Enolase AmEno15, a Promising Candidate for Understanding the Infectious Process of Anaplasma marginale

1. Introduction

The pathogen Anaplasma marginale, the causative agent of bovine anaplasmosis, is a Gram-negative bacterium that infects red blood cells. In cattle, anaplasmosis is characterized by anemia, jaundice, abortions, loss of milk and meat production, and even death [1,2,3]. This disease is primarily transmitted by ticks of the genera Rhipicephalus and Dermacentor, which act as biological vectors, as well as by veterinary surgical materials and blood-sucking flies, considered as mechanical vectors [4]. Bovine anaplasmosis's most significant impact is on the production of dairy and meat derivatives. Therefore, addressing economic losses becomes a priority, not only in the world's tropical regions where vectors are abundant, but also in cold regions to which ticks have migrated due to climate change [5,6]. To date, efforts to develop an effective vaccine for bovine anaplasmosis have been unsuccessful. Thus, the search for new candidates demands powerful and novel bioinformatics strategies in vaccine design based on different proteins from those typically analyzed (MSPs, OMPs, T4SSs) [7,8]. In this regard, recent genomic approaches of A. marginale Mexican strains have proposed moonlighting proteins (MLPs) as potential vaccine candidates against anaplasmosis, due to their role in invasion, adhesion, and virulence, reported in other pathogens [9,10,11]. The MLPs' functional versatility represents an advantage for many pathogens, allowing them to perform multiple functions despite having a reduced genome, as in A. marginale [12,13]. The main characteristic of MLPs is the ability to perform a canonical function in the primary metabolism and a different secondary function in other cellular locations [14]. Due to the above, several MLPs have been proposed as immunogenic targets against bacteria and parasites, like enolase, a protein participating primarily in glycolysis, and in many pathogens has been considered as a virulence factor. Generally, the second function of enolase is a strategy to initiate host tissue invasion through adhesion to host cells, and this mechanism is associated with intracellular pathogens [15,16,17,18,19]. Recently, the identification of enolase AmEno15 in A. marginale has been reported; however, its role in anaplasmosis pathogenesis has not been elucidated [20]. Nevertheless, the enolase role in other pathogenic bacteria is well known,

as the enolase of Mycoplasma bovis and Mycoplasma hyopneumoniae, which binds to plasminogen, activates it into plasmin, and favors invasion of embryonic bovine lung cells and swine tracheal epithelial cells, respectively [21,22]. Additionally, in Mycoplasma suis, a pathogen of porcine red blood cells, recombinant enolase adheres to host tissue and induces an immune response, conferring partial resistance to mycoplasmosis [23,24]. On the other hand, plasminogen is not the only protein to which enolase can bind; it can also bind to proteins of the extracellular matrix, including fibronectin [10], vitronectin [25], laminin [26,27], and collagen [28]. It is worth noting that the process by which A. marginale adheres to and infects bovine erythrocytes remains unknown. Interestingly, molecular docking predictions suggest that AmEno15 might bind to erythrocyte membrane proteins as well as extracellular matrix proteins [29,30]. This fact is relevant for understanding the mechanism related to A. marginale internalization into its target cell. Thus, in this work, we employ a deep immunoinformatics approach to explore the possibility of AmEno15 binding to erythrocyte membrane proteins (spectrin and stomatin), extracellular matrix (ECM) proteins associated with the tick gut tissue (fibronectin), and plasminogen. Additionally, we demonstrate through microplate binding assays that AmEno15 effectively binds to these proteins. These results contribute to understanding the relationship pathogen-vector-host mediated by AmEno15, and its possible role in the adhesion and invasion of A. marginale to erythrocytes, as well as the molecular interaction with tick cells, a key step in the pathogen transmission.

2. Materials and Methods

2.1. Sequence Retrieval

The amino acid sequence of A. marginale enolase was obtained from the NCBI database by browsing the strain MEX-15-099-01 genome annotation (GenBank GCA_008690255.1), and downloading the protein sequence (KAA8472002.1) annotated as phosphopyruvate hydratase. The sequence protein was called AmEno15 and its molecular size was predicted in ProtParam (https://web.expasy.org/protparam/).

2.3. Tridimensional (3D) Modeling

AmEno15 was modeled using SwissModel, a bioinformatics tool that uses protein structure homology to build 3D protein structure models [24]. In this server, models from the AlphaFold DB are appended to the available structures/models [25]. All generated models in SwissModel using AlphaFold DB per-residue scores are transferred to values of pLDDT (predicted local distance difference test) from the underlying template to provide a global model evaluation, where pLDDT is a per-residue measure of local confidence scaled from 0 to 100, with higher scores indicating higher confidence and a more accurate prediction.

2.4. Molecular Docking and Molecular Dynamics Simulation

The docking of the modeled AmEno15 with four protein ligands was performed in the ClusPro server to analyze their binding affinity [31]. The proteins used were plasminogen (PDB: 4DUR), fibronectin (PDB: 3M7P), spectrin (PDB: 3LBX), and stomatin (PDB: 7WH3). For each molecular docking, ten models were generated in ClusPro and downloaded in PDB format, then visualized in UCSF ChimeraX [32]. The model with the highest score was selected.

Molecular dynamics simulation was performed in the iMODS server (https://imods.iqf.csic.es/) to evaluate the docking complex stability (AmEno15 and each ligand) by considering deformability, the B-factor, covariance map, variances, elastic network model, and eigenvalues. These parameters analyze the docked complexes' motions, as they are usually dynamic and entail significant and collective conformational changes [33].

2.5. AmEno15 Recombinant Expression

AmEno15 was subcloned into plasmid pET32a(+), and the expression was carried out in Escherichia coli strain Rosetta (DE3). This plasmid contains a thioredoxin tag (Trx) and a 6-histidine tag (6His), both of which are in frame with the AmEno15 sequence. Cultures transformed with recombinant plasmids were grown in Luria-Bertani medium containing 100 mg/mL ampicillin, and induced with 0.8 mM isopropyl thio-β-D-galactoside (IPTG, Sigma Aldrich). Before induction, cultures were grown at 37°C to an A₆₀₀ of 0.6–0.8, and two conditions of time (4 and 16 h) and temperature (16°C and 37°C) were evaluated. After selecting the induction condition at 37°C for 4 h. Then, the cells were pelleted by centrifugation at 6000 × g for 30 min and stored at -80°C. The cells were lysed by incubation in 50 mM sodium phosphate (10 mL per g wet weight), pH 8.0, containing 300 mM NaCl, 1 mg/mL lysozyme and 1 mM phenylmethanesulfonyl fluoride (PMSF) for 30 min on ice and sonicated for six cycles, 15 s each with 15 s cooling between successive bursts at 5 output in a Branson sonifier 450. The lysate was centrifuged at 20000 × g for 30 min in a Beckman Ultracentrifuge (model LE-80K, 70 Ti rotor). The recombinant AmEno15 was purified from the soluble fraction with a His6-tagged r-Pfen using Ni-NTA (Thermo Scientific) acid affinity chromatography protocol.

2.6. Microplate Binding Assays to Erythrocyte Membrane Proteins, Plasminogen, and Fibronectin

Microplate assays analyzed the interaction of AmEno15 with erythrocyte membrane proteins (stomatin and spectrin), plasminogen, and fibronectin. Briefly, a 96-well ELISA plate (Costar, Fisher Scientific) was coated with 100 μL of 3 μg/mL spectrin (S3644, Sigma-Aldrich, Uniprot P02549), stomatin (ab132752, Abcam, Uniprot P27105), plasminogen (SRP6518, Sigma-Aldrich, Uniprot P00747), or fibronectin (F2006, Sigma-Aldrich, Uniprot P02751) and incubated overnight at 4°C. The plates were washed once with 100 μL of PBST (0.05% Tween-20) (P4417 and P9416, Sigma Aldrich), followed by a blocking with 100 μL of 5% BSA-PBST 0.05% for 2 h at 37°C. After blocking, one wash was performed with 100 μL of PBST 0.05% and 100 μL of recombinant AmEno15 was added at different concentrations (ranging from 10 to 0.039 μg/mL), and incubated for 2 h at 37°C. After incubation, the plate was washed with 100 μL of PBST 0.05%. Then, 100 μL of Trx tag antibody, mAb, mouse (dilution 1:1000, Cat. A00180-40, GeneScript) was added and incubated at 37°C for 1 h. Another wash with PBST 0.05% was performed before incubation with 100 μL of HRP-conjugated goat anti-mouse IgG (H+L) (dilution 1:10,000, Cat. 31430, Thermo Scientific) at 37°C for 1 h. After washing with PBST 0.05%, 100 μL of the substrate 1-Step™ Ultra TMB-ELISA Substrate (Thermo Scientific) was added and incubated for 15 min at 37°C. The reaction was then stopped with 100 μL of 2 M sulfuric acid (Baker), and the absorbance was measured at 450 nm (Multiskan FC, Thermo Scientific). Each binding assay comprised three independent experiments with three replicates each. In the binding assays to spectrin and fibronectin, the AmEno15 concentrations used were 10 μg/mL, 5 μg/mL, 2.5 μg/mL, 1.25 μg/mL, 0.625 μg/mL, and 0.312 μg/mL, whereas, with stomatin and plasminogen, the AmEno15 concentrations were 1.25 μg/mL, 0.625 μg/mL, 0.312 μg/mL, 0.156 μg/mL, 0.078 μg/mL, and 0.039 μg/mL. The specificity control for primary and secondary antibodies were the ligand-coated wells assayed with BSA at different concentrations (ranging from 10 μg/mL to 0.039 μg/mL), and carbonate buffer-coated wells (only AmEno15 at 10 μg/mL to 0.039 μg/mL) were used as controls for AmEno15 specificity.

2.6. Statistical Analysis

The statistical significance of the data was determined using one-way Analysis of Variance (ANOVA) with a post hoc Duncan's Multiple Range Test (DMRT) in STATISTICA software version 7.0 [34]. P values of < 0.05 were considered statistically significant. All measurements were conducted in triplicate with three internal repetitions. Plots were created in RStudio [35].

3. Results

3.1.3D AmEno15 Modeling

A 3D model of AmEno15 monomer was generated using the template of the enolase from A. marginale AlphaFold DB (Q5PAS6.1.A), with an average pLDDT of 95.40%, 100% sequence identity, and 80% coverage. A Ramachandran plot showed that 95.10% of the amino acids fell within the favored regions; hence, the AmEno15 3D model could be considered a structure with good resolution, since the scores of over 90% correspond to protein structures with a resolution of 2-3 Å (Figure 1). The AmEno15 3D model assessment yielded a MolProbity score of 1.24, which reflects its high quality. This score is calculated by combining clashscore (bad steric overlap), Ramachandran, and rotamers, and is normalized to be on the same scale as X-ray resolution.

3.2. AmEno15 Recombinant Expression

The recombinant expression of AmEno15-Trx-6His resulted in a protein with a molecular weight of 64 kDa (Figure 2), which is consistent with ProtParam that predicted a molecular weight of 64.57 kDa, corresponding to 587 amino acids.

3.2. Molecular Docking and Molecular Dynamics Simulation

Four protein–protein dockings were performed using AmEno15 as a receptor, and the proteins spectrin (3LBX), stomatin (7WH3), fibronectin (3M7P), and plasminogen (4DUR), as ligands. The ten generated docking models were retrieved as "models with balanced coefficients since we do not know what forces dominate the complex protein–protein," as recommended in the ClusPro manual. The best docking models predicted correspond to the lowest energy model, which were -928.7 kcal/mol, -880.6 kcal/mol, -1133.4 kcal/mol, and -1046.2 kcal/mol, for AmEno15-3LBX, AmEno15-7WH3, AmEno15-3M7P, and AmEno15-4DUR, respectively. Additionally, all models were evaluated using Ramachandran plots to validate the docking quality in PDBsum. According to this, the obtained values for the most favored regions were 90.6% (AmEno15-3LBX), 87.3% (AmEno15-7WH3), 84.6% (AmEno01-3M7P), and 76.8% (AmEno15-4DUR) (Figure 3, Figure 4, Figure 5 and Figure 6, panels A).

Additionally, molecular dynamics simulations were performed using the iMODs server. The green graph of all dockings showed areas of hinges indicating the deformability degree measured between 0 and 1 Å, which in all models is lower than 1 Å. Notably, the docking AmEno15-7WH3 showed the lowest deformability values, while the docking AmEno15-4DUR showed the highest values (Figure 3, Figure 4, Figure 5 and Figure 6, Panels B). The B-factor (pink graphs) showed the protein flexibility and movement, allowing the identification of mobile or static regions within the structure. Regions with high B-factors indicate greater flexibility and movement, while those with low B-factors suggest static regions. According to this, our results indicate regions with significant flexibility in the four docking (Figure 3, Figure 4, Figure 5 and Figure 6, Panels C). Regarding the eigenvalues (purple graphs), which represent the energy required to deform the structure, we found that the values for each docking were 6.174995 e-06 (AmEno15-3LBX), 6.106602 e-05 (AmEno15-7WH3), 8.125026 e-05 (AmEno01-3M7P), and 9.918514 e-06 (AmEno15-4DUR). As the lower values denote easier deformability, the results suggest that docking AmEno15 with spectrin (3LBX) and plasminogen (4DUR) are the most flexible structures, requiring less energy to be deformed (Figure 3, Figure 4, Figure 5 and Figure 6, Panels D). In the conformational change analysis, the variance map, showed the parts of the docking complex more likely to move together; in this regard, we found in all docked models high values of cumulative variance (cyan bars) representing significant movements, while low individual variances percentages (purple bars), indicated the atoms more likely to remain in place or move together as a rigid unit (Figure 3, Figure 4, Figure 5 and Figure 6, Panels E). The covariance maps represent the movement patterns between linked amino acid residues, which result in molecular communication and interactions within the docked structure, revealing key regions of protein function. In this regard, our results revealed red regions along the diagonal line of the map in the four docked complexes, which represent linked regions among the receptor and ligand amino acids. In contrast, the regions in white and blue were out of the diagonal line, which represent the uncorrelated and anti-correlated amino acids, respectively (Figure 3, Figure 4, Figure 5 and Figure 6, Panels F). Lastly, the elastic network model showed that each dot in the graph corresponded to one spring between a pair of atoms. The more grey the dot, the more rigid the spring (grey-black scale). Our results showed dots in light grey along the diagonal, which represent a low spring's stiffness, resulting in highly flexible docked complexes (Figure 3, Figure 4, Figure 5 and Figure 6, Panels G).

3.3. Microplate Binding Assays to Stomatin, Spectrin, Plasminogen, and Fibronectin

The binding of recombinant AmEno15 to erythrocyte proteins (stomatin and spectrin), fibronectin, and plasminogen was evaluated by a microplate binding assay. We observed that AmEno15 binds to spectrin in a concentration-dependent manner ranging from 10 to 0.312 μg/mL, but not in the BSA control (Figure 7A). Interestingly, the binding assay of AmEno15 to stomatin, with concentrations ranging from 10 to 0.312 μg/mL, showed a similar binding in all cases, but not in a concentration-dependent manner. Nevertheless, when we used lower concentrations (0.312 to 0.039 μg/mL), we did observe binding in a concentration-dependent manner (Figure 7B). As in the previous assay, we did not observe binding to the BSA control. Regarding the AmEno15 to fibronectin binding assay, we observed a concentration-dependent binding in all cases (10 to 0.312 μg/mL), and we did not observe binding to the BSA control (Figure 7C). Finally, in the assay using plasminogen, we found that in concentrations of 10, 5, and 2.5 μg/mL, there is no significant difference in the binding of AmEno15. Notably, at concentrations 1.25, 0.625, and 0.312 μg/mL, the results suggest a concentration-dependent binding. Nevertheless, in BSA control wells, at lower concentrations than 0.312 μg/mL, we did not observe any binding (Figure 4D). It is important to note that the control used in all assays (only AmEno15 at different concentrations) revealed absorbances that corresponded to a non-specific binding, which supports the idea that the binding of AmEno15 depends exclusively on the presence of the ligand.

4. Discussion

So far, bovine anaplasmosis is a disease with many unanswered questions and molecular processes that need to be understood. Several efforts have been made to elucidate the fundamental biological mechanisms of A. marginale, including adhesion and invasion of erythrocytes, as well as adhesion to tick cells [36,37,38]. Nevertheless, research must continue to identify the leading proteins of these key mechanisms and to propose new strategies to develop vaccine candidates. In this regard, only a few studies of the molecules involved in the A. marginale-erythrocyte interaction have been reported. Like Noh et al. [36], who identified A. marginale proteins Msp1a, Mlp3, Am779, Msp3, as adhesins for bovine erythrocytes, and Omp13 as a binding protein to tick cells. Similarly, the recombinant proteins Msp1a and Msp1b of A. marginale expressed in E. coli allowed this bacterium to adhere to bovine erythrocytes [38].

According to this, there is growing research on this topic to understand the A. marginale target cell adhesion, whether in the host or the biological vector. Thus, we performed an immunoinformatics analysis that allowed us to predict the binding potential of AmEno15 to diverse ligand proteins. In this regard, we include as ligands some erythrocyte membrane proteins like spectrin (3LBX) and stomatin (7WH3), and other proteins probably related to bacterial pathogenesis (fibronectin (3M7P9), and plasminogen (4DUR)) [10,39,40]. The bioinformatics analysis of the Ramachandran plot scores for the four docked complexes strongly suggests feasible interactions of AmEno15 with all ligands. Additionally, the molecular dynamics of these complexes revealed the potential interactions between individual amino acid residues, as well as a high degree of deformability. Therefore, the AmEno15 may be able to bind to ligands involved in different pathogenic processes.

Experimentally, we demonstrated that AmEno15 can bind specifically to spectrin and stomatin in a concentration-dependent manner. These results could support the fact that A. marginale binding to the membrane is key in the invasion and establishment inside the erythrocyte, which is fundamental for pathogen survival in the host. Therefore, we propose that A. marginale could utilize AmEno15 to bind to stomatin during the initial adhesion process, and potentially, AmEno15 would bind to spectrin once A. marginale reaches the inside of the erythrocytes, since spectrin is located at the inner cytoplasmic face and is the principal constituent of the erythrocyte membrane skeletal network [41].

In addition, we prove that the AmEno15 also binds to fibronectin and plasminogen in a concentration-dependent manner. In this regard, we hypothesize that A. marginale could use AmEno15 as a virulence factor to invade fibronectin-producing cells [42]. The above is based on the fact that some members of the Anaplasmataceae family, such as A. phagocytophilum, can infect and lyse endothelial cells, which is crucial in the bacterial infection and invasion [43]. Likewise, experiments using dual fluorescence microscopy showed co-localization of A. marginale with an endothelial cell marker in tissue sections from an infected calf, suggesting a potential tissue invasion [37]. Although this process remains unexplored in A. marginale, our findings do not exclude this possibility.

Concerning the vector, we propose that AmEno15 may participate in the fibrinolysis process by binding to circulating plasminogen in host blood, thereby avoiding coagulation and facilitating tick blood intake [44]. Additionally, we think that the AmEno15-fibronectin interaction could be relevant in the context of the A. marginale infection to tick gut cells, similarly to Borrelia burgdorferi, whose membrane proteins interact with a protein containing fibronectin domains in the Ixodes spp gut cells. This process induces a spirochete congregation in the tick gut, facilitating the bacteria's exit to the salivary glands [39]. However, this hypothesis must be studied in greater depth in A. marginale.

Since the identification of the moonlighting proteins in the Mexican strains of A. marginale [11], elucidating the role of AmEno15 was a priority. Our results showed that AmEno15 binds specifically to spectrin, stomatin, fibronectin, and plasminogen; however, the molecular role of AmEno15 in each interaction remains to be explored under in vivo conditions. Therefore, this work represents the first approach to understanding, particularly the role of enolase, in binding to erythrocyte proteins, extracellular matrix components, tick gut proteins, and plasminogen.

It is undeniable that our findings contribute to the knowledge of A. marginale biology and the elucidation of the molecular communication between vector-pathogen-host, a topic of research that may be useful in the development of drugs and/or vaccines against tick-borne diseases.