Nonhuman Transferrin Receptor 1 Is an Efficient Cell Entry Receptor for Ocozocoautla de Espinosa Virus

Abstract

Ocozocoautla de Espinosa virus (OCEV) is a novel, uncultured arenavirus. We found that the OCEV glycoprotein mediates entry into grivet and bat cells through transferrin receptor 1 (TfR1) binding but that OCEV glycoprotein precursor (GPC)-pseudotyped retroviruses poorly entered 53 human cancer cell lines. Interestingly, OCEV and Tacaribe virus could use bat, but not human, TfR1. Replacing three human TfR1 amino acids with their bat ortholog counterparts transformed human TfR1 into an efficient OCEV and Tacaribe virus receptor.

TEXT

Classified arenaviruses have been assigned to four phylogenetic groups, the Old World (OW) viruses and clades A to C New World (NW) viruses (1). Each arenavirus is adapted to a distinct vertebrate host, which maintains the virus chronically without developing disease (2). The arenavirus glycoprotein (GP1 subunit of glycoprotein precursor [GPC]) primarily determines host tropism (3–5) by binding to one of several cell surface receptors. The primary receptor for two pathogenic OW arenaviruses, Lassa virus (LASV) and lymphocytic choriomeningitis virus (LCMV), is α-dystroglycan (6). All classified pathogenic clade B arenaviruses (i.e., Chapare, Guanarito [GTOV], Junín [JUNV], Machupo [MACV], and Sabiá) and at least two nonpathogenic clade B arenaviruses (i.e., Amaparí [AMAV] and Tacaribe [TCRV]) can use transferrin receptor 1 (TfR1) to enter host cells. TfR1 seems to be the primary cell entry receptor for these viruses (7), but TfR1-independent replication has been observed in vitro, as well as in certain laboratory-mouse models (8–12). Importantly, human TfR1 is utilized only by the pathogenic clade B viruses but not by nonpathogenic AMAV or TCRV (13, 14). The GP1s of JUNV and MACV, but not of GTOV, efficiently bind the TfR1 ortholog of the MACV host, the big laucha (Calomys callosus). All three GP1s of GTOV, JUNV, and MACV can utilize the TfR1 ortholog of the GTOV host, the short-tailed zygodont (Zygodontomys brevicauda). JUNV GP1 efficiently binds to the TfR1 ortholog of its host, the drylands laucha (Calomys musculinus), whereas GTOV and MACV GP1 do not (15). TfR1 orthologs from related rodents, the house mouse (Mus musculus) and brown rat (Rattus norvegicus), do not support infection by GTOV, JUNV, or MACV (15). Finally, the nonpathogenic AMAV and TCRV efficiently use the TfR1 orthologs of their primary hosts, the common neacomys (Neacomys spinosus) and the Jamaican fruit-eating bat (Artibeus jamaicensis), respectively. Whereas TCRV can bind to both TfR1s, AMAV cannot use the bat receptor, but both viruses can use short-tailed zygodont and big laucha TfR1 (16). These observations indicate that TfR1, and especially TfR1 orthologs, determine the relative (but not necessarily absolute) permissibility of individual mammal hosts to NW clade B arenaviruses.

Between 1545 and 1815, viral hemorrhagic fever (VHF) epidemics that may have been caused by arenaviruses were recorded in Mexico (17, 18). In 1967, a VHF epidemic occurred in Chiapas State among several hundred Maya people. Low titers of complement-fixing antibodies against arenaviruses found in survivors suggested an arenavirus etiology (19). In 2012, Cajimat et al. sequenced the genome of a novel arenavirus, Ocozocoautla de Espinosa virus (OCEV), found in kidneys of Mexican deer mice (Peromyscus mexicanus) captured in Chiapas State (20). OCEV could not be isolated, but phylogenetic studies placed the OCEV genome firmly in NW clade B. Consequently, Cajimat et al. speculated that OCEV could have been the etiological agent of the 1967 Mexican epidemic.

To identify OCEV's cell entry determinants, we first evaluated whether the OCEV genome encodes a functional GPC. Retroviruses pseudotyped with arenavirus glycoproteins have been used successfully to characterize arenavirus cell entry determinants (10, 13, 15, 16, 21–24). We therefore produced enhanced green fluorescent protein (eGFP)-expressing Moloney murine leukemia virus (MoMLV) pseudotyped with control (TCRV, MACV, or LASV) GPC (15, 16, 25) or OCEV GPC (synthesized by DNA2.0 based on GenBank accession number JN897398) and evaluated transduction efficiencies in cells obtained from ATCC. Grivet (Chlorocebus aethiops) Vero E6 cells (Fig. 1A) and human HeLa cells (Fig. 1B) were exposed to pseudotypes for 5 h and imaged for eGFP 48 h later. As published previously, MACV and LASV pseudotypes transduced Vero E6 (Fig. 1A) and HeLa (Fig. 1B) cells efficiently (10, 16, 21, 22). Although not as efficiently as MACV and LASV pseudotypes, OCEV pseudotypes entered Vero E6 cells at levels comparable to those of TCRV, indicating that the OCEV GPC is functional.

Fig 1.

OCEV glycoprotein GPC mediates cell entry. eGFP-encoding MoMLV was pseudotyped with OCEV, TCRV, MACV, or LASV GPC and used to transduce grivet Vero E6 (A), human HeLa (B), or human NCI-60 panel (C) cells. (A and B) Images of transduced cells (blue, DAPI stained). (C) Quantification of transduction of OCEV pseudotype using FACS. Shown are the results of one of two independent, but highly similar, NCI-60 screens. Error bars indicate standard deviations for duplicate samples.

Since OCEV pseudotypes entered HeLa cells poorly (Fig. 1B), we expanded our screening of the four pseudotypes (MACV, LASV, OCEV, and TCRV) to 53 human NCI-60 panel cancer cell lines (26) and measured transduction efficiencies by fluorescence-activated cell sorting (FACS) (Fig. 1C). As expected, LASV and MACV pseudotypes efficiently transduced most human cancer cells (data not shown). OCEV pseudotype transduction was extremely low in 52 of 53 tested cell lines, with some lines reaching the low level of OCEV susceptibility of Vero E6 cells (Fig. 1C, red line) and one cell line (SNB19) being somewhat more susceptible than Vero E6 cells (Fig. 1C).

To identify cells that are highly permissive to OCEV, we screened human 293T/17, Egyptian rousette (Rousettus aegyptiacus) RoNi/7.2 (27), Büttikofer's epauletted fruit bat (Epomops buettikoferi) EpoNi/22.1 (27), grivet BS-C-1, Chinese hamster (Cricetulus griseus) CHO-K1, Syrian hamster (Mesocricetus auratus) BHK-21, house mouse (Mus musculus) NIH/3T3 and M2-10B4, hispid cotton rat (Sigmodon hispidus) CRL, and domestic dog (Canis familiaris) MDCK cells (Fig. 2). Results obtained with grivet BS-C-1 cells were similar to those obtained with Vero E6 cells. As expected (10, 16, 21), hamster and house mouse cells were largely refractory to transduction with MACV, OCEV, and TCRV pseudotypes but were susceptible to LASV pseudotypes. Hispid cotton rat cells proved to be excellent targets for LASV and MACV pseudotypes but were largely resistant to TCRV and OCEV pseudotype transduction. Dog cells were somewhat susceptible to LASV pseudotype transduction but not permissive to other pseudotypes. Unexpectedly, OCEV pseudotypes transduced RoNi/7.2 and EpoNi/22.1 cells with high efficiency and TCRV, MACV, and LASV pseudotypes also entered both bat cell lines at lower efficiencies (Fig. 2). These results identified the two bat cell lines as suitable tools for further OCEV studies.

Fig 2.

OCEV pseudotypes transduce bat cells with high efficiency. MoMLV pseudotyped with the indicated arenavirus GPCs were used to transduce cells from the indicated mammals, followed by imaging.

Since OCEV clusters with clade B NW arenaviruses, we hypothesized that TfR1 is an OCEV receptor. We cloned Egyptian rousette TfR1 (reTfR1) by extracting cellular RNA from RoNi/7.2 cells using an RNeasy minikit (Qiagen) followed by SuperScript III one-step reverse transcription (RT)-PCR (Life Technologies) using primers 5′-ATGATGGATCAAGCCAGATCAGCAWTCTCT-3′ and 5′-AAAYTCATTGTCAATGTCCCAAAYGTCACCA-3′. Amplified PCR products were cloned by bidirectional TOPO TA cloning (Life Technologies), and five clones were analyzed. A new set of primers, 5′-CACCATGATGGATCAAGCCAGATCAGCA-3′ and 5′-TTACTTATCGTCATCGTCCTTGTAGTCAAATTCATTGTCAATGTCCCA-3′, was used to clone reTfRI with a C-terminal FLAG tag using a pcDNA3.1 directional TOPO expression kit (Life Technologies). Following protocols and reagents previously described (15, 16), we transiently overexpressed FLAG-tagged control human (hTfR1), house mouse (mTfR1), Jamaican fruit-eating bat (ajTfR1), big laucha (ccTfR1), common neacomys (nsTfR1), cat (fTfR1), or reTfR1 or transfected empty plasmid into naturally hTfR1-expressing HEK 293T/17 cells and transduced these cells with OCEV, TCRV, MACV, or LASV pseudotypes (Fig. 3). Anti-FLAG antibody (Sigma-Aldrich) detected surface TfR1s (Fig. 3A). Unsurprisingly, LASV pseudotypes, which are dependent on α-dystroglycan for cell entry, entered HEK 293T/17 cells with roughly equivalent efficiencies independently of TfR1 expression. As expected, hTfR1 and ccTfR1 overexpression increased susceptibility to MACV pseudotypes compared to that for empty plasmid (Fig. 3B). However, these TfR1 orthologs as well as mTfR1 did not appreciably affect cell permissiveness to OCEV or TCRV pseudotypes. Bat (reTfR1 and ajTfR1) and nsTfR1 conferred susceptibility to OCEV and TCRV pseudotypes and increased cell susceptibility to MACV pseudotypes compared to that of the empty-plasmid controls. Interestingly, OCEV could utilize fTfR1, which was also efficiently used by NW clade B hemorrhagic fever arenaviruses (15) and the nonpathogenic AMAV, but not the nonpathogenic TCRV (16). Together, these results indicate that TfR1 is a receptor for OCEV.

Fig 3.

TfR1 is an OCEV receptor. (A) C-terminal FLAG-tagged human (h), Egyptian rousette (re), house mouse (m), Jamaican fruit-eating bat (aj), big laucha (cc), common neacomys (ns), or cat (f) TfR1 was overexpressed in HEK 293T/17 cells, and expression was confirmed using anti-FLAG antibody. Cells overexpressing the indicated TfR1s were transduced with arenavirus pseudotypes and imaged. (B) The same experiment as for panel A, with transduction efficiencies quantified by FACS. Error bars indicate standard deviations for two independent experiments.

A local region in the apical domain of TfR1 (residues 208 to 212), and in particular residue Y211, determines GP1 binding specificity (15, 16, 23). Following published strategies (15, 16, 25), we aligned hTfR1 residues 201 to 216 with those of other orthologs (Fig. 4) and then created three FLAG-tagged h/reTfR1 chimeras. These ORFs were stably overexpressed in CHO-K1 cells using the Flp-In system (Life Technologies) (Fig. 5). Cells were exposed to pseudotypes, imaged (Fig. 5A), and analyzed by FACS (Fig. 5B). As little as three amino acid changes (h/re2TfR1, GRL206 to 209→SPT) transformed hTfR1 into a functional OCEV and TCRV receptor.

Fig 4.

Alignment of TfR1 ortholog residues 201 to 216. Alignment of TfR1 apical domain residues 201 to 216 of the indicated mammals and generated human/Egyptian rousette TfR1 (h/re) chimeras. TfR1 apical domain residues critical for NW clade B hemorrhagic fever arenavirus GP1 binding (residues 208 to 212) are underlined, and critical residue Y211 is shown in green. Residues deviating from the hTfR1 sequence are shown in red.

Fig 5.

Three amino acid changes turn human TfR1 into a functional OCEV receptor. Effect of TfR1 chimera expression on arenavirus pseudotype transduction. (A and B) Stable expression in CHO-K1 cells. (A) Confirmation of expression of FLAG-tagged chimera TfR1 using anti-FLAG antibody (FACS) and images of transduced cells. (B) FACS analysis of transduction efficiency. Error bars indicate standard deviations for two independent experiments.

The results of our study indicate that OCEV, as it currently circulates in Mexico, cannot use human TfR1 efficiently to enter human cells. The amino acid sequence of OCEV GP1 supports this conclusion. In the case of MACV GP1, residues R111, Y122, and D123 were found to be essential and residues D155 and P160 to be important for direct engagement of human TfR1 (28). Of these residues, only D123 and P160 are conserved in OCEV GP1. Furthermore, the three bat ortholog residues that transformed human TfR1 into an efficient OCEV receptor are located within TfR1 motif 5. These human TfR1 residues contact MACV GP1 residues S97 and F98 (23). Interestingly, S97 is conserved among GTOV, JUNV, MACV, and SABV (23) but T occupies the same position in OCEV and TCRV. Mutational studies with OCEV GP1 should be performed to elucidate which, and how many, mutations would be necessary for adaptation of OCEV to human TfR1.