A New Strategy for Mapping Epitopes of LACK and PEPCK Proteins of Leishmania amazonensis Specific for Major Histocompatibility Complex Class I

Leishmaniasis represents a complex of diseases with a broad clinical spectrum and epidemiological diversity, considered a major public health problem. Although there is treatment, there are still no vaccines for cutaneous leishmaniasis. Because Leishmania spp. is an intracellular protozoan with several escape mechanisms, a vaccine must provoke cellular and humoral immune responses. Previously, we identified the Leishmania homolog of receptors for activated C kinase (LACK) and phosphoenolpyruvate carboxykinase (PEPCK) proteins as strong immunogens and candidates for the development of a vaccine strategy. The present work focuses on the in silico prediction and characterization of antigenic epitopes that might interact with mice or human major histocompatibility complex class I. After immunogenicity prediction on the Immune Epitope Database (IEDB) and the Database of MHC Ligands and Peptide Motifs (SYFPEITHI), 26 peptides were selected for interaction assays with infected mouse lymphocytes by flow cytometry and ELISpot. This strategy identified nine antigenic peptides (pL1-H2, pPL3-H2, pL10-HLA, pP13-H2, pP14-H2, pP15-H2, pP16-H2, pP17-H2, pP18-H2, pP26-HLA), which are strong candidates for developing a peptide vaccine against leishmaniasis.


Introduction
Leishmaniases represents a complex of diseases with a broad clinical spectrum and epidemiological diversity, considered a major public health problem. Ninety-two countries or territories are considered endemic or have reported cases of cutaneous leishmaniasis, while visceral leishmaniasis is present in 83 countries. Currently, more than 1 billion people living in endemic areas are at risk of infection. It is estimated that 30,000 new cases of visceral leishmaniasis and more than 1 million new cases of cutaneous leishmaniasis occur annually [1]. Although several attempts have been made to obtain vaccines against leishmaniasis, no product is suitable for human use [2]. Furthermore, treatment issues due to high toxicity, low compliance, and cases of resistance, together with a lack of efficiency in control measures, argue that a vaccine would be the most effective, safe, comprehensive, and economically viable way to control this endemic [3].
Immunological findings against Leishmania amazonensis infection in the murine model highlight is a strong tool for proposing immunology profiles and vaccines for cutaneous leishmaniasis [4,5]. In this context, previous studies with subcellular fractionation of L. amazonensis promastigote forms have shown that the microsomal fraction (FMic) associated with immunomodulators provides protection against experimental infection by this parasite [6]. Furthermore, analyses of the different components of the fraction allowed for the identification of Leishmania spp. The homolog of receptors for activated C kinase (LACK) and phosphoenolpyruvate carboxykinase (PEPCK) proteins are responsible for the observed protection [7], suggesting that the production of recombinant forms of peptides of these proteins could lead to the development of a vaccine against cutaneous leishmaniasis.
In recent years, several Leishmania antigens have been analyzed as vaccine candidates with different protective immune responses in various experimental models [8][9][10]. One strategy that can elicit a strong immune response is the identification and use of peptides containing binding epitopes of the major histocompatibility complex (MHC) [11]. It is known that proteins have several immunodominant epitopes that can induce immunity through presentation via MHC proteins present on the surface of antigen-presenting cells (APCs) [12]. Thus, the identification of these epitopes can be efficient in the development of epitope-based vaccines. Several studies have also revealed that the strategy of epitopebased vaccines, chosen for their binding to human leukocyte antigen, appears to induce more potent responses than whole-antigen vaccines [13].
Using peptides containing immunogenic parts of a protein capable of inducing a specific T-cell response may become a promising strategy in the prophylaxis of leishmaniasis [13][14][15]. In addition, the development and use of bioinformatics tools have made it easier to identify potential T-cell epitopes restricted to human leukocyte antigens (HLA) and murine (H2) immunogenic vaccines. Peptide-based vaccines have advantages, including the absence of infectious materials, specificity, stability, and large-scale production at low cost, and these peptide vaccines have already been successfully tested against several diseases [16][17][18][19][20]. In addition, potential immunogenic peptides have already been identified within proteins and described as vaccine candidates. For example, Leishmania histone H2B, Promastigote Surface Antigen (PSA), and Leishmania major L. form-large RAB GTPase (LmlRAB) induced a predominant Th1 response in individuals immune to L. major or L. infantum [21][22][23].
Therefore, possible immunogenic peptides can be identified within proteins previously described as vaccine candidates. This work identifies potential antigenic peptides from LACK and PEPCK L. amazonensis proteins from a workflow that includes in silico and in vitro approaches. In addition, CD8+ T-lymphocyte epitopes (n = 3 and n = 6, respectively), capable of binding to MHC class I proteins, were identified and designed to build antigenic peptides as vaccine targets. The set of results presented here on antigenic epitope mapping of these proteins contributes to improving the rational development of vaccines based on specific regions of these proteins that can trigger protective immune responses.

Promiscuity Analysis of Human and Murine MHC I Alleles
The distribution of the HLA allele varies among populations, but vaccinal peptides need to be immunogenic for most individuals. The Brazilian population is highly miscegenated, but migration history is different for each region. Therefore, before starting antigenic prediction, the Allele Frequency Net Database was searched for the most frequent allele in each Brazilian state. As a result, HLA-A*0201 was chosen for its high frequency compared to other types of HLA throughout the Brazilian territory. In parallel, H2Db was the allele chosen for predicting immunogenic epitopes because it can be found in C57BL/6J mice [24], a representative strain in various life science fields. In addition, it is frequently used for transgenesis, being permissive for the expression of most mutations, and it was the first strain for which the mouse genome was sequenced [25].

Prediction of LACK and PEPCK Epitopes
Leishmania major immunogenic peptides already deposited in the Immune Epitope Database (IEDB) were used as positive controls to test the prediction parameters. Initially, the amino acid sequences were aligned between the LACK of L. amazonensis and L. major to find the similarity between sequences already known in the database with our predicted sequences (Supplementary Figure S1). Then, the proteins were subjected to two prediction strategies: (i) IEDB server-binding of 11 amino acid peptides with MHC-I in six algorithms without proteasomal cleavage, as the server does not allow it and (ii) Database of MHC Ligands and Peptide Motifs (SYFPEITHI) server-binding of 10 amino acid peptides with MHC-I with proteasomal cleavage. The use of servers with and without proteasomal cleavage increased the analysis range and provided greater reliability in the prediction and evaluation of the servers' applicability.
Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW vaccines based on specific regions of these proteins that can trigger protective i responses.

Promiscuity Analysis of Human and Murine MHC I Alleles
The distribution of the HLA allele varies among populations, but vaccinal p need to be immunogenic for most individuals. The Brazilian population is high cegenated, but migration history is different for each region. Therefore, before antigenic prediction, the Allele Frequency Net Database was searched for the m quent allele in each Brazilian state. As a result, HLA-A*0201 was chosen for its h quency compared to other types of HLA throughout the Brazilian territory. In p H2Db was the allele chosen for predicting immunogenic epitopes because it can b in C57BL/6J mice [24], a representative strain in various life science fields. In add is frequently used for transgenesis, being permissive for the expression of mos tions, and it was the first strain for which the mouse genome was sequenced [25].

Prediction of LACK and PEPCK Epitopes
Leishmania major immunogenic peptides already deposited in the Immune Database (IEDB) were used as positive controls to test the prediction paramet tially, the amino acid sequences were aligned between the LACK of L. amazonensi major to find the similarity between sequences already known in the database w predicted sequences (Supplementary Figure S1). Then, the proteins were subjected prediction strategies: (i) IEDB server-binding of 11 amino acid peptides with M six algorithms without proteasomal cleavage, as the server does not allow it and tabase of MHC Ligands and Peptide Motifs (SYFPEITHI) server-binding of 10 acid peptides with MHC-I with proteasomal cleavage. The use of servers w without proteasomal cleavage increased the analysis range and provided great bility in the prediction and evaluation of the servers' applicability.
To filter potential epitopes, we selected peptides with a score ≥50% in at least 4 ferent algorithms on the IEDB server and a score ≥2 on the SYFPEITHI server. From th parameters, the number of epitopes of LACK [H2Db (IEDB-97.0% and SYFP THI-17.8%) and HLA-A*0201 (IEDB-93.0% and SYFPEITHI-14%)] and PEP [H2Db (IEDB-95.0% and SYFPEITHI-17.0%) and HLA-A*0201 (IEDB-93.7% SYFPEITHI-10.5%)] were reduced ( Figure 2). To identify peptides with a conserved motif in the two servers, the total predi peptides of LACK (H2Db and HLAA*201) and PEPCK (H2Db and HLAA*201) w submitted separately to the MEME Suite server. First, the residues that appeared most in the set of peptides were chosen. Then, peptides that presented conserved am acid residues in the logo plot and E-value <= 2 were selectedThe selected peptides lowed the criteria of conserved amino acid residues in the logo plot and presented best E-values. LACK H2 epitopes were the only ones giving conserved motifs (LEHP and GAKPSECI). Therefore, LACK HLA PEPCK epitopes for H2 (ENVEWG ELVQWA) and HLA (GGCYAK and VLSYAK) were selected by E-value (Supplemen Figure S2 and Table S1). Then, the motifs were aligned in the multialign tool to iden anchor amino acid residues in the peptides (Supplementary Figure S3). Finally, 26 p tides were selected (Table 1).  Only peptides that reached scores ≥ 50% in at least 4 of the 6 IEDB servers and ≥2 for the SYFPEITH server were selected for the next phase.
To identify peptides with a conserved motif in the two servers, the total predicted peptides of LACK (H2Db and HLAA*201) and PEPCK (H2Db and HLAA*201) were submitted separately to the MEME Suite server. First, the residues that appeared the most in the set of peptides were chosen. Then, peptides that presented conserved amino acid residues in the logo plot and E-value ≤ 2 were selectedThe selected peptides followed the criteria of conserved amino acid residues in the logo plot and presented the best Evalues. LACK H2 epitopes were the only ones giving conserved motifs (LEHPIVV and GAKPSECI). Therefore, LACK HLA PEPCK epitopes for H2 (ENVEWG and ELVQWA) and HLA (GGCYAK and VLSYAK) were selected by E-value (Supplementary Figure S2 and Table S1). Then, the motifs were aligned in the multialign tool to identify anchor amino acid residues in the peptides (Supplementary Figure S3). Finally, 26 peptides were selected (Table 1).

Prediction of Physicochemical Parameters
The ProtParam server showed physicochemical results of the epitopes alone, where the molecular weight, theoretical protrusion index (P1), estimated in vitro half-life in human mammalian reticulocytes, and the instability index was calculated, revealing that all peptides can be stable for vaccine targets. In addition, the aliphatic calculated index indicated a thermostable nature at different temperatures and the general average of hydropathicity (GRAVY) ( Table 2).

Validation of Predicted Epitopes in Spleen Cells from Infected Mice by ELISpot
The 26 selected sequences were produced as synthetic peptides to confirm the immunogenicity of the predicted epitopes. After synthesis, the peptides were purified in HPLC coupled to mass spectrometry (Supplementary Figure S4). Of the 26 peptides, the peptide pP25-HLA was lost during purification and was not tested.
Specific T-cell responses, measured by the production of interferon-γ (IFN-γ) after interaction with spleen lymphocytes from L. amazonensis, were observed in infected and the control mice. The immunogenicity of the 25 synthetic peptides was initially validated by groups according to the protein of origin and MHC haplotype used on the prediction. Splenocytes derived from infected mice were collected 83 and 98 days after infection and were stimulated ex vivo using 4 pools of peptides: LACK murine (LACK-H2Db) and human (LACK-HLA-A*0201) haplotypes, and PEPCK murine (PEPCK-H2Db) and human (PEPCK-HLA-A*0201) haplotypes. PEPCK-H2Db and LACK-HLA-A*0201 peptides induced a specific IFN-γ response, especially 98 days after infection (Figure 3). No significant numbers of IFN-γ secreting cells were observed. All cells stimulated with ConA had high numbers of IFN-γ secreting cells, while control, non-stimulated cells presented no significant numbers of IFN-γ secreting cells.

Quantification of the T Lymphocyte Subpopulations Reactive to the Predicted Epitopes
The immunogenicity of the 25 synthetic peptides was individually validated by analyzing their ability to bind to lymph node cells from L. amazonensis infected mice. DimerX/peptide complexes were used in ex vivo assays to assess the percentage of CD3 T lymphocytes that would specifically bind to these supramolecular complexes, which mimic interactions that occur during antigen presentation. Cell/complex binding was measured by flow cytometry analysis. In tests with cells and DimerX-H2Db/peptide complexes, double staining was interpreted as clones of CD3 T lymphocytes, specifically reactive to the complexes and reactive CD8+ lymphocytes, capable of binding to DimerX-H2Db/peptide complexes (Supplementary Figure S5). Of the 25 reactive peptides to the DimerX/PEPTIDES complex, 8 gave a specific MHC class I binding response (Figure 4).

Quantification of the T Lymphocyte Subpopulations Reactive to the Predicted Epitopes
The immunogenicity of the 25 synthetic peptides was individually validated by analyzing their ability to bind to lymph node cells from L. amazonensis infected mice. DimerX/peptide complexes were used in ex vivo assays to assess the percentage of CD3 T lymphocytes that would specifically bind to these supramolecular complexes, which mimic interactions that occur during antigen presentation. Cell/complex binding was measured by flow cytometry analysis. In tests with cells and DimerX-H2Db/peptide complexes, double staining was interpreted as clones of CD3 T lymphocytes, specifically reactive to the complexes and reactive CD8+ lymphocytes, capable of binding to DimerX-H2Db/peptide complexes (Supplementary Figure S5

Discussion
Knowledge of the epitopes responsible for T-lymphocyte-mediated immunity is essential for vaccine studies in Leishmania spp. However, to date, most leishmaniasis vac-

Discussion
Knowledge of the epitopes responsible for T-lymphocyte-mediated immunity is essential for vaccine studies in Leishmania spp. However, to date, most leishmaniasis vaccines have failed due to poor antigen response to selected proteins/peptides, lack of a strong and persistent cellular response, and lack of standardized guidelines for experimental assays [26,27]. To contribute to overcoming these limitations, this work advances into the predictive dissection of T-lymphocyte epitopes present in LACK and PEPCK virulence factors from L. amazonensis, pointing to improve the selection of antigenic determinants from both proteins as sources of potent epitopes.
One of the great challenges for developing vaccines against cutaneous leishmaniasis is the differences found among Leishmania species, even when they cause similar forms of the disease [9,28]. For instance, although lipophosphoglycan (LPG) is considered a virulence factor for L. major, the same does not seem to be the case for L. amazonensis [28]. Similarly, protection elicited with a Th1 response during L. major infection has already been established, whereas, in L. amazonensis, the disease can persist in the presence of Th1 response [29]. Therefore, different virulence factors and immune responses induced by other species must be considered when developing a vaccine. In parallel, the use of the whole parasite has been seen as safe and effective, but there is still a challenge in standardizing doses. Therefore, in conjunction with in vitro/ex vivo validation, an immunome-derived vaccine approach may accelerate the development of candidate vaccines for leishmaniasis [28,[30][31][32][33].
Immunomics tools and databases have favored coping with this barrier as an accurate strategy in epitope mapping in Leishmania spp., as applied here [34]. Although some reports of epitope mapping have been the subject of current studies in reverse vaccinology in Leishmania spp. [35], there is still a need to establish a better strategy for promising mapping epitopes from Leishmania spp. proteins. Controversies about the activation pathway of CD8+ T-lymphocyte, concerning the presentation of Leishmania antigens through MHC class I [36,37], and the absence of consensus on the most appropriate strategy to predict true positive epitopes [38][39][40] must be overcome. In this context, the experimental design proposed here started forecasting with two quantitative matrix methods widely used in the literature: the Immune Epitope Database-IEDB, using different prediction methodologies within the platform itself, including Artificial Neural Network-ANN, Stabilized Matrix Method-SMM, SMMPMBEC, Pickpocket, CONSENSUS, and NetMHCpan; and the SYFPEITHI server. However, as these methods do not differentiate between ligands and non-ligands, the immunogenicity of these peptides must be carefully evaluated through quantitative methods. Score evaluation and haplotype selection are fundamental for choosing possible immunogenic epitopes.
Analysis of the protein sequences yielded many possible epitopes from both proteins, but only a few were selected from all the algorithms used, chosen based upon the efficiency binding to more than one supertype or allele. Peptides that had high and reliable scores were further selected. In addition, the motifs of these peptides and their proteasomal cleavage were observed. Based on these selections, and the MHC class I-restricted epitopes prediction that is considered reliable, we designed 26 synthetic peptides from the two L. amazonensis proteins. These peptides were then submitted to an experimental approach to validate their ability to bind and activate Leishmania-infected lymphocytes from MHC-compatible C57BL/6J mice, further confirmed by molecular docking. This approach narrowed our numbers to 9 selected peptides, capable of inducing IFN-γ production and binding to CD8+ lymphocytes through MHC class I. Developing a high throughput method for screening peptides for MHC binding to determine CD8+ T-lymphocyte responses, such as IFN-γ ELISpot and tetramer/peptide binding technologies, is a proof of concept for prediction validation.
LACK protein has long been identified as an important virulence factor of Leishmania spp. and a potential immunogen and drug target against the parasite [41]. Its sequence is highly conserved among Leishmania species [42], which encourages its exploitation for vaccine development. The data presented here corroborates the antigenic properties of LACK peptides previously proposed for other studies that evaluated the immune response of LACK epitopes as vaccine targets against L. major. Salehi-Sangani et al. showed that a chimeric protein that included the most immunogenic epitopes from the genes of several proteins, including LACK, induced a high level of IFN-γ, partially protecting mice against L. major [11]. Another study identified LACK peptides that are effectively presented via MHC class II molecules from mice of susceptible and resistant backgrounds [43]. Our results indicate that LACK peptides pL1H2, pL3H2, and pL10HLA are highly antigenic for L. amazonensis, all comprising sequences similar to those previously identified for L. major [11,43].
PEPCK is a key player in the gluconeogenesis pathway in Leishmania spp. [44]. Its homology among pathogenic species of Leishmania reaches more than 90%, whereas it has less than 17% of homology to mice and humans [45], making it an excellent source of vaccine immunogens. Previous results show that recombinant PEPCK can protect mice from L. major infection [45]. Similarly, a synthetic DNA vaccine encoding the PEPCK gene was able to elicit a protective response against L. major, generating specific T memory cells [46]. To our knowledge, no epitope screening protocol was previously applied to that protein, although a peptide tetramer (PEPCK 335-351 ) was shown to bind to CD4 + T cells from L. major-infected mice. The present study identified pP13-H2, pP14-H2, pP15-H2, pP17-H2, pP18-H2, and pP26-HLA as strongly antigenic and capable of binding specific CD8+ T lymphocytes and inducing IFN-γ production.
There has been an understanding that the activation of CD8+ cells and IFN-γ production are important for protection against intracellular pathogens [47]. Studies show that CD8+ T-lymphocytes contribute to the destruction of Leishmania-infected cells by activating macrophages for oxidative explosion via cytokines produced upon antigen stimulation [48,49]. Our work shows that LACK and PEPCK peptides induce a CD8+ T-lymphocyte response with either human or murine MHC. Similarly, peptides obtained from the cysteine proteinase B from L. amazonensis could recognize the T cell receptor on the surface of CD8+ lymphocytes, indicating the presence of specific cells on the microenvironment of the draining lymph nodes of the lesion [50].
Studies have identified T cell epitopes using infected macrophages as APCs. They have revealed the existence of potential T lymphocyte epitopes restricted to HLA classes I and II in the amino-terminal region that could stimulate specific cellular immune responses in volunteers infected with L. donovani or L. panamensis [14]. Not only external or secreted Leishmania antigens can be presented in the context of MHC class I proteins, but intracellular proteins can be as well [26,51]. Agallou and collaborators showed that peptides from L. infantum induced a T-cell response mainly characterized by priming CD8+ T lymphocytes and IFN-γ production in immunized mice [52]. Our work shows that the T-cell receptor (TCR) recognized the processed peptides bound to MHC class I, and CD8+ T lymphocytes were identified by binding the TCR-MHC complex through its interaction with non-polymorphic regions of the MHC H2Db/peptide. Despite the mechanism by which the MHC complex binds to peptides not being fully understood, the use of soluble MHC-peptide complexes and tetramers allows epitope mapping and detection of antigen-specific T-lymphocytes from L. amazonensis proteins [50,53]. The data gathered here present candidate peptides for vaccine development and advances new and powerful examples of tetramer use for L. amazonensis epitope mapping. Furthermore, a better understanding of the mechanism underlying the complex interaction of the MHC proteins, with peptides assessed, can drive us to future responses regarding the cellular and humoral immunity of L. amazonensis peptides.

Human Population Coverage and Murine MHC Class I Allele
The distribution of HLA alleles among the endemic population is essential for effective vaccine development. Therefore, the human population coverage analysis tool Allele Frequency Net Database (http://www.allelefrequencies.net (accessed on 1 July 2020)) [54] was used to identify the highest frequent Human Leukocyte Antigen (HLA) allele in the Brazilian population. The search was performed for each Brazilian state. Murine MHC class I binding allele (H2Db-Beta-2-microglobulin) was selected from the chosen strain, C57BL/6J-Mus musculus, in Mouse Haplotype Table [55].

Recovery of Peptide Sequences
LACK (LAMA_000011500) and PEPCK (LAMA_000546600) complete aminoacid sequences from Leishmania amazonensis [56] were selected and retrieved from TriTrypDB-Kinetoplastid Informatics Resources (https://tritrypdb.org (accessed on 1 November 2019)) in FASTA format. Both sequences were inserted in The National Center for Biotechnology Information-NCBI (ncbi.nlm.nih.gov (accessed on 1 August 2020)). Immune Epitope Database and Analysis Resource-IEDB (iedb.org (accessed on 1 August 2020)) was used to verify the peptide sequences. As there are already studies with the prediction of LACK protein epitopes identified and experimentally tested, we used the sequences of these peptides as a control for our starting point [43].

Prediction of T Lymphocyte Epitopes
The sequences of L. amazonensis LACK and PEPCK were submitted to the six servers of the Immune Epitope Database-IEDB platform (Artificial Neural Network-ANN, Stabilized Matrix Method-SMM, SMMPMBEC, Pickpocket, CONSENSUS, NetMHCpan)-(iedb.org (accessed on 1 November 2020)) [57] and the server SYFPEITHI-(syfpeithi.de (accessed on 1 February 2021)) [58] for prediction of T cell peptides (MHC class I). These servers have a database for MHC ligands and peptide motifs in humans and other animal species. In addition, the SYFPEITHI server also predicts proteasome-cleaved peptides, although their content is restricted to published data only. After prediction, peptides with scores ≥50 for at least 4 of the 6 IEDB servers, or ≥2 for the SYFPEITHI server, were selected. Next, those were submitted to the MEME Suite software (https://meme-suite. org/meme (accessed on 1 May 2021)) [59] for analysis of the most predominant amino acids, followed by alignment in the Multalin software (multalin.toulouse.inra.fr (accessed on 1 May 2021)) [60] to observe which peptides had added residues on the N and C terminal chains, and to select possible immunogenic epitopes.

Tap Transport/Proteasomal Cleavage
To predict antigen processing through the class I MHC presentation pathway, we used the NetCTL 1.2 server combined with the Tap transport/proteasomal cleavage tools (https: //services.healthtech.dtu.dk/service.php?NetCTL-1.2 (accessed on 1 January 2022)). The method integrates the prediction of class I MHC peptide binding, C-terminal proteasomal cleavage, and TAP transport efficiency. The C-terminal cleavage weight was set to 0.15, the Tap transport efficiency was set to 0.05, and the epitope ID was set to 0.75 (default values) [61].

Synthesis of Peptides and Purification
The peptides were synthesized by the F-moc strategy in a synthesizer machine (Mul-tipep1; Intavis Bioanalytical Instruments, Köelh, Germany) using the Tentagel ® resin as described elsewhere [64]. Peptides were purified by an HPLC coupled to mass spectrometry. The chromatographic column used was the XBridge BEH C18 from Waters, with a particle size of 2.7 µM and dimensions of 5 cm × 4.6 mm. The mobile phase was comprised of the following eluents: 0.05% formic acid in H 2 O (18 MΩ × cm) and 0.05% formic acid in acetonitrile. The equipment used was a Waters Autopurification System with a Waters 2545 binary gradient pump, a Waters 2998 diode array detector, and a Waters SQ Detector 2 electrospray ionization source mass detector (IBMP 002063). The diode array detector performed monitoring from 200-300 nm, and the mass detector served monitoring from m/z 450-1450 in positive mode. The injection volume used was 20 µL of a sample with an estimated 1 mg/mL concentration. The chromatography column was kept at room temperature.

Mice and Infections
C57BL/6J male mice, aged 4 to 5 weeks, were obtained from ICTB-FIOCRUZ and brought to experimental mouse facilities, where they were acclimatized for at least one week before being used in the experiments. Upon arrival, mice were randomly separated into cages with 4 animals each and maintained in microisolators at random columns and rows. Animals were fed with standard pellet food and water ad libitum and maintained with enrichment items, which were changed every week. All experiments were carried out with the approval of the Ethics Committee for the Use of Laboratory Animals of Instituto Oswaldo Cruz (CEUA-IOC) under number L-014/2019. Mice were infected with 1 × 10 5 L. amazonensis amastigotes in the left, hind footpad. After infection, mice were observed weekly for signs of illness (lethargy, ruffled fur, weight loss, footpad swelling).

Enzyme-Linked Immunospot Assay (ELISpot)
The ELISpot assay was performed with spleen cells obtained from L. amazonensisinfected mice at 83 and 98 days after infection; five mice were each assayed in duplicates. Three non-infected mice were used each time, but the results are shown together as 0 days after infection. Cells (250,000 cells per well) were stimulated with one of 4 different pools (pool 1-H2Db LACK with 3 peptides, pool 2-H2Db PEPCK with 7 peptides, pool 3-HLA LACK with 8 peptides, pool 4-HLA PEPCK with 8 peptides) of peptides (20 µg/peptide) and a RPMI medium with 10% FBS (negative control) or concanavalin-A (Sigma Aldrich, St. Louis, MO, USA) (4 µg/mL) (positive control). The reaction was carried out using the ELISpot Plus: Mouse IFN-γ (ALP)-MABTECH kit, according to the manufacturer's recommendations. Spots generated after stimulation were counted with Immunospot reader S6UV ultra (Cellular Technology Ltd., Cleveland, OH, USA). The number of adjusted IFN-γ, secreting cells per million spleen cells, was expressed as the mean number of spots induced by antigen subtracted by the number of spots caused in non-stimulated wells.

Flow Cytometric Detection of T Cells Binding Complexed with Recombinant Dimers of MHC Class I (H-2Db) Molecule
To select the peptides with higher affinity to bind T lymphocytes, each one was coupled to MHC class I and then put in contact with lymphocytes from L. amazonensis-infected mice. The dimeric MHC class I (H-2Db) was a fusion protein between mouse H-2Db and mouse IgG1, and the complex was performed using a DimerX I: Recombinant Soluble Dimeric Mouse H-2D[b]: Ig kit, according to the manufacturer's protocol (BD Biosciences, San Jose, CA, USA). Next, 10 µg of each peptide was complexed with the MHC molecule by overnight incubation. On the next day, approximately 3.3 × 10 5 mouse lymph node cells/tube were resuspended in 1% FCS-PBS buffer and incubated for 10 min at 4 • C with Mouse BD Fc Block™ (purified CD16/CD32). Then, DimerX/H-2D (1 µg/mL) loaded with one of the 25 desired peptides (0.9 µg/mL) was added to each tube and incubated for 50 min at 4 • C. After washing, PE-conjugated mAb A85-1 (anti-mouse IgG1) and PE-Cy5 conjugated anti-mouse CD3e were added to each tube and incubated for 30 min at 4 • C. Controls with no antibody and no control isotype, with each antibody separate, were also prepared. Ten thousand events were acquired using a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, USA). The following parameters were considered: forward scatter to evaluate the cellular size, side scatter to evaluate cellular complexity, and analysis of cell marker expression with fluorescence analysis. PE (BD Biosciences) and PE-Cy-5 (BD Biosciences) fluorescence were acquired through 585/42 and 690/50 BP, respectively, of the 488 nm-blue laser. Data analysis was performed using CytoExpert software v2.3 (Beckman Coulter, Brea, CA, USA), and Kaluza v.2.1 (Beckman Coulter, Brea, CA, USA) was used to generate the histograms and dot plots (Supplementary Materials).

Conclusions
In silico analysis and prediction techniques of synthetic peptide sequences that are highly conserved and promiscuously bound to murine or human MHC class I molecules make them candidate vaccines against leishmaniasis. These findings provide complementary data on mapping epitopes to L. amazonensis LACK and PEPCK proteins. Data gathered here demonstrate that combining immunoinformatics approaches with experimental validation yields the identification of nine antigenic peptides: pL1-H2, pL3-H2, pL10-HLA, pP13-H2, pP14-H2, pP15-H2, pP17-H2, pP18-H2, and pP26-HLA. Based on these results, future investigations will be carried out to verify the ability of the peptides to induce protection in murine models infected with L. amazonensis.