Regulation of NF-κB by the p105-ABIN2-TPL2 complex and RelAp43 during rabies virus infection

Abstract

At the crossroad between the NF-κB and the MAPK pathways, the ternary complex composed of p105, ABIN2 and TPL2 is essential for the host cell response to pathogens. The matrix protein (M) of field isolates of rabies virus was previously shown to disturb the signaling induced by RelAp43, a NF-κB protein close to RelA/p65. Here, we investigated how the M protein disturbs the NF-κB pathway in a RelAp43-dependant manner and the potential involvement of the ternary complex in this mechanism. Using a tandem affinity purification coupled with mass spectrometry approach, we show that RelAp43 interacts with the p105-ABIN2-TPL2 complex and we observe a strong perturbation of this complex in presence of M protein. M protein interaction with RelAp43 is associated with a wide disturbance of NF-κB signaling, involving a modulation of IκBα-, IκBβ-, and IκBε-RelAp43 interaction and a favored interaction of RelAp43 with the non-canonical pathway (RelB and p100/p52). Monitoring the interactions between host and viral proteins using protein-fragment complementation assay and bioluminescent resonance energy transfer, we further show that RelAp43 is associated to the p105-ABIN2-TPL2 complex as RelAp43-p105 interaction stabilizes the formation of a complex with ABIN2 and TPL2. Interestingly, the M protein interacts not only with RelAp43 but also with TPL2 and ABIN2. Upon interaction with this complex, M protein promotes the release of ABIN2, which ultimately favors the production of RelAp43-p50 NF-κB dimers. The use of recombinant rabies viruses further indicates that this mechanism leads to the control of IFNβ, TNF and CXCL2 expression during the infection and a high pathogenicity profile in rabies virus infected mice. All together, our results demonstrate the important role of RelAp43 and M protein in the regulation of NF-κB signaling.

Fig 1. Mapping of RelAp43 interactors by MS and identification of the p105-ABIN2-TPL2 complex as a major target of M_Tha.

A. Volcano plot of FG-RelAp43 vs FG with V5-CAT (left panel) and FG-RelAp43 vs FG with V5-M_Tha (right panel). HeLa cells were co-transfected with 2 plasmids expressing FG or FG-RelAp43 and V5-CAT or V5-M_Tha. After 24h, 30 μg of protein were purified by TAP and analyzed by MS. The Log2(FG-RelAp43/FG) in X-axis was obtained by a two-sample test, comparing the LFQ between FG-RelAp43 and the control FG. The Log(pValue) in Y-axis was obtained by a multi-sample test, determining if any of the means of several groups are different from each other. Each spot represent a protein significantly identified with RelAp43 (black spot) or failing the statistical analysis (grey spot). The doted lines represent the thresholds for a FDR determined at 5%. Colored spots highlight the proteins of interest (see B). B. LFQ values of the proteins of interest highlighted in (A) and ratio of LFQ values between FG-RelAp43/CAT and FG-RelAp43/M_Tha. C. The presence of the transfected FLAG- and V5-tagged proteins as well as the endogenous p105/p50 proteins in the input and output were revealed by western blot. The results presented are 2 experiments representative of 4 replicates. The “IgH” band corresponds to the immunoglobulin heavy chain from the antibodies used in the TAP experiment.

Fig 2. Characterization of the interactions between RelAp43, p105, ABIN2 and TPL2.

A. Interactions between RelAp43, p105, ABIN2 and TPL2 assessed using BRET technology. HEK 293T cells were transfected for 48h with 2 plasmids encoding for Nluc(★)- and YFP(☆)-tagged proteins before measuring BRET. p50 and STAT1 were used as controls. A threshold (doted line) was determined at 0.025 of netBRET, based on the mean+3SD values obtained from non-interacting control pairs: p50-STAT1, -ABIN2 and -TPL2. Each value represents the mean of 3 independent experiments. Error bars represent the standard deviation. B-G. Interactions between RelAp43, p105, ABIN2 or TPL2 in presence or not of a third helper protein, using PCA technology. HEK 293T cells were transfected with 4 plasmids expressing: a Glu1-, a Glu2-, a cMyc-tagged protein and a firefly luciferase for 48h before measuring both gaussia and firefly activity. Position of the tag is indicated as follow: Glu1-, Glu2- for N-ter and -Glu1, -Glu2 for C-ter. Results are the mean of the logarithm of gaussia NLR values normalized to luciferase activity in at least 4 distinct experiments. A threshold of NLR = 3.5 previously described [54] was used to define the positive results (log 3.5 = 0.54). Error bars represent the standard deviation. ***p<0.001, **p<0.01; *p<0.05.

Fig 3. ABIN2 is excluded from a RelAp43-p105-TPL2 complex during Tha infection.

Interactions between ABIN2 and RelAp43 (A), TPL2 (B) or p105 (C), using PCA technology, in combination with a third protein: mCAT, mp105 or mTPL2, and infected or not by Tha or SAD virus. HEK293T cells were infected at a MOI of 1 and transfected 3h later with 4 plasmids expressing: a Glu1-, a Glu2-, a cMyc-tagged protein and a firefly luciferase for 48h before measuring both gaussia and firefly activity. Position of the tag is indicated as follow: Glu1-, Glu2- for N-ter and -Glu1, -Glu2 for C-ter. Results are the mean of the logarithm of gaussia NLR values normalized to luciferase activity of at least 4 distinct experiments. A threshold of NLR = 3.5 previously described [54] was used to define the positive results (log 3.5 = 0.54). Error bars represent the standard deviation. **p<0.01; *p<0.05.

Fig 4. M_Tha interacts with RelAp43 and ABIN2.

A. Interactions between M_Tha, M_SAD or P and RelAp43, p105, p50, ABIN2 and TPL2 were assessed using BRET technology. P and STAT1 were used as controls. BRET was measured in living HEK 293T cells 48h after transfection. Each value is the mean of all the combinations between Nluc-tagged host proteins and YFP-tagged viral proteins in N- and C-ter position of each protein (excepted P that was only tagged in C-ter), and performed in duplicate. A threshold (doted line) was determined at 0.024% of BRET, based on the mean+3SD values of the M protein-STAT1 and P protein-RelAp43, -p105, -p50, -ABIN2 and -TPL2 pairs of partners. Error bars represent the standard deviation. B-E. M_Tha, M_Th4M and M_SAD interacting with p105 (B), p43 (C), TPL2 (D), and ABIN2 (E), using BRET technology. BRET signal was measured in living HEK 293T cells 48h after transfection of plasmids expressing YFP- and Nluc-tagged proteins at 5 ratios of DNA, from 1:2 to 8:1. A threshold (doted line) was determined for each dilution of DNA (1.2, 1.4, 2.2, 2.5, 3.4% of BRET respectively), based on the mean+3SD values of the M proteins-p105 pairs of partners. Each value is the mean between 3 experiments with error bars representing the standard deviation. *p<0.05. For D, only the non significant (ns) values were highlighted.

Fig 5. Tha virus modulates the response to the infection in infected mice.

A-F. Relative quantification of IFNß (AD), CXCL2 (BE), and TNF (CF) mRNA expression in the brain of mice infected by Tha or Th4M viruses. Six weeks old BALB/c (CDE) or C57BL/6 (FGH) were infected by intramuscular injection and monitored over 21 days. The mice were sacrificed at experimental end point and mRNA was extracted from the brain. Six mice were used per condition. Error bars represent the standard deviation. G. Survival curve of the BALB/c (left panel) and C57BL/6 (right panel) mice used for RNA quantification in A-F.

Fig 6. Model of the control of the RelAp43-p105-ABIN2-TPL2 complex by Tha virus to fine-tune the response to the infection.

A. Interactions between RelAp43, p105, ABIN2 and TPL2 adapted from [3]. RelAp43 and p105 interact through their respective RHD domain. p105 interact with its C-ter part with TPL2 on two different domains: the processing inhibitory domain (PID) of p105 with the C-ter part of TPL2 and the death domain (DD) of p105 with the kinase domain (KD) of TPL2. ABIN2 interacts with TPL2 through its ABIN-homology domain 4 (AHD4) with the same C-ter region of TPL2 involved in the interaction between p105 and TPL2. The circle highlights a region where each protein of the complex might be in a very close proximity, possibly involved in the binding of ABIN2. B. The RelAp43-p105-ABIN2-TPL2 complex is modulated by the matrix protein of pathogenic rabies virus. M proteins destabilize the interaction of ABIN2 with the complex which leads to the regulation of both the NF-κB and the MAPK dependent genes in response to viral infection.