Specific Recognition of β-Galactofuranose-Containing Glycans of Synthetic Neoglycoproteins by Sera of Chronic Chagas Disease Patients

Chagas disease (CD) can be accurately diagnosed by detecting Trypanosoma cruzi in patients’ blood using polymerase chain reaction (PCR). However, parasite-derived biomarkers are of great interest for the serological diagnosis and early evaluation of chemotherapeutic efficacy when PCR may fail, owing to a blood parasite load below the method’s limit of detection. Previously, we focused on the detection of specific anti-α-galactopyranosyl (α-Gal) antibodies in chronic CD (CCD) patients elicited by α-Gal glycotopes copiously expressed on insect-derived and mammal-dwelling infective parasite stages. Nevertheless, these stages also abundantly express cell surface glycosylphosphatidylinositol (GPI)-anchored glycoproteins and glycoinositolphospholipids (GIPLs) bearing nonreducing terminal β-galactofuranosyl (β-Galf) residues, which are equally foreign to humans and, therefore, highly immunogenic. Here we report that CCD patients’ sera react specifically with synthetic β-Galf-containing glycans. We took a reversed immunoglycomics approach that entailed: (a) Synthesis of T. cruzi GIPL-derived Galfβ1,3Manpα-(CH2)3SH (glycan G29SH) and Galfβ1,3Manpα1,2-[Galfβ1,3]Manpα-(CH2)3SH (glycan G32SH); and (b) preparation of neoglycoproteins NGP29b and NGP32b, and their evaluation in a chemiluminescent immunoassay. Receiver-operating characteristic analysis revealed that NGP32b can distinguish CCD sera from sera of healthy individuals with 85.3% sensitivity and 100% specificity. This suggests that Galfβ1,3Manpα1,2-[Galfβ1,3]Manpα is an immunodominant glycotope and that NGP32b could potentially be used as a novel CCD biomarker.

Here we show the synthesis of two β-Galf-containing glycans, which could potentially be terminal partial structures of a TCT GPI anchor ( Figure 1A). The glycans were conjugated to bovine serum albumin (BSA), and the resulting NGPs were immunologically evaluated using sera of CCD patients as well as sera of healthy individuals as a negative control group.

Results and Discussion
We hypothesized that a β-Galf-based glycotope could be identified and exploited as a new specific BMK for CD which, unlike previously discovered glycan-based BMKs for CD, is free of α-Gal, and therefore shows less cross-reactivity with NHS. To address this hypothesis, we took a reversed immunoglycomics approach [40] that began with the synthesis of Galfβ1,3Manpα-CH2CH2CH2SH (G29SH) and Galfβ1,3Manpα1,2-[Galfβ1,3]Manpα-CH2CH2CH2SH (G32SH) ( Figure 1B). These target compounds contain terminal disaccharide and tetrasaccharide partial structures of a TCT GPI-anchor ( Figure 1A) [31,33], respectively, and a 3-thiopropyl moiety allowing for conjugation to maleimide-derivatized BSA. Derivatives of Galfβ1,3Manpα [41,42] and Galfβ1,3Manpα1,2-[Galfβ1,3]Manpα [41], as well as the terminal trisaccharide of a Leishmania type-2 GIPL [40,42], and a Leishmania lipophosphoglycan (LPG) core heptasaccharyl myo-inositol [43], and a T. cruzi LPPG heptasaccharyl myo-inositol [44] have previously been synthesized. Our synthesis of disaccharide G29SH and its conjugation to commercially available maleimide-activated BSA is illustrated in Scheme 1. The benzoylated thiogalactofuranoside donor 1 [45] was used to glycosylate allyl mannoside acceptor 2 [46] to produce disaccharide 3. The regioselectivity using this or a similar mannosyl acceptor had previously been shown [40,42]. Hydrolysis of the benzylidene protecting group and radical addition of thioacetic acid to the terminal alkene afforded thioester 4. Global removal of all acyl groups under Zemplén conditions gave the target disaccharide G29SH, which oxidized to (G29S)2 on air. Upon reduction of the disulfide with TCEP, G29SH was conjugated to commercial maleimide-derivatized BSA to give rise to the neoglycoprotein (NGP) NGP29b with a high payload of 30 units of disaccharide per BSA molecule, as observed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) (see Supplementary Material). Scheme 2 shows the synthesis of tetrasaccharide G32 SH and its conjugation to maleimidederivatized BSA to produce NGP32b. First, previously synthesized disaccharide 3 was acety- lated, then the anomeric allyl group was removed under PdCl 2 catalysis, and the resulting hemiacetal was converted to trichloroacetimidate 5. This donor was used to glycosylate disaccharide acceptor 3 to furnish the fully protected tetrasaccharide 6 in good yield. Hydrolysis of the benzylidene acetal followed by radical addition of thioacetic acid to the terminal alkene gave the thioester S5 (see Supplementary Material). Global deprotection of all ester groups under Zemplén conditions afforded G32 SH , which oxidized to (G32 S ) 2 on air. The disulfide was reduced with TCEP followed by conjugation to commercial maleimide-derivatized BSA to furnish NGP32b, as observed by MALDI-TOF-MS (see Supplementary Material). Scheme 1. Synthesis of disaccharide G29SH and its conjugation to BSA affording NGP29b: (a) NIS, AgOTf, DCM 0 °C to rt, 70%; (b) TFA, H2O, rt, 53%; (c) AcSH, AIBN, THF, UV light (λ = 350 nm), 79%; (d) NaOMe, MeOH, rt, quant.; and (e) TCEP, maleimide-activated BSA, phosphate buffer pH 7.2. Scheme 2 shows the synthesis of tetrasaccharide G32SH and its conjugation to maleimide-derivatized BSA to produce NGP32b. First, previously synthesized disaccharide 3 was acetylated, then the anomeric allyl group was removed under PdCl2 catalysis, and the resulting hemiacetal was converted to trichloroacetimidate 5. This donor was used to glycosylate disaccharide acceptor 3 to furnish the fully protected tetrasaccharide 6 in good yield. Hydrolysis of the benzylidene acetal followed by radical addition of thioacetic acid to the terminal alkene gave the thioester S5 (see Supplementary Material). Global deprotection of all ester groups under Zemplén conditions afforded G32SH, which oxidized to (G32S)2 on air. The disulfide was reduced with TCEP followed by conjugation to commercial maleimide-derivatized BSA to furnish NGP32b, as observed by MALDI-TOF-MS (see Supplementary Material). With the NGP29b and NGP32b in hand, IgG Ab responses using sera from CCD patients could now be measured by CL-ELISA, as previously described [40,[47][48][49]. First, to identify the most suitable CL-ELISA conditions, sera from 10 patients with CCD from With the NGP29b and NGP32b in hand, IgG Ab responses using sera from CCD patients could now be measured by CL-ELISA, as previously described [40,[47][48][49]. First, to identify the most suitable CL-ELISA conditions, sera from 10 patients with CCD from an endemic region (Venezuela) and 10 sera from healthy individuals from a nonendemic region (USA) were pooled. Figure 2 shows the CL-ELISA cross-titrations of varying pooled serum dilutions and NGP antigen loadings [ng/well]. Analysis of the data indicated that a serum dilution of 1:800 and 50 ng of NGP antigen per well gave the best differential IgG responses when comparing CCDSP vs. NHSP. Under these conditions, the difference of IgG reactivity between the CCDSP and the NHSP was a factor of 24 for NGP29b, and a factor of 32 for NGP32b. These differential reactivities are by a factor of roughly 2-3 greater when compared to Galpα1,3Galpβ-containing NGPs observed in similar CL-ELISA studies [47,49]. Only one NGP that contained Galpα1,3Galpβ1,4GlcNAcα ("NGP24b") was able to reach an antibody differential reactivity of a factor of~20 [48]. Our previous studies with CCD and cutaneous leishmaniasis patient sera had shown that the binding of serum components to BSA or the same crosslinker present in NGP29b and NGP32b was negligible [40,49]. The cross-titrations shown in Figure 2 already indicated that both, NGP29b and NGP32b, could potentially be BMKs for CD. However, to obtain sensitivity and specificity data, which determine their suitability as a BMK, CL-ELISAs of individual sera (not pooled sera) are necessary. This experiment aimed to identify CL-ELISA conditions that give the largest differential reactivity between CCD sera (red) and NHS control group (green). Both NGPs were titrated at different concentrations (3.12, 6.25, 12.5, 25, 50, 100, 200, or 400 ng/well) with serum pools from ten CCD patients and from ten healthy individuals, both diluted at 1:200, 1:400, 1:800, or 1:1600. The CL-ELISA was performed as described in the Supplementary Material. CCDSP, CCD serum pool; NHSP, normal human serum pool; RLU, relative luminescence units.
Individual CCD serum samples of 75 patients from the city of Salta, Argentina, and 15 sera from healthy individuals from the USA were assayed. Initially, CL-ELISA endpoint titers of individual samples were defined based on a meaningful statistical method [50], which was originated from each immunoassay microtiter plate by using nine technical replicates of NHSP as negative control to increase accuracy, with a confidence interval (CI) (1-α) ranging 95.0 to 99.9%. In our analyses, a CI = 99.5% with a standard deviation multiplier (f) of 3.537 were chosen to calculate the cutoff for each immunoassay plate (see Supplementary Material). Each individual sample was tested in triplicate. Based on the CL-ELISA data using individual CCD patient sera and individual NHS sera ( Figure 3A), NGP29b diagnosed as positive 60/75 (80.0% sensitivity), whereas NGP32b diagnosed as positive 65/75 (86.7% sensitivity) (Tables 1 and 2). Using a titer cut-off value of 1.000, NGP29b and NGP32b exhibited a specificity of 80.0% and 93.3%, respectively. To compare the sensitivities and specificities of NGP29b and NGP32b, receiver-operating characteristic (ROC) curves were plotted ( Figure 3B). The area under curve (AUC) values of the ROC curves are 0.9124 and 0.9636 for NGP29b and NGP32b, respectively, indicating that NGP32b exhibited a higher sensitivity and specificity than NGP29b. Following a method developed by Greiner et al. [51], the initially selected titer cutoff value of 1.000 (Ci; Figure 3A) for NGP29b and NGP32b was fine-tuned by performing a two-graph ROC (TG-ROC) analysis. The ROC data ( Figure 3B) for sensitivity (Se) and specificity (Sp) were plotted against the cutoff value ( Figure 3C). This TG-ROC plot allows the selection of a cutoff value that best matches the selectivity and specificity needs of the assay. In the case of BMK discovery for CD, there is a need for determining treatment efficacies by serology. A higher specificity of the assay is needed to distinguish antibody responses due to the This experiment aimed to identify CL-ELISA conditions that give the largest differential reactivity between CCD sera (red) and NHS control group (green). Both NGPs were titrated at different concentrations (3.12, 6.25, 12.5, 25, 50, 100, 200, or 400 ng/well) with serum pools from ten CCD patients and from ten healthy individuals, both diluted at 1:200, 1:400, 1:800, or 1:1600. The CL-ELISA was performed as described in the Supplementary Material. CCDSP, CCD serum pool; NHSP, normal human serum pool; RLU, relative luminescence units.
Individual CCD serum samples of 75 patients from the city of Salta, Argentina, and 15 sera from healthy individuals from the USA were assayed. Initially, CL-ELISA endpoint titers of individual samples were defined based on a meaningful statistical method [50], which was originated from each immunoassay microtiter plate by using nine technical replicates of NHSP as negative control to increase accuracy, with a confidence interval (CI) (1-α) ranging 95.0 to 99.9%. In our analyses, a CI = 99.5% with a standard deviation multiplier (f ) of 3.537 were chosen to calculate the cutoff for each immunoassay plate (see Supplementary Material). Each individual sample was tested in triplicate. Based on the CL-ELISA data using individual CCD patient sera and individual NHS sera ( Figure 3A), NGP29b diagnosed as positive 60/75 (80.0% sensitivity), whereas NGP32b diagnosed as positive 65/75 (86.7% sensitivity) (Tables 1 and 2). Using a titer cut-off value of 1.000, NGP29b and NGP32b exhibited a specificity of 80.0% and 93.3%, respectively. To compare the sensitivities and specificities of NGP29b and NGP32b, receiver-operating characteristic (ROC) curves were plotted ( Figure 3B). The area under curve (AUC) values of the ROC curves are 0.9124 and 0.9636 for NGP29b and NGP32b, respectively, indicating that NGP32b exhibited a higher sensitivity and specificity than NGP29b. Following a method developed by Greiner et al. [51], the initially selected titer cutoff value of 1.000 (C i ; Figure 3A) for NGP29b and NGP32b was fine-tuned by performing a two-graph ROC (TG-ROC) analysis. The ROC data ( Figure 3B) for sensitivity (Se) and specificity (Sp) were plotted against the cutoff value ( Figure 3C). This TG-ROC plot allows the selection of a cutoff value that best matches the selectivity and specificity needs of the assay. In the case of BMK discovery for CD, there is a need for determining treatment efficacies by serology. A higher specificity of the assay is needed to distinguish antibody responses due to the presence from residual T. cruzi trypomastigotes from a weak serum cross-reactivity that is observed with NHS, and that is expected after CD patients have been completely cured. For NGP29b, when the titer cutoff value is adjusted to 1.385, only 59/75 (78.7% sensitivity) CCD patients are correctly diagnosed, but the specificity of the assay increases to 93.3% (Tables 1 and 2). In case of NGP32b, an adjustment of the CL-ELISA titer value to 1.130 correctly diagnosed 64/75 (85.3% sensitivity) CCD patients, and the specificity increases to 100% (Tables 1 and 2). These data show that NGP32b, in particular, has an excellent potential as a diagnostic serological BMK. Moreover, the sensitivity and specificity performances of NGP32b suggest that the terminal Galf β1,3Manpα1,2[Galf β1,3]Manpα tetrasaccharide is an immunodominant glycotope present in T. cruzi trypomastigote cell surface glycoconjugates, most likely in the abundant GIPLs and GPI moieties of major glycoproteins like tGPI-mucins [27,52]. The structurally different β-Galf -containing glycans present in the O-linked glycans of GPI-mucins from Meta forms and, potentially, TCT forms of certain T. cruzi strains, particularly those belonging to the genotype or discrete typing unit (DTU) Tc I [30,34,53], could also elicit anti-β-Galf Abs. Such Abs would be expected to have a different specificity, but they could potentially cross-react with NGP29b and NGP32b, and contribute to the CL-ELISA signal. One explanation for NGP32b's superior performance, in both sensitivity and specificity, over NGP29b could be that NGP32b comprises a larger glycotope with two appending β-Galf units, which may be more immunogenic and antigenic than a glycotope that is only half its size. Consequently, the IgG Abs elicited may have a higher titer in the CCD patients' sera, and/or bind more strongly to the tetrasaccharide glycotope present in NGP32b. Another possible explanation could be that CCD patient sera contain Abs that recognize the Galf β1,3Manpα glycotope in a certain conformation that may be preferred in the molecular environment of the Galf β1,3Manpα1,2[Galf β1,3]Manpα tetrasaccharide.  The CCD sera used in this study stemmed from patients of endemic regions that are geographically far apart. The pooled CCD sera were collected in Venezuela, and the individual CCD sera were collected in Argentina. Different T. cruzi DTUs predominate in these areas of sample collection (i.e., TcI in Venezuela and TcV in Argentina) [54]. Since the cross-titration results using pooled sera from Venezuela ( Figure 2) perfectly match the CL-ELISA results using individual patient sera from Argentina ( Figure 3A), one can assume that galactofuranosylated GPI anchors are common in different T. cruzi DTUs. If this proves to be true, new β-Galf -based diagnostic and prognostic tools could potentially be universally applicable throughout Latin America, and even in nonendemic regions like the U.S. and Europe, whose parasite genotype makeup circulating in the T. cruzi-infected population usually reflects that of the country of origin in Latin America. For instance, in Spain the majority of CCD patients are originated from Bolivia and are infected with DTU TcV or TcIV [55].
The reversed immunoglycomics approach for BMK discovery, as used in this study, is at its best when exact structural data are absent and/or when the identity or exact size of a glycotope is not known [40]. Through a combination of synthetic chemistry and serology, this approach can provide support for the existence of certain glycotopes within a larger cell surface glycan expressed by an organism. Classical immunoglycomics is a top-down approach that can, in principle, also lead to the identification of glycotopes, but here, it would require the cultivation of T. cruzi trypomastigotes at a large scale; a number of glycoproteomics techniques, for example, enzymatic and/or chemical cleavage of glycans from a glycocalyx of enormous heterogeneity and complexity; fluorophore labeling; glycoprofiling by high-resolution mass spectrometry (HR-MS), tandem MS, and nuclear magnetic resonance (NMR) spectroscopy. On the other hand, reversed immunoglycomics is a bottom-up approach that combines the synthesis of suspected glycotopes with conjugating them to a carrier protein to produce NGP antigens, and probing them for antigenicity in serological assays using patient sera [40]. The purpose of protein conjugation is to generate antigens that adhere to the wells of microtiter plates. A strong antibody reactivity, and a large differential to negative control sera is indicative of the presence of a glycotope on the parasite cell surface and can lead to BMK discovery.
The presence of anti-β-Galf Abs in the serum of patients with CCD was first observed by Schnaidman et al. [18]. Those authors showed that human CCD sera and a rabbit anti-T. cruzi epimastigote antiserum were able to precipitate galactomannans partially purified from T. cruzi epimastigotes (Tc-GM). The binding of both human CCD and anti-T. cruzi epimastigote rabbit sera to T. cruzi eGM could be completely abolished (100%) by incubation with 40 µM βGalf (1,3)-methyl-α-D-Manp, clearly indicating that the T. cruzi GM preparation contained the same βGalf (1,3)α-D-Man epitope found in the GIPLs (formerly, LLPG). That study also demonstrated that preparations of Aspergillus fumigatus GM (Af -GM) and galactoglucomannans from Dactylium dendroides (Dd-GGM), both known to contain various terminal nonreducing β-Galf epitopes, could partially cross-react with human CCD sera and a rabbit anti-T. cruzi epimastigote antiserum [18,32]. Subsequently, Golgher et al. showed that a rabbit antiserum raised against purified LLPG (=GIPLs) could strongly bind to T. cruzi epimastigote lysates from ten different strains and clones (Y, G, CL-14, CA1, Tulahuén, Ana, Basileu, M226, DM28, and YuA), and at least to one lysate from TCTs of Y strain in western blot, indicating a widespread distribution of the β-Galf epitopes in the parasites from distinct genotypes and geographical origins. The major bands recognized were LPPG/GIPLs and mucins (then known as bands ABC) in epimastigotes, and LPPG/GIPLs and GP80-90 kDa in TCTs. At the same time, Almeida et al. [33] showed that an organic extract (known as fraction F1), obtained from the TCT stage and enriched with βGalf -containing GIPLs (based on the reactivity to an antiserum from a rabbit immunized with purified LPPG from T. cruzi epimastigotes), was recognized by 100% (n = 115) of the sera from a cohort of untreated CCD patients, in a CL-ELISA. Interestingly, patients (n = 28) subjected to BZN chemotherapy but still positive for the conventional serology (CS) while negative for the trypanolytic antibody assay (TAA), thus named dissociated patients [56][57][58][59], showed much lower reactivity to F1, with 4/28 (14%) already negative for F1 by CL-ELISA. Meanwhile, all treated and cured patients (exhibiting both negative CS and TAA) (n = 5) were negative for F1, whereas all treated but not cured patients (n = 15) were positive for F1. Taken together, these previous studies confirmed the immunodominance of anti-βGalf Abs among CCD patients [12,18,20,60] and open the possibility of using purified TCT GIPLs or synthetic βGalf -containing NGPs for accurate diagnosis and early assessment of the chemotherapy of CD.
Finally, while similar reversed immunoglycomics approaches were previously applied by us and others to develop diagnostic and prognostic tools for CD, these research endeavors focused on NGPs that display T. cruzi-derived α-Galp moieties [47][48][49]61,62]. However, an extensive validation of all potential BMKs discovered by this method is necessary using large patient sera panels and large panels of control sera of healthy individuals or patients with heterologous diseases. . To a flask containing thioester S5 (72 mg, 0.043 mmol), 13 mL of 0.5 M NaOMe in MeOH was added under Ar, and the mixture was stirred at rt for 2 h. HRMS showed complete removal of the protecting groups. The reaction mixture was neutralized with Amberlyst-15 ion-exchange resin, and filtered through Celite. The MeOH was removed by rotary evaporation, and the remainder was dissolved in water and lyophilized. Initially, the unprotected 3-thiopropyl disaccharide G32 SH was produced, which oxidized to the disulfide by handling on air within hours. (G32 S ) 2 was obtained as a white powder (25 mg, 75%). 1

Conjugation Protocol
The kit for the conjugation of the thiol-containing glycans (G29 S ) 2 , or (G32 S ) 2 , to BSA (Imject™ Maleimide-Activated BSA, Thermo Fisher Scientific, Waltham, MA, catalog #77116), was purchased from Thermo Fisher Scientific, and the conjugation procedure followed was similar to the one described by the manufacturer and previously published [61]. A 0.05 M TCEP solution was prepared by diluting neutral TCEP solution (Bond-Breaker TCEP solution, 0.5 M, Thermo Fisher Scientific, Waltham, MA, catalog #77720) with conjugation buffer provided in the kit (83 mM sodium phosphate buffer, 0.1 M EDTA, 0.9 M sodium chloride, 0.02% sodium azide, pH 7.2). 0.9 Equiv of the diluted TCEP solution was added to a 1.5 mL micro-centrifuge tube that contained disulfide (G29 S ) 2 (0.6 mg, 3.0 µmol) or (G32 S ) 2 (1 mg, 3.0 µmol), and the mixture was agitated on a shaker for 30 min to furnish thiol G29 SH or G32 SH , respectively. The maleimide-activated BSA (2 mg, 15-25 moles of maleimide/mole BSA) was reconstituted with 200 µL of conjugation buffer to produce a 10 mg/mL solution. The disaccharide solution was added to the reconstituted BSA and agitated at rt for 2-3 h. The conjugation mixture was diluted with HPLC grade water to a volume of 1 mL and desalted using an Amicon Ultra 3K centrifugal filter and was centrifuged for 20 min at 4000× g, rt. The mixture was washed with 1 mL of water 3 times following the same centrifugation procedure. The tube with the filtrate was then removed, and 500 µL of water was added to the NGP29b or NGP32b solution remaining in the filter. Since a small amount of aggregation can occur, the solution/suspension was transferred onto a 2-mL ZebaTM spin desalting column (7K MWCO), provided in the kit, that was previously washed with 1 mL of water 4 times and centrifuged at 1000× g for 2 min at rt. This procedure removed all salts and aggregated protein. The filtrate was lyophilized and can be stored at −50 • C for at least 6 months. In our hands, this combination of filtration and size exclusion chromatography avoids or minimizes aggregation of the NGP. To determine the NGP29b or NGP32b quantity, a solution of 1-2 mg of it in 1-3 mL of ultrapure water was prepared, and the concentration was determined with a Pierce BCA Protein Assay Reagent kit using a spectrophotometer at a detection wavelength of 562 nm. To determine the payload (average number of G29 SH or G32 SH per BSA molecule) comparative MALDI-TOF MS spectra were measured, see Supplementary Material).

Chemiluminescent ELISA
The detailed chemiluminescent ELISA (CL-ELISA) protocol is described in the Supplementary Material.

Conclusions
Using a reversed immunoglycomics approach, a combination of the chemical synthesis of suspected T. cruzi-derived β-Galf -containing glycotopes and their immunological evaluation with sera of CCD patients led to the discovery of two novel BMKs for CD (i.e., NGP29b and NGP32b). Of the two BMKs studied, NGP32b performed particularly well in distinguishing chronic Chagas disease sera from the sera of healthy individuals displaying sensitivity and specificity values of 85.3% and 100%, respectively. These data suggest that the branched tetrasaccharide Galf β1,3Manpα1,2[Galf β1,3]Manpα present in NGP32b is an immunodominant glycotope most likely present in at least one of the GPI anchors of mucins from the mammal-dwelling trypomastigote form of T. cruzi. β-Galactofuranosylation of core GPI anchors could result in common motifs in T. cruzi trypomastigotes, which could be the basis for the development of new diagnostic and potentially prognostic tools for CD.

Supplementary Materials:
The following supporting information can be downloaded online. List of abbreviations; Scheme S1, Synthesis of G29 (3-thiopropyl Galβf 1,3Manα) (overall reaction scheme); Scheme S2, Synthesis of G32 (3-thiopropyl Galβf 1,3Manα1,2-[Galβf 1,3]Manα) (overall reaction scheme); Figure S1, MALDI-TOF mass spectra of NGP29b and NGP32b; detailed description of the syntheses of reaction intermediates S1; S2; S3; S4; S5; description of the CL-ELISA protocol (including titer cutoff value calculation); Appendix with NMR and additional MS spectra of compounds 3; 4; 5; 6; S1; S2; S3; S4; S5; G29 SH /(G29 S ) 2 ; and G32 SH /(G32 S ) 2 .  Institutional Review Board Statement: This study was conducted according to the regulations of International Ethical Guidelines for Biomedical Research Involving Human Subjects, the Good Clinical Practice guidelines, and the Declaration of Helsinki. Panels of positive serum samples used in the study were reviewed and approved by the Institutional Review Board (IRB) of the University of Texas at El Paso, under protocol number 1590350-1 (Biomarkers for Chagas disease, approval date: 28 April 2020), and by the IRB committees at the original institutions, as follows. Individual serum samples from 75 adult patients diagnosed with CCD, were from the Universidad Nacional de Salta, Facultad de Ciencias de la Salud, Salta, Argentina. Moreover, additional serum samples from CCD patients (n = 10) were obtained from the Universidad Central de Venezuela (ICV), Facultad de Medicina, Instituto de Medicina Tropical, Caracas, Venezuela. Negative control serum samples of healthy individuals from a nonendemic area (n = 15) were from the blood bank from the MD Anderson Cancer Center, and kindly donated by Dr. Dapeng Zhou (formerly at MD Anderson Cancer Center, USA, currently at Tongji University, Shanghai, China). All serum samples used in this study were deidentified and coded using a number assigned by the principal investigator at the original institution. At UTEP, the samples received a separate code assigned by the personnel involved in the study. At no time were the PI or the personnel at UTEP able to identify any patients nor were they able to identify the samples. Additionally, the personnel involved in the study performed at UTEP was required to sign a confidentiality agreement.

Informed Consent Statement:
Since this study only used deidentified, stored serum samples, no informed consent was required, as determined by the Exempt IRB Protocol approved by UTEP.
Data Availability Statement: Some of the results described here have been reported in Alba L. Montoya's dissertation, 2020, The University of Texas at El Paso.