J. Cell Sri. 79, 137-149 (1985) 137Printed in Great Britain © The Company of Biologists Limited 1985

DIRECT TRANSFER OF /3-GLUCURONIDASE FROM

MOUSE MACROPHAGES TO OTHER TYPES OF CELL

M. F. DEAN, A. McNAMARA AND B. M. JENNE

Kennedy Institute of Rheumatology, 6 Bute Gardens, Hammersmith, London W6 7DW,UJt

SUMMARYRabbit polyclonal antibodies were raised against /J-glucuronidase purified from mouse liver.

This antiserum immunoprecipitated the /J-glucuronidase secreted by mouse fibroblasts but did notcross-react with the same enzyme isolated from human tissue. The ^-glucuronidase present inmouse 3T3 fibroblasts and mouse peritoneal macrophages was clearly identified by indirectimmunofluorescence, using the antiserum and an FITC-conjugated second antibody, while humanfibroblasts with normal levels of /3-glucuronidase activity did not fluoresce when tested with thesame reagents. A range of human fibroblasts, human neuroblastoma and rat glioma cells did notfluoresce when incubated with the antibody but did fluoresce after they had been co-cultured for24 h with mouse macrophages, showing that mouse /J-glucuronidase had been transferred fromadherent macrophages into adjacent recipient cells. Transfer took place even when receptor-mediated endocytosis was blocked with a suitable competitive ligand, the transferred enzyme beingvisible mainly as a bright punctate fluorescence with a lysosome-like distribution. Macrophagesthus have the potential to act as donors of lysosomal enzymes to a wide range of recipient cells andto transfer enzymes to them during direct cell-to-cell contact.

INTRODUCTION

Lysosomal enzymes such as /3-glucuronidase can be transferred from normal donorcells to deficient recipients by two distinct mechanisms (Dean & Diment, 1984).Indirect transfer involves secretion of enzymes by donor cells, followed by theirbinding to specific receptors present in the plasma membrane of recipient cells(Neufeld, Lim & Shapiro, 1975). Different types of cell express different receptors.Fibroblasts, for example, have receptors that recognize mannose 6-phosphate(Kaplan, Achord & Sly, 1977), while the receptors on macrophages bind enzymeshaving mannose (Stahl, Rodman, Miller & Schlesinger, 1978) or fucose residues(Shepherd, Lee, Schlesinger & Stahl, 1981). Binding is followed by rapid in-ternalization and subsequent localization of the enzyme within lysosomes (vonFigura & Kresse, 1974; Bach & Liebmann-Eisenberg, 1979). In the secondmechanism enzymes can be transferred directly from donor lymphocytes ormacrophages by a process for which cell-to-cell contact is obligatory (Olsen, Dean,Muir & Harris, 1982; Olsen et al. 1983; McNamara, Jenne & Dean, 1985) and theeventual distribution of transferred enzyme within the recipient cell appears to belysosomal.

Most studies on these mechanisms of transfer have involved direct measurement oftransferred enzyme activity within the recipient cells and for this reason have been

Key words: /3-glucuronidase, macrophages, enzyme transfer.

138 M. F. Dean, A. McNamara and B. M. Jenne

limited almost exclusively to enzyme-deficient mutant fibroblasts. In order to extendthese studies to other types of non-enzyme-deficient cells, an effective method ofdistinguishing endogenous donor enzyme from newly acquired enzyme is required.One such method is indirect immunofluorescence, using a first antibody that willreact only with donor enzyme and will not cross-react with endogenous enzymepresent in the recipient cells. In this paper we describe the preparation of one suchspecies-specific antibody raised against purified mouse /3-glucuronidase and its use toshow that this enzyme can be transferred directly from mouse macrophages to arange of other cells with normal endogenous levels of /3-glucuronidase when indirecttransfer is blocked with an appropriate competitive ligand.

MATERIALS AND METHODS

MaterialsMouse 3T3 fibroblasts, rat C6 glial cells and human IMR32 neuroblastoma cells were purchased

from Flow Laboratories, Irvine, Ayrshire, U.K. Human fibroblasts deficient in /3-glucuronidase(GM151) and hexosaminidase-A-deficient firbroblasts (Tay-Sachs Disease, GM0221) were ob-tained from the Human Genetic Mutant Repository, Camden, NJ, U.S.A. Iduronate-sulphatase-deficient fibroblasts from a patient with Hunter disease were a gift from the Paediatric ResearchUnit, Guy's Hospital, London, U.K. and multiple-enzyme-deficient fibroblasts from a patientwith mucolipidosis II was a gift from the Enzyme Laboratory, Institute of Child Health, London,U.K. Tissue culture media, supplements, sera and Dulbecco's phosphate-buffered saline (PBS)were supplied by Flow Laboratories, fluorescein isothiocyanate (FITC)-conjugated goat anti-bodies, anti-rabbit immunoglobulin G (IgG) and Freund's adjuvant from Miles Laboratories,Stoke Poges, U.K. and Concanavalin A-Sepharose from Pharmacia Fine Chemicals, Uppsala,Sweden. Protein reagent and standards were purchased from Bio-Rad Laboratories, Watford,U.K. and 4-methylumbelliferyl /5-D-glucuronide substrate from Koch Light Laboratories, Bucks,U.K.

Purification of f}-glucuronidase antigen/3-Glucuronidase antigen was isolated from batches of 10 g wet weight CBA mouse livers, which

were homogenized in SO ml of 0-lSM-NaCl containing 0-01 M-Tris-HCl (pH7-0) and the enzymewas precipitated in 50% saturated ammonium sulphate. The precipitate was re-dissolved in Trissaline buffer and the /J-glucuronidase was purified some 250-fold by chromatography onConcanavalin A—Sepharose as described previously (Diment & Dean, 1983). Final purification waseffected by chromatography on an antibody affinity column linked to CNBr-activated Sepharose(Dean et al. 1985). The resulting preparation had a specific activity of 40000 units per mg proteinand ran as a single band following electrophoresis on SDS/polyacrylamide gel.

Preparation of antiserum

A 6-month-old New Zealand White rabbit was immunized intradermally with 50 y% of mouseliver /J-glucuronidase in 2 ml of Freund's complete adjuvant. A second injection of a further 50 /igin complete adjuvant was given 6 weeks later and a third injection in incomplete adjuvant after afurther 4 weeks. The rabbit was bled 3 weeks after the final immunization, the blood was allowed toclot overnight at 4CC and was centrifuged for 30min at 2500 £, and serum was stored in batches at-20°C.

Immunoprecipitation of fi-glucunmidaseSamples of /J-glucuronidase from CBA mouse liver, mouse 3T3 fibroblasts and human placenta

were partially purified by affinity chromatography on Concanavalin A-Sepharose and diluted in

Macrophage fi-glucuronidase in other cells 139

PBS (pH7-4) to an activity of 1000 units per ml. The rabbit anti-(mouse /3-glucuronidase)antiserum was diluted 1 in 100 and 1 in 1000 with PBS, 100/zI of each dilution mixed with 100 jd of/3-glucuronidase and 100 /d of a 10% (v/v) suspension of Stapkylococcus aureus protein Aadsorbent (Miles Laboratories Ltd, Sloughf U.K.) and the enzyme was immunoprecipitated asdescribed previously (Dean, Olsen & Muir, 1982). The activity of non-immunoprecipitatedenzyme remaining in solution was determined and compared with controls in which specificantiserum was replaced with pre-immune serum.

Culture of cellsMouse 3T3 fibroblasts and macrophages were grown in medium 199 with 5 % (v/v) foetal calf

serum, rat C6 glial cells in medium 199 containing 15 % (v/v) donor horse serum and 5 % (v/v)foetal calf serum and human IMR-32 neural cells in medium 199 plus 10 % (v/v) foetal calf serum.Human fibroblasts were grown in Eagle's minimum essential medium containing 10 % (v/v) foetalcalf serum. All culture media were supplemented with glutamine (ZmmolP1) Bnd SOi.u.penicillin/SO fig streptomycin per ml.

Endocytosis of secreted fi-glucuronidaseMedium was collected from macrophages either cultured alone for 24 h in 35 mm Petri dishes or

co-cultured with ^-glucuronidase-deficient fibroblasts in the same dish, as described previously(McNamara et al. 1985). Pooled media were then filter-sterilized and 1-ml samples from each wereadded to duplicate confluent cultures of /S-glucuronidase-deficient fibroblasts in 1-5 cm diameterLinbro multi-wells (104 cells per well) to determine the rate of uptake of the /3-glucuronidasesecreted by the macrophages. After a further 24 h cells were washed with PBS, lysed in 1 ml of0-1 % (v/v) Triton X-100 in water and their /3-glucuronidase activities were determined. Non-specific endocytosis was measured in control cultures containing 0-01 M-mannose 6-phosphate toblock specific receptor-mediated endocytosis.

Cell interactionDonor macrophages were elicited from CBA mice by injecting them with 2 ml of 1 % (w/v)

starch in PBS, harvested 3 days later, seeded onto glass coverslips in 2 ml of medium in 35 mmPetri dishes (5X105 cells/dish) and cultured for 3 days. The medium was then replaced with1 ml of fresh medium containing 5X104 recipient cells and either 0-01 M-mannose 6-phosphate(fibroblasts) or 1 mg ml"1 yeast mannan (glial cells) and the cells were cultured together in the samedish for a further 24 h. Non-adherent cells were removed by washing the coverslips vigorously with2x1 ml of PBS and the /J-glucuronidase present in the recipient cells was visualized with thespecific anti-(mouse /J-glucuronidase) antiserum using indirect immunofluorescencc (see below).

Indirect immunofluorescenceAfter interaction, cells were fixed in 1 ml of 3-7% formalin in PBS for 8min at room

temperature, then for 5 min in absolute methanol at 4°C. The coverslips were washed with PBSand incubated for 30 min with anti-/3-glucuronidase antiserum diluted 1 in 500 with PBS or pre-immune antiserum as a control, at 37°C in a humidified incubator. Excess antiserum was removedby three washes in PBS and the cells then incubated with a 1 in 32 dilution of FITC-conjugatedgoat (anti-mouse immunoglobulin) antibody for a further 30 min. The coverslips were washedagain, dried in air and mounted in Gelvatol for photography. Donor macrophages were readilydistinguished from accompanying recipient cells by their smaller size and the spherical shape theyassumed during fixation, while recipient fibroblasts and glial cells retained their characteristicmorphology.

Other methods

The activity of /3-glucuronidase preparations was determined using 4-methyl-umbelliferyl /J-D-glucuronide as a substrate (Dean et al. 1982) when 1 unit of activity hydrolysed 1 nmol substrate

140 M.F. Dean, A. McNamara and B. M. Jenne

perh at 37°C. Protein concentrations were determined by the method of Bradford (1976) usingBio-Rad reagent and bovine y-globulin as a standard.

RESULTS

Specificity ofantiserum

The species specificity of the antiserum was determined by comparing its ability toimmunoprecipitate /3-glucuronidase from different mouse and human sources. Ascan be seen in Table 1, 75 % of the /3-glucuronidase purified from CBA mouse liversby chromatography on Concanavalin A-Sepharose and 87 % of that secreted by 3T3mouse fibroblasts was immunoprecipitated, even at a 1 in 1000 dilution of the rabbit(anti-mouse-/3-glucuronidase) antiserum. None of the /3-glucuronidase present inextracts of human placenta was immunoprecipitated under the same conditions evenat a 1 in 100 dilution of the same antiserum.

As a further test of specificity the antiserum was incubated with cultures of rat C6glioma cells, which contain high endogenous levels of/3-glucuronidase. There was nocross-reactivity with rat enzyme, as evidenced by the complete lack of fluorescencefollowing addition of FITC-conjugated second antibody (Fig. 1). The enzymespecificity of the antiserum was tested by measuring its ability to immunoprecipitatetwo other lysosomal enzymes, /3-hexosaminidase and /3-galactosidase, which werepresent in the partially purified preparations of /3-glucuronidase. Neither of these twoenzymes was immunoprecipitated by the antibody, irrespective of whether they hadbeen isolated from mouse or human tissues. Further evidence of specificity wasprovided by double diffusion of the antiserum against partially purified samples ofenzyme from 3T3 fibroblasts, CBA mouse liver or human placenta (Ouchterlony,1961). Only one precipitin line was formed with mouse liver or fibroblast antigensand this single line stained positively for /J-glucuronidase when incubated withAS-BI-/3-D-glucuronide substrate and post-coupled with Fast Blue (Olsen et ai.1982). No precipitin lines formed during double diffusion against human antiserum(data not shown).

Table 1. Specificity of immunoprecipitation of fi-glucuronidase from mouse andhuman tissues

Enzyme activity precipitated (%)

Source of ff-Glucuronidase 0-Hexosaminidase /3-Galactosidaseenzyme A B A A

CBA mouse liver 89 75 0 03T3 mouse 87 87 0 0

fibroblastHuman 0 0 0 0

placenta

Duplicate samples containing 100 units of each enzyme were immunoprecipitated with S. aureusas described in the Materials and Methods and the percentage activity remaining in the supernatantsolutions determined. Columns A, antiserum diluted 1:100; column B, antiserum diluted 1:1000.

Macrophage f}-glucunmidase in other cells 141

Endocytosis of secreted /3-glucuronidase

Previous experiments with cultured peritoneal macrophages have shown that thesecells secreted /3-glucuronidase, which could be taken up into deficient human fibro-blasts via their mannose 6-phosphate receptors (McNamara et al. 1985). The datashown in Table 2 confirmed this finding and furthermore showed that macrophagescontinued to secrete /S-glucuronidase in a high-uptake form when they were co-cultured with fibroblasts. The enzyme they secreted during co-culture was taken upinto fibroblasts at almost three times the rate of that secreted by macrophagescultured alone, suggesting that a high proportion contains ligands that can berecognized by mannose 6-phosphate receptors.

Nevertheless, receptor-mediated uptake of this enzyme was blocked very effec-tively by the addition of mannose 6-phosphate, which reduced uptake of the /S-glucuronidase secreted from co-cultured cells by 98 % and uptake of that secretedfrom macrophages cultured alone by 92% (Table 2).

Specificity of imtnunofluorescence

As shown in the micrograph in Fig. 2A, cultures of mouse 3T3 fibroblasts showed ahigh degree of fluorescence following incubation with anti-(mouse-/3-glucuronidase)

Fig. 1. Rat glioma cells after incubation with species-specific antiserum. The micrographshows the complete absence of fluorescence in a group of C6 glioma cells after incubationwith anti-(mouse /3-glucuronidase) antiserum, which does not cross-react with rat P-glucuronidase. X900.

142 M. F. Dean, A. McNamara and B. M. jfenne

Table 2. Endocytosis of macrophage fi-glucunmidase by GM151 fibroblasts

Source ofmedium

Macrophages culturedalone

Macrophages +mannose 6-phosphate

Macrophage/fibroblastco-culture

Macrophage/fibroblastco-culture-l- mannose6-phosphate

/S-glucuronidasein medium

(units of activity)

21

17

9

23

Rate ofendocytosis

(%/mg cell protein per h)

3-6'

0-29

10-1

019

Medium from macrophages cultured for 24 h alone or co-cultured with fibroblasts was addedto duplicate plates of /J-glucuronidase-deficient fibroblasta with and without 0-01 M-mannose6-phosphate. Uptake of the /3-glucuronidasc secreted by the macrophages was measured asdescribed in the text.

antiserum and FITC-conjugated second antibody, but did not fluoresce afterincubation with pre-immune rabbit serum (Fig. 2B) . In contrast, normal humanfibroblasts showed no positive fluorescence with the specific antiserum (Fig. 2c),confirming that it did not react with /J-glucuronidase from human tissues. After theyhad been co-cultured for 24 h with mouse peritoneal macrophages and mannose6-phosphate, however, the human fibroblasts showed an overall fluorescence coupledwith brighter punctate staining, indicating that they had acquired considerableamounts of mouse /J-glucuronidase (Fig. 2 D ) . Only that enzyme acquired from thedonor cells was recognized by the antibody, as the endogenous enzyme present inhuman fibroblasts did not cross-react under the same conditions. Furthermore,since receptor-mediated endocytosis via receptors recognizing mannose phosphatewas blocked during co-culture by the addition of 0-01 M-mannose 6-phosphate(McNamara et al. 1985), mouse /3-glucuronidase must have been transferred directlyfrom the macrophages to the fibroblasts during co-culture.

Direct transfer to fibroblasts

A range of human fibroblasts deficient in both /3-glucuronidase and other lyso-somal enzymes were tested against the species-specific antiserum but again did notfluoresce. When co-culture experiments were repeated with mouse macrophages asdonor cells and /3-glucuronidase-deficient human fibroblasts as recipients, however,

Fig. 2. Immunofluorescence of normal mouse and human fibroblasts. A. Confluent mouse3T3 fibroblasts after incubation with specific anti-(mouse /3-glucuronidase) antiserum.X225. B. Negative fluorescence of control culture of 3T3 fibroblasts after incubation withpre-immune rabbit serum. X225. C. Negative fluorescence of normal human fibroblastsafter incubation with specific anti-(mouse /J-glucuronidase) antiserum. X225. D. Normalhuman fibroblasts following co-culture with mouse macrophages. The group of brightlystained small round cells in the top right quadrant of the picture are macrophages, thelarger cells fibroblasts. The fibroblast in the centre of the field shows an area of intensefluorescence immediately adjacent to an attached macrophage. X900.

Macrophage P-glucuronidase in other cells 143

144 M. F. Dean, A. McNamara and B. M. Jenne

Macrvphage fi-glucumnidase in other cells 145

direct transfer of enzyme was observed again (Fig. 3A), resulting in a bright punctatefluorescence throughout the recipient fibroblasts. Human fibroblasts deficient inother lysosomal enzymes such as iduronate sulphatase (Fig. 3B), hexosaminidase A(Fig. 3c) and the multiple enzyme-deficient fibroblasts from a patient withMucolipidosis II (Fig. 3D), all showed similar punctate staining on a background ofgeneral fluorescence following interaction in the presence of mannose 6-phosphate.Direct transfer of /3-glucuronidase from macrophages was not, therefore, confined tonormal cells alone but was equally effective when a range of mutant fibroblasta,deficient in a number of different lysosomal enzymes, were used as recipient cells.The nuclei of recipient cells remained unstained (Fig. 3D) while mouse macrophagesassociated with the fibroblasts generally fluoresced too brightly for any details of theirpattern of staining to be resolved (Fig. 3A). Sometimes, however, in less brightlyfluorescent macrophages the /3-glucuronidase could be seen as clumps or patches nearone pole of the cell (Fig. 3B).

Transfer to neuronal and glial cells

In order to determine if direct transfer of /J-glucuronidase was specific to fibro-blasts, or whether other types of cell could act as recipients, mouse macrophageswere co-cultured with human neuroblastoma cells (IMR32) and with rat glioma cells(C6). Preliminary experiments showed that /3-glucuronidase was not taken up intoneuroblastoma cells via mannose phosphate or mannose receptors but was taken upinto glial cells via mannose receptors (Dean & Jenne, unpublished data).

As can be seen in Fig. 4A, the rat glioma cells acquired mouse /3-glucuronidase bydirect transfer from macrophages in the presence of yeast mannan ( lmgmP1) ,which was included in order to block receptor-mediated endocytosis via mannosereceptors. Fluorescent staining can be seen clearly in a punctate pattern within co-cultured glial cells, together with a brighter halo of peripheral staining. Fig. 4B showsthe complete absence of specific fluorescence in another group of glial cells afterincubation with pre-immune serum as a control. The IMR32 neuroblastoma cellsalso became brightly fluorescent following co-culture either with or without mannose6-phosphate. Fluorescence was visible in a foamy pattern around the nuclei, with thenuclei themselves remaining unstained (Fig. 4c). Again, when these neuronal cellswere incubated with pre-immune- serum followed by FITC-conjugated secondantibody they did not fluoresce (Fig. 4D). Macrophages present in the co-culturewere weakly fluorescent because they were able to bind some FITC-conjugatedantibody non-specifically, presumably via their Fc receptors.

Fig. 3. Immunofluorescence of enzyme-deficient human fibroblasts. A. /3-Glucu-ronidafle-deficient fibroblasts. In the centre of the field is a fibroblast showing charac-teristic punctate staining to which a very brightly fluorescing macrophage is still attached.X900. B. An iduronate-sulphate-deficient fibroblast with an attached macrophage andtwo detached macrophages showing patchy staining. X900. C. A hexosaminidase-A-deficient fibroblast with punctate staining surrounded by a group of macrophages. X22S.D. Fibroblasts from a patient with mucolipidosis II, showing diffuse punctatefluorescence surrounding unstained nuclei. Each fibroblast has one or more macrophages,which appear as brighter staining patches still attached. X250.

M. F. Dean, A. McNamara and B. M. Jenne

Macrophage P-glucuronidase in other cells 147

DISCUSSION

The /3-glucuronidase present in mouse cells differs from human enzyme in anumber of its physicochemical characteristics. For example, it can be eluted fromDEAE-cellulose at a higher concentration of sodium chloride during ion-exchangechromatography (Dean et al. 1982) and is readily separated from human enzyme bygel electrophoresis (Chern, 1977). These differences indicate differences in thestructure of these enzymes that are sufficiently great to affect their antigenicdeterminants. Thus, it was possible to produce an antiserum specific for mouse/3-glucuronidase, which did not cross-react with /S-glucuronidase from human orrat tissues. This specificity was demonstrated by the ability of the antiserum toimmunoprecipitate mouse /3-glucuronidase but not human enzyme when incubatedwith S. aureus and by the bright fluorescent staining seen in mouse cells followingincubation with the antiserum followed by FITC-conjugated second antibody.Human fibroblasts and rat glioma cells containing normal levels of /3-glucuronidasedid not fluoresce when tested with the same antibody.

Following 24 h of co-culture with mouse macrophages, however, normal humanfibroblasts had become positively fluorescent, as were the macrophages present in thesame culture (see Fig. 2D). Some of the /S-glucuronidase present in the mousemacrophages must, therefore, have been transferred to the human fibroblasts.

Enzyme transfer could not have occurred as a result of secretion and subsequentreceptor-mediated endocytosis since uptake of enzyme secreted by co-culturedmacrophages takes place via mannose 6-phosphate receptors. These receptors wereblocked during co-culture by the inclusion of mannose 6-phosphate showing thatenzyme acquisition must have occurred as a result of direct transfer during directcell-to-cell contact.

Direct transfer of /S-glucuronidase and other lysosomal enzymes from lymphocytes(Olsen, Dean, Harris & Muir, 1981; Olsen et al. 1982, 1983) and macrophages(McNamara et al. 1985) to enzyme-deficient fibroblasts has been demonstrated. Thepresent results show that direct enzyme transfer is a more general phenomenon sincemouse /S-glucuronidase was also detected in a range of human fibroblasts havingnormal levels of /3-glucuronidase but deficient in other lysosomal enzymes. More-over, direct transfer was not confined to fibroblasts since rat glioma and humanneuroblastoma cells were also able to act as recipients during co-culture withmacrophages. Rat glial cells appear to have receptors characteristic of cells of thereticuloendothelial system that recognize mannose but not mannose 6-phosphateresidues (Dean & Jenne, unpublished data). Although the /S-glucuronidase secreted

Fig. 4. Immunofluorescence of neuronal and glial cells. A. Group of rat glioma cellsshowing punctate fluorescence and bright peripheral halo of staining surrounding eachcell. X900. B. Negative fluorescence of control culture of glioma cells incubated with pre-immune serum. X225. C. Human neuroblastoma cells showing nuclei surrounded byfluorescent cytoplasm with characteristic foamy appearance. A brightly fluorescingmacrophage is still attached to the neuroblastoma cell in the upper right corner of thefield. X900. D. Negative fluorescence of control culture of neuroblastoma cells incubatedwith pre-immune serum. The small rounded weakly fluorescing cells are macrophages.X225.

148 M. F. Dean, A. McNamara and B. M. Jenne

by co-cultured macrophages is recognized by mannose 6-phosphate receptors, yeastmannan was included in all co-cultures of macrophages with glial cells in order toblock receptor-mediated endocytosis, which might have occurred via mannosereceptors. All of the mouse /J-glucuronidase present in the recipient glial cells must,therefore, have been acquired as a result of direct cell transfer. In contrast, humanIMR32 neuroblastoma cells lack both mannose and mannose 6-phosphate receptorsand do not take up /5-glucuronidase by receptor-mediated endocytosis (Dean &Jenne, unpublished data). They are, however, able to acquire enzyme by directtransfer during cell-to-cell contact.

These data suggest that if cells such as macrophages were able to cross theblood/brain barrier they may be able to transfer lysosomal enzymes to deficient braincells in patients with inherited storage diseases. Direct transfer of lysosomal enzymesmay therefore prove to be a widespread phenomenon not confined to specificinteractions between macrophages and fibroblasts. This species-specific antibody canbe used as a probe with which to examine further the specificity and the mechanismof direct enzyme transfer between different types of cell.

The authors thank Action Research for the Crippled Child and the Medical Research Council fortheir financial support.

REFERENCES

BACH, G. & LIEBMANN-EISENBERG, A. (1979). Intracellular localisation of exogenous /3-glucuronidase in cultured skin fibroblasts. Eur.jf. Biochem. 96, 613-619.

BRADFORD, M. (1976). A rapid and sensitive method for the quantitation of microgram quantitiesof protein utilising the principle of protein-dye binding. Analyt. Biochem. 72, 248-254.

CHERN, C. J. (1977). Detection of active heteropolymeric /3-glucuronidase in hybrids betweenmouse and human fibroblasts with /3-glucuronidase deficiency. Proc. natn. Acad. Set. U.SA. 74,2948-2952.

DEAN, M. F. & DIMENT, S. (1984). Exchange of lysosomal enzymes between cells. Biochem. Soc.Trans. 12, 524-526.

DEAN, M. F., DIMENT, S., OSTLOND, C , JENNE, B. M. & CONTRACTOR, S. (1985). Iodinated

fibroblast /3-glucuronidase as a ligand for receptor-mediated endocytosis. Biochem. J. 229,213-219.

DEAN, M. F., OLSEN, I. & MUIR, H. (1982). Identification of rabbit and mouse /J-glucuronidasesin human fibroblasts following direct interaction with lymphocytes. Biockim. biophys. Acta 721,441-448.

DIMENT, S. & DEAN, M. F. (1983). Receptor-mediated endocytosis of fibroblast /3-glucuronidaseby peritoneal macrophages. Biockim. biophys. Acta 762, 165-174.

KAPLAN, A., ACHORD, D. T. & SLY, W. S. (1977). Phosphohexosyl recognition is a generalcharacteristic of pinocytosis of lysosomal glycosidases by human fibroblasts. Proc. natn. Acad.Set. U.SA. 74, 2026-2030.

MCNAMARA, A., JENNE, B. & DEAN, M. F. (1985). Fibroblasts acquire /3-glucuronidase by directand indirect transfer during co-culture with macrophages. Expl Cell Res. (in press).

NEUFELD, D. F., LIM, T. W. & SHAPIRO, L. J. (1975). Inherited disorders of lysosomalmetabolism. A. Rev. Biochem. 44, 357-376.

OLSEN, I., DEAN, M. F., HARRIS, G. & Mum, H. (1981). Direct transfer of a lysosomal enzymefrom lymphoid cells to deficient fibroblasts. Nature, Land. 291, 244-247.

OLSEN, I., DEAN, M. F., MUIR, H. & HARRIS, G. (1982). Acquisition of 0-glucuronidase activityby deficient fibroblasts during direct contact with lymphoid cells. J. Cell Sci. 55, 211-231.

OLSEN, I., Mum, H., SMITH, R., FENSOM, A. & WATT, D. J. (1983). Direct enzyme transfer fromlymphocytes is specific. Nature, hand. 306, 75—77.

Macrophage fi-glucuronidase in other cells 149

OUCHTERLONY, O. (1961). Immunochemical approaches to problems. In Microbiology (ed. M.Heidelberg & O. Plescia), p. 5. New Brunswick, NJ: Rutgers University Press.

SHEPHERD, V. L., LEE, Y. C , SCHLESINGER, P. H. & STAHL, P. D. (1981). L-Fucose-terminatedglycoconjugates are recognised by pinocytosis receptors on macrophages. Proc. natn. Acad. Sci.U.SA. 78, 1019-1022.

STAHL, P., RODMAN, J. S., MILLER, J. & SCHLESINGER, P. (1978). Evidence for receptor-mediatedbinding of glycoproteins, glycoconjugates and lysosomal glycosidases by alveolar macrophages.Proc. natn. Acad. Sci. U.SA. 75, 1399-1403.

VON FIGURA, K. & KRESSE, H. (1974). Quantitative aspects of pinocytosis and the intracellular fateof iV-acetyl-ar-D-glucosaminidase in Sanfilippo fibroblasts. J. din. Invest. 53, 85-90.

(Received 29 April 1985 - Accepted 7 June 1985)