Biosynthesis of Human Myeloperoxidase

Abstract

Members of Chordata peroxidase subfamily [1] expressed in mammals, including myeloperoxidase (MPO), eosinophil peroxidase (EPO), lactoperoxidase (LPO), and thyroid peroxidase (TPO), express conserved motifs around the heme prosthetic group essential for their activity, a calcium-binding site, and at least two covalent bonds linking the heme group to the protein backbone. Although most studies of the biosynthesis of these peroxidases have focused on MPO, many of the features described occur during biosynthesis of other members of the protein subfamily. Whereas MPO biosynthesis includes events typical for proteins generated in the secretory pathway, the importance and consequences of heme insertion are events uniquely associated with peroxidases.

This Review summarizes decades of work elucidating specific steps in the biosynthetic pathway of human MPO. Discussion includes cotranslational glycosylation and subsequent modifications of the N-linked carbohydrate sidechains, contributions by molecular chaperones in the endoplasmic reticulum, cleavage of the propeptide from proMPO, and proteolytic processing of protomers and dimerization to yield mature MPO. Parallels between the biosynthesis of MPO and TPO as well as the impact of inherited mutations in the MPO gene on normal biosynthesis will be summarized. Lastly, specific gaps in our knowledge revealed by this review of our current understanding will be highlighted.

INTRODUCTION

The members of the Chordata peroxidase protein subfamily include myeloperoxidase (MPO), eosinophil peroxidase (EPO), lactoperoxidase (LPO), and thyroid peroxidase (TPO) [1] and each catalyzes physiologically important reactions. They share critical structural features but are expressed differently, both with respect to intracellular and tissue location. For example, MPO is a soluble protein in the matrix of azurophil granules of neutrophils, and to a much lesser extent monocytes, whereas TPO is a transmembrane protein in the plasma membrane [2]. Despite the biological importance of all of these mammalian peroxidases in human physiology, studies of their biosynthesis have been largely limited to that of MPO, with a few notable exceptions (vide infra and [3–6]0. Consequently, this Review focuses on known details of human MPO biosynthesis, provides an update on information presented earlier [7], summarizes parallels between the biosynthesis of MPO and TPO, and highlights specific gaps in our current understanding. The extent to which features of human MPO generation apply to biosynthesis of MPO from other species or to production of LPO and EPO is unknown but merits exploration.

MPO GENE

The gene encoding human MPO contains twelve exons and is located on chromosome 17q22 [8]. MPO gene expression occurs between the late myeloblast and promyelocyte stages of normal myeloid development [9–13] and rapidly shuts down as cells differentiate [14]. Consequently, MPO biosynthesis under normal conditions occurs only in myeloid precursors residing in bone marrow and not in mature phagocytes in circulation or in tissue. Only one of three identified promoter regions in human MPO operate in vivo, whereas all three are operational in the expression of murine MPO [14–18]. Many factors exert tight developmental and tissue control of MPO gene expression, both in normal myeloid precursors and in the context of acute leukemias [12, 19–21]. A thorough discussion of MPO gene regulation is beyond the scope of this review, but the importance of the -463GA polymorphism in the MPO promoter merits mention.

Reports from the Reynolds lab first described the existence of the -463G/A polymorphism in the Alu receptor response element in the promoter region of the primate MPO gene [22], with the -463G allele associated with greater gene expression. PPAR-γ agonists promote increased MPO gene expression in macrophages, a response that is blocked by estrogens [23] and by statins [24, 25]. The latter observations provide provocative links between augmented MPO expression and macrophage biology in inflammatory diseases such as atherosclerosis, a disease already associated with MPO-dependent modifications of lipoproteins and other elements critical in development of atherosclerosis (e.g. [26–31]), and neurologic diseases [32, 33]. Furthermore, many studies have examined the relationship between the -463GA polymorphism and the risk, either increased or decreased, of a broad spectrum of human disease, including cystic fibrosis, hypertension, malignancies, inflammatory vasculitides and neurodegenerative diseases [21, 34–44], to name a few. Given that many of the disorders of interest are polygenic in origin, the failure to identify clear-cut associations with a polymorphism in a single gene such as MPO has been challenging.

EVENTS IN THE ENDOPLASMIC RETICULUM

Glycosylation

The primary translation product for MPO undergoes cotranslational glycosylation in the endoplasmic reticulum (ER) to generate apoproMPO, the 90-kDa heme-free glycosylated precursor of MPO [45–51] (Figure 1). Endoglycosidase digestion of immunoprecipitates of MPO precursors generated by pulse-chase biosynthetic radioactive labeling by cultured human promyelocytic cell lines demonstrates that glycosylation is exclusively on asparagine residues and that five sites in apoproMPO undergo N-linked glycosylation [52]. However, more precise and direct analysis of MPO precursors identified six asparagine residues that are glycosylated: N139, N323, N355, N391, N483, and N729 [53, 54] (Figure 2). Whereas one site, N139, is in the propeptide, the remaining five carbohydrate sidechains reside at sites destined to be in the heavy subunit of mature MPO. Analyses of tryptic glycopeptides from recombinant MPO precursors and mature MPO reveal heterogeneity among the glycans, the significance of which remains unknown [53]. Partial deglycosylation of recombinant proMPO compromises its enzymatic activity [53], but the contribution of the oligosaccharide moieties at individual sites to enzymatic action or to productive biosynthesis has not been fully elucidated. In preliminary work not yet published, we have examined the importance of N-linked glycosylation on enzymatic activity and biosynthesis by replacing asparagine residues known to be glycosylated with glutamines. Stable transfectants expressing MPO with N139 replaced with Q (N139Q), either alone or in combination with N391Q, N323Q, or N483Q, generated fully active MPO. In contrast, N355Q, alone or in combination with other mutated N-glycosylation sites, compromised MPO activity of transfectants. In general, the importance of glycosylation in the biosynthesis and enzymatic activity of MPO needs additional study.

Figure 1. Overview of MPO biosynthesis.

Depicted are the currently recognized steps in MPO biosynthesis, starting with generation of the primary translation product, preMPO, in the ER. PreMPO undergoes cotranslational glycosylation of six asparagine residues, one in the propeptide and five in what will be the heavy subunit of the mature protein, to generate enzymatically inactive apoproMPO. ApoproMPO associates transiently and reversibly with ER molecular chaperones, including ERp57, calreticulin (CRT), and calnexin (CLN), and acquires heme to become enzymatically active proMPO. Upon exit from the ER, proMPO enters the secretory pathway where the majority undergoes proteolytic processing and dimerization to yield mature MPO and a small fraction (~10%) is secreted extracellularly as monomeric proMPO. In azurophilic granules of human neutrophils, MPO exists as a dimer composed to two identical heavy-light protomers, each linked together solely by three ester bonds with the heme group, with dimerization through a single disulfide bond between C319 on the heavy subunits of each heavy-light protomer.

Figure 2. Conversion from proMPO to mature MPO.

ProMPO undergoes three key modifications in its transformation into the mature form of MPO found in neutrophil granules. (a) An unidentified subtilisin-like convertase cleaves the 116 amino acid propeptide from proMPO, an events that occurs after proMPO has exited the ER. (b) A cysteine protease excises a hexapeptide, ASFVTG, from the monomeric 75-kDa intermediate species. (c) Pairs of heavy-light protomers dimerize via disulfide bond formation between the C319 residues in the heavy subunit of each protomer. Carbohydrate sidechains on N323 and N483 reside on the interfacial surface of the heavy-light protomers.

In addition to effects on intrinsic enzymatic activity and on productive biosynthesis, glycosylation of MPO may contribute to its binding and delivery to targets. In the crystal structure of mature MPO, glycans at N323 and N483 sit on the interface between the two protomers (Figure 2) and perhaps thereby provide stability to the dimer. The remaining sidechains (on N355, N391, and N729) decorate the exposed surface of the heavy subunit of mature dimeric MPO with six carbohydrate groups that are accessible and available to bind to surfaces and thus target MPO-dependent oxidants more precisely. The oligosaccharides on MPO have been implicated in neutrophil activation via engagement of β2 integrins [55] as well as in binding to microbes [56], macrophages [57–60], extracellular matrix [61], and lipids [62]. In the context of MPO and innate immunity against infection [63], glycosylation of MPO may benefit host defense. For example, MPO binds to microbes, especially Gram negative bacteria and some Gram positive microbes, and binding correlates with the relative susceptibility of the microorganisms to MPO-dependent killing [64]. By the same token, binding to host substrates, as occurs in the pathogenesis of atherosclerosis for example, would promote tissue damage and not serve in the best interest of the host (e.g. [65, 66]).

Molecular chaperones in the ER

Studies of MPO biosynthesis by cultured cells in vitro suggest that the half-life of apoproMPO in the ER is long [67–69], thus necessitating stabilization of biosynthetic intermediates during their prolonged stay in the ER. Carbohydrate sidechains on MPO precursors interact with ER resident molecular chaperones, ERp57, calreticulin, and calnexin [70, 71] (Figure 1). Such proteins typically interact transiently with nascent glycoproteins during their transit in the ER and serve to prevent their aberrant and unproductive folding [72–74]. None interacts with mature MPO subunits or with non-glycosylated MPO precursors synthesized in the presence of tunicamycin [75, 76], an antibiotic that blocks N-linked glycosylation [77]. Notably, calnexin associates transiently with both apoproMPO and proMPO, whereas ERp57, calreticulin, and calnexin each associate transiently with only apoproMPO [71]. In itself, the distinct properties of calreticulin and calnexin are not unexpected and have been previously described [78], but the contrast raises the possibility that calnexin binding may participate in some way in the acquisition of heme by apoproMPO and the generation of proMPO, although no data directly support that speculation.

Supportive evidence that calreticulin, calnexin, and ERp57 contribute to quality control during MPO biosynthesis comes from studies of the biosynthesis of missense mutations that recapitulate selected genotypes of inherited MPO deficiency. MPO precursors in transfectants that express R569W, Y173C, M251T, or G501S, all genotypes responsible for inherited MPO deficiency in humans, exhibit prolonged association with calreticulin, calnexin, or both [79–83], findings implicating their role in monitoring productive protein synthesis [84].

Acquisition of heme

Heme acquisition by apoproMPO and the resultant generation of proMPO is the rate-limiting step in MPO biosynthesis when studied in vitro [85–87]. Heme incorporation into apoproMPO is required for proMPO to exit the ER and undergo subsequent maturation and targeting to the azurophilic granule (Figure 1): blocking heme synthesis arrests MPO biosynthesis at the apoproMPO stage [85–87]. Brefeldin A, an agent that disrupts the Golgi network when added to cells in culture [88–90], arrests biosynthesis of MPO at the proMPO stage [85], thereby indicating that apoproMPO acquires heme within the ER. Although the spectral properties of proMPO residing in the ER have not been studied directly, detailed analysis of proMPO constitutively released extracellularly suggests that proMPO already possesses critical structures seen in mature MPO. The recent report of the crystal structure of proMPO and its solution structure derived from small angle X-ray scattering provides seminal insights into the structure and activity of proMPO [54]. The organization of the heme cavity in proMPO is nearly identical to that of mature MPO, suggesting that the heme prosthetic group and its immediate environs are fully developed in proMPO. Consistent with these structural data, recombinant proMPO and mature MPO share nearly identical structural, thermodynamic, and catalytic properties [91–96].

All members of the Chordata peroxidase subfamily in mammals have two ester bonds covalently linking the heme prosthetic group to the protein, with MPO unique in having an additional covalent bond formed by a sulfonium linkage between the 2-vinyl group and Met409 [96–99]. Detailed structural and functional analyses causally link the three covalent bonds in MPO with many of its unique features, including its spectral properties, catalytic activity, and its capacity to generate hypochlorous acid. Elegant studies by Ortiz de Montellano have elucidated the autocatalytic reactions that promote ester bond formation between the heme and the protein (reviewed in [96]). Addition of H₂O₂ to freshly isolated recombinant LPO generated in a baculovirus expression system increases ester bond formation around the heme group [100, 101]. Presumably, the same chemistry promotes sulfonium bond formation in proMPO. The source of H₂O₂ in vivo to promote autocatalytic oxidation in the ER has not been identified but the amount of H₂O₂ must be finely tuned, as high concentrations inactivate mature MPO [102].

EVENTS AFTER LEAVING THE ER

Between the moment that monomeric proMPO exits the ER and reaches its final stop as dimeric mature MPO in azurophilic granules, three processing events must occur: cleavage of the propeptide, excision of the hexapeptide ASFVTG, and dimerization of the two protomers to form mature MPO. Despite the importance of these steps in the biosynthetic pathway of MPO, data identifying either the intracellular site for each step or the relative order in which events occur are extremely limited.

Secreted monomeric proMPO

ProMPO exits the ER as a glycosylated monomer, with its heme group covalently bound and fully functional, and subsequently undergoes one of two different fates. For the bulk of proMPO, a series of proteolytic steps culminates in generation of dimeric mature MPO and its packaging in the azurophilic granules of neutrophils (Figure 1). A small fraction of newly synthesized proMPO is constitutively released extracellularly in cultured human bone marrow [48], cultured myeloid cell lines [52, 75, 103–106], and heterologous expression systems [107–109]. Although all published reports demonstrate that the catalytic properties of intracellular and secreted proMPO are the same [92], the two forms of proMPO differ with respect to their oligosaccharide sidechains. Endoglycosidase H, which attacks only high mannose oligosaccharides, digests the carbohydrate groups on intracellular proMPO and mature MPO, suggesting that both retain the high mannose sidechains that were added cotranslationally in the ER [49, 51, 52]. In contrast, secreted proMPO resists endoglycosidase H digestion, consistent with the presence of complex oligosaccharides [106], likely generated during passage of proMPO through the secretory pathway. The functional consequences of the modification of sidechains on secreted proMPO have not been determined.

Propeptide function

The contribution of the propeptide of proMPO to MPO biosynthesis has been the focus of many early studies in the field. Clearly, MPO production requires the propeptide, since deletion of the 116 amino acid N-terminal domain results in maturation arrest of the pro-deletion mutant in the ER and subsequent degradation [106, 110].

Several studies tested the hypothesis that the propeptide serves as a motif to target nascent MPO to azurophilic granules and not to the other granule compartments in neutrophils [105, 106, 111]. Transfectants expressing a variety of different constructs fused with the propeptide of MPO fail to deliver the fusion protein to an intracellular compartment, demonstrating that the propeptide of MPO does not contribute directly to sorting and targeting. In subsequent studies, Niels Borregaard demonstrated that segregation of neutrophil granule proteins into distinct intracellular storage compartments does not rely on targeting motifs in individual proteins or their precursors. Instead, the contents of individual granule populations reflect the order in which genes encoding the proteins of interest are expressed [112]. According to the “targeting by timing” mechanism, genes expressed early in the promyelocyte stage of granulopoiesis, such as MPO and elastase, are delivered to primary (aka azurophil) granules, as they form during the same developmental stage. By the same token, granules formed later, as in the myelocyte stage, contain granule proteins such as lactoferrin, whose genes are expressed in granulocyte precursors more mature than promyelocytes [113].

If not a targeting determinant, what function does the propeptide of MPO serve? Reasoning that the propeptide might operate as an intramolecular chaperone as seen in the synthesis of other proteins that contain propeptides in their precursor forms [114–117], we coexpressed the propeptide and a construct of MPO in which the propeptide had been deleted, expecting that the rescue of the propeptide deletion mutant by expressing the propeptide in trans would implicate the propeptide as a chaperone. However, no mature or active MPO-related proteins were generated [105], despite both constructs being expressed. These observations suggest that the propeptide does not operate in the strictest sense as an intramolecular chaperone. However, the recent success by the Obinger laboratory to crystallize recombinant monomeric human proMPO secreted by Chinese hamster ovary cells has provided critical insights into the structural organization of proMPO [54]. ProMPO possesses the same six disulfide bridges that exist in mature MPO but has one additional disulfide, namely C158-C319 (Figure 3). Mature MPO isolated from neutrophil azurophilic granules is dimeric, with the two pairs of heavy-light protomers joined via a single disulfide bond between the C319 residues in each heavy subunit. C158, on the other hand, resides in the propeptide of proMPO, bound there by a disulfide bond to C319. Mutation of C319 [54, 118] or C158 [54] interrupts MPO biosynthesis, with mutant species trapped in the ER, bound to molecular chaperones [54]. Thus C158 serves as a place holder, occupying C319 until a C319 from another proMPO monomer replaces it to form a dimeric species.

Figure 3. Cleavage of the propeptide.

Recent structural data demonstrate that C158 in the propeptide binds C319 in monomeric proMPO. In a post-ER compartment, a subtilisin-like proprotein convertase attacks ¹²⁸RKLRSLWR¹³⁶ in proMPO and cleaves the N-terminal propeptide. The otherwise free C319 forms a disulfide bond with a C319 residue in another monomeric heavy-light protomer to create the mature MPO dimer.

Cleavage of the propeptide

Removal of the propeptide from proMPO occurs in a post-ER compartment, since biosynthesis in the presence of brefeldin A is arrested at the proMPO stage, with no production of the 75-kDa intermediate or mature MPO subunits [85]. Proteolytic maturation of the fraction of proMPO destined to become dimeric mature MPO in azurophilic granules requires removal of the 116 amino acid propeptide. Addition of CMK-RVKR, an inhibitor of subtilisin-like proprotein convertases [119], arrests MPO biosynthesis at the stage of proMPO [109]. That is, only proMPO and not the 75-kDa intermediate is synthesized in the presence of CMK-RVKR. Given that proteolytic processing of a variety of protein precursors relies on the action of subtilisin-like proprotein convertases [120], we examined the propeptide of MPO for the presence of a motif for recognition by proprotein convertases, (K/R)-X_n-(K/R), where K/R indicates amino acids lysine or arginine, n represents 0, 2,4 or 6 residues, and X any amino acid (but rarely cysteine) [120–122]. Site-directed mutagenesis suggested that cleavage of proMPO occurs at ¹²⁸RKLRSLWR¹³⁶. Support for this prediction comes from the crystal structure of proMPO [54], which demonstrates that ¹²⁸RKLRSLWR¹³⁶ is accessible for attack by the proprotein convertase. Cleavage of the propeptide generates an enzymatically active monomer of ~75-kDa, the precursor of the heavy-light protomer (Figure 3).

Excision of ASFVTG

Akin and Kinkade described a 74-kDa protein derived from proMPO and recovered from crude granule preparations from HL-60 cells [104]. Whereas proMPO is associated with granules in a pH- and cation-dependent fashion, terminal processing of the 74-kDa intermediate occurs at neutral pH [123]. These data underscore two important aspects of MPO biosynthesis in human promyelocytes. First, they suggest the presence of an acidic compartment in the transporting of proMPO from ER to azurophilic granules. Second, they indicate that the generation of mature MPO from the 74-kDa intermediate occurs at neutral pH (7.5).

We had observed a 75-kDa species during our studies of MPO biosynthesis, but only occasionally in pulse chase experiments, which suggested to us that it was a transient intermediate that was rapidly converted to its product. However, biosynthetic radiolabeling of stable HEK transfectants expressing human MPO in the presence of any of three cysteine protease inhibitors, N-acetyl-leucyl-leucyl-methionine, N-acetyl-leucyl-leucyl-norleucine, or E-64, results in production of only the 75-kDa intermediate; no mature MPO is generated in the presence of cysteine protease inhibition [109]. These data suggest that a cysteine protease generates the heavy-light protomer containing the heavy and light chains of mature MPO joined together only through the covalent bonds to the heme group. The presumed target for excision, ASFVTG, is surface-exposed on the crystal structure and thus accessible to endoproteolytic attack [54]. However, the identity of the cysteine protease or endopeptidase that promotes excision of ASFVTG is not known.

Dimerization

Mature MPO exists as a dimer, composed of a pair of identical protomers containing heavy-light polypeptides that are covalently bound through the single heme group in each protomer, and the two protomers linked together via a single disulfide bond between the two heavy polypeptides [97–99] (Figure 1). The functional advantages of dimerization are not known, since the specific activity of hemi-MPO, created by reductive cleavage of mature MPO, is the same as that for mature MPO [124, 125], thereby suggesting that dimerization does not augment intrinsic peroxidase activity.

In general, disulfide bonds enhance protein stability and disulfide-bonded cysteines (i.e. half-cystines) are consistently favored over unpaired cysteines in proteins. In fact, half-cystines are positively selected during eukaryotic protein evolution and more conserved than is tryptophan, which is the most conserved amino acid [126]. Disulfides form in the ER but shuffling of specific sites accompanies protein folding during biosynthesis and trafficking within the secretory pathway [127–131]. Given these general principles, one would expect that dimerization via the C319-C319 bridge between the heavy polypeptides in two monomeric heavy-light protomers would follow soon after proteolytic cleavage of the propeptide. Such timing would allow for shuffling of the disulfides bonding of C319 from C158 in the propeptide to the C319 in a nearby MPO monomer and minimize the time that nascent MPO has an unpaired cysteine.

Data from the Kinkade laboratory demonstrate that the maturation of MPO precursors by HL-60 cells occurs in a subcellular fraction enriched in granules [104, 123, 132]. Furthermore, dimer formation in that experimental system is a late event and occurs with a t1/2 of ~36 hours, long after excision of ASFVTG and generation of heavy and light subunits, events with a t_1/2 of ~7 hours [104]. Recent structural data allow that removal of ASFVTG could occur either before or after dimerization [54]. Although propeptide cleavage occurs after egress of proMPO from the ER, the precise subcellular location of this event is not known. However, taken together, the available experimental data from studies of HL-60 cells suggest that propeptide cleavage, ASFVTG excision, and dimerization occur in granules, or a subcellular compartment that cosediments with granules.

Events in the biosynthesis of TPO

The only transmembrane protein in the cyclooxygenase peroxidase family, human TPO resides in the apical membrane of thyroid cells where it catalyzes the iodination of thyroglobulin (reviewed in [133]). Underscoring the clinical importance of TPO in thyroid health, mutations in the gene encoding TPO may be the most common cause of permanent congenital hypothyroidism secondary to abnormal thyroid dyshormonogenesis due to defective iodide organification [134, 135]. The significant progress made in identifying steps essential for biosynthesis and targeting of TPO parallels many of the features described above for MPO production. The physiologic presence of TPO as a transmembrane protein on the cell surface provides investigators with a sensitive means to assess successful expression. Biosynthesis of TPO in experimental settings can be evaluated using flow cytometry and antibodies directed against extracellular epitopes. Furthermore, sera from patients with autoimmune forms of thyroid disease provide a source for immunochemical tools to monitor optimal expression of TPO constructs in experimental systems.

Nascent TPO biosynthesis begins in the ER, where heme incorporation is a prerequisite for its exit from the ER and expression of functional enzyme at the cell surface. Inhibition of heme synthesis by succinyl acetone blocks productive expression of TPO, whereas the addition of heme or holotransferrin and aminolevulinic acid increases effective biosynthesis [136]. Expression of functional TPO at the cell surface increases in the presence of H₂O₂ added exogenously to cell culture [136], presumably a consequence of more efficient autocatalytic formation of the two ester bonds with the heme. Expressed in thyroid cells and formally known as Thox or thyroid oxidase [137], the NADPH oxidase protein family member Duox may provide H₂O₂ in the in vivo setting [138, 139]. As with proMPO, N-linked oligosaccharides on proTPO generated in the ER [140] support productive folding and early folding events in the ER depend on apparently competing interactions between proTPO and the molecular chaperones calreticulin, calnexin, ERp57, and BiP [141–143]. The relatively small fraction of human TPO that successfully reaches the apical surface of thyroid cells [144] suggests that complex folding events accompany the maturation of TPO [141].

After exit from the ER, proTPO undergoes O-linked glycosylation and cleavage of its propeptide by a proprotein convertase [140]. The importance of the propeptide in the biosynthesis of proTPO is controversial. The expression of TPO-MPO chimeric constructs in Chinese hamster ovary (CHO) cells is largely unsuccessful, with only one of seven constructs reaching the cell surface and only at very low levels [145] and suggest that the propeptide is not sufficient to target constructs to the plasma membrane. Reports of the impact of propeptide deletion on surface expression of TPO contradict each other. In one report [140], CHO transfectants expressing propeptide deletion mutants of TPO fail to deliver properly folded, enzymatically active TPO to the cell surface. In contrast, other investigators using the same expression system report that the propeptide deletion exhibits normal surface expression and enzyme activity [146]. Although the length of the signal peptide differs in the two constructs (14 vs 20 amino acids), the basis for the opposing results, and hence an understanding of the importance of the propeptide in TPO biosynthesis, remains unknown.

The crystal structure of TPO has not been solved, although crystals have been produced from recombinant TPO expressed in insect cells [147] and human TPO isolated from thyroid tissue from patients with Graves’ disease [148]. However, data from several sources, including recent molecular modeling [149], demonstrate that TPO, like MPO, exists as a homodimer, with C296, the conserved cysteine that is equivalent to C319 in MPO [150], forming the disulfide bond critical for dimerization.

Taken together, these data highlight parallels in the biosynthesis and structural organization of MPO and TPO, two members of the same protein family but with very different cellular locations.

KNOWLEDGE GAPS

Progress in understanding MPO biosynthesis has been considerable over the nearly four decades since Michi Yamada’s early studies of HL-60 cells [46]. However, that same progress prompts many questions, both old and new. For example:

• What is the source of H₂O₂ in the ER that promotes autocatalytic formation of ester bonds between the heme group and the protein backbone?

• How is heme transported from mitochondria, its likely site of synthesis, into the ER for insertion into apoproMPO?

• Is heme acquisition a chaperone-facilitated event? If so, by what chaperone(s)?

• What specific proprotein convertase catalyzes cleavage of the propeptide and in what subcellular compartment?

• What specific cysteine protease is responsible for excision of ASFVTG and where in the cell does it occur?

• What biological function, if any, is served by the secreted monomeric proMPO, locally in bone marrow or systemically in plasma?

• To what extent do the features identified in MPO biosynthesis apply to generation of the soluble peroxidases EPO and LPO, and, for that matter, other heme-containing proteins generated in the ER, such as NADPH oxidase protein [151–154]?

In addition, I think that there is merit in exploration of the possibility that trophic factors generated in specific inflammatory settings could promote resurrection of MPO gene expression in cells that do not typically produce MPO. For example, macrophages possess but do not actively produce MPO. However, in the setting of atherosclerosis, locally generated cytokines might initiate transcription of the MPO gene and promote synthesis and secretion of monomeric proMPO into the atheromatous plaque, thereby fueling additional tissue damage. Analogous events could occur in microglia that participate in neuroinflammatory disorders, as described for microglia in multiple sclerosis [155]. Reawakening of MPO production in such contexts could amplify inflammation and promote disease progression. Evidence for MPO gene expression in such settings might provide a rationale for modulating enzyme activity as a therapeutic intervention in inflammatory diseases.

For certain, there is much to learn about MPO production and biochemistry.