with purified water at a 1:4 volume ratio. Following this, the slides were annealed in a kiln (Sentry Xpress 2.0, Orton Ceramic Foundation, OH) at 500°C for 5 h before the oxygen plasma treatment.
2.5 Supported Lipid Bilayers (SLBs)
SLBs were formed within the channels of microfluidic devices via spontaneous vesicle fusion . Five μL of an SUV solution was injected into each microchannel and incubated for 10 min followed by rinsing with pure PBS buffer to remove excess vesicles. For protein detection measurements, solutions of either avidin or streptavidin were continually flowed through each channel for up to 1 h.
2.6 pH Titration of SLBs
PBS solutions with and without 150 mM NaCl were prepared at pH values ranging from 4.5 to 10.5 by mixing with appropriate amounts of Na2HPO4, NaH2PO4, or Na3PO4. The pH was measured with a standard glass electrode setup having an absolute error of ± 0.1 pH units. Titration curves of the fluorescence from the dye molecules in the SLBs were obtained by systematically changing the pH of the bulk solution by flowing the appropriate pH adjusted buffer solutions through the microfluidic device in a stepwise fashion. The value of the pH for each buffer was determined immediately prior to adding it to the microfluidic channels.
2.7 Fluorescence Microscopy
All fluorescence images were obtained using a Nikon Eclipse Ti-U fluorescence microscope (Tokyo, Japan) equipped with a ProEM 1024 CCD camera (Princeton Instruments). A 10× air objective (N.A. = 0.45) was used for all detection measurements. The light source was a Lumen 200 (Prior Scientific) and a Texas Red filter set (Chroma Technology, Bellows Falls, VT) was employed. All images were processed with MetaMorph software (Version 18.104.22.168, Universal Imaging).
3 Results and discussion
3.1 pH Titration Curves
3.2 Sensing Avidin and Streptavidin
3.3 Altering Buffer Conditions to Optimize Assay Design
Summary of protein signal changes for various experimental conditions
Protein (300 nM)
10 mM PBS Buffer Conditions
Average % Change in Signal upon Protein Binding
pH 7.2 w/ 150 mM NaCl
pH 5.9 w/ 150 mM NaCl
pH 7.2 w/ 150 mM NaCl
As can be seen from Table 1, the results from the low ionic strength experiments were remarkable. In fact, the percentage change in fluorescence upon the introduction of streptavidin was approximately a factor of three greater than when 150 mM NaCl was present. The change for avidin was even larger, a factor of six. This is somewhat curious, as avidin should contain a smaller charge at pH 8.1 compared with 7.2, as it is closer to its isoelectric point. As such, the relatively large change in signal for avidin must be dominated by the increase in the Debye length under these conditions.
Two important considerations needed to be taken into account when employing the pH modulation sensor described herein. First, it is necessary to operate this assay in a pH window where the fluorescence response can be modulated. Second, it is important that proteins of interest be tested sufficiently far away from their pI values to elicit a fluorescence response. Through manipulation of the ionic strength of the operating buffer, it is possible to widen the assay range in order to accommodate proteins with a variety of different pI values. However, it should also be noted that the surface chemistry of the membrane and the underlying solid support play an important role in determining the surface potential. These variables too could be modulated by changing the substrate type or specific lipid membrane chemistry. Additional file 1: Online Resource 5 demonstrates that the addition of charged lipids to the membrane, the addition of osmolytes to the buffer, or even changing the nature of the buffer ions can all affect the titration curve for the dye. Of course, modifications to the dye itself would also change the range in which the assay can be operated.
Although numerous variables can be manipulated to change the operating range of the assay, it should be noted that changing these variables may also modulate the detection limits of this assay. Such behavior could be quite complex. Indeed, in the case of avidin, running the assay at lower ionic strength, but higher pH, (8.1 instead of 7.3) resulted in a greater signal response despite the fact that the protein was actually closer to its isoelectric point (Figure 6). In this particular case, the lower ionic strength was the determining factor. Moreover, it could also be shown that decreasing the ionic strength of the buffer at constant pH increased the sensitivity for streptavidin detection (Additional file 1: Online Resource 6). Further studies will be required to determine if other variables such as analyte charge density or specific chemistry can be the determining factors in fluorescence changes. Such studies are now being undertaken.
This pH modulation sensor represents a facile and rapid method for detecting binding events on supported lipid bilayers. As such, it may serve as a screen for monitoring protein-ligand interactions in biophysical assays of membranes. Further quantification of this technique should be possible by exploiting QCM-D [3, 4, 5, 6] or SPR [7, 8, 9, 10] as complementary approaches to help determine the number density of analyte proteins and perhaps even the charge per analyte. Moreover, complementary techniques may also help quantify potential interference effects and related platform biofouling effects when more complex, real-world biological samples are analyzed.
This work was funded by the Office of Naval Research (N00014-08-1-0467).
- 33.Skander M, Humbert N, Collot J, Gradinaru J, Klein G, Loosli A, Sauser J, Zocchi A, Gilardoni F, Ward TR: Artificial metalloenzymes: (Strept)avidin as host for enantioselective hydrogenation by achiral biotinylated rhodium-diphosphine complexes. J Am Chem Soc 2004,126(44):14411–14418.CrossRefGoogle Scholar
- 38.Israelachvili JN: Intermolecular and Surface Forces. 3rd edition. Academic Press, San Diego; 2011.Google Scholar
- 39.Metzler DE, Metzler CM: Biochemistry: The Chemical Reactions of Living Cells. vol v. 1. Harcourt/Academic Press, San Diego; 2001.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.