Limits...
IFITMs from Mycobacteria Confer Resistance to Influenza Virus When Expressed in Human Cells.

Melvin WJ, McMichael TM, Chesarino NM, Hach JC, Yount JS - Viruses (2015)

Bottom Line: Analysis of sequence elements shared by bacterial IFITMs and IFITM3 identified two hydrophobic domains, putative S-palmitoylation sites, and conserved phenylalanine residues associated with IFITM3 interactions, which are all necessary for IFITM3 antiviral activity.We also demonstrated the ability of a bacterial IFITM to co-immunoprecipitate with IFITM3 suggesting formation of a complex, and also visualized strong co-localization of bacterial IFITMs with IFITM3.However, the mycobacterial IFITMs lack the endocytic-targeting motif conserved in vertebrate IFITM3.

View Article: PubMed Central - PubMed

Affiliation: Department of Microbial Infection and Immunity, Center for Microbial Interface Biology, the Ohio State University, Columbus, OH 43210, USA. wjamesmelvin@gmail.com.

ABSTRACT
Interferon induced transmembrane proteins (IFITMs) found in vertebrates restrict infections by specific viruses. IFITM3 is known to be essential for restriction of influenza virus infections in both mice and humans. Vertebrate IFITMs are hypothesized to have derived from a horizontal gene transfer from bacteria to a primitive unicellular eukaryote. Since bacterial IFITMs share minimal amino acid identity with human IFITM3, we hypothesized that examination of bacterial IFITMs in human cells would provide insight into the essential characteristics necessary for antiviral activity of IFITMs. We examined IFITMs from Mycobacterium avium and Mycobacterium abscessus for potential antiviral activity. Both of these IFITMs conferred a moderate level of resistance to influenza virus in human cells, identifying them as functional homologues of IFITM3. Analysis of sequence elements shared by bacterial IFITMs and IFITM3 identified two hydrophobic domains, putative S-palmitoylation sites, and conserved phenylalanine residues associated with IFITM3 interactions, which are all necessary for IFITM3 antiviral activity. We observed that, like IFITM3, bacterial IFITMs were S-palmitoylated, albeit to a lesser degree. We also demonstrated the ability of a bacterial IFITM to co-immunoprecipitate with IFITM3 suggesting formation of a complex, and also visualized strong co-localization of bacterial IFITMs with IFITM3. However, the mycobacterial IFITMs lack the endocytic-targeting motif conserved in vertebrate IFITM3. As such, these bacterial proteins, when expressed alone, had diminished colocalization with cathepsin B-positive endolysosomal compartments that are the primary site of IFITM3-dependent influenza virus restriction. Though the precise evolutionary origin of vertebrate IFITMs is not known, our results support a model whereby transfer of a bacterial IFITM gene to eukaryotic cells may have provided a selective advantage against viral infection that was refined through the course of vertebrate evolution to include more robust signals for S-palmitoylation and localization to sites of endocytic virus trafficking.

No MeSH data available.


Related in: MedlinePlus

N-terminal HA-tagging of human IFITM3 does not affect its anti-influenza virus activity. (A) HEK293T cells were transfected with empty vector, or vector expressing human HA-IFITM3 or untagged human IFITM3. Cell lysates were examined by Western blotting with anti-IFITM3 antibodies and anti-GAPDH antibodies as a loading control; (B) Cells treated as in (A) were examined by flow cytometry staining with anti-IFITM3 antibodies. Cells are gated on IFITM3-positive cells and the blue numbers indicate the average of triplicate samples +/− standard deviation. Ssc, side scatter; (C) Cells treated as in (A) and (B) were infected with IAV H1N1 strain PR8 at an MOI of 1.0 or mock infected for 6 h. Cells were analyzed for percent infection by flow cytometry based on anti-IAV nucleoprotein staining in comparison to mock-infected controls. Average percent infection of triplicate samples in a representative experiment was graphed. Error bars indicate standard deviation. Results are representative of more than 5 similar experiments. NS, not significant by Student’s t test.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4488726&req=5

viruses-07-02759-f002: N-terminal HA-tagging of human IFITM3 does not affect its anti-influenza virus activity. (A) HEK293T cells were transfected with empty vector, or vector expressing human HA-IFITM3 or untagged human IFITM3. Cell lysates were examined by Western blotting with anti-IFITM3 antibodies and anti-GAPDH antibodies as a loading control; (B) Cells treated as in (A) were examined by flow cytometry staining with anti-IFITM3 antibodies. Cells are gated on IFITM3-positive cells and the blue numbers indicate the average of triplicate samples +/− standard deviation. Ssc, side scatter; (C) Cells treated as in (A) and (B) were infected with IAV H1N1 strain PR8 at an MOI of 1.0 or mock infected for 6 h. Cells were analyzed for percent infection by flow cytometry based on anti-IAV nucleoprotein staining in comparison to mock-infected controls. Average percent infection of triplicate samples in a representative experiment was graphed. Error bars indicate standard deviation. Results are representative of more than 5 similar experiments. NS, not significant by Student’s t test.

Mentions: Given the important differences and striking similarities between the mycobacterial IFITMs and IFITM3 (Figure 1B), we sought to empirically determine if MAB and MAV IFITMs could confer antiviral activity when expressed in human cells. Since these IFITMs have never before been studied, antibodies for detection of their expression do not exist. Epitope tagging has been extensively used for the study of IFITMs [4,6,20,28,29,30,35,36,37,38,39,40,41], but it has been hypothesized that epitope tags may alter IFITM3 activity [7,8,13], though data demonstrating this has not been published. Thus, we first examined N-terminally HA-tagged and untagged human IFITM3 expressed from the same vector for their ability to inhibit influenza virus infection of HEK293T cells. This human cell line was utilized for our studies because it is highly infectable and capable of supporting the full life cycle of influenza virus [42], it expresses low levels of endogenous IFITM3 (Figure 2A,B), it is highly transfectable, and it has proved useful for distinguishing subtle differences in the activities of various IFITM mutants in several studies [4,9,20,28,29,33,36]. HEK293T cells robustly expressed both HA-IFITM3 and IFITM3 as examined by Western blotting with anti-IFITM3 antibodies (Figure 2A). Analysis at the single cell level by flow cytometry confirmed that only a small percentage of HEK293T cells express detectable levels of endogenous IFITM3, and that our transfection efficiencies for the IFITM3 constructs were greater than 70% (Figure 2B). Compared to the vector control transfection, both HA-IFITM3 and IFITM3 were able to significantly inhibit infection by influenza A virus (IAV) H1N1 strain PR8 as measured by anti-IAV nucleoprotein staining by flow cytometry [4,20,33,36]. No statistically significant difference was observed in the abilities of HA-IFITM3 and IFITM3 to inhibit infection (Figure 2C). Thus, we reasoned that utilization of HA-tagging of the mycobacterial IFITMs could provide an effective way to measure their expression without a likely effect on activity.


IFITMs from Mycobacteria Confer Resistance to Influenza Virus When Expressed in Human Cells.

Melvin WJ, McMichael TM, Chesarino NM, Hach JC, Yount JS - Viruses (2015)

N-terminal HA-tagging of human IFITM3 does not affect its anti-influenza virus activity. (A) HEK293T cells were transfected with empty vector, or vector expressing human HA-IFITM3 or untagged human IFITM3. Cell lysates were examined by Western blotting with anti-IFITM3 antibodies and anti-GAPDH antibodies as a loading control; (B) Cells treated as in (A) were examined by flow cytometry staining with anti-IFITM3 antibodies. Cells are gated on IFITM3-positive cells and the blue numbers indicate the average of triplicate samples +/− standard deviation. Ssc, side scatter; (C) Cells treated as in (A) and (B) were infected with IAV H1N1 strain PR8 at an MOI of 1.0 or mock infected for 6 h. Cells were analyzed for percent infection by flow cytometry based on anti-IAV nucleoprotein staining in comparison to mock-infected controls. Average percent infection of triplicate samples in a representative experiment was graphed. Error bars indicate standard deviation. Results are representative of more than 5 similar experiments. NS, not significant by Student’s t test.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4488726&req=5

viruses-07-02759-f002: N-terminal HA-tagging of human IFITM3 does not affect its anti-influenza virus activity. (A) HEK293T cells were transfected with empty vector, or vector expressing human HA-IFITM3 or untagged human IFITM3. Cell lysates were examined by Western blotting with anti-IFITM3 antibodies and anti-GAPDH antibodies as a loading control; (B) Cells treated as in (A) were examined by flow cytometry staining with anti-IFITM3 antibodies. Cells are gated on IFITM3-positive cells and the blue numbers indicate the average of triplicate samples +/− standard deviation. Ssc, side scatter; (C) Cells treated as in (A) and (B) were infected with IAV H1N1 strain PR8 at an MOI of 1.0 or mock infected for 6 h. Cells were analyzed for percent infection by flow cytometry based on anti-IAV nucleoprotein staining in comparison to mock-infected controls. Average percent infection of triplicate samples in a representative experiment was graphed. Error bars indicate standard deviation. Results are representative of more than 5 similar experiments. NS, not significant by Student’s t test.
Mentions: Given the important differences and striking similarities between the mycobacterial IFITMs and IFITM3 (Figure 1B), we sought to empirically determine if MAB and MAV IFITMs could confer antiviral activity when expressed in human cells. Since these IFITMs have never before been studied, antibodies for detection of their expression do not exist. Epitope tagging has been extensively used for the study of IFITMs [4,6,20,28,29,30,35,36,37,38,39,40,41], but it has been hypothesized that epitope tags may alter IFITM3 activity [7,8,13], though data demonstrating this has not been published. Thus, we first examined N-terminally HA-tagged and untagged human IFITM3 expressed from the same vector for their ability to inhibit influenza virus infection of HEK293T cells. This human cell line was utilized for our studies because it is highly infectable and capable of supporting the full life cycle of influenza virus [42], it expresses low levels of endogenous IFITM3 (Figure 2A,B), it is highly transfectable, and it has proved useful for distinguishing subtle differences in the activities of various IFITM mutants in several studies [4,9,20,28,29,33,36]. HEK293T cells robustly expressed both HA-IFITM3 and IFITM3 as examined by Western blotting with anti-IFITM3 antibodies (Figure 2A). Analysis at the single cell level by flow cytometry confirmed that only a small percentage of HEK293T cells express detectable levels of endogenous IFITM3, and that our transfection efficiencies for the IFITM3 constructs were greater than 70% (Figure 2B). Compared to the vector control transfection, both HA-IFITM3 and IFITM3 were able to significantly inhibit infection by influenza A virus (IAV) H1N1 strain PR8 as measured by anti-IAV nucleoprotein staining by flow cytometry [4,20,33,36]. No statistically significant difference was observed in the abilities of HA-IFITM3 and IFITM3 to inhibit infection (Figure 2C). Thus, we reasoned that utilization of HA-tagging of the mycobacterial IFITMs could provide an effective way to measure their expression without a likely effect on activity.

Bottom Line: Analysis of sequence elements shared by bacterial IFITMs and IFITM3 identified two hydrophobic domains, putative S-palmitoylation sites, and conserved phenylalanine residues associated with IFITM3 interactions, which are all necessary for IFITM3 antiviral activity.We also demonstrated the ability of a bacterial IFITM to co-immunoprecipitate with IFITM3 suggesting formation of a complex, and also visualized strong co-localization of bacterial IFITMs with IFITM3.However, the mycobacterial IFITMs lack the endocytic-targeting motif conserved in vertebrate IFITM3.

View Article: PubMed Central - PubMed

Affiliation: Department of Microbial Infection and Immunity, Center for Microbial Interface Biology, the Ohio State University, Columbus, OH 43210, USA. wjamesmelvin@gmail.com.

ABSTRACT
Interferon induced transmembrane proteins (IFITMs) found in vertebrates restrict infections by specific viruses. IFITM3 is known to be essential for restriction of influenza virus infections in both mice and humans. Vertebrate IFITMs are hypothesized to have derived from a horizontal gene transfer from bacteria to a primitive unicellular eukaryote. Since bacterial IFITMs share minimal amino acid identity with human IFITM3, we hypothesized that examination of bacterial IFITMs in human cells would provide insight into the essential characteristics necessary for antiviral activity of IFITMs. We examined IFITMs from Mycobacterium avium and Mycobacterium abscessus for potential antiviral activity. Both of these IFITMs conferred a moderate level of resistance to influenza virus in human cells, identifying them as functional homologues of IFITM3. Analysis of sequence elements shared by bacterial IFITMs and IFITM3 identified two hydrophobic domains, putative S-palmitoylation sites, and conserved phenylalanine residues associated with IFITM3 interactions, which are all necessary for IFITM3 antiviral activity. We observed that, like IFITM3, bacterial IFITMs were S-palmitoylated, albeit to a lesser degree. We also demonstrated the ability of a bacterial IFITM to co-immunoprecipitate with IFITM3 suggesting formation of a complex, and also visualized strong co-localization of bacterial IFITMs with IFITM3. However, the mycobacterial IFITMs lack the endocytic-targeting motif conserved in vertebrate IFITM3. As such, these bacterial proteins, when expressed alone, had diminished colocalization with cathepsin B-positive endolysosomal compartments that are the primary site of IFITM3-dependent influenza virus restriction. Though the precise evolutionary origin of vertebrate IFITMs is not known, our results support a model whereby transfer of a bacterial IFITM gene to eukaryotic cells may have provided a selective advantage against viral infection that was refined through the course of vertebrate evolution to include more robust signals for S-palmitoylation and localization to sites of endocytic virus trafficking.

No MeSH data available.


Related in: MedlinePlus