The Lagomorph family of Leporidae (leporids) originated in the New World (Neoartics/Americas), an area which still is home to the most successful of living leporids i.e. species of Sylvilagus and Lepus. The genera Sylvilagus (cottontail rabbits) and Lepus (jack rabbits or hares) comprise numerous species, with Lepus having conquered also the Old World (Paleoartics/Afro-Eurasia) 1 . In contrast, typical Old World leporid genera tend to be monotypic, inhabiting isolated areas where many of them are listed as endangered 2 . The recent world-wide success of the European rabbit (Oryctolagus cuniculus), which in prehistoric times was confined to the Southwestern parts of the Iberian Peninsula, was largely, if not entirely, due to human activity 3 4 5 6 .

The introduction of Myxoma virus (MYXV) during the midst of last century as a method of rabbit pest control had devastating effects on populations of European rabbit with reported mortality rates approaching 100% in Europe and Australia 7 . This was in sharp contrast to the very mild pathology caused by the virus in its natural host and reservoir, i.e. species of the genus Sylvilagus 8 9 . For Lepus species, only few cases of MYXV infections were reported and experiments in France have shown that most individuals are innately resistant, reviewed in 7 . In nature, infection with MYXV occurs through bites by flying or jumping insects. Replication of virus starts in MHC-II positive dendritic-like cells at the bite lesions and is passed on to T cells of lymph nodes draining the inoculation site 10 . The pathogenesis of MYXV infection apparently depends upon the aptitude of avoiding the spreading of infected cells throughout the lymphatic system. Whereas in cottontail rabbits MYXV infection remains localized, in ‘naïve’ European rabbits (below “Rabbit”)^b, the MYXV infected cells rapidly spread to distal nodes. This results in a generalized leukocyte depletion, particular of CD4+ T cells, which leads to a systemic immunodepression with fatal outcome i.e. myxomatosis 11 12 .

Leukocyte migration and trafficking are mainly governed through interactions of a variety of chemokines with their cellular receptors 13 14 . Insights in the parasite strategies of immune evasion offer major gateways for identifying genetic components of pathways allowing Cottontail rabbit to cope with MYXV infection. Studies of different research teams have shown that this virus encodes a number of proteins that manipulate factors of the innate immune system of the host, among them proteins interfering directly with chemo-attractive functions of the CC chemokines 15 16 . It shows that these proteins have played a role during the process of coadaptation between virus and host, and most likely still do. These findings have been of cardinal guidance in the search for host genes (candidate genes) that could make the difference between susceptible vs. resistant species.

MYXV is a large double-stranded DNA virus of the poxvirus family (genus Leporipoxvirus). There have been indications that the CCR5 receptor might play a crucial role during MYXV infection, as it is the case by HIV infection in human 17 18 , although the experimental evidence for this has been disputed 19 . However as already mentioned, the variation of pathogenicity of the MYXV among leporid species does not depend upon the fact whether or not the virus can enter and replicate in the host cell, but more likely on a constellation of endogenous factors preventing or permitting the dissemination of infected cells throughout the lymphatic system 11 12 . Studies of pathways underlying the contrasting outcomes of MYXV infection may therefore contribute to a more general understanding of pathogenesis due to large DNA viruses in mammals, inclusive humans. In view of the importance of CCR2 and CCR5 receptors in HIV infection, genes controlling these receptors and their ligands might be prefigurative of such ‘candidate genes’. This led to the discovery of a gene conversion that altered the second external loop of Rabbit CCR5. This mutation occurred in the ancestral lineage of the Old World genera including Oryctolagus and Bunolagus, but not in the lineages of Sylvilagus and Lepus species 20 21 . Although these differences at CCR5 obviously do not arbitrate the entry of MYXV for lymphocytes, they might affect CCR5 related pathways of signal transduction 17 18 19 . Note that Bunolagus species being highly endangered, studying their susceptibility to myxomatosis proved impracticable 2 .

We therefore have taken a closer look at the main ligands of the Rabbit CCR2 and CCR5 receptors which are the ‘macrophage inflammatory proteins’ chemokines (MIP’s) and the ‘monocyte chemotactic proteins’ (MCP’s). The excellent recent review of the gene organization of mammalian chemokines by Nomiyama and coworkers 22 , while comprehensive by extending to non-eutherian mammals (Metatheria and Monotremata), did not include Lagomorpha (Rabbits and Hares). Indeed, chemokine data on Rabbit are incomplete and sometimes erratic (see below). The Rabbit Genome Project being recently completed at the Broad Institute at 7x coverage 23 , we have assessed the Rabbit Whole Genome Sequence (WGS) data for orthologs of the mammalian genes of MIP-RANTES and MCP-Eotaxin. Our analyses based on nucleotide sequence similarity revealed that Rabbit possesses proper orthologs of three MCP encoding genes (CCL7, CCL8, and CCL13) which are not identified by gene finder methods used by GenBank. The non-annotation can indicate that in Rabbit these genes may have acquired singularities hampering transcription or disqualifying them as functional proteins. We have searched for such traits and, at the event, verified their presence or absence in species of the leporid genera Oryctolagus, Bunolagus, Sylvilagus and Lepus.

The genes of the CC chemokine ligands (CCL) RANTES/CCL5, MIP-1a/CCL3, and MIP-1b/CCL4 are documented for Rabbit (Oryctolagus cuniculus) [GenBank: NC_013687_REGION:24922000.25085000]. They are located on chromosome 19 as a syntenic group [GenBank: NC_000017_REGION:34198000to34433100] and are in every respect (chromosomal location, gene organization, sequence similarity) orthologs of their mammalian counterparts cf. 22 . In contrast, the GenBank list of Rabbit orthologs of the mammalian MCP-Eotaxin encoding genes is limited to MCP-1/CCL2 and Eotaxin/CCL11 [Genbank: NC_013687 REGION:23720000.23798000].True orthologs of mammalian MCP-3/CCL7, MCP-2/CCL8, and MCP-4/CCL13 have not yet been identified (a print-out of the GenBank Features report is shown in Additional file 1). This is surprising because at least the first two chemokines seem to be functional in most, if not all mammal species studied 22 , and in Human and Mouse are subject of intense investigations due to their importance in regulating inflammatory and anti-tumoral effects and for their role in HIV infection 24 25 26 27 28 .

We will adopt the CCL nomenclature unless we are dealing with proteins. For most mammals, MCP encoding genes are organized as a syntenic group composed of the chemokine genes CCL2 CCL7 CCL11 CCL8 CCL13 CCL1 22 (in that order; CCL1 serves here only as a syntenic marker). In Human the CCL2 CCL1 encompassing syntenic group is located on chromosome 17 [GenBank: NC_000017.1_REGION:32582070.32690817]. The fragment of Rabbit chromosome 19 is homologous to this region [Genbank: NC_013687_REGION:23720000.23798000]. We will refer to it as “R-MCPgb”.

The pronounced interspecies similarity between the coding sequences allowed localizing the orthologs of the CCL2, -7, -11, -8, and -13 genes in the Rabbit genome. The exons positions are listed side by side for Rabbit and Human in Table 1. The sequence alignments can be consulted in Additional files, with annotations for the undocumented genes (CCL7: Additional file 2, CCL8: Additional file 3, CCL13: Additional file 4) or for the entire MCPgb regions as a ‘blunt' 108 kb alignment in FASTA format (Additional file 5).

^$): pos 1 corresponds to position 23720000 of NC_013687.1.

^*): pos 1 corresponds to position 32582070 of NC_000017.1.

The Rabbit genome fragment R-MCPgb containing the CCL2 and CCL1 genes is compared to its Human correlate according to the alignment in Additional file 5. The identification of the Rabbit orthologs of the Human CCL genes based on sequence similarity was checked for consistency using Genscan 29 , which furthermore provided estimates of the quality of intron splice sites (values varying between 0 and 1 are shown under the heading “P”). Positions of transcription initiator and termination sites were estimated by homology with Human. The exons annotated in the GenBank feature file of the Rabbit sequence (Additional file 1) are marked “v” under “GenB”.

We show that both MCP-2/CCL8 and CCR5 were altered in lineage of Oryctolagus and Bunolagus due to apomorphic mutations which did not occur in the lineage comprising species of Sylvilagus and Lepus. It implies that the species known to be reservoirs of MYXV, i.e. S. brasiliensis and S. bachmani, dispose most likely of normal (plesiomorphic) CCR5 and CCL8 genes. Although such deductions are in line with current phylogenetic inference, they might be worthwhile to be verified, as falsification would imply the independent alteration of one or both of these genes in lineages separated in space and time for millions of years. Meanwhile we will assume that the CCR5 and CCL8 genes of the different Sylvilagus species do not differ significantly from those shared by the New World rabbits of this study.

The situation revealed by the presented data might orientate research towards the role of chemokine and their receptors in host species of MYXV and could lead to new insights in processes of parasite-host coadaptation. In mice MCP-2/CCL8 and CXCL12 were found to cooperate to attract hematopoietic progenitors of immune-regulatory dendritic cells 43 while Islam and coworkers 44 describe mouse MCP-2/CCL8 as crucial regulator of T(h)2 cell homing. In Human, MCP-2 is known to bind to chemokine receptors CCR1, CCR2 and CCR5 and can act as a potent inhibitor of HIV-1 infectivity 44 45 46 . More in particular, compared to other chemokines, MCP-2 was found the most efficient inhibitor of the HIV protein gp120 for CCR5 receptor binding 47 48 . Studies of the role of MCP chemokines in Human and Mouse revealed that inflammation is regulated by feedback mechanisms where proteases play an important role 24 33 34 35 . They act by removing the N-terminal tetra or penta peptide of MCP’s which can transform them into MCP antagonists. Different research groups reported that natural occurring posttranslational modified MCP-2/CCL8 products can completely (sic) block the chemotactic effects of intact MCP’s and of RANTES/CCL5, and have identified natural MCP-2(6-76) (cf. Figure 7) as a potent and functional CC chemokine inhibitor 28 33 34 . These studies have highlighted the role played by MCP-2 products in the subtle agonist-antagonist interplay with CC chemokine receptors, including CCR5.

In this context our data might fuel speculations about possible reasons underlying the permanent loss in Old World rabbits of an important gene function, which has been well preserved throughout mammalian evolution (cf. Figures 3 and 4; a CCL8-like gene has also been reported in bony fish [GenBank: BT048349]). One explanation could be that functions of CCL products can be redundant or interchangeable, at least in leporids, which would be at odds with the evolutionary perpetuation of CCL gene identity. More interesting is the hypothesis of a causal link between the appearance within the lineage of Old World rabbits of the alteration at the second external loop of CCR5 on one hand, and the pseudogenization of one of its prime ligands on the other. Both events are highly unusual and none of them did occur in the Sylvilagus-Lepus lineage, nor in any other studied species. Although the argument is somewhat circular - we looked at the CCL8 gene precisely because of the CCR5 alteration - it can offer a plausible explication for the knock-out of an otherwise prominent gene function over a period of more than 4 My (cf. Figure 3). If “lost by accident”, the CCL8 gene could during this period have been repaired by back-mutations or by gene conversion with one of its neighbors. One might consider that receptor alteration occurred first, making CCL8 either useless or even detrimental, allowing or forcing its permanent pseudogenization. Or on the contrary, the CCL8 gene knock-out, and the consecutive perturbation of CCR5-dependent signaling pathways (e.g. due to the loss of a potential (ant)agonist of other CCR5 ligands such as Rantes/CCL5), may have favored structural change at its orphaned target (i.e. at the second external loop of CCR5). Regardless of the scenario, we could be facing an irreversible situation where a “gene knock-out” resulted in a gene “lock-out”. Indeed gene repair would not be favored by selection if the recovered ligand can no longer recognize its receptor (why repair the key if the lock has changed). It might therefore be interesting to know to what extent, if at all, the receptor mutation impairs the affinity of the different CCR5 ligands.

A further question beyond our competence is whether, when, and in which cellular environment CCL8 is expressed in Sylvilagus during MYXV infection and how, at the event, it contributes to the clearance of infected lymphocytes. In this context it might be interesting to mention that MCP-2 expression was down regulated in human HIV infected brain cells by miRNA146a 49 , a microRNA which we found to be present in Rabbit (the miRNA146a sequences are identical among Human and Rabbit and are localized in same chromosomal region; WvdL unpublished observations).

The large number of host-specific immunodulatory proteins encoded by MYXV implies multiple levels of elaborate interactions between the virus and its natural host which can be the outcome of thousands of years of co-adaptive evolution. Identifying the constituents of this interplay remains a huge challenge, as host factors involved can be even more numerous. It might therefore be worthwhile to consider that the knock-out of a single host factor could severely affect this virus-host equilibrium. Given that monocyte chemotactic proteins control patterns of leukocyte migration, which in turn govern the outcome of MYXV infection in rabbits, the observation of a factual correlation between the near absence of MYXV virulence and the concurring presence of a functional MCP-2/CCL8 and an ‘intact’ CCR5 gene, could promote studies on the role played by this particular chemokine ligand-receptor interaction in keeping Myxoma virus under control.

Tissue samples specimens of leporid species belonging to genera Oryctolagus, Lepus and Sylvilagus were provided by CIBIO Lagomorph Tissue Collection maintained by Paulo C. Alves (CIBIO, Vairão, Portugal; http://pcalves@mail.icav.up.pt). Species and sample names are listed in Table 2. All samples are from wild populations.

^c) sources of cDNA, all others were sources of gDNA.

a) MCP-2/CCL8 and similar. Read: either the Monocyte chemotactic

protein

gene

b)

R

H

C

MIP: Macrophage inflammatory proteins; MCP: Monocyte chemotactic proteins; R-MCPgb: Rabbit MCP-Eotaxin WGS fragment [Genbank: NC_013687 REGION: 23720000.23798000]; MYXV: Myxoma virus; MHC-II: Major Histocompability Complex Class 2; HIV: Human Immunodeficiency Virus; CCL: CC chemokine ligand; CCR: CC chemokine receptor; WGS: Whole Genome Sequence; i.e.: In extenso, more in detail; e.g.: Exempli gratia, for example; cf.: Confer; indel: Insert or deletion in sequence comparisons.

The authors declare that they have no competing interests.

WvdL conceived the study and its design and carried out the literature research, data mining and analysis, drafting and editing of the manuscript. The most important findings of present report were nevertheless produced by SA and ALM, who provided the new data which are the core of the paper. SA did the PCR amplifications and sequencing, revealing the existence of a potentially functional CCL8 gene in Lepus and Sylvilagus species. ALM contribution was the cDNA work, putting the corner stone to this study by showing that in these species the CCL8 gene is transcribed. PJE is the leader and JA one of the most inspiring members of the CIBIO Evolutionary Immunogenetic Group, and played an important role in coordinating and supporting the work of ALM and SA, and by critical commenting and stimulating discussion of the manuscript.