Different subcellular localisations of TRIM22 suggest species-specific function
- First Online:
- Cite this article as:
- Herr, AM., Dressel, R. & Walter, L. Immunogenetics (2009) 61: 271. doi:10.1007/s00251-009-0357-z
The B30.2/SPRY domain is present in many proteins, including various members of the tripartite motif (TRIM) protein family such as TRIM5α, which mediates innate intracellular resistance to retroviruses in several primate species. This resistance is dependent on the integrity of the B30.2 domain that evolves rapidly in primates and exhibits species-specific anti-viral activity. TRIM22 is another positively selected TRIM gene. Particularly, the B30.2 domain shows rapid evolution in the primate lineage and recently published data indicate an anti-viral function of TRIM22. We show here that human and rhesus TRIM22 localise to different subcellular compartments and that this difference can be assigned to the positively selected B30.2 domain. Moreover, we could demonstrate that amino acid changes in two variable loops (VL1 and VL3) are responsible for the different subcellular localisations.
Retroviruses can be potently blocked in target cells shortly after infection as part of an innate immune response (Bieniasz 2004; Towers 2007). Such restrictions are often encountered when species borders are crossed in infections, e.g. the human immunodeficiency virus type 1 (HIV-1) is restrained in cells of catarrhine primates (human, apes, Old World monkeys), or cells of platyrrhine primates (New World monkeys) restrict immunodeficiency viruses of Catarrhini (Himathongkham and Luciw 1996; Hofmann et al. 1999; LaBonte et al. 2002; Shibata et al. 1995). Stremlau et al. (2004) showed that TRIM5α is an essential factor for the inhibition of certain retroviruses and since that, the anti-viral function of TRIM5α has been under intensive investigation. Interestingly, this TRIM5α-mediated restriction is species-specific, meaning that different viruses are blocked in different primate species (Bieniasz 2003; Sharp et al. 2001; Stoye 2002). While the TRIM5α protein of rhesus monkeys (Macaca mulatta, rhTRIM5α) potently inhibits an infection with HIV-1, the human and chimpanzee TRIM5α can only moderately block this virus (Stremlau et al. 2005).
TRIM5α belongs to the family of tripartite motif (TRIM containing) proteins, which consist of three domains: a RING, a B-box and a coiled-coil domain (Reddy et al. 1992). The RING domain is involved in protein–protein interactions and exhibits E3 ubiquitin ligase activity, which is indeed a function of some TRIM proteins (Freemont et al. 1991; Joazeiro and Weissman 2000; Nisole et al. 2005). The B-box domains have so far only been identified in TRIM proteins and might be involved in protein–protein interactions as well (Borden 1998; Nisole et al. 2005). Coiled-coil domains are generally involved in forming homo- and hetero-multimers (Reymond et al. 2001). Approximately 60% of all known human TRIM genes also encode a variable fourth domain at the carboxy terminus, the so-called B30.2 domain (Nisole et al. 2005), which was originally discovered in butyrophilin (Henry et al. 1998).
TRIM proteins form homomultimers through their coiled-coil domains and tend to form higher order protein complexes (Reymond et al. 2001). Such complexes are found in different subcellular compartments, thereby determining the specific functions of the various TRIM proteins. As already pointed out, TRIM proteins are involved in innate immune responses. The transcription of various TRIM genes, e.g. TRIM22, can be induced with type I interferons (Barr et al. 2008; Gongora et al. 2000; James et al. 2007; Orimo et al. 2000; Rajsbaum et al. 2008; Rhodes et al. 2002; Toniato et al. 2002). TRIM22 plays a role in limiting HIV-1 infections by binding to the viral gag protein and interfering with virus budding (Barr et al. 2008). Additionally, TRIM22 is able to repress viral gene transcription in HIV-1 (Tissot and Mechti 1995). TRIM34 can moderately inhibit SIVmac infection (Zhang et al. 2006) and TRIM32 interacts with tat proteins of HIV-1 and HIV-2 (Fridell et al. 1995). TRIM1 moderately restricts infections with N-MLV and overexpression of TRIM19 mediates resistance against the human cytomegalovirus, the vesicular stomatitis virus, herpes simplex virus and influenza virus A (Chelbi-Alix et al. 1998; Everett et al. 2006; Tavalai et al. 2006).
Most TRIM proteins, which were reported in an anti-viral context (e.g. TRIM1, 5, 22, 26 and 34), possess a B30.2 domain at the C-terminus. Stremlau et al. (2005) reported that the B30.2 domain of TRIM5α is responsible for HIV-1 restriction. Song et al. (2005) investigated TRIM5α and other related TRIM genes in several primate species and rodents and identified four regions in the B30.2 domain, which showed species-specific sequence and length variation, referred to as variable regions v1, v2, v3 and v4. The TRIM5α protein of primates is especially polymorphic in v1 and v3, with v1 and v3 being elongated in Catarrhini and Platyrrhini, respectively. Moreover, the v1 region contains exactly the amino acid sequence that is essential for the HIV-1 restriction in rhesus macaques (Stremlau et al. 2005).
Here, we report that human and rhesus macaque TRIM22 differ in their subcellular localisation and this difference could be assigned to the polymorphic B30.2 domain. Furthermore, we could demonstrate that polymorphic amino acid residues of two variable loops in the B30.2 domains, which are expected to mediate contacts to the binding partner, are responsible for the different subcellular localisations.
Material and methods
cDNA constructs and transfection
Full-length human and rhesus macaque TRIM22 cDNA sequences (hTRIM22, rhTRIM22) were obtained with primers 5′aag ctt gca atg gat ttc tca gta aag 3′ (forward human and rhesus macaque), 5′aag ctt cca gga gct cgg tgg gca c 3′ (reverse human) and 5′aag ctt gga gct cgg ttg gca cag 3′ (reverse rhesus macaque), which contain HindIII restriction sites (underlined). PCR products were cloned into the pGEM-T Easy PCR cloning vector (Promega). Clones were sequenced and subcloned in the pAcGFP1-N1 expression vector (Clontech) to allow for expression of a TRIM22-AcGFP fusion protein. For the expression of a V5-His-tagged TRIM22 protein, PCR products were directly cloned in pcDNA3.1/V5-His-TOPO® (Invitrogen). The constructs were transiently transfected in COS-7 or HeLa cells by using Metafectene (Biontex) according to the supplier's manual.
Establishment of chimeric constructs
The chimeric constructs hTRIM-rhB30.2-22 and rhTRIM-hB30.2-22 were established by exchanging the exons coding for the TRIM22 B30.2 domains of human and rhesus macaque. The exons were excised from the expression constructs using ScaI and SacII restriction enzymes. In the rhesus macaque construct, an internal ScaI restriction site had to be generated using site-directed mutagenesis (QuikChange site-directed mutagenesis kit, Stratagene), which did not change the amino acid sequence. The excised exons were ligated to the opposed remaining TRIM22 constructs.
As TRIM22 possesses an internal BamHI restriction site, the chimeric constructs hTRIM-rhB30.2(368-498)-22 and rhTRIM-hB30.2(368-498)-22 could be established by excising the B30.2 segments and ligating the excised part to the remaining opposite TRIM22 construct. All other chimeric constructs were established by using site-directed mutagenesis according to the supplier's manual (QuikChange Site Directed Mutagenesis Kit, Stratagene).
Immunoblotting and immunoprecipitation
For co-immunoprecipitation experiments HEK293T cells were transiently co-transfected with hTRIM22-V5-His and rhTRIM22-AcGFP expression constructs. Twenty-four hours post-transfection, 0.5 × 107 cells were lysed with 500 µl NP-40 lysis buffer (50 mM Tris–HCl pH 8.0, 150 mM NaCl, 1% NP-40) containing protease inhibitor cocktail (Roche). After incubation overnight at 4°C with 1 µg monoclonal anti-GFP antibody (Jl-8, BD), 50 µl of precleared protein G sepharose beads was added to the cell lysate and the suspension was incubated for an additional 2 h at 4°C on a rotator. Protein G sepharose beads were washed five times with ice-cold NP-40 lysis buffer to remove unbound protein. Proteins were eluted by adding sample buffer (1% SDS, 100 mM DTT, 50 mM TRIS pH 7.5) to the beads, and after incubation at 95°C for 10 min, proteins were resolved by SDS-PAGE and immunoblotted with horseradish peroxidase-conjugated anti-V5 antibody (Invitrogen).
Immunocytochemistry and microscopy
Transiently transfected cells (HeLa or COS-7) were grown on tissue culture coverslips (Sarstedt) for 24–48 h and fixed with 3% paraformaldehyde for 10 min. Human and rhesus macaque peripheral blood mononuclear cells (PBMC) were isolated from whole blood using Leucosep tubes (Greiner) and fixed with 1.5% paraformaldehyd for 10 min.
After three washes with phosphate-buffered saline (PBS), cells were permeabilised with 0.5% Triton X-100 in PBS for 5 min at room temperature. Non-specific binding was blocked by incubation with 5% bovine serum albumin (BSA) in PBS for 30 min at room temperature. Antibodies were diluted in 0.5% BSA in PBS; primary antibody (anti-α-tubulin, 1:5,000; polyclonal anti-TRIM22, 1:500) was applied overnight at 4°C and secondary antibody (polyclonal rabbit–anti-mouse–TRITC, DAKO, 1:40) for 1 h at room temperature. Nuclei were stained with 4′,6-diamidino-2-phenylindoledihydrochloride (DAPI; Sigma). Cells were rinsed with PBS and mounted with fluorescent mounting medium (DAKO). Fluorescent cells were analysed with a Zeiss Axio Observer confocal microscope through ×20 or ×40 objectives or with a confocal laser scanning microscope LSM 5 PASCAL (Zeiss) through ×40 or ×63 objectives.
Establishment of a mouse anti-human TRIM22 antiserum
A bacterial expression construct of human TRIM22 that lacks the coiled-coil domain was established using the prokaryotic expression vector pGEX-4T3, which allows the expression of a GST-tagged protein. After overexpression of TRIM22-GST, inclusion bodies were isolated and 100 µg was used to immunise C3H mice. Serum was obtained after three immunisations with 2-week intervals. The immunisations had been approved by the local government and were in accordance with institutional guidelines for the welfare of animals. Specificity of the antiserum was analysed by immunoblotting and immunocytochemistry. The antiserum also cross-reacts with rhesus macaque TRIM22.
Subcellular localisation of TRIM22-AcGFP fusion constructs
Recent data indicated that the B30.2 domain of TRIM22 is subject to positive selection in primates (Sawyer et al. 2007). We therefore hypothesised that TRIM22 might show species-specific differences in function, similar to what was published for TRIM5α. In a first approach, we compared the subcellular localisation of human and rhesus macaque TRIM22, as different subcellular localisations are frequently associated with functional differences.
Interaction of hTRIM22 and rhTRIM22
Differences in subcellular localisation are due to polymorphisms in the variable loops of the B30.2 domain
Localisation of TRIM22 polymorphisms in a 3D model
It was previously shown that the anti-viral TRIM5α and TRIM22 genes are subject to positive selection (Sawyer et al. 2005, 2007). In particular the B30.2 domains show signs of positive selection in the primate lineage, including amino acid changes and sequence elongations. These changes lead to species-specific anti-viral activity of TRIM5α (Sawyer et al. 2005; Song et al. 2005). Other TRIM proteins were also shown to exhibit anti-viral activity, although these activities are less specific against particular retroviruses as was reported for TRIM5α (Bonilla et al. 2002; Turelli et al. 2001; Zhang et al. 2006).
We focused our analysis on the subcellular localisation of TRIM22, because variations in localisation are likely associated with functional differences. Interestingly, we found that human and rhesus macaque TRIM22 differ in their subcellular localisation. These differences could be assigned to the positively selected B30.2 domain. Human TRIM22 localises diffuse in the cytoplasm, while rhesus macaque TRIM22 forms cytoplasmic bodies (speckles). TRIM proteins form high molecular weight complexes via the coiled-coil domain and this multimerising is necessary for the correct subcellular localisation. The RING and B-box domains are also involved in the subcellular targeting of TRIM proteins (Reymond et al. 2001). Our data clearly indicate that in addition the B30.2 domain is important for the cellular distribution, at least in TRIM22.
Human TRIM5α tend to form speckles as rhesus macaque TRIM22 does. These speckles of TRIM5α are highly dynamic structures and move along microtubules (Campbell et al. 2007). The structure of rhesus macaque TRIM22 speckles resembles the nuclear bodies formed by TRIM19 (PML), the best characterised TRIM family member. For PML, it was hypothesised that nuclear body formation is necessary for PML function (Dong et al. 2004; Zhu et al. 1997). Mutations in the B30.2 domain of TRIM9 and TRIM18 (MID1) also lead to changes in their subcellular localisation: while both proteins are normally located in microtubuli-associated structures, mutations lead to localisation into cytoplasmic speckles (Short and Cox 2006). The described mutations in the B30.2 domain of TRIM18 are associated with the Opitz syndrome (Cox et al. 2000). Reymond et al. (2001) hypothesised that TRIM domains provide the protein–protein interfaces for the recruitment of other proteins, thereby specifying cellular compartments. The rapid evolution of TRIM22, and in particular its B30.2 domain, might have led to differences in protein–protein interaction interfaces and to species-specific functional differences of TRIM22 as a means to adapt to pathogenic threats.
Song et al. (2005) identified four variable regions (v1–v4) in the B30.2 domain of TRIM proteins. For TRIM5α, several amino acid substitutions and sequence elongations were found in the v1 and v3 regions. Moreover, it was shown that those amino acids responsible for the restriction of HIV are located in the v1 region. James et al. (2007) analysed the crystal structure of the B30.2 domain of TRIM21 and found that the B30.2 domain comprises several conserved β-sheets alternating with variable loops (VL), which correspond to the v1–v4 regions identified by Song et al. (2005). These variable regions are thought to be involved in protein–protein interactions. Interestingly, most amino acid differences between human and rhesus macaque TRIM22 map to VL1 and VL3. We could narrow down the amino acids, which are responsible for the subcellular localisation, to the VL1 and VL3 regions. Those amino acid substitutions might lead to conformational changes that alter the functional properties of TRIM22. Alignment of the TRIM22 VL1 and VL3 amino acid residues with the known 3D structure of the TRIM21 B30.2 domain (James et al. 2007) suggests that these variable loops are in close contact (Fig. 7).
It has been shown by several groups that human TRIM22 is inducible by type I interferons and is able to block HIV-1 replication (Barr et al. 2008; Bouazzaoui et al. 2006; Tissot and Mechti 1995). Moreover, Barr et al. (2008) could demonstrate binding of TRIM22 to HIV-Gag, resulting in alteration of Gag's intracellular trafficking from punctuate structures near the cell membrane to diffuse cytoplasmic localisation. Interestingly, differences in the subcellular localisations were also reported for human and rhesus macaque TRIM5α (Campbell et al. 2007; Stremlau et al. 2004). In addition, the positive selection of the closely linked TRIM5α and TRIM22 genes was found to be mutually exclusive in primates and mutually exclusive presence/absence of these genes is evident in certain non-primate mammals (Sawyer et al. 2007). These data indicate that rhesus macaque TRIM5α and human TRIM22 have assumed similar anti-retroviral functions, but that either TRIM5 or TRIM22 predominate in the anti-retroviral activity of primates.
In summary, our findings showed a difference in the subcellular localisation of human and rhesus macaque TRIM22. This difference could be assigned to the B30.2 domain and was further narrowed down to few polymorphic amino acid residues in the variable loops VL1 and VL3. We hypothesise that these differences in the subcellular localisation are associated with functional variations of TRIM22 and reflect evolutionary adaptations to species-specific anti-retroviral functions.
The authors are grateful to Nicole Otto, Christiane Schwarz and Leslie Elsner for expert technical assistance. This study was supported by the Volkswagenstiftung (‘Evolutionsbiologie’) by a Ph.D. fellowship to A.H.