A Wide-Field Fluorescence Microscope Extension for Ultrafast

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Technology Note

A wide field fluorescence microscope extension for ultrafast screening of one-bead one-compound libraries using a spectral image subtraction approach Wolf Heusermann, Beat Ludin, Nhan Pham, Manfred Auer, Thomas Weidemann, and Martin Hintersteiner ACS Comb. Sci., Just Accepted Manuscript • DOI: 10.1021/acscombsci.5b00175 • Publication Date (Web): 08 Apr 2016 Downloaded from http://pubs.acs.org on April 11, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Combinatorial Science is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

A wide field fluorescence microscope extension for ultrafast screening of one-bead one-compound libraries using a spectral image subtraction approach.

Wolf Heusermann1, Beat Ludin2, Nhan Pham,3 Manfred Auer3, Thomas Weidemann4, and Martin Hintersteiner5* 1

IMCF Biozentrum, Universität Basel Klingelbergstrasse 50/70, 4056 Basel, Switzerland

2

Life Imaging Services, Efringerstrasse 79, 4057-Basel, Switzerland

3

The University of Edinburgh, School of Biological Sciences, CH Waddington Building,

3.07, The King's Buildings, Mayfield Road, Edinburgh, EH9 3JD, UK 4

Max Planck Institute of Biochemistry, Am Klopferspitz 18, 82152 Martinsried

5

Bioseutica BV, Corso Elvezia 4, 6900 Lugano, Switzerland

* E-mail: [email protected]

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 28

ABSTRACT

The increasing involvement of academic institutions and Biotech companies in drug discovery calls for cost effective methods to identify new bioactive molecules. Affinity based on-bead screening of combinatorial one-bead one-compound libraries combines a split-mix synthesis design with a simple protein binding assay operating directly at the bead matrix. However, one bottle-neck for academic scale on-bead screening is the unavailability of a cheap, automated, and robust screening platform that still provides a quantitative signal related to the amount of target protein binding to individual beads for hit bead ranking. Wide-field fluorescence microscopy has long been considered unsuitable due to significant broad spectrum autofluorescence of the library beads in conjunction with low detection sensitivity. Herein we demonstrate how such a standard microscope equipped with LED based excitation and a modern CMOS camera can be successfully used for selecting hit beads. We show that the autofluorescence issue can be overcome by an optical image subtraction approach which yields excellent signal-to-noise ratios for the detection of bead associated target protein. A polymer capillary attached to a semi-automated bead-picking device allows the operator to efficiently isolate individual hit beads in less than 20 seconds. The system can be used for ultra-fast screening of > 200 000 bead bound compounds in 1.5 hours, thereby making highthroughput screening accessible to a wider group within the scientific community.


Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION Today, academic institutions play an increasing role in the development of novel and innovative drugs.1,2 The emergence of chemical biology as a new discipline in the postgenomic era is one contributing factor to that development. Consequently, an ever increasing number of proteins and their interaction networks are being studied.3,4 This calls for efficient, miniaturized and economic approaches to identify new bioactive molecules.5–7 Combinatorial Chemistry and High-throughput screening (HTS) certainly represent two enabling technologies for identifying new bioactive molecules.8 A particularly efficient embodiment of combinatorial chemistry, the one-bead one-compound (OBOC) library concept, as invented by Lam et al.9, allows the generation of large libraries of hundreds of thousands to millions of compounds on polymeric carrier beads. In addition, on-bead screening is a particularly easy to use and efficient affinity based screening concept. In on-bead screening (OBS), hit-beads are selected based on the binding of a soluble, tagged target protein or a cell to the immobilized compound on the bead surface.10–12 Each bead therefore represents a highly miniaturized screening assay. For example, a 90 micrometer TentaGel bead contains about 100 pmol of substance in a bead (= assay) volume of less than 1 nl. Thus the amount of chemical substances required for screening and the corresponding assay volumes are amongst the lowest for industrial screening procedures.13 This very efficient initial screening step can then be combined with different follow-up assays for further characterization and ranking of individual hit beads before resources are committed to the re-synthesis of screening hits.14,15 One of the major hurdles for a wider application of on-bead screening is the lack of simple, automated screening platforms. In its simplest implementation, a fluorescently tagged protein is incubated with library beads to assay surface binding to specific bead-immobilized compounds. Selecting hit beads therefore requires a sophisticated image analysis that can distinguish signal contributions associated with the bound target protein from intrinsic bead 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 28

associated autofluorescence. Since library beads, have gone through multi-step chemical treatments of a combinatorial synthesis, autofluorescence levels show strong variation within the bead population and can makes a detection of hit beads ambiguous. Different strategies have therefore been developed to cope with the variable levels of autofluorescence, for example, immobilization of 3-Nitro-Tyrosine as an internal quencher prior to library synthesis or the use of quantum dots as secondary detection aid.16,17 Also, physical separation of beads using a magnet and magnetic particles, linked to a secondary detection reagent in a sandwich assay, has been applied successfully.18–20 While this represents a fast solution for pre-sorting of beads, a method offering a quantifiable signal related to the extent/affinity of target protein binding is still desirable for hit-bead ranking. Furthermore, non-specific binding of the secondary detection reagent in sandwich assays can be hard to control. Currently, the COPAS instrument is the only commercially available on-bead screening platform.21,22 It essentially works like a fluorescence activated bead sorter and thereby allows a fast real-time manipulation of beads, but suffers from the same protein-signal versus bead autofluorescence problem. Pre-sorting of libraries to remove beads with high autofluorescence has been suggested as one possible, but time-consuming workaround.23 Alternatively, a multi-channel real-time analysis was recently shown to enhance the signal-to-noise ratio on the COPAS instrument.24 Our own group has previously developed the PickoScreen instruments, a confocal nanoscanning and bead-picking platform (CONA), exploiting the high-resolution fluorescence imaging capabilities of confocal microscopy.25,26 However, in CONA, the need for serial scanning of large areas at high resolution limits the screening speed. Furthermore, the use of lasers, sensitive confocal optics and a complex bead-picking robot lead to considerable complexity and high instrumental costs, which render this technology inaccessible to many researchers.


Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Here, we describe a set of technical extensions to turn a standard wide-field fluorescence microscope into a fully enabled on-bead screening device. The instrument is built from a minimal set of accessories, including LED excitation, CMOS camera detection, a polymer capillary and a semiautomatic, kinematically guided mechanical slider as a picker device. We developed an image subtraction approach which is essential to reliably detect hit-beads under typical screening conditions despite a significant and variable autofluorescence background. Image processing, combined with the fast wide field instrumentation and the robust, yet simple bead picking option provides for the first time an on-bead screening platform which can be mounted on any standard fluorescence microscope at limited costs with unprecedented screening speed. RESULTS Library bead fluorescence To study the fluorescence properties of typical TentaGel based one-bead one-compound combinatorial libraries, cyclic hexapeptides of the general form NH2-Pra-Glu[-Pro-X-Y-ZAsp-], being X, Y and Z the combinatorial positions, were synthesized using Fmoc- based solid phase peptide chemistry. Fmoc-Glu-OAll was coupled as the second amino acid after the Pra-OH building block via its side chain COOH, to provide a side-chain anchoring of the cyclic peptide. The peptide sequence was then synthesized using standard Fmoc chemistry. Finally, the head-to-tail cyclization was effected on resin after palladium acetate based allylgroup removal. This treatment resulted in a bead population with different levels of autofluorescence, as seen from a CONA image (Figure 1A). Fluorescence excitation- and emission spectra were measured by cuvette based fluorescence spectroscopy of beads dispersed in solution (Figure 1B). Using UV excitation (360 nm) showed that these cyclic peptide beads emit significant broad band autofluorescence with a maximum at 480 nm. 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 28

When shifting the excitation to 480 nm, the emission maximum was also red shifted, indicating that different populations of intrinsically fluorescent substances give rise to autofluorescence. Thus at all wavelengths subtle intensity changes due to target protein binding would be difficult to detect by conventional fluorescence microscopy. For further characterization, the average fluorescence lifetimes of these library beads were measured using intensity modulation (Figure 1C). In addition, the fluorescence lifetimes of individual beads were measured using a standard wide field fluorescence microscope equipped with a commercially available module for phase-domain based fluorescence lifetime determination (LIFA from- Lambert Instruments) (Figure 1D).

Figure 1: Fluorescence properties of typical TentaGel library beads. A) CONA image of a bead monolayer in a 96-well microtiter plate using 633 nm laser excitation. (B,C) Cuvette 6 ACS Paragon Plus Environment

Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


based measurements of the same beads dispersed in solution. B) Fluorescence excitation spectrum (green, emission at 470 nm) and two emission spectra (red, excitation at 360 nm; black excitation at 480 nm). C) Example for phase domain based fluorescence life-time measurements (see Table 1). D) Corresponding fluorescence life-time imaging (FLIM, blue 2 ns, red 4 ns) using a commercially available phase-domain based accessory mounted on our WIOBS instrument.

Both approaches revealed surprisingly long average fluorescence lifetimes of about 3 ns, rendering the use of fluorescence-lifetime as a detection principle for hit-bead detection impractical (Table 1). We therefore decided to focus on spectral separation of bead autofluorescence and protein-tag derived fluorescence signal for building an optimized fluorescence wide-field microscope for on-bead screening.

Table 1: Fluorescence life-time data determined for the autofluorescence arising from typical one-bead one-compound library beads at different excitation/emission wavelengths, as recorded on a Fluorolog tau-3 spectrofluorometer (Horiba Jobin Yvon).

Instrumental design elements The design of our Wide-field Imaging based On-Bead Screening device (WIOBS) was guided by the use of commercially available components while avoiding complicated robotic systems (Figure 2). A standard inverted microscope body, the Leica DMI6000B, equipped with filter cubes and a motorized scanning stage served as the platform to which the LED light sources and a camera were coupled (Figure 2A). Cooled LEDs reflect a compromise between broad band mercury arc lamps and expensive laser equipment. At the detection side we used a sCMOS technology based camera, which features high-sensitivity while maintaining a large 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 28

field of view. In addition, such cameras offer a reasonable dynamic range to cope with the different intensity levels encountered in OBS, where hit beads with varying autofluorescence levels accumulate different amounts of labeled target protein at their surface (Figure 1A). Screening requires bead picking for decoding and downstream evaluations of the hit compounds. In CONA, the bead-picker was one of the core elements of the PickoScreen instruments, which contained a robotic system for moving picking capillaries in three dimensions. Bead picking required a sophisticated alignment procedure between the stepping motors and the image derived, two-dimensional coordinates of the beads in the screening plate. Moreover the precision needed for picking 90 µm beads was at the expense of lateral speed. We therefore reasoned that a semiautomatic procedure combining manual sliding and magnetic positioning in the x-y plane with motorized movements that control the capillary vertically would be faster. Our picking device is composed of a mechanical slider which allows to manually centering the picking capillary at the optical z-axis within the field of view. We used a micromanipulator that performs a vertical three-step movement to approach the inner well surface. As picking capillary we used a stretch of polyimide tubing which was glued into a short metal ring (Figure 2B). The metal part was then inserted into a silicon lined channel in a transparent acrylic glass disc keeping both ends of the capillary accessible. As oriented at the microscope, the lower end is the bead picking tip, while the upper end was connected via tubing to a hydraulic pump. The metal attached acryl glass holder kept the capillary with sufficient mechanical stability in a vertical position. Once placed above a hit bead, negative pressure pulses then automatically transfer the bead from the screening plate into appropriate vials (Figure 2C).


Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


C)

Figure 2: The WIOBS-1 instrument. A) 3D overview sketch, generated with IronCAD, with bead picker positioned over a 96-well microtiter plate. The individual parts are indicated; the inset shows the bead picker arm (light blue) between the condenser and the scanning stage. B) Close up of the bead-picker composed of the picker capillary mounted into an acrylic glass disc. C) Photograph of the instrument with the bead picker in the slide-in position above the scanning stage.

Wide field fluorescence based quantitative on-bead screening For testing the suitability of WIOBS-1 for on-bead screening we first generated model hit beads by decorating standard 90 µm TentaGel S beads with biotin. Prior to imaging, these biotinylated beads were incubated with three differently labeled streptavidin conjugates: 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Alexa-488-streptavidin,

Tetramethylrhodamine(TMR)-streptavidin

Page 10 of 28

and

Alexa-647-

streptavidin. After thorough washing, the beads were transferred into a 96-well microtiter plate and fluorescence images were recorded using constant acquisition parameters (Figure 3A). Streptavidin binding to the beads was clearly detectable for all three colors, as seen from the pronounced fluorescent rings around the beads. An analysis of the average pixel intensities for each image shows that the fluorescence signal responds linearly to the increasing incubation concentrations of all tested streptavidin conjugates (Figure 3B). Thus, our WIOBS instrument proved to be able to specifically and quantitatively detect the binding of a fluorescently labeled target protein to TentaGel beads under typical on-bead screening conditions.


Page 11 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3: WIOBS on-bead screening images and signal dynamics. A) images of biotinylated beads incubated with increasing concentrations of differently labeled streptavidin-conjugates. Left: Alexa-488, Middle: TMR, Right: Alexa-647. Top-to-bottom: 50 nM, 150 nM, 250 nM, 500 nM labeled streptavidin, after 6 h incubation. B) Linear data fit of average pixel intensities obtained from entire images (R > 0.9).

Image merging and screening speed Next, we addressed the imaging speed and performance at different optical magnifications. Because the beads distribute within wells of a 96-well microtiter plate over a much larger area than the field of view, an efficient tile-image merging procedure had to be established. 9, 32, or 84 tile-images are required to cover the entire well when imaging with a 5x, 10x or 20x objective, respectively (Figure 4). Due to inhomogeneity in excitation and detection efficiency within a single image frame, each of the individual tile images was first corrected using a previously recorded reference image. The time needed to record and process individual tile-images for wells of a 96-well microtiter plate was tested using different parameters for operating the instrument (Table 2). The image acquisition times including autofocus determination ranged from 26 seconds per well at 5x magnification to 2.5 minutes per well at 20x magnification. Thus, recording an entire 96-well microtiter plate, containing 200 000 beads, takes about 41 minutes at 5x magnification. Even at 5x magnification, the sensitivity proved to be perfectly suitable for detecting Alexa-64711 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 28

streptavidin binding to the biotinylated beads after incubation with 50 nM labeled protein for 6 hours (Figure 4). While the sensitivity at 20x magnification is clearly improved as compared to lower magnifications, the recording time for the 84 tile-images adds up to 4 hours (96 x 150 sec) plus an additional 8 hours (96 x 290 sec) of computing time for tile-image merging on a standard PC. To gain additional speed, the time required for the image processing can be allocated to a second computer and processed in parallel. Even without this feature, the 10x objective allowed us to acquire data from an entire 96-well plate in less than 2 hours, representing a good compromise between speed and sensitivity. To demonstrate the hit detection efficiency and performance of the system under suboptimal imaging conditions, we decided to use the 5 x magnification in subsequent experiments. The procedure would also make sense for real screening situations, since the operator would always have the choice to revisit individual wells and image the area of interest at a higher resolution.

Table 2: Typical parameters for imaging of entire wells of a 96-well microtiter plate on the WIOBS-1 instrument at different magnifications.


Page 13 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4: Images recorded on the WIOBS 1 instrument at different magnifications and after tile-image merging. Left: entire wells of a 96-well microtiter plate containing a monolayer of 90 µm TentaGel beads; Middle: zoomed section, Right: one individual bead, exemplifying the resolution differences. Theoretical resolutions (x,y) of the three magnifications were: 5x: 2.6 µm, for 10x: 1.3 µm, and for 20x: 0.65 µm. 13 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 28

Bead picking Our bead-picking device enables a fast isolation and sorting of beads, both in brightfield illumination mode (Figure 5A) or under fluorescence imaging settings (Figure 5B). Bead picking is achieved in sequence of six steps: Step 1 - Alignment: as individually cut picking capillaries vary in length, two appropriate zaxis positions have to be determined before the picking session: (a) the bead selection

position which is 10 to 50 µm above the bead monolayer and allows the picking capillary to be seen in the field of view and (b) the picking position, which is a few micrometers above the well bottom. Step 2 - Bead Selection: The stage-coordinates of the image frame are stored. A mouse click in the image then allows the operator to select the respective beads and causes the scanning stage to move to these x-y positions, bringing the selected bead into the center of the field of view. Step 3 – Capillary movement. The capillary tip is carefully lowered above the bead monolayer to Z-position (a) bead selection position and becomes visible in the field of view. It allows fine adjustment of the capillary above the selected bead by moving the x,y-stage with minimal drag forces to maintain the relative bead positions in the well. Step 4 – Picking: The capillary tip is further lowered to the Z-position (b) picking position where it engulfs the selected bead. For 90 µm beads a predefined volume of about 4 µl liquid is sucked by a negative pressure pulse that draws the selected bead into the capillary. The liquid volumes for bead picking have to be adjusted to the bead diameter. Step 5 – Capillary withdrawal: The capillary is moved out of the well to the resting position, from where the slider has to be manually moved aside for dispensing the bead. 14 ACS Paragon Plus Environment

Page 15 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Step 6 – Bead dispensing: A volume of 10 µl is dispensed through the capillary by a positive pressure pulse into a glass vial, a well of a microtiter plate, or any other vessel. The entire picking procedure for a single bead takes approximately 20-30 seconds.

Figure 5: Bead picking on the WIOBS instrument. A) Bright field images documenting the picking process at the plane of the bead monolayer in a 96-well plate. The tip of the picking capillary appears dark and larger than the beads. B) The same procedure under fluorescence imaging settings.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 28

Image Subtraction Approach for on-bead screening of libraries So far, the data showed that wide-field fluorescence imaging can reliably detect streptavidin binding at the surface of biotinylated TentaGel beads. However, these beads have only undergone a single coupling step and therefore exhibit a quite uniform and low autofluorescence. To model the contrast issue in a more realistic screening scenario, we used the small Fmoc-based solid phase peptide library containing cyclic hexapeptides as described above. We reasoned that a second fluorescence image, taken at excitation and emission wavelengths spectrally distant from the absorption and emission maxima of the target protein label can be used to subtract bead autofluorescence (Figure 6A). To test this approach with the WIOBS instrument, beads from the cyclic peptide library were mixed with biotinylated beads (3%), swollen in buffer and homogenized by shaking, followed by incubation with 50 nM fluorescently tagged streptavidin conjugate for six hours. To discriminate signal from autofluorescence, entire wells of a 96-well microtiter plate were imaged with two different excitation and emission wavelength settings (Figure 6B). For the Alexa-647 labeled streptavidin, the 488/520 nm channel was used for autofluorescence correction, whereas for the Alexa-488 labeled streptavidin the 455/485 nm channel was used to correct for autofluorescence. Exposure time and laser power for the correction channel were adjusted to the level of average signal intensities in the corresponding target protein channel. This adjustment can easily be done with a population of library beads in the absence of target prior to screening. The two set of images were then subtracted to generate an autofluorescence corrected image. We tested under various conditions that this procedure reproduced images in which only the hit beads with ring-shaped signals arising from the surface bound target protein remained visible (Figure 6C). Even for the lowest 5x magnification the approach proved sensitive enough to discriminate Alexa-488-streptavidin decorated beads from the 16 ACS Paragon Plus Environment

Page 17 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


bulk; a worst case scenario, as the peak maxima of the bead autofluorescence and Alexa-488 signal match at around 490 nm.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 28

Figure 6: Image subtraction approach for on-bead screening on WIOBS. All images were recorded using 5 x magnification. Wells were filled with beads from a one-bead onecompound library of cyclic hexapeptides, containing about 3% biotinylated beads and incubated with different streptavidin conjugates (50 nM, 6 h). A) Schematic representation of the image subtraction approach. B) Example for tile-images of a well incubated with Alexa647-streptavidin. C) Entire well of a 96-well microtiter plate. Top row: Alexa-647streptavidin conjugate. Bottom row: Alexa-488-streptavidin conjugate.

Cell based screening One conceptual advantage of using a standard wide-field microscope for on-bead screening is the screening of cell-surface binding ligands with cells. To test the feasibility of our WIOBS instrument for cell-based on-bead screening, we used HeLa and Huh-7 cells and compared the behavior of two different peptidic ligands, a cyclic disulfide bridged peptide, CRKRLDRNC27, as well as the transferrin binding peptide HAIYPRH28.. Both peptides were synthesized on TentaGel beads using Fmoc-chemistry. The disulfide bridged peptide was cyclized on-bead by iodine mediated disulfide formation. Beads carrying either one of the two peptides or biotin as a ligand were incubated with HeLa cells or Huh-7 cells at 37° C. The beads were then transferred into a microtiter plate and imaged on the WIOBS microscope using bright field illumination (Figure 7). As soon as five minutes after incubation HeLa cells attached specifically to beads with the cyclic CRKRLDRNC peptide whereas, no attachment was observed for the other two ligands. Also, Huh-7 cells did not bind to any of the ligands. This demonstrates, that the WIOBS instrument is also suitable for cell-based on-bead screening e.g. for identification of cell-type specific, extracellular ligands.


Page 19 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7: Cell-based on-bead screening using WIOBS. All images were recorded using bright-field illumination with 5x magnification. Left: beads carrying the cyclic peptide CRKRLDRNC27, Middle: beads carrying a linear peptide, HAIYPRH28, Right, biotinylated beads. Top row: HELA cells, bottom row: Huh-7 cells.

DISCUSSION

In our attempt to set up a simplified OBS procedure, we have developed a wide field fluorescence microscope extended by an accessory bead picking device (WIOBS). With WIOBS motorized movements are reduced to only one spatial dimension (z), thereby simplifying the outlay for hardware, as well as software that controls the instrument. After imaging the incubated bead monolayers, hit beads are selected by the operator in images. The corresponding x-y stage coordinates then guide a picker arm for transfer of the selected bead. For a 96-well microtiter plate containing approximately 200 000 TentaGel S 90 µm beads, the screening times for the entire plate recorded with the WIOBS instrument was just under 90 minutes, when using a 5x objective and still less than 3.5 hours when a 10x objective was used, which is 5 - 10 times faster than CONA or COPAS screening. Furthermore, the procedure is flexible in that it allows changing the magnification for certain wells of interest.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 28

Although wide field imaging provides a huge gain in speed as compared to confocal scanning, one has to deal with signal-to-noise limitations. Low contrast of fluorescent target protein binding is especially aggravated due to intrinsic autofluorescence of the library beads. Three main factors contribute to this: (1) TentaGel, the most common solid support used in on-bead screening has a substantial autofluorescence, arising from the polystyrene co-polymer. (2) During chemical synthesis, in-complete couplings and side reactions of reagents lead to byproducts with can produce pronounced fluorescence emission across the entire spectrum. (3) Some library compounds themselves may be fluorescence if they contain larger conjugated pi-electron systems. Only the last factor can be taken into consideration during library design e.g. by avoiding certain building blocks.

Therefore it was crucial to increase the signal over noise for hit bead detection by a separate image processing step. We found that autofluorescence could be measured in a second, spectrally adjacent color channel that, after rescaling, is subtracted from raw images. Although this image subtraction approach doubles the imaging time per se, the speed gain associated with the use of fluorescence wide field imaging more than compensates for this. Ideally, the correction channel should locate at shorter wavelengths close to the specific fluorescence emission channel of the labeled target protein but sufficiently separated to avoid signal contributions from the target protein. To our surprise this rather simple image subtraction approach led to quite remarkable detection efficiencies. It was even successful using short wavelength dyes for labeling target protein (e.g. Alexa-488) that emit in a spectral range where autofluorescence of TentaGel beads have a pronounced peak. Thus, our approach provides superior flexibility for the choice of fluorescent labels in the presence of autofluorescence as compared to many conventional systems.


Page 21 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In summary, our presented OBS screening procedure is simple to operate, robust, and comes at reasonable costs. The development may have potential to make OBS accessible to a larger number of researchers who want to identify new bioactive molecules from combinatorial chemical libraries.

MATERIALS AND METHODS

All chemicals were bought from Sigma-Aldrich. TentaGel beads (standard 90 µm diameter) with either free amine, Rink linker or HMBA linker were bought from Rapp Polymers, Germany. Labeled Streptavidin conjugates were bought from Life Technologies Corporation, USA.

Chemical syntheses:

All syntheses were carried out using manual solid phase peptide synthesis equipment and standard Fmoc chemistry, following standard procedures as recommended by Novabiochem, Germany and described previously.14,29 In brief, resin loading for HMBA resin was carried out using the MSNT/1-methylimidazole method. For coupling the individual building blocks, an excess of the respective building block (5.8 equ.) was dissolved in DMF, HATU (6 equ.) was used as coupling reagent and DIPEA (12 equ.) as the base. All couplings were carried out twice with each coupling time 40 min. Fmoc deprotection after extensive washing was carried by four repetitive treatments of the resin with piperidine (20% in DMF) for 5 min. For the cyclic peptide library a head-to-tail cyclization on resin was carried out following Pd(0) catalyzed removal of an ally protection group on Glutamic acid (Fmoc-Glu-OAll) and onresin cyclization with HATU. Final side-chain deprotection of peptides was carried out using TFA:TIS:water (95:2.5:2.5) for 2 hours. Biotinylated beads were generated by coupling Biotin to TentaGel S NH2 resin using the HATU coupling procedure. The CRKRLDRNC 21 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 28

peptide was synthesized on TentaGel S HMBA resin, using C-terminal 8-amino-3,6dioxaoctanoic acid as C-terminal spacer. The peptide was cyclized after on-resin side-chain deprotection by DMSO treatment for 24 hours.

Fluorescence excitation and emission spectra of TentaGel beads

Excitation and emission spectra of TentaGel beads containing a one-bead one compound library of cyclic peptides were recorded on a Fluorolog tau-3 spectrofluorometer (Horiba Jobin Yvon). All measurements were performed with a Hellma 1.5 ml cuvette under stirring. The beads were suspended in a methanol: water mixture (1:3). Instrumental parameters were: polarizers in magic angle settings (55°), integration time 2 seconds, excitation and emission slit widths 3 nm and 4 nm, respectively. All spectra shown were corrected for the blank, dark and reference contributions as well as for spectral instrumental characteristics.

Fluorescence lifetime measurements of TentaGel beads

Lifetime measurements were performed on the Fluorolog tau-3 in lifetime mode. Ludox (silica particles in water) was used as the reference. The procedure for obtaining lifetime data was according to the manufacturer’s guidelines. Excitation and emission slits settings were optimized to obtain maximal signal from the bead sample. The measurement was performed with parameters for integration time set to 10 seconds in the frequency range 20 – 200 MHz. The appropriate excitation and emission wavelengths were determined from the excitation and emission spectra of the library beads. The data was fitted using Horiba’s own “lifetime modelling” software.


Page 23 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Instrumental description of WIOBS instrument

The WIOBS instrument is based on an inverted fluorescence wide field microscope (DMI6000B, Leica Microsystems, Wetzlar/Germany) equipped with a sCMOS NEO (Andor Technology, Belfast/UK) camera, a high precision motorized stage (C7018-9012K VEXTA Stepping Motor - Oriental Motor CO, Ltd. Tokyo, Japan), a cooled pE-2 LED fluorescence illuminator (4 LEDs 400/490/550/635 nm) connected by 2 liquid light-guides and a TransLED white light source for bright field imaging. The light path was equipped with objectives HC PL APO 10x/0.40, HC Plan APO 20x/0.70, and HCX PL APO 40x/1.25-0.75 oil immersion, as well as 4 fluorescence filter sets (LF 405- Ex 390/40 Em 452/45 Di405, GFP 3035- Ex 472/30 Em 520/35 Di 495, TRITC- Ex 542/20 Em 620/52 Di 570, Cy5 4040Ex 628/40 Em 692/40 Di660) for fluorescence excitation and emission separation.

The Bead picker system was designed by Life imaging services (Basel, Switzerland) and consisted of a stiff and light carbon arm, mounted on a high speed (350 mm/sec) high precision motorized z-axis, M-683.2U4, plus controller PI C-867 (both from Physik Instrumente GmbH & Co. KG, Karlsruhe/Germany). The high precision of the motorized z-axis featured a linear encoder with sensor resolution of 0.1 µm and a bidirectional repeatability ±1 µm over the entire travel range of 50 mm. The end of the arm frames a translucent polycarbonate disc which slides up to medial kinematic end position defined by a magnetic stop block on a manual x-axis (medial-lateral mobility 32 cm) underneath the Leica S70 WD condenser in the light path. The disc allows for bright field illumination and connects an exchangeable capillary of polyimide tubing (Accellent Inc. MA, USA) by a silicon adaptor to the high- precision syringe pump VersaPump3 (Kloehn Inc. USA Las Vegas, NV). The stable end position of the bead picker capillary can be manually fine-tuned in the field of view by two screws in horizontal x-y directions. 23 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 28

All components of the WIOBS where controlled by MethaMorph Microscopy Automation and Image Analysis Software (Molecular Devices LLC. CA/USA Version 7.7.5.0 64 bit Image acquisition, processing and analysis package). For the integration of the Kloehn pump and the additional bead picker PI z-axis in the MethaMorph software, the common shared communication port was used. Automated data recording for a full 96-well microtiter plate was achieved by a combination of standard MethaMorph macros and recording of new journals. Image analyses, quantification, and image processing like stitching and image subtraction routines were accomplished with the MetaMorph Image Analysis software package.

Preparing beads for screening

Bead aliquots in 2 ml Eppendorf tubes were allowed to swell for several minutes in 1 ml of PBS, 0.005% Tween 20. After sonication (3 times 10 sec). the beads were spun down using a table-top centrifuge and the supernatant was replaced by buffer containing varying concentrations of dye-labeled streptavidin. Target protein binding was allowed to proceed under constant agitation at RT (6 hours). Prior to fluorescence imaging and bead picking, the incubation buffer was removed and the beads were washed at least 3 times with PBS, 0.005% Tween 20. Beads where then transferred to the different wells of a 96-well microtiter plate. To generate bead monolayers for screening, plates containing 2000 TentaGel beads per well (~1 mg of dry resin) in incubation solution were covered with sealing film and vigorously agitated for several seconds such that all beads were floating in suspension before allowing them to settle.


Page 25 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Cell-based on-bead screening assay

Adherent cells (HeLa & Huh7) were detached by incubation with Cell Dissociation Buffer (enzyme free, PBS-based - Invitrogen) for 8 min, centrifuged for 2 min, 300 x g, followed by a careful re-suspension in FCS free medium at 106 cells / ml). 1 mg beads were washed with PBS and re-suspended in 1 ml FCS free cell medium. The beads were mixed with 5x 105 cells in a total volume of 750 µl in a 1.5 ml Eppendorf tube. After incubation for 10 min at 37°C on a rotator wheel, freely floating cells in the supernatant were removed. After addition of 200 µl of medium the beads were transferred into the wells of a 96-well microtiter plate. Images from each well were recorded immediately, and again after incubation of 24 hours at 37°C.

Fluorescence imaging, bright-field imaging and image merging on WIOBS instrument

Instrumental settings for fluorescence imaging using streptavidin conjugates were: Alexa-488: excitation 490 nm, exposure time 200 ms; TAMRA: excitation 550 nm, exposure time 250 ms; Alexa-647: 635 nm, exposure time 350 ms. Camera settings: frames to average 3; Center Quad. For bright-field imaging with TransLed illumination (laser power 30-80%) exposure times were set between 2 to 20 ms and automated illustration (color depth, thresholds) was turned on. Images showing entire wells of a 96-well plate in 5x, 10x and 20x magnification were obtained by a three step protocol operating on the tile images: recording, stacking, and merging.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 28

REFERENCES (1) (2) (3) (4)

(5)

(6) (7) (8) (9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17) (18)

Kneller, R. The Importance of New Companies for Drug Discovery: Origins of a Decade of New Drugs. Nat. Rev. Drug Discov. 2010, 9, 867–882. Frearson, J.; Wyatt, P. Drug Discovery in Academia- the Third Way? Expert Opin. Drug Discov. 2010, 5, 909–919. Ho, Y. et al. Systematic Identification of Protein Complexes in Saccharomyces Cerevisiae by Mass Spectrometry. Nature 2002, 415, 180–183. Jones, R. B.; Gordus, A.; Krall, J. A.; MacBeath, G. A Quantitative Protein Interaction Network for the ErbB Receptors Using Protein Microarrays. Nature 2006, 439, 168– 174. Meisner, N. C.; Hintersteiner, M.; Uhl, V.; Weidemann, T.; Schmied, M.; Gstach, H.; Auer, M. The Chemical Hunt for the Identification of Drugable Targets 1. Curr.Opin.Chem.Biol. 2004, 8, 424–431. Kodadek, T. Rethinking Screening. Nat. Chem. Biol. 2010, 6, 162–165. az-Mochon, J. J.; Bialy, L.; Bradley, M. Dual Colour, Microarray-Based, Analysis of 10 000 Protease Substrates. Chem. Commun. 2006, 3984–3986. Kodadek, T. The Rise, Fall and Reinvention of Combinatorial Chemistry. Chem. Commun. 2011, 47, 9757. Lam, K. S.; Salmon, S. E.; Hersh, E. M.; Hruby, V. J.; Kazmierski, W. M.; Knapp, R. J. A New Type of Synthetic Peptide Library for Identifying Ligand-Binding Activity. Nature 1991, 354, 82–84. Sweeney, M. C.; Wang, X.; Park, J.; Liu, Y.; Pei, D. Determination of the Sequence Specificity of XIAP BIR Domains by Screening a Combinatorial Peptide Library. Biochemistry 2006, 45, 14740–14748. Gocke, A. R.; Udugamasooriya, D. G.; Archer, C. T.; Lee, J.-Y.; Kodadek, T. Isolation of Antagonists of Antigen-Specific Autoimmune T Cell Proliferation. Chem Biol. 2009, 16, 1133–1139. Doran, T. M.; Gao, Y.; Mendes, K.; Dean, S.; Simanski, S.; Kodadek, T. Utility of Redundant Combinatorial Libraries in Distinguishing High and Low Quality Screening Hits. ACS Comb. Sci. 2014, 16, 259–270. Auer, M.; Moore, K. J.; Meyer-Almes, F. J.; Guenther, R.; Pope, A. J.; Stoeckli, K. A. Fluorescence Correlation Spectroscopy: Lead Discovery by Miniaturized HTS. Drug Discov. Today 1998, 3, 457–465. Hintersteiner, M.; Kimmerlin, T.; Kalthoff, F.; Stoeckli, M.; Garavel, G.; Seifert, J.-M.; Meisner, N.-C.; Uhl, V.; Buehler, C.; Weidemann, T.; Auer, M. Single Bead Labeling Method for Combining Confocal Fluorescence On-Bead Screening and Solution Validation of Tagged One-Bead One-Compound Libraries. Chem Biol 2009, 16, 724– 735. Hintersteiner, M.; Knox, A. J.; Mudd, G.; Auer, M. Towards Mimicking Short Linear Peptide Motifs: Identification of New Mixed Α,β-Peptidomimetic Ligands for SLAMAssociated Protein (SAP) by Confocal on-Bead Screening. J. Chem. Biol. 2012, 5, 63– 79. Townsend, J.; Do, A.; Lehman, A.; Dixon, S.; Sanii, B.; Lam, K. S. 3-Nitro-Tyrosine as an Internal Quencher of Autofluorescence Enhances the Compatibility of Fluorescence Based Screening of OBOC Combinatorial Libraries. Comb. Chem. High Throughput Screen. 2010, 13, 422–429. Olivos, H. J.; Bachhawat-Sikder, K.; Kodadek, T. Quantum Dots as a Visual Aid for Screening Bead-Bound Combinatorial Libraries. Chembiochem 2003, 4, 1242–1245. Astle, J. M.; Simpson, L. S.; Huang, Y.; Reddy, M. M.; Wilson, R.; Connell, S.; Wilson, J.; Kodadek, T. Seamless Bead to Microarray Screening: Rapid Identification 26 ACS Paragon Plus Environment

Page 27 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(19)

(20)

(21)

(22) (23)

(24)

(25)

(26)

(27)

(28)

(29)

of the Highest Affinity Protein Ligands from Large Combinatorial Libraries. Chem. Biol. 2010, 17, 38–45. Liu, T.; Qian, Z.; Xiao, Q.; Pei, D. High-Throughput Screening of One-Bead-OneCompound Libraries: Identification of Cyclic Peptidyl Inhibitors against Calcineurin/NFAT Interaction. ACS Comb. Sci. 2011, 13, 537–546. Mendes, K.; Ndungu, J. M.; Clark, L. F.; Kodadek, T. Optimization of the Magnetic Recovery of Hits from One-Bead–One-Compound Library Screens. ACS Comb. Sci. 2015, 17, 506–517. Mueller, K.; Gombert, F. O.; Manning, U.; Grossmueller, F.; Graff, P.; Zaegel, helene; Zuber, jean F.; Freuler, F.; Tschopp, C.; Baumann, G. Rapid Identification of Phosphopeptide Ligands for SH2 Domains. Screening of Peptide Libraries by Fluorescence-Activated Bead Sorting. J. Biol. Chem. 1996, 271, 16500–16505. Pulak, R. Techniques for Analysis, Sorting, and Dispensing of C. Elegans on the COPAS Flow-Sorting System. Methods Mol Biol 2006, 351, 275–286. Reddy, M. M.; Bachhawat-Sikder, K.; Kodadek, T. Transformation of Low-Affinity Lead Compounds into High-Affinity Protein Capture Agents. Chem. Biol. 2004, 11, 1127–1137. Hintersteiner, M.; Auer, M. A Two-Channel Detection Method for Autofluorescence Correction and Efficient on-Bead Screening of One-Bead One-Compound Combinatorial Libraries Using the COPAS Fluorescence Activated Bead Sorting System. Methods Appl. Fluoresc. 2013, 1, 017001. Hintersteiner, M.; Buehler, C.; Uhl, V.; Schmied, M.; Muller, J.; Kottig, K.; Auer, M. Confocal Nanoscanning, Bead Picking (CONA): PickoScreen Microscopes for Automated and Quantitative Screening of One-Bead One-Compound Libraries. J Comb Chem 2009, 11, 886–894. Hintersteiner, M.; Buehler, C.; Auer, M. On-Bead Screens Sample Narrower Affinity Ranges of Protein-Ligand Interactions Compared to Equivalent Solution Assays. Chemphyschem 201213, 3472-3480. Hong, H.; Lee, H. Y.; Kwak, W.; Yoo, J.; Na, M.-H.; So, I. S.; Kwon, T.-H.; Park, H.S.; Huh, S.; Oh, G. T.; Kwon, I.-C.; Kim, I.-S.; Lee, B.-H. Phage Display Selection of Peptides That Home to Atherosclerotic Plaques: IL-4 Receptor as a Candidate Target in Atherosclerosis. J. Cell. Mol. Med. 2008, 12, 2003–2014. Lee, J. H.; Engler, J. A.; Collawn, J. F.; Moore, B. A. Receptor Mediated Uptake of Peptides That Bind the Human Transferrin Receptor. Eur. J. Biochem. 2001, 268, 2004–2012. Hintersteiner, M.; Kimmerlin, T.; Garavel, G.; Schindler, T.; Bauer, R.; Meisner, N.C.; Seifert, J.-M.; Uhl, V.; Auer, M. A Highly Potent and Cellularly Active B-Peptidic Inhibitor of the p53/hDM2 Interaction. ChemBioChem 2009, 10, 994–998.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 28

A wide field fluorescence microscope extension for ultrafast screening of one-bead one-compound libraries using a spectral image subtraction approach. Wolf Heusermann, Beat Ludin, Nhan T. Pham, Manfred Auer, Thomas Weidemann, and Martin Hintersteiner


A Wide-Field Fluorescence Microscope Extension for Ultrafast

Recommend Documents