MONOCLONAL ANTIBODIES IN IMMUNODIAGNOSIS AND IMMUNOTHERAPY Volume 33, Number 2, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/mab.2013.0088

A Mouse Monoclonal Antibody Against Alexa Fluor 647 Irene Wuethrich, Eduardo Guillen, and Hidde L. Ploegh

Fluorophores are essential tools in molecular and cell biology. However, their application is mostly confined to the singular exploitation of their fluorescent properties. To enhance the versatility and expand the use of the fluorophore Alexa Fluor 647 (AF647), we generated a mouse monoclonal antibody against it. We demonstrate its use of AF647 for immunoblot, immunoprecipitation, and cytofluorimetry.

Introduction

T

he possible uses of an antibody specific for a fluorophore are several fold. First, such antibodies make single-purpose molecules (fluorescence) multi-functional, for example for use in immunoprecipitation, affinity purification, and quantification by enzyme-linked immunosorbent assay (ELISA). A singularly modified molecular species can now be interrogated with different methods, making assays easier to compare through elimination of experimental variables. In combination with enzymatic stoichiometric labeling methods such as sortagging,(1) in which fluorophores are installed at precisely specified molecular locations and can be combined with cross-linkers, STORM probes, glycosylation, and sulfation markers or other functional handles, anti-fluorophore antibodies can turn dyes into multi-faceted and versatile tools. Second, a fluorophore can be used directly as a platform for signal amplification, allowing for increased assay sensitivity and detection of low-abundance molecules. A widely used fluorescent dye is Alexa Fluor 647 (AF647), originally developed by Molecular Probes and currently marketed by Life Technologies. Its maximum absorbance is at 650 nm, the maximum emission at 665 nm. A general advantage of dyes that emit towards the red end of the visible spectrum ( > 600 nm) is less background fluorescence compared to fluorophores that emit at shorter wavelengths, which suffer from interference by autofluorescence in biological samples. AF647 is considered more photostable(2) and, as reported by the manufacturer, less self-quenching than Cy5, whose absorption and emission spectra it closely matches. Total fluorescence of the secondary antibody conjugates prepared with the Alexa Fluor 647 dye is significantly higher than that of Cy5, and unlike the Cy5 dye, the Alexa Fluor 647 dye shows very little change in absorbance or fluorescence emission spectra when conjugated to proteins, oligonucleotides, and nucleic acids.(3,4) Sulfonation confers onto AF647 a negative charge, which makes it hydrophilic and thus wellsuited for application in an aqueous milieu.

We describe here the generation of a monoclonal mouse anti-AF647 antibody by microengraving screening. The ability of this antibody to react in immunoblots suggests it could be an excellent candidate to perform correlated labeling studies, for example, through a combination of immunoelectron microscopy and standard fluorescence microscopy. Materials and Methods Conjugation of Alexa Fluor 647 to carrier proteins

Alexa Fluor 647 (AF647) succiminidyl ester (Invitrogen, Carlsbad, CA) was covalently conjugated to bovine serum albumin (BSA) or KLH (keyhole limpet hemocyanin) (Pierce, Rockford, IL) following the manufacturer’s instructions. Mouse immunization

Experiments were performed in accordance with institutional, state, and federal guidelines. Female 3-monthold BALB/c mice were immunized intraperitoneally with 75 mg conjugate emulsified in complete Freund’s adjuvant (day 0), followed by an intraperitoneal boost with 50 mg conjugated protein emulsified in Freund’s adjuvant (day 12), and a second intravenous boost with 50 mg conjugated protein in phosphate-buffered saline (PBS, on day 19). Blood serum sample was taken (tail vein) for a direct ELISA assay (day 19). Lymphocyte harvest and hybridoma cell generation

Splenic lymphocytes were harvested (day 22) and fused to plasmacytoma cells (P3/NSI/1-Ag4-1 [NS-1], ATCC TIB18; ATCC, Manassas, VA), as described previously.(19) HAT selection

The resulting hybridoma cells were grown until day 32 in HAT selection medium (Dulbecco’s Modified Eagle’s Medium [DMEM] with 20% v/v inactivated fetal calf serum

Whitehead Institute for Biomedical Research, Cambridge, Massachusetts.

109

110

[IFS], 20 mM HEPES, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids [Gibco, Grand Island, NY], 50 mM 2mercaptoethanol, 10% v/v HCF [hybridoma cloning factor; BioVeris, Gaithersburg, MD], hypoxanthine-aminopterinthymidine [HAT; Sigma Aldrich, St. Louis, MO] according to the manufacturer’s instructions) at 37C (5% CO2). Cell culture

12CA5 cells (a subclone of the H26DO8 hybridoma clone(6)) were maintained in DMEM, 10% fetal bovine serum (FBS), 2 mM L-glutamine, 1 mM sodium pyruvate, and nonessential amino acids (Gibco). Cells were cultured at 37C, 5% CO2. Direct ELISA

Wells of 96-well microtiter plates were coated overnight at room temperature with AF647 conjugated to carrier protein (BSA or KLH) or polyclonal rabbit anti-Ig and anti-IgG in 0.1 M phosphate buffer (pH 9.2; 1 mg/mL; 50 mL/well). Wells were washed 2x with PBST (PBS, 0.05% v/v Tween-20). Serum was diluted 1:200 in binding buffer (PBS, 0.5% calf serum, 0.05% Tween-20), subsequently diluted in 1:2 dilution series, and incubated 1 h at 4C. Unbound antibody was removed by washing three times with PBST. Bound antibody was detected after addition of horseradish-peroxidase-coupled (HRP) goat anti-mouse antibody (BD Biosciences, Franklin Lakes, NJ) diluted 1:3000 in binding buffer and HRP substrate 3,3’5,5’-tetramethylbenzidine (Sigma Aldrich). The reaction was stopped by the addition of 0.5 M H2SO4. Assays were visually inspected, and absorbance was measured at 450 nm. Isotyping by sandwich ELISA

Hybridoma-secreted antibodies were captured with rabbit anti-Ig. HRP-coupled isotype-specific antibodies (BD Biosciences) were used as secondary antibodies.

WUETHRICH ET AL.

was deposited on the surface of a slide under a coverslip (LifterSlip; Erie Scientific Company, Portsmouth, NH) and incubated overnight at 4C. After incubation, the slides were immersed in blocking buffer (PBS, 0.01% w/v NaN3 1% w/v BSA) for 30 min at room temperature and stored overnight at 4C. The slides were rinsed in PBST, PBS, and then deionized water. They were spun dry for 5 min at 100 g immediately before being sealed against the micro-well arrays. Loading arrays of micro-wells

A suspension of cells was diluted to 1 · 105 cells/mL in serum-containing media and 0.5–1 mL was pipetted onto the surface of the micro-well array. The cells were allowed to settle for 3 min. Excess cell suspension was eliminated by vacuum aspiration at one edge of the array while tilting the array. The percentage of wells filled and the average number of cells per well were determined by counting the number of cells in randomly determined viewing fields with a 10 · lens and averaging data collected from multiple micro-well arrays. Engraving microarrays using loaded micro-wells

An array of micro-wells filled with cells was placed wellside down on the surface of a treated, dry glass slide. The combination of the array and glass slide was sandwiched together in a hybridization chamber (DT-1001; Die-Tech, Bensenville, IL); the screws used to clamp the chamber together were tightened just until finger-tight. The entire assembly was incubated at 37C for 20 min. After incubation, the treated glass slide was removed from the surface of the micro-well array and immediately immersed in blocking buffer (1% BSA, 0.05% Tween-20, PBS), then incubated on a rocker for 1 h at 25C. After placing the glass slide in blocking buffer, the micro-well array was quickly immersed in a bath of pre-warmed medium before medium contained in the micro-wells completely evaporated and further incubated in selection medium for 2 days at 37C (0.5% CO2).

Microengraving

Anti-AF647 antibody secreting clones were detected and isolated by microengraving as described previously.(5) Micro-well array fabrication

Micro-well arrays were fabricated in PDMS (Sylgard 184; Dow Corning, Midland, MI) using photolithography and replica molding. PDMS was cast onto a master, cured for 2 h at 60C and peeled away. The PDMS was allowed to swell in hexanes for 24 h, de-swell in acetone for 24 h, and then dry in an oven at 130C for 24 h. The resulting stamps contained arrays of 50 mm diameter wells. The micro-well arrays were treated with oxygen plasma (PDC-32G; Harrick Scientific, Pleasantville, NY) for 20 s; this process sterilizes the array and increases hydrophilicity. The plasma-treated device was immersed in a solution of 1% w/v BSA (0.01% sodium azide) for 1 h at 25C and then rinsed with sterile PBS (Gibco). Glass slide preparation

Glass slides (cat. ID SME2; Arrayit, Sunnyvale, CA) were used. A solution of capture antibody (20 mg/mL goat antimouse Ig and anti- mouse IgG [BD Biosciences]) in borate buffer (50 mM borate [pH 9.0], 80 mM sucrose, 50 mM NaCl)

Interrogation of microarray

After blocking, glass slides were rinsed with PBST, PBS, and then deionized water for 5 min each; the slides were spun dry for 5 min at 100 g. A solution of goat anti-mouse secondary antibody (Alexa Fluor 532; Invitrogen) and fluorescent antigen (10 mg/mL KLH-Alexa Fluor 647 conjugate) in array deposition buffer (PBS, 0.1% BSA, 0.05% Tween-20, 0.05% NaN3) was deposited on the surface of a slide under a coverslip (LifterSlip; Erie Scientific) and incubated in the dark for 1 h at 25C. After antigen exposure, glass slides were blocked in blocking buffer for 30 min at room temperature on rotating shaker. The slides were rinsed with PBST and PBS two times and spun dry for 5 min at 100 g. The slides were imaged with a GenePix 4000B microarray scanner (Molecular Devices, Sunnyvale, CA) using 532 and 635 nm lasers and factory-installed emission filters. The lasers were used at 100% power and the PMT gain was set between 600 and 900 to maximize the dynamic range of the detector without saturation. Images of the microarrays were analyzed using GenePix Pro 6.1 (Molecular Devices). Double positive wells were scored. Color ratio images were generated in GenePix and saved in the red and green channels of a 24-bit TIFF file. Background intensities were subtracted using median values measured in

ANTI-ALEXA FLUOR 647 MONOCLONAL

111

regions between individual spots of the array. The signal-tonoise ratio for a given positive spot in the array to negative (or background) spots within the same subarray was calculated by dividing the sum of the median intensity values from each channel for a given spot by the average sum of median intensity values for the negative spots determined from at least 20 negative spots. The mean values reported are the average of at least 128 positive spots from more than 12 subarrays.

Sortase-labeling

Microscopy and micromanipulation

Proteins resolved by SDS-PAGE were semi-dry blotted on polyvinylidene difluoride (PVDF) membranes (Bio-Rad, Hercules, CA) for 1.5 h (1.2 mA/cm2). After blocking PVDF membranes with PBS, 0.1% Tween-20, and 5% bovine serum albumin (BSA; Sigma Aldrich), membranes were incubated for 1 h at 37C or overnight at 4C with serum or purified a-AF647 antibody in PBS, 0.1% Tween-20, and 2.5% BSA followed by 1 h incubation at 37C, with HRP-conjugated goat anti-mouse Ig secondary antibodies (BD Biosciences) in PBS, 0.1% Tween-20, and 2.5% BSA. Bound antibody was visualized by chemoluminescence using Western Lightning ECL detection kit (PerkinElmer) and exposure to XAR-5 films (Kodak).

Double-positive cells were picked from their PDMS stamp wells. Phase-contrast images were acquired using Metamorph software (v6.3r3; Molecular Devices) and an inverted microscope (Eclipse TE2000-E; Nikon, Tokyo, Japan) equipped with a ORCA AG camera (Hamamatsu Photonics, Hamamatsu, Japan). Cells were retrieved from micro-wells using an automated system for micromanipulation (CellCelector; Aviso, Jena, Germany). To withdraw the contents of a well, the array of micro-wells was positioned on the microscope under a layer of medium, and a capillary with an outer diameter of 100 mm (inner diameter 50 mm) was positioned directly over the top of an appropriate well. A small volume (1–5 mL) was withdrawn with the affixed syringe until the cells were removed from the well successfully. The tip was then transferred into a well of a 96-well plate containing 200 mL medium (10% hybridoma cloning factor) and the cell(s) expelled into the volume. Both extraction from the micro-well and deposition of the cells into another container (96-well plate) were monitored visually to ensure the transfer of the cells into and out of the tip.

Sortase-labeling reactions were performed as described elsewhere.(1,20,21) Briefly, E. coli cytolethal distending toxin subunit B was equipped with a C-terminal LPETG motif and purified protein conjugated to GGG-Alexa Fluor 647 via sortase-mediated transpeptidation. Western blot analysis

Coomassie staining

Proteins resolved by SDS-PAGE were visualized by Coomassie stain. Incubation 2–4 h in 50 mL Coomassie staining solution (1.25 g Brilliant Blue R, methanol [200 mL], water [250 mL], acetic acid [50 mL]). Destaining was accomplished 2–24 h in destaining solution (water, ethanol, and acetic acid in a ratio of 6:3:1).

Limiting dilution

Fluorescence imaging

Hybridoma cells were diluted to an average of 10 cells per well, and antibody secretion and antigen specificity were tested by ELISA. Positive wells were reseeded in 96-well plates, 1 cell per well.

Fluorescence scans were obtained using a variable mode imager (Typhoon 9200; GE Healthcare, Fairfield, CT).

Antibody purification from serum

Secreted anti-AF647 antibodies were affinity purified from hybridoma culture medium. After centrifugation (10,000 g, 30’ at 4C), cell culture supernatant was diluted 1:1 with 100 mM sodium acetate (pH 5.0) and applied to protein G column, washed with 3 column volumes of sodium acetate. Bound protein was eluted with glycine-HCl (pH 2.8) into neutralization buffer (250 mM Tris [pH 9.0]). Elution fractions were pooled and dialyzed against PBS ON at 4C. Protein concentration was determined with a Bradford assay. Metabolic labeling and immunoprecipitation

3 · 106 cells per sample were starved for 30 min in 1 mL methionine/cysteine-free DMEM at 37C prior to labeling. Cells were labeled for 10 min at 37C with 10 mCi/mL [35S] methionine/cysteine (expressed protein mix; PerkinElmer, Waltham, MA) to each sample. After resuspension in 1 mL pre-warmed ‘‘cold’’ culture medium, cells were incubated 5.5 h at at 37C. Immunoprecipitation with protein G beads was performed from cell medium supernatant for 3 h at 4C with gentle agitation. Samples were eluted by boiling in reducing sample buffer, subjected to reducing SDS-PAGE, and visualized by autoradiography using DMSO-PPO and exposure to XAR-5 films (Kodak, Rochester, NY).

Cross-reactivity assay

4 ng each of commercially available fluorophores coupled to different carrier proteins (BD Biosciences; Invitrogen; and Molecular Probes, Eugene, OR) were resolved by SDS-PAGE and analyzed by Western blot with anti-AF647 antibody. Molecular weight shift assay

4 mg beta-2 microglobulin (b2M)-AF647 were incubated with 40 mL cell culture supernatant from a-AF647 secreting hybridoma cells or culture supernatant from control cells (12CA5 cells) at 4C for 1 h. Gel filtration fractions of V = 0.5 mL were collected, 100 mL thereof aliquoted into a 96-well plate, and emission at 670 nm (excitation 630 nm) measured with a Spectramax plate reader (Molecular Devices, Sunnyvale, CA). Gel filtration

¨ KTA FPLC Protein mixtures were resolved using an A system using a Superdex 6 column (GE Healthcare) equilibrated with PBS (pH 7.5). FACS

Surface-bound IgM of murine splenocytes was stained with 1 mg/mL anti-IgM coupled to AF647, or left unstained for 30 min at 4C. After washing, and a second incubation

112

with hybridoma culture supernatant of either anti-AF647 or anti-HA secreting hybridoma (12CA5), or medium only for 30 min at 4C, cells were washed with media and analyzed by flow cytometry on a FACSCalibur HTS (BD Biosciences) with CellQuest Pro (BD Biosciences). Data were analyzed with FlowJo (Tree Star Inc., Ashland, OR). Results Generation of anti-AF647 antibody-secreting hybridoma cells

Alexa Fluor 647 (AF647) can be considered a hapten, a small molecule with the molecular weight of 1.3 kDa that by itself does not elicit an immune response. To render it immunogenic, we conjugated AF647 to the carrier protein bovine serum albumin (BSA), which is large (68 kDa), displays numerous immunogenic epitopes, and contains an abundance

WUETHRICH ET AL.

of lysine residues that allow coupling of haptens. Figure 1A outlines the mouse immunization protocol, which consisted of intraperitoneal administration of the antigen with Freund’s adjuvant, followed by an intraperitoneal boost including Freund’s incomplete adjuvant and a second intravenous boost excluding adjuvant. The immune response was monitored by ELISA using mouse serum collected on day 19 after primary immunization (Fig. 1B). We detected a strong response against the AF647 hapten, both when coupled to BSA or when conjugated to keyhole limpet hemocyanin (KLH). Control serum from a non-immunized mouse showed no reactivity, neither against hapten nor carrier protein. We harvested splenocytes from one mouse immunized with AF647-BSA and fused them to immortalized BALB/c-derived NS-1 plasmacytoma cells using standard protocols (Fig. 1A). The resulting hybridoma cells were selected in HAT-containing medium.

FIG. 1. Generation of anti-AF647 antibody-secreting hybridoma cells. Immunization of BALB/c mice with Alexa Fluor 647 (AF647) conjugated to carrier protein. (A) Schedule for administration of the antigen, harvest of splenocytes from a AF647-BSA immunized mouse, generation of hybridoma cells, and selection for immortalized clones. KLH, keyhole limpet hemocyanin; BSA, bovine serum albumin; IP, intraperitoneal; IV, intravenous; HAT, hypoxanthine-aminopterinethymidine. (B) Monitoring of the immune response against AF647. Mouse serum was tested on day 19 after the first immunization by ELISA. Non-immunized mice did not mount an immune response, whereas serum from mice immunized with BSA-AF647 tested positive for reaction against the fluorophore, and to a lesser extent against the carrier protein.

ANTI-ALEXA FLUOR 647 MONOCLONAL Microengraving and selective hybridoma picking

Hybridoma cells derived from NS-1/splenocyte fusions were seeded in four polydimethylsiloxane (PDMS) stamps as described by Love and colleagues,(5) aiming for an average of 1 cell per micro-well. Each array contains 25,000 microwells, so that four of these stamps correspond to the equivalent of *1000 96-well plates. The arrays were then placed upside-down on a glass slide, coated with a mixture of capture antibodies for mouse Ig and IgG (for a schematic representation, see Fig. 2A). After 20 min of incubation at 37C, the stamps with micro-wells, which retain the cells after removal of the print, were immersed in medium and returned to the incubator for continued cell culturing. The glass slide print was incubated with goat anti-mouse antibody coupled to Alexa Fluor 532 and KLH-Alexa Fluor 647, washed, and imaged with a GenePix 4000B microarray scanner (Molecular Devices). Images were analyzed using GenePix Pro 6.1 software (Molecular Devices). Double-positive wells were scored. Figure 2B shows a section of a scanned slide. The absence of a signal corresponds to micro-wells that contain either no cells or cells that do not secrete immunoglobulin. An AF532-positive signal signifies cells that secrete immunoglobulin (Ig), regardless of Ig specificity. A doublepositive signal for AF532 and AF647 indicates the presence of cells that secrete anti-AF647 immunoglobulins. The greater the AF532 or AF647 signals, the greater concentration of total or AF647-specific antibodies respectively. We selected double-positive cells to be picked from their corresponding micro-wells. With the help of 2-D coordinates, each well could be precisely matched up with its signal engraved on the glass slide and located for cell extraction. A robotic cell selector was used for the cell picking process. A capillary (inner diameter 50 mm) was positioned over the selected micro-well of a PDMS stamp immersed in cell culture medium. A volume of 1–5 mL was withdrawn, which included any cell present. The capillary content was transferred to a 96-well plate, and the picking process was repeated. Both extraction from the micro-well and deposition of the cells into another container (96-well plate) were monitored visually. At the end of the picking process, 96-well plates were placed in the incubator for continued culturing, and the PDMS stamps were discarded. Of the four PDMS stamps scored, the number of doublepositives varied between *16% and 43% (Fig. 2C). This amounts to a total number of 142,041 AF532 single-positive, or 23,135 double-positive wells. One caveat regarding these numbers is that only qualitatively good sections of the microwell arrays were scored. Towards the edges and corners of the stamp engravings on the glass slides, signal was sometimes low or insufficiently discrete. A total of 202 cells were picked with the cell selector. Of those, about 33% (or 67 cells clones) tested positive for secreted a-AF647 antibodies after a week of culturing. Some clones of the negatively tested population grew at a slower rate. After 3 weeks of culture, an additional nine wells scored positive (data not shown). Confirmation of anti-AF647 activity and isotype determination

Reactivity of antibodies against AF647 secreted by the hybridoma clones isolated by microengraving was confirmed by ELISA (Fig. 3A). Plates were coated with either KLH-

113

AF647 or a mix of anti-mouse Ig/IgG. All clones secreted immunoglobulin and all showed binding to AF-647. Clones 8 and 10 showed a lower signal for AF647 binding than the other clones. Since their immunoglobulin titer was comparatively lower as well, this is likely due to decreased Ig secretion, perhaps due to a lower growth rate, rather than decreased affinity. To determine the isotype of the secreted aAF647 Ig, ELISA plates were coated with antibodies specific for individual IgG isotypes. The clones analyzed were all of the IgG1 isotype with kappa light chains (Fig. 3B). Purification of anti-AF647 antibody from hybridoma culture supernatant

Clones isolated by microengraving were compared sideby-side with clones obtained from a limiting dilution approach done in parallel (clones L9, L10). Both L9 and L10 tested positively for anti-AF647 activity (data not shown). The four clones derived from microengraving (Fig. 4A) were each picked from a different PDMS stamp. The migration pattern of the heavy chains (around 50 kDa) was comparable for all anti-AF647 antibodies. The light chains migrated distinctly from the standard mouse light chain, usually found at *25 kDa as seen for the positive control, a hybridoma line secreting the monoclonal mouse anti-HA antibody (12CA5, a subclone of the H26DO8 hybridoma clone(6)), indicating clones of distinct origin. The mobility of the different light chains was somewhat variable, especially when comparing clone 5 with clones 8, L9, and L10, indicative of molecularly distinct light chain species. Clone 3 shows two distinct polypeptides for the light chain, one migrating slightly below 25 kDa, the other above that marker. No signal was observed for the supernatant of cultured, non-fused NS-1 cells. Anti-AF647 antibody was isolated from cell culture supernatant of four clones picked by microengraving (clones 5, 8, 9, 10) on a larger scale using protein G beads, and eluted at low pH. Reducing SDS-PAGE of the purified Ig showed acceptable purity of the different batches after visualization by Coomassie stain (Fig. 4B). Shown are the products from four clones isolated by microengraving. Again, as apparent in Figure 4A, mobility of the heavy chains is similar for the different clones, whereas there is a difference in the migration patterns of the light chains. Use of anti-AF647 antibody for Western blot analysis

The remainder of the experiments was performed using anti-AF647 antibodies secreted from clone 8 (a-AF647-8). aAF647-8 was used to probe size exclusion chromatography fractions of cytolethal distending toxin B subunit labeled with AF647 and resolved by SDS-PAGE (Fig. 5A). The antiAF647-8 antibody detects a polypeptide at *28 kDa, the expected molecular weight for EcCDTB-AF647, which corresponds to the signal observed by fluorescence scan. Figure 5B shows a titration series of AF647 coupled to b2microglobulin (b2M) with a molecular weight of 13 kDa, a part of the MHC I complex. b2M-AF647 was mixed with cell lysate to probe for the possible recognition of a background signal by a-AF647-8. The observed a-AF647-8 immunoblot signal corresponded to that of the fluorescence scan. An additional polypeptide at 25 kDa was observed at 500 ng b2M-AF647 for high concentrations of the antigen-carrier

114

WUETHRICH ET AL.

FIG. 2. Microengraving and selective hybridoma picking. From micro-wells to cell isolation. (A) Schematics for microengraving. Cells were seeded into PDMS stamps with an average of one cell per micro-well and allowed to settle. A glass-slide coated with capture antibody was fixed tightly onto the chamber and incubated for 20 min. Immunoglobulins secreted by hybridoma clones bound to capture antibody and were stained in a next step with fluorescently labeled antimouse Ig and IgG antibody, as well as KLH-AF647. (B) The glass slide was scanned, and double-positive fluorescent wells scored. A double-positive signal corresponds to a hybridoma clone secreting Ig/IgG that binds to AF647. (C) Cells were picked from four PDMS stamps. From a total of roughly 23,000 double-positive cells, 202 clones were extracted from the micro-wells. About 33% of those picked clones were later tested positive for AF647 binding. Only PDMS stamp sections with good signal quality were counted to obtain the listed numbers.

FIG. 3. Confirmation of a-AF647 activity and isotype determination. The secreted antibodies were characterized by enzyme-linked immunosorbent assay (ELISA). (A) Wells of a 96-well plate were coated either with a mix of antibodies against mouse Ig and IgG or KLH-AF647 conjugate and incubated with cell culture supernatant from picked hybridoma clones 1 week post-picking. Of the depicted clones, most showed a strong signal for both Ig secretion and anti-AF647 activity. Control, medium only. (B) All clones investigated were of IgG1 isotype, paired with a kappa light chain.

FIG. 4. Purification of a-AF647 antibody from serum. Patterns of a-AF647 antibodies secreted by different hybridoma clones were compared. (A) Metabolic labeling of cells with 35S followed by immunoprecipitation of secreted immunoglobulin with protein G beads. Eluted proteins were resolved by reducing SDS-PAGE and visualized by autoradiography. (B) Coomassie stain of protein G bead purified anti-AF647 antibody from serum, eluted with low pH, dialyzed against PBS, and resolved by reducing SDS-PAGE. Each lane contains 3 mg total protein. 115

116

WUETHRICH ET AL.

FIG. 5. Use of anti-AF647 antibody for Western blot anlysis. Anti-AF647 antibody can be used for Western blot. (A) Samples of gel filtration fractions of AF647-labeled E. coli cytolethal distending toxin (EcCDTBAF647) were resolved by reducing SDS-PAGE, transferred to a PVDF membrane after a fluorescence scan, and probed with anti-Alexa 647 antibody. Following incubation with HRP-coupled anti-mouse Ig secondary antibody, anti-AF647 binding was revealed by enhanced chemiluminescence (ECL). The Western blot signal corresponds to AF647 fluorescence. (B) Anti-AF647 antibody reveals AF647 conjugated to beta-2 microglobulin (b2MAF647) by Western blot. conjugate, probably corresponding to b2-microglobulin dimers, as b2-microglobulin has a tendency to dimerize.(7)

molecular weight fraction and indicative of successful complex formation.

Cross-reactivity with other fluorophores

Use of a-AF647-8 antibody for FACS

Commercially available fluorophores coupled to carrier proteins IgG, transferrin (TF), or ovalbumin (Ova) were resolved by SDS-PAGE and transferred onto a PVDF membrane. Subsequent immunoblotting with a-AF647-8 antibody showed that of the tested array of commonly used fluorophores, a-AF647-8 bound to two of them (Fig. 6A). The signals for Cy5 and Alexa Fluor 660 (AF660) were comparable to the signal obtained for AF647. Inspection of the molecular structures of AF647 and Cy5 (Fig. 6B) shows that both possess an identical ‘‘core,’’ highlighted in red. Differences consist of the attachment site of the linker and additional sites of sulfonation. No crossreactivity was observed with Cy5.5. While sharing the same core as Cy5 and AF647, Cy5.5 has an additional ring system conjugated to each benzene instead of the sulfonic acid groups.

Application of the a-AF647-8 antibody for fluorescenceactivated cell sorting (FACS) was tested next. Surface IgM of live mouse splenocytes was labeled with anti-IgM-AF647 antibody and subsequently incubated with hybridoma supernatant containing a-AF647-8 antibody, an IgG control (12CA5), or cell culture medium only (Fig. 8). Unstained splenocytes appear as a control population at the lower end of the red (APC) channel. AF647 labeled splenocytes incubated with either the control IgG or medium show roughly 2.5 orders of greater magnitude fluorescence signal. IgMAF647 plus medium-treated splenocytes also display a peak overlapping with the unstained population. However, no difference was observed for the maximal fluorescence intensity between the two conditions. Splenocytes labeled with IgM-AF647 and incubated with a-AF647-8 antibody show an increase in fluorescence *3.0 orders of magnitude over the unstained population. The ability of the a-AF647-8 antibody to perform in FACS in principle allows the distinction between surface-disposed fluorophore-conjugated targets and their internalized counterparts, a trait that might be useful when tracking endocytosis of fluorescently labeled molecules.

Immunoprecipitation with anti-AF647-8 antibody

To explore the stability of the a-AF647-8 antibody interaction with its epitope, b2M-AF647 was incubated with hybridoma cell culture supernatant containing either a-AF647-8 antibody or an IgG control (12CA5). The reaction mixtures were separated by size exclusion chromatography, and AF647 fluorescence was measured for individual fractions (Fig. 7). The signals for b2M-AF647 only and b2M-AF647 incubated with non-binding antibody were virtually identical. The profile for b2M-AF647 incubated with a-AF647-8 antibody showed an additional peak, corresponding to a higher

Discussion

We generated an anti-Alexa Fluor 647-specific antibody that can be used in a variety of applications (Western blot, immunoprecipitation, FACS), independent of the protein to which the fluorophore is conjugated. Immunization of mice

ANTI-ALEXA FLUOR 647 MONOCLONAL

117

FIG. 6. Cross-reactivity with other fluorophores. Specificity of anti-AF647 antibody was tested. (A) An array of commercially available proteins were tested by Western blot with anti-AF647 antibody. Cross-reactivity was observed to Cy5 and AF660. TF, transferrin; Ova, ovalbumin; AF, Alexa Fluor; FITC, fluorescein isothiocyanate; PerCP, peridinin chlorophyll. (B) Structure formulas of AF647, Cy5, and Cy5.5. The shared ‘‘core’’ structure is highlighted in red.

with AF647 coupled to a carrier protein, followed by hybridoma generation and the use of microengraving to specifically isolate clones with the desired attributes, has proven to be a useful strategy to generate anti-AF647 monoclonal antibodies. The fact that non-immunized mice do not show reactivity against the fluorophore or the carrier proteins indicates that neither contains epitopes that mice are naturally exposed to, essential for the interpretation of the ELISA results. Serum from mice exposed to AF647-BSA showed reactivity against

AF647-KLH and vice versa. AF647 is thus a potent hapten when coupled to either BSA or KLH. Microengraving provides a fast and accurate approach to isolate reasonable numbers of clones with specific secretion profiles(5,8) or to study immune response at the single cell level,(9) with the added benefit of being able to recover viable relevant cells. Of the hybridomas analyzed by microengraving, about a fifth of immunoglobulin secreting cells also tested positive against AF647. In terms of picking clones from the PDMS stamps, there is a bias towards cells that grow

118

WUETHRICH ET AL.

FIG. 7. Toward immunoprecipitation with anti-AF647 antibody. Beta-2 microglobulin (b2MAF647) was incubated with either anti-AF647 antibody or an unrelated IgG (12CA5, specific for the HA epitope). Incubation with anti-AF647 resulted in increase of apparent molecular weight of b2M-AF647, as demonstrated by size exclusion chromatography. AF647 fluorescence was retained. fast and/or secrete a high titer of antibodies. High secretors may not necessarily be the fastest growers, and these cells might later be lost during ELISA tests that are done after short periods of incubation. This partly explains why, of the actual clone picks, only *33% scored positive for both Ig secretion and AF647 binding in a later ELISA assay, which was done one week after cell picking. Indeed, some of the clones testing negative initially tested positive after an additional 2 weeks of culturing. It was reported that mouse antibody production by murine hybridomas is typically unstable in the

early stages after fusion, but repeated cloning normally produces stable clones.(10) In line with expectations, IgG1 is the predominant isotype of the antibodies secreted by the selected AF637-specific hybridomas. One clone showed an additional polypeptide running below the dominant light chain. It might be the consequence of a triple fusion of two different splenocytes with a hybridoma cell, or less likely, an unusual form of partial light chain processing. In favor of the first hypothesis are two

FIG. 8. Use of anti-AF647 antibody for FACS. Surface IgM of murine splenocytes was stained with anti-IgMAF647 antibody (control, unstained) and subsequently incubated with supernatant from either anti-AF647 or anti-HA antibody secreting hybridoma clones or medium only. The anti-IgMAF647/anti-AF647 antibody treated cells is shifted to the right on the red channel, which corresponds to an increase in fluorescence intensity.

ANTI-ALEXA FLUOR 647 MONOCLONAL

publications that report the same phenomenon observed in hybridoma cells.(11,12) Zack and colleagues traced their phenotype back to a double-fusion event. There is even a report of double-heavy chain producing hybridoma cell lines.(13) Expression of multiple light or even heavy chains is not unheard of when generating hybridomas. The diffuse nature of the strong band of the heavy chain is due to heterogeneity of glycosylation, a common modification of IgGs except in various disease-related circumstances.(14) Our a-AF647-8 antibody works very well in immunoblot with minimal background signal. a-AF647-8 recognizes its epitope seemingly equally well, regardless of the protein to which AF647 is conjugated. It therefore seems to be contextindependent for use in immunoblot. The explanation for cross-reactivity of a-AF647-8 antibody with Cy5 must lie in the related structures of the two fluorophores. AF647 was developed on the backbone of Cy5. Some anti-Cy5 antibodies are therefore likely to recognize AF647 as well. Until very recently (2012), the structure of AF647 was proprietary and at the time we generated the anti-AF647 antibodies, details of the fluorophore’s structure were not available. The structure of AF660 has not been made public to date. a-AF647-8 antibody fails to bind to Cy5.5. While Cy5.5 shares the same core with Cy5 and AF647, it has an additional ring system conjugated to each benzene instead of the sulfonic acid groups, which give it a dramatically different structure. Taken together, this suggests that the (sulfonated) edge of the substituted benzene in Cy5 and AF647 might be part of the epitope recognized by a-AF647-8 antibody. a-AF647-8 antibody binds its cognate antigen in solution. The shift in apparent molecular weight of b2M-AF647 observed by gel filtration after incubation with a-AF647-8 is explained by complex formation with the antibody. The complex is sufficiently stable to persist in size exclusion chromatography. This suggests that a-AF647-8 may also be suitable for use in immunoprecipitation. The antibody has indeed been used for that purpose (Melanie Brinkmann, personal communication). Murine splenocytes surface-labeled with AF647 shifts toward higher fluorescence intensity when incubated with aAF647-8 antibody. While many anti-fluorophore antibodies quench fluorescence,(15,16) an increase in fluorescence intensity by anti-fluorophore antibody binding is not unheard of. For example, the anti-dansyl antibody A-6398 (Life Technologies) enhances the fluorescence of the fluorophore to which it binds.(17) We conclude that the a-AF647-8 antibody described here is specific for its antigen, and is suitable for application in a variety of biochemical techniques. The a-AF647-8 antibody remains to be tested for use in other applications, such as immunohistochemistry or correlative light electron microscopy (CLEM). With CLEM, the same sample is first viewed by fluorescence microscopy and then by immunoelectron microscopy. Another fluorophore, Alexa Fluor 488, is already successfully being used(18) for this purpose. The a-AF647-8 antibody thus expands the range of possible applications of AF647. Acknowledgments

We thank J. Christopher Love and Adebola O. Ogunniyi for instruction in the use of the microfabrication and cellpicking devices.

119 Author Disclosure Statement

The authors have no financial interests to disclose. References

1. Popp MW, Antos JM, Grotenbreg GM, Spooner E, and Ploegh HL: Sortagging: a versatile method for protein labeling. Nat Chem Biol 2007;3:707–708. 2. Ballard JL, Peeva VK, deSilva CJS, Lynch J, and Swanson NR: Comparison of Alexa Fluor and CyDye for practical DNA microarray use. Mol Biotechnol 2007;36:175–183. 3. Alexa Fluor dyes spanning the visible and infrared spectrum—Section 1.3. Life Technologies. Available at: http:// www.lifetechnologies.com/us/en/home/references/molecularprobes-the-handbook/fluorophores-and-their-amine-reactivederivatives/alexa-fluor-dyes-spanning-the-visible-and-infraredspectrum.html 4. Berlier JE, Rothe AA, Buller GG, Bradford JJ, Gray DR, Filanoski BJ, Telford WG, Yue SS, Liu JJ, Cheung CY, Chang WW, Hirsch JD, Beechem JM, Haugland RP, and Haugland RP: Quantitative comparison of long-wavelength Alexa Fluor dyes to Cy dyes: fluorescence of the dyes and their bioconjugates. J Histochem Cytochem 2003;51:1699– 1712. 5. Love JC, Ronan JL, Grotenbreg GM, van der Veen AG, and Ploegh HL: A microengraving method for rapid selection of single cells producing antigen-specific antibodies. Nat Biotechnol 2006;24:703–707. 6. Niman HL, Houghten RA, Walker LE, Reisfeld RA, Wilson IA, Hogle JM, and Lerner RA: Generation of protein-reactive antibodies by short peptides is an event of high frequency: implications for the structural basis of immune recognition. Proc Natl Acad Sci USA 1983;80: 4949–4953. 7. Mendoza VL, Antwi K, Baro´n-Rodrı´guez MA, Blanco C, and Vachet RW: Structure of the preamyloid dimer of beta2-microglobulin from covalent labeling and mass spectrometry. Biochemistry 2010;49:1522–1532. 8. Yamanaka YJ, Szeto GL, Gierahn TM, Forcier TL, Benedict KF, Brefo MS, Lauffenburger DA, Irvine DJ, and Love JC: Cellular barcodes for efficiently profiling single-cell secretory responses by microengraving. Anal Chem 2012; 84:10531–10536. 9. Varadarajan N, Kwon DS, Law KM, Ogunniyi AO, Anahtar MN, Richter JM, Walker BD, and Love JC: Rapid, efficient functional characterization and recovery of HIVspecific human CD8 + T cells using microengraving. Proc Natl Acad Sci USA 2012;109:3885–3890. 10. Castillo FJ, Mullen LJ, Grant BC, DeLeon J, Thrift JC, Chang LW, Irving JM, and Burke DJ: Hybridoma stability. Dev Biol Stand 1994;83:55–64. 11. Krishnan IS, Hansen HJ, Gold DV, Goldenberg DM, and Leung SO: Co-secretion of two distinct kappa light chains by the mu-9 hybridoma. Hybridoma 1999;18:325–333. 12. Payelle-Brogard B, Ragimbeau J, Avrameas S, and Christodoulou C: Immunoglobulin double isotype-producing hybridomas isolated from an autoimmune (NZB x NZW) F1 mouse. Hybridoma 1998;17:289–297. 13. Zack DJ, Wong AL, Stempniak M, and Weisbart RH: Two kappa immunoglobulin light chains are secreted by an antiDNA hybridoma: implications for isotypic exclusion. Mol Immunol 1995;32:1345–1353. 14. Parekh R, Isenberg D, Rook G, Roitt I, Dwek R, and Rademacher T: A comparative analysis of disease-associated

120

15. 16.

17. 18.

19.

WUETHRICH ET AL.

changes in the galactosylation of serum IgG. J Autoimmun 1989;2:101–114. Wei AP, and Herron JN: Bifluorophoric molecules as fluorescent beacons for antibody-antigen binding. J Mol Recog 2002;15:311–320. Anti dye and anti hapten antibodies—Section 7.4. Life Technologies. Available at: http://www.lifetechnologies .com/us/en/home/references/molecular-probes-the-handbook/ antibodies-avidins-lectins-and-related-products/anti-dye-andanti-hapten-antibodies.html#head2 tools.lifetechnologies.com. Available at: http://tools .lifetechnologies.com/content/sfs/manuals/mp06398.pdf van Rijnsoever C, Oorschot, V, and Klumperman, J: Correlative light-electron microscopy (CLEM) combining livecell imaging and immunolabeling of ultrathin cryosections. Nat Methods 2008;5:973–980. Ko¨hler G, and Milstein C: Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 1975;256:495–497.

20. Popp MW, Antos JM, and Ploegh HL: Site-specific protein labeling via sortase-mediated transpeptidation. Curr Protocol Protein Sci Ch 15, Unit 15:3, 2009. 21. Guimaraes CP, Witte MD, Theile CS, Bozkurt G, Kundrat L, Blom AEM, and Ploegh HL: Site-specific C-terminal and internal loop labeling of proteins using sortase-mediated reactions. Nat Protocol 2013;8:1787–1799.

Address correspondence to: Prof. Hidde L. Ploegh Whitehead Institute for Biomedical Research 9 Cambridge Center Room 361 Cambridge, MA 02142 E-mail: [email protected] Received: December 6, 2013 Accepted: February 25, 2014