hh That Are Localized to Different Cytoplasmic Organelles

We describe two dynein heavy chain (DHC)like polypeptides (DHCs 2 and 3) that are distinct from the heavy chain of conventional cytoplasmic dynein (DHC1) but are expressed in a variety of mammalian cells that lack axonemes. DHC2 is a distant member of the "cytoplasmic" branch of the dynein phylogenetic tree, while DHC3 shares more sequence similarity with dynein-like polypeptides that have been thought to be axonemal. Each cytoplasmic dynein is associated with distinct cellular organelles. DHC2 is localized predominantly to the Golgi apparatus. Moreover, the Golgi disperses upon microinjection of antibodies to DHC2, suggesting that this motor is involved in establishing proper Golgi organization. DHC3 is associated with as yet unidentified structures that may represent transport intermediates between two or more cytoplasmic compartments. Apparently, specific cytoplasmic dyneins, like individual members of the kinesin superfamily, play unique roles in the traffic of cytomembranes. M ICROTUBULE-dependent motor enzymes play critical roles in many types of intracellular transport. There are two principal classes of microtubule motors: kinesins and dyneins. The kinesins comprise a superfamily of proteins, defined by sequence identity in a "motor domain." They are associated with a surprising diversity of cytoplasmic movements: spindle pole separation, chromosome movement, axonal transport, Golgi vesicle traffic, migrations in and out of axonemes, and probably a wealth of additional movements that are still to be characterized (Goldstein, 1993; Skoufias and Scholey, 1993). The family of dyneins is less diverse, but even so, it includes more than a dozen heavy chain polypeptides that contribute to making an axoneme (Gibbons et al., 1994; Rasmusson et al., 1994) and a cytoplasmic isoform that has been suggested to play roles in vesicle transport, membrane dynamics, chromosome movement, and the positioning of some organelles (for review see Holzbaur and Vallee, 1994). The importance of dynein function in axoneme motility is unequivocal, but our understanding of dynein's role in intracellular motility is less complete. Several lines of evidence implicate this enzyme in retrograde axonal transport (Schroer, 1992). Localization studies have placed both the heavy chain and a 74-kD intermediate chain of cytoplasmic dynein at kinetochores, where a minus-enddirected motor enzyme could contribute to anaphase A (Steuer et al., 1990; Pfarr et al., 1990), but there is as yet no evidence that dynein plays such a role. Microinjection of Address all correspondence to Eugeni A. Vaisberg, Department of Molecular, Cellular, and Developmental Biology, Campus Box 347, University of Colorado at Boulder, Boulder, CO 80309-0347. Tel.: (303) 4928534. Fax: (303) 492-7744. e-mail: [email protected]. function-blocking antibodies raised against bacterially expressed fragments of cytoplasmic dynein heavy chains from two species blocked the separation of mitotic centrosomes, but they had no effect on chromosome attachment to the spindle (Vaisberg et al., 1993). Genetic analysis of cytoplasmic dynein function in fungi has implicated this enzyme in both the positioning and the elongation of mitotic nuclei in Saccharomyces cerevisiae (Li et al., 1993) and in postmitotic nuclear movements in other fungi (Xiang et al., 1994; Plamann et al., 1994). Cytoplasmic dynein has also been implicated in the motion of vesicles derived from the Golgi apparatus to the centrosomal region of lysed cell models (Corthesy-Theulaz et al., 1992), and in the transport of endocytic vesicles (Bomsel et al., 1990; Fath et al., 1994; Lafont et al., 1994; Oda et al., 1995), but the relationship between these in vitro systems and membrane movements in vivo is still undefined. The diversity of functions attributed to cytoplasmic dynein, together with the ineffectiveness of our motilityblocking dynein antibodies in disrupting both chromosome movement and Golgi dynamics in vivo (Vaisberg, E.A., and J.R. Mclntosh, unpublished results), led us to suspect that cytoplasmic dynein might be more complex than was previously thought. We hypothesized that there were novel isoforms of cytoplasmic dynein involved in several aspects of cell motility, so experimental approaches focused on the one known isozyme were bound to fail. We have therefore undertaken a search for novel dynein heavy chains that are expressed in cells making neither cilia nor flagella. Here we report the identification of two such isoforms, cytoplasmic dynein heavy chains (DHCs) 1 2 and 3 1. Abbrevia t ions used in this paper:. BFA, brefeldin A; DHC, dynein heavy chain; RT, reverse transcription. © The Rockefeller University Press, 0021-9525/96/05/831/12 $2.00 The Journal of Cell Biology, Volume 133, Number 4, May 1996 831-842 831 on N ovem er 9, 2017 jcb.rress.org D ow nladed fom (DHC2 and 3), whose intracellular localizations differ markedly from that of "conventional" cytoplasmic dynein heavy chain (DHC1) and from each other. DHC2 is associated predominately with the Golgi apparatus and is likely to be involved in the organization of this organelle. DHC3 localizes to cytoplasmic structures that do not correspond well to any of the well-defined membranous compartments of cytoplasm. Materials and Methods Reagents and Antibodies Unless otherwise specified, biochemical reagents were purchased from Sigma Chemical Co. (St. Louis, MO). Enzymes for molecular biology were purchased from Promega Corp. (Madison, WI) or Boehringer Mannheim Biochemicals (Indianopolis, IN). The mAb p58-9 (Bloom and Brashear, 1989) was a gift from Dr. G. Bloom (University of Texas Southwestern Medical Center, Dallas, TX); the mAb to mannose-6-phosphate receptor (Lombardi et al., 1993; Dintzis et al., 1994) was a gift from Dr. S. Pfeffer (Stanford University, Stanford, CA); the mAb to et-mannosidase II (Burke et al., 1982) was purchased from BAbCO (Richmond, CA). Texas red-conjugated human transferrin vital fixable markers for lysosomes (LysoTracker Red DHD-99) and mitochondria (MitoTracker Green FM) were purchased from Molecular Probes, Inc. (Eugene, OR). Reverse Transcription, PCR, and Analysis of Clones Purification of mRNA, cDNA synthesis, and PCR with degenerate primers were performed as described previously (Vaisberg et al., 1993) with two modifications: PCR was performed with an annealing temperature of 37°C, and the pDK101 vector (Kovalic et al., 1991) (No. 77406; American Type Culture Collection, Rockville, MD) was used to clone the resulting reaction products. RNA Walk, Library Screening, Molecular Biological Techniques, and Sequence Analysis A random primed eDNA library from a human adenocarcinoma cell line (Stratagene, La Jolla, CA) was screened with the DHCl-specific probe Hp22 (Vaisberg et al., 1993) using the Genius 1 labeling system (Boehringer Mannheim Biochemicals) according to the manufacturer's protocol. eDNA for a 5' "walk" was primed with a degenerate oligonucleotide that encodes the MNPGYAG sequence (primer "b" in Vaisberg et al., 1993). Subsequent tailing with dCTP and amplification by two rounds of PCR with sequence-specific primers (1.1 and 1.2 for DHC1 and 2.1 and 2.2 for DHC2; see Fig. 1 B) and an anchor dG oligonucleotide were performed as described in (Lee et al., 1993), except dITP residues were incorporated into the adapter (Schuster et aL, 1992). The resulting products were cloned into the pDK101 vector and sequenced using the ATaq Cycle Sequencing kit (United States Biochemical Corp., Cleveland, OH). To account for possible PCR errors during reverse transcription (RT)-PCR and RACE, sequences encoding DHC2 were verified by sequencing three independently isolated clones. The sequence of the 3' 2 kb of the library clone for DHC1 was additionally verified by sequencing the corresponding RACE clone. Sequence alignment and analysis was performed using GCG Wisconsin Package Version 8.1-UNIX (Program Manual for the Wisconsin Package, Version 8, September 1994; Genetics Computer Group, Madison, WI). For Southern blot analysis, genomic DNA was isolated from HeLa cells as described in Sambrook et al. (1989). 20 p~g of human genomic DNA was digested with restriction endonucleases as indicgted, blotted, and hybridized with DHC probes at high stringency as described in Vaisberg et al. (1993). For Northern blotting, total RNA was isolated using RNA Isolator (Genosys Biotechnologies Inc., The Woodlands, TX). 10 Izg of RNA was separated on a 0.8% formaldehyde-denaturing agarose gel, transferred to a nylon membrane, and probed as described in Vaisberg et al. (1993). Preparation of Immunogens, Purification of Antibodies, and Immunoblotting cDNA clones used for the expression of the antigens were obtained by direct PCR from HeLa RNA using isoform-specific primers, thus minimizing the possibility of accumulating PCR errors in cycles of amplification. The primer pairs used were: 1.6-1.3 to produce DHC1/1; 2.6-2.3 to produce DHC2/1; 3.4-3.3 to produce both DHC3/1 and DHC3/2; and 1.5-1.4 and 2.5-2.4 to produce DHCI/2 and DHC2/2, respectively (Fig. 1). The identity of the clones was verified by end sequencing. Gene fragments encoding the antigens were cloned into the pET5b expression vector (Studier et al., 1990) (DHC1/1, DHC1/2, DHC2/2, and DHC3/2) or pRSET vector (Invitrogen, San Diego, CA) (DHC2/1 and DHC3/1) and expressed in Escherichia coil strain BL21(DE3). Inclusion bodies containing the expressed protein were purified (Lin and Cheng, 1991) and fractionated by gel electrophoresis. Bands of expressed proteins were excised and electroeluted using an Elutrap device (Schleicher & Schuell, Inc., Keene, NH). Immunization, preparation of columns for affinity depletion and purification, and chromatography were performed as previously described (Vaisberg et al., 1993). Serum from rabbits immunized with each of the smaller (see Fig. 4) DHC fragments was depleted for cross-reactivity with the two other isoforms by repeated passages through the columns with coupled larger fragments until no cross-reactivity was detected by immunoblotting. This preparation was affinity purified on a column with the larger antigen representing the same isoform used for immunization. For example, anti-DHC2 antibody was prepared by immunization of rabbits with DHC2/2 fragment, sequential immunodepletion of the resulting serum on columns with DHC1/1 and DHC3/1 polypeptides, and then affinity purification on the column with DHC2/1 polypeptide. Immunoblotting was performed with 4% dry milk as a blocking reagent and developed using an ECL development reagent (Amersham Corp., Arlington Heights, IL) according to manufacturer's protocol. Antibodies to DHC1, 2. and 3 were used at concentrations of 0.2-0.5 ~g/ml. Preparation of CeU Extracts and Fractionation on Sucrose Gradients HeLa and NRK cells were collected by scraping into ice-cold PBS; washed two times with PBS and once with PME buffer (100 mM Pipes, 5 mM MgSO4, 1 mM EGTA, 1 mM DT'I?; pH 6.9); resuspended in PME, supplemented with protease inhibitors (10 i~g/ml of leupeptin, pepstatin A, and aprotinin; 1 mM PMSF) to the final density 108 cells per ml; lysed by gentle sonication or by passing 10-15 times through a 26-gauge needle; and clarified by centrifugation in an SS34 rotor (Sorvall Instruments Div., Newton, CT) at 17,000 g~v for 15 min. 250 I.tl of the resulting extract was loaded on the top of 5-20% gradient of sucrose in PME and centrifuged in an SW60Ti rotor (Beckman Instruments, Inc., Fullerton, CA) at 360,000 gay for 5.2 h. Thyroglobulin (19S), catalase (llS), and cytochrome C (2.2S) were used as standards. Fractions from the gradients were analyzed by immunoblotting as described above. Cell Culture, Immunofluorescence, and Microinjection NRK, COS-7, and HeLa cells were grown in DME (Sigma Chemical Co.), supplemented with 10% FCS (GIBCO BRL, Gaithersburg, MD), 50 U/ml penicillin G, and 50 U/ml streptomycin sulfate. A431 cells were grown in DME, supplemented with 10% calf serum (GIBCO BRL), and CFPAC-1 cells were cultured in IMEM (Sigma Chemical Co.), supplemented with 10% FCS. For immunofluorescence, cells were plated on glass coverslips, and allowed to grow for 24-36 h, fixed with 4% paraformaldehyde in PBS for 15 min, permeabilized with 0.1% Triton X-100 in PBS for 5 rain, and processed for immunofluorescence as previously described (Nislow et al., 1990). All anti-DHC antibodies were used at a concentration of 5-10 ~g/ml. During the course of this study we experimented with several fixation protocols for immunofluorescence. Our experience with the microtubule cytoskeleton had led us to favor fixation in cold (-20°C) methanol, usually followed by cold acetone. Upon application of this protocol to cells destined for staining by DHC3 antibodies, we found label specifically associated with mitochondria. Aldehyde fixation, followed by detergent permeabilization, on the other hand, showed little or no such association. By comparing these localizations with various marker antibodies, we discovered that cold methanol fixation makes staining with several markers, such as 13COP, congruent with mitochondrial markers, a result that is almost certain to be erroneous. We therefore fixed specimens by ultrarapid freezing and freeze substitution. Cultured cells were plated upon formvarcoated, carbon-stabilized gold grids, as used for EM. Individual grids were plunge frozen in liquid ethane, freeze substituted in acetone at -90°C, and fixed in paraformaldehyde and glutaraldehyde as the specimen was warmed to 0°C; then the cells were rinsed in pure acetone, rehydrated in PBS, and processed for immunofluorescence. These specimens showed the expected distributions for 13COP, mitochondrial, and Golgi markers, and DHC staining was far more similar to that described here with aideThe Journal of Cell Biology, Volume 133, 1996 832 on N ovem er 9, 2017 jcb.rress.org D ow nladed fom h serving cytomembranes for immunofluorescent localization. Microinjection of antibodies to DHC2 and preimmune immunoglobulins (both at 4.7 mg/ml) was performed as previously described (Vaisberg et al., 1993). After injection, NRK cells were incubated at 37°C for 3 h, and then fixed and processed for immunostaining with antibody to ct-mannosidase II (FITC) and to the injected antibody (Texas red). The dispersion of the Golgi apparatus in injected and control cells was assessed by eye.