Background

Panthera is a genus within the family Felidae, comprising five species: P. leo, P. onca, P. uncia, P. pardus and P. tigris. Sadly, Panthera spp. are now endangered due to illegal hunting, loss of habitat and trading [1]. Furthermore, Panthera spp. have been subjected to outbreaks of babesiosis which is a zoonotic, hemoprotozoan disease caused by tick-borne piroplasmids of the genus Babesia [2]. Babesia spp. are the second most common haemoparasites of mammals after trypanosomes. They have a wide range of hosts including humans (Homo sapiens), bovines (e.g. Bos taurus), pantheras (e.g. P. tigris, P. leo, P. pardus), equines, and a number of other mammal and bird species [3,4,5]. To date, more than 100 species of Babesia have been identified, which may be attributed to the fact that Babesia is not strictly host-specific. While babesiosis infection is most commonly asymptomatic, the disease can cause fever, fatigue and haemolytic anaemia that at times can be fatal [5]. Babesiosis is believed to be a major cause of mortality for big cats held in captivity [6,7,8,9,10,11,12,13,14,15,16,17] (Table 1). Species of Babesia infecting Panthera spp. are usually Babesia leo or Babesia canis [12]. A number of effective therapeutics are currently in use which include imidocarb dipropionate, diminazene aceturate, clindamycin, azithromycin and antiparasitic drugs such as atovaquone-azithromycin [14, 18]. The current drug application is often accompanied by intravenous fluids and blood transfusion; these are difficult to administer in the wilderness [14, 18] and therefore developing new and effective therapeutics against babesiosis infection is very important for animal conservation.

Table 1 Reported Babesia spp. infections in Panthera spp
Full size table

Being a crucial part of the protein translation machinery, aminoacyl tRNA synthetases (aaRSs) are currently being studied as drug targets for several pathogens [19,20,21,22,23,24,25,26,27,28]. The aaRSs catalyse the addition of amino acids to respective tRNAs, and are usually multi-domain proteins with an anti-codon binding domain (ABD) as well as a conserved catalytic domain. Often aaRSs possess editing domains that remove incorrectly charged tRNAs. The 20 aaRSs fall into either Class I or Class II based on their modes of substrate binding and structural folds. Class I aaRSs contain a Rossmann fold which is characterized by KMSKS and HIGH motifs. Class II aaRSs have three motifs and another different β-sheet arrangement [29]. The aaRSs can localise to different subcellular compartments where they are responsible for protein synthesis [30]. More recently, aaRSs have been explored from several eukaryotic pathogens such as Plasmodium, Toxoplasma and Leishmania as potential druggable targets [19,20,21,22,23,24,25,26,27,28]. In the present study we used bioinformatics tools to investigate the genomes of Babesia spp. (B. bovis, B. microti and B. bigemina) and P. tigris and annotated their aaRSs. We identified 33 (B. bovis), 34 (B. microti), 33 (B. bigemina) and 33 (P. tigris) aaRSs in these organisms. We have analysed the aaRSs sequences and predicted their cellular localizations. Finally, we determined the percentage sequence identity in aaRSs from Babesia spp. with respect to P. tigris to identify divergent aaRSs. This work hence provides a resource for launching experimental investigations of druggable Babesia aaRSs in an effort to control babesiosis.

Methods

Open reading frames (ORFs) for the annotated B. bovis, B. microti, B. bigemina and P. tigris were downloaded from the National Center for Biotechnology Information (NCBI) database of genomes and computationally translated. All 20 aaRSs were individually used to perform BLAST searches against non-redundant databases available at the NCBI (Additional file 1: Dataset S1). For each, hidden Markov models (HMMs) [31] were generated using homologs identified from BLAST searches. Sequence alignment and modeling software system (SAM) was used to generate multiple sequence alignment (MSA), and the HMMER package was used for building profile HMM for each MSA. Significantly similar matches for each HMM profile were identified within the genomic sequences for the three Babesia spp. and for P. tigris. Genome sequences were analysed using blastx and four additional aaRSs were identified. Pfam domains were assigned to computationally translated sequences using the ‘Pfam web server for analysis of domains’ [32]. Details on domain functions were also obtained from the Pfam database. Subcellular localizations were predicted using Target P1.1 [33], SignalP 3.0 [34], TMHMM Server v.2.0 [35] and Deeploc 1.0 web server [36]. The presence of secretory signals was detected using Target P1.1 and SignalP, while transmembrane domains were detected using TMHMM. Furthermore, DeepLoc-1.0 was used to predict the subcellular localisations. If a protein was predicted to have a signal peptide and at least one transmembrane domain and predicted to be either mitochondrial or apicoplastic via DeepLoc-1.0 it was annotated as an organellar protein. Homology modelling for protein sequences was performed using Phyre2 server [37] and Chimera [38] was used for structure visualization and analyses.

Results

The Babesia spp. that infect Panthera spp. are usually B. leo or B. canis [12]. Since the genomes of B. leo and B. canis are currently unavailable, we used the open reading frames (ORFs) and genome sequences of B. bovis, B. microti and B. bigemina that are known to infect bovine species and at times Panthera spp. as well [39]. Genome sizes for B. bovis, B. microti and B. bigemina are ~ 8.2, ~ 6.3 and ~ 13.8 Mbp, respectively. The aaRSs in the genomes of B. bovis, B. microti, B. bigemina and P. tigris were identified via HMM-based searches [31]. We annotated aaRSs corresponding to each of the 20 amino acids in these four studied genomes (Tables 2, 3, 4, 5, Additional file 2: Tables S1–S4). Our analyses have annotated 33 aaRSs in B. bovis, 34 aaRSs in B. microti, 33 aaRSs in B. bigemina and 33 aaRSs in P. tigris (Fig. 1a).

Table 2 Aminoacyl-tRNA synthetase (aaRS) domains in the genome of B. bovis
Full size table
Table 3 Aminoacyl-tRNA synthetase (aaRS) domains in the genome of B. microti
Full size table
Table 4 Aminoacyl-tRNA synthetase (aaRS) domains in the genome of B. bigemina
Full size table
Table 5 Aminoacyl-tRNA synthetase (aaRS) domains in the genome of P. tigris
Full size table
Fig. 1
figure 1

Aminoacyl-tRNA synthetases (aaRSs) and their sites. Number of annotated aaRSs (a) and predicted subcellular localisations (b) for the studied Babesia spp. (B. bovis, B. microti and B. bigemina) and Panthera tigris

Full size image

The predicted compartmentalisations of all aaRSs were analysed based on the presence of signal sequences using Target P1.1 [33], SignalP 3.0 [34], TMHMM v.2.0 [35] and Deeploc 1.0 web servers [36]. Our analyses predicted subcellular localisation for multiple aaRSs with gene copies in both cytoplasm and in an organelle (apicoplast/mitochondria) for Babesia spp. (Fig. 1b, Tables 2, 3, 4). For P. tigris, the predicted localisations cover both cytoplasm and mitochondria (Fig. 1b, Table 5). Hereafter, the aaRSs with a predicted subcellular localisation to apicoplast in Babesia spp. will be referred to as organellar. We emphasize the need for experimental validation to positively assess the localisations as aaRSs may co-reside in mitochondria [40,41,42]. In B. bovis, the predictions are for 16 aaRSs as cytoplasmic, 12 aaRSs as organellar and 5 aaRSs [cysteinyl-tRNA synthetase (CRS), alanyl-tRNA synthetase (ARS), aspartyl-tRNA synthetase (DRS), glycyl-tRNA synthetase (GRS), threonyl-tRNA synthetase (TRS)] as potential co-localisers between the cytoplasm and organelle (Fig. 1b, Table 2). In B. microti, 16 aaRSs are cytoplasmic, 14 aaRSs are organellar and 4 aaRSs (CRS, ARS, DRS, GRS) are possibly dual-localised (Fig. 1b, Table 3). In B. bigemina, 16 aaRSs are cytoplasmic, 12 aaRSs are organellar and 5 aaRSs (CRS, ARS, DRS, GRS, TRS) may reside in both the cytoplasm and organelle (Fig. 1b, Table 4). Our HMM profile-based analyses failed to predict a complete set of 20 organellar aaRSs for three aaRSs [glutaminyl-tRNA synthetase (QRS), valyl-tRNA synthetase (VRS), phenylalanyl-tRNA synthetase (FRS)] in B. bovis, three aaRSs (VRS, FRS, TRS) in B. microti and three aaRSs (QRS, VRS, FRS) in B. bigemina. In an attempt to predict the missing organellar aaRSs within the Babesia spp., we searched the non-redundant protein sequence databases. Our searches resulted in identification of tRNA synthetases (isoleucine, leucine, methionine and valine; I, L, M and V, respectively) in B. bovis, B. microti and B. bigemina with ~ 55%, 54% and 55% sequence identity, respectively, with, for example, predicted VRS from each of the three species (Tables 2, 3, 4). Searches using Blastx resulted in the identification of four additional aaRSs: methionyl-tRNA synthetase (MRS) (cytoplasmic) in B. bovis; QRS (organellar) and tryptophanyl-tRNA synthetase (WRS) (organellar) in B. microti; and tyrosyl-tRNA synthetase (YRS) (cytoplasmic) in B. bigemina. Based on previous reports, a non-discriminating organellar glutamyl-tRNA synthetase (ERS) is known to convert tRNAGln into a misacylated Glu-tRNAGln [43]. The misacylated Glu-tRNAGln is then converted into Gln-tRNAGln by an organellar glu-tRNAGln amidotransferase [43]. While this explains the absence of QRS in the organelles of the three Babesia spp., there is a need to identify the missing organellar TRS in B. microti to complete its set of 20 aaRSs. A plausible reason for failure to predict a complete set may be poor sequence identity with aaRS homologs that were used for building profile HMMs.

In contrast to Babesia spp., for P. tigris, 19 aaRSs are present within the cytoplasm including a bi-functional aaRS, namely glutamyl-prolyl-tRNA synthetase (EPRS). In addition, 14 aaRSs are mitochondrial in P. tigris (Fig. 1b, Table 5). These observations are supported by previous reports which show that despite the occurrence of protein synthesis in the three distinct compartments in Plasmodium falciparum, this apicomplexan does not possess three complete sets of 20 aaRSs for each compartment (i.e. not 60 but has 36 aaRSs) [42]. Furthermore, while cytoplasmic aaRSs supposedly can drive translation in that compartment, several studies have revealed that some aaRSs have the potential to be multi-localized [30]. When a complete set of 20 aaRSs is not available for organellar protein synthesis, as an example, the mitochondria are able to import charged tRNAs, as shown in yeast, Leishmania, Trypanosoma, Plasmodium and Toxoplasma [44,45,46,47,48]. The requirement for their own mitochondrial aaRSs in these organisms is therefore bypassed as charged tRNAs are transported between the cellular compartments, thereby compensating for the absence of any aaRS(s) [40,41,42].

One aaRS enzyme is particularly interesting as it can occur as heterodimers of two different genes: the FRSs. It is well established that FRSs can exist in (αβ)2 subunit organization, with exceptions of mitochondrial FRSs from yeast and Homo sapiens and possibly others that are monomeric [40]. Our analyses predicted cytoplasmic FRS α and FRS β subunit in all the four studied genomes. A third gene for FRS was predicted in B. bovis and B. bigemina to localise within the cytoplasm, while the gene for FRS in B. microti was predicted to localise within the organelle (Tables 2, 3, 4). Mitochondria of the apicomplexan P. falciparum import tRNAs along with an active FRS, and so we analysed the FRSs from B. bovis, B. microti and B. bigemina using Target P1.1 [33] and SignalP 3.0 [34] for the presence of signal peptide and transmembrane domain [35]. The encoded cytoplasmic FRSs from B. bovis and B. bigemina showed no identifiable organelle targeting sequence or transmembrane domains. However, a transmembrane domain was predicted within the FRS from B. microti (Table 3). These predicted FRSs need to be characterised experimentally for validation.

Domain architectures of aminoacyl tRNA synthetases in Babesia spp. and Panthera tigris

Based on HMM searches, we identified a total of 33 aaRSs in B. bovis, 34 aaRSs in B. microti, 33 aaRSs in B. bigemina and 33 aaRSs in P. tigris. Amongst these, 18 aaRSs belong to Class I and 15 aaRSs to Class II in B. bovis; 18 to Class I and 16 Class II in B. microti; 18 to Class I and 15 to Class II in B. bigemina; and 16 to Class I and 17 to Class II in P. tigris (Tables 2, 3, 4, 5). Pfam server predicted catalytic domains for all these aaRSs. Additionally, anticodon-binding domains (ABD), N-terminal domain (NTD) and C-terminal domain (CTD), DALR [aspartate (D), alanine (A), leucine (L), arginine (R)], DHHA [aspartate (D), histidine (H), histidine (H), alanine (A)], SAD (second additional domain), WHEP-TRS (Wh-T) and glutathione S-transferase (GST) domains were predicted.

Panthera tigris aaRSs are also predicted to contain DALR and DHHA domains (Fig. 2). DALR is an all alpha-helical anti-codon binding domain named after the characteristic conserved amino acids: aspartate (D), alanine (A), leucine (L) and arginine (R). The DHHA domain is named after conserved amino acids: aspartate (D), histidine (H), histidine (H), alanine (A) in the cytoplasmic version of ARS [49]. Furthermore, in P. tigris a GST-like domain was predicted at the N-termini of dual-localised MRS, EPRS and cytoplasmic VRS (Fig. 2). While the functional implications of the GST domains may vary, these seem to be involved in protein assembly and folding [50, 51]. Several studies have reported that GST or GST-like domains play an important role in the formation of complexes between aaRSs and multifunctional factors (p18, p38, p43) [52]. Additionally, Wh-T domain was also predicted in cytoplasmic MRS, WRS, GRS and in the dual-localised HRS/EPRSs (Fig. 2). Wh-T domain has a characteristic helix-turn-helix motif, with consensus lysine and arginine residues. These residues are critical for protein-protein interactions within the multi-synthetase complex [53,54,55,56].

Fig. 2
figure 2

NCBI reference sequence and pfam assigned domains in B. bovis. Abbreviations: aaRSs, aminoacyl-tRNA synthetase; CD, catalytic domain; ABD, anticodon binding domain; SAD, second additional domain; NTD, N-terminal domain; CTD, C-terminal domain; ATED, aminoacyl tRNA editing domain; TBA, tRNA binding arm; GST, glutathione S-transferase

Full size image

We observed no distinct domains amongst the aaRSs from B. bovis, B. microti and B. bigemina (Figs. 3, 4, 5). In Babesia spp., RRS and SRS were found to have an additional NTD, while their PRSs have an additional CTD, with no predicted function for either of them (Figs. 3, 4, 5) [57, 58]. Interestingly, for B. microti, a ZnF domain was annotated for the organellar IRS. The C-terminal peptide (CP) and C-terminal enzyme-bound zinc participate in aminoacylation of tRNAIle (Fig. 4). These domains were not found for other two species, B. bovis and B. bigemina. The differences observed within the aaRSs of three Babesia spp. and P. tigris provide a unique opportunity to exploit some aaRSs as potential drug targets.

Fig. 3
figure 3

NCBI reference sequence and domain annotations of aaRSs in B. microti. Abbreviations: aaRSs, aminoacyl-tRNA synthetases; ABD, anticodon binding domain; SAD second additional domain; NTD, N-terminal domain; CTD, C-terminal domain; ATED, aminoacyl tRNA editing domain; GST, glutathione S-transferase domain

Full size image
Fig. 4
figure 4

Domain annotations and NCBI reference sequence for aaRSs in B. bigemina. Abbreviations: aaRSs, aminoacyl-tRNA synthetases; ABD, anticodon binding domain; SAD, second additional domain; NTD, N-terminal domain; CTD, C-terminal domain; ATED, aminoacyl tRNA editing domain; TBA, tRNA binding arm, GST, glutathione S-transferase domain

Full size image
Fig. 5
figure 5

NCBI reference sequence and Pfam assigned domains in P. tigris. Abbreviations: aaRSs, aminoacyl-tRNA synthetases; ABD, anticodon binding domain; TBA, tRNA-binding arm; NTD, N-terminal domain; CTD, C-terminal domain; ATED, aminoacyl tRNA editing domain; GST, glutathione S-transferase domain; WT, WHEP-TRS domain

Full size image

We subsequently studied the sequence identities between Babesia spp. (B. bovis, B. microti and B. bigemina) and P. tigris (Tables 2, 3, 4, 5). While the aaRSs identified in the three Babesia spp. share similarities with their homologs in P. tigris, our analyses reveal poor identity in the range of ~ 20–50% (Tables 2, 3, 4, 5). The aaRSs with less than 30% identity with the corresponding protein in P. tigris include ERS, IRS, MRS, WRS, YRS, HRS, NRS and SRS (Tables 2, 3, 4, 5). Out of eight aaRSs with sequence identity less than 30% between Babesia spp. and P. tigris, ERS, MRS, YRS and NRS are cytoplasmic. Among the aaRSs with sequence identity less than 30%, IRS and LRS are the targets of drugs currently in use, i.e. IRS (bacterial: muciprocin, available as a 2% topical preparation) and LRS (fungal: tavaborole/AN2690, available as a 5% topical preparation) [19, 24, 59,60,61].

Structure-based analysis of cladosporin binding sites in Babesia and Panthera KRSs

The Plasmodium KRS (for cladosporin, CLD) and PRS (for halofuginone) are currently being studied as potential drug targets [19, 24, 59,60,61]. We analysed CLD which, along with its analogs, is being investigated as an inhibitor of P. falciparum-KRS (Pf-KRS) and of various other pathogen KRSs [21, 62]. Several amino acid residues in CLD binding pocket are highly conserved except at two positions near the ATP binding pocket and adjacent to CLD methyl moiety (Fig. 6a). The basis for CLD selectivity has been ascribed to these two key selectivity residues that show clear divergence across species. As proof of concept, we conducted a structural analysis of CLD bound Homo sapiens KRS (Hs-KRS) (PDB: 4YCU) and Plasmodium falciparum KRS (Pf-KRS) (PDB: 4PG3) in comparison with the three Babesia spp. and P. tigris (Fig. 6b). We built a three-dimensional structure model for KRSs using Phyre2 (Protein Homology/AnalogY Recognition Engine)-based protein structure prediction [37]. The two key residues bestowing selectivity to CLD were analysed in Babesia spp. and P. tigris. In Pf-KRS, valine and serine (VS) occupy the two selectivity residue positions thereby providing tight binding. Our analysis here reveals the presence of structurally smaller residues, cysteine and serine (CS), in B. bovis and B. bigemina, with an exception of B. microti that has a smaller valine and a bigger non-favourable threonine (VT) at the site (Fig. 6b). In Hs-KRS, these two positions are occupied by bulkier glutamine-threonine (QT) residues that likely hamper CLD binding. The corresponding position in P. tigris is also occupied by these bulkier QT residues (Fig. 6b). It is noteworthy that biochemical analysis of recombinant KRSs has previously shown that CLD displays a nanomolar range potency of inhibition (IC50 ~ 40–90 nM) against Pf-KRS, which is ~ 500-fold higher when compared with Hs-KRS [20, 62, 63]. This suggests potential poor selectivity for P. tigris KRS due to the same bulkier residues (QT) at the CLD binding pocket in comparison to the smaller, more favourable residues (CS/VT) in Babesia spp. This analysis suggests that small molecule targeting of KRS active site in Babesia spp. is an attractive avenue from the perspective of developing anti-infectives.

Fig. 6
figure 6

Analysis of CLD binding site in KRSs from Babesia spp. and P. tigris. a Sequence alignment of KRSs. The two key residues are highlighted in a black box and other conserved residues responsible for CLD binding are shown in blue boxes. b Structural superimposition of Hs-KRS (PDB: 4YCU), Pf-KRS (PDB: 4PG3), P. tigris and three Babesia spp. (Bb, B. bovis; Bm, B. microti and Bg, B. bigemina) (built structure model, this study). The smaller CLD selectivity residues of cysteine-serine (for Bb, Bg) and valine-threonine (for Bm) may accommodate CLD in the binding pocket of Babesia KRSs. The bulkier glutamine-threonine (for Hs, H. sapiens; Pt, P. tigris) potentially hinder high potency CLD binding. Abbreviations: CLD, cladosporin; KRS, lysyl-tRNA synthetase

Full size image

Discussion

The treatment of babesiosis presents an emerging challenge. It is a hemoprotozoan disease whose causative agents are apicomplexan Babesia spp. Panthera spp. have been subjected to outbreaks of babesiosis, caused by Babesia spp., the second most common haemoparasites of mammals after trypanosomes, with a worldwide distribution. The current regime of therapy for targeting babesiosis requires the use of antibiotics and antiparasitic drugs. While atovaquone plus azithromycin is used to treat most cases, clindamycin plus quinine is used in more severe cases with chances of relapses [19, 24, 59,60,61]. Therefore, new drugs with high specificity and low toxicity are desirable. The recent availability of Babesia spp. (B. bovis, B. microti and B. bigemina) genomes has paved the way for screening of new chemotherapy targets. In this study, we focused on the housekeeping enzymes aminoacyl-tRNA synthetases that are essential for protein synthesis and cell viability. Lately, parasitic, microbial and fungal aaRSs have been explored for druggability [19,20,21,22,23,24,25,26,27]. A bacterial IRS inhibitor, mupirocin (marketed as Bactroban), and a fungal LRS inhibitor, 5-fluoro-1,3-dihydro-1-hydroxy-2,1-benzoxaborole (AN2690), have been developed for human use [19, 24, 59,60,61]. Recent studies on potential anti-malarial compounds like CLD (against KRS) and halofuginone (against PRS) are also promising [19, 20, 22, 24, 59, 64]. Considering this evidence, we have identified Babesia aaRSs with poor sequence identity to Panthera aaRSs as proteins of interest [65, 66]. As proof-of-concept, we performed a structure-based analysis of P. tigris and Babesia KRSs and have indicated a potential for selective drug targeting (Fig. 6). Hence, our work here lays a foundation for the future to further investigate and exploit Babesia aaRSs as potential targets.

Conclusions

In the present study, we provide data on genome-wide identification and annotation of aaRSs from Babesia spp. and P. tigris. Poor sequence identity (~ 20–50%) between pathogen/host aaRS pairs offers a window for specific studies to explore druggability. This detailed genomic cataloguing of aaRSs from pathogenic Babesia merits future experiments to validate new drug targets against Babesia spp.