Nrf1 can be processed and activated in a proteasome-independent manner

In Brief:

Vangala et al. demonstrate that the ER-bound transcription factor Nrf1 can be proteolytically cleaved and activated when the proteasome is completely inhibited, thus pointing to the existence of a proteasome-independent pathway to generate functional Nrf1.

In response to proteasome inhibition, transcription factor Nrf1 facilitates de novo synthesis of proteasomes by inducing proteasome subunit (PSM) genes [1, 2]. Previously, we showed that as a membrane-bound protein in the endoplasmic reticulum (ER) with the bulk of its polypeptide in the lumen, p120 Nrf1 activation involves its retrotranslocation into the cytosol in a manner that depends on the AAA-ATPase p97/VCP [3]. This is followed by proteolytic processing and mobilization of the transcriptionally active p110 form of Nrf1 to the nucleus. A subsequent study suggested that site-specific proteolytic processing of Nrf1 by the proteasome yields an active, 75 kD fragment [4]. We show here that under conditions where all three active sites of the proteasome are completely blocked, p120 Nrf1 can still be proteolytically cleaved to the p110 form, which is translocated to the nucleus to activate transcription of PSM genes. Thus, our results indicate that a proteasome-independent pathway can promote release of active p110 Nrf1 from the ER membrane.

Using cell lines including the neuroblastoma line SH-SY5Y, Sha and Goldberg found that whereas low-dose proteasome inhibitor treatments for 16 hr stimulate the formation of a transcriptionally-active, 75 kD form of Nrf1, a high-dose of proteasome inhibitors in that same treatment period blocked processing and activation [4]. To determine whether the proteasome mediates the processing of membrane-bound p120 Nrf1 (inactive precursor) to soluble p110 (processed active form) that we observed upon treatment with proteasome inhibitors, we first analyzed the protein levels of these species in human SH-SY5Y and mouse NIH-3T3 cells subjected to increasing doses of three unrelated proteasome inhibitors – Bortezomib (BTZ), Carfilzomib (CFZ), and MG132 for either 4 or 16 hr. We found no evidence for impaired formation of p110 Nrf1 at any of the doses of the three proteasome inhibitors that we used regardless of the time of exposure (Figures 1A and S1A). We also tracked the Chymotrypsin-like, Trypsin-like, and Caspase-like activities of the proteasome in cell lysates derived from 4 hr proteasome inhibitor treatments and observed a dose-dependent suppression which turned into an almost complete inhibition at higher doses for all inhibitors (Figures S1B and S1C). Overall, we found no correlation between proteasome activity and the ability of Nrf1 to be proteolytically processed to p110 in these cells.

Figure 1. Generation of p110 from p120 is unaffected by proteasome inhibition.

(A) SH-SY5Y human neuroblastoma and NIH-3T3 mouse fibroblast cells were treated for 4 hr with different concentrations of Bortezomib (BTZ), Carfilzomib (CFZ) or MG132 as indicated. The whole cell lysates were immunoblotted to detect p120 and p110 forms of Nrf1. β-Actin served as a loading control. (B) Schematic representation of the pulse-chase experiment to track the processing of Nrf1 is shown. (C) HEK-293-Nrf1^3×Flag cells were pretreated for 2 hr with 10 μM NMS-873 (NMS) [7] to accumulate Flag-tagged Nrf1 p120. During the second hour, the cells were additionally exposed to 12.5 μM CFZ (which completely inhibits all three active sites of the proteasome consistent with previous observations [8]; Also see Figure S1D) and pulse-labeled with 50 μM L-azidohomoalanine (AHA). After washing out the NMS, the cells were then chased with 12.5 μM carfilzomib (CFZ), 100 μg/mL cycloheximide (CHX), and excess of methionine. The cells were harvested at various time points as indicated and immunoprecipitation of lysates was performed with anti-Flag beads. Immunoprecipitants were labeled by Biotin-PEG4-alkyne based click chemistry method. Immunoblot analysis was performed with Neutravidin-HRP to detect the Nrf1 species. (D) HeLa cells were transfected with siRNA against PSMD4 or PSMB4. A non-targeting siRNA was used as control (Ctrl). Three days after transfection, cell lysates were analyzed by immunoblotting to detect the indicated proteins. (* non-specific signal).

The above experiments, although suggestive, are not entirely conclusive since it is possible that the p110 Nrf1 that we observed could have been generated before the proteasome was completely blocked by the inhibitors. To clarify this issue, we set up a pulse-chase experiment to track the formation of p110 from newly synthesized p120 under conditions of complete proteasome inhibition (Figure 1B). During the chase period, we observed conversion of pulse-labeled Nrf1 p120 into the processed p110 form (Figure 1C), thus indicating that this species of Nrf1 can be generated independently of proteasome activity.

Next, in an orthogonal approach, when we depleted PSMB4 (a 20S subunit) or PSMD4 (a 19S subunit) using siRNA, we detected robust accumulation of TCF11/Nrf1 in these cells, but mainly in the processed form (Figure 1D). Sha and Goldberg proposed that formation of p75 occurs because chemically-inhibited proteasome is partially crippled in its enzymatic sites, such that there remains sufficient activity to non-processively clip Nrf1, but insufficient activity to processively degrade it. This mechanism can only work if, on the level of individual proteasome molecules, there is reduced cleavage activity. However, if instead each proteasome molecule retains normal activity but the total number of assembled proteasomes is reduced, Nrf1 should accumulate, but only in the unprocessed form because Nrf1 molecules that engage the remaining proteasomes should be fully degraded. The efficient formation of p110 in cells partially depleted of either PSMB4 or PSMD4 is inconsistent with the idea that this cleavage is performed by crippled proteasomes.

Next, we asked if Nrf1 p110 generated under conditions of complete proteasome inhibition is capable of exerting its biological function as a transcription factor. To this end, we used SH-SY5Y and NIH-3T3 cells and first confirmed that 12.5 μM CFZ for 1 hr results in near-zero values for all three activities of the proteasome (Figure S2A). We then performed similar pulse-chase assays (Figure S2B) as described above to track endogenous Nrf1. We analyzed the cells collected at the beginning and end of the chase period by immunofluorescence microscopy and detected increased Nrf1 nuclear signal in the latter samples regardless of the cell type (Figure S2C). Consistent with this observation, we found that a representative set of PSM genes were upregulated at the end of the chase period compared to the beginning in both cell types tested, implying that the processed pool of endogenous Nrf1 in the nucleus was able to transcriptionally induce its target genes (Figure S2D). Accordingly, this effect was completely abolished in cells with Nrf1 knocked out (Figures S2D and S2E).

We do not understand the basis for the divergence in results observed here compared to those reported by Sha and Goldberg. One notable difference is the Nrf1 species being followed. Whereas Sha and Goldberg report a p75 form of Nrf1 as the active species, in our hands we have consistently seen a correlation between the p110 form and the transcriptional competence of Nrf1 in line with other studies [5]. However, our results do not exclude the possibility that different pathways may mediate processing and activation of Nrf1, depending upon the physiological state of the cell.

Identification of the protease involved in the generation of Nrf1 p110 could be valuable not only from a mechanistic point of view, but also from the perspective of cancer therapy [6] since this protease can then be targeted to suppress the Nrf1-mediated proteasome recovery pathway mobilized during proteasome inhibitor treatments.