Increased Urinary 3-Mercaptolactate Excretion and Enhanced Passive Systemic Anaphylaxis in Mice Lacking Mercaptopyruvate Sulfurtransferase, a Model of Mercaptolactate-Cysteine Disulfiduria

Mercaptopyruvate sulfurtransferase (Mpst) and its homolog thiosulfate sulfurtransferase (Tst = rhodanese) detoxify cyanide to thiocyanate. Mpst is attracting attention as one of the four endogenous hydrogen sulfide (H2S)/reactive sulfur species (RSS)-producing enzymes, along with cystathionine β-synthase (Cbs), cystathionine γ-lyase (Cth), and cysteinyl-tRNA synthetase 2 (Cars2). MPST deficiency was found in 1960s among rare hereditary mercaptolactate-cysteine disulfiduria patients. Mpst-knockout (KO) mice with enhanced liver Tst expression were recently generated as its model; however, the physiological roles/significances of Mpst remain largely unknown. Here we generated three independent germ lines of Mpst-KO mice by CRISPR/Cas9 technology, all of which maintained normal hepatic Tst expression/activity. Mpst/Cth-double knockout (DKO) mice were generated via crossbreeding with our previously generated Cth-KO mice. Mpst-KO mice were born at the expected frequency and developed normally like Cth-KO mice, but displayed increased urinary 3-mercaptolactate excretion and enhanced passive systemic anaphylactic responses when compared to wild-type or Cth-KO mice. Mpst/Cth-DKO mice were also born at the expected frequency and developed normally, but excreted slightly more 3-mercaptolactate in urine compared to Mpst-KO or Cth-KO mice. Our Mpst-KO, Cth-KO, and Mpst/Cth-DKO mice, unlike semi-lethal Cbs-KO mice and lethal Cars2-KO mice, are useful tools for analyzing the unknown physiological roles of endogenous H2S/RSS production.


Establishment of Three Independent Mpst Mutant Lines
The CRISPR/Cas9 genome editing was used to delete exon 2 of Mpst which encodes 67% of the Mpst open reading frame ( Figure 1A; and Supplementary Figure S1A for full DNA sequences) in one hundred C57BL/6J fertilized zygotes; from which 14 mice (9 males and 5 females) were born (14% birthrate). The Mpst deletion was apparent in two females and single male as revealed by tail DNA PCR ( Figure 1B) and confirmed by direct sequencing. The targeted region was deleted in the 1st and 3rd lines but a substantial portion of random DNA repair was found in the 2nd line (Supplementary Figure S1B-E). All three lines were successful in germline transmission. Mating of their progeny produced both heterozygous and homozygous KO mice (Het and KO, respectively) as manifested by tail DNA PCR ( Figure 1C). Mpst-Het and Mpst-KO mice were generally obtained with the expected frequency without marked sexual bias (Table 1). and is located proximal to its homolog Tst gene. The upstream (u) and downstream (d) crRNAs were designed to delete exon 2 which contains the start ATG codon and 67% of the entire open reading frame. Three independent mouse lines (1st, 2nd, and 3rd) were established. (B) Initial screening of 1st-3rd mouse lines from 14 independent mice (9 males and 5 females) that originated from individual fertilized zygotes electroporated with Cas9 protein, tracrRNA and crRNAs (u and d). PCR with forward (f) and reverse (r) primers detected the deletion of Mpst exon 2 in the 1st-3rd lines. (C) PCR detection of 1st and 3rdtype deletion using 1, 3, and r primers and 2nd-type deletion using 2, 4, and r primers from tail DNAs of wild-type (WT), Mpst-heterozygous (Het), and Mpst-homozygous (KO) mutant mice. and is located proximal to its homolog Tst gene. The upstream (u) and downstream (d) crRNAs were designed to delete exon 2 which contains the start ATG codon and 67% of the entire open reading frame. Three independent mouse lines (1st, 2nd, and 3rd) were established. (B) Initial screening of 1st-3rd mouse lines from 14 independent mice (9 males and 5 females) that originated from individual fertilized zygotes electroporated with Cas9 protein, tracrRNA and crRNAs (u and d). PCR with forward (f) and reverse (r) primers detected the deletion of Mpst exon 2 in the 1st-3rd lines. (C) PCR detection of 1st and 3rd-type deletion using 1, 3, and r primers and 2nd-type deletion using 2, 4, and r primers from tail DNAs of wild-type (WT), Mpst-heterozygous (Het), and Mpst-homozygous (KO) mutant mice.

Mpst, Tst, Cbs, and Cth Expression in Mpst Mutant Livers
We developed an anti-mouse Mpst rabbit polyclonal antibody by immunizing a rabbit with purified mouse full-length recombinant Mpst protein used in our assays. Mpst expression was widespread similar to its homolog Tst. All 26 mouse organs we tested expressed 33-kDa Mpst proteins (as well as 38-kDa Tst) to greater or lesser degrees. This is in stark contrast to the other two H 2 S/RSS-producing enzymes, Cbs and Cth (Figure 2A). In all 3 germ lines, Mpst-Het livers displayed approximately 50% Mpst expression levels of wild-type (WT) livers. Mpst-KO livers did not express Mpst, as expected ( Figure 2B). Therefore, the random DNA repair in the 2nd line (Supplementary Figure S1D) likely did not confer the expression of unexpected Mpst fragments. We found our Mpst antibody cross-reacts with 38-kDa Tst because pre-incubation of this antibody with recombinant Tst proteins abolished 38-kDa bands but not 33-kDa Mpst bands in the western analyses (Supplementary Figure S2). Using this antibody, we observed that hepatic Tst expression levels are not altered in all Mpst mutants ( Figure 2B). Furthermore, Western blot analyses of liver proteins using a specific anti-Tst rabbit monoclonal (that does not cross-react with mouse Mpst), anti-Cbs rabbit polyclonal, anti-Cth mouse monoclonal, and anti-glutathione peroxidase 1 (Gpx1, another major anti-oxidant protein) antibodies, did not detect any significant differences for Tst, Cbs, Cth, Gpx1 expression within Mpst mutant mice ( Figure 2C). Mpst and Tst are homologous proteins (63.5% similarity cDNA, 57.6% similarity protein sequence). Further, Mpst has some Tst activity and Tst has some Mpst activity [9]. Indeed, in vitro Mpst and Tst enzyme assays using 3-mercaptopyruvate (3-MP) and thiosulfate as substrates revealed that Mpst recombinant protein has some Tst activity while Tst recombinant protein has some Mpst activity at high substrate concentrations ( Figure 3A,B). However, at lower concentrations (5.95 mM 3-MP for Mpst and 25 mM thiosulfate for Tst), both enzymes displayed specific activities ( Figure 3A,B). At these reaction conditions and substrate concentrations, liver homogenates from Mpst mutant mice displayed specific Mpst and Tst activities that match well with their Mpst and Tst protein expression levels ( Figure 3A,B).

Mpst, Tst, Cbs, and Cth Expression in Mpst Mutant Livers
We developed an anti-mouse Mpst rabbit polyclonal antibody by immunizing a rabbit with purified mouse full-length recombinant Mpst protein used in our assays. Mpst expression was widespread similar to its homolog Tst. All 26 mouse organs we tested expressed 33-kDa Mpst proteins (as well as 38-kDa Tst) to greater or lesser degrees. This is in stark contrast to the other two H2S/RSS-producing enzymes, Cbs and Cth (Figure 2A). In all 3 germ lines, Mpst-Het livers displayed approximately 50% Mpst expression levels of wild-type (WT) livers. Mpst-KO livers did not express Mpst, as expected ( Figure 2B). Therefore, the random DNA repair in the 2nd line (Supplementary Figure S1D) likely did not confer the expression of unexpected Mpst fragments. We found our Mpst antibody cross-reacts with 38-kDa Tst because pre-incubation of this antibody with recombinant Tst proteins abolished 38-kDa bands but not 33-kDa Mpst bands in the western analyses (Supplementary Figure S2). Using this antibody, we observed that hepatic Tst expression levels are not altered in all Mpst mutants ( Figure 2B). Furthermore, Western blot analyses of liver proteins using a specific anti-Tst rabbit monoclonal (that does not cross-react with mouse Mpst), anti-Cbs rabbit polyclonal, anti-Cth mouse monoclonal, and anti-glutathione peroxidase 1 (Gpx1, another major anti-oxidant protein) antibodies, did not detect any significant differences for Tst, Cbs, Cth, Gpx1 expression within Mpst mutant mice ( Figure 2C). Mpst and Tst are homologous proteins (63.5% similarity cDNA, 57.6% similarity protein sequence). Further, Mpst has some Tst activity and Tst has some Mpst activity [9]. Indeed, in vitro Mpst and Tst enzyme assays using 3-mercaptopyruvate (3-MP) and thiosulfate as substrates revealed that Mpst recombinant protein has some Tst activity while Tst recombinant protein has some Mpst activity at high substrate concentrations ( Figure 3A,B). However, at lower concentrations (5.95 mM 3-MP for Mpst and 25 mM thiosulfate for Tst), both enzymes displayed specific activities ( Figure 3A,B). At these reaction conditions and substrate concentrations, liver homogenates from Mpst mutant mice displayed specific Mpst and Tst activities that match well with their Mpst and Tst protein expression levels ( Figure 3A,B).   (Mpst) and related proteins in adult wild-type and Mpst mutant mice. All samples were taken from males except for the ovary, uterus, and mammary gland. (A) Western blot analysis of Mpst, Tst (thiosulfate sulfurtransferase), Cbs (cystathionine β-synthase), Cth (cystathionine γ-lyase), Gpx1 (glutathione peroxidase 1), and Gapdh (glyceraldehyde-3-phosphate dehydrogenase; as a loading control) in various mouse tissues. (B) Western blot analysis of the liver Mpst (and Tst) in wild-type (WT), heterozygous (Het), and homozygous (KO) Mpst mutant mice (1st-3rd lines) using the anti-mouse Mpst rabbit polyclonal antibody produced in this study. Relative amounts of Mpst protein was expressed as % of the WT samples (mean ± SD; n = 3 each). (C) Hepatic expression of Tst, Cbs, Cth, Gpx1, and Gapdh using specific antibodies in WT, Het, and KO mice (1st-3rd lines). Relative expression of each protein was expressed as % of the WT samples (mean ± SD; n = 3 each).
(B) Western blot analysis of the liver Mpst (and Tst) in wild-type (WT), heterozygous (Het), and homozygous (KO) Mpst mutant mice (1st-3rd lines) using the anti-mouse Mpst rabbit polyclonal antibody produced in this study. Relative amounts of Mpst protein was expressed as % of the WT samples (mean ± SD; n = 3 each). (C) Hepatic expression of Tst, Cbs, Cth, Gpx1, and Gapdh using specific antibodies in WT, Het, and KO mice (1st-3rd lines). Relative expression of each protein was expressed as % of the WT samples (mean ± SD; n = 3 each).

Increased Urinary Excretion of 3-Mercaptolactate in Mpst-KO Mice
Serum amino acid/thiol compound levels for all lines of Mpst-KO mice were indistinguishable from those of WT mice, which was in marked contrast to Cth mice; however, all Mpst-KO mice excreted 5.5-7.3 times the normal amount of 3-mercaptolactate (3-ML) in urine ( Table 2). Though not significant, urinary (total) cysteine excretion tended to be higher in all Mpst-KO mice when comparted to WT mice ( Table 2). These results suggest that Mpst-KO mice displayed MCDU. In contrast, serum biochemistry (albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), creatine phosphokinase (CPK), creatinine (CRE), lactate dehydrogenase (LDH), total bilirubin (T-bilirubin), total protein (T-protein), uric acid (UA), and Thiobarbituric Acid Reactive Substances activity (TBARS)) did not reveal significant differences between WT and our Mpst-KO mice lines (Table 3), suggesting generally normal organ function, including the liver and kidney.

Generation of Mice Lacking Both Mpst and Cth
We have previously generated Cth-KO mice [23] that require cyst(e)ine as an essential amino acid [23]; display increased vulnerability to cadmium/methyl mercury [24], paraquat [23], acetaminophen [25], dietary methionine [26], cardiac ischemia/reperfusion injury [27], unilateral ureteral obstruction-induced kidney fibrosis [28], and decreased contraction responses to oxytocin [29]. Physiologically, they appear normal under laboratory conditions and are fertile [23]. Therefore, we crossbred Mpst (1st)-KO mice with Cth-KO mice to generate Mpst/Cth-DKO mice. Mpst/Cth-DKO mice were obtained at the expected frequencies without sexual bias (Table 1) and developed normally with some slightly altered serum biochemical parameters for (hepatic) functions (e.g., albumin, ALT, AST, CPK, LDH, and T-protein) when compared to the Mpst (1st)-KO mice (Table 3). We observed that Cth gene deletion only affects Cth expression while Mpst gene deletion only affects Mpst expression in the liver when compared to Tst, Cbs, and Gpx1 ( Figure 4). This indicates that compensatory expression of these proteins by Mpst/Cth deletion did not occur; although slight increases in hepatic TBARS activity were observed in Mpst/Cth-DKO mice (Table 3). Even breeding between Mpst/Cth-DKO males and females, as well as Mpst-het/Cth-KO males and females, could produce substantial numbers of pups although nearly halves of the breeding resulted in the deaths of all neonates, which is attributable to phenotypes (defective lactation) in Cth-KO mothers [29].

Enhanced PSA Response in Mpst-KO Mice
H2S has been shown to regulate various inflammatory responses [30,31], but its implications on anaphylaxis, "a serious generalized or systemic, allergic or hypersensitivity reaction that can be life-threatening or fatal [32,33]," remain unknown. Here we investigated dinitrophenyl (DNP)-conjugated human serum albumin (HSA)-induced PSA and passive cutaneous anaphylaxis (PCA) after pre-sensitization with anti-DNP IgE in WT, Mpst-KO, and Cth-KO mice as described in Materials and Methods. First, PSA was induced and changes in rectal temperature were followed. The hypothermic response associated with PSA was more pronounced in Mpst-KO mice compared to WT or Cth-KO mice; the maximum response was greater and the recovery was delayed ( Figure 5A). Next, PCA was induced and differences in Evans blue dye extravasation between left (not pre-sensitized) and right ears (pre-sensitized with anti-DNP IgE) were evaluated. In contrast to PSA, PCA responses were similar in all mice ( Figure 5B).

Enhanced PSA Response in Mpst-KO Mice
H 2 S has been shown to regulate various inflammatory responses [30,31], but its implications on anaphylaxis, "a serious generalized or systemic, allergic or hypersensitivity reaction that can be life-threatening or fatal [32,33]," remain unknown. Here we investigated dinitrophenyl (DNP)-conjugated human serum albumin (HSA)-induced PSA and passive cutaneous anaphylaxis (PCA) after pre-sensitization with anti-DNP IgE in WT, Mpst-KO, and Cth-KO mice as described in Materials and Methods. First, PSA was induced and changes in rectal temperature were followed. The hypothermic response associated with PSA was more pronounced in Mpst-KO mice compared to WT or Cth-KO mice; the maximum response was greater and the recovery was delayed ( Figure 5A). Next, PCA was induced and differences in Evans blue dye extravasation between left (not pre-sensitized) and right ears (pre-sensitized with anti-DNP IgE) were evaluated. In contrast to PSA, PCA responses were similar in all mice ( Figure 5B). Differences between pre-sensitized versus not pre-sensitized (vehicle alone) measurements were significant at ### P < 0.001 as well as those between genotypes at $$$ P < 0.001 in a two-way repeated-measured ANOVA with Bonferroni's multiple comparison test (Differences in each time point were significant at * P < 0.05, ** P < 0.01, and *** P < 0.001). (B) Analysis of ear edema in IgE-antigen-induced PCA in WT, Mpst-KO, and Cth-KO mice after antigen challenge. Differences in Evans blue dye extravasation between left (not pre-sensitized) and right ears (pre-sensitized with anti-DNP IgE) were evaluated. The ratio differences were significant (* P < 0.05 and *** P < 0.001, presensitized versus not pre-sensitized) in ratio paired t-tests. Differences in the ratios were not significant between the genotypes in a one-way ANOVA with Tukey's multiple comparison test.

Discussion
MCDU has historically been attributed to MPST deficiency, but studies containing direct experimental evidence are lacking. We provide here the first direct evidence that Mpst deficiency in mice causes enhanced secretion of 3-ML in urine, a hallmark trait of MCDU (Table 2). Cysteine is first metabolized to 3-MP by AST. Following, 3-MP is metabolized to pyruvate by Mpst or to 3-ML by LDH; and thus, Mpst deletion could lead to enhanced 3-ML excretion to urine [11]. The quantitative detection of 3-mercaptolactate-cysteine disulfide in urine samples is currently underway. MCDU patients are rarely found; thus, previous investigations into 3-ML or 3-mercaptolactate-cysteine disulfide quantities in MCDU patient urine samples have yet to be performed. Therefore, we do not know whether the 5.5-7.3-fold level increase of 3-ML in urine (Table 2) is mirrored in MCDU patients.
Previously generated Mpst-KO mice exhibited increased anxiety-like behaviors associated with increased levels of serotonin and its major metabolite (5-hydroxyindoleacetic acid) in the prefrontal cortex and decreased levels of dopamine in the hippocampus [7]. Further, 2-3-fold increases in hepatic (mitochondrial) Tst expression have been observed [10]. Because Tst has some Mpst activity (that can metabolize 3-ML, Figure 3 and [9]), the potential overexpression of Tst may have an Differences between pre-sensitized versus not pre-sensitized (vehicle alone) measurements were significant at ### p < 0.001 as well as those between genotypes at $$$ p < 0.001 in a two-way repeated-measured ANOVA with Bonferroni's multiple comparison test (Differences in each time point were significant at * p < 0.05, ** p < 0.01, and *** p < 0.001). (B) Analysis of ear edema in IgE-antigen-induced PCA in WT, Mpst-KO, and Cth-KO mice after antigen challenge. Differences in Evans blue dye extravasation between left (not pre-sensitized) and right ears (pre-sensitized with anti-DNP IgE) were evaluated. The ratio differences were significant (* p < 0.05 and *** p < 0.001, pre-sensitized versus not pre-sensitized) in ratio paired t-tests. Differences in the ratios were not significant between the genotypes in a one-way ANOVA with Tukey's multiple comparison test.

Discussion
MCDU has historically been attributed to MPST deficiency, but studies containing direct experimental evidence are lacking. We provide here the first direct evidence that Mpst deficiency in mice causes enhanced secretion of 3-ML in urine, a hallmark trait of MCDU (Table 2). Cysteine is first metabolized to 3-MP by AST. Following, 3-MP is metabolized to pyruvate by Mpst or to 3-ML by LDH; and thus, Mpst deletion could lead to enhanced 3-ML excretion to urine [11]. The quantitative detection of 3-mercaptolactate-cysteine disulfide in urine samples is currently underway. MCDU patients are rarely found; thus, previous investigations into 3-ML or 3-mercaptolactate-cysteine disulfide quantities in MCDU patient urine samples have yet to be performed. Therefore, we do not know whether the 5.5-7.3-fold level increase of 3-ML in urine (Table 2) is mirrored in MCDU patients.
Previously generated Mpst-KO mice exhibited increased anxiety-like behaviors associated with increased levels of serotonin and its major metabolite (5-hydroxyindoleacetic acid) in the prefrontal cortex and decreased levels of dopamine in the hippocampus [7]. Further, 2-3-fold increases in hepatic (mitochondrial) Tst expression have been observed [10]. Because Tst has some Mpst activity (that can metabolize 3-ML, Figure 3 and [9]), the potential overexpression of Tst may have an unexpected influence (e.g., alterations in neurotransmitter levels) on behavioral phenotypes in Mpst-KO mice. In this regard, our Mpst-KO mice lines have a great advantage over previous KO mice as hepatic Tst expression levels in our mice were comparable to that of WT mice ( Figure 2B,C). Indeed, in our preliminary analyses, our Mpst-KO mice were free of any behavioral disorders common to all mutant lines such as increased anxiety (data not shown), which is in good agreement with a previous report that showed the absence of obvious physical and mental disabilities in two MCDU sisters (11 and 13 years old) [6]. Expression of Cbs, Cth, and Gpx1 were also comparable between all Mpst-KO mice and WT mice ( Figure 2C). It should be noted that hepatic expression of Gpx1 (the anti-oxidant "selenoprotein") was not altered. This is important because Mpst may participate in selenium delivery to selenoproteins [34]. Accordingly, hepatic GSSG/tGSH ratios as well as TBARS activities (as the measures of oxidative stress/injury) in Mpst-KO mice were equivalent to those of WT mice ( Table 2). The phenotypic differences between previous Mpst-KO mice and our Mpst-KO mice could be solely attributable to different Mpst gene deletion strategies as the mouse Tst gene is located near the Mpst gene ( Figure 1A) and their genetic backgrounds are identical (C57BL/6J) [7].
In this study, we analyzed, for the first time, the systemic anaphylactic responses in Cth-KO and Mpst-KO mice and found greater PSA (but not PCA) responses in Mpst-KO mice ( Figure 5). In anaphylaxis, mast cells and macrophages secrete various inflammatory mediators including histamine, platelet-activating factor, and cysteinyl leukotrienes; thereby eliciting vasodilation and increased vascular permeability which leads to hypothermia [35,36]. Passive anaphylactic responses originate from mast cells that are pre-sensitized with antigen-specific IgE antibodies [33]. Because vascular permeability levels (in PCA assay) were equivalent between Mpst-KO and WT mice, enhanced hypothermic responses (in PSA assay) are likely caused by abnormal vasodilation and autonomous thermoregulation rather than mast cell responses. A recent study demonstrated that the administration of propargylglycine into mice, a non-specific inhibitor of Cth, interferes with LPS (lipopolysaccharide)-induced and brown adipose tissue-mediated thermogenesis [37]; suggesting a possible role for H 2 S/RSS in general thermoregulation. Further studies are necessary to delineate the roles of Mpst/Cth and produced H 2 S/RSS in the thermoregulation.
Furthermore, we generated Mpst/Cth-DKO mice that develop normally (with generally normal serum biochemical parameters; Table 3) which are fertile. Although Cth is located upstream of Mpst in cysteine metabolism, both enzymes could provide active sulfane sulfur from cysteine in parallel for cyanide detoxification to thiocyanate [13]. General incidences for inherited diseases of amino acid metabolism by single gene deletion are estimated to be less than 1/250,000 [38]; thus, the occurrence of MCDU combined with "Cth-deficient" cystathioninuria should be extremely rare. Nonetheless, Mpst/Cth-DKO mice as well as Mpst-KO and Cth-KO mice are a very useful in vivo analytical tool, which contrasts with (semi-)lethal Cbs-KO [39,40] and Cars2-KO mice [22]. Mpst/Cth-DKO mice generally displayed the combined phenotypes of Mpst-KO (increased 3-ML levels in urine; Table 2) and Cth-KO mice (e.g., increased serum levels of His, cystathionine, citrulline, and total homocysteine (Hcy); increased urinary levels of total Cys, total Hcy, total cysteinylglycine (Cys-Gly), and total γ-glutamylcysteine (γGlu-Cys); and reduced total glutathione (GSH) and increased GSSG/total GSH ratios in the liver) ( Table 2). We previously reported that Cth-KO mice look normal but display severe disorders upon various intervention, and our Mpst-KO mice and Mpst/Cth-DKO mice may be the similar. Although we are experiencing technical difficulties in obtaining sufficient numbers of Cth-KO pups from Cth-KO mothers due to defective milk ejection [29] irrespective of coexisting Mpst mutations, further studies including the examination of Mpst/Cth-DKO mice for PSA/PCA alterations are needed.
In conclusion, we generated Mpst-KO mice as an animal model of MCDU and also Mpst/Cth-DKO mice as a useful tool for analyzing functions of H 2 S/RSS that both enzymes may produce in vivo. Moreover, we found a novel phenotype in Mpst-KO mice, the enhanced passive systemic anaphylaxis responses. Physiological significance and roles of Mpst are yet unknown and the further studies using

PSA and PCA Assays
PSA and PCA assays were performed as previously described [49]. For PSA assays, WT, Mpst-KO, and Cth-KO mice were intravenously administered with 3 µg of anti-2,4-dinitrophenyl (DNP) mouse monoclonal IgE (SPE-7; Sigma-Aldrich) in 200 µL PBS or PBS alone through tail veins. At 24 h, the mice were challenged intravenously with 500 µg of DNP-conjugated human serum albumin (DNP-HSA (filtered through 0.22-µm filter (Millipore) before use); Sigma-Aldrich) in 200 µL PBS. After allergen challenge, rectal temperature was monitored every 5 min for 2 h using a BAT-12R microprobe thermometer equipped with a Mouse RET ISO Rectal Probe (Physitemp Instruments, Clifton, NJ, USA). For our PCA assays, mice were passively sensitized by intradermal injection with 10 ng of anti-DNP mouse monoclonal IgE (SPE-7) in 20 µL PBS. After 24 h, the mice were challenged by intravenous injection (through tail veins) with 20 µg of DNP-HSA in 200 µL saline containing 1 mg of Evans blue dye. Evans blue dye extravasation in the ear after 30 min was evaluated after sacrifice. Ears were removed and incubated at 37 • C overnight in 500 µL of 1 N KOH. The lysates were then mixed with 17.05 µL of 44 N phosphoric acid and 975 µL of acetone, and quantitative analysis of the dye in the extracts was performed within a linear range from 0.2 to 25 µg measuring the absorbance at 620 nm.

Statistical Analyses
Data were expressed as mean ± SD (n: sample numbers). Statistical comparison was performed using Prizm 5 software (GraphPad, San Diego, USA). Two-way repeated-measured ANOVA with Bonferroni's multiple comparison test was used in PSA assays ( Figure 5A) and ratio (pre-sensitized/not pre-sensitized) paired t-test was used in PCA assays ( Figure 5B). One-way ANOVA with Tukey's multiple comparison tests were used in all other experiments. All P values less than 0.05 denoted a significant difference.
Supplementary Materials: Supplementary Data can be found at http://www.mdpi.com/1422-0067/21/3/818/s1. Figure S1: Mouse Mpst gene and CRISPR/Cas9-mediated deletion of exon 2, Figure S2: Pre-incubation of anti-mouse Mpst rabbit polyclonal antibody with recombinant mouse Tst proteins abolishes 38-kDa (Tst) but not 33-kDa (Mpst) bands, Figure S3: Representative chromatograms in the measurements of free amino acids in serum samples, Figure S4: Representative chromatograms in the measurements of tryptophan in serum samples, Figure S5: Representative chromatograms in the measurements of thiol compounds in serum samples, Figure S6: Representative chromatograms in the measurements of 3-mercaptolactate in urine samples.