coiled coil (domain)
RING finger (domain)
The tripartite motif (TRIM) proteins are important in a variety of cellular functions additional to anti-viral activity. We systematically analysed mRNA expression of representative TRIM molecules in mouse macrophages, myeloid and plasmacytoid dendritic cells, and a selection of CD4+ T cell subsets. We defined four clusters of TRIM genes based on their selective expression in these cells. The first group of TRIM genes was preferentially expressed in CD4+ T cells and contained the COS-FN3 motif previously shown to be involved in protein interactions. Additional TRIM genes were identified that showed up-regulation in macrophages and dendritic cells upon influenza virus infection in a type I IFN-dependent manner, suggesting that they have anti-viral activity. In support of this notion, a subset of these TRIM molecules mapped to mouse chromosome 7, syntenic to human chromosome 11, where TRIM family members such as TRIM5, shown to have anti-viral activity, are localized. A distinct group of TRIM was constitutively expressed in plasmacytoid dendritic cells independently of viral infection or signalling through the type I IFN receptor. Our findings on expression and regulation of TRIM genes in cells of the immune system that have different effector functions in innate and adaptive immune responses, may provide leads for determining functions of this diverse family of molecules.
The tripartite motif (TRIM) family of proteins are characterized by the presence of a RING finger (R) domain, one or two B-boxes (B) and a coiled coil (CC) domain 1. The R domain of many TRIM has been shown to have E3 ubiquitin ligase activity, whereas the B and CC domains may be involved in protein interactions and homo/heterodimerization 2. The RBCC domain structure can be followed by a variety of C-terminal domains. Most commonly present is the B30.2 domain, a conserved region thought to be involved in protein-protein interactions or RNA binding 3, also suggested to confer anti-viral activity by direct interaction with retroviruses 4, 5. The B30.2 domain can be sometimes associated with a fibronectin (FN)3 motif, which is also involved in molecular recognition 6.
TRIM proteins can have a variety of cellular functions including cell proliferation, differentiation, oncogenesis, apoptosis, immune signalling and have been shown to play a role in development, and implicated in autoimmune disease 2. Furthermore, a number of TRIM proteins have been found to have anti-viral activity and are possibly involved in innate immunity 7. TRIM5α directly blocks replication of the N-tropic murine leukaemia virus in humans, and of HIV-1 in the African green monkey and macaques by a B30.2-dependent mechanism 8, 9. However, TRIM19 (or PML), which is known to be involved in acute promyelocytic leukemia 10, lacks a B30.2 domain but also inhibits replication of a wide range of viruses including influenza by as yet unknown mechanisms 11. Since TRIM19 has also been shown to regulate cytokine signalling 12, 13, this may suggest indirect mechanisms for anti-viral activity. Although some TRIM have been shown to be up-regulated in a variety of cell types by addition of type I interferons (IFN), a requirement for type I IFN for the up-regulation of TRIM expression in response to viruses has not been shown 14–18. Interestingly, TRIM25 is itself required for optimal induction of type I IFN by signalling through retinoic acid-inducible gene I, which may explain some of its anti-viral activity 19.
Despite their potential roles in the immune response to viruses, little is known regarding expression of TRIM genes in primary cells of the innate or adaptive immune systems. Macrophages and dendritic cells (DC) initiate immune responses against pathogens following activation by microbial products through pattern recognition receptors including Toll-like receptors (TLR) 20. Mouse macrophages, myeloid DC (mDC) and plasmacytoid DC (pDC) respond to single-stranded RNA encoded in viruses like influenza and also to unmethylated (CpG) oligonucleotides, abundantly present in microbial genomes, to induce different levels of cytokines 21. pDC are highly specialized cells that are major producers of type I IFN upon viral infection 22, 23. Some TRIM genes (TRIM14/PUB, TRIM20/MEFV, TRIM35/MAIR) have been reported to be expressed in macrophages and suggested to be involved in immune responses 24–26, but very little has been reported with respect to TRIM expression in DC 27.
Once activated by microbes and/or their products, DC then activate T cells, including CD4+ T cells of the adaptive immune response, and induce their differentiation into CD4+ T helper (Th) effector cells. These include Th1 cells producing the hallmark cytokine IFN-γ, essential for eradication of intracellular pathogens 28, and Th2 cells producing the key cytokines IL-4, IL-5 and IL-13, important for anti-helminth responses 29. These effector responses are regulated by naturally occurring Foxp3+ CD25+ T regulatory cells (CD25+ Treg) that do not produce pro-inflammatory cytokines upon in vitro stimulation, and/or by the immunosuppressive cytokine IL-10, which can be produced by many cells of the immune system including antigen-driven Treg (IL-10 Treg) 30. Little is known about the expression and potential function of TRIM proteins in these CD4+ cellular subsets.
We hypothesised that expression of TRIM mRNA in effector cells of the immune system, which produce different cytokines, may help to predict the potential function of TRIM protein in innate and adaptive immune responses. We defined four clusters of TRIM genes on the basis of their distinct expression in either CD4+ T cells or macrophages and DC. A group of TRIM genes was preferentially expressed in CD4+ T cells. TRIM members were further subdivided into clusters on the basis of their up-regulation by influenza viruses via a type I IFN-dependent mechanism in macrophages and DC, or in contrast their constitutive expression in pDC independently of type I IFN production. This knowledge of patterns of expression and regulation of TRIM genes in immune cells may help to uncover strategies for determining their function.
TRIM are differentially expressed in CD4+ T cell subsets, macrophages and DC
To determine whether TRIM mRNA followed a particular pattern of expression which may help to predict their function, we systematically analysed the expression of TRIM family members in mouse CD4+ T cell subsets, macrophages and DC, which produce different patterns of cytokines. For this study, we selected 29 mouse TRIM genes, spread across the mouse genome, including representatives located in clusters on different chromosomes (chr) (Supporting Information Fig. 1) and representing different subfamilies based on their C-terminal domain composition 1, 31. Using quantitative real-time PCR we determined the expression of TRIM genes in CD4+ naive T cells, neutral CD4+ T cells (activated CD4+ T cells without polarization), Th1, Th2, IL-10 Treg and CD25+ Treg upon TCR activation, derived as previously described 32. We also analysed macrophages, mDC and pDC derived as described 33 after infection with two different influenza viruses, A/Puerto Rico/8/34 (PR/8) and A/New Caledonia/20/99 (CAL), and stimulation with unmethylated CpG DNA.
As expected the stimulation of CD4+ T cells, macrophages and DC led to distinct cytokine profiles in the different cells (Fig. 1A–C, 32, 33). A heat map representing the relative values of TRIM mRNA expression was derived as explained in Supporting Information Fig. 2. The data in Fig. 1D show the values of TRIM mRNA detected at the peak of expression (6 h for T cells and 24 h for macrophages/DC, determined in pilot kinetic experiments).
Grouping of TRIM genes by patterns of mRNA expression identified four different clusters (C-1 to C-4; Fig. 1D). TRIM genes in cluster C-1 (TRIM9, TRIM1, TRIM18, TRIM46, TRIM16) showed high expression in CD4+ T cells with much lower to undetectable levels in macrophages and DC. TRIM9 showed a unique pattern of expression in resting IL-10 Treg that was not further up-regulated after TCR triggering; however, up-regulation was also observed in Th2 cells, suggesting that its expression may not be attributable to cytokines produced by the different cells after activation. TRIM1 expression, however, was up-regulated in IL-10-producing T cells upon stimulation (Th2 and IL-10 Treg) (Fig. 1D, C-1) and less so in Th1 cells, which produce lower amounts of IL-10 upon stimulation, suggesting a correlation with the IL-10 mRNA expression profile in these different effector T cell subsets (Fig. 1C, 32). However, TRIM1 expression was also induced upon stimulation of CD25+ Treg, which, although they have the capacity to produce IL-10 during in vivo inflammatory conditions, do not produce IL-10 subsequent to in vitro stimulation.
Strikingly, the COS-FN3 domain was contained only in TRIM in C-1 (Fig. 1D) (TRIM9, TRIM1, TRIM18, TRIM46) and not in TRIM in clusters C-2 to C-4, suggesting that the COS-FN3 domains may provide these TRIM genes with specific characteristics which co-ordinately regulate functions in CD4+ T cells. Although TRIM16 was also preferentially expressed in CD4+ T cells, this TRIM protein does not possess an R domain, suggesting that its regulation of expression in T cells may differ from the COS-FN3-containing TRIM molecule.
A distinct group of TRIM shown in clusters C-2 and C-3 (Fig. 1D) was most highly expressed in macrophages and DC (TRIM2, TRIM6, TRIM3, TRIM20, TRIM35, TRIM25, TRIM14, TRIM45, TRIM19, TRIM23, TRIM21, TRIM30, TRIM26, TRIM34, TRIM8), and these TRIM genes were further up-regulated in response to influenza virus infection. However, this group was further subdivided since a number of TRIM genes were additionally expressed in naive CD4+ T cells and CD25+ Treg (Fig. 1D, C-3; TRIM14, TRIM45, TRIM19, TRIM23, TRIM21, TRIM30, TRIM26, TRIM34, TRIM8) but did not show an increase in expression upon stimulation. Of note, expression of TRIM genes in C-2 and C-3 was especially high under conditions which induced the highest levels of type I IFN (Fig. 1A). CpG stimulation of macrophages and DC resulted in lower production of IFN-β as compared to infection with viruses (Fig. 1A). This correlated with lower induction of TRIM expression, suggesting a possible role of type I IFN in this up-regulation. However, expression of TRIM2 and TRIM6 was restricted to macrophages and mDC and was low to undetectable in pDC, even upon stimulation, despite the fact that pDC are the highest producers of type I IFN.
Cluster C-4 comprised a group of TRIM (TRIM24, TRIM27, TRIM28, TRIM37, TRIM39, TRIM65, TRIM68, TRIM44, TRIM59) which were expressed in pDC at high levels prior to stimulation, and for the most part their expression was either down-regulated or not affected upon stimulation. Only low levels of expression of these TRIM molecules were seen in macrophages and mDC (Fig. 1D, C-4) in the presence or absence of stimulation. Thus macrophages, mDC and pDC appear to have different intrinsic capacities to express particular TRIM genes, regardless of the cytokines produced upon stimulation with microbes such as viruses and their products. This cluster of TRIM (Fig. 1D, C-4) was also expressed constitutively in naive and CD25+ Tregs.
B30.2 domains are present in members of the TRIM family, in addition to other molecules of diverse function, and play a major role in protein-protein interactions 34. Furthermore, a group of TRIM containing B30.2 domains have been suggested to have evolved to restrict viral infection 35. We show here that TRIM molecules containing the B30.2 domain organization are not restricted to any particular cluster of expression, but are spread throughout C-1 to C-4 inclusively (Fig. 1D, solid circles). These findings support previous reports that B30.2 domains are responsible for a broad set of cellular functions in addition to anti-viral restriction.
In summary, TRIM genes that possess a COS-FN3 motif are highly expressed in T cells with little to no expression in macrophages and DC (C-1). Another group of TRIM were expressed in macrophages and DC and up-regulated by viral infection and CpG, in correlation with type I IFN production (C-2 and C-3). These TRIM were further divided based on additional expression in CD4+ T cells (C-3). Finally, a distinct cluster of TRIM (C-4) was expressed constitutively at a high level in pDC with very low expression in mDC and macrophages, and expression of these TRIM was not further up-regulated by viruses.
Type I IFN-dependent and -independent expression of TRIM in macrophages and DC
Our data show that the expression of TRIM genes in clusters C-2 and C-3 is up-regulated in macrophages and DC upon viral infection and CpG (Fig. 1D), and that the level of up-regulation appeared to correlate with the induction of type I IFN production (Fig. 1A). In contrast, a distinct group of TRIM molecules comprising C-4 was expressed constitutively at high levels in pDC (Fig. 1D), and yet for the most part was not further up-regulated by viruses, although high levels of type I IFN were induced in these cells (Fig. 1A). To investigate potential mechanisms for regulation of TRIM expression in these cells, macrophages, mDC and pDC were obtained from mice lacking the type I IFN-αβ receptor (IFN-α/βR–/–) and their TRIM expression was compared to equivalent wild-type (WT) cells obtained from control mice, under the conditions of stimulation described earlier.
Expression of TRIM within clusters C-2 and C-3, inducible upon virus infection or CpG stimulation in macrophages and DC, was completely dependent on type I IFN production since their expression was not up-regulated in the IFN-α/βR–/– cells, regardless of the level of expression (Fig. 2; C-2, C-3). Although expression of TRIM20 and TRIM35 was up-regulated upon stimulation, in contrast to the rest of the TRIM in C-2 and C-3, these increases were not completely dependent on type I IFN (Fig. 2, C-2, asterisks and black box). It is worth noting that production of cytokines including IL-12, TNF, IL-10 and IFN-β was not significantly affected in IFN-α/βR–/– macrophages and mDC infected with influenza virus as compared to controls (data not shown). However, an autocrine role of type I IFN signalling was observed in pDC stimulated with CpG and influenza virus, since pDC from IFN-α/βR–/– mice produced less IFN-α, TNF and IL-12p70 as compared to controls (data not shown), consistent with previous reports 36, 37.
In keeping with our findings on the up-regulation of TRIM expression in C-2 and C-3 by virus and CpG, stimulation with LPS and dsRNA [poly(I:C)] (data not shown), which signal via additional or different intracellular adaptor proteins downstream of TLR to produce type I IFN 20, also led to up-regulation of these TRIM in macrophages and mDC via a type I IFN-dependent mechanism. Collectively these data show that expression of the majority of TRIM molecules within clusters C-2 and C-3 is exclusively dependent on type I IFN signalling regardless of whether macrophages and DC are infected by different influenza virus strains or stimulated with TLR ligands. In contrast, the TRIM molecules contained within cluster C-4 were constitutively expressed at high levels in pDC, and were not affected by a complete absence of signalling through the type I IFN receptor.
TRIM located on chromosome 7 are up-regulated in macrophages/DC by type I IFN
We have defined clusters of TRIM (C-1 to C-4) based on their levels of expression in the different cell types (Fig. 1D) and up-regulation in macrophages and DC in a type I IFN-dependent manner (Fig. 2). To determine whether TRIM genes may have co-evolved, we searched for co-regulation of expression of closely linked TRIM genes. Although TRIM proteins are spread across the human genome, previous studies have suggested that these proteins have evolved by gene duplication leading to groups of closely related TRIM on individual chr which may share functional similarities 1, 34, 38. Similarly, mouse TRIM are found on almost all of the mouse chr and groups of closely related TRIM are observed on chr-7, in the MHC region on chr-17, and less closely related groups on chr-11 and chr-3 (Fig. 3; Supporting Information Fig. 1) 35, 39.
TRIM genes mapping to mouse chr-3, chr-11 and chr-17 were expressed broadly in the different cell types with no distinct pattern of expression (Fig. 3C). In contrast, the majority of TRIM examined that map to mouse chr-7 (TRIM3, TRIM6, TRIM21, TRIM30 and TRIM34) showed expression in macrophages and DC but not in T cells following stimulation (Fig. 3D). Up-regulation of these TRIM in macrophages and DC by viruses was completely inhibited in cells deficient in type I IFN signalling (Fig. 3E). TRIM6, TRIM21, TRIM30 and TRIM34 are the most closely related and group tightly in the F2 region of mouse chr-7 (Fig. 3A, B). The fact that these TRIM are phylogenetically related and are co-regulated by type I IFN suggests that they may have co-evolved to co-ordinate important anti-viral functions. In keeping with a role in anti-viral function, TRIM in this region (F2) of mouse chr-7 show high sequence similarity with TRIM in an equivalent region (p15.4) of the syntenic human chr-11, which have been demonstrated to have anti-viral activity 8, 18, 40, 41.
We have defined four clusters of TRIM molecules on the basis of their distinct expression in either CD4+ T cells or macrophages and DC, which have different innate and adaptive immune functions to an extent determined by their cytokine profile. A group of TRIM genes was preferentially expressed in CD4+ T cells and exclusively contained the COS-FN3 motif associated with protein-protein interactions. Additional clusters of TRIM were defined on the basis of their up-regulation by influenza viruses via a type I IFN-dependent mechanism in macrophages and DC, suggesting that this large group of TRIM may have anti-viral functions. Conversely, a distinct group of TRIM genes was constitutively expressed in pDC independently of type I IFN production.
TRIM1, TRIM9, TRIM18 and TRIM46 (Fig. 1D, C-1) were highly expressed in CD4+ T cells, but less so or not at all in macrophages and DC, and exclusively contained the COS-FN3 motif which has been reported to bind microtubules 31. Thus, TRIM18, which has also been implicated in signalling pathways 42, 43, may be involved in immune function in addition to its previously reported role in microtubule dynamics in the context of the development of the ventral midline 44. Based on our findings that the COS-FN3 domain was only found in TRIM expressed in T cells, it is of interest to speculate that TRIM containing this motif may have similar signalling functions in T cells. Similarly the function of TRIM9, since it is highly expressed in activated Th2 and in IL-10 Treg, may not be restricted to its reported role in the central nervous system 45.
Human TRIM1 has been shown to inhibit N-tropic murine leukaemia virus replication 9. Our findings that mouse TRIM1 is mainly expressed in CD4+ T cells and not macrophages and DC suggests that it may act as a restriction factor specifically in T cells, perhaps explaining the relative increased resistance of T cells to retroviral infection 46, 47. Alternatively, our data may indicate that TRIM1 may have additional functions to anti-viral activities.
We define two clusters of TRIM genes (C-2 and C-3) based on their preferential induction in macrophages and DC upon infection with influenza virus via a type I IFN-dependent mechanism. Unlike the majority of TRIM genes in C-2 and C-3, expression of TRIM20 and TRIM35 in macrophages and DC was not exclusively dependent on type I IFN, in keeping with previous reports that expression of these TRIM can be up-regulated by TNF or IL-10, or M-CSF, respectively 25, 26. TRIM in cluster C-3 were distinguished from those in C-2 since they were also expressed in certain CD4+ T cell subsets, albeit to a lower extent (Fig. 1D). TRIM8 and TRIM21, which we show here to fall in cluster C-3, have previously been shown to be expressed in T cells and suggested to be specifically involved in signalling pathways required for IFN-γ and IL-2 production 48–50. Our data that these TRIM genes are also expressed in macrophages and DC, which do not produce these cytokines, suggest that they may have a broader function in innate and adaptive immune responses.
We show that TRIM genes containing a B30.2 domain are not confined to a particular cluster defined by their expression and/or up-regulation by viral infection via type I IFN, in line with a broad function of B30.2 domains in protein-protein interactions. Indeed, the B30.2 domain can be found in proteins that belong to ten different families additional to the TRIM family, some of which have been shown to play a role in signalling in immune cells and proposed to have been selected as a component of immune defence 34.
Expression of TRIM14, TRIM19, TRIM21, TRIM25, TRIM26 and TRIM34 (all in C-2 and C-3) has previously been shown to be up-regulated by influenza virus infection in a human epithelial cell line and it was inferred that this up-regulation was caused by production of type I IFN 51. Furthermore, expression of TRIM8, TRIM19, TRIM20, TRIM21, TRIM25, TRIM30 and TRIM34 (all in C-2 and C-3) was earlier shown to be up-regulated upon addition of either type I or type II IFN to a variety of cultured cells 15, 17, 18, 24, 48, 49, 52. However, an exclusive requirement for type I IFN in the induction of TRIM expression has not been addressed.
We now show that the expression of a large number of TRIM genes (C-2 and C-3) is up-regulated by influenza virus and TLR ligation in macrophages and DC, for the most part via a type I IFN-dependent mechanism (Fig. 2). Macrophages and DC are susceptible to influenza virus infection but are known to limit productive viral replication 53, 54, using a number of IFN-inducible anti-viral proteins including Mx, PKR and possibly TRIM19 11, 55, 56. We propose that the large group of TRIM that we defined as clusters C-2 and C-3 may all function to limit viral replication in macrophages and DC, possibly by different mechanisms. For example, TRIM25 by ubiquitinating retinoic acid-inducible gene I contributes to the signalling pathway required for IFN-β production 19. Our findings that induction of TRIM25 by influenza virus infection is exclusively dependent on type I IFN in macrophages and DC indicates, however, that a tight autocrine loop is necessary for TRIM25 expression and IFN-β production in these cell subsets.
Our results suggest that macrophages, mDC and pDC may have different intrinsic capacities to express certain TRIM molecules regardless of their cytokine profile (Fig. 1D, C-4). For example, those in C-4 were expressed constitutively in pDC at high levels as compared to macrophages and mDC and yet their expression was not up-regulated by viruses, nor dependent on signalling by type I IFN (Fig. 2, C-4). This begs the question as to their function, particularly since pDC have been strongly implicated as dominant in the innate immune response to limit viral infections as a result of their secretion of large amounts of type I IFN upon viral infection or stimulation with viral products 57, 58. Therefore expression of these TRIM genes in pDC may reflect additional anti-viral function of TRIM molecules in this specialized cell type. Our observations that this cluster C-4 of TRIM genes is also expressed in T cells, may reflect the close relationship suggested between pDC and lymphoid cells from observations that pDC express a number of markers of the lymphoid lineage 59.
Among the large number of TRIM dependent on type I IFN for their expression upon viral infection, was a group of homologous TRIM that mapped to the F2 region on mouse chr-7 (TRIM3, TRIM6, TRIM21, TRIM30, TRIM34) (Fig. 3B), syntenic to the region p15.4 of human chr-11, containing TRIM5, TRIM6, TRIM21, TRIM22 and TRIM34 35, which have been reported to have anti-viral restriction activity 8, 18, 40, 41. Furthermore, mouse TRIM6, TRIM30 and TRIM34 are phylogenetically related to human TRIM5, TRIM6, TRIM22 and TRIM34 35. With the exception of mouse TRIM30, none of these mouse TRIM proteins have so far been shown to have anti-viral activity. Taken together our findings suggest that mouse TRIM3, TRIM6, TRIM21, TRIM30 and TRIM34 located on chr-7 may have evolved similar mechanisms for viral restriction as the human TRIM genes located on chr-11.
The human TRIM molecule located on human chr-11 have been suggested to exert their anti-viral restriction by B30.2 domain-dependent interactions 41. In keeping with this, four out of five TRIM genes (exception TRIM3) located on mouse chr-7 contain a B30.2 domain, suggesting that as in human these TRIM genes have co-evolved to restrict viruses. However, TRIM genes without the B30.2 domain, like TRIM19, can also be involved in anti-viral functions 7. Therefore, our classification of a large number of TRIM defined on the basis of their induction by viruses via a type I IFN-dependent mechanism in macrophages and DC may be a better predictor of anti-viral activity than the presence of a B30.2 domain or their chromosomal location.
Consistent with our findings that up-regulation of type I IFN determines the up-regulation of certain TRIM molecules, up-regulation of TRIM genes in human macrophages was mainly observed under conditions which resulted in the induction of IFN-β (in this case LPS and IFN-γ), as observed using GeneSpring analysis of a previously published microarray study 60 (Supporting Information Fig. 3A). In contrast, this was not observed in human macrophages stimulated with IL-4, which did not induce IFN-β production (Supporting Information Fig. 3A, indicated with an asterisk). TRIM genes, which we found to be constitutively expressed in mouse pDC (Fig. 2, C-4), were not up-regulated in human macrophages under these conditions 60 (Supporting Information Fig. 3B).
Strikingly, expression of TRIM3, TRIM5, TRIM6, TRIM21, TRIM22 and TRIM34, located on human chr-11, as previously discussed, and shown previously to have anti-viral activity 8, 18, 40, 41, is up-regulated in human macrophages under conditions that lead to induction of IFN-β 60 (Supporting Information Fig. 3A, C). This is in keeping with our data that the mouse TRIM genes located on the syntenic chr-7 are up-regulated via a type I IFN-dependent mechanism (Fig. 3), supporting our hypothesis that these mouse and human TRIM genes located in the specific regions on chr-7 and chr-11, respectively, have co-evolved to combat viruses.
The majority of the TRIM genes, which we showed in mouse macrophages to be up-regulated by virus in a type I IFN-dependent manner (Fig. 2), were also up-regulated in human macrophages producing IFN-β (Supporting Information Fig. 3A, and Table 1). Additionally, human TRIM5, TRIM17, TRIM31, TRIM33, TRIM48 and TRIM62, which map to different chr that are either non-existent in mouse or were not tested in our study, were up-regulated in human macrophages under these conditions. Taken together, our study provides data to support the hypothesis that TRIM expressed in the context of type I IFN signalling may be a broad predictor of anti-viral activity of TRIM molecules in both human and mouse.
|TRIM||Changes in TRIM expression in human macrophages by microarray (Martinez et al.60)||Changes in TRIM expression in mouse macrophages by RT-PCR (our study)|
|TRIM1||Not affected||Not affected|
|TRIM5||Up-regulated||No homologue in mouse|
|TRIM9||Not affected||Not affected|
|TRIM22||Up-regulated||No homologue in mouse|
|TRIM24||Not affected||Not affected|
|TRIM28||Not affected||Not affected|
|TRIM30||No homologue in human||Up-regulated|
|TRIM37||Not affected||Not affected|
|TRIM39||Not affected||Not affected|
|TRIM48||Up-regulated||No homologue in mouse|
|TRIM49||Not affected||No homologue in mouse|
|TRIM59||Not affected||Not affected|
In this study we have shown groups of mouse TRIM genes that are expressed either in CD4+ T cells, or alternatively a distinct group expressed in macrophages and DC. Clusters of TRIM expression were further subdivided on the basis of their up-regulation by influenza viruses via a type I IFN-dependent mechanism in macrophages and DC, or in contrast their constitutive expression in pDC independently of type I IFN production. This grouping of TRIM genes based on their expression and regulation may provide leads to delineate the potential functions of this diverse family of proteins.
Materials and methods
129Sv/Ev (WT and IFN-α/βR-knockout), BALB/c and C57BL/6 mice were used to provide macrophages and DC. BALB/c mice were used to obtain CD4+ T cell subsets. 129Sv/Ev and IFN-α/βR-knockout mice were purchased from B&K Universal Ltd. All mice were bred at the National Institute for Medical Research (NIMR, London, UK) under UK Home Office regulations and housed under specific pathogen-free conditions. Female mice were used between 8 and 12 wk of age.
Reagents and isolation of primary cell subsets
Phosphorothioate CpG DNA (CpG1018) was purchased from Invitrogen Life Technologies. Influenza A virus (H1N1) strains PR/8 and CAL were grown at NIMR. The reagents used for macrophage, DC and T cell preparation have been described before 32, 33. Cells were purified using a MoFlo flow cytometer (DakoCytomation) and differentiated into neutral, Th1, Th2 and IL-10 Treg as described 32. BM-derived macrophages were generated in L cell-conditioned medium containing M-CSF; BM-derived mDC were generated in GM-CSF (Schering Plough); BM-derived pDC were generated in FLT3 ligand (Shanghai Genomix, China) and purified by flow cytometry as described by Boonstra et al.33. The purity was always ⩾98%.
In vitro stimulation of DC and macrophages, and quantitation of cytokine production
pDC, mDC or macrophages were cultured (5×105/mL–1×106/mL) in flat-bottom culture plates (96 wells, 48 wells, 24 wells, respectively) (Nunc) as described before 33 with medium alone, CpG1018 DNA (0.5 µM) or influenza virus PR/8 or CAL at 100 hemagglutinin U/mL. After 24 h stimulation, supernatants were collected to determine cytokine levels by ELISA and the cells were used to extract total RNA. Commercially available ELISA kits were used for assay of IL-12p70, TNF, IL-10 (Ready-Set-Go; eBioscience) and IFN-β (PBL supplier). IFN-α was measured with an anti-IFN-α capture mAb (F18; Hycult), and a rabbit anti-IFN-α polyclonal Ab (PBL supplier) followed by goat anti-rabbit HRP (Sigma-Aldrich).
Real-time quantitative PCR
Generation of heat maps for data presentation
The cycle threshold of each TRIM was normalized by the cycle threshold of HPRT to obtain the relative values as previously described 32. These values were imported to GeneSpring GX 7.3.1 (Agilent Technologies) software, and individual TRIM genes further normalized to the median of all samples for each gene. A value of 1.0 represents the median (yellow); high expression relative to the median is shown in red, and low expression in green (for further details please see Supporting Information Fig. 2).
Phylogenetic analysis of TRIM proteins
The amino acid sequences of all mouse TRIM proteins reported up to date were obtained from the Mouse Genome Informatics database (http://www.informatics.jax.org). Amino acid alignments of the full-length protein sequence of all TRIM molecules or genes were obtained using the online ClustalW server (http://www.ebi.ac.uk/clustalw) 61. Phylogenetic analysis and neighbour joining bootstrap analysis was performed on the amino acid alignment using the NJplot setting (from http://www.informatics.jax.org) and 500 replications to place all TRIM in groups according to sequence similarity. Based on this analysis we then chose representative TRIM members for analysis of mRNA expression (Supporting Information Fig. 1, asterisks). Because a high number of TRIM proteins contain the B30.2 domain and this domain may be a hotspot for evolutionary selection 35, 62, we repeated the analysis described above using the protein sequence of the RBCC motif without the B30.2 domain or any other C-terminal sequence (Fig. 3A), and still obtained the same pattern of TRIM sequence similarity.
We thank John Shoemaker and Christine Graham for input into this study, Chris Atkins, Graham Preece and Aaron Rae for cell sorting, and George Kassiotis and John McCauley for scientific input and reading of the manuscript. This work was funded by the Medical Research Council, UK, and the Sixth EU Framework Programme for Research and Technological Development (EU FP6).
Conflict of interest: The authors declare no financial or commercial conflict of interest.