Dual host-virus arms races shape an essential housekeeping protein

Abstract

Transferrin Receptor (TfR1) is the cell-surface receptor that regulates iron uptake into cells, a process that is fundamental to life. However, TfR1 also facilitates the cellular entry of multiple mammalian viruses. We use evolutionary and functional analyses of TfR1 in the rodent clade, where two families of viruses bind this receptor, to mechanistically dissect how essential housekeeping genes like TFR1 successfully balance the opposing selective pressures exerted by host and virus. We find that while the sequence of rodent TfR1 is generally conserved, a small set of TfR1 residue positions has evolved rapidly over the speciation of rodents. Remarkably, all of these residues correspond to the two virus binding surfaces of TfR1. We show that naturally occurring mutations at these positions block virus entry while simultaneously preserving iron-uptake functionalities, both in rodent and human TfR1. Thus, by constantly replacing the amino acids encoded at just a few residue positions, TFR1 divorces adaptation to ever-changing viruses from preservation of key cellular functions. These dynamics have driven genetic divergence at the TFR1 locus that now enforces species-specific barriers to virus transmission, limiting both the cross-species and zoonotic transmission of these viruses.

Figure 1. TFR1 evolution in rodents has been shaped by two separate host-virus arms races.

(A) A cladogram illustrates the evolutionary relationship of the rodent species analyzed. These species fall into two major families: Muridae and Cricetidae. The retrovirus (MMTV) and arenaviruses known to be harbored by these rodents in nature are also indicated. Three of the rodent arenaviruses (Guanarito, Machupo, and Junin) are zoonotically transmitted to humans. (B) Red stars represent the six rapidly evolving codon positions identified in rodent TFR1, mapped to a linear schematic of the TfR1 ectodomain. The amino acid encoded by human TFR1 at each of these positions is indicated. Residue 109 was also identified as being under positive selection (Table S1). Although potentially of functional relevance, this residue lies outside of the structure of the TfR1 ectodomain and therefore was not analyzed further in the current study. (C) Residue positions under positive selection are indicated in red on the structure of human TfR1 (PDB 1CX8) . TfR1 is a homodimer, and the six sites of positive selection are indicated on the outer edge of each monomer. Known binding regions on TfR1 for Machupo virus GP and MMTV Env are indicated in gray and blue, respectively, and the small region where they overlap is indicated with crosshatching. To the right is shown a rotated view of one edge of the TfR1 dimer.

Figure 2. Residues under positive selection on the receptor binding surface of Machupo virus GP1.

(A) A diagram of the Machupo virus surface glycoprotein precursor protein, GP. This protein is cleaved into three subunits: the stable signal peptide (SSP), the receptor-binding component GP1, and a transmembrane component GP2. The 16 residue positions that directly contact TfR1, as defined previously , are shown with blue lines positioned at the bottom of the diagram. An alignment of codons 24–256, spanning part of the SSP and almost all of GP1, was analyzed for codons with dN/dS>1. Residues corresponding to dN/dS>1 codons are indicated with red lines positioned above the diagram. Asterisks indicate four residues that both directly contact the receptor and are under positive selection. (B) A maximum likelihood tree of the 13 Machupo virus sequences analyzed. All of these viruses were isolated in Bolivia, in the years and regions indicated, from either humans (bold sequences) or Calomys mice (other sequences). The tree is unrooted. (C) The crystal structure of Machupo GP1 (PDB 2WFO) showing residues that contact TfR1 (blue), residues under positive selection (red), and the four residues that both contact TfR1 and are under positive selection (purple).

Figure 3. Fossil MMTV-like endogenous retroviruses (ERVs) identified in divergent rodent taxa.

(A) A diagram of the MMTV genome is shown, with genes drawn on three levels to indicate different reading frames. Proteins produced from each gene are listed underneath. An MMTV-like ERV was found in the genome of the deer mouse, Peromyscus maniculatus. Five MMTV-like ERVs are also evident in the genome of brown rat, R. norvegicus, as has previously been noted . Multiple rat ERV copies allowed us to construct the consensus sequence approximating the sequence of the exogenous rat virus (“RMTV”) that gave rise to these ERVs. Genetic distances between each rat ERV and the RMTV consensus (0.008–0.018 substitutions/site), combined with the neutral substitution rate observed in the rat genome (0.00506 substitutions/site/MY) , support an RMTV infection of rats that lasted from 3.6 to 1.6 million years ago. A diagram of the RMTV genome is also shown. LTR, long terminal repeat. (B) A beta-retrovirus phylogeny constructed from an alignment of approximately 900 nucleotides in the region of pro-pol. In bold are exogenous viral sequences. All others are endogenous viral sequences found integrated in the genomes of the indicated host species. The tree shows that the brown rat and deer mouse (green star) ERVs discussed in the text are more closely related to MMTV than any other virus reported in GenBank. The predicted position of the ancestral RMTV virus is shown (red star). The red branches indicate a family of viruses that we refer to as MMTV-like viruses. A maximum likelihood tree is shown. On each node are bootstrap values, given as percentage of 1,000 replicates. The tree was rooted with feline immunodeficiency virus (FIV), a lentivirus that is not in the beta-retrovirus family. JSRV, Jaagsiekte sheep retrovirus; SRV4, Simian retrovirus 4; MPMV, Mason-Pfizer monkey virus. See also Figure S1. (C) A small portion of the aligned Pol protein translation is shown to demonstrate the degree of sequence similarity between the three MMTV-like viruses discussed.

Figure 4. Cellular entry of MMTV and Machupo virus is permitted by some but not all rodent TfR1 orthologs.

MDCK cells were transduced to stably express the TfR1 of various Cricetidae and Muridae species used in the evolutionary analysis, or human TfR1. An extracellular FLAG tag was added to each receptor and cell surface expression was monitored on live cells by flow cytometry. These cells were infected with GFP-encoding retroviral vectors pseudotyped with the surface glycoproteins of (A) MMTV or (B) Machupo virus. Relative entry is scored by the mean fluorescent intensity (m.f.i.) of GFP. As previously reported, the surface protein of Machupo virus mediates cellular entry through the TfR1 of the Machupo virus host species, the vesper mouse (green line), and to a lesser extent through TfR1 of zygodont (orange line) and human TfR1 (red line) . In contrast, cellular entry of MMTV was supported strictly by the TfR1 of house mouse.

Figure 5. Mutations at sites of positive selection alter MMTV entry through TfR1.

(A) A partial TfR1 sequence alignment shows the three residue positions under positive selection (highlighted in yellow) located in the MMTV binding region. Asterisks indicate completely conserved residue positions, while positions under positive selection are highly variable. The viruses that have been previously shown to enter cells via each of these receptors are also summarized (although in the case of brown rat, cellular entry of MMTV via the rat TfR1 does not lead to productive infection , consistent with TfR1 usage being a necessary but not sufficient determinant of host range in the wild). In the remaining panels, amino acids are swapped between the species indicated at these three positions under positive selection. In one case (blue graphs), these three positions in the house mouse TfR1 were altered to encode the amino acids found in the vesper mouse TfR1. In the second case (orange graphs), these three positions in the zygodont TfR1 were altered to encode the amino acids found in the house mouse TfR1. (B and D) MDCK cells stably expressing the indicated TfR1-FLAG were infected with GFP-encoding retroviral vectors pseudotyped with the surface glycoprotein of MMTV. Virus entry was scored by measuring the percentage of GFP positive cells using flow cytometry. (C and E) Cell surface expression of TfR1 (mean fluorescent intensity) measured on live cells with a fluorescently labeled α-FLAG antibody. (F) Cellular entry of retroviral vectors pseudotyped with the surface glycoproteins of three different arenaviruses (Machupo, Junin, and Guanarito). In all experiments, three replicates were performed and error bars indicate one standard deviation.

Figure 6. Mutations at sites of positive selection do not alter TfR1 association with host proteins.

Co-crystal structures of human TfR1 in complex with (A) human transferrin (1SUV) and (B) human HFE (1DE4) illustrate that sites of positive selection (red) fall at a distance from these protein–protein interaction surfaces. Thus, mutations at these sites are not predicted to affect important host-beneficial functions of TfR1. MDCK cells stably expressing wild-type and mutant TfR1 were incubated with media containing FITC-labeled iron-loaded mouse transferrin. The cells were then washed and analyzed by flow cytometry for the mean fluorescent intensities (m.f.i.) of (C) FITC-transferrin and (D) a fluorescently labeled α-FLAG antibody measuring TfR1 surface expression. In all experiments, three replicates were performed and error bars indicate one standard deviation.

Figure 7. A human SNP in TfR1 is protective against Machupo virus entry.

(A) The apical domain of human TfR1 is shown in green, in a co-crystal structure with the Machupo GP1 shown in grey (PDB 3KAS) . The βII-1–βII-2 species-specific virus binding motif (residues 204–212) is highlighted in yellow. The side chains of residue positions identified as evolving under positive selection are shown in red. A human SNP (L212V; rs41301381) has been reported at position 212, shown in blue. (B) MDCK, (C) HEK293, or (D) HEL299 cells were stably transduced to express either 212L TfR1 or 212V TfR1 (or an empty vector) and then infected with various amounts of Machupo pseudovirus. Virus entry is scored by percentage of cells that become GFP positive (+). The error bars in the HEL299 experiment are large due to difficulty in sorting these cells. Nonetheless, this pattern of relative entry between the different TFR1 alleles expressed in HEL299 cells was observed in four independent experiments (not shown). (B, C, D, right-hand panels) Relative cell-surface expression of human TfR1 variants in each cell line was measured on live cells with a fluorescently labeled α-FLAG antibody. In all experiments, three replicates were performed and error bars indicate one standard deviation.