IFNα treatment of PBMCs induces TC-specific deaminase activity. A, PBMCs from a healthy human donor were cultured overnight and either left untreated (Un), or treated with anti-CD3 and anti-CD28 antibodies (3/28), PMA + IL-2 (PMA), IFNα (α), or IFNγ (γ). Cells were lysed at the indicated time points and analyzed for deaminase activity in the presence of RNase A using oligonucleotides containing motifs for either ACCC (left graph) or ATTC (right graph). Lysates from 293T cells or 293T cells expressing A3G were analyzed in parallel (data not shown). Activity is graphed as RFU. Lysates shown in the activity graphs were also analyzed by Western blot (WB) with antisera against A3G, A3F, or calreticulin (cal), which served as a loading control. Lysates from 293T cells (−) or 293T cells transiently expressing untagged A3G, A3A, or A3F were analyzed in parallel (293T). Bar, lanes containing IFNα-treated samples. B, PBMCs from the same donor as above were cultured in the presence of neutralizing antibodies against IFNα and the IFNα/β receptor (anti-IFNα/β). Cells were then either left untreated or treated with αCD3/28, PMA + IL-2, IFNα, or IFNγ. Cells were lysed and analyzed for deaminase activity as in A using oligonucleotides containing motifs for either ACCC (left graph) or ATTC (right graph). These lysates were also analyzed by Western blot as in A. The arrows in both blot panels show the expected positions of A3G, A3F, A3A, and calreticulin. For deaminase assays, error bars represent S.E. of triplicate reactions. All samples in A and B are from the same experiment and were analyzed together. Graphs in this figure are presented in color to help with clarity.

Treatment with PHA plus IFNα results in a transient increase in deaminase activity. PBMCs from a healthy human donor were cultured overnight and then treated with either anti-IFNα/β, PHA alone, IFNα alone, IFNγ alone, IFNα + PHA, or IFNγ + PHA. Cells were lysed at the indicated time points and analyzed for deaminase activity in the presence of RNase A using oligonucleotides containing ATTC. Activity is graphed as RFU, and error bars represent S.E. of triplicate reactions. Lysates from all time points were analyzed by Western blot (WB) with antiserum against A3G or calreticulin (cal) as a loading control. The arrows show the expected positions of A3G, A3A, and calreticulin. All samples are from the same experiment and were analyzed together.

IFNα induces TC-specific deaminase activity and A3A isoforms in monocytes and macrophages but not in T cells. A, PBMCs from a healthy human donor were cultured overnight and then subjected to negative selection to obtain three populations enriched for either CD4⁺ T cells, CD8⁺ T cells, or monocytes (for subset analysis, see supplemental Fig. 1). Cell subsets were either left untreated or treated with IFNα for 24 h. The cells were then lysed and analyzed for deaminase activity in the presence of RNase A using the ATTC-containing oligonucleotide. Activity is graphed as RFU. Lysates from these cells were analyzed by Western blot (WB) with antiserum against A3G or calreticulin (cal) as a loading control. B, PBMCs from a healthy human donor were cultured overnight and then subjected to negative selection for monocytes. Both the monocytes and the positively selected cells were treated with either IFNα or anti-IFNα/β, as described above. Cells were lysed at the indicated time points and analyzed for deaminase activity as described in A. Activity is graphed as RFU. Lysates from monocytes (mono) or positively selected cells (other) treated with IFNα (α) or anti-IFNα/β (ab) were analyzed by Western blot with antiserum against A3G or calreticulin as a loading control. C, PBMCs from a healthy human donor were cultured overnight and then subjected to negative selection for isolation of monocytes, which were differentiated into MDMs by treating with M-CSF for 7 days. After differentiation, MDMs were treated with either IFNα or anti-IFNα/β, as described above. Cells were lysed at the indicated time points and analyzed for deaminase activity in the presence of RNase A using either the ACCC- or ATTC-containing oligonucleotide. Activity is graphed as RFU. Lysates from these cells were analyzed by Western blot with antiserum against A3G, or calreticulin as a loading control. In all blots, arrows show expected positions of A3G, A3A, and calreticulin. In each panel, samples are from one experiment and were analyzed together. For deaminase assays, error bars represent S.E. of triplicate reactions.

A3A in PBMCs does not require RNase treatment for activity. PBMCs from a healthy human donor were cultured overnight and then either left untreated or treated with IFNα. Cells were lysed at the indicated time points and analyzed for deaminase activity in the absence (−RNase) or presence (+RNase) of RNase A using oligonucleotides containing the ATTC motif. Activity is graphed as RFU. Error bars represent S.E. of triplicate reactions.

Consistency of IFNα effect across donors and effect of TLR ligands that induce IFNα. A, effect of IFNα on deaminase activity in PBMCs from three different donors. PBMCs from three different human donors (indicated by three different symbols) were cultured overnight and then either left untreated or treated with PHA or IFNα and then analyzed for deaminase activity as in Fig. 1 using the ACCC-containing (left) or ATTC-containing (right) oligonucleotides. B, TLR ligands that induce IFNα production cause sustained TC-specific deaminase activity in PBMCs. PBMCs from a healthy human donor were cultured overnight and then either left untreated or treated with IFNα or ligands for TLR3 (poly(I-C)), TLR 7 (imiquimod), TLR8 (single-stranded poly(U)), or TLR9 (ODN2016). Cells were lysed at the indicated time points and analyzed for deaminase activity in the presence of RNase A using oligonucleotides containing ATTC. Activity is graphed as RFU. Lysates from untreated PBMCs (−), or PBMCs treated with IFNα (α) or ligands for TLR3 (3), TLR 7 (7), TLR8 (8), or TLR 9 (9) were analyzed by Western blot (WB) with antiserum against A3G, A3F, or calreticulin (cal) as a loading control. The arrows show expected positions of A3G, A3F, A3A, and calreticulin. All samples in panel B are from the same experiment and were analyzed together. In deaminase assays, error bars represent S.E. of triplicate reactions.

Western blots with A3A/B antiserum confirm that the ∼20-kDa proteins likely represent A3A isoforms. Lysates containing 10 μg of total cellular protein from untreated PBMCs, IFNα-treated PBMCs (PBMC + IFNα), and 293T cells transiently transfected with an A3A plasmid engineered to contain an optimized Kozak consensus (A3A-Koz) were analyzed by Western blot (WB) with a commercial antiserum directed against A3A and -B (A3A/B) and an antiserum directed against A3G. Markers show the expected positions of A3B, A3G, and A3A, all of which share significant sequence homology, accounting for cross-reactivity of antisera.

siRNA knockdown of A3A prevents induction of TC-specific deaminase activity and A3A isoforms in monocytes. PBMCs from a healthy human donor were subjected to negative selection for monocytes, and then the monocytes were cultured overnight. The following day, the monocytes were transfected with either a control siRNA oligonucleotide (control) or siRNA oligonucleotides specific for A3G (G) or A3A (A) using the Amaxa Nucleofector system and then allowed to rest in serum-containing recovery medium for 4 h. The cells were then either left untreated or treated with either IFNα or neutralizing antibodies against IFNα and the IFNα receptor (anti-IFNα/β). Cells were lysed at 48 h post-transfection and analyzed for deaminase activity in the presence of RNase A using oligonucleotides containing the ATTC motif. Activity is graphed as RFU. Error bars represent S.E. of triplicate reactions. Lysates were also analyzed by Western blot (WB) with antisera against A3G or calreticulin (cal) as a loading control (bottom). The arrows show expected positions of A3G, A3A, and calreticulin. All samples are from the same experiment and were analyzed together.

TC-specific deaminase activity of an A3A isoform generated by translation from an alternative start site. A, 293T cells were transfected with plasmid encoding either wild-type A3A with its native translational start site (A3A-nat; GCACATGGAA, start codon at position 1 underlined), A3A with an engineered optimal Kozak consensus at position 1 (A3A-Koz; CACCATGGAA, start codon at position 1 underlined), or A3A encoding point mutation M1A, M13A, or E72Q. 293T cells transfected with empty vector (mock) and PBMCs treated with IFNα were analyzed in parallel. Cells were lysed 48 h post-transfection and analyzed by Western blot (WB) with antisera against A3G and A3A/B, with calreticulin (cal) as a loading control. Molecular weights are shown to the right. All blots are from the same experiment and were analyzed together. Lysates were also analyzed for deaminase activity in the presence of RNase A using the ATTC-containing oligonucleotide. Activity is graphed as RFU for the indicated amounts of total protein input. Error bars represent S.E. of triplicate reactions. B, ClustalW alignment of the first 43 amino acids of the predicted amino acid sequence of human (Homo sapiens, NM_145699 in GenBank) A3A with predicted sequences of A3A from macaque (Macaca mulatta, XP_001096739 in GenBank), sheep (Ovis aries, ACH92046 in GenBank), and cow (Bos taurus, NP_001157408 in GenBank). Amino acids that are either unique or similar between species are shaded. The arrows indicate the positions of the Met¹ start site and the Met¹³ alternate starting site of human A3A.