DAVID J. T. VAUX AND SIAMON GORDONSir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford0X1 3RE, U.K.
SUMMARY
Monoclonal antibodies were prepared to study the cytoplasmic face of latex phagolysosomesisolated from thioglycollate-elicited mouse peritoneal macrophages. Phagolysosomes obtained bysucrose flotation contained latent P-glucuronidase activity and tightly associated cellular proteinsand glycoproteins. Fluorescence-activated cell sorter analysis, scanning and transmission electronmicroscopy showed that the particle preparation contained > 98 % monomers and dimers, investedwith a smooth layer of membrane and minimally contaminated with cytoplasmic adhesions. Serafrom immunized rats bound preferentially to isolated phagolysosomes rather than intact cells andmonoclonal antibodies PL-1 and PL-4 were isolated on this basis. Indirect fluorescent, radio- andperoxidase immunobinding assays with intact and methanol-permeabilized cells confirmed thatantigens PL-1 and PL-4 were exclusively intracellular and that well-washed phagolysosomes boundboth antibodies. These antigens were found in a variety of cells from several species and inmacrophages not fed latex. Although the PL-1 antigen could not be immunoprecipitated,intracellular staining was characteristic of intermediate filament distribution, that is, it was in theform of a fine intersecting network, which collapsed, reversibly, in a rim round the nucleus upontreatment with colcemid. The staining pattern was undetectable in cells 1 h after adherence to asubstratum, but gradually appeared after 6—12 h. The PL-4 antibody has been shown elsewhere todefine a Ca2+-binding protein of approximately 20000 molecular weight, which is phosphorylatedduring phagocytosis. This antibody stained stress fibres and revealed a widespread punctatedistribution of antigen within cells at all stages after adhesion. The nature of the association betweenthese intracellular antigens and phagolysosomes and their possible role in phagocytosis are notknown.
INTRODUCTION
The structure of many biological membranes is now widely accepted as consistingof a bilayer of phospholipid in which are distributed a variety of proteins andglycoproteins (Singer & Nicolson, 1972). There is good evidence that the structureof these membranes is not symmetrical (Rothman & Lenard, 1977) and it is generallyheld that glycoproteins, for example, are universally orientated to the exterior or itstopological equivalent within the cell (Hirano et al. 1972). Much is now known aboutboth the structure and the function of many externally disposed components of plasmamembranes in a diverse range of cell types. The macrophage has received particularscrutiny both because of its major role in host defence and pathological processes andbecause it lends itself particularly well to the study of the dynamics of plasma
membrane movements (Steinman, Mellman, Muller & Cohn, 1983). In particular,the study of the macrophage has shed light on processes such as phagocytosis,receptor-mediated endocytosis, exocytosis and secretion and, more recently, theproblem of plasma membrane recycling. Many externally exposed components havebeen described on the surface of the mouse peritoneal macrophage, and characterizedboth functionally and biochemically (for a review, see Vaux & Gordon, 1981). Thecytoplasmic face of the plasma membrane has received much less attention, althoughthere is now good evidence that the cytoplasmic contractile machinery responsible forthe movement of plasma membrane during endocytosis is physically attached to it(Larsen, Tung, Murray & Swenson, 1979). As a consequence of the widespreadendocytic activity of the macrophage, it is a particularly suitable cell type for the studyboth of the cytoplasmic face of the plasma membrane per se and its interaction withthe contractile machinery within the cell.
Taking advantage of a latex flotation technique originally described forAcanthamoeba by Wetzel & Korn (1969) and applied to macrophage phagocytosis byWerb & Cohn (1972), we have made an attempt to probe the cytoplasmically orien-tated face of the macrophage plasma membrane by the preparation of monoclonalantibodies against isolated phagolysosomes.
MATERIALS AND METHODS
Reagents and mediaReagents were of analytical grade and supplied by Sigma (London, U.K.) or BDH (Poole,
Dorset, U.K.). Radioisotopes were obtained from Amersham International (Amersham, U.K.).The standard buffer, PBA, consisted of Dulbecco's phosphate-buffered saline solution 'A' (PBS,deficient in Mg2"1" and Ca2+) with lOmM-sodium azide and 0-1% (w/v) bovine serum albumin(BSA). Foetal bovine serum (Gibco Europe, Paisley, Scotland) was heat inactivated at 56°C for 30min and used to supplement Dulbecco's or Iscove's modifications of Eagle's minimum essentialmedium. Antibiotics were included in culture media (KSP: 100 [Jg/ml kanamycin, 50 /ig/ml strep-tomycin and 50/i/ml penicillin).
Animals and primary cellsSwiss Pathology Oxford (PO) mice of either sex, weighing 25-35 g, were used as a source of
peritoneal macrophages, spleen cell preparations and thymocytes. Specific pathogen-free HO.B2strain rats were obtained from the MRC Cellular Immunology Research Unit, Oxford. Residentperitoneal cells (RPC) were obtained by lavage of untreated mice and thioglycollate-elicited cells(TPC) 4 days after an intraperitoneal injection of Brewer's complete thioglycollate broth (DifcoLabs., Detroit, MI). Resident peritoneal macrophages (RPM) and thioglycollate-elicitedmacrophages (TPM) were isolated by adherence and cultivated by standard methods on tissueculture plates (Falcon Plastics, Oxnard, Ca) in medium supplemented with 5 % heat-inactivatedfoetal bovine serum (I+5). £acj7/i«-Calmette-Guerin (BCG)-activated peritoneal macrophageswere obtained as described by Ezekowitz, Austyn, Stahl & Gordon (1981).
Antibodies (ab)Rabbit F(ab')2 anti-rat Fab (RAR) was prepared as by Jensenius & Williams (1974). The rat
monoclonal antibody F4/80 was used to reveal a macrophage-specific antigen expressed on the outersurface of the plasma membrane (Austyn & Gordon, 1981). The F(ab')2 fragment of the mouse
Phagolysosome antigens 111
monoclonal anti-rat immunoglobulin G (IgG), 0X12, was a gift from Dr S. V. Hunt, Universityof Oxford. RAR was radioiodinated with Na'25! by the method of Jensenius & Williams (1974).RAR or 0X12 F(ab')2 was conjugated with tetramethyl rhodamine isothiocyanate (TMRITC;Nordic Laboratories, Maidenhead, England) in 50 mia-disodium tetraborate (pH9-3) containing0-4 M-sodium chloride by dialysing the antibody at 1 mg/ml against TMRITC at 30 Mg/ml for 12 hat 4 °C. The conjugated protein was dialysed repeatedly against PBS to remove excess fluorochrome.
Indirect radioimmune binding assays (IBA)
Trace IBA on live or glutaraldehyde-fixed cells were performed in 96-well plates (Linbro76—003-05) by the method of Austyn & Gordon (1981). IBA on methanol-fixed cells in suspensionwere performed in disposable plastic tubes (LP-3; Luckham Ltd), on isolated latex phagolysosomesin Beckman microfuge tubes. Assays were routinely carried out in triplicate at 4°C, with two PBAwashes between first and second antibody and two further washes in PBA to remove unboundradioiodinated second antibody. Values are mean ± 1 standard deviation.
Preparation of labelled latex microspheresCarboxylate-modified latex microspheres (0-9Jim diameter; CLB-i, Sigma Chemicals) were
labelled with [3H]tyramine by the method of Ito, Ralph & Moore (1979) and stored in PBA at 4CC.Subsequent testing showed that > 9 8 % of the radiolabel pelleted with the latex particles.Rhodamine-labelled latex microspheres were prepared by first conjugating carboxylate-modifiedlatex microspheres with BSA using N-hydroxysuccinimide and then labelling this latex-proteinconjugate with TMRITC as above.
Isolation of latex phagolysosomes (LP)Monolayers of mouse peritoneal macrophages in 90 mm tissue culture dishes were incubated for
2h at 37 °C with I+5 containing 1 % (v/v) of latex microsphere stock (LB-11; Sigma Chemicals) of1"1 /tfn diameter. The monolayers were then washed twice in sterile PBS and reincubated in I+5overnight. Cells were harvested by scraping in 2 ml/plate of ice-cold harvesting buffer (PBS con-taining 200mM-KCl, 10mM-NaN3, 2mM-EDTA, 1 mM-phenylmethylsulphonyl fluoride and 1 %Trasylol) and pelleted by centrifugation at 1000# for 5 min. The pellet was resuspended in ice-coldharvesting buffer containing 10 % sucrose at approximately 4 X 107 cell equivalents/ml and disrup-ted with 80 strokes in a pre-chilled tight-fitting Dounce homogenizer. The resulting homogenate wasmixed with an equal volume of cold harvest buffer containing 60 % sucrose and loaded as the bottom4 ml layer in Beckman cellulose nitrate tubes. This was overlaid with 8 ml of ice-cold harvest buffercontaining 25 % sucrose and finally 2-5 ml of ice-cold harvest buffer containing 5 % sucrose. Thesediscontinuous density gradients were centrifuged in the Beckman SW41 swingout rotor at 35 000rev./min for 1 h at 4°C. The 5 % to 25 % interface was collected with a fine needle, diluted with ice-cold harvest buffer and washed by pelleting in a Beckman microfuge for 2-5 min. The isolated LPwere then either presoaked in PBA for 1 h on ice before use in binding assays or pelleted again andresuspended in the appropriate assay buffer or lysis buffer (harvest buffer without the KC1, butcontaining 0-5 % NP-40) at 4CC for biochemical characterization.
Radiolabelling of cells and isolated LPFor the incorporation of f5S]methionine or [75Se]methionine into cellular protein, cells were
washed in PBS and resuspended in methione-free MEM containing 2-5 % dialysed and 2-5 %undialysed heat-inactivated FBS and plated at 5 X 10* cells/dish in 60 mm tissue culture dishes.[35S]methionine was added to a final concentration of 10-50 ^Ci/ml and the plates were incubatedat 37°C for 12-24 h. Cells were harvested, after being washed three times with PBS, by scrapingthem from the dish with a rubber policeman in the smallest possible volume of lysis buffer (0-5 ml/dish). To obtain [35S] methionine-labelled isolated LP, the labelled monolayers were fed up to 10 % ofstock latex microspheres in methionine-free medium for 2h before washing in PBS and harvestingin ice-cold harvesting buffer. The labelled homogenate was then processed as for unlabelled LP.
Radioiodinated isolated LP were prepared by chloramine T iodination of freshly isolated LPusing the protocol of Jensenius & Williams (1974), separated from excess 125I by extensive dialysis
112 D.jf.T. Vaux and S. Gordonagainst PBS + 10 mM-NaN3 at 4°C, pelleted in a Beckman microfuge at 4°C and extracted with ice-cold lysis buffer by vigorous vortex mixing. The latex microspheres were removed from theradiolabelled lysate by centrifugation and the lysate was then used directly either for electrophoresisor immunoprecipitation.
ImmunoprecipitationLysates and antibody preparations were cleared by microfuge centrifugation for 15 min at 4°C
before use. A 200 /il sample (4 X 106 cell equivalents) of lysate was incubated with 4 (Jg of antibodyon ice for 60-120 min. Antigen—antibody complexes were precipitated with25 (A of crude rabbit anti-rat Fab antiserum, collected by microfuge centrifugation, washed twice in lysis buffer and resus-pended in 25 ̂ 1 of 8iu-urea and 5% (w/v) sodium dodecyl sulphate (SDS) in 50mM-Tris-HClat pH 7-2 by brief sonication. This was mixed with 25 /il of sample buffer and boiled for 5 min beforeelectrophoresis.
SDS/polyacrylamide gel electrophoresis (SDS/PAGE)SDS/PAGE was performed by the method of Laemmli (1970) using 0-5 mm thick slab gels of
10 % acrylamide with 3 % acrylamide stacking gels. For autoradiography, gels were either vacuumdried at this stage (for 125I-containing gels) or further processed for scintillant impregnation accord-ing to Bonner & Laskey (1974) and then dried. Separated glycoproteins were visualized with
I-labelled concanavalin A (ConA), a gift from Dr M. Bramwell in our department. Followingdestaining, gels were washed exhaustively in PBS containing 0-4 M-NaCl and then soaked overnightin [I25I] ConA in the same buffer on a shaker. Gels were washed in several changes of the same buffer,dried and autoradiographed.
Production of monoclonal abHO.B2 rats were immunized with freshly isolated LP at about 5 X 107 cell equivalents/rat
intramuscularly with complete Freund's adjuvant. The rats received subsequent injections of freshLP intravenously without adjuvant at intervals over the next 6 months. Four days after the finalinjection spleen cells were fused with either NS-1 mouse myeloma cells (Kohler & Milstein, 1976)or Y-3 rat myeloma cells (Galfre, Milstein & Wright, 1979) and cultivated essentially as describedby Galfre et at. (1977). Supernatants were screened for binding activity to isolated LP versus intactmacrophages and any culture positive by these criteria on two consecutive screens was expanded andcloned twice in 0-33 % soft agar.
Indirect immunofluorescenceCells were seeded onto sterile 11 mm coverslips at between 105 and 5 X 105 cells per coverslip in
100lA of I+5 and allowed to adhere for 1 h. The coverslips were then washed in PBS, refed I+5,incubated overnight and washed in PBS before fixation in methanol for 6 min at — 20 CC. Saturatingconcentration of first antibody was used at 10 /il per coverslip for 1 h at 4 °C and 10 fd of OX12 F(ab')2conjugated to tetramethylrhodamine isothiocyanate, at a concentration of 25/*g/ml, was added ina humid atmosphere at 4 C for 1 h. Coverslips were then washed, dried in air, mounted on a dropof 0-22fiM filtered Mcllvain's buffer (50% glycerol in 0-2id-citric acid, 0 1 M-Na2HPO4, pH5-5)and examined using a Leitz microscope equipped with epifluorescent illumination and a 63 X oil-immersion objective lens. Results were photographed on 400 ASA Ektachrome with exposure timesbetween 10 s and 30 s.
Indirect immunoperoxidaseCells on coverslips were fixed as above and processed for indirect immunoperoxidase staining
(Hume, Loutit & Gordon, 1984) with avidin—biotin enhancement, using a commercially availablekit (Vectastain, PK4004, Seralab., U.K.).
Fluorescence-activated cell sorter (FACS) analysisCells or particles were used for IBA as above but with saturating concentrations of a
Phagolysosome antigens 113
fluorochrome-conjugated second antibody. This was either a fluorescein isothiocyanate (FITC)-conjugated RAR or a FITC- or TMRITC-conjugated 0X12 F(ab')2. Fluorochrome-conjugatedlatex particles were also used to analyse mixed LP populations. FITC-latex was obtained commer-cially (from Sigma or Polysciences) and TMRITC-BSA-latex was prepared as described above.Labelled cells, LP and latex particles were analysed on a Becton Dickinson FACS II. Whenanalysing isolated LP the threshold cutoff was inactivated and the neutral density filter in the low-angle forward scatter photomultiplier light-path was removed. The limit of operation of the machinein this state appeared to be particles of about 600 nm diameter. The krypton laser was used to excitetetramethylrhodamine fluorescence. Profiles from 105 scatter events were photographed directly orstored on magnetic tape for analysis.
Immunochemiccd techniquesMonoclonal antibody was concentrated from culture supernatant by precipitation with amm-
onium sulphate, dialysed against PBS with 10mM-NaN3 and centrifuged before use. Monoclonalantibody was also purified by affinity chromatography.
Biochemical techniquesProtein was estimated using a spectrophotometric method based on Coomassie Blue with a
commercial kit (Bio-Rad reagent kit number 500-0001), using BSA as a standard. (J-Glucuronidaseactivity was measured by a spectrophotometric method based on phenolphthalein glucuronic acid(Sigma P.0376).
Scanning electron microscopySamples were allowed to settle onto poly-L-lysine-coated glass coverslips, which were rinsed
gently by flooding with PBS and fixed in 2-5 % glutaraldehyde in cacodylate buffer (pH7-4) for30min at 4°C. The fixed glass coverslips were then washed in cacodylate buffer followed bymethanol, and then critical-point dried and coated with gold in a Polaron E5100 sputter coater(Polaron Equipment Ltd, Watford, Herts.). The samples were examined using a Jeol 100CXscanning electron microscope (SEM) with an accelerating voltage of 40 kV.
RESULTS
Characterization of phagolysosomes
Isolation. Homogenates of thioglycollate broth-elicited peritoneal cells (TPM)loaded with tritiated latex particles were separated by discontinuous sucrose gradientcentrifugation. Fig. 1A shows that at least 80 % of the latex applied to the gradient wasisolated at the 5% to 25% interface. Similar experiments using fJ'SeJmethionine-labelled TPM demonstrated the presence of cell protein at this interface (Fig. 1B) andassays of f3-glucuronidase activity confirmed the presence of appreciable lysosomalhydrolase activity in this fraction (Fig. lc). The fractions obtained from the gradientexhibited no P-glucuronidase activity unless assayed in the presence of 0*5 % TritonX-100; this latency suggesting that the latex microspheres were invested with an intactphagolysosomal membrane.
Biochemical analysis. The spectrum of proteins contained within, and firmlyassociated with, the membrane investing the phagolysosomes was characterized bySDS/PAGE and autoradiography after chloramine T and biosynthetic labelling (datanot shown). The use of [3SS]methionine label confirmed the presence of a large
114 D.jf. T. Vaux andS. Gordon
LP
8 12 16 20
16Fraction
Fig. 1. Isolation of latex phagolysosomes. A. Distribution of tritiated latex beads acrossfractions of discontinuous sucrose gradients after centrifugation of lysates of TPM fed withradiolabelled latex, B. Distribution of [75Se]methionine counts across discontinuoussucrose gradients after centrifugation of lysate of TPM labelled with [75Se]methionine andfed latex microspheres. c. P-Glucuronidase activity of latex-fed TPM after centrifugationof lysate across discontinuous sucrose density gradients. Assay in the presence of 0-5 %Triton X-100. In each case the broken line represents the position of the latexphagolysosomes at the S % to 25 % sucrose boundary.
Phagolysosome antigens 115
IM
Scatter Scatter
cquc
ncy
i
A
n
1Scatter Fluorescence
Fig. 2. FACS analysis of isolated latex phagolysosomes. A. A linear profile of low-angleforward scatter of isolated LPs. This is in effect a histogram of particle size against relativefrequency and shows a major monomer and minor dimeric peak. B. A plot of fluorescenceagainst low-angle forward scatter for isolated LP made with fluorescein-labelled latexmicrospheres. c. A logarithmic profile of low-angle forward scatter against relativefrequency from the events producing the plot in B. D. A logarithmic profile of fluorescenceagainst relative frequency for the same events as in c. In each case the profiles are the resultof counting 10s events.
n u m b e r of endogenous protein components , including bands that comigrated with
standards of actin (44xlO3Mr) and myosin (200x10^,-). The presence ofglycoproteins in the phagolysosomal membrane was established both by prelabellingTPM surfaces by the Gahmberg [3H]periodate method, and by developing SDS/polyacrylamide gels of unlabelled LP protein with [125I]ConA (not shown). Theseglycoproteins could not be removed from intact phagolysosome preparations withtrypsin or neuraminidase.
FACS analysis. FACS analysis of isolated LP provided a convenient method forassessment of the purity and characteristics of the phagolysosome population, asshown in Fig. 2. A linear scatter profile of isolated LP (Fig. 2A) shows that > 98 %of all particles detected in the preparation consisted of monomeric or dimericphagolysosomes. The dot plot of relative fluorescence against scatter (Fig. 2B) for a
116 D.jf.T. Vaux and S. Gordon
population of LP prepared with endogenously fluorescent latex microspheres confir-med the presence of dimeric forms by revealing that the dimers, by scatter criteria,expressed twice the fluorescence of the monomer peak. This result is shown graphic-ally as paired relative frequency histograms in Fig. 2c,D. Channel markers have beenadded at exact multiples of the monomer peak channel using linear scales throughout.
Confirmation of the existence of dimeric forms was obtained by using the FACS tosort isolated LP preparations into monomer and dimer fractions by means of the scattersignals and examining the resulting fractions by scanning electron microscopy. Mono-mers and dimers in > 200 fields counted at X 5000 magnification revealed a twenty-foldenrichment for morphologically dimeric forms in the FACS sorted dimer fraction.
Fig. 3. Scanning electron microscopy of latex microspheres and latexphagolysosomes.A. Untreated 4*7/an diameter latex microspheres. X 12000. B. Isolated LP made fromTPM fed 4-7 fan diameter latex microspheres. X 11 000. c. Dimer of isolated LP madefrom TPM fed 1-1 fan diameter latex microspheres. X 43 000. D. Mixed dimer from LPsisolated from TPM fed successively with 4-7 jtfn and 1-1 /an diameter latexmicrospheres. X 21 000.
Phagolysosome antigens 117
Scanning electron microscopy. Preliminary observations using transmissionelectron microscopy (not shown) suggested that the isolated LP consisted of latexmicrospheres enveloped in a closely applied membrane with very few cytoplasmiccomponents attached to the external surface. Scanning electron microscopy was car-ried out to assess further the appearance of isolated LP. Fig. 3A documents theappearance of untreated latex microspheres processed as in Materials and Methodsand Fig. 3B shows similar particles after conversion into LP. The presence of dimericforms is demonstrated by Fig. 3c. Fig. 3D shows that the ingested particles from twosuccessive waves of phagocytosis have access to a common intracellularphagolysosome pool. TPM were fed 4-7 pan diameter latex microspheres for 2h,washed, incubated overnight to ensure complete internalization and then challengedwith 1-1/im diameter latex microspheres. Mixed dimers, as seen in Fig. 3D, werefound in the subsequently isolated LP and these were not seen when separate prepara-tions of TPM fed either 4-7/im or 1-1/im latex microspheres were homogenizedtogether.
Preparation of anti-LP monoclonal antibodies
Fresh, unfixed preparations of LP from TPM were used to immunize rats atintervals of about 4 weeks, and progress was monitored by indirect radioimmunoassayon serum samples. By the second immunization the rat sera began to show selectivity,with improved titres against LP targets relative to intact TPM targets. The serafrequently showed saturation binding to LP but not to intact TPM at this stage. Titresrarely improved further after the fourth immunization.
From assays of approximately 400 hybridoma-containing wells, four stablehybridomas that produced antibodies against determinants expressed on the LP, butnot on intact macrophages were finally isolated. These were designated PL-1, PL-2,PL-3 and PL-4. Two of these antibodies (PL-2 and PL-3) showed very variablebinding to isolated phagolysosomes and an indirect immunofluorescence pattern con-sistent with a ribonucleoprotein antigen and are considered no further in this paper.
Characterization of the specificities of antibodies PL-1 and PL-4
The absence of binding to the external surface of mouse peritoneal macrophagesseen in the first screen was confirmed by a further indirect radioimmunoassay on livecells either in suspension or adherent to the wells of microtitre plates. These resultswere extended by the use of indirect immunofluorescence microscopy and FACSanalysis of cells after indirect saturation binding assays. In no case was significant(more than twice background) binding to intact cells seen. Positive controls consistedof identical cells permeabilized by fixation with methanol and, for the macrophagepopulations, of F4/80 on intact cells. Negative controls were either buffer (PBA) ormedium conditioned by the myeloma parent.
Fixation in methanol at low temperature was used to permeabilize cells and makeintracellular antigens accessible. A variety of cell types permeabilized in this way wasused to probe the tissue and species specificity of both the monoclonal antibodies.Both indirect radioimmunoassay on adherent and suspension cell populations and
118 D. J. T. Vaux and S. Gordon
12-
xE6.
•ac
2F4/8C)
,PL-1
30-
20-
10-
PL-1
PBA
1/10 I/1K0 1/13 1/9 1/27 1/81 PL-11/KXX) 1
Antibody dilution
Fig. 4. Indirect radioimmunoassays. A. PL-1 and F4/80 assayed on intact TPM. Thehighest concentration of antibody F4/80 used here was saturating (not shown), B. PL-1assayed on isolated LP. Antibody F4/80 did not bind to intact isolated phagolysosomes(not shown). Each point is the mean of triplicate determinations and the standard devia-tion is < 10% of the mean.
FACS analysis after saturation indirect immunofluorescence assays were used. Fig.4 shows the results obtained by indirect radioimmunoassay using the antibody PL-1on intact macrophages (A) and isolated LP (B). In Fig. 4A, binding is compared withthat of antibody F4/80, a determinant expressed on the external surface of themacrophage. The standard preparation of PL-1 antibody was clearly able to saturateantigenic sites when assayed in this way on isolated LP.
FACS analysis after indirect immunofluorescence binding assays was performed onlive or methanol-nxed whole cells and on isolated LP. Antibody PL-1 failed to bindto intact cells (Fig. 5A) but did bind to methanol-permeabilized cells (Fig. 5B). In Fig.5c we demonstrate binding of antibody PL-1 to isolated LP, by using scatter gatingto limit analysis to the size range of the latex microspheres.
Table 1 presents the results from many such assays and demonstrates expressionof the antigens detected by antibodies PL-1 and PL-4 in different tissues and species.In each case the results are based on the use of at least two assay techniques and thecut-off for a positive value was taken as four times the background seen in the absenceof the monoclonal antibody. It is clear that both monoclonal antibodies readily recog-nize determinants present within different cell types. Furthermore, it is apparent thatthe determinants show little species restriction. In particular, it is notable that the PL-1 antibody recognizes its antigen in the rat, the species in which the monoclonalantibody was prepared.
Phagolysosome antigens 119
uc
r req
i
i
/PL-1
\ .PBA
A
1Fluorescence
-PBA
PL-1
Fluorescence
Fluorescence
Fig. 5. FACS analysis of indirect immunofluorescence assays, A. Binding of PL-1 tointact TPM produces a fluorescence profile superimposable on that of the medium blank.B. Binding of PL-1 to methanol-permeabilized Vero cells, c. Binding of PL-1 to isolatedlatex phagolysosomes. All fluorescence profiles generated from 105 dected particles.
ttfi D. J. T. Vaux and S. Gordon
Phagolysosome antigens 121
Table 1. Species and tissue distribution of antigens PL-1 and PL-4
Cell lines were obtained from stocks at the Sir William Dunn School of Pathology.
Intracellular distribution of the PL-1 and PL-4 antigens
Binding assays on isolated phagolysosomes showed clearly that the PL-1 antigenremained associated with the cytoplasmic face of the phagolysosome membrane afterextensive washing in high salt, EDTA-containing buffers (not shown). However, itwas apparent from indirect immunofluorescence microscopy on methanol-fixed cellmonolayers that the PL-1 antigen was present in cells that are at best very weaklyphagocytic (such as the African green monkey renal epithelium cell line, Vero, as wellas in macrophages that had not been fed latex particles). In Vero cells the PL-1 antigenwas distributed as a fine network of intersecting filaments (Fig. 6A).
In an attempt to characterize the nature of the filament network further, the res-ponse of the PL-1 antigen to pretreatment of the cells with colcemid was examined.Colcemid causes the disaggregation of microtubules into tubulin monomers (Margolis& Wilson, 1978) and also disrupts intermediate filaments (Hynes & Destree, 1978) buthas no effect on the actin-based microfilaments. Fig. 6B shows that treatment of Verocells with colcemid caused an extensive disruption of the PL-1 antigen network, whichappeared to collapse to form a rim around the nucleus. This effect was freely revers-ible, however, and 90 min after removal of colcemid the PL-1 network had consider-ably recovered (Fig. 6c). By 225 min after removal of colcemid the PL-1 networkwithin Vero cells appeared morphologically to have returned to normal. It is sig-nificant to note that this recovery was not accomplished by the development ofmicrotubule-organizing centres (Brinkley et al. 1971; Watt & Harris, 1980).
Fig. 6. Intracellular distribution of PL-1 antigen by indirect immunofluorescence. A. PL-1 distribution in methanol-fixed Vero cells, B. PL-1 distribution in Vero cells methanol-fixed immediately after colcemid treatment, c. PL-1 distribution in Vero cells methanol-fixed 90 min after washing colcemid out of the incubating medium.
122 D. jf. T. Vaux and S. Gordon
Fig. 7. The effect of adherence on.PL-1 antigen distribution revealed by indirect imm-unoperoxidase labelling, A. PL-1 distribution in TPM fixed with methanol after 1 h ofadherence, B. PL-1 distribution in TPM fixed with methanol after 24h of adherence.
Indirect immunofluorescence using the PL-4 antibody on methanol-fixed cellsproduced diffuse cytoplasmic staining, sparing the nucleus, and a variable amount offine punctate staining. The use of Triton-paraformaldehyde fixation to ensurepreservation of stress fibres enabled us to demonstrate clear staining of stress fibresby the PL-4 ab. However, even when clear stress-fibre labelling was seen, diffusecytoplasmic staining and widespread punctate labelling was still apparent.
Phagolysosome antigens 123
Effects of cell adherence on the expression of the Pl-J antigen
Chance observation suggested that in the first minutes after first adherence, cellsin culture did not show an extended filamentous distribution of the PL-1 antigen. Thisobservation was followed up by indirect immunofluorescence and immunoperoxidaseexperiments using early time points of adherence with a variety of cultured cells.These experiments confirmed that in resident, thioglycollate-elicited and BCG-activated peritoneal macrophages the PL-1 antigen was not present in a morphologic-ally distinguishable form in the first hour after adherence. By 6 hours of adherence,the usual filamentous network was partially developed and by 12 h the full networkpattern was apparent. Fig. 7 shows TPM at 1 h of adherence (Fig. 7A) and 24 h ofadherence in culture (Fig. 7B). The difference cannot be explained in terms of opticalproblems with rounded cells or inadequate reagent penetration since the cells labelledapparently equally at both time points when stained with PL-4 antibody.
Biochemical characterization of the PL-1 and PL-4 antigens
The biochemical nature of the antigens recognized by the monoclonal antibody wasexamined by immunoprecipitation. The results obtained with PL-4 antibody havebeen reported elsewhere (Vaux & Gordon, 1982). This antibody defines a Ca2+-bind-ing protein of 20 X 103 M,, which is phosphorylated by phagocytosis. Despite repeatedattempts, adequate immunoprecipitation of the PL-1 antigen has not been achieved.This may be related to problems of antigen solubilization, since the PL-1 antigen isknown to be distributed as an organized cytoskeletal network.
DISCUSSION
Techniques for the examination of externally disposed membrane componentscannot always be modified for examination of the more elusive cytoplasmic face of theplasma membrane. Shearing adherent monolayers to expose 'footprints' of cytoplas-mic surface (Boyles & Bainton, 1981), thin sectioning of monolayers parallel to thesubstrate (Singer, 1979) and selective extraction followed by high-resolution electronmicroscopy of platinum replicas (Heuser & Kirschner, 1980) have all given insightsinto the morphology of the cytoplasmic surface of membranes.
An alternative approach is to isolate a population of intracellular membrane-boundorganelles whose exposed outer surface is topologically equivalent to the cytoplasmicface of a bilaminar membrane. We have used a technique described for Acanthamoebaby Wetzel & Korn (1969), and used by various groups to study phagocytosis andmembrane internalization (Werb & Cohn, 1972; Willinger, Gontas & Frankel, 1979;Muller, Steinman & Cohn, 1980; Oates & Touster, 1976; Segal, Darling & Coade,1980). Other groups have also attempted to isolate monoclonal antibodies that reactwith isolated organelles, including lysosomes (Reggio et al. 1984) and synaptic ves-icles (Matthew, Outwater, Tsavaler & Reichardt, 1982). The antigens we detectedwere tightly associated with isolated LP, but were widely distributed within cells inthe absence of a phagocytic stimulus. Intracellular constituents can interact with LP
124 D. Jf. T. Vaux and S. Gordon
during phagocytosis or adsorb during cell fractionation and mask underlying mem-brane constituents, thus complicating the immunological strategy we have adopted.
Use of a discontinuous sucrose gradient gave a yield of about 80 % of a given latexmicrosphere load at the collected interface. The remainder of the particles remainedtrapped in incompletely disrupted cells or within large aggregates of debris and werefound in the pellet. We established that the latex particles collected at this interfacewere associated with proteins synthesized by the cell, and demonstrated the importantpoint that at least some of the particles had fused with lysosomes to becomephagolysosomes. The latency of this lysosomal hydrolase activity strongly suggeststhat the latex microspheres were surrounded by an intact plasma membrane. Theintegrity of the enveloping membrane is critical to the interpretation of dual screenbinding assay results when binding to intact cells is compared with binding to isolatedLP. Transmission and scanning electron microscopy revealed that the LP prepara-tions used for immunization consisted almost entirely of latex particles enclosed in asmooth layer of membrane, minimally contaminated with cytoplasmic adhesions, andthe FACS profile measured this by showing that > 98 % of the particulate matter inthe sample consisted of monomeric or dimeric latex particles.
The assay of crude rat antisera against LP showed a considerably higher titreagainst LP than against macrophage surface membranes after only two immuniza-tions. The presence of anti-surface-membrane activity in these assays and in somesubsequent fusion assays may result from a number of causes — contamination of theLP preparation with very small quantities of highly immunogenic 'right-side-out'vesicles, the presence of some determinants on both sides of the membrane or theability of the rat immune system to raise antibody responses to initially cryptic anti-gens. Whatever the cause, anti-cell-surface reactivity was in the minority, both at thestage of the crude antisera and subsequently in the supernatants from the fusion wells.Several independent immunodetection systems were used to characterize antibodybinding to intact and permeabilized cells and isolated phagolysosomes. Both PL-1 andPL-4 antibodies bound to the interior, but not the external surface of macrophages,fibroblasts, renal epithelial cells and hepatoma cells. This suggests that theintracellular antigens recognized do not show tissue specificity, which contrasts shar-ply with the often exquisite tissue specificity of monoclonal antibodies against extern-ally exposed antigens.
The FACS has yet to be widely used for probing subcellular structures, but wedemonstrate here that it can be used satisfactorily to estimate the size distribution ofparticles in the LP preparation. The isolated LP preparation contained particles of1-1 /im diameter (95 % of the total) and a much smaller number of particles with anaverage diameter of about 2^un (3 % of the total). SEM examination of these twopopulations after sorting confirmed that the larger peak represented monomeric andthe smaller peak dimeric LP. Other studies showed that successive waves of particlesphagocytosed by macrophages reached the same compartment - these mixed dimericforms could also be demonstrated using the FACS, which can differentiate differentpopulations of latex microspheres either by size or by fluorescence. Since it is possible
Phagolysosome antigens 125
to modify the surface of the latex microspheres to engage specific immunologicalpathways of phagocytosis (Silverstein, Steinman & Cohn, 1977), this facility of iden-tification provides a powerful tool for the dissection of post-endocytic compartmen-talization of phagosomes.
The PL-1 antigen was present both on the LP by binding assay and as a fine networkof intersecting filaments throughout the cytoplasm by indirect immunofluorescencemicroscopy. There is evidence that in the mouse macrophage the secondary lysosomeis associated with arrays of both intermediate filaments and microtubules in thecytoplasm (Phaire-Washington, Silverstein & Wang, 1980). There has been consider-able argument as to whether the intermediate filaments have any active role in thecontrol of the intracellular movement of organelles (Bhisey & Freed, 1971; Pesanti &Axline, 1975; Collot, Lovard & Singer, 1984) or whether they act as passive"integrators of cellular space" (Lazarides, 1980).
The response of the PL-1 antigen to pre-treatment of the cells with colcemid isexactly as described by Lazarides (1980) for intermediate filaments. Intermediatefilaments are notoriously difficult to solubilize, and are left as the main cytoskeletalelements when cells are extracted with detergent in high salt without microtubulestabilizers (e.g. see Heuser & Kirschner, 1980). This could well explain the difficultyexperienced with immunoprecipitation — a combination of poor yield of antigen intothe lysate, and a tendency for the antigen to form high molecular weight aggregateswhen immunoprecipitated.
If the PL-1 antigen is indeed an intermediate filament subunit, or even an inter-mediate filament-associated protein, then the demonstration that the antigen is foundintimately associated with isolated phagolysosomes is of considerable interest. Sucha result would be in good agreement with the observations of Phaire-Washington etal. (1980). The association of the phagolysosome with both a non-contractile struc-tural element (the intermediate filament) and a potential motive force (themicrotubules) seen within the cell might represent a second system for movement ofintracellular organelles. This would also help to reconcile observations associatingintermediate filaments and organelles with apparently contradictory results, showingthat phagocytosis and subsequent phagolysosomal fusion are not inhibited by disrup-tors of intermediate filament organization (Kielian & Cohn, 1981).
The PL-1 antigen appears to be assembled into an intracellular filament networkafter cultured cells have become adherent to a substrate. The relatively slow time-course of this process over a number of hours may enable useful information to beobtained about the production and assembly of the network, together with necessaryconditions and the effect of membrane perturbations. It is not yet clear whether thelow levels of labelling with PL-1 seen in newly adherent cells represent a relativeabsence of the PL-1 protein or the fact that the PL-1 epitope is present only onfilamentously assembled PL-1 proteins. Further experiments to unravel the role of thePL-1 antigen both in phagocytosis and phagosome processing and in the events ofcellular adhesion are in progress.
126 D. J. T. Vaux nnd S. Gordon
This work was supported in part by a grant from the Medical Research Council, U.K. We thankMaxine Hill for expert technical assistance and Mrs Elwena Gregory for typing the manuscript.
REFERENCES
AUSTYN, J. M. & GORDON, S. (1981). F4/80, a monoclonal antibody directed specifically againstthe mouse macrophage. Eur.J. Immun. 11, 805-815.
BHISEY, A. N. & FREED, J. J. (1971). Amoeboid movement induced in cultured macrophages bycolchicine or vinblastine. Expl Cell Res. 64, 419-429.
BONNER, W. M. & LASKEY, R. A. (1974). A film detection method for tritium-labelled proteins andnucleic acids in polyacrylamide gels. Eur. J. Biochem. 46, 83-88.
BOYLES, J. & BAINTON, D. F. (1981). Changes in plasma membrane-associated filaments duringendocytosis and exocytosis in polymorphonuclear leukocytes. Cell 24, 905-914.
BRINKLEY, B. R., COX, S. M., PEPPER, D. A., WIBLE, L., BRENNER, S. L. & PARDUE, R. L.
(1971). Tubulin assembly sites and the organisation of cytoplasmic microtubules in culturedmammalian cells. J. Cell Biol. 90, 554-562.
COLLOT, M., LOUVARD, D. & SINGER, S. J. (1984). Lysosomes are associated with microtubulesand not with intermediate filaments in cultured fibroblasts. Proc. natn. Acad. Set. U.SA. 81,788-792.
EZEKOWITZ, R. A. B., AUSTYN, J. M., STAHL, P. D. & GORDON, S. (1981). Surface properties ofBCG activated mouse macrophages. Reduced expression of mannose specific endocytosis, Fcreceptors and antigen F4/80 accompanies induction of la. .7. exp. Med. 154, 60—76.
GALFRE, G., HOWE, S. C , MILSTEIN, C , BUTCHER, G. W. & HOWARD, J. C. (1977). Antibodiesto major histocompatibility antigens produced by hybrid cell lines. Nature, hand. Dob, 550-552.
GALFRE, G., MILSTEIN, C. & WRIGHT, B. (1979). Rat x rat hybrid myelomas and a monoclonalanti-Fd portion of mouse IgG. Nature, Land. 277', 131-133.
HEUSER, J. E. & KIRSCHNER, M. W. (1980). Filament organisation revealed in platinum replicasof freeze dried cytoskeletons. J. Cell Biol. 86, 212-234.
HIRANO, H., PAJUCHOUSE, B., NICOLSON, G. L., LENNOX, E. S. & SINGER, S. J. (1972).Distribution of saccharide residues on membrane fragments from a myeloma-cell homogenate:Implications for membrane biogenesis. Proc. natn. Acad. Set. U.SA. 69, 2945-2949.
HUME, D. A., LOUTIT, J. F. & GORDON, S. (1984). The mononuclear phagocyte system of themouse defined by immunohistochemical localisation of antigen F4/80. Macrophages of bone andconnective tissue. J . Cell Set. 66, 189-194.
HYNES, R. O. & DESTREE, A. T. (1978). 10 nm filaments in normal and transformed cells. Cell 13,151-163.
ITO, M., RALPH, P. & MOORE, M. A. S. (1979). In vitro stimulation of phagocytosis in amacrophage cell line measured by a convenient radio-labelled latex bead assay. Cell. Immun. 46,48-56.
JENSENIUS, J. C , & WILLIAMS, A. F. (1974). The binding of anti-immunoglobulin antibodies torat thymocytes and thoracic duct lymphocytes. Eur.J. Immun. 4, 91-97.
KIELIAN, M. C. & COHN, Z. A. (1981). Modulation of phagosome-lysosome fusion in mousemacrophages. J . exp. Med. 153, 1015-1020.
KOHLER, G. & MILSTEIN, C. (1976). Derivation of specific antibody producing tissue culture andtumour lines by cell fusion. Eur.J. Immun. 6, 511.
LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of bacteriophage T4.Nature, Land. 227, 680-685.
LARSEN, W. J. TUNG, H., MURRAY, S. A. & SWENSON, C. A. (1979). Evidence for the participa-tion of actin microfilaments and bristle coats in the internalisation of gap junction membranes.J. Cell Biol. 83, 576-587.
LAZARIDES, E. (1980). Intermediate filaments as mechanical integrators of cellular space. Nature,Land. 283, 249-256.
MARGOLIS, R. L. & WILSON, L. (1978). Opposite end assembly and disassembly of microtubulesat steady state in vitro. Cell 13, 1-8.
Phagolysosome antigens 127
MATTHEW, W. D., OUTWATER, E., TSAVALER, L. & REICHARDT, L. F. (1982). Monoclonal
antibodie8 to synaptic constituents. In N'euro-immunology (ed. J. Brockes), pp. 91—111. NewYork: Plenum Press.
MULLER, W. A., STEINMAN, R. M. & COHN, Z. A. (1980). The membrane proteins of the vacuolarsystem I. Analysis by a novel method of intralysosomal iodination. J. Cell Biol. 86, 292-303.
OATES, P. J. & TOUSTER, O. (1976). In vitro fusion of Acanthamoeba phagolysosomes. I.Demonstration and quantitation of vacuole fusion in Acanthamoeba homogenates. J. Cell Biol.68, 319-338.
PESANTI, E. L. & AXLINE, S. (1975). Phagolysosome formation in normal and colchicine-treatedmacrophages. J. exp. Med. 142, 903-913.
PHAIRE-WASHINGTON, L., SILVERSTEIN, S. C. & WANG, E. (1980). Phorbol myristate acetatestimulates microtubule and 10 nm filament extension and lysosome redistribution in mousemacrophages. 7. Cell Biol. 86, 641-655.
REGGIO, H., BAINTON, D., HARMS, E., COUDRIER, E. & LOUVARD, D. (1984). Antibodies againstlysosomal membranes reveal a 100,000-mol-wt. protein that crossreacts with purified H+, K+
ATPase from gastric mucosa.J. Cell Biol. 99, 1511-1526.ROTHMAN, J. & LENARD, J. (1977). Membrane asymmetry. Science 195, 743-753.SEGAL, A. W., DARLING, J. & COADE, S. (1980). Kinetics of fusion of the cytoplasmic granules
with phagocytic vacuoles in human polymorphonuclear leukocytes. J. Cell Biol. 85, 42-59.SILVERSTEIN, S. C , STEINMAN, R. M. & COHN, Z. A. (1977). Endocytosis. A. Rev. Biochem. 46,
669-722.SINGER, 1.1. (1979). The fibronexus: A transmembrane association of fibronectin-containing fibres
and bundles of 5 nm microfilaments in hamster and human fibroblasts. Cell 16, 675—685.SINGER, S. J. & NICHOLSON, G. L. (1972). The fluid mosaic model of the structure of cell
membranes. Science 175, 720-731.STEINMAN, R. M., MELLMAN, I. S., MULLER, W. A. & COHN, Z. A. (1983). Endocytosis and the
recycling of the plasma membrane. J. Cell Biol. 96, 1-28.VAUX, D. J. T. & GORDON, S. (1981). Overview: the macrophage in cell biology. In Methods for
StudyingMononuclearPhagocytes (ed. H. Koren, P. Edelson & D. Adams), pp. 821-860. NewYork: Academic Press.
VAUX, D. J. T. & GORDON, S. (1982). Monoclonal antibody defines a macrophage intracellularCa++-binding protein which is phosphoryated by phagocytosis. Nature, Land. 299, 70-72.
WATT, F. M. & HARRIS, H. (1980). Microtubule organising centres in mammalian cells in culture.J. Cell Sd. 44, 103-121.
WERB, Z. & COHN, Z. A. (1972). Plasma membrane synthesis in the macrophage followingphagocytosis of polystyrene latex particles. J . biol. Chem. 246, 2439-2446.
WETZEL, M. G. & KORN, E. D. (1969). Phagocytosis of latex beads by Acanthamoeba castellanii.III . Isolation of the phagocytic vesicles and their membranes. J. Cell Biol. 43, 90-104.
WILLINGER, M., GONTAS, N. & FRANKEL, F. R. (1979). Fate of surface proteins of rabbit polymor-phonuclear leukocyte during phagocytosis. II. Internalisation of proteins. J. Cell Biol. 82, 45-56.
