Inositol pyrophosphate mediated pyrophosphorylation of AP3B1 regulates HIV-1 Gag release

Abstract

High-energy inositol pyrophosphates, such as IP(7) (diphosphoinositol pentakisphosphate), can directly donate a beta-phosphate to a prephosphorylated serine residue generating pyrophosphorylated proteins. Here, we show that the beta subunit of AP-3, a clathrin-associated protein complex required for HIV-1 release, is a target of IP(7)-mediated pyrophosphorylation. We have identified Kif3A, a motor protein of the kinesin superfamily, as an AP3B1-binding partner and demonstrate that Kif3A, like the AP-3 complex, is involved in an intracellular process required for HIV-1 Gag release. Importantly, IP(7)-mediated pyrophosphorylation of AP3B1 modulates the interaction with Kif3A and, as a consequence, affects the release of HIV-1 virus-like particles. This study identifies a cellular process that is regulated by IP(7)-mediated pyrophosphorylation.

Fig. 1.

AP3B1 is pyrophosphorylated by IP₇. (A) Schematic representation of the AP-3 complex and domain structure of AP3B1. The region amino acid 576 to 901 represents the bait used in the yeast two-hybrid screen. The acidic regions containing potential IP₇-targeted serines are underlined (regions I–III). (B) In vitro phosphorylation of AP3B1. Protein extracts of kcs1Δ yeast expressing GST-AP3B1 or derivatives were incubated with 5β[³²P]IP₇ and resolved by NuPAGE; autoradiography was used to determine phosphorylation and immunoblotting with anti-GST antibody confirmed protein equal loading. (C) In vitro pyrophosphorylation of purified AP3B1. GST-AP3B1 (576–902) was expressed and purified from E. coli (BL21). Purified GST-AP3B1 (576–902) immobilized on glutathione beads was preincubated as follows: (i) Without, with inactive (boiled) or with active CK2 and ATP and subsequently treated with 5β[³²P]IP₇ as in B (lanes 1, 2, and 3, respectively); (ii) with active CK2 and ATP, treated with boiled or active λ-phosphatase, and subsequently phosphorylated with 5β[³²P]IP₇ as in B (lanes 4 and 5, respectively). (D) In vitro pyrophosphorylation of AP3B1 depends on the endogenous levels of IP₇. Protein extracts of vip1Δ, wild-type (WT), and kcs1Δ yeast expressing GST-AP3B1 were incubated with 5β[³²P]IP₇ and processed as in B. (E) IP₇ pyrophosphorylation of endogenous AP3B1. AP3B1 was immunoprecipitated from WT MEF (+/+) or mocha (mh/mh) cell lines, treated with 5β[³²P]IP₇, and processed as in B. (F) Intracellular pyrophosphorylation of AP3B1 results in a gel mobility shift. Quickly prepared cell extracts from WT and kcs1Δ yeast expressing GST-AP3B1 were boiled in sample buffer and resolved by NuPAGE. Gels are representative of at least three independent experiments.

Fig. 2.

AP3B1 interacts with Kif3A. (A) Yeast two-hybrid interaction between AP3B1 (amino acid 576 to 902; in pGBKT7 vector) and the C-terminal non-motor domain of Kif3A (amino acid 601 to 702; in pACT2 vector). Yeast AH109 expressing both vectors were serially diluted and spotted on yeast synthetic media lacking either leucine and tryptophane (SD-LT⁻, to assess growth of cotransformed yeast) or leucine, tryptophane, histidine, and adenine (SD-LTHA⁻, to assess interaction strength), and grown at 30 °C for 3 days. (B) Interaction between GST-AP3B1 and Myc-Kif3A in mammalian cells. HeLa cells were cotransfected with Myc-Kif3A, Myc-Kif3A (1–342), or Myc-Kif3A (354–702) together with GST-AP3B1, GST-AP3B1 (577–1094), or GST vector control. Proteins were extracted 24 h after transfection and subjected to pull-down with glutathione beads. Interactions were detected by immunoblotting with an anti-Myc antibody (Top) and an anti-GST antibody (Bottom). The inputs are shown on the Right. (C) Endogenous AP3 complex interacts with endogenous Kif3A in mammalian cells. AP3B1 and AP3D1 were immunoprecipitated (IP) from wild-type MEF and mocha (mh/mh) cell lines and immunoblotted with antibodies against Kif3A, AP3B1, and AP3Da. Actin was used as loading control.

Fig. 3.

AP3B1 pyrophosphorylation reduces interaction with Kif3A. (A) Lack of AP3B1 pyrophosphorylation increases its interaction with Kif3A in yeast cells. GST-pull downs (Left) and respective quantification (Right). Protein extracts from WT yeast expressing GST-AP3B1 or GST vector control, or from kcs1Δ yeast expressing GST-AP3B1 were incubated with protein extracts of kcs1Δ yeast expressing HA-Kif3A (601–702). Protein extracts were subjected to pull-down with glutathione beads. Interactions were detected by immunoblotting with anti-HA and anti-GST antibodies. Inputs are shown on the right. Quantification was done by taking the ratio between the bands intensities of the HA-Kif3A and the GST-AP3B1 and normalizing kcs1Δ expressed GST-AP3B1 against WT expressed protein. Data represent means ± standard deviation from three independent experiments. (B) AP3B1 pyrophosphorylation reduces its interaction with Kif3A in mammalian cells. GST-pull downs (Left) and respective quantification (Right). HeLa cells were triple transfected with Myc-Kif3A, GST-AP3B1, and either Myc-IP₆K1, Myc-IP₆K2, or Myc vector control. Protein extracts were subjected to pull-down with glutathione beads. Interaction between AP3B1 and Kif3A was detected by immunoblotting with anti-Myc and anti-GST antibodies. Inputs are shown on the right. Quantification was done by taking the ratio between the bands intensities of the Myc-Kif3A and the GST-AP3B1 and normalizing for IP₆K1–2 overexpression compared with vector-transfected cells. Data represent means ± standard deviation of the mean from three independent experiments. (C) Interaction between AP3B1 and Kif3A is not affected by overexpression of kinase dead IP₆K1 or IP₆K2. HeLa cells were triple transfected with Myc-Kif3A, GST-AP3B1, and either Myc vector control, Myc-IP₆K1K/A, or Myc-IP₆K2K/A. Interaction between AP3B1 and Kif3A was detected as in B. (D) IP₇-mediated pyrophosphorylation of AP3B1 is sufficient to reduce the interaction with Kif3A. Purified recombinant GST-AP3B1 (576–902) conjugated to glutathione beads was sequentially phosphorylated with CK2 [GST-AP3B1 (576–902)*], washed, and either subject to pyrophosphorylation by 5β[³²P]IP₇ or kept under the same conditions but without the addition of 5β[³²P]IP₇. Glutathione conjugated GST-AP3B1 (576–902)* was subsequently washed, incubated with purified recombinant His-Kif3A, and run on SDS-PAGE after being thoroughly washed. Interaction between AP3B1 and Kif3A was detected by immunoblotting with anti-His antibodies.

Fig. 4.

Kif3A is required for HIV-1 VLP release. (A) Kif3A depletion reduces VLP release. HeLa cells transiently expressing siRNA-Kif3A were transfected with Gag-GFP. Culture supernatant and cells lysate were collected 16–18 h later. Kif3A silencing was analyzed by immunoblotting with anti-Kif3A antibody. Gag intracellular expression and release was analyzed by immunoblotting with an anti-p24 antibody. The percentage of Kif3A silencing (see text) was analyzed by quantifying Kif3A protein from siRNA-Kif3A against siRNA-control cells both normalized against an internal endogenous control (actin). (B) Mouse Kif3A complements Kif3A in HeLa cells. HeLa cells transfected with siRNA-Kif3A were subsequently cotransfected with Gag-GFP, and increasing amounts of Myc-mKif3A (mouse Kif3A) samples were analyzed as in A. (C) AP3B1 and Kif3A dominant negative constructs reduced VLP release. HeLa cells were cotransfected with Gag-GFP and Myc-Kif3A (354–702), Myc-AP3B1 (576–902), or empty vector control. Expression of Myc-tagged proteins was analyzed by immunoblotting with an anti-Myc antibody. Gels are representative of analyses performed in duplicate (lanes A and B) in three independent experiments.

Fig. 5.

IP₇-mediated pyrophosphorylation of AP3B1 affects HIV-1 VLP release. Immunoblots analyzing how the modulation between AP3B1 and Kif3A interaction effects HIV-1 VLP release. (A) Increased levels of IP₇ in mammalian cells decrease VLP release. HeLa cells were cotransfected with Gag-GFP and Myc-IP₆K1, Myc-IP₆K2, or empty vector control. Culture supernatants and cell lysates were collected 24 h posttransfection. Detection of Myc-tagged proteins was analyzed by immunoblotting with an anti-Myc antibody. Gag expression, VLP release, and data analysis were performed as in Fig. 4A. (B) Overexpression of kinase dead IP₆K1 or IP₆K2 has no effect on VLP release. HeLa cells were cotransfected with Gag-GFP and Myc-IP₆K1K/A, Myc-IP₆K2K/A, or empty vector control. The experiment was performed as described in A. (C) Decreased levels of IP₇ in mammalian cells increase VLP release. MEF WT (+/+) and IP₆K1 knock-out MEF (ip₆k1/ip₆k1) cell lines show increased HIV-1 VLP release upon Gag-GFP transfection. MEF cells IP₆K1 knock-out were transfected with Gag-GFP and GST empty vector (as a control for transfection efficiency) and analyzed 48 h posttransfection. Gag expression, VLP release, and data analysis were performed as in Fig. 4A.