Published on: Author: Colin Thorpe

The term “sulfhydryl oxidase” was introduced more than 50 years ago to describe an activity characterized from mammalian skin that catalyzed the reaction ///

2 R-SH +  O2  ->  R-S-S-R  +  H2O2   ///

Later, enzymes from milk (1967) and seminal vesicles (1979) were found to catalyze the same reaction. In 1999 the Thorpe and Coppock laboratories recognized that all three oxidases were members of what we called the Quiescin-sulfhydryl oxidase (QSOX) family of flavin-linked oxidases. QSOXs are found in most eukaroyotes – from the smallest free-living eukaryote (Ostreococcus tauri) to humans. QSOXs are absent in the fungi. The evolutionarily unrelated Ero1 flavoprotein sulfhydryl oxidases have been found in all eukaryotes so far examined.

A chronology of some key events in the characterization of the flavin-linked sulfhydryl oxidases (together with other selected developments in oxidative protein folding) is presented below.  Our contributions can be found in “Publications“.



Year Entry



The term “sulfhydryl oxidase” was introduced by Rony et al. to describe an enzyme isolated from skin homogenates catalyzing the oxidation of thiols to disulfide bonds.



Sulfhydryl oxidase activity in skim milk was described by Kiermeier and Petz.  Subsequently Janolino and Swaisgood ascribed this activity to an iron-dependent enzyme.



Ostrowski and Kistler isolated a yellow protein from rat seminal vesicle and found that it was a sulfhydryl oxidase.
1996 Hoober et al. isolated a FAD-binding protein (subsequently found to be a sulfhydryl oxidase) from thousands of mutant chicken egg whites.
1998 The Kaiser and Weissman labs independently found that Ero1p was essential for oxidative protein folding in the yeast endoplasmic reticulum.
1999 First recognition of QSOX as a new enzyme family  …
The avian flavoprotein sulfhydryl oxidase (see 1996) was found by the Thorpe and Coppock laboratories to be homologous to human Quiescin Q6 (a protein secreted from cells as they approach confluence).  The preferred substrates appear to be reduced conformationally-mobile proteins.Why the abbreviation QSOX? – the obvious candidates: SOX, PSOX, SHOX had already been taken!  QSOX1 and QSOX2 are now the accepted HUGO gene names.

Tu et al. showed that Ero1p activity was dependent on FAD.

Benham et al. commence a characterization of human Ero1L-alpha.

Lisowsky et al. found that the yeast growth factor Erv1 was a FAD linked sulfhydryl oxidase.


Benayoun et al. independently found that a rat seminal vesicle sulfhydryl oxidase was homologous to Quiescin Q6.

ALR (the human analog of yeast Erv1) was found to be an FAD-linked sulfhydryl oxidase. [PubMed]

The Kaiser laboratory showed that Erv2p was a FAD-linked enzyme participating in oxidative folding in yeast. [PubMed]

Gerber et al. independently showed that Erv2p was a FAD-linked enzyme participating in oxidative folding in yeast.


Skin sulfhydryl oxidase was cloned by Matsuba et al. and found to be homologous to QSOX (it was previousy believed to be a copper-dependent oxidase).

The crystal structure of yeast Erv2p was determined [PubMed] by the Fass lab in collaboration with the Kaiser group. They proposed a model for catalysis involving a peripheral disulfide shuttling reducing equivalents from reduced protein disulfide isomerase to the proximal disulfide (that next to the flavin ring). Such “shuttles” are a widespread feature of catalysis by sulfhydryl oxidases.

2003 Wu et al. published the crystal structure of the first mammalian sulfhydryl oxidase: rat augmenter of liver regeneration (ALR).
2004 In a key development, the Fass and Kaiser laboratories reported [PubMed] the crystal structure of yeast Ero1p and showed the placement of shuttle and proximal disulfides within the flavoprotein.
2005 Farrell and Thorpe [PubMed] characterized recombinant human augmenter of liver regeneration. They found that cytochrome c was a much better oxidizing substrate than oxygen.They suggested that ALR located in the mitochondrial space might function as a cytochrome c reductase – thereby minimizing ROS.
2006 Gross et al. characterized the oxidative half-reaction of yeast Ero1p and showed that hydrogen peroxide was the immediate product of the reduction of molecular oxygen.

Sevier et al. proposed a model for regulation of yeast Ero1 in which the oxidase responds to the redox poise of the ER.

All of the sulfhydryl oxidase activity in skim milk that Jaje et al. could purify was the flavoprotein bovine QSOX1 – no iron-dependent disulfide bond generating enzyme was found (in contrast to prior reports),

The Bulleid group provided the first evidence that the long form of human QSOX1 functions intracellularly and is located primarily in the Golgi apparatus.


Appenzeller-Herzog et al. proposed a disulfide switch mechanism to regulate human Ero1L-alpha activity in the ER.

Baker et al. also purified human Ero1L-alpha and characterized its activity and potential regulatory behavior.

In an important step, Heckler et al. expressed human QSOX1b for the first time and probed the internal redox steps by mutagenesis.


The first human disease associated with a sulfhydryl oxidase mutation is an autosomal recessive myopathy caused by a point mutation in augmenter of liver regeneration (ALR/GFER) [PubMed]

The first structures of viral FAD-dependent sulfhydryl oxidases have been determined by Hakim and Fass: one from African swine fever virus (ASFV), and the other from a mimivirus.  Curiously, they do not share the same dimer interface.

Lake and colleagues correlated elevated levels of QSOX1 peptides in plasma with overexpression of the oxidase in human pancreatic cancer tumors.


Alon et al. determined that the helix rich region (HRR) domain of human QSOX is derived from the Erv/ALR FAD-binding domain – by gene duplication and loss of the flavin binding site.

The first single-thioredoxin QSOX (from the protist parasite Trypanosoma brucei) was enzymologically characterized [PubMed].

Zito et al. show that Ero1α and Ero1β paralogs are not essential in the mouse.  These oxidases play a surprisingly minor role in the folding and secretion of immunoglobulins.

Tavender et al. and Zito et al. independently report an alternate pathway for disulfide bond generation involving peroxiredoxin 4 coupled to the reduction of hydrogen peroxide.

Inaba and colleagues report the crystal structure of human Ero1-alpha and shed insight into the regulation of the oxidase and its interaction with PDI.

2011 Katchman et al.  investigated the role of QSOX1 in the invasion of pancreatic cancer cells.

Codding et al. explored the factors that modulate the in vitro activity of QSOX towards protein and peptide substrates.

In a major development for QSOX enthusiasts – the first crystal structures of QSOX (Trypanosoma brucei and mouse) were reported by Fass and colleagues


Ilani et al. identify a role of QSOX1 in the elaboration of the extracellular matrix in mammalian cells.

Grossman et al. develop a monoclonal inhibitory antibody that binds to the first thioredoxin domain of QSOX1.

Lake and colleagues showed that higher expression levels of QSOX1 are associated with a poorer prognosis for luminal B breast cancer patients.

Bulleid and coworkers found that proteolytic processing of the long forms of QSOX led to efficient secretion of the active oxidase from mammalian cells.


Israel et al. explored the thermodynamic coupling of disulfide exchange reactions in QSOX catalysis – “going through the barrier …”

A sensitive new assay for QSOX – suitable for a variety of biological samples, including blood – was developed by Israel et al.

A new crystal structure of rat QSOX1 shows two CxxC motifs aligned for efficient disulfide exchange reactions (Gat et al.)


The exploration of multivalent arsenical inhibitors of oxidoreductases using CxxC motifs (Sapra et al.)

Grossman et al. studied conformational sampling in QSOX by single molecule FRET.


Print Friendly, PDF & Email