The interferon-inducible p47 (IRG) GTPases in vertebrates: loss of the cell autonomous resistance mechanism in the human lineage

Abstract

Background: Members of the p47 (immunity-related GTPases (IRG) family) GTPases are essential, interferon-inducible resistance factors in mice that are active against a broad spectrum of important intracellular pathogens. Surprisingly, there are no reports of p47 function in humans.

Results: Here we show that the p47 GTPases are represented by 23 genes in the mouse, whereas humans have only a single full-length p47 GTPase and an expressed, truncated presumed pseudo-gene. The human full-length gene is orthologous to an isolated mouse p47 GTPase that carries no interferon-inducible elements in the promoter of either species and is expressed constitutively in the mature testis of both species. Thus, there is no evidence for a p47 GTPase-based resistance system in humans. Dogs have several interferon-inducible p47s, and so the primate lineage that led to humans appears to have lost an ancient function. Multiple p47 GTPases are also present in the zebrafish, but there is only a tandem p47 gene pair in pufferfish.

Conclusion: Mice and humans must deploy their immune resources against vacuolar pathogens in radically different ways. This carries significant implications for the use of the mouse as a model of human infectious disease. The absence of the p47 resistance system in humans suggests that possession of this resistance system carries significant costs that, in the primate lineage that led to humans, are not outweighed by the benefits. The origin of the vertebrate p47 system is obscure.

Figure 1

Genomic positioning and phylogenetic relationship of mouse Irg GTPases. (a) Disposition of the 23 Irg genes on the mouse karyotype. Individual Irg genes are listed in correct gene order in each cluster. (b) Positioning and orientation of Irg genes in the mouse chromosome 11 and 18 clusters. Positions of genes refer to the location in Mouse ENSEMBL release (v28.33d.1, February 2005) [61] of the first G of the glycine codon of the G1 motif (GKS or GMS) of the GTP-binding domain of each gene. The segments of the chromosome 11 cluster indicated with square brackets are regions of uncertain structure. Gene orientation is given by black arrows. The shaded region of the chromosome 11 map is a duplication introduced in Mouse ENSEMBL v28.33d.1 (February 2005) in an attempt to resolve a region of high ambiguity indicated by the longer square bracket. In our view this duplication does not resolve the ambiguities consistently, and we see no justification at present for the duplicated Irgb5 and Irgb6 genes. The sibling genes Irgb3 and Irgb4 differ by only nine nucleotides; in this case, however, the independent existence of the two genes is proved by the proximity of the PA28βψ retropositioned pseudo-gene to Irgb3 but not to Irgb4, in addition to consistent sequence differences. We have left the duplication of the Irgb5/Irgb6 region in the map for consistency of the base numbering with this release of ENSEMBL. *Indicates minor sequence differences presumably due to sequencing errors. (c) Unrooted tree (p-distance based on neighbour-joining method) of nucleotide sequences of the G-domains of the 23 mouse Irg GTPases, including the two presumed pseudo-genes Irga5 and Irgb7. The sources of all Irg sequences are given in Additional data file 1, and the nucleotide and amino acid sequences themselves are collected in the p47 (IRG) GTPase database from our laboratory website [14]. (d) Phylogenetic tree of the amino acid sequences of the G-domains of 21 mouse Irg GTPases rooted on the G-domain of H-Ras-1 (accession number: P01112). The products of the two presumed pseudo-genes Irga5 and Irgb7 are excluded from the analysis.

Figure 2

Genomic and promoter structure of mouse Irg GTPases. (a) Genomic structure of mouse Irg genes. Green blocks indicate coding exons and blue blocks indicate 5'-untranslated exons. Orange arrows identify putative promoter regions. Stars identify exons shown to be excluded in alternative splice forms. The scale bar is measured in base pairs up to the first base of the long coding exon. Note the presence of two promoters for Irga6 and Irgd. (b) Interferon response elements in the promoter regions of mouse Irg genes. γ-Activated sequences (GAS; pale blue blocks) and interferon-stimulated response element (ISRE; red blocks) sequences were identified in the promoters shown in panel a (also see Additional data file 7). Dark blue blocks downstream of each promoter represent the most 5' exon. The yellow block identifies a putative Sox1 transcription factor binding site in the proximal promoter region of Irgc. The scale bar is measured in base pairs from the first base of the 5' exon.

Figure 3

Interferon responsiveness of mouse and human p47 (IRG) GTPase. (a) Interferon (IFN)-γ responsiveness of eight new mouse Irg genes. Inducibility of eight further Irg genes (also see Boehm and coworkers [13]) in L929 fibroblasts induced for 24 hours with IFN-γ, demonstrated by RT-PCR. D refers to a positive control genomic DNA template; O refers to a negative control of the same genomic template after DNAse1 treatment; and + and - refer to RT-PCR on DNAse1-treated RNA templates from IFN-γ-induced and IFN-γ-noninduced cells, respectively. The sibling genes of the Irgb series could not be individually amplified because of their close sequence similarity. The identities of the amplified genes responding to interferon induction, indicated by vertical arrows, were subsequently established by sequencing of multiple clones from the PCR product. (b) Irgc is not induced by interferon or infection but is constitutively expressed in testis. (i, left) Mouse L929 fibroblasts were induced for 24 hours with IFN-β or IFN-γ or left uninduced (-). Irgc could not be detected by RT-PCR even after 50 amplification cycles in L929 cells. Irga2 after 50 cycles was used as a positive control for the interferon-induced L929 RNA. RNA from mouse testis provided a positive control for Irgc. (i, right) RT-PCR for Irgc and Irga2 (50 and 30 amplification cycles respectively) on RNA from tissues of uninfected mice (-) or mice infected 24 hours previously with Listeria monocytogenes (+). Irga2 was induced in all tissues and Irgc in none. RNA from mouse testis provided a positive control for Irgc, which is detected after 50 cycles. Testis expression of Irga2 was barely detected after 30 cycles (compare with i, left, showing Irga2 in testis after 50 cycles). (Panel ii, left) Human IRGC is not induced by 24 hours of stimulation with IFN-β or IFN-γ in human cell lines (induction of GBP-1 [accession number P32455] was assayed as a positive control) and (Panel ii, right) is constitutively expressed only in human testis. GAPDH was used as control.

Figure 4

Amino acid alignment of the mouse Irg GTPases. Sequences of all 23 mouse Irg GTPases showing the close homology extending to the carboxyl-terminus, aligned on the known secondary structure of Irga6 (indicated in blue above sequence alignment). The sequences of notional products of the two pseudo-genes Irga5 and Irgb7 have been partially reconstructed; premature terminations are indicated by red highlighting. In the C57BL/6 mouse the sequence of the Irga8 gene is damaged by an adenine insertion, indicated by the red highlighted K at position 204. (The sequence given after this point is that given after correcting the frameshift, and is identical to that of the CZECHII [Mus musculus musculus] sequence BC023105 that lacks the extra adenine.) The turquoise-highlighted M in M1 and M2 are initiation codons that are dependent on alternative splicing (also see Figure 2a); the unusual methionine residues in the G1 motif of GMS proteins are highlighted in green. The blue background Q residue of Irgb5 and Irgb2 at positions 405 and 396 indicate the point at which tandem splicing occurs to Irgb4 and Irgb1, respectively. Canonical GTPase motifs are indicated by red boxes.

Figure 5

Extended phylogeny of the G domains of IRG and related proteins. The phylogeny relates all of the IRG sequences described in this report and reveals the distinct clades on which the nomenclatural fine structure is based. All except the mouse sequences are labeled with the species of origin. Dog IRG sequences are found in the B, C, D and M clades, and human sequences only in clades C and M. The mouse and human quasi-IRG proteins, IRGQ (FKSG27), could not be included in the phylogeny because they are so deviant in the G-domain (see Figure 6 and Additional data file 6).

Figure 6

Extended alignment of the vertebrate IRG proteins. Individual sequences are given in full and are labeled as in Figure 5. Unusual residues in the G1 motif are highlighted (M of the GMS proteins in green and two deviant residues in the zebrafish irgq sequences in pink). The essential structural relationship between IRG genes and quasi-IRG genes is apparent in the alignment despite the modified G-domains. For mouse and human IRGQ the long carboxyl-terminal coding exons that contain the p47 homology were used for the alignment. In human IRGQ the sequence ENPKGESLKNAGGGGLENALSKGREKCSAGSQKAGSGEGP was removed from the alignment between positions 210 and 211 (highlighted in turquoise) to prevent extensive gap formation. The position of the intron present in pufferfish and zebrafish irgf genes is indicated by two adjacent residues highlighted in blue.

Figure 7

Synteny relationships between the human and mouse IRG genes (a) Synteny between mouse chromosome 7 and human chromosome 19 in the region of the IRGC and IRGQ genes. The figures indicate distances from the centromere in megabases. The locations of three further syntenic markers are given. Gene orientation is given by black arrows. (b) Complex synteny relationship between human chromosome 5 and mouse chromosomes 11 and 18 in the regions containing the mouse Irg genes. Figures indicate distances from the centromere in megabases. The locations of IRG genes are shown in the yellow panels. Positions of diagnostic syntenic markers are also indicated. Syntenic blocks are given in full color, and the rest is shaded.

Figure 8

Structure and expression of the human IRGM gene. (a) (left panels) RT-PCR analysis of expression of IRGM in HeLa and GS293 cells. The b and c splice variants were amplified simultaneously by the same primer pair (IRGMs1-rGMS). A different downstream primer (IRGMs1-r1) internal to all the 3' splice forms was used to show differences in the overall expression level of IRGM in the two cell lines. No RT' indicates that no reverse transcriptase is included in cDNA preparation. The band immediately below the IRGMc band in GS293 cell material, indicated with an asterisk, is a nonspecific band amplified only in this cell line. The band was sequenced and is unrelated to IRGM. (right panel) Analysis of IRGM expression in human brain, liver and testis by RT-PCR. GAPDH was used as a control. (b) Five splice forms of the IRGM gene have been identified, as indicated: IRGM(a)-IRGM(e). The promoter and 5'-untranslated regions of the gene are associated with an ERV9 retroviral LTR. Scale-bar is given in base pairs.