Lipase maturation factor 1 affects redox homeostasis in the endoplasmic reticulum

Abstract

Lipoprotein lipase (LPL) is a secreted lipase that clears triglycerides from the blood. Proper LPL folding and exit from the endoplasmic reticulum (ER) require lipase maturation factor 1 (LMF1), an ER-resident transmembrane protein, but the mechanism involved is unknown. We used proteomics to identify LMF1-binding partners necessary for LPL secretion in HEK293 cells and found these to include oxidoreductases and lectin chaperones, suggesting that LMF1 facilitates the formation of LPL's five disulfide bonds. In accordance with this role, we found that LPL aggregates in LMF1-deficient cells due to the formation of incorrect intermolecular disulfide bonds. Cells lacking LMF1 were hypersensitive to depletion of glutathione, but not DTT treatment, suggesting that LMF1 helps reduce the ER Accordingly, we found that loss of LMF1 results in a more oxidized ER Our data show that LMF1 has a broader role than simply folding lipases, and we identified fibronectin and the low-density lipoprotein receptor (LDLR) as novel LMF1 clients that contain multiple, non-sequential disulfide bonds. We conclude that LMF1 is needed for secretion of some ER client proteins that require reduction of non-native disulfides during their folding.

Keywords: disulfide bonds; lipase maturation factor 1; lipoprotein lipase; protein folding; redox.

Figure 1. Identification of new LMF1‐interacting partners

1. Experimental schematic of DSP crosslinking experiment.

2. Schematic of pBPA crosslinking.

3. Western blot of LMF1 partners after DSP crosslinking and affinity tag purification using LMF1's C‐terminal His tag. Tetracycline induces LMF1 expression. Loading dye with 50 mM DTT was used to break the disulfide bonds between LMF1 and its interacting partners. LMF1 complexes are labeled with asterisks and an arrow points to LMF1.

4. Western blot of LMF1 partners after pBPA photocrosslinking and affinity tag purification. The DMSO panel does not include pBPA and serves as a negative control. Fractions were eluted with increasing concentrations of imidazole. An arrow points to LMF1.

5. SYPRO orange protein stain of DSP crosslinked samples. Samples without tetracycline but with DSP and DTT (negative control), and with tetracycline, DSP, and DTT were sent for LC‐MS/MS. LMF1 complexes are labeled with asterisks and an arrow points to LMF1.

6. SYPRO orange stained gel of pBPA samples. An arrow points to LMF1.

Source data are available online for this figure.

Figure 2. siRNA knockdown of ERp44, ERp72, ERdj5, UGGT1, UGGT2, and TRX decreases LPL secretion

1. A, B

   HEK293 cells stably expressing LPL‐V5 or PL‐V5 were transfected with scrambled negative control siRNA (NC) or siRNA against LMF1‐interacting partners. Media fractions were probed for the lipase V5 tags, and lysate fractions were probed for V5 and the LMF1‐interacting partners. GAPDH is a loading control.

2. C

   siRNA knockdowns were performed in triplicate, and lipase secretion was quantified relative to NC siRNA. Error bars indicate the standard error.

3. D

   Quantification of the percent of each protein remaining from three independent experiments. Error bars indicate the standard error.

4. E

   HEK293 cells stably expressing LPL‐V5 or PL‐V5 were transfected with NC siRNA or siRNA against TRX. Lipase and TRX levels were tested as for Fig 1A.

5. F

   Txnip overexpression in HEK293 cells stably expressing LPL‐V5 or PL‐V5 hinders LPL, but not PL, secretion.

Source data are available online for this figure.

Figure 3. LPL forms disulfide‐bonded aggregates when LMF1 is missing

1. A

   A Western blot against the C‐terminal V5 tags of both lipases shows that LPL, but not PL, forms intermolecular disulfide bonds in cells. Lipase monomers are marked with a single arrow, and aggregates are marked with double and triple arrows. The LPL aggregates are not present when samples are treated with DTT. There is no PL in the pellet fraction because PL does not aggregate.

2. B

   A Western blot of a non‐reducing gel of the lysate and pellet fraction of LPL grown in cells lacking LMF1 (HEK293∆LMF1), normal HEK293 cells, or cells with additional LMF1 (+LMF1). LPL monomers and aggregates are marked as above.

3. C

   Co‐translational LPL folding was carried out in the presence of SP HEK293 or HEK293∆LMF1 cells. At the indicated time points, AMS was added to reactions to differentiate reduced (R) and oxidized (O) LPL. The position of the stacker layer is indicated.

4. D–F

   The amount of reduced/total LPL intensity (D), the amount of oxidized/total LPL (E), and the intensity of the stacker/total LPL (F) were quantified. Each point is the average of three independent experiments, and error bars represent the standard deviation. Significance was determined by a two‐tailed Student's t‐test.

5. G, H

   Data from (D–F) were combined to show the trends in LPL folding over time for reactions using HEK293 or HEK293∆LMF1 cells. Each point is the average of three independent experiments, and error bars represent the standard deviation.

Source data are available online for this figure.

Figure 4. LMF1 contributes to redox homeostasis in the ER

1. A, B

   Differential sensitivity of HEK293 cells containing and lacking LMF1 (A) and cld/cld and wt/cld MEFs (B) to drugs that perturb cellular redox homeostasis. Cells were treated with BSO, DTT, or tunicamycin as described in Materials and Methods. Treated cell and untreated cells were counted, and the fraction of surviving cells was calculated. Three wells were counted and averaged for each data point, and three independent trials were carried out per condition. Significance was determined by a two‐tailed Student's t‐test.

2. C

   The redox sensor ERroGFP‐S4 shows that HEK293∆LMF1 cells have a more oxidized ER than HEK293 and HEK293 + LMF1 cells both untreated and when stressed by expression of LPL‐mCherry. A two‐tailed Student's t‐test was used to calculate the significance of the difference in the ratio of oxidized/reduced GFP fluorescence intensity between cell types. Each data point is one of three spots measured per cell, and 17 cells were measured per condition.

Source data are available online for this figure.

Figure 5. Conserved cysteines in LMF1 are important to LPL secretion

1. The LMF1 protein sequence from multiple species was aligned using Clustal Omega (Sievers et al, 2011).

2. Schematic of the membrane topology of LMF1 showing the location of all the cysteines.

3. Secretion of LPL‐V5 from HEK293∆LMF1 cells transiently transfected with LMF1 bearing the indicated cysteine to alanine mutations. Lipase levels in the lysate and pellet, and LMF1 levels in the pellet (α‐His), are also probed. GAPDH serves as a loading control.

Source data are available online for this figure.

Figure 6. Analysis of LMF1's redox‐dependent interactions

1. α‐His Western blot to detect the redox state of LMF1. In the first two lanes, samples were untreated and resolved in reducing or non‐reducing loading dye. The next four lanes were subject to the indicated treatments prior to cell lysis then treated with AMS prior to non‐reducing SDS–PAGE.

2. α‐His Western blot to detect acid trapped partners of LMF1. Samples were TCA precipitated, treated with NEM, and LMF1 was purified via its C‐terminal His tag. Samples were resolved by reducing or non‐reducing SDS–PAGE. The lanes marked ‐LMF1 are HEK293∆LMF1 cells included as a negative control.

3. Samples were prepared as in (B), but were probed with an antibody for TRX.

Source data are available online for this figure.

Figure 7. Fibronectin and LDLR depend on LMF1 for secretion

1. Bottom, center, and top slices from a z‐stack of a cld/wt or cld/cld MEF cell. YPet‐tagged fibronectin is in green, and mCherry‐Sec61β, an ER marker, is in red. Note fibronectin assembly into the extracellular matrix in cld/wt cells.

2. cld/wt cells had significantly higher amounts of fibronectin outside of the ER than cld/cld cells by a two‐tailed Student's t‐test.

3. Bottom, center, and top slices show that LDLR is retained in the ER in cells HEK293∆LMF1 cells. LDLR‐GFP is in green, and mCherry‐Sec61β is in red. Note the clumping of LDLR in the ER in HEK293∆LMF1 cells.

4. The ratio of LDLR within the ER/total LDLR was significantly higher for HEK293∆LMF1 cells than HEK293 or HEK293 + LMF1 cells by a two‐tailed Student's t‐test.

Data information: For all panels > 10 cells were measured per genotype. All scale bars are 10 μm.Source data are available online for this figure.