The effects of ?-2-microglobulin on the infectivity of murine cytomegalovirus

Arch Virol (1992) 123:59 72

_Archives

Vi rology © Springer-Verlag 1992 Printed in Austria

The effects of p-2-microglobulin on the infectivity of murine

cytomegalovirus

Michelle N. Wykes, P. Price, and G. R. Shellam

Department of Microbiology, University of Western Australia, Nedlands, Australia

Accepted August 12, 1991

Summary. The role of [3-2-microglobulin (j32m) in murine cytomegalovirus (MCMV) infection of susceptible (H-2 d) and resistant (H-2 k) murine embryo fibroblasts (MEF) and peritoneal macrophages was evaluated using serum-free virus (SF-MCMV). The infectivity of SF-MCMV was significantly lower than virus propagated in serum, although the concentrations of virions were similar. Infection of cells with SF-MCMV was assessed by measuring the proportion of cells expressing viral antigens, the sizes of plaques formed in fibroblast monolayers and TCIDs0 titers. Infection of susceptible fibroblasts was significantly increased 1.64.7 fold by the addition of whole FCS, a < 20 kDa FCS fraction, or purified human [32m. These supplements also significantly enhanced infection of susceptible macrophages and increased TCIDs0 titers by 3.5-10 fold in susceptible MEF. In relatively resistant H-2 k cells, the TCIDs0 titer and the proportion of cells expressing viral antigens after infection with SF-MCMV were not affected by 132m or FCS, but plaque sizes were increased 2.5-3 fold in resistant BALB.K MEF.

When human or murine 132m was added to infected cultures, immunogold electron microscopy revealed these proteins to be always associated extracellularly with the tegument material of disrupted multicapsid virions, but rarely with the envelope of intact virions. However, no murine ~2m was found in association with the envelope or tegument of SF-MCMV. These relatively modest effects of j32m which were restricted to genetically susceptible cells, may be due to tegument-bound 132m facilitating infection by capsids, or the stabi- lisation of the conformation of Class 1 molecules by exogenous J32m, promoting MCMV binding to the target cell.

Introduction

Murine cytomegalovirus (MCMV) has been used extensively as a model for human CMV (HCMV) disease as their virion structures and the resultant disease

60 Michelle N. Wykes et al.

syndromes are similar [10]. Both viruses infect cells in most major organs [6] and replicate in vitro in cultured fibroblasts, endothelial cells and to a more limited extent in monocytes [11, 18, 20].

An association between HCMV and [32m was first postulated by McKeating et al. [15, 16] after demonstrating that the protein inhibited detection of the virus by certain monoclonal antibodies and co-purified with the virus from the urine of HCMV-infected patients. Further studies showed extracellular bovine or human p2m enhanced HCMV infection, which suggested that infection was initiated by the binding of virion-bound 132m to the heavy chains of Class 1 major histocompatibility complex (MHC) molecules on the cell surface [9]. Immunoprecipitation studies suggested that the viral envelope bound to 132m after release from infected fibroblasts and not during the budding process [8-]. Stannard [21] detected virion-associated ~2m using immunogold staining and electron microscopy, but suggested that the protein may be localised in the tegument material rather than on the envelope of extracellular HCMV. How- ever, Beersma et al. [3] were not able to detect 132m either on HCMV proteins by Western blotting or on virion envelopes by electron microscopy.

We have previously shown that the sensitivity of cultured murine macrophages or embryo fibroblasts (MEF) to MCMV is determined by their H-2 genotype, where H-2 d cells are sensitive and H-2 k cells are resistant [11, 17, 18]. This correlates with the resistance of the donor strains to MCMV infection in vivo [4]. In addition, MCMV infection of MEF has a reversible temperature- insensitive and an irreversible temperature-sensitive phase [12]. If we assume the latter represents endocytosis of the virus, it would seem plausible that the affinity of binding of virion-bound 132m to different H-2 molecules may determine their relative susceptibility.

Our present study investigates the effect of extracellular 132m on infection by MCMV of cells of differing H-2 associated susceptibility, and uses immunogold electron microscopy to determine if [32m binds to the virion. Infection of MEF and 3T3 cells was assessed after 6 h as a measure of initial infection, and after 48 and 120 h to assess virus release and secondary spread of infection within cultures. Initial infection ofmacrophages was assessed after 48 h, as these cells have a longer lag phase before the expression of viral antigens. Congenic strains were used to control for non-H-2 effects [7].

Materials and methods

Animals

BALB/c (H-2d), BALB.K (H-2k), C3H/HeJ (H-2k), and Swiss (non-inbred) mice were obtained from the Animal Resources Centre (Murdoch, Western Australia) as advanced pregnant females for the preparation of embryo fibroblasts, or at 8 weeks of age as donors of peritoneal macrophages. BALB/c and Swiss mice are genetically susceptible to MCMV infection, while BALB.K and C3H/HeJ are genetically resistant [4, 18].

13-2-microglobulin and MCMV 61

Cell culture

MEF were obtained from 15-17 day old embryos by trypsin dispersion, cultured in Eagles minimal essential medium (MEM; Gibco, USA) with antibiotics and 10% fetal calf serum (FCS; Commonwealth Serum Laboratories, Australia), dislodged from the flasks by tryp- sinization, aliquoted and stored in the gas phase of liquid nitrogen. Thawed cells were cultured in MEM with 10% FCS or 2% Ultroser G (UG; IBF, France). BALB/c-derived 3T3 cells of a low passage number were grown in Dulbecco's MEM with 2% UG.

Production of '~erum-free" MCMV

MEF grown in MEM with 2% UG were infected with 2 x 106 plaque forming units (pfu) of salivary gland derived MCMV [2] using centrifugal enhancement. Briefly, MCMV was added to flasks of Swiss MEF, which were centrifuged (700 g, 30 rain) and incubated at 37 °C for 90 rain. The cells were then cultured in 2% UG and serum-free (SF)-MCMV was harvested after 3 days. The SF-MCMV used in this study contained 5 x 105 pfu/ml when titrated on Swiss MEF after dilution in MEM with 2% FCS [2]. Titers quoted in the text as pfu/ml are derived from this assay.

Supplementation of UG cultures

FCS was separated into fractions with molecular weights greater than 20 kDa (> 20kD FCS) and less than 20 kDa (< 20kD FCS), using Sartorius-Centrisart tubes (Germany) according to manufacturers instructions. The filters provided accurate separation of serum proteins as assessed by gel electrophoresis. Serum fractions reconstituted to the original volume and whole FCS were added at 5% to MEM, with 2% UG to ensure adequate levels of all nutrients, and human [32m (Sigma, USA) was used at 5 gg/ml in MEM with 2% UG, since preliminary experiments established that higher concentrations of any of these supplements did not increase levels of infection further.

Infection of MEF and 373 cells

Cells were grown to confluency, trypsinized and seeded onto sterile coverslips in 24 well culture plates at 2.5 × 105 cells per well in MEM with 2% UG. After 24h, confluent monolayers were treated with SF-MCMV diluted 10-1 and 10 - 3 in MEM containing whole FCS, UG alone, or UG plus fractionated FCS or ~32m for 1 h at 37 °C. The viral inoculum was then removed and replaced with fresh medium containing the same supplements for a 48 h incubation or UG alone for a 6 h incubation. After 6 to 48 h, coverslips were fixed in cold 85% acetone (7 min), dried and stored at -70 °C.

TCIDso assay

BALB/c and BALB.K MEF grown to confluency were trypsinized and seeded in 96 well plates at 1 × 105 cells per well in MEM supplemented with 2% UG. After 24 h, cells were infected with 0.2 ml SF-MCMV serially diluted 10- 2 to 10-9, in MEM plus the supplements as described above. The viral inoculum was removed after 1 h at 37 °C and replaced with fresh medium containing the same supplements. After 120h the monolayers were fixed with 0.05% methylene blue in 10% formalin. TCIDs0 titer of MCMV infectivity was calculated by applying the Kaeber equation to 12 replicates,

Collection and infection of macrophages

Peritoneal cells pooled from 4-8 mice were resuspended in N-2-hydroxyethylpiperazine-N- 2-ethane-sulphonic acid (HEPES)-buffered RPMI 1640, dispensed into 24 well plates con-


taining sterile coverslips and infected with SF-MCMV using centrifugal enhancement [17], with the supplements described above. The inoculum was replaced with HEPES-RPMI containing 0,5% normal mouse serum and 4 x 10- 5 M 2-mercaptoethanol. After 24 h non- adherent cells were dislodged by gentle pipetting, aspirated and fresh medium was added to the adherent cells. Cells were fixed with cold acetone after 48 h.

Enumeration of infected cells at 6 and 48 h

Cells expressing viral antigen were visualised using hyperimmune serum from BALB/c mice infected with MCMV. Bound antibody was detected with biotin-avidin immunoperoxidase reagents (Amersham, UK). Mean numbers of infected cells per x 40 field and the sizes of infected foci in mm 2 (assessed with an eye piece graticule) were determined over at least 20 fields in several separate experiments and standard errors were calculated. Results obtained within each experiment were compared using Student's t-test and representative studies were selected for presentation. Evaluation of infection of macrophages was based on 300-700 cells per slide and subjected to ~2 analysis.

Enumeration of virions using electron microscopy

Supernatants were collected from MCMV-infected MEF cultures, aliquoted and stored at - 70 °C. Aliquots were centrifuged at 1200 g for 20 rain to reduce cell debris, and triplicate 50 gl volumes were spun onto copper grids coated with Fomvar (Sigma, USA), at 22 psi in a Beckman airfuge. Grids were stained for 2 rain with 4% sodium silica tungstate (pH 6.9) and dried overnight [5, 13]. Virus particles comprising multiple capsids were counted over at least ninety ( x 4500) fields per sample to obtain a mean and standard error. Student's t-test was applied to assess differences between samples.

ImmunogoM staining of virus for electron microscopy

MEF grown to confluency were infected with MCMV in 5% FCS for t h at 37 °C. The viral inoculum was replaced with 5% fetal human serum, 5% mouse serum, 2% UG or 2% UG with 5 ~g/ml human f32m. Infection was allowed to proceed until 80% of the monolayer showed cytopathic changes. The supernatant was clarified by centrifugation at 1000 g, overlaid onto a sucrose cushion and re-centrifuged at 12 000 g for 2 h. The surface of the sucrose phase was collected and stored at - 70 °C until required. Electron microscopy grids were coated with virus as described above. To detect human 132m, grids were transferred onto drops of an immunoglobulin fraction of chicken anti-human ~2m (provided by Dr. A. Sanderson, London, UK) for 1 h, then washed through ten changes of tris-buffered saline and placed on drops of rabbit anti-chicken immunoglobulin (Cappel, UK). This was followed by protein A-colloidal gold conjugate with t2 nm gold particles (Amersham, UK). Mouse [32m was detected with an immunoglobulin fraction of rabbit anti-mouse 132m serum (provided by Dr. A. Lew, Melbourne, Australia) followed by protein A-colloidal gold conjugate. Control grids were treated with non-immune chicken or rabbit serum as ap- propriate. Finally, all grids were negatively stained as described above.

Results

Effect o f serum on the quantity and infectivity o f extracetlular virus released from infected fibroblasts

Swiss M E F grown in 2% U G were infected with 1 x 105pfu S F - M C M V using centrifugal enhancement . The duplicate monolayers were incubated in 5% FCS or 2% U G following infection, and aliquots o f the superna tan t media were

~-2-microglobulin and MCMV 63

collected on days 2, 3 and 5. From these, virus contained in 50 t~1 volumes was centrifuged onto grids for electron microscopy. Very few virions were detected on days 2 and 3, however supernatants collected on day 5 from cultures maintained with UG or FCS contained 10.7 ± 0.4 or 9.8 ± 0.4 virions per x 4500 field respectively (p > 0.05). In addition no difference in the amount of viral antigen in the supernatant of these cultures was detected by complement fixation assay (data not shown). These findings suggest that FCS did not affect the number of virions released from infected cells.

The infectivity of these two virus preparations was investigated by infection of confluent coverslip cultures of Swiss MEF monolayers previously grown in 2% UG. The virus preparations were diluted 1 in 10 to assess expression of viral antigen and 1 in 1000 to assess plaque sizes, in either 5% FCS or 2% UG to correspond with the medium used during the production of these virus preparations. After infection, cultures were incubated in the same media. There were significantly higher numbers of cells (15.9 ± 0.7 cells / x 40 field) expressing viral antigen in cultures infected with virus grown and maintained in FCS compared to cultures treated with SF-MCMV and maintained in the absence of serum (11.1 ± 1.2 cells / x 40field; p < 0.01). Furthermore, foci formed in cultures containing FCS averaged 0.43 ± 0.03mm 2 in size compared with 0.18 ± 0.04mm 2 in serum-free cultures (p < 0.05). These results indicate that the infectivity of MCMV propagated and assessed in FCS in higher than the infectivity of SF-MCMV measured in the absence of serum, although the concentration of virion particles was equal in both preparations. This suggested a factor in serum moderately enhanced viral infectivity. Subsequent experiments were designed to identify the factor(s) responsible and to determine whether the effect was greater under other conditions of culture.

Effect of human fl2m or FCS on M C M V infection of MEF and 3T3 cells

Viral antigen expression

BALB/c, BALB.K and C3H/HeJ MEF, and 3T3 cells were grown to confluency in MEM plus 2% UG on coverslips and infected with SF-MCMV at 1.7 x 102 or 1.7 4- 104pfu, supplemented with either 5% FCS, 2% UG with 5% < 20kD FCS, 2% UG with 5% > 20kD FCS or 2% UG with 5 gg/ml purified human 132m. Viral antigen expression was assessed after 6 and 48 h.

Cells expressing MCMV antigens were counted after 6 h (Table 1). BALB/ c MEF and 3T3 cultures displayed higher numbers of infected cells per field when virus was supplemented with whole FCS (t-test p < 0.025), < 20kD FCS (p < 0.001) or purified ~32m (p < 0.005), compared to infection in UG alone or with > 20kD FCS. The number of BALB.K and C3H/HeJ cells infected were low and not affected significantly by FCS or J32m.

After 48 h, foci of infected cells had developed in most of the monolayers. Infected cells were generally rounded in FCS and J32m preparations, and cells in the centre of the foci had lifted off, forming plaques. However, loci formed


Table 1. Effect of FCS and 132m on frequencies of expression of viral antigens ~ after 6 h in MEF and 3T3 cells treated with SF-MCMV

Cell type None > 20kD FCS b < 20kD F C S b Whole FCS b human 132m u

3T3 7.4±0.1 4.3±0.3 31.3i0.5 ° 24.2+0.3 ~ 26.0±0.6 ° BALB/c MEF 7.9 ~0.5 7.34-0.5 13.0± 1.0 ° 14.14-0.9 ~ 16.2±0.9 c BALB.K MEF 1.8±0.3 1.6±0.3 2.3:t:0.3 2.0±0.8 1.3±0.3 C3H/HeJ MEF 2.0±0.2 1.6±0.2 2.34-0.3 2.6±0.4 1.3±0.2

a Mean number of cells (+ S.E.M.) expressing immediate early antigens per x 40 HPF after 6 h, counted over 20-100 fields in 2-4 experiments

b Supplements used during 1 h co-incubation with the virus (see Materials and methods) °Significant potentiation of infection relative to unsupplemented cultures (t-test,

p < 0.05)

in the presence of only U G were small and cells retained their normal mor- pho logy (Fig. 1). Consis tent with the effect o f supplements on the f requency of infected cells, the sizes of foci in B A L B / c M E F and 3T3 monolayers were also increased (2-4 fold larger) in the presence o f whole FCS (p < 0.0005), < 20kD FCS (p < 0.025) or purif ied 132m (p < 0.0005), compared with foci fo rmed in the presence o f U G alone or > 20kD F C S (Table 2). The sizes o f foci in B A L B . K M E F were also increased by the addi t ion o f < 20kD FCS, whole F C S or h u m a n ~2m (p < 0.05) relative to unsupp lemented cultures, bu t foci in C 3 H / H e J M E F

were too small to detect any effect o f serum supplements . C 3 H / H e J and Swiss endothel ial cells were also infected as described above.

Infect ion of susceptible Swiss cells was increased by supplementa t ion with < 20kD FCS, FCS and purif ied 132m, bu t no effect was seen with C 3 H / H e J endothel ial cells (data not shown). This is consis tent with results ob ta ined with f ibroblasts f rom these strains.

Table 2. Effect of FCS and 132m on sizes of plaques a generated after 48 h in MEF and 3T3 cells infected with SF-MCMV

Cell type None > 20kD F C S b < 20kD F C S b Whole FCS b h u m a n ~2m b

3T3 0.194-0.03 <0.01 0.33±0.06 c 0.33+0.08 ° 0.75:t:0.10 ° BALB/c MEF 0.06±0.01 0.06+0.01 0.25±0.02 ° 0.28+0.02 c 0.25±0.03 c BALB.K MEF 0.02±0.00 0.02±0.00 0.064-0.01 ° 0.05±0.01 c 0.06±0.01 c C3H/HeJ MEF 0.01 ± 0.00 0.0i 4- 0.00 0.01 4- 0.00 0.01 ± 0.00 0.0I + 0.00

Mean sizes of plaques in mm 2 after 48 h. At least 20 plaques were measured in 2.4 experiments

b Supplements used during 1 h co-incubation with the virus and for 48 h post infection (see Material and methods)

°Significant potentiation of infection relative to unsupplemented cultures (t-test, p < 0.05)

[3-2-microglobulin and MCMV 65

Fig. 1. Representative plaques formed 48 h pi in 3T3 cell monolayers, infected with SF- MCMV, supplemented with a no supplement, b whole FCS or e human ]32m. Infected cells were detected using hyperimmune antibody to MCMV and immunoperoxidase staining. Sizes of foci approximate to the mean values for 3T3 cells described for the various supplements in Table 2. Plaques were small in the absence of serum supplements (a) and

increased in size by the addition of whole FCS (b) or human 132m (e). x 100


Table 3. Effect of FCS or 132m on TCIDs0 titers in BALB/c and BALB.K MEF infected with SF-MCMV

Supplements a - loglo TCIDso/0.2 ml

BALB/c ° BALB.K c BALB/c

None 6.20 4.26 5.00 > 20kD FCS fraction - - 5.08 < 20kD FCS fi'action - - 5.55 FCS 7.22 b 4.12 5.58 Human 132 - - - 6 . 0 0 b

a Supplements used during 1 h co-incubation with virus and for 120 h post infection (see Material and methods)

b Significant increase in infection relative to unsupplemented cultures, see text

c Results of an experiment in which cells of both strains were infected simultaneously under identical condition

TCIDso assay

BALB/c and BALB.K M E F cultured in 2% U G were infected simultaneously with serial dilutions of S F - M C M V in 2% U G or 5% FCS. The results shown in Table 3 indicate that FCS had no effect on the TCIDs0 titer in BALB.K MEF, while it increased the titer 10.5 fold in BALB/c M E F (p < 0.005). To analyse the effect of FCS further, BALB/c M E F were cultured in 2% U G and treated simultaneously with serial dilutions of a separate stock of S F - M C M V in either 2% U G or 5% whole FCS, 5% < 20kD FCS, 5% > 20kD FCS, or 5 gg/ml of purified h u m a n 132m. TCIDs0 titers in BALB/c M E F were 3.5 fold, 3.8 fold or 10 fold (p < 0.05) higher in the presence of < 20 kD FCS, whole FCS or h u m a n [32m respectively, compared to U G alone (Table 3). These experiments were repeated several times for < 20kD FCS, > 20kD FCS, and FCS, and increases in TCIDs0 titer were always observed when < 20kD FCS or FCS (7-10 fold; p < 0.05) were added.

Effect of human fi2m and FCS on MCMV infection of macrophages

Peritoneal macrophages f rom BALB/c and BALB.K mice were infected with 2.5 x 104 or 5 x 104pfu S F - M C M V (Table4). The virus was diluted in M E M with 2% U G alone or with whole or fract ionated FCS, or h u m a n 132m. After 48 h, viral antigen-positive cells were enlarged and rounded, and occurred at higher frequencies in preparat ions f rom BALB/c than f rom BALB.K mice. At the lower concentra t ion of virus the frequency of infection of BALB/c macrophages was moderately increased by whole FCS, < 20kD FCS or h u m a n 132m (;(2, p < 0.05) relative to cultures mainta ined in U G alone or with > 20kD FCS. In contrast, the p ropor t ion of BALB.K macrophages infected with


Table 4. Effect of FCS or 132m on frequencies (%) of viral antigen expression a in BALB/c and BALB.K macrophages 48 h after infection with SF-MCMV

Supplements BALB/c macrophages BALB.K macrophages

2.5 x 104b 5 × 105b 2.5 × 104b 5.10 x 104b pfu pfu pfu pfu

None 72 90 3 8 > 20kD FCS 67 NT NT 3 < 20kD FCS 87 ° NT NT 3 Whole FCS 91 c 96 2 4 Human [32m 90 c NT NT 5

a Percentage of cells expressing viral antigens after 48 h (300-700 cells counted) b Final concentration of virus in 0.5 ml ° Significant increase in infection relative to unsupplemented cultures, (;~2, p < 0.05) NT Supplements not tested

M C M V was low and not apparently increased by FCS or 132m irrespective of virus concentration.

Immunogold staining to identify virion-associated fl2m

M C M V was grown in cells cultured in either 5% fetal human serum, 2% U G with 5 gg/ml human 132m, 5% mouse serum or 2% U G alone and treated with antibody to human or murine 132m, followed by an immunogold conjugate to detect the protein using electron microscopy. Human 132m was used to obtain evidence that 132m could associate with the virus extracellularly, since any mouse 132m detected in association with virions from cells cultured in mouse serum could have been derived endogenously. Virions consisted of packages of 2-20 capsids surrounded by a dense, amorphous tegument material and an outer envelope, although in some cases the envelope was ruptured or absent. Mul- ticapsid virions are a feature of MCMV. Virus treated with non-immune chicken or rabbit sera did not bind colloidal gold, shown here for control chicken serum (Fig. 2 a), whereas antibodies to f32m appeared to localise colloidal gold particles with the virus. In those virions in which the envelope was either absent (Fig. 2 b) or ruptured (Fig. 2 c), the gold particles appeared to be concentrated in the tegument material surrounding the capsids. However, intact virions were not usually associated with 132m (Fig. 2 d), and 132m binding to viral envelopes was only rarely observed. When SF-MCMV obtained from cells cultured in 2% U G was examined using rabbit anti-mouse 132m antibody, no significant colloidal gold binding was detected with ruptured (Fig. 2 e) or intact particles (not shown), indicating an absence of association of [32m with these virions.

Discussion

This report analyses the role of serum factors and human 132m in the infection of cells with MCMV. Our data shown that FCS added to the culture media


Fig. 2. Detection of ]32m associated with multicapsid virions of MCMV from cells grown in human (a, b) or murine (e, d) sera or UG (e) treated with normal chicken serum (a), chicken anti-human ~2m serum (b), or rabbit anti-mouse 132m (e, d, e) followed by immunogold reagents. Bar: 50 nm

did not affect the virus output from cultured cells but did enhance the infectivity of extracellular virus. In this study, the proport ion of cells infected by the addition of a standardised viral inoculum after 6 h and the spread of infection within cultures after 48-120 h were assessed independently. The addition of whole FCS enhanced initial infection of SF-MCMV and the spread of the virus in susceptible BALB/c M E F and 3T3 cells. Initial infection of BALB/c macrophages was also enhanced. The factor(s) in FCS which potentiated infection had molecular weight(s) less than 20kDa and at a concentration of 5%, was


as efficient as 5 Ixg/ml of human j32m. In contrast, infection of BALB.K and C3H/HeJ MEF and BALB.K macrophages, when assessed for early viral antigen expression, was not significantly affected by FCS or 132m. While this result suggests that 132m only potentiates the initial infection of cells of susceptible H-2 d strains, it should be interpreted with caution. The difficulty of achieving equivalent levels of infection in cells of resistant H-2 k strains limits comparison of the effect of 132m on cells of both donor types. However, the spread of infection in BALB.K MEF was increased by the addition of serum supplements, and may be due to the amplification of a very small effect of ~2m on the initial infection of BALB.K MEF. Using a TCIDs0 assay, FCS increased the virus titer 10.5 fold in BALB/c MEF while not influencing the titer in BALB.K MEF. This strongly supports the notion that 132m exerts a selective effect on MCMV infection of H-2 d cells, and may reflect a greater affinity of binding of ~2m or virion-~2m complexes to H-2 a than H-2 k Class 1 molecules. A 3.5-3.8 fold increase in TCIDs0 titer was seen in susceptible BALB/c MEF following the addition of a < 20 kDa fraction of FCS or whole FCS to the serum-free viral inoculum and cell cultures. This effect was less than that reported with HCMV [9]. We observed a 10 fold increase in TCIDs0 titer when human 132m was used with BALB/c cells.

However, despite the reproducible and significant effects of ~2m or serum factors on MCMV infection, it is clear that a substantial level of infection occurs in the absence of these supplements, as exemplified by the data in Tables 3 and 4. Thus the effects of exogenous 132m or serum factors in potentiating infection must be viewed as relatively modest. The infection of cells in cultures deprived of 132m or serum factors therefore appears to occur by a non-132m-associated mechanism. The recently described 30 kDa and 92.5 kDa cells membrane proteins which are apparently unrelated to 132m or HLA Class 1 molecules, have been shown to bind HCMV [1, 14, 24], but evidence that they initiate uptake into the cell is lacking. We chose not to examine the effects of [32m or serum factors on the binding of MCMV to cells because of the likelihood that not all virus binding leads to infection, and because the lack of agreement about the level of binding of HCMV to Daudi and Raji cells [9, 23] indicates difficulties with binding assays involving CMV. While our results suggest the presence of two mechanisms of entry for MCMV, one [32m-associated and a second unrelated mechanism, caution is required in interpreting the data because the ubiquity of J32m and its production by cultured cells limits attempts to study the role of exogenous [32m in the infectious process. Indeed, the relatively modest effects of exogenous 132m or serum factors reported here may be a reflection of this problem.

The mechanism by which J32m potentiates infection was earlier postulated to involve the extracellular binding of 132m to virions, and the displacement of native ~2m from Class 1 molecules of the major histocompatibility complex by virion-132m complexes [9], triggering receptor mediated endocytosis. However, evidence obtained at the ultrastructural level in the present investigation does


not suport this view. We rarely detected 132m on the envelopes of intact virions of MCMV, confirming the observations of Stannard [21] and Beersma I-3] with HCMV. Thus, despite the limitations of ultrastructurat studies, the facil- itation of CMV infection by 132m is unlikely to involve envelope-associated binding of 132m to intact virions. However we did observe substantial binding of 132m to the tegument material surrounding the capsids of disrupted virions. Stannard compared the location of viral glycoproteins and [32m within HCMV virions and also reported that [32m lay within the tegument. However chro- matographic studies employing fractionated HCMV preparations have shown that 132m separates with the envelope proteins [8]. It has been reported that the 132m binding proteins of the tegument also bind the Fc portion of immunoglobulin G [22], suggesting a role for tegument proteins both in the infectious process and in immune evasion.

There are several possible explanations which suggest a role for 132m in the tegument material of virions, and for the modest effect of this protein on the infectivity of SF-MCMV. The first is that the binding of [32m to tegument proteins in ruptured envelopes facilitates the attachment of those capsids which are encased in tegument material to either H-2 Class 1 molecules or to other possible [32m-binding proteins on the cell surface. Alternatively, the infection- promoting effect of [32m may be due to the need for exogenous 132m to maintain the configuration of Class 1 molecules at the cell surface, thus permitting virion binding to these molecules. In this event, the greater affinity of binding of 132m to H-2 d than H - 2 k Class 1 molecules [19] may be associated with greater con- figurational stability of H - 2 d Class 1 molecules. Both mechanisms involve a role for Class 1 molecules which we believe is justified by the available evidence. Our laboratory has previously established that H-2 molecules influence infection with MCMV in vivo and in vitro [4, 17, 18] and that susceptibility is linked to the expression of H-2D d molecules [17, 183. Similarly, Grundy etal. have suggested a role for Class t molecules in HCMV infection [9], and Beersma et al. [3] have demonstrated in transfection studies that HLA B27 may uniquely confer susceptibility of HCMV infection. Finally, our preliminary evidence shows that Class 1 negative cells are relatively resistant to MCMV and that this defect can be repaired by the transfection of H-2 Class 1 genes. The two mechanisms proposed for the potentiating effect of 132m are not mutually exclusive and could complement each other. Further experiments are required to elucidate the role of f32m in CMV infection.

Acknowledgements

The authors thank Professor J. M. Papadimitrou, Mr. T. Robertson and Dr. G. B. Harnett for assistance with the electron microscopy, Dr. A. Lew (Melbourne, Australia) and Dr. A. Sanderson (London, UK) for antibodies to murine and chicken 132m, respectively, Professor J. B. Hudson for his discussion of the work, and Dr. L. Stannard for permission to quote her work before publication. P. Price is supported by the National Health and Medical Research Council of Australia.


References

1. Adlish JD, Lahijani RS, St Jeor SC (1990) Identification of a putative cell receptor for human cytomegalovirus. Virology 176:337-345

2. Allen JE, Shellam GR (1984) Genetic control of murine cytomegalovirus infection: virus titers in resistant and susceptible strains of mice. Arch Virol 81:139-150

3. Beersma MFC, Wertheim-van Dillen PME, Feltkamp TEW (1990). The influence of HLA-B27 on infectivity of cytomegalovirus for mouse fibroblasts. Scand J Rheumatol [Suppl] 87:102-t03

4. Chalmer JE, Mackenzie JS, Stanley NF (1981) Influence of H-2 and non H-2 genes on resistance to murine cytomegalovirus infection. Infect Immun 32: 2770. 2860

5. Derrick KS (1973) Quantitative assay for pant viruses using serologically specific electron microscopy. Virology 56:652-653

6. Griffiths PD, Grundy JE (1988) The status of CMV as a human pathogen. Epidemiol Infect I00:1-15

7. Grundy (Chalmer)JE, Trapman J, Allen JE, Shellam GR, MeliefCJM (1982) Evidence for a protective role of interferon in resistance to murine cytomegalovirus and its control by non-H-2 linked genes. Infect Immun 37:143-150

8. Grundy JE, McKeating JA, Griffiths PD (1987) Cytomegalovirus strain AD169 binds [32 microglobulin in vitro after release from cell. J Gen Virol 68:777-784

9. Grundy JE, McKeating JA, Ward PJ, Sanderson AR, Griffiths PD (1987) [~2 microglobulin enhances the infectivity of cytomegalovirus and when bound to the virus enables class 1 HLA-molecules to be used as a virus receptor. J Gen Virol 68:793-803

10. Hamilton JD (1982) Cytomegalovirus and immunity. Monogr Virol 12:1-108 11. Harnett GB, Shellam GR (1982) Variation in murine cytomegalovirus replication in

fibroblasts from different mouse strains in vitro: correlation with in vivo resistance. J Gen Virol 62:3947

12. Hodgkin PD, Scalzo AA, Swaminathan N, Price P, Shellam GR (1988) Murine cytomegalovirus binds reversibly to mouse embryo fibroblasts: implications for quanti- tation and an explanation of centrifugal enhancement. J Virol Methods 22:215-230

13. Janson J, Harnett GB, Bucens MR (1985) Electron microscopy after direct ultracen- trifugation. Pathology 17:29-30

14. Keay S, Merrigan TC, Rasmussen L (1989) Identification of cell surface receptors for the 86-kilodalton glycoprotein of human cytomegalovirus. Proc Natl Acad Sci USA 86:10100-10103

15. McKeating JA, Grundy JE, Vargese Z, Griffiths PD, Grundy JE (1986) Detection of cytomegalovirus by ELISA in urine samples is inhibited by 13-2-microglobulin. J Med Virol 18:341-348

16. McKeating JA, Griffiths PD, Grundy JE (1987) Cytomegalovirus in urine specimens has host 132 microglobulin bound to the viral envelope: a mechanism of evading the host immune response? J Gen Virol 68:785-792

17. Price P, Winter JG, Nikoletti S, Hudson JB, Shellam GR (1987) Functional changes in murine macrophages infected with cytomegalovirus relate to H-2 determined sensitivity to infection. J Virol 61:3602-3606

18. Price P, Gibbons AE, Shellam GR (1990) H-2 Class 1 loci determine sensitivity to MCMV in macrophages and fibroblasts. Immunogenetics 32:20-26

19. Schmidt W, Festenstein H, Ward PJ, Sanderson AR (1981) Interspecies exchange of 132-microglobulin and associated MHC and differential antigens. Immunogenetics 13: 483-491

20. Smiley LM, Mar EC, Huang CS (1988) Cytomegalovirus infections and viral induced transformation of human endothelial cells. J Med Virol 25:213-226

72 Michelle N. Wykes et al.: [3-2-microglobulin and MCMV

21. Stannard L (1989) [32-microglobulin binds to the tegument of cytomegalovirus: an immunogold study. J Gen Virol 70:2179-2184

22. Stannard LM, Hardie DR (1991) An Fc receptor for human immunoglobulin G is located within the tegument of human cytomegalovirus. J Virol 65:3411-3415

23. Taylor HP, Cooper NR (1989) Human cytomegalovirus binding to fibroblasts is receptor mediated. J Virol 63:3991-3998

24. Taylor HP, Cooper NR (1990) The human cytomegalovirus receptor on fibroblasts is a 30 kilodalton membrane protein. J Virol 64:2484-2490

Authors' address: Michelle N. Wykes, Department of Microbiology, University of Western Australia, Nedlands, W.A. 6009, Australia.

Received June 19, 1991

Documents

The effects of ?-2-microglobulin on the infectivity of murine cytomegalovirus