Dephosphorylation of CDK9 by protein phosphatase 2A and protein phosphatase-1 in Tat-activated HIV-1 transcription

Abstract

Background: HIV-1 Tat protein recruits human positive transcription elongation factor P-TEFb, consisting of CDK9 and cyclin T1, to HIV-1 transactivation response (TAR) RNA. CDK9 is maintained in dephosphorylated state by TFIIH and undergo phosphorylation upon the dissociation of TFIIH. Thus, dephosphorylation of CDK9 prior to its association with HIV-1 preinitiation complex might be important for HIV-1 transcription. Others and we previously showed that protein phosphatase-2A and protein phosphatase-1 regulates HIV-1 transcription. In the present study we analyze relative contribution of PP2A and PP1 to dephosphorylation of CDK9 and to HIV-1 transcription in vitro and in vivo.

Results: In vitro, PP2A but not PP1 dephosphorylated autophosphorylated CDK9 and reduced complex formation between P-TEFb, Tat and TAR RNA. Inhibition of PP2A by okadaic acid inhibited basal as well as Tat-induced HIV-1 transcription whereas inhibition of PP1 by recombinant nuclear inhibitor of PP1 (NIPP1) inhibited only Tat-induced transcription in vitro. In cultured cells, low concentration of okadaic acid, inhibitory for PP2A, only mildly inhibited Tat-induced HIV-1 transcription. In contrast Tat-mediated HIV-1 transcription was strongly inhibited by expression of NIPP1. Okadaic acid induced phosphorylation of endogenous as well transiently expressed CDK9, but this induction was not seen in the cells expressing NIPP1. Also the okadaic acid did not induce phosphorylation of CDK9 with mutation of Thr 186 or with mutations in Ser-329, Thr-330, Thr-333, Ser-334, Ser-347, Thr-350, Ser-353, and Thr-354 residues involved in autophosphorylation of CDK9.

Conclusion: Our results indicate that although PP2A dephosphorylates autophosphorylated CDK9 in vitro, in cultured cells PP1 is likely to dephosphorylate CDK9 and contribute to the regulation of activated HIV-1 transcription.

Figure 1

PP2A dephosphorylates CDK9 in vitro. A, Dephosphorylation of CDK9 by PP2A and PP1. Recombinant CDK9/cyclin T1 was incubated with γ-(³²P) ATP to allow autophosphorylation (lane 1). The kinase activity was blocked by 7 mM EDTA and CDK9 was used as a substrate for PP1 (lanes 2 and 3) or PP2A (lanes 4 and 5). Dephosphorylated CDK9 was resolved on 10% SDS-PAGE and quantified on PhosphoImager (lower panel). B, Phosphorylase-a phosphatase activity of PP1 and PP2A at concentrations corresponding to panel A, presented as the amount of phosphorylase-a remained in the reaction after the treatment with the phosphatase. C, Pre-treatment with PP2A increases autophosphorylation of CDK9. Recombinant CDK9/cyclin T1 was incubated without (lane 1) or with PP1 (lanes 2 and 3) or PP2A (lanes 4 and 5) at concentrations corresponding to Panel A. After incubation, the phosphatases were blocked with 1 μM okadaic acid and CDK9/cyclin T1 was subjected to the autophosphorylation with γ-(³²P) ATP (lanes 1 to 5). Phosphorylated CDK9 was resolved on 10% SDS-PAGE and exposed to the PhosphoImager screen.

Figure 2

Binding of Tat to TAR RNA and CDK9/cyclin T1. Precipitation of biotin TAR RNA with purified Tat and with CDK9/cyclin T1. Lane 1, control denatured TAR RNA. Lane, control without Tat. Lane 3, untreated CDK9. cyclin T1. Lane 4, CDK9/cyclin T1 treated with PP2A (panel A) or with PP1 (panel B). Precipitated proteins and TAR RNA were recovered in SDS-loading buffer, resolved 12% SDS-PAGE and immunoblotted with indicated antibodies. Position of TAR RNA was determined by Ponceau-S staining.

Figure 3

Contribution of PP2A and PP1 to Tat-activated transcription in vitro. A, In vitro transcription reactions were carried with the indicated amounts of recombinant Tat. Lane 1, no DNA template; lane 2, no Tat added; lanes 3–5, Tat added at 10 ng, 50 ng and 100 ng correspondingly. Transcription product was resolved on 5 % Urea-PAGE, exposed to the PhosphoImager screen and quantified. B, Inhibition of PP1 and PP2A by okadaic acid in phosphorylase-a dephosphorylation assay. PP1 and PP2A were inhibited by okadaic acid with IC₅₀ = 70 nM and 0.4 nM concentration of inhibitor respectively. C, Okadaic acid inhibits basal and Tat-activated transcription. Lane 1, no DNA template; lane 2, no Tat added; lane 3, transcription with 50 ng of Tat; lanes 4 and 5, transcription in the absence of Tat and with 10 nM or 1 μM of okadaic acid; and lanes 6 and 7, transcription in the presence of 50 ng of Tat and with 10 nM or 1 μM of okadaic acid. Transcription products were resolved on 5 % Urea-PAGE, exposed to the PhosphoImager screen and quantified. D, NIPP1 inhibits Tat-activated transcription. Lane 1, no DNA template; lane 2, no Tat added; lane 3, transcription with 50 ng of Tat; lane 4, transcription in the absence of Tat and with 100 ng NIPP1; lane 5, transcription in the presence of 50 ng of Tat and 100 ng NIPP1. Transcription products were resolved on 5 % Urea-PAGE, exposed to the PhosphoImager screen and quantified.

Figure 4

Okadaic acid modestly inhibits Tat-induced HIV-1 transcription in cultured cells. A, COS-7 cells were co-transfected without (lane 1) or with Tat-expressing vector and HIV-1 LTR-LacZ (lanes 2–10). Cells were also treated with indicated concentrations of okadaic acid (lanes 3–10). Expression of β-galactosidase was analyzed using ONPG-based assay. B, Quantification of the inhibition of Tat-induced transcription by okadaic acid using Prism. C, COS-7 cells were transfected with mutant HIV-1 LTR with a deletion of the fragment encoding TAR RNA (HIV-1 LTRΔTAR) without (lanes 1 and 3–10) or with Tat-expression plasmid (lane 2) and treated with the indicated concentrations of okadaic acid (lanes 3–10).

Figure 5

Expression of NIPP1 inhibits Tat-dependent HIV-1 transcription in COS-7 cells. A, Lane 1, COS-7 cells grown in 24-well plate were transfected with the indicated amount of JK2 using Ca²⁺-phosphatase method. Lane 2, COS-7 cells were transfected with 25 ng of JK2 and indicated amount of Tat expression plasmid. Lane 3 and 4, COS-7 cells were transfected with 25 ng of JK2, 50 ng of Tat expression vector and indicated amounts of wild type or mutant NIPP1. B, NIPP1 and mutant NIPP1 equally affect HIV-1 transcription in the absence of Tat. COS-7 cells were transfected with 50 ng of JK2 or JK2ΔTAR and with indicated amounts of NIPP1 or mutant NIPP1.

Figure 6

CDK9 is dephosphorylated by PP1 in cultured cells. A, Immunoprecipitation of CDK9. Lane 1, CDK9 was precipitated from HeLa cell extract with anti-cyclin T1 antibodies, resolved on 10% SDS-PAGE and immunoblotted with anti-CDK9 antibodies; Lane 2, immunoprecipitation of recombinant CDK9/cyclin T1; Lanes 3, input recombinant CDK9/cyclin T1; lane 4, input HeLa cell extract. B, HeLa cells were labelled with (³²P) orthophosphate in the absence (lane 1) and in the presence of 1 μM okadaic acid (lane 2) and cellular extracts were immunoprecipitated with anti-cyclin T1 antibodies, resolved by 10% SDS-PAGE and transferred to PVDF membrane. Position of CDK9 was determined by probing the membrane with anti-CDK9 antibodies using 3,3'-Diaminobenzidine enhancer system. The picture is autoradiogram of the membrane exposed to phosphor imager screen. C, 293T cells were labeled with (³²P) orthophosphate in the absence (lane 1) and in the presence of 100 nM okadaic acid (lane 2) and cellular extracts were immunoprecipitated with anti-CDK9 antibodies, resolved by 10% SDS-PAGE and transferred to PVDF membrane. Position of CDK9 was determined by probing the membrane with anti-CDK9 antibodies using 3,3'-Diaminobenzidine enhancer system. The picture is an autoradiogram of the membrane exposed to phosphor imager screen. D, 293T-cdNIPP1 cells stably expressing central domain of NIPP1 (143–224) were transfected with Flag-CDK9 expression vector and labeled with (³²P) orthophosphate in the absence (lane 1) and in the presence of 100 nM okadaic acid (lane 2). Cellular extracts were immunoprecipitated with anti-Flag antibodies, resolved by 10% SDS-PAGE and transferred to PVDF membrane. Position of CDK9 was determined by probing the membrane with anti-CDK9 antibodies using 3,3'-Diaminobenzidine enhancer system. The picture is autoradiogram of the membrane exposed to phosphor imager screen.

Figure 7

Determination of CDK9 residues dephosphorylated in cultured cells. A, 293T cells were transfected with expression vectors Flag-tagged CDK9 (lanes 2 and 3), CDK9 T186A mutant (lanes 4 and 5) and CDK9 C8A mutant (lanes 6 and 7). Cells were labeled with (³²P) orthophosphate in the absence (lanes 2, 4 and 6) and in the presence of 100 nM okadaic acid (lane 3, 5 and 7). Lane 1, mock transfected cells. Cellular extracts were immunoprecipitated with anti-Flag antibodies, resolved by 10% SDS-PAGE and transferred to PVDF membrane. Upper panel shows expression of CDK9 determined by probing the membrane with anti-CDK9 antibodies using 3,3'-Diaminobenzidine enhancer system. Lower panel is an autoradiogram of the membrane exposed to phosphor imager screen. B, quantification of the Phosphor Imager panel.