Virokinin, a bioactive peptide of the tachykinin family, is released from the fusion protein of bovine respiratory syncytial virus.

Tachykinins, an evolutionary conserved family of peptide hormones in both invertebrates and vertebrates, are produced by neuronal cells as inactive preprotachykinins that are post-translationally processed into different neuropeptides such as substance P, neurokinin A, and neurokinin B. We show here that furin-mediated cleavage of the bovine respiratory syncytial virus fusion protein results in the release of a peptide that is converted into a biologically active tachykinin (virokinin) by additional post-translational modifications. An antibody directed to substance P cross-reacted with the C terminus of mature virokinin that contains a classical tachykinin motif. The cellular enzymes involved in the C-terminal maturation of virokinin were found to be present in many established cell lines. Virokinin is secreted by virus-infected cells and was found to act on the tachykinin receptor 1 (TACR1), leading to rapid desensitization of this G protein-coupled receptor as shown by TACR1-green fluorescent protein conjugate translocation from the cell surface to endosomes and by co-internalization of the receptor with beta-arrestin 1-green fluorescent protein conjugates. In vitro experiments with isolated circular muscle from guinea pig stomach indicated that virokinin is capable of inducing smooth muscle contraction by acting on the tachykinin receptor 3. Tachykinins and their cognate receptors are present in the mammalian respiratory tract, where they have potent effects on local inflammatory and immune processes. The viral tachykinin-like peptide represents a novel form of molecular mimicry, which may benefit the virus by affecting the host immune response.

Bovine respiratory syncytial virus (BRSV) 1 and human respiratory syncytial virus (HRSV), two closely related RNA vi-ruses of the family Paramyxoviridae, replicate in the lower respiratory tract, causing similar diseases in their respective hosts (1). Young calves and infants are particularly affected and suffer from obstructive bronchitis, wheezing, and hypoxemia. Damage of respiratory epithelial cells due to the infection itself probably plays a minor role in RSV pathogenesis (2). Rather, the inflammatory response triggered by cellular chemokines and cytokines that are released from RSV-infected epithelial cells and lymphocytes is believed to contribute to the severity of the disease (3)(4)(5)(6)(7)(8)(9).
Several observations indicate that the adaptive immune response to RSV is incomplete and of short duration. Maternally derived virus-neutralizing antibodies do not provide sufficient protection from infection (10,11). Reinfections, even by RSV of the same serogroup, occur repeatedly during childhood and throughout life (12,13). There is also evidence that infection with HRSV contributes to increased airway hyperresponsiveness and asthma (14 -17). Recent data suggest that RSV has immunosuppressive properties. In the mouse model, HRSV infection causes suppression of lung CD8 ϩ T-cell effector activity and inhibition of pulmonary CD8 ϩ T-cell memory development (18). In BRSV-infected lambs, mitogen-induced proliferation of peripheral blood lymphocytes was found to be reduced (19 -21), and lymphopenia as well as an increased susceptibility to opportunistic infections were observed (21,22).
The envelope glycoprotein F plays an important role in the RSV life cycle. It mediates fusion between the viral and cellular membrane and also has receptor binding activity (23)(24)(25). Recent work has shown that the F proteins of HRSV and BRSV cause contact inhibition of mitogen-induced T-cell proliferation (26). Moreover, the F protein is a major virus antigen that induces neutralizing antibodies in the host (1).
The primary sequence of the F protein between different serotypes of HRSV and BRSV is highly conserved but shows only little homology with other paramyxovirus fusion proteins. A property that is also found with other enveloped viruses is the synthesis of the fusion protein as an inactive precursor F 0 that has to be proteolytically cleaved to become fusion-active (27). This cleavage occurs at a multibasic amino acid motif (RKRR 136 2) proximal to the so-called fusion peptide and is mediated by the ubiquitous endoprotease furin of the trans-Golgi network (28). A unique feature of the RSV F proteins is the cleavage of F 0 at a second highly conserved furin recognition site (RA(R/K)R 109 2) 27 amino acids upstream of the fusion peptide. Cleavage at both sites results in the release of an N-glycosylated peptide, pep27 (Fig. 1A) (29,30). Recent studies using recombinant BRSV revealed that both the distal furin recognition motif and pep27 are dispensable for virus replication in cell culture (31). We here demonstrate that pep27 is subject to further post-translational modifications and is converted into virokinin, a bioactive peptide of the tachykinin family. Virokinin acts on specific G protein-coupled receptors and shows characteristic tachykinin-like activity such as induction of smooth muscle contraction. The proinflammatory and immunomodulatory properties of tachykinins in general suggest a potential role of virokinin in the pathogenesis of BRSV.
Peptides and Antibodies-Peptides were prepared by Fmoc (N-(9fluorenyl)methoxycarbonyl) solid phase synthesis and purified by reversed-phase high pressure liquid chromatography to 95% purity (Biosynthan GmbH, Berlin, Germany). The biotin-labeled tracer used in the competitive ELISA was synthesized by coupling a biotin group via a 6-amino-n-hexanoic acid to the peptide TRNSTKKFYGLM-amide. Substance P, neurokinin A, and neurokinin B (purity Ն96%) were purchased from Calbiochem-Merck (Darmstadt, Germany). For generation of a monospecific anti-virokinin (anti-VK) serum, the peptide CTRNST-KKFYGLM-amide was covalently linked by an N-(␣-maleimide)-succimide ester to keyhole limpet hemocyanin (Sigma) as carrier protein and used for immunization of rabbits with ABM-ZK (Linaris GmbH, Bettingen, Germany) as adjuvant. The obtained sera were tested by ELISA, immunofluorescence, and Western blot. The rabbit monospecific antibody SP3 directed to substance P has been described previously (32).
Generation of TACR1, TACR1-GFP, and ␤-Arrestin 1-GFP Constructs-cDNA was generated from human lung mRNA (Clontech) by reverse transcription using the enzyme from Moloney murine leukemia virus (Expand reverse transcriptase; Roche Applied Science) and random hexanucleotides for priming. The open reading frame of the human TACR1 was amplified by PCR (Expand Long Template PCR System; Roche Applied Science) using the forward primer 5Ј-TTTTGG-TACCGAAATGGATAACGTCCTCCCGGTGGAC-3Ј (KpnI restriction site underlined, nucleotides encoding the TACR1 N terminus in boldface type) and the reverse primer 5Ј-TTTTGGATCCCTAGGAGA-GCACATTGGAGGAGAAGC-3Ј (BamHI restriction site underlined and nucleotides encoding the TACR1 C terminus in boldface type). The PCR product was treated overnight with BamHI and KpnI, purified, and ligated with pCEP4 plasmid (Invitrogen) to obtain pCEP4-TACR1. For generation of a TACR1-GFP fusion protein, the TACR1 open reading frame without the stop codon was amplified from the recombinant pCEP4 plasmid using the same forward primer as before and the reverse primer (5Ј-GGCCGGGGATCCCGGGAGAGCACATTGGAGGA-GAAGCTGAAGC-3Ј) (BamHI restriction site underlined, nucleotides encoding the TACR1 C terminus in boldface type). The PCR product was treated with the restriction endonucleases as before, purified, and ligated into the pEGFP-N1 vector (Clontech) to obtain pEGFP-N1-TA-CR1. The ␤-arrestin 1 open reading frame without the stop codon was amplified from human lung cDNA using the forward primer 5Ј-GGCC-GGAAGCTTGCCACCATGGGCGACAAAGGGACGCGAGTGTT-3Ј (HindIII restriction site underlined, nucleotides encoding the ␤-arrestin 1 N terminus in boldface type) and the reverse primer 5Ј-GGCCGGC-CGCGGTCTGTTGTTGAGCTGTGGAGAGCCGG-3Ј (SacII restriction site underlined, nucleotides encoding the ␤-arrestin 1 C terminus in boldface type). The PCR product was treated with HindIII and SacII overnight and ligated with pEGFP-N1 vector to obtain pEGFP-N1-ARR1. DNA sequencing was performed for all of the three constructs to rule out any amplification errors.
Stable Cell Lines and Translocation Assays-CHO-K1 cells were transfected with the pEGFP-N1-TACR1 plasmid and grown in selection medium containing 10% of dialyzed FCS and 1 mg of G418 sulfate/ml for 14 days. Cell clones were isolated by limiting dilution and screened for GFP expression by fluorescence microscopy. Positive clones were recloned once. A HeLa cell line that stably expresses the human TACR1 plasmid was established by transfecting the cells with the pCEP4-TACR1 plasmid. This vector contains elements from the Epstein-Barr virus providing episomal replication of the plasmid in human cells. The cells were selected with medium containing 10% dialyzed FCS and 250 g of hygromycin B/ml (Roche Applied Science) for 14 days and maintained in medium with half-concentrated hygromycin B. Expression of TACR1 was shown by immunofluorescence using a rabbit polyclonal antibody directed to human TACR1 (Novus Biologicals, Littleton, CO). Because of the episomal replication of the plasmid, selection of cell clones was not required. Stable expression of human ␤-arrestin 1 was achieved by transfection of the HeLa-TACR1 cells with the pEGFP-N1-ARR1 plasmid and selection of the cells with G418 sulfate as described above. To induce translocation of GFP-tagged proteins, cells were washed twice with serum-free medium and incubated for 30 -60 min at 37°C with conditioned cell culture supernatant from BRSV-or HRSVinfected Vero cells or with medium containing synthetic peptides. Thereafter, the cells were washed once with PBS and fixed with 3% paraformaldehyde for 20 min. Fluorescence microscopy was performed with a Zeiss Axioplan 2 microscope equipped with a digital video camera (INTAS Focus Imager, Göttingen, Germany).
Immunofluorescence-Cells were fixed with 3% paraformaldehyde 40 h after infection with recombinant BRSV and double-stained with the rabbit SP3 antiserum (1:500) and a mouse monoclonal antibody directed to the RSV fusion protein (1:200; Serotec, UK). The primary antibodies were detected with Cy3-conjugated anti-rabbit Ig serum (Sigma) and fluorescein isothiocyanate-conjugated anti-mouse Ig serum (Amersham Biosciences). For competition analysis, the SP3 antiserum was preincubated with synthetic peptides (10 g/ml) for 15 min prior to cell staining. Fluorescence was visualized with a Leica confocal laserscanning microscope.
Western Blot Analysis-Vero cells were infected with BRSV at a multiplicity of infection of 0.1 and maintained in serum-deficient medium for 3 days at 37°C. The cell culture supernatant was harvested and subjected to low speed centrifugation (2000 ϫ g, 15 min, 4°C) to remove detached cells. The clarified supernatant was desalted using Sephadex G25 columns (Amersham Biosciences) and concentrated by lyophilization. The samples were run on a Tricine-SDS 15% polyacrylamide gel under reducing conditions and transferred to nitrocellulose by the semidry blotting technique. The membrane was blocked overnight and subsequently incubated with the rabbit monospecific antibody directed to virokinin (1:1000 in PBS), a biotinylated goat antirabbit immunoglobulin serum (Amersham Biosciences; 1:1000 in PBS), and a streptavidin-peroxidase complex (Amersham Biosciences; 1:2000 in PBS). Finally, the nitrocellulose was incubated for 1 min with a chemiluminescent peroxidase substrate (BM chemiluminescence blotting substrate; Roche Applied Science), and the resulting light emission was detected and documented with a supercooled CCD camera (Chemi-Doc System; Bio-Rad).
Competitive ELISA-The rabbit monospecific anti-virokinin serum or preimmune serum that served as a negative control was diluted 1:10,000, bound to protein A-coated microtiter plates (Pierce) overnight, and incubated for 3 h with 25 l of cell culture supernatant and 25 l of biotinylated virokinin (10 ng/ml) as tracer. The wells were washed, incubated for 1 h with 50 l of streptavidin-peroxidase complex (1:1000; Amersham Biosciences), washed, and then incubated for 1 h with 50 l of 2,2Ј-azino-di-[3-ethylbenzthiazoline sulfonate (6)] diammonium salt peroxidase substrate (Roche Applied Science). Substrate conversion was photometrically quantified at 405 nm. A calibration curve was assessed using synthetic virokinin in the range of 0.01-100 ng/ml. Four parallel measurements were performed for each sample.

RESULTS
A sequence homology search using the BLAST algorithm revealed that within the C-terminal part of BRSV pep27, there exists a significant homology with preprotachykinins ( Fig. 1B). Most significantly, BRSV pep27 contains the motif FYGLM 129 , which matches the tachykinin signature sequence FXGLM, where X represents an aromatic or branched aliphatic amino acid (33). In addition, the sequences downstream of this motif contain all of the information necessary for the maturation of a tachykinin peptide. The basic motif KKRKRR 136 provides a target for endoproteolytic cleavage by prohormone convertases (34) and subsequent trimming by carboxypeptidases (35). The glycine residue, Gly 130 , carboxyl-terminal to the tachykinin motif provides the NH 2 group for C-terminal amidation by peptidylglycine ␣-amidating mono-oxygenase (36). In contrast, the N-terminal sequence of the viral peptide bears no significant similarity to mammalian tachykinins (Fig. 1C). On the basis of these observations, we speculated that a novel 20amino acid peptide with tachykinin-like biological activity is generated and released by BRSV-infected cells. Because of its viral origin, we refer to this putative tachykinin as virokinin. Despite the overall high homology of HRSV and BRSV F proteins, a tachykinin signature sequence is absent from pep27 of HRSV.
In order to analyze the C-terminal maturation of BRSV pep27, we took advantage of a recombinant BRSV mutant, in which the distal furin motif has been changed from RAKR 109 to RANN 109 , rendering this site resistant to furin (31). pep27 of this mutant remains attached to the small subunit F 2 and is transported to the cell surface as part of the disulfide-linked F 1 -F 2 complex. Cleavage at the proximal furin motif is not affected and allows any normal processing of the C terminus of pep27. Calf kidney PT-11 cells were infected with either the BRSV cleavage mutant or the parental virus, and the cells were analyzed 2 days after infection by cell surface immunofluorescence using a monospecific antiserum directed to substance P (SP) and a monoclonal antibody directed to the F protein. Cells infected with the parental virus showed expression of the F antigen but were not recognized by the SP antibody ( Fig. 2A). In contrast, cells infected with the cleavage mutant strongly reacted with both antibodies. Binding of the SP antibody was abolished when the cells were pretreated with low amounts of trypsin, which results in cleavage at the modified motif RANN 109 and in elution of pep27 from the cell surface (30). When the substance P antibody was incubated with the cells in the presence of the peptide KFYGLM-amide, which corresponds to the postulated C terminus of VK, binding of the antibody was inhibited (Fig. 2B). In contrast, the C-terminally elongated peptide KFYGLMG-amide did not interfere with antibody binding. Even the peptide KFYGLM, terminating with the correct amino acid but lacking amidation of the C terminus, had no inhibitory activity. These results indicate that cellular enzymes convert pep27 into a shorter peptide that terminates with the classical tachykinin motif. The presence of this highly conserved sequence in the viral peptide explains cross-reaction with the SP antibody.
To determine whether the enzymes responsible for C-terminal maturation of pep27 are expressed in a cell type-specific manner, a variety of cell lines originating from different species and different tissues were infected with the BRSV cleavage mutant and analyzed for reactivity with the SP antibody (Table  I). All cell lines were susceptible to infection by BRSV as indicated by cell surface expression of the viral fusion protein.
In most cell lines, infection also led to presentation of the tachykinin epitope on the cell surface, indicating that the enzymes required for the maturation of VK have a broad tissue expression profile. However, calf thymus epithelial cells (TEK) and human bronchial epithelial cells were infected but did not react with the SP antibody. Immunoprecipitation analysis of the fusion protein from infected TEK cells showed that the proximal furin consensus sequence of the protein was efficiently processed, whereas the modified distal furin consensus motif was not cleaved (data not shown), suggesting that subsequent steps in the C-terminal maturation of pep27 are impaired in this cell line.
To allow detection of secreted virokinin, a high affinity antibody directed to the mature molecule was generated by im- munizing rabbits with the synthetic peptide CTRNST-KKFYGLM-amide coupled to keyhole limpet hemocyanin. The antiserum was found to strongly react with cells that had been infected with the BRSV cleavage mutant but not with cells infected with the parental virus (data not shown). The antibody was then used to detect virokinin in the supernatant of BRSVinfected cells by Western blot analysis. Fig. 3A shows that the antibody reacted with synthetic virokinin that migrated as a band of 2.3 kDa (lane a). In the supernatant of BRSV-infected cells, a single band of about 10 kDa was detected (lane b). The higher molecular mass is due to the N-linked oligosaccharide side chain attached to the central asparagine of the peptide (29,30). We have recently shown that most of pep27 can be deleted from the BRSV F protein without affecting virus replication in cell culture (31). Supernatant from cells that had been infected with this deletion mutant did not show any reactivity with the VK-specific antibody (lane c). Likewise, the antibody did not react with supernatant from HRSV-infected (lane d) and from noninfected Vero cells (lane e). In addition, virokinin was not secreted by cells that had been infected with the recombinant BRSV cleavage mutant (see above). Rather, the virus particles released into the medium were found to react with the antibody. Under nonreducing conditions, a band of 72 kDa was detected that corresponded to the disulfide-linked F1-F2 complex (data not shown). Under reducing conditions, we observed a 38-kDa band that represented pep27 linked to the F2 subunit of the fusion protein (30).
For quantitative estimation of virokinin in the cell culture supernatant, the VK-specific antibody was used in a competitive immunoassay with biotinylated virokinin as a tracer (Fig.  3, B-D). This assay allowed us to detect virokinin in a concentration range from 0.5 to 50 ng/ml (Fig. 3B). The specificity of the antibody was confirmed using different synthetic peptides for competition with the tracer (Fig. 3C). Only peptides that mimicked the C terminus of virokinin were able to inhibit binding of the tracer to the immobilized antibody (compare bar a with bars b and c). In contrast, a virokinin-like peptide with a nonamidated C terminus (bar d) or virokinin-like peptides containing additional amino acids at the C terminus (bars e and f) did not compete with the tracer. This shows that the VKspecific antibody like the substance P-specific antibody requires a completely processed virokinin for recognition. The competitive ELISA allowed us to detect VK in the medium of calf kidney cells as early as 24 h after infection with parental BRSV (Fig. 3D). In the course of infection, the VK concentration increased, reaching 70 ng/ml at day 3 and decreased at day 4 to 35 ng/ml probably because of the progressing cytopathic effect and release of lysosomal enzymes. For comparison, physiological concentrations of neurokinin A in the bronchoalveolar lavage of healthy persons are between 0.35-1.4 ng/ml (37). VK was not detected in the medium of cells that had been infected with the BRSV cleavage mutant; nor was it found in the supernatant of cells infected with the BRSV deletion mutant that lacked the intervening pep27. Biological responses to agonists of G protein-coupled receptors are usually rapidly terminated by receptor phosphorylation, uncoupling from heterotrimeric G proteins, and receptor endocytosis (38). G protein-coupled receptor kinases and ␤-arrestins play an important role in these processes. The cellular translocation of green fluorescent protein (GFP)-tagged receptors and GFP-tagged ␤-arrestins provides an elegant way to study G protein-coupled receptor activation and desensitization (39 -41). To determine whether VK exhibits tachykinin biological activity, we made use of this approach and examined how a GFP-tagged ␤-arrestin 1 (␤ARR1-GFP) responded to native and synthetic virokinin. We generated a HeLa cell line that stably expressed both the human tachykinin receptor 1 (TACR1) and ␤ARR1-GFP. The GFP-tagged protein was found to be uniformly distributed in the cytoplasm and in the nucleus (Fig. 4a). Incubation of the cells in the presence of 10 nM SP for 45 min caused translocation of ␤ARR1-GFP into endosomes (Fig. 4b). This effect was not observed when the neuropeptides neurokinin A (NK-A) and neurokinin B (NK-B) were applied at

FIG. 2. Surface expression of a substance P-related epitope by cells infected with a recombinant BRSV mutant.
A, calf kidney cells were infected with either parental BRSV or a recombinant mutant virus (RANN 109 ) where the distal furin consensus sequence RAKR 109 of the F protein was changed into RANN 109 (see Fig. 1B). Double immunofluorescence analysis was performed 40 h after infection using a mouse monoclonal antibody directed to the fusion protein (green) and a rabbit monospecific antiserum directed to substance P (red). RANN 109 ϩ trypsin, cells infected with the cleavage mutant were pretreated with 1 g/ml acetylated trypsin for 60 min at 37°C prior to antibody staining. B, calf kidney cells infected with the BRSV cleavage mutant were incubated with monospecific anti-substance P serum in the presence of the indicated peptides (10 g/ml). Bar, 40 m.

TABLE I Generation of virokinin in different cell lines
The indicated cell lines were infected with the BRSV cleavage mutant and analyzed by double immunofluorescence for reactivity with antibodies directed to the viral fusion protein (anti-F) or substance P (anti-SP) as described in the legend to the same concentration (Fig. 4, c and d). These neuropeptides are known to preferentially act on the tachykinin receptors 2 and 3 (TACR2 and TACR3), respectively (33). However, ␤ARR1-GFP was translocated into endosomes when the cells were incubated with 10 nM of synthetic VK, indicating that the viral peptide, like SP, acts on TACR1 (Fig. 4e). To elucidate whether native, glycosylated virokinin also exhibits tachykinin-like activity, we incubated the reporter cells with conditioned medium from noninfected (Fig. 4f) or BRSV-infected cells (Fig. 4g). Only in the latter case did we observe efficient translocation of ␤ARR1-GFP even when the medium was diluted (1:10). This effect was abolished when the cells had been pretreated with Ac-Trp-3,5-bis(trifluoromethyl)benzyl ester, a competitive TACR1 antagonist (42) (Fig. 4h). As additional controls, we also used cell culture supernatant from cells either infected with the BRSV deletion mutant that lacked pep27 or with HRSV. With both, we could not detect any translocation of ␤ARR1-GFP even when the supernatant was applied undiluted (Fig. 4, i and j).
A similar approach was used to directly visualize ligand-induced internalization of TACR1. GFP was fused to the C terminus of the receptor, and this construct, TACR1-GFP, was used for stable transfection of CHO-K1 cells. In unstimulated cells, TACR1-GFP was localized primarily at the plasma membrane. This distribution was not changed when the cells were incubated with conditioned medium from noninfected cells (Fig. 5a). However, the distribution of TACR1-GFP rapidly changed when diluted cell culture supernatant from BRSVinfected cells was applied. Following a 45-min incubation, the GFP-tagged receptor was detected in a prominent pool of vesicles with a perinuclear localization (Fig. 5b, see arrow). A similar effect was observed with 10 nM synthetic VK (Fig. 5c), whereas there was no activity with conditioned medium from cells infected with either the BRSV deletion mutant (Fig. 5d) or with HRSV (Fig. 5e). These results indicate that BRSV-infected cells secrete a glycosylated tachykinin-like peptide that acts on TACR1 and leads to ␤-arrestin-mediated rapid receptor internalization and therefore receptor desensitization. Since G protein-coupled receptor activation ultimately terminates with the association of ␤-arrestins and receptor, the agonist-induced translocation of the GFP-tagged receptor or of the GFP-tagged ␤-arrestin 1 also indicates that both glycosylated and synthetic VK are able to cause TACR1 activation (39).
To determine whether VK exhibits tachykinin biological activity, we studied its ability to induce smooth muscle contraction, a typical property of tachykinins (43,44). First, we treated isolated circular smooth muscle strips from the guinea pig stomach with conditioned medium from virus-infected cells. It turned out that the virus-induced cytopathogenic effect resulted in the release of tachykinin-unrelated substances that caused smooth muscle contraction. This effect was not inhibited by TACR antagonists. Therefore, the experiments were performed with synthetic tachykinins. Treatment of the muscle with a 0.5 M concentration of either SP or VK evoked a strong, reversible contraction, leading to a significant increase in tone (Fig. 6). N-terminally truncated versions of VK, which may represent additional proteolytic fragments of the peptide, were also tested, and fragment 10 -20 and fragment 15-20 were found to be as effective as VK. On the contrary, the related peptides KFYGLM and KFYGLMG-amide, the C termini of which slightly differ from fragment 15-20, did not induce smooth muscle contraction (data not shown). Also, pep27 from the HRSV fusion protein (Fig. 6) as well as an unrelated peptide from HIV-1 gp120 (data not shown) did not reveal any tachykinin activity. The TACR3 receptor antagonist SR-142801, but not the TACR1 receptor antagonist CP 99,994, dramatically reduced the rise in tone evoked by VK, fragment 10 -20 and fragment 15-20 (Fig. 6). These findings suggest that VK acted through a neural mechanism, since TACR3 is the predominant receptor subtype in the stomach and is located on enteric nerves (45). DISCUSSION In this report, we have shown that proteolytic processing of a viral membrane protein results in the release of a peptide that is subject to further post-translational modifications and is secreted by infected cells as a biologically active tachykinin designated VK. The generation of biologically active VK is the result of an intimate relationship between the virus and the host cell as several cellular enzymes are involved in the maturation of the peptide. Following release of pep27 from the F protein by action of the ubiquitous trans-Golgi network protease furin, all of the six basic amino acids at the C terminus of the peptide are removed. This trimming reaction is probably mediated by a carboxypeptidase of the trans-Golgi network (35). The C terminus of VK is then amidated by use of the glycine residue adjacent to the tachykinin signature sequence. Most likely, this conversion is performed by peptidyl-glycine ␣-amidating mono-oxygenase, a bifunctional enzyme of the trans-Golgi network and secretory granules that also acts on several other peptide hormones (36). It is remarkable that this amidating activity could be detected in many conventional cell lines that were not known to produce peptide hormones or to dispose of a regulated secretory pathway. Our finding that VK is correctly processed in these cell lines suggests that amidation of peptides may not be restricted to specialized secretory cells but is a more common property of mammalian cells. The attachment of an N-linked oligosaccharide to the central asparagine residue of the peptide is a modification that occurs cotranslationally when pep27 is still part of the F molecule (30). The presence of an N-linked oligosaccharide is a unique feature of VK, since other members of the tachykinin family are not modified in this way. Tachykinins are very short lived molecules that are rapidly degraded, for example by action of neutral endopeptidase (46). A possible function of the N-linked oligosaccharide might be the protection of VK from proteolytic attack and inactivation.
Tachykinins are signaling molecules that bind to specific G protein-linked receptors on the target cell, thereby causing activation of phospholipase C and/or adenylate cyclase (47,48). Downstream intracellular pathways then lead to activation of protein kinase C, mitogen-activated protein kinase, calmodulin kinases, and/or protein kinase A with profound effects on transcriptional activity and cellular functions (49 -51). Three mammalian tachykinin receptor subtypes have been characterized, TACR1, TACR2, and TACR3, which show preferential but not absolute selectivity for SP, NK-A, and NK-B, respectively (33,44). Our data indicate that VK, like SP and hemokinin, acts on TACR1 but is also capable of activating TACR3. With respect to the tachykinin signature sequence, VK also resembles hemokinin and SP, since all three tachykinins contain an aromatic amino acid in the motif (see Fig. 1C). This is different from NK-A and NK-B, which both have a valine residue at this position. A detailed pharmacological study will show which receptor subtypes are the preferred targets for virokinin. Tachykinin receptors are expressed by many different cell types that respond to tachykinins in a cell type-specific manner. VK may contribute to the pathogenicity of BRSV, since several symptoms associated with severe BRSV infection like bronchoconstriction, excessive mucus secretion, histamine release, influx of lymphocytes, and edema formation are known to be inducible by substance P and other tachykinins during the process of neurogenic inflammation (52). The TACR3 subtype appears to be also involved in this process, since bronchial hyperresponsiveness to acetylcholine, bronchial microvascular permeability, hypersensitivity to histamine, and cough are inhibited with selective NK3 receptor antagonists (53).
There is also a peptide of 27 amino acids proteolytically released from the fusion protein of HRSV (30). However, this peptide neither contains a tachykinin signature sequence nor revealed tachykinin-like activity. In this respect, BRSV and HRSV, although highly related, differ from each other. Nevertheless, HRSV is able to induce the production of substance P in the respiratory tract of infected mice (54). The mechanism of this induction is not understood, but it appears to depend on the presence of the viral G and/or SH protein. In addition, HRSV infection has been shown to cause up-regulation of TACR1 in the respiratory tract (55,56). Consequently, treatment of HRSV-infected mice with anti-substance P antibody reduced pulmonary inflammation (57). Thus, the pathogenicity of both BRSV and HRSV appears to be closely associated with the action of tachykinins.
The primary sequence of VK is highly conserved in all BRSV strains, suggesting that the virus benefits from this bioactive peptide. There are some features of tachykinins that might be useful for virus survival in the host. First, tachykinins are able to inhibit apoptosis in some cells (58). Since viral replication often induces programmed cell death, a delay in this process due to VK should support virus propagation, or it might even favor the establishment of a persistent infection (59). Second, tachykinins are potent immunomodulators (60 -62). Not only are tachykinin receptors expressed by dendritic cells, macrophages, eosinophils, and mast cells as well as B-and T-cells, but many of these cells also produce tachykinins that act in an autocrine or paracrine fashion (63). In most cases, tachykinins have a stimulating or potentiating effect on lymphocyte proliferation and differentiation, cytokine secretion, and immunoglobulin production (49,(63)(64)(65)(66)(67). The permanent secretion of VK from BRSV-infected cells may lead to its accumulation at high concentrations at the site of infection. This, in turn, could cause desensitization of local tachykinin receptors (68,69) and thus dampen down the local immune response. The importance of cellular tachykinins for controlling virus infection has been recently demonstrated using tachykinin-deficient (PPT-A Ϫ/Ϫ ) mice (70). In these animals, clearance of murine gammaherpesvirus 68 (MHV-68) was delayed by several days, although infection was ultimately controlled.
In order to escape immune surveillance and to increase virus survival in the host, many large DNA viruses express homologues of cellular proteins like cytokines, chemokines, or receptors thereof (71). Our study reveals a novel means by which an RNA virus exploits the cellular biosynthetic machinery to generate a bioactive peptide related to cellular neuropeptides. The study of VK not only will improve our understanding of virushost interaction but will also enhance our knowledge about the immunoregulatory properties of tachykinins. The production of bioactive peptides probably is not confined to BRSV. It should be considered that pep27 from the related HRSV fusion protein might also harbor a biological function even if it is not a tachykinin.