1. Biochemistry and Chemical Biology

Hypoxia truncates and constitutively activates the key cholesterol synthesis enzyme squalene monooxygenase

  1. Hudson W Coates
  2. Isabelle M Capell-Hattam
  3. Ellen M Olzomer
  4. Ximing Du
  5. Rhonda Farrell
  6. Hongyuan Yang
  7. Frances L Byrne
  8. Andrew J Brown  Is a corresponding author
  1. School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Australia
  2. Prince of Wales Private Hospital, Australia
  3. Chris O’Brien Lifehouse, Australia
https://doi.org/10.7554/eLife.82843
Share this article
Cite this article
  1. Hudson W Coates
  2. Isabelle M Capell-Hattam
  3. Ellen M Olzomer
  4. Ximing Du
  5. Rhonda Farrell
  6. Hongyuan Yang
  7. Frances L Byrne
  8. Andrew J Brown
(2023)
Hypoxia truncates and constitutively activates the key cholesterol synthesis enzyme squalene monooxygenase
eLife 12:e82843.
https://doi.org/10.7554/eLife.82843
  2. Download BibTeX
  3. Download .RIS

Abstract

Cholesterol synthesis is both energy- and oxygen-intensive, yet relatively little is known of the regulatory effects of hypoxia on pathway enzymes. We previously showed that the rate-limiting and first oxygen-dependent enzyme of the committed cholesterol synthesis pathway, squalene monooxygenase (SM), can undergo partial proteasomal degradation that renders it constitutively active. Here, we show hypoxia is a physiological trigger for this truncation, which occurs through a two-part mechanism: (1) increased targeting of SM to the proteasome via stabilization of the E3 ubiquitin ligase MARCHF6 and (2) accumulation of the SM substrate, squalene, which impedes the complete degradation of SM and liberates its truncated form. This preserves SM activity and downstream pathway flux during hypoxia. These results uncover a feedforward mechanism that allows SM to accommodate fluctuating substrate levels and may contribute to its widely reported oncogenic properties.

Editor's evaluation

Cholesterol biosynthesis is a highly oxygen-intensive process as the synthesis of one molecule of cholesterol consumes 11 molecules of oxygen. This valuable paper provides a new link between oxygen sensing and cholesterol synthesis by showing that under conditions of hypoxia (oxygen deprivation), a key cholesterol synthesis enzyme called squalene monooxygenase (SM) is partially degraded to a truncated form that is constitutively active. The supporting evidence is solid and suggests that unregulated activation of SM under oxygen-deficient conditions could reduce the toxicity of squalene and other sterol intermediates.

https://doi.org/10.7554/eLife.82843.sa0

eLife digest

Cells need cholesterol to work properly but too much cholesterol is harmful and can contribute to atherosclerosis (narrowing of blood vessels), cancer and other diseases. Cells therefore carefully control the activity of the enzymes that are involved in making cholesterol, including an enzyme known as squalene monooxygenase.

When the level of cholesterol in a cell rises, a protein called MARCHF6 adds molecules of ubiquitin to squalene monooxygenase. These molecules act as tags that direct the enzyme to be destroyed by a machine inside cells, known as the proteasome, thereby preventing further (unnecessary) production of cholesterol.

Previous studies found that squalene monooxygenase is sometimes only partially broken down to make a shorter (truncated) form of the enzyme that is permanently active, even when the level of cholesterol in the cell is high. However, it was unclear what triggers this partial breakdown.

The process of making cholesterol uses a lot of oxygen, yet many cancer cells thrive in tumours with low levels of oxygen. Here, Coates et al. used biochemical and cell biology approaches to study the effect of low oxygen levels on the activity of squalene monooxygenase in human cells. The experiments revealed that low oxygen levels trigger squalene monooxygenase to be partially degraded to make the truncated form of the enzyme.

Firstly, MARCHF6 accumulates and adds ubiquitin to the enzyme to accelerate its delivery to the proteasome. Secondly, as the proteasome starts to degrade the enzyme, a build-up of squalene molecules impedes further breakdown of the enzyme. This mechanism preserves squalene monooxygenase activity when oxygen levels drop in cells, which may compensate for temporary oxygen shortfalls and allow cells to continue to make cholesterol.

Squalene monooxygenase is overactive in individuals with a wide variety of diseases including fatty liver and prostate cancer. Drugs that block squalene monooxygenase activity have been shown to stop cancer cells from growing, but unfortunately these drugs are also toxic to mammals. These findings suggest that reducing the activity of squalene monooxygenase in more subtle ways, such as stopping it from being partially degraded, may be a more viable treatment strategy for cancer and other diseases associated with high levels of cholesterol.

Introduction

Cholesterol is an essential component of mammalian cell membranes, yet its aberrant accumulation is detrimental (Baigent et al., 2010). Most cellular cholesterol arises from an energetically expensive biosynthetic pathway requiring eleven oxygen molecules and over one hundred ATP equivalents per molecule of product (Brown et al., 2021). Furthermore, many intermediates of this pathway are toxic in excess (Porter and Herman, 2011). Coordinated regulation of cholesterol synthesis enzymes is therefore vital to ensure the pathway is active only when required, and sufficient substrates and cofactors are available to maintain flux through the full length of the pathway.

Squalene monooxygenase (SM, also known as squalene epoxidase or SQLE, EC:1.14.14.17) catalyzes the rate-limiting conversion of squalene to monooxidosqualene in the committed cholesterol synthesis pathway (Gill et al., 2011; Chua et al., 2020). This reaction is the first in the pathway to require molecular oxygen, with the introduced epoxide group ultimately forming the signature C3-hydroxyl group of cholesterol. SM can also act a second time on monooxidosqualene to produce dioxidosqualene, the precursor of the potent regulatory oxysterol 24(S),25-epoxycholesterol (Bai et al., 1992). As a flux-controlling enzyme, SM is subject to metabolic regulation at both the transcriptional level via sterol regulatory element-binding proteins (Horton et al., 2003) and the post-translational level via ubiquitination and proteasomal degradation (Gill et al., 2011). The latter is mediated by the N-terminal regulatory domain of SM (SM-N100), which senses lipid levels in the endoplasmic reticulum (ER) membrane and accelerates or attenuates SM degradation in response to excess cholesterol or squalene, respectively (Chua et al., 2017; Yoshioka et al., 2020). These reciprocal feedback and feedforward loops fine-tune SM activity according to metabolic supply and demand. SM is typically fully degraded by the proteasome; however, incomplete proteolysis produces a truncated form of SM (trunSM) that lacks a large portion of the lipid-sensing SM-N100 domain but retains the full catalytic domain (Coates et al., 2021). This renders trunSM cholesterol-resistant and therefore constitutively active. Although truncation is induced by the SM inhibitor NB-598, human cell lines express similar levels of full-length and truncated SM (Coates et al., 2021). This points to the existence of an unknown physiological trigger for truncation.

Clarifying the mechanisms of SM regulation is particularly pertinent given the importance of the enzyme, and cholesterol more generally (Kuzu et al., 2016), in oncogenesis. Overexpression of the SM gene SQLE is associated with greater invasiveness and lethality in breast (Brown et al., 2016), prostate (Stopsack et al., 2016; Pudova et al., 2020), and pancreatic cancers (Bai et al., 2021), amongst others. At the protein level, aberrant SM expression is implicated in colorectal cancer progression (He et al., 2021; Jun et al., 2021) and the development of both nonalcoholic steatohepatitis and hepatocellular carcinoma (Liu et al., 2021; Liu et al., 2018). Given its key role in oxygen-dependent cholesterol synthesis, SM may be particularly critical for cancer cell survival during hypoxia, which is common in the poorly vascularized cores of solid tumors and often associated with poor prognosis (Rankin and Giaccia, 2016). In support of this idea, SM inhibition sensitizes breast and colorectal cancer cells to hypoxia-induced cell death (Haider et al., 2016). Although hypoxic cells tend to accumulate cholesterol, there are conflicting reports on changes in biosynthetic flux (Mukodani et al., 1990; Parathath et al., 2011; Wu et al., 2020). Furthermore, with the notable exception of the early pathway enzyme 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR) (Nguyen et al., 2007), the effects of hypoxia on individual biosynthetic enzymes are unknown. It is also unclear if these might be perturbed in a tumor context to favor continued cholesterol synthesis and cell proliferation.

Here, we show hypoxic conditions induce SM truncation in a variety of cell lines through a combination of accelerated proteasomal degradation and inhibition of its complete proteolysis. This occurs due to the accumulation of both MARCHF6, the major E3 ubiquitin ligase for SM, and squalene, which impedes SM degradation through a mechanism involving the SM-N100 regulatory domain. Taken together, our findings point towards a role for the constitutively active trunSM in adaptations to hypoxic conditions and suggest it may contribute to the oncogenic impacts of SM activity.

Results

Oxygen availability regulates SM truncation

We previously showed that SM is post-translationally regulated by its substrate squalene and pathway end-product cholesterol (Gill et al., 2011; Yoshioka et al., 2020). The enzyme also undergoes partial proteasomal degradation of its N-terminus to liberate a truncated protein (trunSM) that is cholesterol-resistant and thus constitutively active (Figure 1A; Coates et al., 2021), although physiological triggers are unknown. As SM is a rate-limiting enzyme of cholesterol synthesis and catalyzes its first oxygen-dependent reaction (Figure 1—figure supplement 1), we tested if SM protein levels are affected by oxygen availability. Incubation of HEK293T cells under hypoxic conditions (1% O2) stabilized hypoxia-inducible factor-1α (HIF1α; Figure 1B) and upregulated its target genes VEGF and CA9 (Figure 1—figure supplement 2A), confirming the induction of a hypoxic response. We also noted a striking increase in SM truncation caused by the disappearance of full-length SM and a four-fold accumulation of trunSM (Figure 1B). This led to trunSM becoming the predominant SM variant, as indicated by the elevated trunSM:SM ratio (Figure 1—figure supplement 2B). Hypoxia-induced truncation of SM increased over time (Figure 1C, Figure 1—figure supplement 2C) and according to the magnitude of oxygen deprivation (Figure 1D, Figure 1—figure supplement 2D), with the net result of increased total enzyme levels (expressed as the sum of full-length SM and trunSM levels; Figure 1—figure supplement 1C). Importantly, trunSM accumulation was greater under the severely hypoxic conditions characteristic of solid tumors (0.5–2% O2) than the ‘physoxic’ conditions experienced by normal human tissues in situ (3–7.5% O2) (Figure 1D, Figure 1—figure supplement 2D; McKeown, 2014). This suggested that increased SM truncation is a feature of pathophysiological hypoxia. We also noted our experiments, which for technical reasons used a variety of cell seeding densities, showed variation in the normoxic trunSM:SM ratio. Indeed, we confirmed SM truncation is increased at higher cell densities and accompanied by slight stabilization of HIF1α (Figure 1—figure supplement 2E), consistent with other reports (Sheta et al., 2001; Dayan et al., 2009). This further supported the phenomenon of hypoxia-induced truncation.

Figure 1 with 3 supplements see all
Download asset Open asset
Oxygen availability regulates SM truncation.

(A) Simplified overview of SM truncation. Full-length SM contains an N-terminal domain mediating feedback regulation by cholesterol. Ubiquitinated SM is targeted to the proteasome, where proteolysis is prematurely halted within the regulatory domain. This liberates a truncated protein (trunSM) that no longer responds to cholesterol and is therefore constitutively active. (B) HEK293T cells were incubated under normoxic (21% O2) or hypoxic (1% O2) conditions for 24 hr. (C) HEK293T cells were incubated under normoxic or hypoxic conditions for the indicated times. Changes in HIF1α levels over time are consistent with other reports (Jantsch et al., 2011; Bartoszewski et al., 2019). (D) HEK293T cells were incubated under the indicated oxygen concentrations for 24 hr. Each set of immunoblots was obtained in a separate experiment. (B–D) Immunoblotting was performed for SM and trunSM (red). Graphs depict densitometric quantification of SM and trunSM protein levels normalized to the normoxic condition, which was set to 1 (dotted line). In (D), oxygen concentrations considered hypoxic (hyp.), ‘physoxic’ (phys.) or normoxic (norm.) (McKeown, 2014) are indicated in blue. Data presented as mean ± SEM from n=3–4 independent experiments (**, p≤0.01; two-tailed one-sample t-test vs. hypothetical mean of 1).

Figure 1—source data 1

Uncropped immunoblots for Figure 1.

https://cdn.elifesciences.org/articles/82843/elife-82843-fig1-data1-v1.zip

We also surveyed SM levels in a panel of cell lines and found hypoxia-induced accumulation of trunSM was generalizable to all, although full-length SM levels did not decline in MDA-MB-231 breast cancer cells (Figure 1—figure supplement 2F). As HIF1α and hypoxia-inducible factor-2α (HIF2α) transcriptionally regulate the cellular response to hypoxia, we next tested if their activity is required for SM truncation. However, knockdown of HIF1A and HIF2A expression to a level sufficient to reduce target gene activation (Sena et al., 2014) had no effect on the magnitude of hypoxia-induced SM truncation in HEK293T cells (Figure 1—figure supplement 3A, B). This ruled out the involvement of HIF1α, HIF2α and their target genes in this phenomenon.

Hypoxia transcriptionally and post-translationally reduces full-length SM levels

As SM is truncated via partial proteasomal degradation (Coates et al., 2021), we reasoned that hypoxia promotes this through a two-step mechanism: (1) targeting of full-length SM to the proteasome, and (2) inhibition of its complete proteolysis. To confirm the first step of this mechanism, we investigated the reason for the decline in full-length SM levels during hypoxia. SQLE transcripts were downregulated in hypoxic HEK293T cells, as were transcripts encoding the upstream cholesterol synthesis enzyme HMGCR (Figure 2A). Downregulation of SQLE transcripts was not observed in MDA-MB-231 cells (Figure 2—figure supplement 1A), accounting for the unchanged full-length SM levels in this cell line. Although the reduction in SQLE and HMGCR transcripts in HEK293T cells likely reflected a broad transcriptional suppression of cholesterol synthesis during hypoxia, as reported previously (Dolt et al., 2007; Cao et al., 2014), the magnitude of SQLE downregulation was unlikely to fully explain the large reduction in SM protein levels (Figure 1B). Moreover, levels of a constitutively expressed SM construct ([HA]3-SM-V5) were markedly reduced during extended hypoxic incubations with no associated change in mRNA levels (Figure 2B, Figure 2—figure supplement 1B). We concluded that hypoxia reduces the levels of full-length SM through both transcriptional downregulation and accelerated post-translational degradation.

Figure 2 with 1 supplement see all
Download asset Open asset
Hypoxia transcriptionally and post-translationally reduces full-length SM levels.

(A) HEK293T cells were incubated under normoxic or hypoxic conditions for 24 hr. Levels of the indicated transcripts were quantified, normalized to the levels of RPL11, GAPDH and ACTB housekeeping transcripts and adjusted relative to the normoxic condition, which was set to 1 (dotted line). (B) HEK293T cells were transfected with (HA)3-SM-V5 for 24 hr and incubated under normoxic or hypoxic conditions for the indicated times. (C) HEK SM-N100-GFP-V5 cells were treated with or without 20 µM MG132 and 20 nM bafilomycin A1 under normoxic or hypoxic conditions for 16 hr. (D) HEK293T cells were transfected with the indicated constructs for 24 hr and incubated under normoxic or hypoxic conditions for 16 hr. (B–D) Graphs depict densitometric quantification of protein levels normalized to the respective normoxic conditions for each timepoint, treatment or construct, which were set to 1 (dotted line). (A–D) Data presented as mean ± SEM from n=3–6 independent experiments (*, p≤0.05; **, p≤0.01; [A, B] two-tailed one-sample t-test vs. hypothetical mean of 1; [C, D] two-tailed ratio paired t-test).

Figure 2—source data 1

Uncropped immunoblots for Figure 2.

https://cdn.elifesciences.org/articles/82843/elife-82843-fig2-data1-v1.zip

The basal and metabolically-regulated degradation of SM occurs through the ubiquitin-proteasome system and is mediated by the SM-N100 regulatory domain (Chua et al., 2017; Yoshioka et al., 2020). Therefore, we tested the effect of hypoxia on HEK293 cells stably expressing an SM-N100 fusion protein (SM-N100-GFP-V5). Like full-length SM, levels of SM-N100-GFP-V5 were reduced by hypoxic conditions (Figure 2C). Proteasomal inhibition using MG132 increased the levels of SM and SM-N100-GFP-V5 (Figure 2—figure supplement 1C) and blocked their hypoxia-induced degradation (Figure 2C), confirming this degradation occurs via the proteasome. We also noted that protein levels and hypoxia-induced accumulation of trunSM were ablated by MG132 (Figure 2—figure supplement 1C, D), consistent with this protein arising from partial proteasomal proteolysis of SM (Coates et al., 2021). Although hypoxia can trigger autophagy (Bellot et al., 2009), this did not play a role in SM degradation as inhibition of lysosomal acidification using bafilomycin A1 had no additive effect with MG132 (Figure 2C). To identify residues required for hypoxia-induced degradation of SM, we utilized protein constructs with mutations of previously identified ubiquitination sites. The magnitude of hypoxic degradation was blunted by disruption of Lys-82/90/100, a cluster of redundant ubiquitination sites previously found to promote truncation (Coates et al., 2021), but not by disruption of Lys-290 (Figure 2D; Hornbeck et al., 2015). Non-canonical cysteine, serine and threonine ubiquitination sites required for the cholesterol-induced degradation of SM (SM-N100 C/S/T) (Chua et al., 2019) also contributed to hypoxia-induced degradation, suggesting multiple ubiquitin signals are involved. Contrary to the expectation that loss of ubiquitination would stabilize SM, the mutation of Lys-82/90/100 reduced SM levels under normoxic conditions (Figure 2—figure supplement 1E). Therefore, this cluster of residues may be specifically involved in hypoxia-induced, rather than basal, degradation of SM.

Hypoxia-induced degradation of full-length SM requires the E3 ubiquitin ligase MARCHF6

To investigate how hypoxia promotes SM ubiquitination, we considered the possible role of proline hydroxylation. This oxygen-dependent modification, catalyzed by prolyl hydroxylases, is required for the ubiquitination and degradation of HIF1α under normoxic conditions (Ivan et al., 2001), although there is conflicting evidence for the existence of substrates beyond the HIF proteins (Cockman et al., 2019). Indeed, treatment with the prolyl hydroxylase inhibitors DMOG and FG-4592 had no effect on the basal levels nor hypoxia-induced degradation of SM and SM-N100-GFP-V5, despite stabilizing HIF1α (Figure 3—figure supplement 1A).

SM and SM-N100 are targeted for proteasomal degradation by the E3 ubiquitin ligase MARCHF6 (Foresti et al., 2013; Zelcer et al., 2014); therefore, we tested if increased MARCHF6 activity could account for the hypoxia-induced degradation of SM. To do so, we depleted MARCHF6 expression using siRNA that achieves a 60–70% reduction in transcript levels in HEK293 cells (Zelcer et al., 2014). SM and SM-N100-GFP-V5 levels were dramatically increased (Figure 3—figure supplement 1B) and the hypoxic decline in SM levels was blocked (Figure 3A), supporting the involvement of MARCHF6 in hypoxia-induced degradation. The basal levels and hypoxic accumulation of trunSM were also reduced (Figure 3—figure supplement 1B, C), consistent with MARCHF6 contributing to the proteasomal targeting, and therefore partial degradation, of SM (Coates et al., 2021). Hypoxia-induced accumulation of trunSM was not completely abolished, however, indicating SM can be truncated even when targeted to the proteasome by hypoxia-independent mechanisms. Surprisingly, there was no effect of MARCHF6 knockdown on hypoxia-induced degradation of SM-N100-GFP-V5 (Figure 3A), suggesting SM and the isolated SM-N100 domain are degraded through different proteasome-dependent routes under these conditions. We elected to further investigate the MARCHF6-dependent degradation of full-length SM, as the endogenous protein has greater physiological relevance.

Figure 3 with 1 supplement see all
Download asset Open asset
Hypoxia-induced degradation of full-length SM requires the E3 ubiquitin ligase MARCHF6.

(A) HEK SM-N100-GFP-V5 cells were transfected with control or MARCHF6 siRNA for 24 hr and incubated under normoxic or hypoxic conditions for 16 hr. (B) HEK MARCHF6-V5 cells were incubated under normoxic or hypoxic conditions for the indicated times. MARCHF6-V5 appears as two bands that were quantified collectively, as we have done previously (Sharpe et al., 2019). (A, B) Graphs depict densitometric quantification of protein levels normalized to the respective normoxic condition for each siRNA or timepoint, which were set to 1 (dotted line). Data presented as mean ± SEM from n=3–4 independent experiments (*, p≤0.05; **, p≤0.01; [A] two-tailed ratio paired t-test; [B] two-tailed one-sample t-test vs. hypothetical mean of 1).

Figure 3—source data 1

Uncropped immunoblots for Figure 3.

https://cdn.elifesciences.org/articles/82843/elife-82843-fig3-data1-v1.zip

To study MARCHF6 levels in hypoxic cells, we used a previously generated HEK293 cell line stably expressing a V5-tagged form of the protein. This construct was used due to the poor performance of endogenous MARCHF6 antibodies (Sharpe et al., 2019) and to eliminate transcriptional effects on protein levels. We examined the response of MARCHF6-V5 to hypoxia and found it accumulated during prolonged hypoxic incubations, which correlated with the maximal decline in full-length SM levels (Figure 3B). Therefore, increased MARCHF6 protein levels and activity likely account for the accelerated ubiquitination and degradation of SM during hypoxia.

Hypoxia-induced squalene accumulation promotes partial degradation of SM

Having established that SM undergoes accelerated proteasomal degradation during hypoxia, we next investigated how low oxygen levels favor its partial rather than complete proteolysis to yield trunSM. As there is extensive precedent for metabolic regulation of cholesterol synthesis enzymes and the pathway contains multiple oxygen-dependent reactions, we considered if accumulation of a pathway intermediate might underlie this phenomenon. Hypoxia-induced accumulation of trunSM occurred in cells incubated under both lipoprotein-replete and lipoprotein-deficient conditions, in which the cholesterol synthesis pathway is active (Figure 4A). However, the magnitude of this accumulation was diminished when lipoprotein-deficient cells were co-treated with a statin to inhibit HMGCR and the early cholesterol synthesis pathway. By contrast, there was no effect of sterol depletion on the hypoxia-induced reduction in full-length SM. This indicated that an intermediate or end-product of cholesterol synthesis promotes the partial rather than complete degradation of SM at the proteasome. We therefore turned our attention to the SM substrate squalene, as it allosterically regulates SM degradation (Yoshioka et al., 2020) and is the substrate for the first oxygen-dependent step of cholesterol synthesis. Squalene accumulated over the course of a hypoxic incubation (Figure 4B, Figure 4—figure supplement 1), consistent with reduced SM activity under low-oxygen conditions. This accumulation was strikingly well-correlated with the previously observed increase in trunSM levels (Figure 4C), suggesting the two effects may be linked.

Figure 4 with 4 supplements see all
Download asset Open asset
Hypoxia-induced squalene accumulation promotes partial degradation of SM.

(A) HEK293T cells were incubated in medium containing fetal calf serum (FCS), lipoprotein-deficient FCS (LPDS) or LPDS containing 5 µM mevastatin and 50 µM mevalonolactone (LPDS +statin) for 8 hr, refreshed in their respective medium and incubated under normoxic or hypoxic conditions for 16 hr. (B) HEK293T cells were incubated under normoxic or hypoxic conditions for the indicated times. Non-saponifiable lipids were extracted, and squalene levels were determined using gas chromatography-mass spectrometry and adjusted relative to the normoxic condition, which was set to 1 (dotted line). The maximal squalene level detected was 0.66±0.12 ng per µg of total protein. (C) Pearson correlation between squalene levels in (B) and trunSM levels in Figure 1C. Blue line indicates linear regression. (D) HEK293T cells were treated with or without 300 µM squalene (squ.), monooxidosqualene (MOS) or dioxidosqualene (DOS) for 16 hr. (E) HEK293T SQLE-knockout (SQLE-KO) clone 10 (c10) cells were transfected with the indicated constructs for 24 hr, then treated with or without 1 µM NB-598 or 300 µM squalene for 16 hr. (A, D, E) Graphs depict densitometric quantification of trunSM or truncated protein levels normalized to the (A) respective normoxic conditions for each serum type or (D, E) vehicle conditions, which were set to 1 (dotted line). (A–E) Data presented as mean ± SEM from n=3–5 independent experiments (*, p≤0.05; **, p≤0.01; [A] two-tailed ratio paired t-test; [D, E] two-tailed one-sample t-test vs. hypothetical mean of 1).

Figure 4—source data 1

Uncropped immunoblots for Figure 4.

https://cdn.elifesciences.org/articles/82843/elife-82843-fig4-data1-v1.zip

Delivery of exogenous squalene induced trunSM accumulation in normoxic HEK293T and Huh7 cells (Figure 4D, Figure 4—figure supplement 2A, B), confirming its ability to promote partial degradation of SM. Accumulation of trunSM was also induced by the oxygenated squalene derivatives monooxidosqualene and dioxidosqualene (Figure 4D) but not by its saturated analogue squalane (Figure 4—figure supplement 2C), which has similar biophysical properties (Hauß et al., 2002). This indicated truncation is promoted by squalene and its structurally related molecules in a specific manner, rather than through bulk membrane effects caused by lipid accumulation. To address the possibility that exogenous squalene is converted to a downstream product responsible for truncation, we generated SQLE-knockout HEK293T cells (Figure 4—figure supplement 3) and transfected them with a catalytically inactive SM Y195F mutant (Padyana et al., 2019) to prevent the metabolism of added squalene. The truncated form of the Y195F mutant accumulated upon squalene treatment in SQLE-knockout cells, confirming squalene alone can directly induce truncation (Figure 4—figure supplement 2D). There was no significant accumulation of the truncated fragment in cells transfected with wild-type SM, likely due to clearance of exogenous squalene by the overexpressed protein and downstream enzymes.

To confirm if endogenously synthesized squalene is sufficient to trigger SM truncation, cells were treated with inhibitors of the relevant cholesterol synthesis enzymes (Figure 4—figure supplement 2A). The SM inhibitor NB-598 was excluded because of its ability to induce truncation through direct binding and stabilization of the SM catalytic domain that renders it resistant to proteasomal unfolding (Padyana et al., 2019; Coates et al., 2021). Inhibiting squalene synthesis from farnesyl diphosphate (TAK-475) increased trunSM levels but abolished its hypoxia-induced accumulation, whereas significant accumulation still occurred under conditions where squalene synthesis was preserved: inhibition of lanosterol synthesis from monooxidosqualene (BIBB 515), or inhibition of lanosterol demethylation (GR70585X) (Figure 4—figure supplement 2E). This confirmed that lanosterol, which also accumulates during hypoxia (Nguyen et al., 2007), has no effect on SM truncation. We further noted that the inhibition of squalene or lanosterol synthesis, but not lanosterol demethylation, increased the levels of full-length SM and SM-N100-GFP-V5 under normoxic conditions. This was consistent with our previous finding that farnesyl-containing molecules, including monooxidosqualene, dioxidosqualene and a squalene-derived photoaffinity probe, stabilize SM via its regulatory domain in a similar manner to squalene itself (Yoshioka et al., 2020). The increase in normoxic trunSM levels upon treatment with TAK-475 and BIBB 515 (Figure 4—figure supplement 2E) suggested farnesyl-containing cholesterol synthesis intermediates can also induce SM truncation. Nevertheless, as the primary substrate of oxygen-dependent SM catalysis, squalene is likely to be the major driver of this process under hypoxic conditions.

SM contains two known squalene binding sites: the SM-N100 regulatory domain and the active site of the catalytic domain (Yoshioka et al., 2020). As the SM inhibitor NB-598 induces SM truncation by binding and stabilizing the catalytic domain (Coates et al., 2021), we considered if squalene exerts its effects on truncation through a similar mechanism. To eliminate the contribution of the SM-N100 domain, we transfected SQLE-knockout cells with an ectopic form of trunSM (SM[ΔN65]-V5). Consistent with past findings (Yoshioka et al., 2020; Padyana et al., 2019), this construct was stabilized by NB-598 (Figure 4E). However, significant stabilization did not occur when cells expressing an inactive SM(ΔN65)-V5 mutant (Y195F) were treated with squalene. We concluded that squalene promotes SM truncation via the SM-N100 regulatory domain, rather than direct binding and stabilization of the SM catalytic domain. Previous work showed squalene directly binds the SM-N100 domain (Yoshioka et al., 2020), with the hydrophobic re-entrant loop (residues ~ 15–40) as the most logical interaction site. Aromatic residues and leucine residues line the SM active site and are required for catalysis (Padyana et al., 2019; Abe et al., 2007), suggesting they directly interact with squalene. Similar residues in the SM-N100 re-entrant loop may likewise be involved in squalene binding, and possibly the partial degradation of SM. We therefore mutated phenylalanine and leucine residues in the re-entrant loop to test if they are required for SM truncation. However, these mutations, either alone or in combination, did not prevent the hypoxia-induced accumulation of truncated SM (Figure 4—figure supplement 4). This indicated that partial degradation of SM is independent of these putative squalene-binding residues, with further work required to clarify the mechanism.

SM activity is preserved during hypoxia

trunSM has a long half-life and is constitutively active (Coates et al., 2021). Therefore, we hypothesized that hypoxia-induced truncation of SM is a compensatory mechanism to preserve enzymatic activity during oxygen shortfalls. To investigate this possibility, Huh7 cells were radiolabeled with [14C]-acetate and downstream flux through cholesterol synthesis was assayed using thin-layer chromatography. This cell line was chosen due to its tissue of origin, the liver, being a major site of cholesterol synthesis (Turley et al., 1995), as well as its robust hypoxia-induced truncation of SM (Figure 1—figure supplement 2F). Cholesterol synthesis was markedly reduced in hypoxic cells (Figure 5), as expected given the high oxygen demand of the pathway. Acute lanosterol accumulation was detectable within 6 hr, consistent with previous reports (Nguyen et al., 2007; Kucharzewska et al., 2015), and continued over the duration of the hypoxic incubation. By contrast, upstream accumulation of squalene was minimal even after 24 hr. This confirmed that despite its dependence on oxygen, SM activity is preserved during hypoxia, likely due to its truncation to a degradation-resistant form.

Figure 5
Download asset Open asset
SM activity is preserved during hypoxia.

Huh7 cells were labelled with 2 µCi [14C]-acetate and incubated under normoxic or hypoxic conditions for the indicated times. Synthesis of [14C]-squalene, [14C]-lanosterol, and [14C]-cholesterol was determined using thin-layer chromatography. Double-headed arrows indicate multiple enzymatic steps, and arrowhead on right of image indicates band corresponding to the SM product [14C]-monooxidosqualene (Capell-Hattam et al., 2022). Graphs depict densitometric quantification of [14C]-cholesterol and [14C]-lanosterol levels normalized to the conditions with the greatest abundance of each analyte, which were set to 1 (dotted line). Data presented as mean ± SEM from n=3 independent experiments (*, p≤0.05; two-tailed ratio paired t-test vs. respective normoxic condition for each timepoint).

Figure 5—source data 1

Uncropped thin-layer chromatography image for Figure 5.

https://cdn.elifesciences.org/articles/82843/elife-82843-fig5-data1-v1.zip

Discussion

Cholesterol synthesis is tightly regulated by metabolic supply and demand, with the lipid-sensing and rate-limiting enzyme SM a key point at which this regulation is exerted. We previously showed that partial proteasomal degradation of SM produces a truncated and constitutively active form of the enzyme (Coates et al., 2021). In this study, we identified hypoxia as a physiological trigger for trunSM formation (Figure 1). Hypoxia-induced truncation occurs through a two-part mechanism: (1) increased levels of MARCHF6, an E3 ubiquitin ligase that targets SM to the proteasome (Figure 2, Figure 3), and (2) accumulation of squalene, which impedes the complete degradation of SM and yields trunSM (Figure 4). Truncation of SM preserves its activity and facilitates downstream pathway flux during hypoxia (Figure 5). Taken together, our results point towards SM truncation as an adaptive mechanism to clear excess substrate, as well as a likely contributor to the widely reported oncogenic properties of SM.

Cholesterol synthesis during hypoxia

Hypoxia places great strain on metabolic processes and necessitates the strict allotment of available oxygen and energy reserves. Cholesterol synthesis is a particularly resource-intensive pathway requiring eleven oxygen molecules and over one hundred ATP equivalents per molecule of product. However, there are conflicting reports on changes in overall flux from acetyl-CoA to cholesterol during hypoxia (Mukodani et al., 1990; Parathath et al., 2011), suggestive of cell type-specific responses. The small number of studies into individual biosynthetic enzymes nevertheless indicate their activity is suppressed by hypoxia at multiple regulatory levels, and this is supported by the accumulation of various pathway intermediates (Wu et al., 2020; Nguyen et al., 2007; Kucharzewska et al., 2015). Lanosterol 14α-demethylase, which requires three oxygen molecules for catalysis, is transcriptionally downregulated by HIF2α and the hypoxia-induced long non-coding RNA lincNORS, contributing to the characteristic accumulation of lanosterol under hypoxic conditions (Wu et al., 2020; Zhu et al., 2014). Lanosterol in turn triggers ubiquitin-dependent degradation of the early cholesterol synthesis enzyme HMGCR (Nguyen et al., 2007), suppressing further oxygen consumption by the pathway. Our study establishes that oxygen availability also regulates SM, a rate-limiting enzyme of cholesterol synthesis and the first to require molecular oxygen.

We found that SM is transcriptionally downregulated in hypoxic HEK293T cells but not MDA-MB-231 cells, consistent with previously reported cell-type specific changes in SQLE expression (Haider et al., 2016). This accounted in part for the hypoxia-induced decline in full-length SM levels, although reduced SQLE translation through mechanisms such as mTOR suppression (Liu et al., 2006) cannot be ruled out as a contributing factor. We also found that like HMGCR, the SM protein is targeted for proteasomal degradation during hypoxia. However, in stark contrast to HMGCR, the net result of SM degradation under these conditions is the preservation of, or even an increase in, the number of enzyme molecules available for catalysis. Indeed, SM activity is preserved during hypoxia, with only minimal accumulation of its substrate squalene. This is enabled by increased partial proteolysis of SM to form trunSM, which lacks a functional SM-N100 regulatory domain and has a dramatically extended half-life (Coates et al., 2021). Disruption of the SM-N100 domain also renders trunSM resistant to cholesterol-induced degradation, which is central to the metabolic regulation of full-length SM (Gill et al., 2011; Coates et al., 2021). Therefore, hypoxia-induced SM truncation ensures total enzyme levels remain constant even under cholesterol-replete conditions that typically trigger its degradation.

Although preservation of oxygen-dependent SM activity during hypoxia appears paradoxical, there are numerous advantages (Figure 6). During transient or low-level hypoxia, it likely promotes compensatory cholesterol synthesis and maintenance of cell viability. Supporting this idea, hypoxia-induced cell death is exacerbated by SM inhibition (Haider et al., 2016). Furthermore, the longevity of trunSM would enable rapid resumption of pathway activity when normoxia is restored. During long-term or severe hypoxia, where there is insufficient oxygen for flux through downstream cholesterol synthesis, the role of trunSM in hypoxia may shift towards efficient clearance of squalene. While generally considered an inert intermediate, excess squalene induces ER stress and is toxic in cells lacking SM activity or the ability to sequester squalene to lipid droplets (Hong et al., 2022; Mahoney et al., 2019). A secondary effect of squalene clearance is its downstream conversion to lanosterol, which accelerates degradation of HMGCR (Nguyen et al., 2007). We and others (Nguyen et al., 2007; Kucharzewska et al., 2015) observed a stark contrast between acute lanosterol accumulation and minimal squalene accumulation during hypoxia, supporting both of these putative functions for trunSM. Thus, hypoxia-induced SM truncation is likely critical for curtailing HMGCR activity and flux through oxygen-intensive cholesterol synthesis. Stabilization of MARCHF6, which targets lanosterol 14α-demethylase for degradation (Scott et al., 2020), may also contribute to hypoxic lanosterol accumulation. SM activity also promotes synthesis of dioxidosqualene and ultimately 24(S),25-epoxycholesterol, a potent suppressor of cholesterol accretion (Wong et al., 2008). However, these are yet to be studied in the context of hypoxia.

Figure 6 with 1 supplement see all
Download asset Open asset
Model of hypoxia-induced SM truncation.

(A, B) Hypoxic conditions stabilize the E3 ubiquitin ligase MARCHF6 and inhibit SM-catalyzed conversion of squalene to monooxidosqualene, leading to squalene accumulation. (C) Increased MARCHF6 activity promotes ubiquitination of SM and its targeting to the proteasome. (D) Squalene impedes the complete proteasomal degradation of SM via a mechanism involving the SM-N100 domain, (E) yielding the constitutively active trunSM. During transient or low-level hypoxia, trunSM activity may facilitate continued cholesterol synthesis to compensate for the oxygen shortfall. During long-term or severe hypoxia, trunSM activity may reduce squalene-induced toxicity and promote downstream synthesis of lanosterol, which suppresses an early step of the cholesterol synthesis pathway. During pathophysiological hypoxia, cholesterol synthesis enabled by trunSM may contribute to oncogenic cell growth and survival.

Molecular mechanism of truncation

The first step of SM truncation is its delivery to the proteasome (Coates et al., 2021). This is enhanced during hypoxia by a novel ubiquitin signal at the Lys-82/90/100 cluster and accumulation of MARCHF6, which facilitates the basal and metabolically-regulated degradation of SM (Foresti et al., 2013; Zelcer et al., 2014). Non-canonical ubiquitination sites required for cholesterol-induced SM degradation (Chua et al., 2019) also contribute to its hypoxic degradation, which can be reconciled by a combination of increased MARCHF6 levels and the sterol-replete conditions in which these experiments were performed. The mechanism by which hypoxia stabilizes MARCHF6 requires further study. Interestingly, NADPH binds MARCHF6 and activates its E3 ubiquitin ligase activity (Nguyen et al., 2022). Hypoxic cells upregulate NADPH production to counter oxidative stress (Samanta et al., 2016), which may contribute to MARCHF6-dependent proteasomal targeting of SM under these conditions.

A key finding of this study is that upon hypoxic delivery of SM to the proteasome, accumulated squalene inhibits its complete degradation and liberates the constitutively active trunSM. This feedforward mechanism, by which the substrate of SM preserves its activity, is mediated by the SM-N100 regulatory domain. It also functions under normoxic conditions, where it likely buffers against transient squalene fluctuations. Hypoxia-induced trunSM accumulation persists even when MARCHF6 is depleted; therefore, squalene also impedes SM degradation when other E3 ubiquitin ligases target it to the proteasome. These other regulators, and the factors controlling their ubiquitination of SM, await future discovery and may shed light on how the SM-N100 domain undergoes MARCHF6-independent proteasomal degradation during hypoxia. Taken together, our findings expand on the ability of the SM-N100 domain to sense ER membrane lipids and regulate SM degradation (Figure 6—figure supplement 1).

Although hypoxia and lipid accumulation trigger ER stress and changes to proteostasis (Bi et al., 2005), the ability of squalene and other farnesyl-containing compounds to allosterically bind the SM-N100 domain (Yoshioka et al., 2020) strongly suggests truncation is induced directly by squalene, rather than bulk membrane effects. The specificity of this effect is supported by the inability of the saturated squalene analog, squalane, to induce truncation. Squalene is the most likely of the farnesyl-containing cholesterol synthesis intermediates to accumulate under both normal and hypoxic conditions, owing to the rate-limiting and oxygen-dependent nature of SM activity. However, our current and previous (Yoshioka et al., 2020) data suggest a farnesyl group alone is sufficient to promote truncation. Precisely how this lipophilic moiety interferes with SM degradation is unknown, although interaction with the membrane-embedded SM-N100 re-entrant loop (Howe et al., 2015) is a likely possibility. We ruled out putative squalene-binding residues within this loop as a requirement for hypoxia-induced truncation, although a stronger understanding of the exact squalene binding site in SM-N100 is needed to further test its role in SM truncation.

Previously, we reported farnesyl-containing compounds also blunt the MARCHF6-mediated ubiquitination of SM-N100, leading to stabilization of full-length SM (Yoshioka et al., 2020). The existence of dual mechanisms enabling squalene to sustain SM activity is intriguing, particularly as trunSM formation is itself dependent on ubiquitination (Coates et al., 2021). One possibility is that truncation is stimulated at a lower squalene threshold than the inhibition of SM ubiquitination, allowing for a biphasic response to accumulating substrate. We found squalene is at the lower limit of detection in normoxic cells, and only the highly sensitive gas chromatography-mass spectrometry method could detect its accumulation during acute hypoxia. The strong trunSM-squalene correlation thus suggests truncation is triggered by very small quantities of squalene. Another possibility is that squalene-induced truncation is a ‘failsafe’ mechanism to preserve SM activity. This may be necessary when a reduction in MARCHF6-mediated ubiquitination is insufficient to completely prevent targeting of SM to the proteasome. It may also apply when SM ubiquitination is promoted by other, simultaneous cellular stimuli. For instance, hypoxic MARCHF6 stabilization may override the inhibitory effects of accumulated squalene.

Physiological consequences of truncation

Overexpression and overactivity of SM occurs in a wide range of malignancies including hepatocellular carcinoma and prostate cancer, where it is positively correlated with severity and lethality (Liu et al., 2018; Kalogirou et al., 2021). As hypoxia is common in the interior of solid tumors and often associated with poor prognosis (Rankin and Giaccia, 2016), trunSM formation may contribute to oncogenesis. Hypoxia also occurs within atherosclerotic plaques (Marsch et al., 2013). Truncation of SM is yet to be specifically studied in human tissues, and would be worthwhile to test in cancers with a propensity for both hypoxia and SM overexpression, such as prostate and pancreatic cancer (McKeown, 2014). trunSM retains full catalytic activity (Coates et al., 2021) and can likely fulfil cholesterol synthesis-dependent functions of SM in oncogenesis. However, its competency in the suite of other oncogenic SM activities is unknown. These include activation of ERK signaling (He et al., 2021) and interactions with carbonic anhydrase 3 (Liu et al., 2021), GSK3β and p53 (Jun et al., 2021). Cholesterol-independent functions such as these are likely to be most critical during severe hypoxia, where cholesterol synthesis cannot proceed.

Finally, although this study was largely performed using extremely low oxygen levels characteristic of pathophysiological hypoxia, we also observed that SM truncation was responsive to a range of ‘physoxic’ oxygen levels that occur within normal tissues in situ. This reinforces the concept of truncation as a buffer against normal substrate fluctuations, which may be important for preventing squalene-induced dysfunction in tissues with low oxygen perfusion, such as the brain (McKeown, 2014). Squalene fluctuations occur diurnally (Miettinen, 1982) and with changing sterol status (Gill et al., 2011), and its levels vary dramatically between different tissues (Liu et al., 1976). Thus, the ability of squalene to stimulate the truncation and constitutive activation of SM may play a key role in regulating flux through cholesterol synthesis under various physiological conditions.

Materials and methods

Key resources table
Reagent type
(species) or
resource		DesignationSource or referenceIdentifiersAdditional information
Gene (Homo sapiens)SQLERefSeq6713, NM_003129.4
Cell line (H. sapiens)HEK293TGift from UNSW School of
Medical Sciences (UNSW
Sydney, Australia)		Highly transfectable human embryonic kidney cells
Cell line (H. sapiens)HCT116Gift from Dr Ewa Goldys
(UNSW Sydney, Australia)		Epithelial colorectal carcinoma cells
Cell line (H. sapiens)Huh7Gift from Centre for
Cardiovascular Research
(UNSW Sydney, Australia)		Epithelial-like hepatocellular carcinoma cells
Cell line (H. sapiens)HeLaTGift from Drs Louise Lutze-
Mann and Noel Whitaker
(UNSW Sydney, Australia)		Highly transfectable cervical adenocarcinoma cells
Cell line (H. sapiens)MDA-MB-231Gift from Drs Louise Lutze-Mann
and Noel Whitaker (UNSW
Sydney, Australia)		Epithelial breast
adenocarcinoma cells
Cell line (H. sapiens)HEK SM-N100-GFP-V5Generated previously using the
Flp-In T-REx system
(Zelcer et al., 2014)		HEK293 cells stably expressing the N-terminal 100 amino acids of SM (SM-N100) fused with green fluorescent protein and a V5 epitope tag. Controlled by a cytomegalovirus (CMV) promoter.
Cell line (H. sapiens)HEK MARCHF6-V5Generated previously using the
Flp-In T-REx system
(Sharpe et al., 2019)		HEK293 cells stably expressing MARCHF6 fused with a V5 epitope tag. Controlled by a CMV promoter.
Cell line (H. sapiens)HEK293T SQLE-knockout
clone 10 (c10)		This studySee Materials and Methods, Generation of SM knockout cells
Cell line (H. sapiens)HEK293T SQLE-knockout
clone 12 (c12)		This studySee Materials and Methods, Generation of SM knockout cells
Cell line (H. sapiens)HEK293T SQLE-knockout
clone 14 (c14)		This studySee Materials and Methods, Generation of SM knockout cells
Transfected
construct (H. sapiens)		MISSION universal
negative control #1 siRNA		Sigma-AldrichSIC001
Transfected
construct (H. sapiens)		MARCHF6 siRNASigma-AldrichSASI_Hs01_00105239
Transfected
construct (H. sapiens)		HIF1A siRNASigma-AldrichSASI_Hs01_00332063
Transfected
construct (H. sapiens)		EPAS1 (HIF2A) siRNASigma-AldrichSASI_Hs01_00019152
AntibodyAnti-SM(SQLE) (rabbit polyclonal)Proteintech12544-1-AP4 °C for 16 hr (1:2500)
AntibodyAnti-HIF1α (rabbit polyclonal)Proteintech20960-1-APRoom temperature for 1 hr (1:1000)
AntibodyAnti-GAPDH (rabbit monoclonal)Cell Signaling Technology21184 °C for 16 hr (1:2000)
AntibodyAnti-V5 (mouse monoclonal)Thermo Fisher ScientificR960-25Room temperature for 1 hr (1:5000)
AntibodyIRDye 680RD anti-
rabbit IgG (donkey polyclonal)		LI-COR BiosciencesLCR-926-68073Room temperature for 1 hr (1:5000)
AntibodyIRDye 800CW anti-
mouse IgG (donkey polyclonal)		LI-COR BiosciencesLCR-926-32212Room temperature for 1 hr (1:10,000)
AntibodyPeroxidase-conjugated
AffiniPure anti-rabbit
IgG (donkey polyclonal)		Jackson ImmunoResearch Laboratories711-035-152Room temperature for 1 hr (1:10,000)
AntibodyPeroxidase-conjugated
AffiniPure anti-mouse
IgG (donkey polyclonal)		Jackson ImmunoResearch Laboratories715-035-150Room temperature for 1 hr (1:10,000)
Commercial
assay or kit		Lipofectamine RNAiMAX
transfection reagent		Thermo Fisher Scientific13778150
Commercial
assay or kit		Lipofectamine 3000
transfection reagent		Thermo Fisher ScientificL3000001
Commercial
assay or kit		Bicinchoninic acid assay kitThermo Fisher Scientific23225
Commercial
assay or kit		TRI reagentThermo Fisher ScientificAM9738
Commercial
assay or kit		SuperScript III First-
Strand Synthesis kit		Thermo Fisher Scientific18080051
Commercial
assay or kit		PureLink Genomic
DNA Mini kit		Thermo Fisher ScientificK182001
Commercial
assay or kit		QuantiNova SYBR
Green PCR kit		Qiagen208052
Commercial
assay or kit		Immobilon Western chemiluminescent
HRP substrate		MilliporeWBKLS0500
Chemical
compound, drug		[14C]-Acetic
acid sodium salt		PerkinElmerNEC084H001MC
Chemical
compound, drug		2,3,22,23-DioxidosqualeneEchelon BiosciencesS-0302
Chemical
compound, drug		2,3-Oxidosqualene (mono-oxidosqualene)Echelon BiosciencesS-0301
Chemical
compound, drug		5α-CholestaneSigma-AldrichC8003
Chemical
compound, drug		Bafilomycin A1Sigma-AldrichB1793
Chemical
compound, drug		BIBB 515Cayman Chemical Company10010517
Chemical
compound, drug		DMOGSigma-AldrichD3695
Chemical
compound, drug		FG-4592Cayman
Chemical Company		15294
Chemical
compound, drug		GR70585XGlaxoSmithKleinN/A
Chemical
compound, drug		MevalonolactoneSigma-AldrichM4667
Chemical
compound, drug		MevastatinSigma-AldrichM2537
Chemical
compound, drug		MG132Sigma-AldrichC2211
Chemical
compound, drug		NB-598ChemsceneCS-1274
Chemical
compound, drug		SqualaneSigma-Aldrich234311
Chemical
compound, drug		SqualeneSigma-AldrichS3626
Chemical
compound, drug		TAK-475Sigma-AldrichSML2168
Chemical
compound, drug		N,O-Bis-(trimethylsilyl)
trifluoroacetamide		SupelcoT6381
Software, algorithmThermo Xcalibur softwareThermo Fisher Scientificv2.2 SP1.48
Software, algorithmImage Studio Lite softwareLI-COR Biosciencesv5.2.5
Software, algorithmGraphPad Prism softwareGraphPad Software Incv9.0
OtherOpti-MEM I reduced
serum medium		Thermo Fisher Scientific31985062Used to deliver transfection complexes. See Materials
and Methods, siRNA
and Plasmid Transfection

Cell culture

Request a detailed protocol

The cell lines used in this study are listed in the Key Resources Table and were routinely tested to ensure they were mycoplasma-free. Cells were maintained in a humidified Heraeus BB 15 incubator at 37 °C, 5% CO2, and 21% O2 (normoxia) in maintenance medium (DMEM-HG, 10% [v/v] fetal calf serum [FCS], 100 U/mL penicillin, and 100 μg/mL streptomycin). To improve HEK293 and HEK293T cell surface adhesion, culture vessels were treated with 25 μg/mL polyethyleneimine in phosphate-buffered saline (PBS) for 15 min at 37 °C prior to cell seeding. Plasmid and siRNA transfections were performed in maintenance medium lacking penicillin and streptomycin. Sterol depletions were performed in maintenance medium containing lipoprotein-deficient serum (LPDS, 30 mg/mL protein), which was prepared from FCS by density gradient centrifugation and dialysis, as described in Goldstein et al., 1983. Hypoxic incubations at 0.5–10% O2 were performed in a humidified Binder CB 150 incubator at 37 °C and 5% CO2. For all treatments (listed in the Key Resources Table), appropriate solvent controls were used (water [DMOG]; ethanol (mevalonolactone, GR70585X); DMSO [MG132, bafilomycin A1, FG-4592, mevastatin, TAK-475, BIBB-515, NB-598]; dimethyl sulfoxide containing 1% [v/v] Tween-20 [squalene, monooxidosqualene, dioxidosqualene, squalane]) and the final concentration of solvent did not exceed 1% (v/v) in cell culture medium. Treatments were delivered in full medium refreshes, and all experiments were 48–72 hr in duration.

Plasmids

Plasmids encoding Cas9 and SQLE-targeting guide RNAs were generated by BbsI cloning into a PX458 vector as described in Ran et al., 2013. Amino acid substitutions within expression vectors were generated using the overlap extension cloning method, as described previously (Stevenson et al., 2013). The identity of all plasmids was confirmed via Sanger sequencing. The plasmids used in this study are listed in Appendix 1—table 1, and the primer sequences used for DNA cloning are listed in Appendix 1—table 2.

siRNA and plasmid transfection

Request a detailed protocol

To downregulate gene expression or transiently overexpress proteins, cells were seeded into 12-well plates. The next day, cells were transfected with 15 pmol siRNA using Lipofectamine RNAiMAX (15 pmol siRNA: 2 µL reagent) or 1 µg expression vector using Lipofectamine 3000 (1 µg DNA: 2 µL reagent with 2 µL P3000 supplemental reagent), delivered in Opti-MEM I reduced serum medium. After 24 hr, cells were refreshed in maintenance medium and treated as specified in figure legends. The siRNA used in this study are listed in the Key Resources Table.

Protein harvest and immunoblotting

Request a detailed protocol

To quantify protein levels, cells were seeded into 6- or 12-well plates and treated as specified in figure legends. For detection of SM, total protein was harvested in 2% SDS lysis buffer (10 mM Tris-HCl [pH 7.6], 2% [w/v] SDS, 100 mM sodium chloride, 2% [v/v] protease inhibitor cocktail), passed through a 21-gauge needle until homogenous, and vortexed at room temperature for 20 min. For detection of MARCHF6-V5, cells were scraped in ice-cold PBS, pelleted, and lysed in modified RIPA buffer (50 mM Tris-HCl [pH 8.0], 0.1% [w/v] SDS, 1.5% [w/v] IGEPAL CA-630, 0.5% [w/v] sodium deoxycholate, 150 mM sodium chloride, 2 mM magnesium chloride, 2% [v/v] protease inhibitor cocktail), passed 20 times through a 22-gauge needle, rotated at 4 °C for 30 min, and centrifuged at 17,000 g and 4 °C for 15 min to obtain the supernatant. Lysate protein content was quantified using the bicinchoninic acid (BCA) assay, and sample concentrations were normalized by dilution in the appropriate lysis buffer and 0.25 vol 5×Laemmli buffer (250 mM Tris-HCl [pH 6.8], 10% [w/v] SDS, 25% [v/v] glycerol, 0.2% [w/v] bromophenol blue, 5% [v/v] β-mercaptoethanol). For SM detection, normalized samples were heated at 95 °C for 5 min.

Proteins were separated on 10% (w/v) Tris-glycine SDS-PAGE gels (prepared in-house), electroblotted onto nitrocellulose membranes, and blocked in 5% (w/v) skim milk powder in PBS containing 0.1% (v/v) Tween-20 (PBST). Immunoblotting was performed using the antibodies listed in the Key Resources Table, which were diluted in 5% (w/v) bovine serum albumin in PBST containing 0.02% (w/v) sodium azide, except for anti-HIF1α and peroxidase-conjugated antibodies, which were diluted in 5% (w/v) skim milk powder in PBST. Fluorescence-based detection of SM, trunSM, GAPDH, SM-N100-GFP-V5, and (HA)3-SM-V5 was performed using an Odyssey CLx imager (LI-COR Biosciences), and enhanced chemiluminescence-based detection of HIF1α and MARCHF6-V5 was performed using Immobilon Western chemiluminescent HRP substrate (Millipore) and an ImageQuant LAS 500 imager (Cytiva Life Sciences). Densitometry analysis was performed using Image Studio Lite software.

RNA harvest and qRT-PCR

Request a detailed protocol

To quantify gene expression, cells were seeded into 12-well plates and treated as specified in figure legends. Total RNA was harvested using TRI reagent and polyadenylated RNA was reverse transcribed using the SuperScript III First Strand Synthesis kit. cDNA products were used as the template for quantitative reverse transcription-PCR (qRT-PCR) in technical triplicate using the QuantiNova SYBR Green PCR kit and primers listed in Appendix 1—table 2. mRNA levels were normalized to the geometric mean of RPL11, GAPDH, and ACTB for hypoxia experiments, or PBGD for validation of siRNA knockdowns and gene knockout, using the comparative CT method (Schmittgen and Livak, 2008). Normalized data were adjusted relative to the control condition, as specified in figure legends.

Gas chromatography-mass spectrometry

Request a detailed protocol

To quantify cellular squalene levels, gas chromatography-mass spectrometry was performed as described previously (Yoshioka et al., 2020). Briefly, cells in 6-well plates were lysed in 0.05 M sodium hydroxide, total protein was quantified using the BCA assay, and samples were adjusted to the lowest protein concentration using 0.05 M sodium hydroxide plus 4 µg 5α-cholestane as an internal standard, in a total volume of 1 mL. Lysates were mixed with 1 mL 100% (v/v) ethanol, 500 µL 75% (w/v) potassium hydroxide, 1 µL 20 mM butylated hydroxytoluene, and 20 µL 20 mM EDTA, and saponified at 70 °C for 1 hr. Non-saponifiable lipids were extracted by mixing with 1 mL 100% (v/v) ethanol and 2.5 mL hexane, centrifuging at 4,000 g for 5 min, and collection of 2 mL of the organic phase. Lipids were dried in a vacuum centrifuge, resuspended in 50 µL N,O-bis(trimethylsilyl)-trifluoroacetamide, and derivatized at 60 °C for 1 hr.

Derivatized lipids (1.5 µL) were injected via a heated (300 °C) splitless with surge (38.0 psi for 0.50 min) inlet into a Thermo Trace gas chromatograph fitted with a Trace TR-50MS GC column (60 m×0.25 mm, 0.25 µm film thickness, Thermo Fisher). Analytes were separated with helium as the carrier gas at a constant flow of 1.2 ml/min with vacuum compensation, and temperature programming as follows: 70 °C for 0.70 min, 20 °C/min to 250 °C, 3 °C/min to 270 °C, 1.5 °C/min to 315 °C, then hold for 10 min. The GC column was coupled to a Thermo DSQIII mass spectrometer, with a transfer line temperature of 320 °C and an ion source temperature of 250 °C. For mass spectrometry analysis, the electron energy was 70 eV, the emission current was 130 µA and the detector gain was 3.0×105. Squalene and 5α-cholestane standards were analyzed in scan mode (34–600 Da) to identify peaks and retention times, and identity was confirmed using the National Institute of Standards and Technology databases. Experimental samples were analyzed in selective ion monitoring mode to detect squalene (m/z=81.0, 410.4) and 5α-cholestane (m/z=149.1, 217.2, 372.4), with a detection width of 0.1 and dwell time of 200 ms. Chromatograph peaks were integrated using Thermo Xcalibur software and the peak area of squalene was normalized to the 5α-cholestane internal standard. A squalene standard curve ranging from 3.125–100 ng/µL was used to quantify squalene levels, with data adjusted to the total protein content of the cell lysate.

Generation of SM knockout cells

Request a detailed protocol

To knock out SM using CRISPR/Cas9, guide RNAs targeting the SQLE proximal promoter (hg38 chr8:124998048–124998067; AATGGAAACGTTCCGACCCG) and first exon (hg38 chr8:124999588–124999607; ATCCGAGAAGAGGGCGAACT) were designed using CHOPCHOP (Labun et al., 2019) and cloned into the Cas9- and GFP-encoding PX598 vector using BbsI restriction enzyme cloning, as described in Ran et al., 2013, and primers listed in Appendix 1—table 2. HEK293T cells were seeded into 10 cm dishes and transfected with 5 µg of both vectors using Lipofectamine 3000, as described above. After 24 hr, cells were trypsinized, washed with PBS and resuspended in FACS buffer (5% [v/v] FCS and 10 mM EDTA in PBS). Cells were sorted based on GFP fluorescence using a BD FACSAria III at the UNSW Flow Cytometry Facility. GFP-positive cells were seeded into 10 cm dishes at a low density (6,000–12,000 cells/dish) and allowed to adhere for 2–3 weeks. Single colonies were picked, expanded, and screened for SQLE mRNA expression and SM protein expression via qRT-PCR and immunoblotting, as described above. To identify the genomic lesion, genomic DNA was isolated from clones using the PureLink Genomic DNA Mini kit. The CRISPR/Cas9 target region was amplified and cloned into a pcDNA3.1 vector for Sanger sequencing.

Cholesterol synthesis assays

Request a detailed protocol

To assay flux through cholesterol synthesis, radiolabeling and thin-layer chromatography were performed as described previously (Capell-Hattam et al., 2022). Briefly, cells were seeded into 6-well plates and incubated in maintenance medium containing 2 µCi/well [14C]-acetic acid sodium salt as specified in figure legends. Cells were lysed in 1 mL 0.05 M sodium hydroxide and protein content was determined using the bicinchoninic acid assay. Samples were normalized to the lowest protein concentration within each experiment by discarding the required volume and making up to 1 mL with 0.05 M sodium hydroxide. Lysates were mixed with 1 mL ethanol, 500 µL 75% (w/v) potassium hydroxide, 1 µL 20 mM butylated hydroxytoluene, and 20 µL 20 mM EDTA, and saponified by incubation in a 70 °C water bath for 1 hr. Once cooled to room temperature, non-saponifiable lipids were extracted by mixing samples with 1 mL ethanol and 2.5 mL hexane, vortexing for 30 s, and centrifuging at 2,000 g for 5 min. The upper organic phase (2 mL) was collected and evaporated to dryness in a fume cupboard. Dried lipids were resuspended in 50 µL hexane and separated by thin layer chromatography on TLC Silica gel 60 F254 plates (Supelco) using a heptane:ethyl acetate (2:1, v/v) mobile phase. Silica plates were dried, exposed to BAS-IP SR phosphor screens (Fujifilm) for 5–7 days, and imaged using a Typhoon FLA 9500 imager (GE Healthcare). Densitometry analysis was performed using Image Studio Lite software and levels of each analyte were normalized to the condition with the highest abundance.

Data analysis and presentation

Request a detailed protocol

Data were normalized as described in figure legends. To quantify overall changes in the predominant SM variant, the trunSM:SM ratio was calculated. To compare hypoxia- or squalene-induced changes in individual protein levels under different conditions, data were normalized to the respective normoxic condition for each variable. To compare basal (normoxic) protein levels under different conditions, data were normalized to a single control condition. All data were obtained in n≥3 independent experiments, and visualization and statistical testing were performed using GraphPad Prism software as specified in figure legends. Grey lines on graphs indicate statistical comparisons. Where multiple statistical comparisons were made in a single experiment, p-values were corrected using the Benjamini-Hochberg method (Benjamini and Hochberg, 1995) with a false discovery threshold of 5%. Thresholds for statistical significance were defined as: *, p≤0.05; **, p≤0.01. Values of 0.05<p < 0.075 are indicated in text on graphs, and n.s. indicates p≥0.075. Figures were assembled using Adobe Illustrator software (Adobe Inc).

Appendix 1

Appendix 1—table 1
Plasmids used for transfection.
PlasmidDescription
pCMV-(HA)3-SM-V5pcDNA3.1/V5-His TOPO vector containing the coding sequence of human squalene monooxygenase (SM; NM_003129.4) fused with three N-terminal HA epitope tags and C-terminal V5 and 6×His epitope tags, under the transcriptional control of a constitutive cytomegalovirus (CMV) promoter. Generated previously by the Brown laboratory.
pCMV-(HA)3-SM-V5 K82 / 90 / 100RpCMV-(HA)3-SM-V5 containing SM K82R, K90R and K100R substitutions, which blunt SM truncation (Coates et al., 2021). Generated previously (Coates et al., 2021).
pCMV-(HA)3-SM-V5 K290RpCMV-(HA)3-SM-V5 containing a K290R substitution, which disrupts a known ubiquitination site (Hornbeck et al., 2015). Generated in this study.
pCMV-(HA)3-SM-V5 SM-N100 C/S/T>ApCMV-(HA)3-SM-V5 containing SM T3A, T9A, T11A, S43A, C46A, S59A, S61A, S67A, S71A, S83A and S87A substitutions, which blunt cholesterol-induced degradation of SM (Chua et al., 2019). Generated previously (Coates et al., 2021).
pCMV-(HA)3-SM-V5 Y195FpCMV-(HA)3-SM-V5 containing an SM Y195F substitution, which causes loss of catalytic activity (Padyana et al., 2019). Generated previously (Coates et al., 2021).
pCMV-SM(ΔN65)-V5pCMV-(HA)3-SM-V5 containing a deletion of the N-terminal (HA)3 tag and first 65 residues of SM, which are lost upon truncation (Coates et al., 2021). Generated previously (Coates et al., 2021).
pCMV-SM(ΔN65)-V5 Y195FpCMV-SM(ΔN65)-V5 containing an SM Y195F substitution, which causes loss of catalytic activity (Padyana et al., 2019). Generated in this study.
pCMV-(HA)3-SM-V5 F20ApCMV-(HA)3-SM-V5 containing a F20A substitution within the N-terminal re-entrant loop (Howe et al., 2015). Generated in this study.
pCMV-(HA)3-SM-V5 L30ApCMV-(HA)3-SM-V5 containing a L30A substitution within the N-terminal re-entrant loop (Howe et al., 2015). Generated in this study.
pCMV-(HA)3-SM-V5 F35ApCMV-(HA)3-SM-V5 containing a F35A substitution within the N-terminal re-entrant loop (Howe et al., 2015). Generated in this study.
pCMV-(HA)3-SM-V5 L42ApCMV-(HA)3-SM-V5 containing a L42A substitution within the N-terminal re-entrant loop (Howe et al., 2015). Generated in this study.
pCMV-(HA)3-SM-V5 F20A / L30A / F35A / L42ApCMV-(HA)3-SM-V5 containing F20A, L30A, F35A and L42A substitutions within the N-terminal re-entrant loop (Howe et al., 2015). Generated in this study.
PX458 SQLE-1 (promoter)PX458 containing a guide RNA sequence targeting the SQLE proximal promoter (hg38 chr8:124998048–124998067). Generated in this study.
PX458 SQLE-2 (exon 1)PX458 containing a guide RNA sequence targeting SQLE exon 1 (hg38 chr8:124999588–124999607). Generated in this study.
Appendix 1—table 2
Primers used for qRT-PCR and DNA cloning.

Non-annealing nucleotides for restriction enzyme cloning and DNA substitution are indicated in lowercase.

qRT-PCR primer pairPrimer sequence (5′‒3′)Source
VEGFForward
Reverse		CCTGGTGGACATCTTCCAGG
CTGTAGGAAGCTCATCTCTCC		Coates et al., 2019
CA9Forward
Reverse		GTGCCTATGAGCAGTTGCTGTC
AAGTAGCGGCTGAAGTCAGAGG		Origene
RPL11Forward
Reverse		AGAGTGGAGACAGACTGACGCG
CGGATGCCAAAGGATCTGACAG		Origene
GAPDHForward
Reverse		GTCTCCTCTGACTTCAACAGCG
ACCACCCTGTTGCTGTAGCCAA		Origene
ACTBForward
Reverse		CACCATTGGCAATGAGCGGTTC
AGGTCTTTGCGGATGTCCACGT		Origene
HIF1AForward
Reverse		CAGTAACCAACCTCAGTGTGG
CAGATGATCAGAGTCCAAAGC		Dr Laura Sharpe (Brown laboratory)
EPAS1 (HIF2A)Forward
Reverse		CTGTGTCTGAGAAGAGTAACTTCC
TTGCCATAGGCTGAGGACTCCT		Origene
SQLEForward
Reverse		GCTTCCTTCCTCCTTCATCAGTG
GCAACAGTCATTCCTCCACCA		Coates et al., 2021
HMGCRForward
Reverse		TTGGTGATGGGAGCTTGCTGTG
AGTCACAAGCACGTGGAAGACG		Wong et al., 2007
DNA cloning primer pairPrimer sequence (5′‒3′)Method
SQLE gRNA 1 (promoter)Forward
Reverse		caccgAATGGAAACGTTCCGACCCG aaacCGGGTCGGAACGTTTCCATTcBbsI cloning into PX598 vector (as in Ran et al., 2013)
SQLE gRNA 2 (exon 1)Forward
Reverse		aaacAGTTCGCCCTCTTCTCGGATc caccgATCCGAGAAGAGGGCGAACT
SM K290RForward
Reverse		GATGGGCTTTTCTCCAgGTTCAGGAAAAGCCTG
GCGATGCAATTTCCTCATTT		Overlap extension cloning (as in Stevenson et al., 2013)
SM Y195FForward
Reverse		CCAGGTTGTAAATGGTTtCATGATTCATGATCAGG
GCGATGCAATTTCCTCATTT
SM F20AForward
Reverse		GTTCGGGGACgcCATCACTTTGG
GCGATGCAATTTCCTCATTT
SM L30AForward
Reverse		GGAGGTCCTGgcGTGCGTGCTGGTG
GCGATGCAATTTCCTCATTT
SM F35AForward
Reverse		GTGCTGGTGgcCCTCTCGCTG
GCGATGCAATTTCCTCATTT
SM L42AForward
Reverse		CTGGGCCTGGTGgcCTCCTACCGCTG
GCGATGCAATTTCCTCATTT
SM F20A / L30A / F35A / L42AForward
Reverse		Combination of above mutagenic primers
GCGATGCAATTTCCTCATTT

Data availability

All data generated or analyzed during this study are included in the manuscript. Uncropped immunoblots and thin-layer chromatography images are accessible as source data.

References

  10. Book
    1. Brown AJ
    2. Coates HW
    3. Sharpe LJ
    (2021)
    Cholesterol synthesis
    In: Ridgway ND, McLeod RS, Neale R, Roger MC, editors. In Biochemistry of Lipids, Lipoproteins and Membranes. Elsevier Science. pp. 317–355.
    1. Liu GCK
    2. Ahrens EHJ
    3. Schreibman PH
    4. Crouse JR
    (1976)
    Measurement of squalene in human tissues and plasma: validation and application
    Journal of Lipid Research 17:38–45.

Decision letter

  1. Arun Radhakrishnan
    Reviewing Editor; University of Texas Southwestern Medical Center, United States
  2. David Ron
    Senior Editor; University of Cambridge, United Kingdom

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Hypoxia truncates and constitutively activates the key cholesterol synthesis enzyme squalene monooxygenase" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and David Ron as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

Overall, the reviewers were positive about this work and felt that the induction of truncated squalene monooxygenase (truncSM) by hypoxia (relieving it from cholesterol-mediated degradation) was a convincing and exciting result. This result is the core of the manuscript and should be shored up as much as possible. To improve the clarity and presentation, the reviewers suggest some key revisions, which are listed below. Please address these in full and also address the other points raised in the detailed reviews which are appended below.

1) All reviewers were concerned/confused by some of the immunoblot quantification. The interpretations would be simplified if the authors were to show results of quantification for each band in the immunoblots (higher quality immunoblots should be presented if possible). In many cases the graphs of the quantifications seem to be off, it is not clear how the graphs are derived from the blots, which should be representative – see further comments by Reviewers 2 and 3 below on this point.

2) The MARCH6 connection is preliminary/confusing and the inhibitor studies are not clear. A new band appears upon inhibition of VCP/p97 and MG-132 treatment causes a decrease in the protein's expression. This needs to be addressed. The interpretation of Figure 2E is somewhat overstated. Please also see other points about MARCH6 experiments raised below. If it cannot be clarified, the reviewers felt that the MARCH6 angle could be significantly toned down (with some panels removed) without detraction from the main point about hypoxia inducing truncSM.

3) Finally, results shown in Figure 5 showing that truncation of SM correlates with hypoxia in endometrial cancer tissues is a little preliminary. Multiple bands are detected in SM immunoblots, which interferes with interpretation. This experiment could be removed and speculated upon in the discussion.

Reviewer #1 (Recommendations for the authors):

I do have a couple of expository suggestions.

1) You have GOT to include a picture or schematic of the sterol pathway with the relevant enzymes. I have been studying aspects of this pathway for over 30 years (I am old) and I still need one. The people who are new to these ideas will be even more appreciative of this graphic concession.

2) Can you make some kind of summary cartoon of the various regulation processes happening with this key enzyme? I am not sure what it would look like, but it would be a nice summary of the complex and cool story.

Reviewer #2 (Recommendations for the authors):

Is it a clear accumulation of truncSM that results in increased flux through the pathway and sterol metabolic changes? If yes, does this provide a significant advantage to cells under hypoxia? The accumulation of squalene in hypoxia argues against significant truncSM activity. Even if truncSM can be active, oxygen unavailability possibly limits its activity. This possibility is in fact acknowledged by the authors in Line 313 ("consistent with reduced SM activity under low-oxygen conditions").

Another important open question is the role of squalene in promoting truncSM. Under normoxia, MARCHF6 induces the degradation of full-length SM. How is the generation of truncSM favoured by squalene? Does it act directly on MARCHF6? Any additional information to address these issues would significantly strengthen this study. The analysis and some of the data on the relative abundance of SM and truncSM could also be improved.

1. There is huge variability between the control condition (cells grown at 21% O2). For example, in Figure 1D, at .5% O2 the ratio between SM/truncSM is big (probably >5). In contrast, at 3% or 10% O2 the SM/truncSM ratio is probably ~1. It would be important to understand the causes of variability that are observed throughout the manuscript.

2. Another important aspect is the quantification and presentation of the data. The levels of SM and truncSM at 21%O2 are used as references and the two versions of the protein are quantified separately. This is misleading because often cells have different levels of total protein (SM+truncSM) (see for example figure 2). Throughout the manuscript, the levels of truncated protein should be presented in a ratio to full length to provide a better indication of their relative abundance. Another example of strange quantification is in Figures 2C (and 2E). The gel on the left shows very high levels of SM-N100-GFP-V5 in cells treated with MG132. However, in the quantification on the right, those conditions appear to have less SM-N100-GFP-V5 than untreated cells under 21% O2.

3- Figure 2E shows a clear increase in truncSM in MARCHF6 depleted cells. The effect is not insignificant if the truncSM/SM ratio is analysed. In contrast, the effect of MARCHF6 depletion is clear on full-length accumulation. Is this result compatible with the conclusion that "truncation occurs post-ubiquitination by MARCHF6"?

Reviewer #3 (Recommendations for the authors):

The availability of data, code, reagents and other issues is adequate.

https://doi.org/10.7554/eLife.82843.sa1

Author response

Essential revisions:

Overall, the reviewers were positive about this work and felt that the induction of truncated squalene monooxygenase (truncSM) by hypoxia (relieving it from cholesterol-mediated degradation) was a convincing and exciting result. This result is the core of the manuscript and should be shored up as much as possible. To improve the clarity and presentation, the reviewers suggest some key revisions, which are listed below. Please address these in full and also address the other points raised in the detailed reviews which are appended below.

1) All reviewers were concerned/confused by some of the immunoblot quantification. The interpretations would be simplified if the authors were to show results of quantification for each band in the immunoblots (higher quality immunoblots should be presented if possible). In many cases the graphs of the quantifications seem to be off, it is not clear how the graphs are derived from the blots, which should be representative – see further comments by Reviewers 2 and 3 below on this point.

Clarifying our quantification method is clearly essential. Below we address the specific recommendations of Reviewers #2 and #3, and the revised manuscript incorporates these where possible. However, in some cases, we believe alternative quantification methods would distract from our aim to identify factors required for hypoxia-induced SM degradation and trunSM accumulation. We have nevertheless made major updates to the text, figure legends, and axis labels to improve clarity, as outlined below.

Presenting data for all immunoblot lanes: We agree this quantification method, suggested by Reviewer #3, is informative in situations where a condition dramatically affects protein levels at the normoxic baseline. See Author response image 1 for an example based on the original Figure 2C, where normalizing to a single control condition (the normoxic vehicle treatment, set to 1) clearly demonstrates the MG132-induced accumulation of SM and SM‑N100‑GFP‑V5, as seen in the accompanying immunoblot.

Author response image 1
Download asset Open asset

However, a limitation of this normalization method is that the magnitude of hypoxia-induced changes in protein levels cannot be directly compared across multiple treatments or conditions. This was the principle we used to investigate the mechanism of SM truncation, and the rationale behind our original quantification method where protein levels were normalized to the normoxic control for each individual treatment (set to 1). We acknowledge our original visualization of this data, with a single bar representing all normoxic conditions in an experiment (see Author response image 2), was confusing and potentially misleading. The main figures of the revised manuscript still contain quantification of hypoxia-induced protein changes, but the graphs are updated to show all normoxic conditions used for normalization, and thus provide a numerical value for each immunoblot lane. See Author response image 2 for an example of the revised Figure 2C (now Figure 3A) compared with the quantification in the original submission. Where treatments cause normoxic protein accumulation with the potential to confuse interpretation, the supplemental data is updated to include quantification relative to a single control condition, as in the example on the previous page. We have elected to include this additional quantification for the revised Figure 2C (as Figure 2—figure supplement 1C), Figure 2D (as Figure 2—figure supplement 1E) Figure 3A (as Figure 3—figure supplement 1B), and Figure 4—figure supplement 2E, which were flagged by reviewers as points of particular concern.

Author response image 2
Download asset Open asset

We have also updated axis labels and figure legends to improve clarity about the normalization process. For instance, the above graphs replace the original “relative protein levels” label with a more informative “hypoxia-induced reduction” label. The Materials and methods section is also updated to better describe the quantification methods used in the manuscript.

Presenting the trunSM:SM ratio: This quantification method, suggested by Reviewer #2, is useful for visualizing the overall shift from full-length to truncated SM during hypoxia, as was the focus of Figure 1. Expressing the data as a ratio increases variability, but nevertheless supports our original conclusions from these experiments. See Author response image 3 for an example of trunSM:SM quantification from Figure 1D, which Reviewer #2 flagged as a concern due to differences in the normoxic ratio. We also note that cell confluency influences the trunSM:SM ratio—for more information, please refer to our response to Reviewer #2, point #1. The revised manuscript incorporates trunSM:SM quantification for all Figure 1 experiments as Figure 1—figure supplement 2B–D, F and Figure 1—figure supplement 3B.

Author response image 3
Download asset Open asset

When investigating the mechanism of truncation in Figure 2 onwards, we believe presenting the trunSM:SM ratio risks obscuring factors that specifically affect only one of the two steps in the truncation process: (1) proteasomal targeting and (2) partial degradation. This is best illustrated by Figure 4A, which uses different serum types. Here, individual quantification of the two SM variants clearly shows that statin treatment blocks the hypoxia-induced generation of trunSM without an effect on the disappearance (i.e., proteasomal targeting) of full-length SM (Author response image 4, left). This observation was critical, as it led directly to our discovery that squalene promotes the partial, rather than complete, degradation of SM. However, there remains a significant increase in the trunSM:SM ratio in the presence of statin due to the continued disappearance of full-length SM (Author response image 4, red box). Expressing the data in this way makes the differential effects on each SM variant much less apparent, and we believe it may lead to confusion about our conclusions. In the revised manuscript, the Materials and methods section is updated to justify our decisions regarding this quantification method.

Author response image 4
Download asset Open asset

2) The MARCH6 connection is preliminary/confusing and the inhibitor studies are not clear. A new band appears upon inhibition of VCP/p97 and MG-132 treatment causes a decrease in the protein's expression. This needs to be addressed. The interpretation of Figure 2E is somewhat overstated. Please also see other points about MARCH6 experiments raised below. If it cannot be clarified, the reviewers felt that the MARCH6 angle could be significantly toned down (with some panels removed) without detraction from the main point about hypoxia inducing truncSM.

We accept that some of our MARCHF6 experiments are preliminary, and the two MARCHF6‑V5 bands confound their interpretation. The revised manuscript contains only the MARCHF6 knockdown experiment in Figure 2E (showing MARCHF6 is required for hypoxia-induced SM degradation) and the observation of hypoxic MARCHF6 stabilization in Figure 3A. These results are merged into an updated Figure 3. Experiments to determine the mechanism of MARCHF6 stabilization (the original Figure 3B–C, Supplemental Figure S3, Supplemental Figure S5F) are removed and will be the focus of a future study. The working model in Figure 6 and results and discussion sections are revised accordingly.

3) Finally, results shown in Figure 5 showing that truncation of SM correlates with hypoxia in endometrial cancer tissues is a little preliminary. Multiple bands are detected in SM immunoblots, which interferes with interpretation. This experiment could be removed and speculated upon in the discussion.

While the correlation between HIF1α levels and SM truncation in endometrial cancer tissues supports our cell culture work, we acknowledge it could be seen as somewhat preliminary. The original Figure 5 is removed from the revised manuscript and the Discussion section is revised to highlight the role of trunSM in SM-related oncogenesis as a future direction for study.

Reviewer #1 (Recommendations for the authors):

I do have a couple of expository suggestions.

1) You have GOT to include a picture or schematic of the sterol pathway with the relevant enzymes. I have been studying aspects of this pathway for over 30 years (I am old) and I still need one. The people who are new to these ideas will be even more appreciative of this graphic concession.

As suggested, the revised manuscript includes a schematic of the cholesterol synthesis pathway highlighting key enzymes mentioned in the study. This is incorporated as Figure 1—figure supplement 1.

2) Can you make some kind of summary cartoon of the various regulation processes happening with this key enzyme? I am not sure what it would look like, but it would be a nice summary of the complex and cool story.

As suggested, the revised manuscript includes a graphical summary of metabolic regulation of SM. This is incorporated as Figure 6—figure supplement 1 and the Discussion section is updated accordingly.

Reviewer #2 (Recommendations for the authors):

Is it a clear accumulation of truncSM that results in increased flux through the pathway and sterol metabolic changes? If yes, does this provide a significant advantage to cells under hypoxia? The accumulation of squalene in hypoxia argues against significant truncSM activity. Even if truncSM can be active, oxygen unavailability possibly limits its activity. This possibility is in fact acknowledged by the authors in Line 313 ("consistent with reduced SM activity under low-oxygen conditions").

We agree it is important to determine if trunSM activity is preserved during hypoxia. In the revised manuscript, we use [14C]‑acetate labelling and thin-layer chromatography to assay flux through cholesterol synthesis in hypoxic Huh7 cells. This liver-derived cell line shows robust hypoxia-induced truncation of SM (Figure 1—figure supplement 2F). While cholesterol synthesis is reduced during hypoxia, we observed acute accumulation of lanosterol (consistent with other reports, e.g., Nguyen et al. 2007 JBC) yet minimal upstream accumulation of squalene even after a prolonged incubation. This confirms that despite trunSM being oxygen-dependent, its catalytic activity is preserved during hypoxia. We posit that a major function of SM truncation in this context is to ensure efficient clearance of squalene to form lanosterol, which induces degradation of the early cholesterol synthesis enzyme HMGCR and thus suppresses oxygen-intensive pathway flux.

Although we detected squalene accumulation at early hypoxic timepoints in earlier gas chromatography-mass spectrometry experiments (Figure 4B), we attribute this to the much greater sensitivity of mass spectrometry compared with thin-layer chromatography. Indeed, squalene levels are at the lower level of detection in normoxic cells (Figure 4—figure supplement 1) and the strong correlation between trunSM and squalene levels, as determined by mass spectrometry, points towards truncation responding to extremely small changes in squalene levels. The revised manuscript incorporates the new thin-layer chromatography experiment as Figure 5, and the working model in Figure 6 and results, discussion and Materials and methods sections are updated accordingly. Please note the addition of a new author, Isabelle M. Capell-Hattam, who contributed to these experiments.

Another important open question is the role of squalene in promoting truncSM. Under normoxia, MARCHF6 induces the degradation of full-length SM. How is the generation of truncSM favoured by squalene? Does it act directly on MARCHF6? Any additional information to address these issues would significantly strengthen this study.

As suggested, the revised manuscript tests another possible mechanism by which squalene induces SM truncation. Previous work showed squalene directly binds the SM‑N100 regulatory domain (Yoshioka et al. 2020 PNAS), with the hydrophobic re-entrant loop (residues ~15–40) as the most logical binding site. Aromatic residues and leucine residues line the SM active site and are required for catalysis (Padyana et al. 2019 Nat Commun, Abe et al. 2007 BBRC), suggesting they directly interact with squalene. Similar residues in the SM‑N100 re-entrant loop may likewise be involved in squalene binding, and possibly the partial degradation of SM. We therefore mutated phenylalanine (Phe‑20, Phe‑35) and leucine (Leu‑30, Leu‑42) residues in the re-entrant loop to test if they are required for SM truncation. However, these mutations, either alone or in combination, did not prevent the hypoxia-induced accumulation of truncated SM. A better understanding of the precise squalene-binding site in SM‑N100 is needed to further investigate its possible role in SM truncation, but this is beyond the scope of the current study. The revised manuscript incorporates this new SM mutant experiment as Figure 4—figure supplement 4, and the results and discussion sections are updated accordingly.

The analysis and some of the data on the relative abundance of SM and truncSM could also be improved.

Please refer below for our responses to specific concerns, and to our response to Essential Revisions comment #2 for overall changes to data normalization and visualization.

1. There is huge variability between the control condition (cells grown at 21% O2). For example, in Figure 1D, at .5% O2 the ratio between SM/truncSM is big (probably >5). In contrast, at 3% or 10% O2 the SM/truncSM ratio is probably ~1. It would be important to understand the causes of variability that are observed throughout the manuscript.

Due to equipment limitations affecting the number of oxygen concentrations testable at one time, experiments in Figure 1D were staggered across multiple days after cell seeding. This meant that some cells were incubated for longer than others prior to their hypoxic exposure. While steps were taken to obtain a similar cell density at the start of each hypoxic incubation, unavoidable differences in confluency may explain the variability in basal trunSM:SM ratios. We have formally tested the effect of HEK293T seeding density on SM truncation and found a higher confluency indeed promotes a greater trunSM:SM ratio. This is accompanied by a small increase in HIF1α levels that likely reflects increased oxygen consumption of confluent cell monolayers, as reported by other studies (Sheta et al. 2001 Oncogene, Dayan et al. 2009 JCP). Therefore, this observation lends additional support to our model of hypoxia-induced truncation. In other manuscript figures, differences in the basal trunSM:SM ratio likely reflect variation amongst HEK sublines (e.g., HEK293T in Figure 1A, HEK SM‑N100‑GFP‑V5 in Figure 2C, and HEK MARCHF6‑V5 in Figure 3B). The revised manuscript incorporates the new cell confluency experiment as Figure 1—figure supplement 2D, and the Results section is updated accordingly. To improve readability, the original Supplemental Figure S1D–E is moved to a new Figure 1—figure supplement 3.

2. Another important aspect is the quantification and presentation of the data. The levels of SM and truncSM at 21%O2 are used as references and the two versions of the protein are quantified separately. This is misleading because often cells have different levels of total protein (SM+truncSM) (see for example figure 2). Throughout the manuscript, the levels of truncated protein should be presented in a ratio to full length to provide a better indication of their relative abundance. Another example of strange quantification is in Figures 2C (and 2E). The gel on the left shows very high levels of SM-N100-GFP-V5 in cells treated with MG132. However, in the quantification on the right, those conditions appear to have less SM-N100-GFP-V5 than untreated cells under 21% O2.

The revised manuscript includes quantification of the trunSM:SM ratio for experiments in Figure 1, and quantification of protein levels for all immunoblot lanes, including in Figure 2C. It also contains updates to the text, figure legends, and axis labels to improve clarity about data normalization. For more information, please refer to our response to Essential Revisions comment #1.

3- Figure 2E shows a clear increase in truncSM in MARCHF6 depleted cells. The effect is not insignificant if the truncSM/SM ratio is analysed. In contrast, the effect of MARCHF6 depletion is clear on full-length accumulation. Is this result compatible with the conclusion that "truncation occurs post-ubiquitination by MARCHF6"?

The reviewer is correct that hypoxia-induced trunSM accumulation still occurs in MARCHF6-depleted cells. This is also reflected in the trunSM:SM ratio, which remains significantly increased (Author response image 5). Our interpretation is that MARCHF6 is not the sole E3 ubiquitin ligase promoting trunSM formation, and although it is responsible for hypoxia-induced degradation of SM, other proteasomal degradation pathways can cause partial degradation of SM when hypoxia-induced squalene accumulation occurs. To better convey this idea, we have modified our conclusion as follows: “The basal levels and hypoxic accumulation of trunSM were also reduced, consistent with MARCHF6 contributing to the proteasomal targeting, and therefore partial degradation, of SM. Hypoxia-induced accumulation of trunSM was not completely abolished, however, indicating SM can be truncated even when targeted to the proteasome by hypoxia-independent mechanisms.”

Author response image 5
Download asset Open asset

Reviewer #3 (Recommendations for the authors):

The availability of data, code, reagents and other issues is adequate.

The revised manuscript includes availability information for all new experiments.

https://doi.org/10.7554/eLife.82843.sa2

Article and author information

Author details

  1. Hudson W Coates

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Conceptualization, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6506-5249

  2. Isabelle M Capell-Hattam

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared

  3. Ellen M Olzomer

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Methodology
    Competing interests
    No competing interests declared

  4. Ximing Du

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Resources, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6648-017X

  5. Rhonda Farrell

    1. Prince of Wales Private Hospital, Randwick, Australia
    2. Chris O’Brien Lifehouse, Camperdown, Australia
    Contribution
    Resources, Funding acquisition
    Competing interests
    No competing interests declared

  6. Hongyuan Yang

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Resources, Funding acquisition
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8482-6031

  7. Frances L Byrne

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Resources, Funding acquisition, Writing - review and editing
    Competing interests
    No competing interests declared

  8. Andrew J Brown

    School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, Australia
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Methodology, Writing - review and editing
    For correspondence
    aj.brown@unsw.edu.au
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4475-0116

Funding

Australian Government (Research Training Program Scholarship)

  • Hudson W Coates

University of New South Wales (Scientia PhD Scholarship)

  • Isabelle M Capell-Hattam

RANZCOG Research Foundation (Mary Elizabeth Courier Research Scholarship)

  • Rhonda Farrell

Cancer Institute NSW (Career Development Fellowship)

  • Frances L Byrne

Australian Research Council (Grant DP170101178)

  • Andrew J Brown

NSW Health (Investigator Development Grant)

  • Andrew J Brown

Cancer Institute NSW (2021/CDF1120)

  • Frances L Byrne

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Dr Martin Bucknall for technical assistance with gas chromatography-mass spectrometry and the members of the Brown laboratory for critically reviewing this manuscript.

Senior Editor

  1. David Ron, University of Cambridge, United Kingdom

Reviewing Editor

  1. Arun Radhakrishnan, University of Texas Southwestern Medical Center, United States

Version history

  1. Preprint posted: August 19, 2022 (view preprint)
  2. Received: August 19, 2022
  3. Accepted: December 13, 2022
  4. Version of Record published: January 19, 2023 (version 1)

Copyright

© 2023, Coates et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,096
    Page views
  • 145
    Downloads
  • 2
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Hudson W Coates
  2. Isabelle M Capell-Hattam
  3. Ellen M Olzomer
  4. Ximing Du
  5. Rhonda Farrell
  6. Hongyuan Yang
  7. Frances L Byrne
  8. Andrew J Brown
(2023)
Hypoxia truncates and constitutively activates the key cholesterol synthesis enzyme squalene monooxygenase
eLife 12:e82843.
https://doi.org/10.7554/eLife.82843

Categories and tags

Research organism

Further reading

    1. Biochemistry and Chemical Biology
    2. Evolutionary Biology

    Evolution towards simplicity in bacterial small heat shock protein system

    Piotr Karaś, Klaudia Kochanowicz ... Krzysztof Liberek
    Research Article

    Evolution can tinker with multi-protein machines and replace them with simpler single-protein systems performing equivalent functions in an equally efficient manner. It is unclear how, on a molecular level, such simplification can arise. With ancestral reconstruction and biochemical analysis, we have traced the evolution of bacterial small heat shock proteins (sHsp), which help to refold proteins from aggregates using either two proteins with different functions (IbpA and IbpB) or a secondarily single sHsp that performs both functions in an equally efficient way. Secondarily single sHsp evolved from IbpA, an ancestor specialized in strong substrate binding. Evolution of an intermolecular binding site drove the alteration of substrate binding properties, as well as the formation of higher-order oligomers. Upon two mutations in the α-crystallin domain, secondarily single sHsp interacts with aggregated substrates less tightly. Paradoxically, less efficient binding positively influences the ability of sHsp to stimulate substrate refolding, since the dissociation of sHps from aggregates is required to initiate Hsp70-Hsp100-dependent substrate refolding. After the loss of a partner, IbpA took over its role in facilitating the sHsp dissociation from an aggregate by weakening the interaction with the substrate, which became beneficial for the refolding process. We show that the same two amino acids introduced in modern-day systems define whether the IbpA acts as a single sHsp or obligatorily cooperates with an IbpB partner. Our discoveries illuminate how one sequence has evolved to encode functions previously performed by two distinct proteins.

    1. Biochemistry and Chemical Biology
    2. Computational and Systems Biology

    Death by a thousand cuts through kinase inhibitor combinations that maximize selectivity and enable rational multitargeting

    Ian R Outhwaite, Sukrit Singh ... Markus A Seeliger
    Research Article

    Kinase inhibitors are successful therapeutics in the treatment of cancers and autoimmune diseases and are useful tools in biomedical research. However, the high sequence and structural conservation of the catalytic kinase domain complicates the development of selective kinase inhibitors. Inhibition of off-target kinases makes it difficult to study the mechanism of inhibitors in biological systems. Current efforts focus on the development of inhibitors with improved selectivity. Here, we present an alternative solution to this problem by combining inhibitors with divergent off-target effects. We develop a multicompound-multitarget scoring (MMS) method that combines inhibitors to maximize target inhibition and to minimize off-target inhibition. Additionally, this framework enables optimization of inhibitor combinations for multiple on-targets. Using MMS with published kinase inhibitor datasets we determine potent inhibitor combinations for target kinases with better selectivity than the most selective single inhibitor and validate the predicted effect and selectivity of inhibitor combinations using in vitro and in cellulo techniques. MMS greatly enhances selectivity in rational multitargeting applications. The MMS framework is generalizable to other non-kinase biological targets where compound selectivity is a challenge and diverse compound libraries are available.

    1. Biochemistry and Chemical Biology
    2. Microbiology and Infectious Disease

    Indole produced during dysbiosis mediates host–microorganism chemical communication

    Rui-Qiu Yang, Yong-Hong Chen ... Cheng-Gang Zou
    Research Article Updated

    An imbalance of the gut microbiota, termed dysbiosis, has a substantial impact on host physiology. However, the mechanism by which host deals with gut dysbiosis to maintain fitness remains largely unknown. In Caenorhabditis elegans, Escherichia coli, which is its bacterial diet, proliferates in its intestinal lumen during aging. Here, we demonstrate that progressive intestinal proliferation of E. coli activates the transcription factor DAF-16, which is required for maintenance of longevity and organismal fitness in worms with age. DAF-16 up-regulates two lysozymes lys-7 and lys-8, thus limiting the bacterial accumulation in the gut of worms during aging. During dysbiosis, the levels of indole produced by E. coli are increased in worms. Indole is involved in the activation of DAF-16 by TRPA-1 in neurons of worms. Our finding demonstrates that indole functions as a microbial signal of gut dysbiosis to promote fitness of the host.