Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Mar 2.
Published in final edited form as: Traffic. 2015 Sep 2;16(10):1108–1126. doi: 10.1111/tra.12312

Functional roles of N-linked glycosylation of human matrix metalloproteinase 9

Tyler Duellman 1,2, John Burnett 2, Jay Yang 1,2,3
PMCID: PMC4774645  NIHMSID: NIHMS719471  PMID: 26207422

Abstract

Matrix metalloproteinase-9 (MMP-9) is a secreted endoproteinase with a critical role in the regulation of the extracellular matrix and proteolytic activation of signaling molecules. Human (h)MMP-9 has two well-defined N-glycosylation sites at residues N38 and N120, however, their role has remained mostly unexplored partly because expression of the N-glycosylation-deficient N38S has been difficult due to a recently discovered SNP-dependent miRNA-mediated inhibitory mechanism. hMMP-9 cDNA encoding amino acid substitutions at residues 38 (mS38) or 120 (N120S) were created in the background of a miRNA binding site disrupted template and expressed by transient transfection. hMMP-9 harboring a single mS38 replacement secreted well, whereas N120S, or a double mS38/ N120S hMMP-9 demonstrated much reduced secretion. Imaging indicated ER-retention of the non-secreted variants and co-IP confirmed an enhanced strong interaction between the non-secreted hMMP-9s and the ER-resident protein calreticulin. Removal of N-glycosylation at residue 38 revealed an amino acid-dependent strong interaction with calreticulin likely preventing unloading of the misfolded protein from the ER chaperone down the normal secretory pathway. As with other glycoproteins, N-glycosylation strongly regulates hMMP-9 secretion. This is mediated, however, through a novel mechanism of cloaking an N-glycosylation-independent strong interaction with the ER-resident calreticulin.

Introduction

The matrix metalloproteinases (MMPs) constitute a family with over 20 members characterized by a conserved Zn2+-dependent protease domain involved in its proteolytic activity. The MMP superfamily is further divided into groups based on the inherent substrate affinity of the enzyme and includes the gelatinases, collagenases, stromelysins, matrilysins, and membrane-type MMPs (1). The gelatinase subfamily of MMP consists of MMP-9 (gelatinase B) and MMP-2 (gelatinase A). Once considered to be only involved in extracellular matrix remodeling, MMP-9 is now recognized to play a pleiotropic role in the extracellular environment. With numerous substrates such as cell surface proteins (2), cryptic signaling molecules (3-5), latent extracellular protease inhibitors (6), and a growing list of intracellular substrates (7), MMP-9 is now considered to be a dynamic modifier of both extracellular and intracellular events. As expected for a proteinase with such a diversity of substrates, an increase or aberrant expression of MMP-9 has been associated with a number of inflammatory, autoimmune, neoplastic, degenerative and cardiovascular diseases (reviewed in (8)). A detailed understanding of all steps in the life cycle of MMP-9 is required to better understand its role in normal biology and pathobiology.

In most cells, with the exception of neutrophils, MMP-9 expression is tightly regulated at the transcriptional level by cytokines, growth factors, hormones, neurotransmitters, and epigenetic mechanisms, to name a few (reviewed in (8)). The transcribed mRNA is now also known to be regulated by non-coding RNAs, including miRNA, targeting the 3'UTR and the coding exon (9, 10). Once translated, the protein is secreted via small Golgi-derived vesicles in a microtubule-dependent manner in macrophages, neurons, and glial-cells (11-13). Once secreted the pro-enzyme is activated by auto- or hetero-proteolysis of the pro-domain and the enzymatic activity is further regulated by interaction with circulating α2-macroglobulin, reversion-inducing-cysteine-rich protein with kazal motifs (RECK) or tissue-resident tissue inhibitors of metalloproteinases (TIMPs) 1-4. Catabolism or cellular internalization and recycling by Low Density Lipoprotein Receptor Related Protein-1 (LPR1) and other scavenger receptors (14) clear MMP-9 from the circulation. Much work has been reported on most steps of the MMP-9 life cycle, but our understanding of the early secretory mechanism is just emerging. MMP-9 is a glycoprotein with multiple O-glycosylation sites in the linker region between the catalytic and hemopexin domains (15). This O-glycosylation-mediated inter-domain flexibility strongly influences the interaction of MMP-9 with its inhibitors TIMP-1, LRP1, and megalin (15). The human (h)MMP-9 also has three potential N-glycosylation sites identified by the consensus sequence (N-X-S/ T, X indicating any amino acid except proline). An extensive analysis of the oligosaccharide location and structure revealed N38 (N39 in mouse) located within the pro-domain and N120 (N120 in mouse) situated in the catalytic domain (16, 17) to be N-glycosylated whereas N127 is not. The distinct functional roles, however, of these N-glycosylation sites remain to be explored.

Despite the well-described roles of N-glycosylation in the secretion or enzymatic kinetics of glycoproteins (18), little is known about the individual functions of the dual N-glycosylation sites within hMMP-9. For hMMP-9, a naturally occurring missense single nucleotide polymorphism (SNP) (rs41427445) results in an N38S amino acid residue change, thus eliminating the N-glycosylation at this site. However, there has been no literature report characterizing the role of N-glycosylation at this residue presumably because it has been difficult to express this N-glycosylation-deficient hMMP-9 protein due to a recently described novel regulatory mechanism involving miRNA binding sites created by this SNP (10). During the course of our previous study, we discovered that a robust expression of an N-glycosylation-deficient hMMP-9 protein could be attained by disrupting the miRNA-binding site through silent mutagenesis of the mRNA nucleotide sequence while disrupting the N-glycosylated N38 residue. In the present work we took advantage of our ability to express mutations that abolish N-glycosylation while escaping nucleotide-specific miRNA down-regulation to examine the individual roles of N38 and N120 glycosylation in the early secretory process of hMMP-9.

Results

The early process of many N-glycosylated glycoprotein secretion follows an orderly series of events whereby the nascent protein is folded in the ER through interactions with the chaperone proteins followed by the recognition of the properly folded protein by a carrier protein, and transported to the Golgi apparatus via Cop II-coated vesicles (19). We first wanted to confirm the validity of this canonical early secretory pathway in the secretion of MMP-9 in the RAW 264.7 mouse macrophage cell line. MMP-9 expression is regulated by many ligands including phorbol esters (20-22) and phorbol 12-myristate 13-acetate (PMA) stimulation resulted in a large increase in secreted MMP-9 in the cell culture media (Figure 1A). Pharmacological inhibition of the glucose-trimming essential for the chaperone protein recognition of an N-glycosylated protein with castanospermine (CST), inhibition of N-glycosylation with tunicamycin (TUN), and the inhibition of Cop II-mediated vesicular transport by brefeldin-A (BFA) all attenuated the PMA-induced increase in endogenous MMP-9 secretion. Likewise, the same pharmacological treatments reduced the secretion of exogenously expressed hMMP-9-mCherry fusion protein in transfected HEK293 cells (Figure 1B). Fluorescent microscopy revealed the expected accumulation of the mCherry fluorescence in the perinuclear endoplasmic reticulum (ER) upon inhibition of the secretion of a hMMP-9-mCherry fusion protein, while in the condition of no drug treatment the hMMP-9-mCherry appeared in post-Golgi apparatus vesicles similar to what was observed in activated macrophage cells (11) (Figure 1C). A complementation assay based on the co-expression of YFP1-X and YFP2-Y fusion proteins, where the two YPFs fragments represent half of the fluorescent protein, and X and Y are the interacting proteins, has been described (23). Two YFP fragments complement, which reconstitutes the fluorescent reporter when proteins X and Y interact. By choosing an ER-specific protein such as YFP1-calreticulin (CALR) as the bait and YFP2-hMMP-9 as the prey, an organelle-specific interaction between CALR and hMMP-9 in the ER can be assessed. The strength of the complementation signal was dependent on whether the protein was N- or C-terminal tagged, with N-tagging of hMMP-9 giving the larger signal (Supplemental Figure 1E). Such tag-orientation dependence of the complementation assay is expected since the two complementing YFP fluorophores must be in close proximity to yield a signal and the nature of protein interaction will dictate the optimal placement of the tag. Cells co-transfected with the YFP1-CALR and YFP2-hMMP-9 demonstrated ER-localized fluorescence (Supplemental Figure 1D). Furthermore, the N-glycosylation-dependent interaction between CALR and hMMP-9 was confirmed by this YFP-complementation assay since CST diminished the fluorescence signal resulting from the interaction between CALR and hMMP-9 in the ER compartment (Figure 1D). These results confirmed the likely validity of the canonical early secretory events in the ER mediating secretion of both endogenous and transfected exogenously -expressed hMMP-9 allowing us to further investigate the roles of N-glycosylation in hMMP-9 secretion.

Figure 1. Secretion of endogenous MMP-9 from stimulated RAW 264.7 cells and exogenous hMMP-9 in transfected HEK293 cells follow the canonical early secretory pathway in the endoplasmic reticulum.

Figure 1

Open in a new tab

(A) Mouse macrophage RAW 264.7 cells were stimulated with PMA (200 nM) and the indicated inhibitors. The cell culture media was harvested at 24 hrs after the drug treatment and probed for the secreted mMMP-9 with a Western blot. Each lane was loaded with 35 μL cell culture media and the membrane was probed with 1:1000 anti-MMP-9 antibody. The bar plot is the mean ±S.E.M. of densitometric quantitation normalized to the PMA-alone condition from n=3 blots. (B) Same assay as (A) except that HEK293 cells were transfected with 250 μg hMMP-9 plasmid per one well of a 12-well plate. The densitometric quantitation is from n=3 blots. (C) Photomicrograph of HEK293 cells expressing hMMP-9-mCherry treated with different drugs at the indicated concentrations. Top: no drug, next: CST (1 mM), next: tunicamycin (250 ng/ml), and bottom: BFA (5 μg/ml). The bar is 15 μm. (D) YPF fluorescence intensity (arbitrary units) measured on a microplate reader from cells transfected with the complementing CALR.nYFP1 and hMMP-9.nYFP2 plasmids. The open bar is control and the solid bar is after treatment with CST (1 mM) for 15 hrs.

hMMP-9 residues N38 in the pro-peptide domain and N120 in the catalytic domain are N-glycosylated (Figure 2A). We created an N38-glycosylation-deficient hMMP-9 cDNA while disrupting the miRNA-binding sites, preventing the expression of the naturally occurring SNP, and successfully expressed hMMP-9 with an N38S residue switch (10). We name this miRNA-binding disrupted N-glycosylation-deficient construct modified-S38 (mS38). We also created a single N-glycosylation-deficient N120 mutant (N120S) and a double N-glycosylation-deficient mutant (mS38/ N120S). These hMMP-9 constructs were non-tagged or C-terminal tagged with the mCherry fluorescent reporter to allow detection of the expressed protein in transfected live cells and secreted protein in the cell culture supernatant (24). The N-glycosylation-deficient hMMP-9 mutants expressed at levels similar to wild-type-hMMP-9 after transfection, and the intracellular localization of the hMMP-9-mCherry fusion proteins in HEK293 cells viewed with a fluorescent microscope were qualitatively indistinguishable (Figure 2B). The total lysates from cells transfected with untagged-hMMP-9 constructs were harvested, treated with Endo H or PNGase F endoglycase, and a Western blot was probed with an anti-MMP-9 antibody (Figure 2C). The Western blot confirmed that all N-glycosylation-deficient mutants expressed similar to wild-type-hMMP-9, consistent with the visual impression of the mCherry-tagged constructs. The N-glycosylation-deficient hMMP-9 demonstrated the expected apparent molecular mass with a slight mobility shift observed, consistent with a loss of one (mS38 or N120S) or two (mS38/ N120S) N-glycosylation sites. PNGase F treatment collapsed the bands to the maximally mobility-shifted species for the wild-type-, mS38-, and N120S- identical to the mS38/ N120S-hMMP-9. Residual PNGase F-resistant higher molecular weight bands (noted by *, Figure 2C) for wild-type, mS38-, and N120S-hMMP-9 indicate that these species are O-glycosylated, a process that occurs in the Golgi apparatus. Endo H treatment mostly collapsed the bands identical to PNGase F, however, with a faint band remaining (noted by arrows, Figure 2C), suggestive of a small Endo H-resistant component to be expected for proteins acquiring complex terminal glycosylation patterning that occurs in the Golgi apparatus. Together this data indicates that wild-type-, mS38-, and to a lesser extent N120S-hMMP-9 are able to proceed from the ER to the Golgi apparatus. A Western blot probe of the extracellular media showed a robustly secreted mS38-hMMP-9 signal equivalent to wild-type-hMMP-9, but much reduced presence of N120S-hMMP-9 and no recovery of mS38/ N120S-hMMP-9 (Figure 2D, right). In contrast, the intracellular protein levels of all constructs were comparable (Figure 2D, left) in agreement with the fluorescent images. We took advantage of the bright fluorescence of the mCherry-tag and performed a time-course analysis of the various MMP-9 secreted into the culture media using a fluorescent microplate reader. Over a time course of 32 hours after transfection, the extracellular media contained an increasing amount of secreted hMMP-9-mCherry with the wild-type and mS38 equally abundant, N120S about a third of the maximum secretors, and mS38/ N120S showing only 1/10 maximum (Figure 2E). The hMMP-9-mCherry present in the culture media was essentially eliminated by BFA, indicating that the protein was secreted via a Cop II-dependent process (Figure 2F). The N-glycosylation-dependence, in particular the dependence of secretion after terminal glucose trimming of the oligosaccharide (25), was confirmed by the decrease in secretion of hMMP-9 observed after CST treatment (Figure 2F), which inhibits glucosidase I and II. The decrease in secreted hMMP-9 from CST treatment was less for the mS38/N120S mutant compared to the wild type, as expected, since this mutant has no N-glycosylated sites. The BFA-induced decrease in secreted hMMP-9 was robust for all constructs indicating the role of Cop II coated vesicles in the secretion regardless of the N-glycosylation status. The increased intracellular accumulation of hMMP-9-mCherry after BFA or CST treatment was notable on the fluorescent images of the cells (data not shown). N-glycosylation has been reported to play major roles in protein folding, intracellular trafficking, plasma membrane dynamics, glycoprotein half-life, and enzymatic characteristics of a number of extracellular proteases (reviewed in (18)). While we focused on the robust effect of N-glycosylation on secretion, we also investigated the possible effects on the activation of the pro-enzyme, enzymatic activity, and the interaction or inhibition by TIMP-1. Results comparing the abundantly secreted wild-type- and mS38-hMMP-9 were negative in that the loss of N-glycosylation at residue 38 had no detectable effects on these phenotypes (Supplemental Figure 2). We were unable to perform a similar analysis on N120S- or mS38/ N120S-hMMP-9 due to limited protein amounts secreted into the culture media for these mutants but performed a fluorescent-substrate MMP-9 enzymatic assay on the cell lysate (i.e. intracellular hMMP-9) and determined that the secretion deficient mutants demonstrated substantially reduced enzymatic activity (Supplemental Figure 3).

Figure 2. Glycosylation-deficient hMMP-9 result in reduced secretion.

Figure 2

Open in a new tab

(A) Schematic representation of the hMMP-9 domain structure and the locations of N-glycosylation (Y). The engineered N-glycosylation-deficient mutants, mS38 (N38S mutation), N120S (N120S mutation) and mS38/ N120S (N38S and N120S mutations) are indicated below along with the fully N-glycosylated wild-type. SP: signal peptide, OG: O-glycosylated domain, C: cysteine switch. (B) A fluorescence microscopic image of the transfected cells expressing wild-type and N-glycosylation-deficient hMMP-9 tagged with the mCherry reporter. Cells were stained with DAPI to visualize the nuclei. Below each image is the expanded highlighted area with higher magnification. The bar is 25 μm. (C) Western blot analysis of enzymatic deglycosylation of the various intracellular hMMP-9 proteins treated with Endo H or PNGase F. Endo H resistant bands indicating complex N-oligosaccharide structure, a process occurring in the Golgi apparatus, are noted with arrows. PNGase F resistant bands indicating O-glycosylation, a process occurring in the Golgi apparatus, are indicated with *. (D) Western blot analysis of intracellular (left) and extracellular (right) hMMP-9 24 hrs after transfection. GAPDH served as the loading control for the whole cell lysates. A fixed volume of cell culture media was loaded for the extracellular blot. (E) MMP-9 secretion assay showing relative secretion levels of mCherry tagged hMMP-9 N-glycosylation-deficient mutants. Cells were co-transfected with the mCherry-tagged hMMP-9 and ssT-Cad.eGFP encoding a secreted eGFP reporter. Culture media were collected at the indicated times and the fluorescence signal from the secreted mCherry-tagged hMMP-9 protein normalized by the eGFP signal in the media. Mean ± S.E.M, n=3 with * denoting P<0.05. (F) Effects of BFA (5 μg/ mL) or CST (1 mM) on MMP-9 secretion. The drugs were added to the cell culture media 4 hrs after transfection and the secreted mCherry-tagged MMP-9 levels normalized by the eGFP signal was quantified 24 hours post-transfection. Mean ± S.E.M, n=3 with **** denoting P<0.001, *** denoting P<0.005, and * denoting P<0.05.

Since the intracellular expression of the N-glycosylation-deficient hMMP-9 was robust with a reduced extracellular secretion, the protein must be accumulating within the cell. We performed immunocytochemistry in an attempt to localize the subcellular distribution of the secretion-deficient hMMP-9. Cells were co-transfected with organelle-targeting eGFP-tagged constructs localizing to the ER, mitochondria, or the Golgi apparatus and with vectors expressing the various N-glycosylation-deficient hMMP-9 cDNA C-terminally tagged with mCherry. Figure 3 shows an eGFP image (green, left column) of cells transfected with the Golgi apparatus-targeting (top row), mitochondria-targeting (middle row), and ER-targeting (bottom row) vectors; hMMP-9-mCherry (red, middle column); and overlap images (right column). Panels A-D show images from cells co-transfected with the wild-type-, mS38-, N120S-, or mS38/ N120S-hMMP-9-mCherry cDNA. The organelle-targeted eGFP expression in the expected organelles was confirmed by immunostaining with organelle-specific markers (Supplemental Figure 4). The secreted wild-type- and mS38-hMMP-9 showed no specific co-localization while the secretion-deficient N120S- and mS38/ N120S-hMMP-9 showed greater overlap with the ER-marker, suggesting that the N-glycosylation-deficient hMMP-9 might be trapped in the ER unable to progress down the normal secretory pathway through the Golgi apparatus via the Cop II-mediated process and unable to progress to the post–Golgi apparatus vesicles (11).

Figure 3. Co-localization of glycosylation-deficient hMMP-9 with an ER marker.

Figure 3

Open in a new tab

Fluorescence microscopic images of mCherry tagged: wild-type-MMP-9 (A), mS38-hMMP-9 (B), N120S-hMMP-9 (C) and mS38/ N120-hMMP-9 (D) (middle columns) and organelle localization using organelle-targeting eGFP (left columns). Top panel (Golgi apparatus; Golgi), middle panel (mitochondria; Mito) and lower panel (endoplasmic reticulum; ER) show organelle localization. Merged analysis (right columns) shows hMMP-9 mutants (red) and organelle markers (green). The bar is 10 μm.

We investigated the idea of selective retention of N-glycosylation-deficient and non-secreted hMMP-9 in the ER by performing a co-IP assay between hMMP-9 and ER-resident proteins. Calreticulin (CALR) and calnexin (CANX) are ER-resident chaperone proteins critical for the quality control and secretion of glycoproteins in the CANX/ CALR cycle (26). Nascent glycoproteins are co-translationally N-glycosylated and subsequently traverse through the CANX/ CALR cycle until the properly folded protein is passed on to a carrier lectin for transport to the Golgi apparatus and subsequent secretion. Terminally misfolded ER-retained proteins despite cycling through the CANX/CALR cycle are subsequently degraded via the ERAD pathway. Alternatively, secretion can proceed via the bulk flow mechanism through pinching of the ER membrane without involvement of a transport lectin and secretion is precluded only when misfolded proteins are retained within the ER (27). The two prevailing hypotheses of early secretory mechanisms differ in the key point that the properly folded protein destined for secretion is recognized in the ER by a lectin carrier (carrier-mediated secretion) vs. the secretion of the properly folded protein by bulk-flow (26, 28, 29). Both mechanisms predict ER recognition and retention of misfolded non-secreted proteins. While both CANX and CALR chaperone proteins are abundantly present in the ER, CALR is thought to play a major role in interacting with soluble proteins (30) such as MMP-9, while CANX is thought to play a major role in interacting with membrane bound proteins (31, 32). Additionally, complementation assay indicated that wild-type-MMP-9 interacted with CALR in a CST-dependent manner more robustly than CANX (Supplemental Figure 1E), therefore, we focused on the MMP-9: CALR interaction. Lysates from cells transfected with a HA-tagged CALR and hMMP-9 were subjected to immunoprecipitation with an anti-HA antibody and a Western blot was probed with the anti-MMP-9 antibody. The co-IP assay demonstrated little interaction between the secreted hMMP-9 (wild-type or mS38) and CALR, but significant protein-protein interaction noted between the secretion-deficient hMMP-9 (N120S or mS38/ N120S) and the ER-resident CALR (Figure 4A, B). A control co-IP between hMMP-9 variants and TIMP-1-HA demonstrated no effect of N-glycosylation on this protein-protein interaction (Supplemental Figure 2E) as expected for an interaction known to be mediated by the C-terminal hemopexin domain of MMP-9 (33, 34).

Figure 4. Glycosylation-deficient hMMPs interact with the ER-protein CALR.

Figure 4

Open in a new tab

(A) Cells were co-transfected with plasmids encoding the noted hMMP-9 and CALR. Total cell lysates were prepared 24 hrs after transfection and 20 μg protein loaded in each lane. The top panel shows a Western blot of the input cell lysate probed for hMMP-9, CALR, and GAPDH. The middle panel shows the co-IP complex product probed for hMMP-9 and CALR. The bottom panel shows the densitometric quantitation of the co-IP product. For the co-IP experiments, 75 μg of total protein lysate was used following the protocol described in Methods. Densitometric analysis presented as mean ± S.E.M, n=4. (B) CALR complementation assay with wild-type (WT) or N-glycosylation-deficient hMMP-9 mutants. CALR.nYFP1 was co-transfected with the N-tagged YFP2 (nYFP2) or the C-tagged YFP2 (cYFP2) wild-type or N-glycosylation mutants and fluorescence assayed 36 hrs after transfection. Transfection of the individual complementation components, which gave no signal, served as a negative control. Bars are mean ± S.E.M., n=3 with * denoting P<0.05.

The co-IP assay provides information on potential MMP-9 and ER-resident protein: protein interaction. However, this assay provides no information on the subcellular localization and the dependence on N-glycosylation of the putative interaction. A complementation assay based on the co-expression of YFP1-CALR and YFP2-hMMP-9, as shown in Figure 1, was repeated for the N-glycosylation-deficient hMMP-9 constructs in an attempt to demonstrate organelle- and N-glycosylation-specific interaction between CALR and hMMP-9-mutants in the ER. Quantitation of the fluorescence intensity showed the CALR: wild-type-hMMP-9 interaction to be robust (Figure 4B). CALR: N-glycosylation-deficient hMMP-9 complementation signals were present, but much weaker suggesting that the nature of interaction between CALR and the N-glycosylation-deficient hMMP-9 is fundamentally different from the interaction with the N-glycosylated wild-type. Nevertheless, a positive signal between N-glycosylation-deficient-hMMP-9 and CALR adds to the evidence that hMMP-9, in an N-glycosylation-independent manner, can interact with the ER-resident CALR.

The CALR: hMMP-9 co-IP and complementation data (Figure 4) is consistent with the immunofluorescence subcellular localization data (Figure 3), indicating ER-retention of the secretion-deficient MMP-9. ER-resident lectins are known to interact in a “lectin-only” binding model or a “dual-binding” model where both the carbohydrate and polypeptide domain are responsible for proper lectin interaction with its partner (reviewed in (28)). Examination of the co-IP data demonstrated robust interaction between the mS38/ N120S-hMMP-9 devoid of N-glycosylation and the ER-resident proteins, indicating a novel glycosylation-independent interaction between the non-N-glycosylated hMMP-9 and CALR. While N120-glycosylation seems necessary for efficient hMMP-9 secretion, N38-glycosylation appears to be dispensable. We wanted to identify if the absence of N-glycosylation at N38 might be revealing a stronger polypeptide-dependent interaction, and if so, whether the specific amino acid substitution at residue 38 could influence interaction with CALR and subsequent secretion. We systematically altered residue 38, replacing the wild-type N with all 20 amino acids and correlated the secretion with the physical-chemical side chain properties. Additionally, we created an N-glycosylation-deficient hMMP-9, retaining the N38 residue by introducing a T40A mutation and disrupting the N-X-T/S N-glycosylation consensus sequence.

The mCherry-tagged residue 38X-mutant-hMMP-9 constructs were transfected and the supernatant assayed for secreted hMMP-9. The secreted amount of hMMP-9 24 hours after transfection varied between the residue 38-mutants (Figure 5A). This amino acid-dependent effect on hMMP-9 secretion may exist for the N120 residue as well (Figure 5B), however, this point was not pursued further in the current study. The proper intracellular expression of all N38X-hMMP-9 point-mutants was confirmed by visual observation of the mCherry expression and a Western blot of the total cell lysate (data not shown). The secreted amount was plotted against the different amino acid side chain physical-chemical properties, including the molecular volume, hydrophobicity, polarity, and charge (Figure 5C-F). Molecular volume of the substituted amino acids side chain demonstrated significant correlation with the amount secreted suggesting the importance of the molecular volume of the amino acid at residue 38. Hydropathy index, charge, or polarity had little effect. Specifically, smaller amino acids at residue 38 (S, G, P) allowed robust secretion similar to wild-type-hMMP-9. However there were exceptions, most notable being E which secreted well despite having a moderate side chain volume.

Figure 5. Effects of hMMP-9 residue 38 amino acid replacement on secretion.

Figure 5

Open in a new tab

(A) Amino acid residue replacements were made in the mS38-MMP-9 backbone (miRNA-binding disrupted) at residue 38. N38X-mCherry tagged mutants were transfected and 24 hours later the cell supernatant was quantified for secreted hMMP-9. Bars denote the mCherry signal of N38X-MMP-9 mutants compared to the wild-type-hMMP-9. Mean ± S.E.M. n=4. (B) S and Q residues replaced N120 and secretion quantified as above. Mean ± S.E.M. n=3. (C-F) Correlation plots of amino acid side chain physical-chemical properties with the mCherry signal in the culture supernatant. The percent secretion (normalized to wild-type) was plotted against (C) Volume (56), (D) Hydrophobicity (57), (E) Polarity, and (F) Charge. The slope, P value, and R2 value of each regression are indicated as inserts within the graph. * denotes P<0.05 compared to a slope of 0.

If retention in the ER through interaction with an ER-resident protein is the mechanism preventing secretion, the secretion-competent N38G-hMMP-9, with a very small molecular volume, should show less interaction with CALR while the secretion-deficient N38V- or N38W-hMMP-9, with large to very large molecular volume, should show a strong interaction with the ER-resident protein. A co-IP assay between CALR and selected N38X-hMMP-9-mutants (Figure 6A) supports the idea that a stronger interaction between the ER-resident CALR and N-glycosylation-deficient N38X-hMMP-9 correlated with less secretion (Figure 6B) in a residue 38 side-chain volume-dependent manner.

Figure 6. Secretion inversely correlates with the strength of interaction of N-glycosylation-deficient hMMP-9 and CALR.

Figure 6

Open in a new tab

(A)Western blot of cell lysates of cells transfected with selected N38X-hMMP-9 constructs and HA-tagged CALR. 20 μg / well of lysate protein was loaded and probed for MMP-9, CALR-HA, and GAPDH (top). 75 μg of lysate was used as the input, immunoprecipitated with anti-HA antibody, and probed with either anti-HA (CALR) or anti-MMP-9 (middle). Densitometric quantitation of the co-IP experiments (bottom), mean ± S.E.M., n=4. The particular amino acids at residue 38 were selected as they represent a wide range of molecular volume. (B) The secretion data from Figure 5 was combined with the above data to present a correlation plot between the amount of secreted hMMP-9 at 24 hrs and CALR co-IP.

Discussion

MMP-9 is a secreted glycoprotein and a deeper understanding of the secretory mechanism is critical in better understanding the role of this multifactorial protease in diseases and normal biology. Thus far, some information on post-Golgi apparatus mechanisms and the important role of stable microtubule formation in MMP-9 secretion has been reported (11) but the details of the early secretory process in the ER remain unknown. The present work sheds light on the role of N-glycosylation and the early secretory process of hMMP-9.

Secretion of both endogenous MMP-9 from RAW 264.7 cells and exogenous hMMP-9 in transfected HEK293 cells appear to be mediated by an N-glycosylation-dependent pathway. Study of N38 and N120 N-glycosylation-deficient hMMP-9 revealed the importance of both sites in secretion. Deletion of either of the N-glycosylation sites resulted in deficient secretion and accumulation of intracellular hMMP-9, mostly in the ER and dependent on the physical-chemical nature of the unmasked amino acid now devoid of N-glycosylation. The co-IP assay demonstrated an increase in the N-glycosylation-deficient and secretion-compromised hMMP-9 (N120S and mS38/ N120S) interaction with CALR, suggesting retention of the compromised proteins in the ER as the mechanism leading to secretion deficiency. The major finding from this study is the demonstration that hMMP-9 can interact with CALR in an N-glycosylation-independent but molecular volume-dependent manner. A systematic mutagenesis study, replacing residue 38 with several amino acids, revealed that this apparent N-glycosylation independent effect on secretion critically depends on the amino acid present at residue 38. In particular, it appears that residues with small molecular volume at residue 38 support secretion. Such an amino acid residue-dependent effect on secretion and interaction with ER-resident lectins is indicative of two proteins interacting via a well-delineated protein interaction domain. This may not be a surprise given the previous report that a single-residue mutation of the lectin-carrier alters N-glycosylation-dependent interaction with client proteins (35). CALR interacts and prevents non-glycosylated protein aggregation (reviewed in (28, 36)) and according to the “dual-binding” model CALR binds to newly folded glycoproteins assisted by interaction through both lectin sites as well as through polypeptide-binding sites. Hydrophobic residues favor while β–folding potential disfavored polypeptide binding to CALR (37). The significance of our finding with respect to the idea of dual-binding is the novel discovery that N-glycosylation of residue 38 in hMMP-9 shields binding of this protein to the polypeptide-binding site of CALR and that the non-N-glycosylation-dependent CALR: MMP-9 interaction appear to depend on the molecular volume rather than the hydrophobicity of residue 38. Limited substitution of N120 with S and Q amino acids indicate that specific physical-chemical properties of the amino acid are also likely to play a role in secretion and its interaction with CALR at this site too. Figure 7 summarizes our findings highlighting the N-glycosylation-dependent and -independent binding of hMMP-9 to CALR including the residue 38 molecular volume-dependent effect of the revealed amino acid and how these bindings could affect secretion. Identification of the precise protein interaction domains mediating the misfolded N-glycosylation deficient hMMP-9 and ER-resident proteins should not only shed light on the early secretory events of hMMP-9 secretion but also provide a general insight on ways to allow cells to secrete misfolded proteins to escape the initiation of unfolded protein response and even cell death.

Figure 7. Model of the CALR-mediated selective ER retention of the N-glycosylation-deficient hMMP-9 N38 mutants.

Figure 7

Open in a new tab

CALR exhibits amino acid selective binding and ER retention of the N-glycosylation-deficient unmasked residue 38 mutants dependent on the molecular volume of the amino acid. Fully N-glycosylated hMMP-9 as well as mutants with small molecular volume (blue circle) at residue 38 are able to fold properly and proceed to be secreted normally. Amino acids with large volume (red star) are selectively retained in the ER due to an increased interaction with CALR and fail to be efficiently secreted from the cell.

What mechanism is responsible for the secretion of fully N-glycosylated wild-type-hMMP-9? The two prevailing theories for early secretory events in the ER are “bulk flow” and “carrier-mediated” mechanisms. Both mechanisms predict ER retention of misfolded non-secreted proteins, but differ in the key point that the properly folded protein destined for secretion is recognized in the ER (carrier-mediated secretion) vs. the bulk-flow secretion without a specific recognition event. Both mechanisms predict repetitious entry of misfolded protein into the CALR/ CANX cycle until ultimate degradation by the ERAD pathway (26, 28, 29). The co-IP assay (Figure 4) failed to definitively demonstrate interaction between wild-type-hMMP-9 and CALR. However, the more sensitive and specific complementation assay did show a robust signal supporting interaction of wild-type-hMMP-9 and CALR (Figures 1 and 4), revealing the possibility that hMMP-9 may exit the ER via the bulk flow mechanism or be secreted by the carrier-mediated mechanism. Furthermore, lectin, mannose-binding 1 (LMAN1), perhaps the best characterized lectin carrier protein, appears to be a potential carrier protein for hMMP-9 as revealed by decreased MMP-9 secretion in a cell line where LMAN1 was knocked out by genomic editing (38). Chloroquine inhibition of hMMP-9 secretion, as demonstrated for the secretion of other LMAN1-mediated secreted proteins (39) and the loss of the N-glycan-specific interaction between the lectin-deficient LMAN1-N156A mutant and MMP-9 also adds to the evidence for a carrier-mediated secretion mechanism of hMMP-9 (38).

Our pharmacological experiments intended to block N-glycosylation-dependent secretion (Figure 1) confirmed the presence of residual drug-resistant secretion. Such residual secretion could be due to incomplete pharmacological inhibition of the intended target (i.e. poor pharmacological efficacy) or due to the presence of a true N-glycosylation-independent secretion. We selected the drug concentrations based on prior publications ((40) (41) for tunicamycin; (11) (42) for BFA; (43-46) for CST). However, 1 mM CST may be insufficient to completely inhibit the interaction between the secreted protein and CALR For example, 1 mM CST was unable to completely inhibit the secretion of influenza hemagglutinin (HA) (43), CD3 γ subunit (44), acid phosphatase (46), coagulation factors V and VII (45). The authors show a reduced but still present HA interaction with the ER chaperon proteins suggesting that CST at 1mM concentration is unable to completely inhibit α-glucosidase trimming (43). Our own data from the complementation assay demonstrates a much reduced but persistent interaction between MMP-9 and CALR after CST treatment (Figure 1D). Tunicamycin, on the other hand, appears to completely inhibit N-glycosylation of MMP-9 as assessed by the mobility shifted MMP-9 detected on a Western blot (Figure 1A, B). What mechanism could be responsible for the apparent N-glycosylation-independent secretion? Bulk flow secretion via pinched-off membrane vesicles is a possibility but according to this theory, misfolded proteins (i.e. N-glycosylation deficient) are retained in the ER destined for degradation and should not be secreted. A quantitative examination of the bulk flow indicates that this process is fast (29) and the possible secretion of misfolded protein by bulk flow prior to degradation by the ERAD pathway cannot be dismissed. BFA appears to completely inhibit the secretion of hMMP-9 but both carrier-mediated and bulk flow secretions are Cop II-dependent and sensitive to BFA inhibition providing no insight on the mechanism responsible for the residual drug-resistant secretion.

The precise mechanism of MMP-9 secretion and the specific carrier protein mediating the transit through the early secretory process could be distinct for different cell types since MMP-9 secretion is constitutive in cells such as macrophages after de novo transcription (11) in contrast to neutrophil cells which store MMP-9 in granules near the cell periphery and upon appropriate stimulus MMP-9 is secreted in a regulated fashion (47). Furthermore, MMP-2 which is the most closely related member of the MMP family to MMP-9, though non-N-glycosylated, secretes well indicating that different members of the MMP family may secrete via different mechanisms.

We chose to eliminate N-glycosylation at the selected residues through manipulation of the cDNA sequence and provide no information on the actual glycans decorating the N-glycosylation tree. However the identity of the glycan composition at the two hMMP-9 N-glycosylation sites have been extensively studied and reported by others (16, 48). They used the classical approach of serial digestion with exo- and endoglycase followed by a mass spectrometry identification of the glycans. The major conclusions are that the core GlcNAC-Mannose trees are modified to complex glycan trees with mostly 2 antennae structures, but some 3 antennae also present, with GlcNAc, Gal, and terminal sialic acid glycans. Importantly, both the N38 and N120 glycans are similar if not identical. Molecular dynamics structure simulation identified the two glycan trees in diametrically opposite ends of the MMP-9 N-terminal globular domain suggesting that the functional difference between the glycan trees at N38 and N120 may be the result of their specific location within the MMP-9 protein rather than due to the specific glycan composition.

Activity and functionality of many enzymes are affected by N-glycosylation (18). Our data documented that N38-glycosylation-deficient hMMP-9 retained normal gelatinolytic activity, a normal time course of zymogen activation, and normal binding and inhibition by TIMP-1 (Supplemental Figure 2). N120-glycosylation-deficient hMMP-9, although not secreted efficiently, had reduced but very substantial enzymatic activity. N-glycosylation deficient hMMP-9 could have other significant functional consequences on unexplored properties such as substrate specificity, sensitivity to other inhibitors, or thermostability of the protease. The fact that secretion-deficient N-glycosylation-deficient MMP-9 had residual enzymatic activity opens up the possibility that intracellularly-retained non-secreted MMP-9 is catalytically active. Much of the focus of biological activity of MMP-9 has been on extracellular targets, but an increasing number of intracellular MMP-9 substrates have been demonstrated (7). Intracellularly-retained but catalytically-active MMP-9 could play a significant role in the altered cell biology, and possibly pathobiology, resulting from secretion-deficient MMP-9 acting on intracellular targets. In a more general sense, glycosylation of proteins in the ER is an emerging target for therapeutic intervention and novel drug development (49, 50). Further clarification of the role of N-glycosylation in secretion, glycosylation-independent interaction with ER lectins, and identification of a carrier-protein mediating the normal secretion of MMP-9 will contribute towards opening up the early secretory events as a novel target for regulating the activity of MMP-9.

Materials and Methods

Generation of molecular constructs

A cartoon of all newly created molecular constructs in the eukaryotic expression vector pCI/ neo (Promega, Madison, WI) and the relevant restriction enzyme sites used for subcloning are shown in Supplemental Figure 5. Supplemental Table 1 lists all oligonucleotide PCR primers used and Supplemental Table 2 lists the sequences of gene fragments synthesized by a vendor (GenScript, Piscataway, NJ).

Organelle targeting constructs

Mitochondrial, ER, and Golgi apparatus-targeting generic vectors were created and the eGFP cDNA without the start ATG and/ or stop TAA codons were subcloned in-frame using engineered restriction enzyme sites. The mitochondrial targeting sequence consisted of an N-terminal 29 amino acid sequence obtained from the subunit VIII of human cytochrome C oxidase (51). The ER targeting sequences are a 17 amino acid sequence of the N-terminal leader sequence from calreticulin (52) and a C-terminal KDEL ER retention signal. The Golgi apparatus targeting sequence is an N-terminal 81 amino acid sequence of the human beta 1,4-galactosyltransferse (53). Oligonucleotides for the 5’-to-3’ and 3’-to-5’ DNA strands encoding the targeting amino acid sequences were purchased, annealed to each other, and subcloned into the Nhe I/ Mlu I or Mlu I restriction enzyme sites in the multiple cloning site of pCI/neo.

HA/mCherry tagged constructs

Primers designed to remove the stop codon and incorporate an in-frame Mlu I site were used to PCR amplify corresponding cDNA templates to produce gene products which were ligated into PCR Blunt (Invitrogen, Carlsbad, CA) and sequenced. Once correct sequences were confirmed, genes were placed upstream and in-frame of an HA or an mCherry epitope-tag domain within pCI/ neo.

MMP-9 mutant constructs

Starting with the wild-type-hMMP-9 complementary deoxyribonucleic acid (cDNA; Accession BC006093, Image Clone MGC: 12688), we engineered the desired nucleotide switch by overlap PCR with primers listed in Supplemental Table 1 or through gene synthesis (Supplemental Table 2) of the desired gene fragment and subcloned them into the wild-type-hMMP-9 background using the vector Xho I restriction enzyme site and the internal Bsm I site within the hMMP-9 coding sequence. Constructs were sequence confirmed and subcloned into the pCI/ neo expression vector.

Secreted eGFP

The N-terminal signal sequence of a GPI-anchored T-cadherin was fused to the N-terminal of eGFP after removal of the endogenous start methionine (54). The resulting T-cadherin signal sequence-eGFP (ssT-Cad.eGFP) chimera cDNA was subcloned into the pCI/ neo expression vector. When transfected into cells, the fusion eGFP protein was secreted into the cell culture media and could be quantified by a fluorescent microplate reader. ssT-Cad.eGFP was used as a transfection control.

YFP complementation constructs and assay

The fragmentation strategy for YFP was as described by Nyfeler et al. (23). Briefly, synthesized nucleotides encoding the eYFP fragment 1 (YFP1; YFP amino acids 1-158) and eYFP fragment 2 (YFP2; YFP amino acids 159-239) were introduced 5’ or 3’ of a 10 amino acid (GGGGS)2 linker. Additionally, the N-terminal fusion constructs incorporated the CALR signal sequence for proper secretion into the ER lumen. eYFP fragments were subcloned in pCI/ neo to create the parent fusion vectors. cDNA encoding CANX (Clone ID: 4342732, #202757719) was purchased from GE Dharmacon (Lafayette, CO). cDNA encoding CALR (#HG13269-G) was purchased from Sino Biological Inc. (Beijing, China). CALR, CANX, and MMP-9 lacking the signal sequence were PCR amplified with flanking Mlu I sites and cloned into pCR Blunt (Invitrogen). Constructs were sequence verified and subcloned into the YFP fragment pCI/ neo fusion vectors using Mlu I. Efficient complementation between the bait and the prey eYFP fusion constructs depends on the geometry of the two interacting proteins. Therefore, all permutations of the placement of the YFP2 fragments on the N- or the C-termini were examined for MMP-9. HEK293 cells grown overnight (~70-90% confluence) in a 12-well plate were transfected with “bait” tagged with nYFP1 and MMP-9 either c-terminally or n-terminally tagged with YFP2 using Lipofectamine 2000 (Invitrogen). 24 hours post-transfection cell media was aspirated, cells were suspended in 200 μL PBS, and transferred to a black 96-well plate where YFP fluorescence was recorded using an excitation wavelength/bandwidth of 485/20 and emission wavelength/bandwidth of 528/20 with a Synergy2 microplate reader (BioTek Inc).

Cell Culture

Human embryonic kidney (HEK) 293 (#CRL-1572) and murine macrophage RAW 264.7 (#TIB-71) cells (ATCC, Manassas, VA) were grown in Dulbecco's Modification of Eagle's Medium (Mediatech, Inc, Manassas, VA) with 4.5 g/ L glucose, L-glutamine, sodium pyruvate, and supplemented with 10% heat inactivated fetal bovine serum (Mediatech Inc), 100 U/ mL penicillin (Mediatech Inc), and 0.1 mg/ mL streptomycin (Mediatech Inc).

Reagents

4-aminophenylmercuric acetate (APMA) (#A9563), Gelatin Type-A (#G8150), Phorbol 12-myristate 13-acetate (PMA) (#79346) and Chloroquine diphosphate (#C6628) were purchased from Sigma Aldrich (St. Louis, MO). PNGase F (#P0704) and Endo H (#P0702) were purchased from New England Biolabs (Ipswich, MA). Castanospermine (CST) (#BML-S107) and recombinant MMP3 (#ALX-201-042) were purchased from Enzo Life Sciences (Farmingdale, NY). Brefeldin A (BFA) (#9972) was purchased from Cell Signaling Technology (Danvers, MA). Recombinant Human TIMP-1 (#410-01) was purchased from Peprotech (Rocky Hill, NJ).

Enzymatic Protein Deglycosylation

20 μg of intracellular MMP-9 protein lysate was added to 10× Glycoprotein Denaturing Buffer in a 10 μL volume and incubated at 100°C for 10 minutes. Oligosaccharides were removed with the addition of 10× G7 Reaction Buffer, 10% NP-40 and PNGase F or with addition of 10× G5 Reaction Buffer and Endo H and incubation at 37°C for 5 hours.

Western Blot

At various times post-transfection cell culture media and cell pellets were collected for analysis. 20 μg of cell lysate and an equal volume of cell media were loaded in 10% SDS-PAGE gels. This method gave a linear range of correlation between the total cell lysate and the culture cell media (10). Samples were run and transferred onto 0.45 μm nitrocellulose membranes that were blocked using 5% milk and washed in TBST (Tris buffered saline with 0.1% Triton X-100). Membranes were probed with appropriate antibodies and imaged using the Quality One software (BioRad, Hercules, CA) with a ChemiDoc XRS+ Molecular Imager (BioRad). Densitometric quantitation of Western blots was performed using Image Lab 4.1 software (BioRad).

Co-Immunoprecipitation

50 μL Dynabeads Protein G (Invitrogen, Carlsbad, California) were magnetically precipitated and washed with PBST (PBS with 0.2% Tween 20). A 1:150 dilution of mouse anti-HA antibody (Covance, Princeton, New Jersey) was incubated on a rotator at room temperature for 10 minutes. The PBST/ antibody solution was aspirated on a magnet. The beads were re-suspended in a solution of 75 μg of intracellular protein and rotationally incubated for 1 hour. The beads were magnetically separated and the flow-through (FT) was collected. The beads were washed three times with 200 μL PBS, then re-suspended in PBS and transferred to a new tube. 20 μL of 5× loading buffer was used to re-suspend the beads. This solution was boiled at 70°C for 10 minutes. The suspended immunoprecipitated (IP) and FT fractions were boiled and analyzed by Western blot. Mouse anti-MMP-9 (UC-Davis, Davis, California) and rabbit anti-HA (Santa Cruz Biotechnology, Dallas, TX), were used in a 1:1000 ratio for immunoblotting.

Fluorescence microscopy of subcellular localization

hMMP-9 constructs C-terminally tagged with mCherry were co-transfected with eGFP organelle-localizing constructs. 24 hours post-transfection, cells were fixed with 4% PFA and stained with DAPI nuclear stain. Cells were imaged with a Sensicam camera (PCO-Tech Inc. Romulus, MI) mounted on an Olympus IX50 microscope (Olympus Corporations, Tokyo, Japan) and processed using IPLab3.6 software (Spectra Services, Ontario, NY). Pseudo-color of eGFP (green), DAPI (blue), or mCherry (red) was accomplished using Adobe Photoshop CS5.1 (Mountain View, CA).

Secretion Assay

Endpoint and time-course secretion assays were performed as described previously (24) using carboxy-terminally mCherry tagged hMMP-9 N-glycosylation mutants and a secretable form of eGFP (ssT-Cad.eGFP).

Gelatin Zymography

Culture cell media was harvested 24 hours post-transfection, hMMP-9 was activated by treatment with APMA for 1 hr at 37°C, and loaded in 0.75mm SDS-PAGE gels prepared with the addition of 1 mg/ mL gelatin. Denatured, non-reduced samples were run at constant voltage. Gels were allowed to renature with 4 washes in enzyme renaturing buffer containing 2.5% v/ v Triton X-100 for 1 hour. Gels were then incubated at 37° C for extended periods of time, followed by staining with Coomassie Brilliant Blue 250-R. Images were taken using Quality One software (BioRad) with a ChemiDoc XRS+ Molecular Imager (BioRad). Enzymatic activity was quantified by densitometric analysis of the inverted image using Image Lab 4.1 software (BioRad).

Fluorometric MMP-9 enzymatic activity assay

The SensoLyte 520 MMP-9 Assay Kit (Ana Spec, Inc., #71155, Fremont, CA) was used to determine the enzymatic activity of hMMP-9 N-glycosylation-deficient mutants. 24 hours post-transfection cellular lysate was harvested in lysis buffer and diluted 1:10 in MMP-9 Dilution Buffer (50 mM Tris-HCl, 200 mM NaCl, 5 mM CaCl2, 1 μM ZnCl2, 0.05% Brij-35, 0.05% NaAzide, pH 7.0). 100 μL of hMMP-9 samples were supplemented with 1 μM ZnCl2 and 2 mM APMA, mixed, added to a black 96-well plate, and incubated at 37°C for 1 hr. After incubated a final concentration of 20 μM of the fluorogenic substrate (Enzo Life Sciences) in 100 μL of 2X Assay Buffer (100 mM HEPES, 20 mM CaCl2, 0.1% Brij-35, pH 7.0) was added to the activated hMMP-9 and reaction kinetics were quantified using a Synergy 2 microplate reader (BioTek Inc) with an excitation wavelength/ bandwidth of 360/40 and emission wavelength/bandwidth of 460/40 at 37°C for 1 ½ hrs.

APMA activation and TIMP-1 inhibition assays

MMP-9 exists as a pro-enzyme where the catalytic domain is protected by the N-terminal domain. Therefore enzymatic activation of MMP-9 requires a conformational change induced by AMPA resulting in auto activating cleavage or cleavage of the N-terminal domain by another protease such as MMP-3. A detailed description can be found in (55). Briefly, APMA or recombinant MMP-3 was used to activate cell media containing hMMP-9 by incubation at 37°C for various amounts of time. hMMP-9 activation was assessed by the mobility shift detected by Western blot, indicating pro-domain cleavage. TIMP-1 inhibition was quantified using varying amounts of recombinant TIMP-1 (0.1-300 ng/mL) added to the cell media containing hMMP-9 and activated in the presence of APMA for 1 hr at 37°C. Enzymatic kinetic measurements were then taken every minute for 1 hour at 37°C. Quantification of the cleaved fluorogenic MMP-9 substrate (Enzo Life Sciences) was measured using a microplate reader with an excitation wavelength/ bandwidth of 360/40 and emission wavelength/bandwidth of 460/40 using a Synergy2 microplate reader (BioTek Inc).

Statistical Analysis

A minimum of 3 biological replicates was obtained with each biochemical assay, consisting of 1-3 experimental replicate measurements per assay. Statistically significant difference in the mean values were defined as P<0.05 by t-test.

Supplementary Material

01

Acknowledgements

This study was supported by the Clinical Translational Cardiovascular Training Grant, T32 HL07936-12 (TD), Bamforth Endowment Fund (JY) from the Department of Anesthesiology, UW Madison, and NIH RO1 GM107054 and GM105665 (JY). We thank Mary Roth for editorial assistance.

Abbreviations used in this paper

APMA

4-aminophenylmercuric acetate

BFA

brefeldin A

CALR

calreticulin

CANX

calnexin

CST

castanospermine

HEK293

Human embryonic kidney 293 cells

LMAN1

lectin, mannose-binding 1

MMP-9

matrix metalloproteinase 9

mS38

modified serine 38

PMA

phorbol 12-myristate 13-acetate

SNP

single nucleotide polymorphism

TIMP

tissue inhibitor of metalloproteinase

Footnotes

Conflict of Interest

None to declare.

References

  • 1.Nagase H, Visse R, Murphy G. Structure and function of matrix metalloproteinases and TIMPs. Cardiovasc Res. 2006;69(3):562–573. doi: 10.1016/j.cardiores.2005.12.002. [DOI] [PubMed] [Google Scholar]
  • 2.Cauwe B, Van den Steen PE, Opdenakker G. The biochemical, biological, and pathological kaleidoscope of cell surface substrates processed by matrix metalloproteinases. Crit Rev Biochem Mol Biol. 2007;42(3):113–185. doi: 10.1080/10409230701340019. [DOI] [PubMed] [Google Scholar]
  • 3.Van den Steen PE, Proost P, Wuyts A, Van Damme J, Opdenakker G. Neutrophil gelatinase B potentiates interleukin-8 tenfold by aminoterminal processing, whereas it degrades CTAP-III, PF-4, and GRO-alpha and leaves RANTES and MCP-2 intact. Blood. 2000;96(8):2673–2681. [PubMed] [Google Scholar]
  • 4.Bergers G, Brekken R, McMahon G, Vu TH, Itoh T, Tamaki K, Tanzawa K, Thorpe P, Itohara S, Werb Z, Hanahan D. Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nat Cell Biol. 2000;2(10):737–744. doi: 10.1038/35036374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Ferreras M, Felbor U, Lenhard T, Olsen BR, Delaisse J. Generation and degradation of human endostatin proteins by various proteinases. FEBS Lett. 2000;486(3):247–251. doi: 10.1016/s0014-5793(00)02249-3. [DOI] [PubMed] [Google Scholar]
  • 6.Xu D, McKee CM, Cao Y, Ding Y, Kessler BM, Muschel RJ. Matrix metalloproteinase-9 regulates tumor cell invasion through cleavage of protease nexin-1. Cancer Res. 2010;70(17):6988–6998. doi: 10.1158/0008-5472.CAN-10-0242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Cauwe B, Opdenakker G. Intracellular substrate cleavage: a novel dimension in the biochemistry, biology and pathology of matrix metalloproteinases. Crit Rev Biochem Mol Biol. 2010;45(5):351–423. doi: 10.3109/10409238.2010.501783. [DOI] [PubMed] [Google Scholar]
  • 8.Vandooren J, Van den Steen PE, Opdenakker G. Biochemistry and molecular biology of gelatinase B or matrix metalloproteinase-9 (MMP-9): the next decade. Crit Rev Biochem Mol Biol. 2013;48(3):222–272. doi: 10.3109/10409238.2013.770819. [DOI] [PubMed] [Google Scholar]
  • 9.Yan L, Borregaard N, Kjeldsen L, Moses MA. The high molecular weight urinary matrix metalloproteinase (MMP) activity is a complex of gelatinase B/MMP-9 and neutrophil gelatinase-associated lipocalin (NGAL). Modulation of MMP-9 activity by NGAL. J Biol Chem. 2001;276(40):37258–37265. doi: 10.1074/jbc.M106089200. [DOI] [PubMed] [Google Scholar]
  • 10.Duellman T, Warren C, Yang J. Single nucleotide polymorphismspecific regulation of matrix metalloproteinase-9 by multiple miRNAs targeting the coding exon. Nucleic Acids Res. 2014;42(9):5518–5531. doi: 10.1093/nar/gku197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Hanania R, Sun HS, Xu K, Pustylnik S, Jeganathan S, Harrison RE. Classically activated macrophages use stable microtubules for matrix metalloproteinase-9 (MMP-9) secretion. J Biol Chem. 2012;287(11):8468–8483. doi: 10.1074/jbc.M111.290676. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Gueye Y, Ferhat L, Sbai O, Bianco J, Ould-Yahoui A, Bernard A, Charrat E, Chauvin JP, Risso JJ, Feron F, Rivera S, Khrestchatisky M. Trafficking and secretion of matrix metalloproteinase-2 in olfactory ensheathing glial cells: A role in cell migration? Glia. 2011;59(5):750–770. doi: 10.1002/glia.21146. [DOI] [PubMed] [Google Scholar]
  • 13.Sbai O, Ferhat L, Bernard A, Gueye Y, Ould-Yahoui A, Thiolloy S, Charrat E, Charton G, Tremblay E, Risso JJ, Chauvin JP, Arsanto JP, Rivera S, Khrestchatisky M. Vesicular trafficking and secretion of matrix metalloproteinases-2, -9 and tissue inhibitor of metalloproteinases-1 in neuronal cells. Mol Cell Neurosci. 2008;39(4):549–568. doi: 10.1016/j.mcn.2008.08.004. [DOI] [PubMed] [Google Scholar]
  • 14.Hahn-Dantona E, Ruiz JF, Bornstein P, Strickland DK. The low density lipoprotein receptor-related protein modulates levels of matrix metalloproteinase 9 (MMP-9) by mediating its cellular catabolism. J Biol Chem. 2001;276(18):15498–15503. doi: 10.1074/jbc.M100121200. [DOI] [PubMed] [Google Scholar]
  • 15.Van den Steen PE, Van Aelst I, Hvidberg V, Piccard H, Fiten P, Jacobsen C, Moestrup SK, Fry S, Royle L, Wormald MR, Wallis R, Rudd PM, Dwek RA, Opdenakker G. The hemopexin and O-glycosylated domains tune gelatinase B/MMP-9 bioavailability via inhibition and binding to cargo receptors. J Biol Chem. 2006;281(27):18626–18637. doi: 10.1074/jbc.M512308200. [DOI] [PubMed] [Google Scholar]
  • 16.Kotra LP, Zhang L, Fridman R, Orlando R, Mobashery S. N-Glycosylation pattern of the zymogenic form of human matrix metalloproteinase-9. Bioorg Chem. 2002;30(5):356–370. doi: 10.1016/s0045-2068(02)00501-1. [DOI] [PubMed] [Google Scholar]
  • 17.Van den Steen P, Rudd PM, Proost P, Martens E, Paemen L, Kuster B, van Damme J, Dwek RA, Opdenakker G. Oligosaccharides of recombinant mouse gelatinase B variants. Biochim Biophys Acta. 1998;1425(3):587–598. doi: 10.1016/s0304-4165(98)00113-5. [DOI] [PubMed] [Google Scholar]
  • 18.Skropeta D. The effect of individual N-glycans on enzyme activity. Bioorg Med Chem. 2009;17(7):2645–2653. doi: 10.1016/j.bmc.2009.02.037. [DOI] [PubMed] [Google Scholar]
  • 19.Helenius A, Aebi M. Intracellular functions of N-linked glycans. Science. 2001;291(5512):2364–2369. doi: 10.1126/science.291.5512.2364. [DOI] [PubMed] [Google Scholar]
  • 20.Masure S, Billiau A, Van Damme J, Opdenakker G. Human hepatoma cells produce an 85 kDa gelatinase regulated by phorbol 12-myristate 13-acetate. Biochim Biophys Acta. 1990;1054(3):317–325. doi: 10.1016/0167-4889(90)90103-k. [DOI] [PubMed] [Google Scholar]
  • 21.St-Pierre Y, Couillard J, Van Themsche C. Regulation of MMP-9 gene expression for the development of novel molecular targets against cancer and inflammatory diseases. Expert Opin Ther Targets. 2004;8(5):473–489. doi: 10.1517/14728222.8.5.473. [DOI] [PubMed] [Google Scholar]
  • 22.Shin Y, Yoon SH, Choe EY, Cho SH, Woo CH, Rho JY, Kim JH. PMA-induced up-regulation of MMP-9 is regulated by a PKCalpha-NF-kappaB cascade in human lung epithelial cells. Exp Mol Med. 2007;39(1):97–105. doi: 10.1038/emm.2007.11. [DOI] [PubMed] [Google Scholar]
  • 23.Nyfeler B, Michnick SW, Hauri HP. Capturing protein interactions in the secretory pathway of living cells. Proc Natl Acad Sci U S A. 2005;102(18):6350–6355. doi: 10.1073/pnas.0501976102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Duellman T, Burnett J, Yang J. Quantitation of secreted proteins using mCherry fusion constructs and a fluorescent microplate reader. Anal Biochem. 2014 doi: 10.1016/j.ab.2014.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Cannon KS, Helenius A. Trimming and readdition of glucose to N-linked oligosaccharides determines calnexin association of a substrate glycoprotein in living cells. J Biol Chem. 1999;274(11):7537–7544. doi: 10.1074/jbc.274.11.7537. [DOI] [PubMed] [Google Scholar]
  • 26.Anelli T, Sitia R. Protein quality control in the early secretory pathway. Embo J. 2008;27(2):315–327. doi: 10.1038/sj.emboj.7601974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Wieland FT, Gleason ML, Serafini TA, Rothman JE. The rate of bulk flow from the endoplasmic reticulum to the cell surface. Cell. 1987;50(2):289–300. doi: 10.1016/0092-8674(87)90224-8. [DOI] [PubMed] [Google Scholar]
  • 28.Williams DB. Beyond lectins: the calnexin/calreticulin chaperone system of the endoplasmic reticulum. J Cell Sci. 2006;119(Pt 4):615–623. doi: 10.1242/jcs.02856. [DOI] [PubMed] [Google Scholar]
  • 29.Thor F, Gautschi M, Geiger R, Helenius A. Bulk flow revisited: transport of a soluble protein in the secretory pathway. Traffic. 2009;10(12):1819–1830. doi: 10.1111/j.1600-0854.2009.00989.x. [DOI] [PubMed] [Google Scholar]
  • 30.Wada I, Imai S, Kai M, Sakane F, Kanoh H. Chaperone function of calreticulin when expressed in the endoplasmic reticulum as the membrane-anchored and soluble forms. J Biol Chem. 1995;270(35):20298–20304. doi: 10.1074/jbc.270.35.20298. [DOI] [PubMed] [Google Scholar]
  • 31.Ou WJ, Cameron PH, Thomas DY, Bergeron JJ. Association of folding intermediates of glycoproteins with calnexin during protein maturation. Nature. 1993;364(6440):771–776. doi: 10.1038/364771a0. [DOI] [PubMed] [Google Scholar]
  • 32.Danilczyk UG, Cohen-Doyle MF, Williams DB. Functional relationship between calreticulin, calnexin, and the endoplasmic reticulum luminal domain of calnexin. J Biol Chem. 2000;275(17):13089–13097. doi: 10.1074/jbc.275.17.13089. [DOI] [PubMed] [Google Scholar]
  • 33.Olson MW, Gervasi DC, Mobashery S, Fridman R. Kinetic analysis of the binding of human matrix metalloproteinase-2 and -9 to tissue inhibitor of metalloproteinase (TIMP)-1 and TIMP-2. J Biol Chem. 1997;272(47):29975–29983. doi: 10.1074/jbc.272.47.29975. [DOI] [PubMed] [Google Scholar]
  • 34.Bode W, Maskos K. Structural basis of the matrix metalloproteinases and their physiological inhibitors, the tissue inhibitors of metalloproteinases. Biol Chem. 2003;384(6):863–872. doi: 10.1515/BC.2003.097. [DOI] [PubMed] [Google Scholar]
  • 35.Kamiya Y, Kamiya D, Yamamoto K, Nyfeler B, Hauri HP, Kato K. Molecular basis of sugar recognition by the human L-type lectins ERGIC-53, VIPL, and VIP36. J Biol Chem. 2008;283(4):1857–1861. doi: 10.1074/jbc.M709384200. [DOI] [PubMed] [Google Scholar]
  • 36.Lederkremer GZ. Glycoprotein folding, quality control and ERassociated degradation. Curr Opin Struct Biol. 2009;19(5):515–523. doi: 10.1016/j.sbi.2009.06.004. [DOI] [PubMed] [Google Scholar]
  • 37.Sandhu N, Duus K, Jorgensen CS, Hansen PR, Bruun SW, Pedersen LO, Hojrup P, Houen G. Peptide binding specificity of the chaperone calreticulin. Biochim Biophys Acta. 2007;1774(6):701–713. doi: 10.1016/j.bbapap.2007.03.019. [DOI] [PubMed] [Google Scholar]
  • 38.Duellman T, Burnett J, Shin A, Yang J. LMAN1 (ERGIC-53) is a potential carrier protein for matrix metalloproteinase-9 glycoprotein secretion. Biochem Biophys Res Commun. 2015 doi: 10.1016/j.bbrc.2015.06.164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Appenzeller-Herzog C, Roche AC, Nufer O, Hauri HP. pH-induced conversion of the transport lectin ERGIC-53 triggers glycoprotein release. J Biol Chem. 2004;279(13):12943–12950. doi: 10.1074/jbc.M313245200. [DOI] [PubMed] [Google Scholar]
  • 40.Elbein AD. Inhibitors of the biosynthesis and processing of N-linked oligosaccharide chains. Annu Rev Biochem. 1987;56:497–534. doi: 10.1146/annurev.bi.56.070187.002433. [DOI] [PubMed] [Google Scholar]
  • 41.Duksin D, Mahoney WC. Relationship of the structure and biological activity of the natural homologues of tunicamycin. J Biol Chem. 1982;257(6):3105–3109. [PubMed] [Google Scholar]
  • 42.Strous GJ, van Kerkhof P, van Meer G, Rijnboutt S, Stoorvogel W. Differential effects of brefeldin A on transport of secretory and lysosomal proteins. J Biol Chem. 1993;268(4):2341–2347. [PubMed] [Google Scholar]
  • 43.Zhang JX, Braakman I, Matlack KE, Helenius A. Quality control in the secretory pathway: the role of calreticulin, calnexin and BiP in the retention of glycoproteins with C-terminal truncations. Mol Biol Cell. 1997;8(10):1943–1954. doi: 10.1091/mbc.8.10.1943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.van Leeuwen JE, Kearse KP. Calnexin associates exclusively with individual CD3 delta and T cell antigen receptor (TCR) alpha proteins containing incompletely trimmed glycans that are not assembled into multisubunit TCR complexes. J Biol Chem. 1996;271(16):9660–9665. doi: 10.1074/jbc.271.16.9660. [DOI] [PubMed] [Google Scholar]
  • 45.Pipe SW, Morris JA, Shah J, Kaufman RJ. Differential interaction of coagulation factor VIII and factor V with protein chaperones calnexin and calreticulin. J Biol Chem. 1998;273(14):8537–8544. doi: 10.1074/jbc.273.14.8537. [DOI] [PubMed] [Google Scholar]
  • 46.Jannatipour M, Callejo M, Parodi AJ, Armstrong J, Rokeach LA. Calnexin and BiP interact with acid phosphatase independently of glucose trimming and reglucosylation in Schizosaccharomyces pombe. Biochemistry. 1998;37(49):17253–17261. doi: 10.1021/bi981785c. [DOI] [PubMed] [Google Scholar]
  • 47.Masure S, Proost P, Van Damme J, Opdenakker G. Purification and identification of 91-kDa neutrophil gelatinase. Release by the activating peptide interleukin-8. Eur J Biochem. 1991;198(2):391–398. doi: 10.1111/j.1432-1033.1991.tb16027.x. [DOI] [PubMed] [Google Scholar]
  • 48.Rudd PM, Mattu TS, Masure S, Bratt T, Van den Steen PE, Wormald MR, Kuster B, Harvey DJ, Borregaard N, Van Damme J, Dwek RA, Opdenakker G. Glycosylation of natural human neutrophil gelatinase B and neutrophil gelatinase B-associated lipocalin. Biochemistry. 1999;38(42):13937–13950. doi: 10.1021/bi991162e. [DOI] [PubMed] [Google Scholar]
  • 49.McLaughlin M, Alloza I, Quoc HP, Scott CJ, Hirabayashi Y, Vandenbroeck K. Inhibition of secretion of interleukin (IL)-12/IL-23 family cytokines by 4-trifluoromethyl-celecoxib is coupled to degradation via the endoplasmic reticulum stress protein HERP. J Biol Chem. 2010;285(10):6960–6969. doi: 10.1074/jbc.M109.056614. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Dalziel M, Crispin M, Scanlan CN, Zitzmann N, Dwek RA. Emerging principles for the therapeutic exploitation of glycosylation. Science. 2014;343(6166):1235681. doi: 10.1126/science.1235681. [DOI] [PubMed] [Google Scholar]
  • 51.Rizzuto R, Brini M, Pizzo P, Murgia M, Pozzan T. Chimeric green fluorescent protein as a tool for visualizing subcellular organelles in living cells. Curr Biol. 1995;5(6):635–642. doi: 10.1016/s0960-9822(95)00128-x. [DOI] [PubMed] [Google Scholar]
  • 52.Smith MJ, Koch GL. Multiple zones in the sequence of calreticulin (CRP55, calregulin, HACBP), a major calcium binding ER/SR protein. Embo J. 1989;8(12):3581–3586. doi: 10.1002/j.1460-2075.1989.tb08530.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Watzele G, Berger EG. Near identity of HeLa cell galactosyltransferase with the human placental enzyme. Nucleic Acids Res. 1990;18(23):7174. doi: 10.1093/nar/18.23.7174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Goubaeva F, Giardina S, Yiu K, Parfyonova Y, Tkachuk VA, Yang J. T-cadherin GPI-anchor is insufficient for apical targeting in MDCK cells. Biochem Biophys Res Commun. 2005;329(2):624–631. doi: 10.1016/j.bbrc.2005.02.020. [DOI] [PubMed] [Google Scholar]
  • 55.Duellman T, Warren CL, Peissig P, Wynn M, Yang J. Matrix metalloproteinase-9 genotype as a potential genetic marker for abdominal aortic aneurysm. Circ Cardiovasc Genet. 2012;5(5):529–537. doi: 10.1161/CIRCGENETICS.112.963082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Zamyatnin AA. Protein volume in solution. Prog Biophys Mol Biol. 1972;24:107–123. doi: 10.1016/0079-6107(72)90005-3. [DOI] [PubMed] [Google Scholar]
  • 57.Zimmerman JM, Eliezer N, Simha R. The characterization of amino acid sequences in proteins by statistical methods. J Theor Biol. 1968;21(2):170–201. doi: 10.1016/0022-5193(68)90069-6. [DOI] [PubMed] [Google Scholar]
  • 58.Lahtinen U, Dahllof B, Saraste J. Characterization of a 58 kDa cis-Golgi protein in pancreatic exocrine cells. J Cell Sci. 1992;103(Pt 2):321–333. doi: 10.1242/jcs.103.2.321. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS719471-supplement-01.pdf (2.2MB, pdf)

RESOURCES