In Situ Hybrid Hybridization at the Electron Microscope Level" Detection by Autoradiography and Colloidal Gold

In situ hybridization has become a standard method for localizing DNA or RNA sequences in cytological preparations. We developed two methods to extend this technique to the transmission electron microscope level using mouse satellite DNA hybridization to whole mount metaphase chromosomes as the test system. The first method devised is a direct extension of standard light microscope in situ hybridization. Radioactively labeled complementary RNA (cRNA) is hybridized to metaphase chromosomes deposited on electron microscope grids and fixed in 70% ethanol vapor; hybridization sites are detected by autoradiography. Specific and intense labeling of chromosomal centromeric regions is observed even after relatively short exposure times. Interphase nuclei present in some of the metaphase chromosome preparations also show defined patterns of satellite DNA labeling which suggests that satellite-containing regions are associated with each other during interphase. The sensitivity of this method is estimated to be at least as good as that at the light microscope level while the resolution is improved at least threefold. The second method, which circumvents the use of autoradiographic detection, uses biotin-labeled polynucleotide probes. After hybridization of these probes, either DNA or RNA, to fixed chromosomes on grids, hybrids are detected via reaction with an antibody against biotin and secondary antibody adsorbed to the surface of colloidal gold particles (~20 nm in diameter). Gold particles bind specifically both directly over centromeric heterochromatin and along the associated peripheral fibers. Labeling is on average ten times that of background binding. This method is rapid and possesses the potential to allow precise ultrastructural localization of DNA sequences in chromosomes and chromatin. In situ hybridization has become a classical method for mapping DNA sequences in cytological preparations and it has been used extensively to map repeated genes in polytene and metaphase chromosomes at the light microscope (LM) level (see reference 1 for review). Recent refinements, which enhance the efficiency of hybridization and the sensitivity of hybrid detection, have made it possible to identify and localize specific cellular or viral DNA and RNA sequences even when present in low copy number (2-12). Furthermore, by using high molecular weight radioactive-probe networks (3), or by including dextran sulfate in the hybridization mixture (13, 14), unique sequences have been mapped to metaphase chromosomes after autoradiographic exposure of 5-22 d. A few attempts have been made to extend in situ hybridization to the electron microscope (EM) level. However, most of these involved hybridization to tissue sections, tissue blocks, or whole cells (see reference 15 for review); no general method is available for hybridization to whole mount metaphase chromosomes or chromatin spreads. Such a method would (a) allow gene mapping in systems where the metaphase chromosomes are very small (e.g., Drosophila, avian microchromosomes, double minute chromosomes), (b) provide a technique for fme structure mapping of genes, (c) permit investigation of ultrastructural features of chromosome organization, and (d) facilitate the analyses of nascent transcripts in chromatin preparations made by the Miller spreading technique (16, 17), thus permitting positive identification of actively transcribing sequences in chromatin and allowing a detailed study of their THE JOURNAL OF CetL BIOLOGY • VOLUME 95 NOVEMBEr 1982 609-618 © The Rockefeller University Press . 0021-9525/82/11/0609/10 $1.00 609 on O cber 0, 2017 jcb.rress.org D ow nladed fom organization as well as of the structure of the chromatin underlying these regions. Because of the potential use of a general technique for EM level in situ hybridization, we set out to develop the necessary methodology using mouse satellite DNA and whole mount metaphase chromosomes as a simple, well-characterized test system. Since these DNA sequences reside in the centromeric heterochromatin of all mouse chromosomes except for the Y (18, 19), we could rapidly evaluate experiments by examining the centromere region of any metaphase chromosome without the requirement for specific chromosome identification. We first show that it is possible to perform in situ hybridization to chromosomes on EM grids using autoradiographic detection. Specific labeling of centromeric heterochromatin is obtained with at least a threefold improvement in resolution over the LM. To achieve the variety of objectives described above, it is essential to employ a method for detecting hybridized probes with greater speed and resolution than that provided by EM autoradiography. Several groups have attempted to develop sensitive procedures for detecting nonradioactively labeled polynucleotides. Fluorescence detection methods have been described that employ antibodies specific for DNA:RNA hybrids (20, 21) or RNAs that are Y-end labeled with fluorochromes (22-24). However, fluorescence techniques are unsuitable for EM studies. Manning et al. (25) developed an innovative scanning EM method based on avidin-polymer sphere binding to biotin-coupled nncleic acid probes hybridized to polytene chromosomes. Recently, Langer et aL (26) synthesized nucleotide derivatives directly modified with biotin. Such derivatives are incorporated enzymatically into DNA and RNA molecules that can function efficiently as hybridization probes (26-28). Hybrids are detected in the LM after reaction with a rabbit antibody specific for biotin (RAB) and then either fluorescent or enzymatic markers coupled to a second antibody against rabbit IgG. This approach for preparing and detecting nucleic acids labeled with biotin seemed the one most readily adaptable for EM studies. Thus, we developed a method for detecting biotin-substituted probes hybridized to chromosomes on grids which employs RAB and secondary antibody adsorbed to colloidal gold particles. This technique provides a simple and rapid method for mapping DNA sequences on chromosomes at the EM level. MATERIALS AND METHODS Cell Culture and Chromosome Preparation Mouse L929 cells (American Type Culture Collection, Rockville, MD) were maintained either in Joldik's modified suspension medium (Gibeo Laboratories, Grand Island Biological Co., Grand Island, NY) supplemented with 10% fetal calf serum (Irvine Scientific, h'vine, CA) or in synthetic medium, Synmed (Centaurus, Anaheim, CA), plus glutamine. Metaphase cells were obtained by selective detachment from semiconfluent cultures after incubation with Colcemid (Gibco Laboratories) at 50-80 ng/ml for 4-12 h. Cells were pelleted by centrifugation at 1,000-2,000 rpm for 10 min at room temperature in a Sorvall GLC-2 centrifuge. Pellets were resuspended in a small volume of growth medium (e.g., ceLLs from a T25 flask were resuspended in 1-1.5 ml). Cells were lysed by gently mixing an equal volume of cell suspension and 1.0% Nonidet P-40 (NP-40, BRL, Bethesda, MD) (pH 8-9). Alternatively, ceils were lysed by 1:150 dilution in 1 mM Tris (pH 7.4); this mode of lysis results in somewhat less condensed chromosomes. Gold EM grids of either 400-mesh (Ted Pella, Inc., Tustin, CA) or 460-mesh hexagonal pattern (Polaron, Doylestown, PA) were coated with 1% parlodion and a thin layer of carbon. Samples of lysed metaphase cells were deposited onto grids by centrifugation through 0.5 M sucrose (pH 8-9), essentially as described by Rattner et al. (29). Grids were rinsed in 0.1% Photoflo (Eastman Kodak, Rochester, NY) and air-dried while held in forceps. Dry grids were stored on filter paper in petri dishes at room temperature 610 THE JOURNAL Of CELL BIOLOGY-VOLUME 95, 1982 and used within 12 wk. AU electron microscopy was performed on a Siemens 1A operated at 60 or 80 kV. Preparation of Mouse Satellite DNA and 3 H-cRNA Mouse satellite DNA was a gift from Dr. Lorraine Lica (University of California, Irvinc). SateLlite sequences were purified from L929 cells and separated from main band DNA by three cycles of equihbrium centrifugation in cesium chloride plus Hoechst A33258 (Calbiochem-Behring Corp., La Jolla, CA), according to Manuelidis (30). Tritium-labeled RNA complementary to this DNA was prepared using Escherichia coil RNA polymerase (the generous gift of Dr. Michael Chamberlin, University of California, Berkeley) and 3H-nucleoside triphosphates (New England Nuclear, Boston, MA), essentially according to Pardue and Gall (31). RNA complementary to Clostridium pe~fringens DNA (Sigma Chemical Co., St. Louis, MO) was synthesized using the same protocol. The specific activity of radio-labeled cRNA was estimated to be 106 cpm//~g. Bio-cRNA was prepared according to an identical protocol, except that biotinlabeled UTP (Bio-UTP) (26) was substituted for UTP. Nick Translation of Mouse Satellite DNA Nick-translated DNK probes were prepared essentially according to Rigby et al. (32), except that most reactions were run with biotin-substituted deoxy-UTP (Bio-dUTP). I /tg of DNA was labeled by E. coli DNA polymerase 1 in the presence of 25 ~M each dATP, dCTP, and Bio-dUTP, plus 10/~Ci 3H-dATP (14.2 Ci/mmol), after DNase 1 nicking. Control reactions were carried out in parallel with dTTP instead of BIn-dUTP. Reactions were terminated by the addition of EDTA and the DNA was purified by Sephadex G-50 chromatography. Samples were made 0.1 M sodium acetate (pH 5), tRNA was added (150 #g/ml final concentration), and the DNA was ethanol precipitated. One DNA sample was labeled using a BRL nick translation kit according to the suppher's instructions with the substitution of Bio-dUTP for dTTP. This reaction was terminated with EDTA and the DNA was phenol-extracted. Most of the resulting phenol phase was discarded and the remainder of the sample was dialyzed against 0.01 M Tris (pH 8), 0.1 mM EDTA. DNA containing biotin nucleotide is referred to as Bio-DNA or Bio-sat DNA (for biotin-labeled satellite DNA). DNA labeled in parallel control reactions with dTTP is referred to as TDNA. Colloidal Gold Preparation and Labeling Gold colloids were prepared as described by Geoghegan and Ackerman (33). 500 ml of aqueous 0.01% wt/vol gold chloride (HAuCh. 3H20; Fisher Scientific, Pittsburgh, PA) was reduced with 1% wt/vol sodium citrate using 13 ml or 20 mL The resulting average particle sizes, 20 and 18.5 am, respectively, were determined relative to 85-nm latex size markers. The protein concentration required for stable labeling of gold colloids was determined by the salt flocculation assay (33) with gold sols adjusted to pH 7.5-8.5. For preparation of stock antibody-labeled gold particles, 10 ml of gold colloid was adjusted to the appropriate pH and mixed with the appropriate amount of secondary antibody (10-20/~g protein/ml colloidal gold, as determined above). After 10 rain, polyethylene glycol 20M (PEG 20M, Union Carbide Corp., San Diego, CA) was added to a final concentration of 0.5 mg/ml and unbound antibody was removed by centrifugation at 53,000 g for 1.5 h at 4°C in polycarbonate centrifuge tubes in a Beckman J21B centrifuge (Beckman Instruments, Spinco Div., Pale Alto, CA). The resulting pellet was usually very soft, although small dark clumps were occasionally observed. The supernatant, containing unbound antibody, was removed carefully and the pellet resuspended in 10 ml of buffer B (0.15 M NaCI, 0.05 M Tris [pH 7.0], I mM MnCI2, 0.5 mg/ml PEG 20M). A second pellet was obtained by centrifugation as before and it was also resuspended in buffer B. Aggregates were removed by low speed centrifugation (400-700 g) for 10-20 rain. Antibody-labeled colloids were stored at 4°C. These stocks usually remained active and free of precipitates or aggregates for many weeks. Preparations which turned dark or developed precipitates were discarded. LM In Situ Hybridization and C-banding Metaphase squashes were prepared from Colcemid-arrested L929 cells using 45% acetic acid fixation; these squashes were hybridized to 3H-cRNA against mouse satellite DNA according to Pardue and Gall (31) with the addition of the 2 x SSC incubation (1 x SSC = 0A5 M NaC1, 0.015 M Na citrate [pH 7.01) described by Bonnet and Pardue (34). Approximately 50,000 cpm was applied to each shde. For C-banding, chromosomes were treated as for in situ hybridization with the elimination of the RNase treatments, and incubated in 6 x SSC without nucleic acid probe. After this mock hybridization, slides were stained with on O cber 0, 2017 jcb.rress.org D ow nladed fom Giemsa's (Harleco Azure B, Scientific Products, McGaw Park, IL) in 0.01 M phosphate buffer (pH 7.2). Slides were analyzed using an Olympus BHA microscope and photographed on Kodak High Contrast Copy film. EM In Situ Hybridization With Autoradiographic