Virus–receptor interactions of human parainfluenza viruses types 1, 2 and 3

Microbial Pathogenesis 1999; 27: 329–336

Article available online at http://www.idealibrary.com on

Article No. 1999.0313

MICROBIAL PATHOGENESIS

Virus–receptor interactions of human parainfluenza viruses types 1, 2 and 3 C. Ah-Tye, S. Schwartz, K. Huberman, E. Carlin & A. Moscona∗ Departments of Pediatrics and Cell Biology/Anatomy, Mount Sinai School of Medicine, 1 Gustave L. Levy Place, New York, NY 10029-6574, U.S.A. (Received March 19, 1999; accepted in revised form June 10, 1999)

Human parainfluenza viruses types 1, 2 and 3 (HPF 1, 2 and 3) are important pathogens in children. While these viruses share common structures and replication strategies, they target different parts of the respiratory tract; the most common outcomes of infection with HPF3 are bronchiolitis and pneumonia, while HPF 1 and 2 are associated with croup. While the HPF3 fusion protein (F) is critical for membrane fusion, our previous work revealed that the receptor binding hemagglutinin– neuraminidase (HN) is also essential to the fusion process; interaction between HN and its sialic acid-containing receptor on cell surfaces is required for HPF3 mediated cell fusion. Using our understanding of HPF3 HN’s functions in the cell-binding and viral entry process, we are investigating the ways in which these processes differ in HPF 1 and 2, in part by manipulating receptor availability. Three experimental treatments were used to compare the HN–receptor interaction of HPF 1, 2 and 3: infection at high multiplicity of infection (m.o.i.); bacterial neuraminidase treatment of cells infected at low m.o.i.; and viral neuraminidase treatment of cells infected at low m.o.i. (using Newcastle disease virus [NDV] neuraminidase or UV irradiated HPF3 as sources of neuraminidase). In cells infected with HPF3, we have shown that infection with high m.o.i. blocks fusion, by removing sialic acid receptors for the viral HN. However, in cells infected with HPF 1 and 2, infection with high m.o.i. did not block fusion; the fusion increases with increasing m.o.i.. In cells infected with HPF 1 and 2, neither bacterial nor NDV neuraminidase blocked cell fusion, using amounts of neuraminidase that completely block fusion of HPF3 infected cells. However, when inactivated HPF3 was used as a source of viral neuraminidase, the treatment inhibited fusion of cells infected with HPF 1 and 2 as well as 3. The differences found between these viruses in terms of their interaction with the cell, ability to modulate cell–cell fusion and reponse to exogenous neuraminidases of various specificities, may reflect salient differences in biological properties of the three viruses.  1999 Academic Press

Key words: human parainfluenza viruses, virus receptors, hemagglutinin-neuraminidase, sialic acid receptors, viral fusion.

Introduction Human parainfluenza viruses types 1, 2 and 3 (HPF 1, 2 and 3) are important pathogens in ∗ Author for correspondence. 0882–4010/99/110329+08 $30.00/0

children. While these three viruses share common structural features and replication strategies, they target different parts of the respiratory tract; the most common outcomes of infection with HPF3 are bronchiolitis and pneumonia, while HPF 1 and 2 are associated  1999 Academic Press

330

with the croup syndrome. Using our understanding of the functions of the HPF3 HN molecule in the cell-binding and viral entry process, we are investigating the ways in which these processes differ among the three types of HPF, to determine whether differences in HN–receptor interaction may contribute to tissue tropism. While paramyxoviruses have been considered to have many biological properties in common with one another, a number of lines of investigation indicate that there are important differences between viruses of this family in regard to viral entry and fusion. The paramyxoviruses differ from each other in the requirements for promotion of membrane fusion [7]. For a number of the paramyxoviruses that have been studied, both the HN and F proteins participate in the fusion process [9]. For HPF 3 [13], 1 and 2 [8], promotion of membrane fusion by the F protein requires participation of the HN protein. However, for SV5, the F protein can mediate membrane fusion in the absence of HN, and coexpression of F and HN has little effect on the degree of syncytium formation [1, 7, 18]. There are also important differences between paramyxoviruses in the receptor requirements for virus entry and spread. We have shown that HPF3 establishes persistent infection with no cell fusion at a m.o.i. of 5 [11, 13], however, NDV demonstrates widespread cell fusion at the same m.o.i. and mediates fusion from without at higher m.o.i. [3]. In comparing HPF3 to SV5, we found that at low m.o.i., the pathogenesis of HPF3 and SV5 are similar, but in a high m.o.i. infection the CPE resulting from infection with each virus is strikingly different [14]. Our finding that the fusion behaviour of HPF3 and SV5 are different at high m.o.i. is in accordance with the differences observed in the fusogenic activity of the expressed F glycoproteins of these viruses. For HPF3, the degree of cell fusion that occurs in a monolayer of infected cells depends on the m.o.i.; as this increases the extent of syncytium formation decreases. Cells can be persistently infected with HPF3 at a high m.o.i. of 5, and these pi cells do not show the usual cytopathic effect of cell fusion. This phenomenon is caused by the neuraminidase component of the HPF3 HN glycoprotein in the inoculum, cleaving sialic acid residues from the cell surface in proportion to the size of the viral inoculum [12, 14]. Larger inocula result in cleavage of a higher percentage of sialic acid residues, preventing interaction of viral HN with its receptor, thus blocking cell

C. Ah-Tye et al.

fusion by HPF3. The same effect can be produced at low m.o.i. by adding exogenous neuraminidase to the cells. Removal of sialic acid moieties, by addition of bacterial neuraminidase or UV irradiated HPF3 in amounts comparable to those in a high m.o.i. infection, reproduces the effect of infection without cell fusion. Our studies have extended to determining whether HPF1 and HPF2 behave similarly to HPF3 in terms of interaction of HN with its sialic acid containing receptor. The experiments involved manipulation of receptor availability, scrutiny of conditions required to establish persistent infection and comparison of the neuraminidase activities of the three types of virus. The results show that HPF 1, 2 and 3 differ from each other with respect to several properties that might be pertinent to pathogenic mechanisms.

Results Three experimental treatments were used to compare the HN–receptor interaction of HPF 1, 2 and 3: infection at high m.o.i., bacterial neuraminidase treatment of cells infected at low m.o.i. and exogenous viral neuraminidase treatment of cells infected at low m.o.i. (using NDV neuraminidase or UV irradiated HPF3 as sources of viral neuraminidase).

Effect of high m.o.i. on the outcome of infection with the three viruses Fig. 1 shows that while for HPF3 infection with high m.o.i. blocks fusion, for HPF 1 and 2 infection with high m.o.i. did not block fusion. The fusion mediated by HPF1 and HPF2 increases with increasing m.o.i. Cells were infected with human parainfluenza virus type 1, 2 or 3 at the indicated m.o.i. (0.1, 1 and 5). The photographs show that the degree of fusion at the higher m.o.i. differs dramatically between HPF 1, 2 and 3. This experiment provided the first indication of major differences in the receptor interactions between these viruses.

Effect of treatment with exogenous bacterial neuraminidases on the outcome of infection with the three viruses In these experiments cells were infected with the three viruses at low m.o.i., and treated with

Parainfluenza virus–receptor interactions m.o.i.

0.1

331 1

5

HPF1

HPF2

HPF3

Figure 1. Effect of m.o.i. on the degree of cell fusion produced by HPF 1, 2 and 3. Cells were infected with human parainfluenza virus type 1, 2 and 3 at the indicated m.o.i. (0.1, 1 and 5). The photographs show the progression of cell fusion in a monolayer after infection.

0.1 unit (per 60 mm dish) of C. perfringens neuraminidase (which has a wide range of sialic acid linkage specificity), or A. ureafaciens neuraminidase (which has moderate specificity for the 2–6 linkage). Fig. 2 shows a clear difference between the viruses; fusion mediated by HPF 3 was inhibited by the neuraminidases, while fusion mediated by HPF 1 or 2 was unaffected.

Effect of treatment with exogenous viral neuraminidases on the outcome of infection with the three viruses Newcastle disease virus (NDV) neuraminidase has a strict specificity for hydrolysis of the NeuAca2-3Gal linkage, with no hydrolysis of the NeuAca2-6Gal linkage [19]. Fig. 3 shows that in cells infected with HPF 1 and 2, NDV neuraminidase did not block cell fusion, using amounts that completely block fusion of HPF3 infected cells. However, when inactivated HPF3 was used as a source of viral neuraminidase (in amounts to simulate a high m.o.i.), the treatment inhibited fusion of cells infected with HPF 1, 2 and 3. Thus, only HPF3 viral neuraminidase inhibits the fusion mediated by all three viruses,

and NDV neuraminidase activity blocks only fusion mediated by HPF3. We then tested whether UV inactivated HPF 1 and 2 (in amounts equivalent to an m.o.i. of 10) are capable of blocking fusion in cells infected with HPF 1, 2 or 3. Neither HPF1 or HPF2 neuraminidase can block fusion in cells infected with any of the three viruses (data not shown). This finding is especially striking in view of the assays of the neuraminidase activities for HPF 1, 2 and 3 discussed below. Table 1 provides a summary of the effects of the experimental treatments used to compare the HN–receptor interaction of HPF 1, 2 and 3. The results of infection at high m.o.i., bacterial neuraminidase treatment of cells infected at low m.o.i. and exogenous viral neuraminidase treatment of cells infected at low m.o.i. are compared for the three viruses.

Relative neuraminidase activities of the three viruses Assessment of the relative neuraminidase activities for the three viruses reveals dramatic differences. Fig. 4(a) shows that HPF 1 and 2

332

C. Ah-Tye et al. Control

C. Perfringens

A. Ureafaciens

HPF1

HPF2

HPF3

Figure 2. Effect of bacterial neuraminidases on the degree of cell fusion produced by HPF 1, 2 and 3. Cells were infected with HPF 1, 2 or 3 at low m.o.i.. Bacterial neuraminidase (0.1 unit/60 mm dish) was added 90 min later. The photographs show the extent of cell fusion in infected cells incubated with or without bacterial neuraminidase.

Control

NDV

UV-HPF3

HPF1

HPF2

HPF3

Figure 3. Effect of exogenous viral neuraminidases on the degree of cell fusion produced by HPF 1, 2 and 3. Cells were infected with HPF 1, 2 or 3 at low m.o.i.. NDV neuraminidase or UV inactivated HPF3 was added 90 min later. The photographs show the extent of cell fusion in infected cells incubated with or without viral neuraminidases.

Parainfluenza virus–receptor interactions

333 3.5

Table 1. Summary of the effects of the experimental treatments used to compare the HN–receptor interaction of HPF 1, 2 and 3

(a) 3.0

Cell fusion HPF1

HPF2

HPF3

+ +

+ +

− +

Low m.o.i.+neuraminidases: – bacterial + – NDV, HPF1, HPF2 + – HPF3 −

+ + −

− − −

High m.o.i. Low m.o.i.

OD 549

2.5

The experiments described above indicated that the nature of the HN–receptor interaction for HPF 1 and 2 was very different from HPF3; in particular, high m.o.i. infection with these viruses resulted in greater CPE, whereas high m.o.i. infection with HPF3 causes no CPE and results in persistent infection (pi) [11]. This suggested that the mechanism for establishment of pi was different for these viruses. In examining this question, we took advantage of our finding that very low m.o.i. (<0.1 pfu/cell) infections with HPF2 in CV-1 cells result in minimal fusion. We infected cell monolayers with HPF2 at multiplicities of 0.0001–0.1, and tested for spread of virus through the culture by serial HAD. At the lowest multiplicities, HPF2 caused virtually no CPE but spread throughout the culture, as demonstrated by positive HAD. These cells were persistently infected, releasing HPF2 into the supernatant fluid, and the synthesis of viral protein and RNA continued after passage (data not shown). Since this method for establishment of pi was different from that for HPF3, we hypothesized that establishment of pi with HPF2 does not involve depletion of cell surface receptors for

1.5 1.0 0.5

0

have much higher activity than HPF3. We also determined that the neuraminidases of all three viruses demonstrate a preference for the 2–3 over the 2–6 linkage of neuraminic acid (Table 2). All have similar acidic pH optima [Fig. 4(b)], although the optimum pH of HPF 1 and 2 is slightly higher (5.2) than that of HPF3 (4.7).

30

60

90

120 150 Time (min)

180

210

240

1.2 (b) 1.0

Relative activity

Establishment of persistent infection by HPF2 requires different conditions than HPF3

2.0

0.8

0.6

0.4

0.2

0.0

3

4

5

6 pH

7

8

9

Figure 4. Relative neuraminidase activity of HPF 1, 2 and 3. (a) The amount of viral protein in sucrose gradient purified virus was quantitated on stained, scanned polyacrylamide gels. Aliquots of HPF 1 (Α), 2 (Β) and 3 (Χ) preparations with the same viral protein content were used to determine neuraminic acid release (see OD) as a function of time. (b) Product formation by each virus’ neuraminidase [HPF 1 (Α), 2 (Χ) and 3 (Β)] at different pHs.

HN. To test the prediction that these cells would not undergo fusion when seeded with uninfected CV-1 cells, CV-1 cells (3×106 cells) were added to confluent monolayer cultures of HPF 2 infected cells in 60 mm dishes. Fusion of the cells with the HPF2 infected cells was assessed at various times after incubation at 37°C [13]. We observed no fusion of uninfected cells with

334

C. Ah-Tye et al.

Table 2. Relative neuraminidase activities of the three viruses assayed with either 2,3- or 2,6-neuraminyllactose as substrate Substrate HPF1 HPF2 HPF3

2–3 NANL

2–6 NANL

1 1 1

0.001 0.024 0.12

the HPF2 infected cells (data not shown). This finding confirms that, contrary to HPF3, the mechanism of establishment of pi does not involve sialic acid receptor depletion.

Discussion and Conclusions The results presented indicate that HPF 1, 2 and 3 differ in their ability to modulate cell–cell fusion and in the effects of exogenous neuraminidases of various specificities on this process. Cell fusion mediated by HPF 3 is blocked by bacterial or viral neuraminidase treatments that remove a portion of the sialic acid containing receptors [14]. For HPF 1 and 2, cell fusion is not blocked by either high m.o.i. infection or by bacterial or NDV neuraminidase treatment, but is blocked by inactivated HPF3. While these differences do not necessarily translate into determinants of tropism, they may point to important differences in virus–host cell interaction. In spite of the much higher activity of the neuraminidases of HPF 1 and 2 relative to HPF3, only HPF3 neuraminidase blocks fusion by HPF 1, 2 and 3. Thus, the differential efficacy of the neuraminidases must lie in properties such as target selectivity, that are not revealed by the standard quantitative assay. Our previous selection experiments [16] suggest that HPF3 acts on specific sialic acid containing receptor(s) only, in contrast to bacterial neuraminidase which removes a diverse array of sialic acid containing molecules. The neuraminidases of HPF 1 and 2 may also have wider specificities than that of HPF3, and much of their relatively high total activity is directed to cleaving the sialic acid moieties on the large number of molecules that do not function as viral receptors. Several reports have documented HPF2 persistent infections in humans; however, despite evidence that HPF2 persistent infections occur there is little information to indicate whether

persistence is caused by factors relating to the virus, the host, or the environment. The ability of a wide variety of other paramyxoviruses to establish persistent infections in vivo and in vitro has been documented [2, 5, 6, 11, 17, 20–22], suggesting that pi of their respective hosts may be a general characteristic of this group of viruses. The herein described establishment of persistent infection with HPF2 in vitro indicates that, unlike HPF3 which establishes pi at higher m.o.i. due to depletion of sialic acid containing receptors, HPF2 establishes pi at very low m.o.i. without depletion of receptors. These data lend further support to our conclusion that these parainfluenza viruses differ greatly in their biology: establishment of pi in cell culture, which in the case of HPF3 relates directly to HN–receptor interactions, operates via a different route in HPF2. In addition, our results suggest that HPF3 neuraminidase treatment of HPF1 or HPF2 infected cells may select for HN variants of these viruses that have higher avidity for receptor. We have previously used bacterial neuraminidase to select for high avidity variants of HPF3, and these variants have provided information about functional sites on the HPF3 HN molecule [15]. In the case of HPF 1 and 2, such variants could be extremely useful in dissecting the structure– function relationship for the HN molecules.

Materials and Methods Viruses Stocks of wt HPF3 were made in CV-1 cells from virus that was plaque purified four times. HPF 1 and 2 were obtained from the ATCC and stocks were made in CV-1 cells after plaque purifying four times. Virus was released after freeze-thawing 36–48 h post-infection and stored at −80°C. HPF1 infections were carried out in the presence of trypsin as described [4]. Virus titre was determined by a plaque assay (see below) with CV-1 cells.

Cells CV1 (African green monkey kidney) cells were maintained with Eagle minimal essential medium supplemented with 10% fetal bovine serum and antibiotics.

Parainfluenza virus–receptor interactions

335

Determination of neuraminidase enzymatic activity

Acknowledgments

Neuraminidase activities were determined by the thiobarbituric acid assay [10] at the pH optima of 4.7 for HPF3 and 5.2 for HPF 1 and 2. 2,3- and 2,6-neuraminyllactose substrates were purchased from Sigma Chemical Corp. For these experiments, sucrose gradient purified virus was used. In each experiment, the input amounts of viral protein were quantitated on stained, scanned polyacrylamide gels, so that the amount of viral protein used in each assay was the same for HPF 1, 2 and 3.

We thank Olga Greengard and Richard Peluso for helpful conversations, Natasha Poltoratskaia for technical assistance, and are grateful to Frederick J. Suchy for support and encouragement. This work was supported by an NIH grant to A.M. (RO1 AI 31971).

Neuraminidase treatment Monolayer cultures of infected cells were washed with medium lacking serum. Neuraminidase preparations (from Clostridium perfringens; Sigma Chemical Co., St. Louis, MO, U.S.A.) were then added in serum free medium and incubation continued for 12–36 h. Within each experiment, the incubation period was the same for all three viruses and for uninfected cells; the experiment shown in Fig. 2 was incubated for 24 h.

Treatment with UV irradiated virus Virus was treated to eliminate its infectivity by irradiation with 254 nm wavelength light at a distance of 10 cm for 7 min on ice. The loss of infectivity was confirmed by plaque assay. Cells were infected with HPF 1, 2 or 3. The amount of UV irradiated virus added to these infected cells was equivalent to an m.o.i. of 10. Incubation was continued for 90 min after addition of the infectious and irradiated virus. Medium lacking serum was then added.

Hemadsorption assays Monolayers were washed with cold medium lacking serum and then incubated with human erythrocytes at 4°C for 75 min [14]. Nonadherent cells were removed by washing with cold medium, and the cells were photographed through a phase contrast microscope.

References 1 Bagai S, Lamb R. Quantitative measurement of paramyxovirus fusion: differences in requirements of glycoproteins between SV5 and human parainfluenza virus 3 or Newcastle disease virus. J Virol 1995; 69: 6712–9. 2 Basle M, Russell W, Goswami K, Rebel A, Giraudon P, Wild FA, Filmon RA. Paramyxovirus antigens in osteoclasts from Paget’s bone tissue detected by monoclonal antibodies. J Gen Virol 1985; 66: 2103–10. 3 Bratt M, Gallaher W. Preliminary analysis of the requirements for fusion from within and fusion from without by Newcastle disease virus. Proc Natl Acad Sci USA 1969; 64: 536–43. 4 Gorman W, Pridgen C, Portner A. Glycosylation of the Hemagglutinin–Neuraminidase glycoprotein of human parainfluenza virus type 1 affects its functional but not its antigenic properties. Virology 1991; 183: 83–90. 5 Goswami K, Cameron K, Russell W, Lange L, Mitchell D. Evidence for the persistence of paramyxoviruses in human bone marrows. J Gen Virol 1984; 65: 1881–8. 6 Gross P, Green R, McCrea Curnen M. Persistent infection with parainfluenza type 3 virus in man. American Review of Respiratory Disease 1973; 108: 894–8. 7 Horvath CM, Paterson RG, Shaughnessy MA, Wood R, Lamb RA. Biological activity of paramyxovirus fusion proteins: factors influencing formation of syncytia. J Virol 1992; 66: 4564–9. 8 Hu X, Ray R, Compans RW. Functional interactions between the fusion protein and hemagglutinin– neuraminidase of human parainfluenza viruses. J Virol 1992; 66: 1528–34. 9 Lamb RA. Paramyxovirus fusion: a hypothesis for changes. Virology 1993; 197: 1–11. 10 Merz D, Prehm P, Scheid A, Choppin PW. Inhibition of the neuraminidase of paramyxoviruses by halide ions: a possible means of modulating the two activities of the HN protein. Virology 1981; 112: 296–305. 11 Moscona A, Galinski MS. Characterization of human parainfluenza virus type 3 persistent infection in cell culture. J Virol 1990; 64: 3212–8. 12 Moscona A, Peluso RW. Persistent infection with human parainfluenza virus 3 in cell culture: analysis of the role of defective interfering particles. Virology 1993b; 194: 399–402. 13 Moscona A, Peluso RW. Fusion properties of cells persistently infected with human parainfluenza virus type 3: participation of hemagglutinin–neuraminidase in membrane fusion. J Virol 1991; 65: 2773–7. 14 Moscona A, Peluso RW. Fusion properties of cells in-

336

15

16

17

18

fected with human parainfluenza virus type 3: receptor requirements for viral spread and virus-mediated membrane fusion. J Virol 1992; 66: 6280–7. Moscona A, Peluso RW. Relative affinity of the human parainfluenza virus 3 hemagglutinin–neuraminidase for sialic acid correlates with virus-induced fusion activity. J Virol 1993a; 67: 6463–8. Moscona A, Peluso RW. Analysis of human parainfluenza virus 3 receptor binding variants: evidence for the use of a specific sialic acid-containing receptor. Microbial Pathogenesis 1996; 20: 179–84. Muchmore H, Parkinson A, Humphries J, Scott E, McIntosh L, Cooney M, Miles J. Persistent parainfluenza virus shedding during isolation at the South Pole. Nature 1981; 289: 187–9. Paterson RG, Hiebert SW, Lamb RA. Expression at the cell surface of biologically active fusion and hemagglutinin/neuraminidase proteins of the para-

C. Ah-Tye et al.

19

20

21

22

myxovirus simian virus 5 from cloned cDNA. Proc Natl Acad Sci USA 1985; 82: 7520–4. Paulson J, Weinstein J, Dorland L, van Halbeek H, Vliegenthart J. Newcastle disease virus contains a linkage-specific glycoprotein sialidase. J Biol Chem 1982; 257: 12734–8. Pringle C, Wilkie M, Elliott RM. A survey of respiratory syncytial virus and parainfluenza virus type 3 neutralising and immunoprecipitating antibodies in relation to Paget disease. J Med Virol 1985; 17: 377–86. Ter Meulen V, Koprowski H, Iwasaki Y, Kackell Y, Muller D. Fusion of cultured multiple sclerosis brain cells with indicator cells: presence of nucleaocapsids and virions and isolation of parainfluenza type virus. Lancet 1972 2: 1–5. Walker D, Hinze H. A carrier state of mumps virus in human conjunctival cells. I. General considerations. J Exp Med 1962; 116: 739–49.