Progress on Receptors of Canine Distemper Virus

ZHU Liqi, ZHU Chen, SUN Jingyu, CAO Shuyang, XIE Lingzhi, QIAN Miao,YIN Jun, ZHANG Quan

(1. Institute of Comparative Medicine, College of Veterinary Medicine, Yangzhou University, Yangzhou 225009, China; 2. Jiangsu Co-innovation Center for Prevention and Control of Important Animal Infectious Diseases and Zoonoses, Yangzhou University, Yangzhou 225009, China)

Abstract: Canine distemper (CD) is an infectious disease caused by canine distemper virus (CDV). CDV can infect many species of animals and cause huge economic losses in canine and other fur animal industry. The major CDV receptors are the signaling lymphocyte activation molecule (SLAM) and the poliovirus receptor-like protein-4 (PVRL4, also known as Nectin-4), which are closely related to the pathogenesis of CD. In addition to these two distinct receptors, CDV may also infect animals through other unknown receptors. This article reviews advances in the knowledge on CDV cell receptors to get a better understanding of the pathogenesis of CDV.

Key words: Canine distemper virus; receptor; pathogenesis

Canine distemper (CD) is a severe infectious disease caused by Canine distemper virus (CDV),belongs to the genus Morbillivirus of the family Paramyxoviridae[1], including Measles virus (MeV),Rinderpest Virus (RPV), Pestedes petits ruminants pestivirus (PPRV), Phocine distemper virus (PDV),and Porpoise morbillivirus (PMV) . In terms of viral structure and ultrastructure, CDV is almost identical to MeV and RPV. These three viruses are closely related and have similar molecular biological characteristics and antigenic relationships[2].

CDV can infect dogs of different ages, sexes,and breeds. Besides canidae, many species of animals such as the Felidae, Mustelidae, P rocyonidae, Ursidae and Primates are also susceptible to CDV[3]. Similar to the MeV, CDV first targets on lymphoid tissues and causes similar lymphopenia and immunosuppression.Subsequently, CDV spreads to the epithelial tissue,infecting a wide range of epithelial cells, which in turns the infect nerve tissue[4]. There have been reports of CDV strains causing large-scale outbreaks of canine distemper in Chinese and Japanese macaques[5], but no detectable cross-protective antibodies against MeV and CDV have been found. However, CDV infection promotes elevated levels of MeV neutralizing antibodies,suggesting the presence of cross-reactive antibodies.The high similarity between CDV and MeV, as well as the success of the campaign to eradicate MeV, may cause failure of CDV cross-protection provided by measles vaccines or natural infections, and may also trigger a large number of CDV zoonotic possibilities research[6]. Although there had been no report that wild type CDVs can utilize human SLAM, and the mutation of H is necessary for the virus to use human SLAM[7],CDV should be considered as a potential threat to human beings seriously, and the development of effective antiviral drugs and vaccines against CDV are of great practical significance.

1 Virus introduction

CDV particles are diverse in shape, mostly spherical, sometimes filamentous, with a diameter of 120-350 nm. The virus center contains a helical nucleocapsid composed of genomic RNA and nucleocapsid protein subunits with a broad diameter of about 15-17 nm. CDV is a non-segmented, nonoverlapping single-stranded negative-strand RNA virus. The viral genome is 15690 bp in length; from the 3' end to the 5' end are the 3' leader sequence,Nucleocapsid (N), Phosphoprotein (P) (containing two non-structural protein genes, C and V), Matrix protein(M), Fusion protein (F), Hemagglutinin protein (H),Large protein (L), and a 5' trailer sequence. The 3'-end leader sequence and the 5'-end tail sequence direct the transcriptional replication p rocess of the gene[8].

The N protein wraps the surface of the RNA to form a nucleocapsid which protects RNA from degradation. The N protein sequence is highly conserved and is an antigenic protein with a cross-immunological reaction between genus Morbillivirus. Studies have shown that the two non-structural proteins V and C encoded by P protein in different open reading frames have the function of blocking the host interferon pathway and participating in evading the host immune response during the infection of MeV. The L protein is considered to be a multifunctional enzyme unit of a single-strand negative-chain non-segmented RNA virus, performing enzymatic p rocesses involved in transcription and replication, including initiationextension-termi nation of ribonucleotide polymerization,mRNA transcription and the addition of Cap,polymerase-related cofactors, P protein methylation and specific phosphorylation. P protein, N protein and L protein constitute a ribonucleoprotein (RNP) complex together which is the basic infectious unit of virus transcription and replication; M protein is the smallest protein in CDV, mainly involved in virus assembly,budding and release. In addition, studies have shown that M protein affects the localization of H protein and F protein on the capsule, which reduces the binding ability of H protein to receptor and thus affects the pathogenicity of CDV.

The F protein is one of the glycoproteins on the surface of the virus, which mediates the fusion between the envelope and cell membrane. Studies have shown that altering the position of the F protein signal peptide in the leader sequence can affect F protein function and thus affect CDV viral fusion and neurotoxicity. In addition, the binding of the H protein to the cellular receptor triggers a series of conformational changes in the F protein, directly promoting the fusion activity of the viral cells. Among members of genus Morbillivirus,MeV, CDV and RPV have group-specific epitopes,mainly on the F protein. The F protein has a high degree of epitope homology and is the major crossprotective antigen of heterotypic immunity[9]. There are two important Th lymphocyte epitopes on the F protein,one of which is present on the N-terminal sequence of F1. This Th cell epitope may be the highest priority presentation site for canine antigen presenting cells(APCs). Therefore, F protein has certain significance in inducing CDV protective immune response[10].

The H protein belongs to the member of the typeⅡ glycoprotein family and is the main component of the vesicular fibroid. It has the highest mutation rate among the eight proteins encoded by CDV, which determines the host specificity of CDV. The H protein assists the F protein in invading the host cell by membrane fusion of the viral envelope with the host cell. Briefly, the H protein first adsorbs the host cell SLAM receptor, and then the conformation of the H protein is altered, and produces a signal for the F protein to cause membrane fusion between the viral envelope and the host cell to invade the host cell[11]. Moreover, the extent and efficiency of membrane fusion is also dependent on the H protein. Previous studies[12]showed that the H gene was changed before and after the adaptation to Vero cells of CDV isolates and vaccine strains. Amino acid derivation confirmed the presence of amino acid substitutions in the H proteins after adapting to Vero cells, especially in the H-4 epitope, N 481 turnning to Y481.

2 Virus receptors

Virus receptors are cell surface sites that mediate virus invasion into cells by binding to virus adsorbing proteins specifically. The host tissues and cell surface receptors are the main factors that determine the virus invasion pathways, diffusion patterns, and host disease characteristics. At present, two cellular receptors,signaling lymphocyte activation molecule (SLAM)and the poliovirus receptor-like protein-4 (PVRL4,Nectin-4) have been identified as CDV receptors. SLAM is expressed only on lymphocytes, while Nectin-4 is present on epithelial cells. When CDV invades, both receptors interact with the viral H protein β-propellershaped structure through the immunoglobulin variable region (V) at the distal membrane end, and virus replicates in lymphoid and epithelial tissues through SLAM and Nectin-4 respectively[13], and then transmits to the nervous system and replicates, forming persistent infection. Understanding the CDV receptors helps to understand the infection mechanism of CDV, which provides new ways for the treatment of this virus and the development of vaccines.

2.1 SLAM In 2000, Tatsuo et al.[14]discovered the second MeV receptor-signal lymphocyte-activating molecule (SLAM), also known as CD150. The receptor can be recognized by both clinical MeV strains and vaccine strains. SLAM can be expressed on lymphocytes, especially on activated and memory T cells, B cells, platelets, monocytes, NKT cells, and mature DC cells. SLAM belongs to the immunoglobulin superfamily and other members including 2B4 (CD244),NTB-A (Ly108, SF2000), CD84, Ly9 (CD299),CRACC, 19A. Its extracellular components contain one V-type and one C2-type immunoglobulin-like (Iglike) domain. Binding of the V domain at the N-terminus to the MeV H protein mediates viral adhesion. As its own ligand, SLAM participates in various immune functions, including: co-stimulation of T cells and B cells, secretion of IFN from Th1 cells, and inhibition of apoptosis of B cells[15]. In 2001, Tatsuo et al. further demonstrated SLAM is a receptor for CDV[16]. By using indirect immunofluorescence to perform SLAM cell localization, semi-quantitative PCR and indirect immunohistochemistry (IHC) to detect the distribution of SLAM in the related tissues of healthy minks and infected minks[17], a large number of CDV antigen positive cells were observed in liver, spleen, lung,kidney, and brain tissues of infected minks. Moreover,the results of PCR demonstrated that SLAM was expressed in minks′ liver, spleen, lungs, kidneys and brain[17]. From this research, we can understand that SLAM receptors are expressed in the cell membrane and cytoplasm, and are also distributed in liver, spleen, lung,kidney and brain tissues.

Lan et al.[18]transfected CDV strains 007Lm,S124C, Ac96I and Vero cell adapted strains of CDV in Vero-DST (Vero-Dog SLAM Tag) cells for 20 passages. Although the CPE of the 20thgeneration virus was slightly weaker than the original strain, its growth condition was better. The sequencing results showed that the H protein amino acid sequence of each strain produced changes at different positions compared to the original strains, but the P, M, and L proteins did not change. Another research provided evidence that three key residues in the virulent canine distemper virus A75/17 H protein (Y525, D526, and R529),clustering at the rim of a large recessed groove created by-propeller blades 4 and 5, control SLAM-binding activity without drastically mod ulating protein surface expression or SLAM-independent F triggering[19]. These results indicate that the mutation in the CDV H protein is an important factor when interacting with the canine SLAM.

The functional location and size of CDV proteins are in good agreement with MeV. From the perspective of the three-dimensional structure of the receptor, the interaction between hSLAM and virus is mainly focused on the MeV H protein (MeV-H) and its receptor.In general, the binding of SLAM does not lead to conformational changes in the MeV H protein. Two H molecules are also present in the complex structure to form a typical dimer structure[20]. However, through careful analysis, Hashiguchi et al.[21]found that the H/SLAM dimer in the structure can be further assembled to form two forms of tetrameric conformation. The corresponding full-length MeV H protein was expressed in 293T cells, and the presence of tetramers was indeed observed.

The three-dimensional structure prediction of SLAM receptors showed that the SLAM molecule consists of 21 amino acids, two of which are spatially close to the β-sheet and form a binding surface with the virus H protein, and 8 of them (located in aa 63, 66,69, 72, 84, 119, 121 and 130 of the SLAM molecule)are presumed to be essential amino acids that determine the binding of viral H protein to SLAM[22]. Hashiguchi et al.[21]successfully crystallized the crystal structure of MV-H and hSLAM-V domain 4 polymer complexes,and identified 13 key amino acid positions of hSLAM-V in spatial structure interacting with MV-H protein, and 7 of which are located in the predicted 21 amino acid residues of Ohishi et al[23]. The amino acid homology between canine and human SLAM is more than 75%[24].The three-dimensional structure prediction shows that the spatial structure of the 8 interaction with CDV-H is similar to that of hSLAM-V-MV-H[19]. Macaca SLAM (maSLAM)-V domain presents a typical V-type immunoglobulin-like structure consisting of two sets of folded sheets of BED and AGFCC 'C''. In the complex structure, maSLAM-V interacts extensively with the 4,5, 6 paddles of the H protein (H-4, H-5, H-6) through its AGFCC'C'' folded sheet, mediating the binding of two molecules; H-5 formed the binding center with SLAM, which is highly consistent with the results of previous mutation experiments.

It was confirmed that three CDV wild strains could easily replicate on Vero cells and caused CPE.Further sequence analysis outcomes showed that there was no change in the H gene[25]. Neither SLAM expression nor mRNA transcription was detected in Vero cells[26]. Since H protein is the only virus receptor binding protein according to the current study for canine distemper virus and even measles virus, it suggests that there are other CDV binding receptors in addition to SLAM. Fujita et al. incubated B95a cells with SLAM monoclonal antibody, and then used rCDV-EGFP to infect B95a cells (expressing positive human SLAM cells), the cell infection rate was 90% or 26% when the antibody absent or present, respectively[27]. These results indicated that other receptors might exist for CDV infection. Immunohistochemistry showed that CDV immunofluorescence staining in the footpad epithelial cells was strongly positive, but no SLAM positive cells were present in the footpad epithelium, which also indicates the presence of additional cell receptors[28].

2.2 Nectin-4 Although SLAM has been found to be a receptor for CDV and MeV, there is continuing evidence that CDV and/or MeV can infect epithelial cells and some tumor cells independent of SLAM. Using gene microarray technology to compare the difference in MeV susceptibility and non-susceptible gene expression levels, the other receptor molecule of MeV was discovered. Two research groups reported in 2012 that nectin-4 acts as a MeV receptor in epithelial cells and tumor cells[29]. Nectin-4, also known as PVRL-4,is a member of the immunoglobulin superfamily, which includes nine members: 4 Nectin molecules (Nectin 1-4) and five Nectin-like molecules (Nectin-like 1-5, Necl 1-5). All family members are typical type I transmembrane proteins consisting of the extracellular segment, the transmembrane region, and the intracellular segment. The extracellular domain contains three Ig domains, and the N-terminus is a V domain[30]. The intracellular domain of the Nectin molecule binds to the Afadin containing PDZ protein, causing aggregation of actin, but Necl does not[31]. Members of the Nectin family are able to form homodimers or heterodimers involved in calcium-independent cell adhesion proce sses, and are associated with cell migration,proliferation, and synapse formation[32]. Nectin-4 is expressed on placental trophoblast cells, Gastric gland cells, lung, breast and ovarian adenomas, and also has a certain level of expression on tonsils, oral mucosa,and esophagus and airway epithelial cells[33]. Antibodies against the Nectin-4 V domain completely blocked the invasion of MeV, indicating that Nectin-4 mediates viral-host interactions with the N-terminal V domain.In addition to MeV using Nectin-4 as a receptor,some members of the Nectin family are also used by viruses, for example, Necl-5 (CD155) is a receptor for poliovirus[34]. Nectin-4 was first reported to be associated with CDV infection in 2012[4]. As one of the CDV receptors, Nectin-4 has a very similar mechanism of infection with SLAM, particularly via the V-domain of the far-membrane of Nectin-4 interacted with CDV H protein. Interestingly, although Nectin4-MeV induced less syncytia formation compared to SLAM, the interaction of Nectin4 with MeV H was stronger than that of MeV H with SLAM[29]. It was demonstrated that canine Nectin-4 as a CDV receptor could be expressed on epithelia of different organs as well as neurons in the brain[35]. Therefore, the neurological symptoms presented after CDV infection may be related to this receptor. The inhibition of canine Nectin-4 by RNAi or neutralizing antibody can reduce CDV titer and EGFP fluorescence[13], and canine Nectin-4 can promote the formation of syncytia. Interestingly, the CDV Ac96I strain hardly replicates in the NCI-H358 cell line expressing human Nectin-4 but can grow adaptively in it. There was a truncated C protein in the Ac96I strain,and it was found that the growth ability of the strain with truncated C protein was lower in NCI-H358 cells after comparing with other CDV strains with intact protein C. The binding of CDV to Nectin-4 receptor had an important relationship with protein C. Interestingly, the growth ability of the CDV in SLAM positive cells did not affected by the integrity of C protein[36].

SLAM and Nectin-4 as the most recognized two types of CDV receptor had an important relationship with the mechanism of CDV infected cells. In the H protein of CDV, P493/Y539 residues are involved in canine nectin-4 mediated fusion, and at the same time,the F132/P133/A134/G135 residues in the V domain of Nectin-4 also play key roles in the fusion of virus with cells[37]. In 2012, the complex crystal structure of the wild type MeV strain (IC-B strain) H protein with Nectin-4 was successfully analyzed[38]. Similar to the previously analyzed vaccine strain H protein, the H of the wild-type strain also exhibited a typical sixpropeller blade structure; a dimer similar to vaccine strain H was obtained by symmetric operation. The Nectin-4 molecule binds to the groove between the β4 and β5 paddles of the H protein through its own N-terminal V-type immunoglobulin-like domain by hydrophobic interactions.

2.3 Other receptors The first measles virus receptor CD46 was discovered in 1993, and two different research groups reported that transfection of CD46 in measles virus-insensitive cells can cause infection with the measles virus (Edmonston vaccine strain)[39]. After that Sareen provided evidence that neither vaccine nor wild type strains of CDV bound to CD46[40]. However,there is evidence suggesting that CDV-H protein interacts with an unknown cellular receptor(s) regulated by CD9, which has been shown to form a complex with CD46, integrin beta 1 and heparin-binding EGF-like growth factor (pro-HB-EGF)[41]. In addition, it was proved that pro-HB-EGF is a cellular receptor for PDV.Although PDV belonged to the genus Morbillivirus,CDV and measles virus cannot bind to this receptor[42].Heparin-like receptors are considered as potential receptors for CDV, while the research on heparinlike receptors is very limited. The cell infection rate dropped from 88% to 81% after incubation of rCDVEGFP with heparin[27]. When treated with anti-SLAM antibody, the effect was even more effective.Terao Muto[43]reported infection of SLAM-negative HEK-293 cells and hepatocytes after treatment with recombinant heparin (rM V-EGFP) of MV wild-type strain HL strain expressing enhanced green fluorescent protein with soluble heparin. The rate has dropped significantly. There was no significant inhibition of 293 cells expressing SLAM. At the same time, heparin affinity chromatography test proved that the purified rM V-EGFP virions could directly bind to heparin.The protein bound to heparin is H glycoprotein, not F protein, which means in the case of the presence of SLAM, the effect of the heparin-like receptors is small. Heparin sulfate was present on the cell surface and can serve as the first receptor of various viruses such as vaccinia virus, foot-and-mouth disease virus,dengue virus, and pseudorabies virus. The interaction of some viruses with heparin sulphate induced a cellular signal response to promote the reorganization of the cytoskeleton or virus infection of cells. However, the mechanism of heparin sulfate as a canine receptor was not clear.

In addition, other receptors mainly mediate virus adhesion rather than entry into cells. For example,DC-SIGN and Langerin can bind to the H protein of MeV and promote the infection of MeV, while, their interaction cannot cause the membrane fusion process of the virus[44].

2.4 Unknown neuro receptors In the clinical case of CD, the diseased animal got high fever, cough,purulent nasal discharge, secondary gastrointestinal disease, high degree of mental depression, and symptoms such as epilepsy, circling, unstable standing,etc[45]. This provides a basis for the existence of CDV neuro receptors. MeV, CDV and PPRV are not only genetically similar, but also cause similar symptoms in their respective hosts, such as fever, respiratory tract and gastrointestinal diseases with secondary bacterial infections[46]. Interestingly, measles virus rarely causes neurological disease, but CDV is frequent. There is no connection between centra l nervous system syndrome and PPRV infection[47]. In canine distemper demyelinating leukoencephalitis (DL) caused by CDV,astrocyte s represent as the main virus target[48]. The comparison between single-cycle H protein knockout and complete H protein virus infection indicted that the complete H-protein is required for CDV persistently infection and cell-to-cell transmission[49]. Virus transfer that occurring selectively during astrocyte process es may contribute to the long-range coverage of the virus in astral neural networks. Lisa[50]found that Nectin-4 receptor and SLAM receptor could not be detected in primary cultured astrocyte s, but the combination of viral fusion mechanisms could lead to the existence of undiscovered neuro receptors. Further experiments also showed that CDV persistent infection in the canine brain white matter is independent of the Nectin-4 receptor[50].Besides, Nectin-4 receptor can be detected in ependymal cells, choroid plexus epithelial cells, meningeal cells,nerve cells, granulocytes, and Purkinje's cells[51], but it cannot be detected in microglial cells and astrocytes,supporting the idea that CDV has an unknown neuro receptor. Possible pathways for CDV transmission from the periphery to the central nervous system have been proposed, such as the hematogenous spread of virusinfected lymphocytes across the blood-brain barrier and the blood-cerebrospinal fluid (CSF) barrier[52]. The spread of virus along olfactory nerves along the nasal concha is another possible pathway. Contrary to the elucidated CDV entry mechanism, little is known about the pattern of viral transmission after invasion of the CNS. CDV mainly targets neurons in the brainstem,cortex and hippocampal brain parenchyma, acute phase neurons. In addition, EGFP fluorescence in brain slices shows continuous infectious progression in the central nervous system: CDV mainly infected the neuroependymal cells lining the ventricular wall and the neurons of the hippocampus and cortex adjacent to the ventricle, and it then progressed to an extensive infection of the brain surface, followed by the parenchyma and cortex. The ependymal and meningeal barriers are in intimate contact with the underlying astrocyte layer (glial cell restriction). Progressive loss of infected neurons in these areas is associated with the development of neurological signs, especially seizures and circling. In the acute phase, CDV invading the cerebrospinal fluid may primarily infect neurons and induce trans-synaptic transmission, and subsequent neurological complications of demyelinating encephalitis will be caused by the spread of astrocytes between viral cells[49].

2.5 Potential transmission to human Early reports have reported that CDV was isolated in patients with multiple sclerosis. Since both SLAM and Nectin-4 receptors are present in human cells, whether CDV infects human cells is an important research. Na Feng isolated a virus from the lung tissues of rhesus monkeys that died of CDV and infected them with three cell lines,Vero-monkey SLAM, Vero-canine SLAM, and Verohuman SLAM[53]. The results showed that the virus is capable of infecting Vero cell lines expressing dogs and monkeys SLAM, but it cannot infect Vero cell lines expressing human SLAM. The Nectin4 sequences from humans and dogs are almost identical at the amino acid level and both V domains contain the FPAG motif[54].The infection of the H358 cell line with recombinant CDV-A75/17red confirmed that the virus could bind to human Nectin-4 receptor, and the gene sequences of H and F proteins did not change. After that, the virus was successfully used to infect Vero cell line expressing human SLAM. It was found that position 540 of the H protein sequence have changed from Asp to Gly(D540G) after sequence alignment. Structural modelli ng suggests that the adaptive mutation D540G in H reflects the sequence alteration from canine to human SLAM at position 70 and 71 from Pro to Leu (P70L) and Gly to Glu (G71E)[55]. These results suggested that the mutation of nucleotide on the CDV-A75/17 genome gave it ability to bind to Vero-human SLAM. It is speculated that CDV may be capable of cross-host infection of hSLAM through the mutation of certain key amino acids, and thus becoming the pathogen of infection in humans. However, a recent study showed that additional mutations may require full virulence in non-native hosts, and there was a difference between the immune system of humans and monkeys[6]. From the perspective of receptors, humans are easily infected by CDV, but the colonization and growth of the virus are related to many factors. Suitable receptors can only be used as basic conditions for CDV infection in humans.

3 Conclusion

CDV cell receptors have been a vital topic of research. The understanding of CDV receptors can provide new directions for the treatment of CDV and the development of vaccines. At present, the study of CDV infection mechanism most relies on the interaction of H protein with SLAM and nectin-4 receptors, but research on the genetic variation of other CDV proteins(such as M, P and V) and their pathogenic mechanisms in the host is few. Therefore, studying the interaction of viruses with receptors through the construction of CDV infectious clones by reverse genetics, elucidating the molecular mechanisms of viral infection and assessing the susceptibility of humans and other primate hosts are imperative. To comprehensively understand the pathogenesis of CDV, it is important to continue to explore the neural receptors of CDV. In addition,the continued development of vaccines to cope with the constantly mutating CDV is also the basis for the prevention and control of Canine distemper.