In silico analysis of putative drug and vaccine targets of the metabolic pathways of Actinobacillus pleuropneumoniae using a subtractive/comparative genomics approach

The respiratory disease known as porcine respiratory disease complex (PRDC) is a widespread problem on intensive pig farms worldwide. PRDC is a polymicrobial disease that is caused by various viral and bacterial agents, including Mycoplasma hyopneumoniae and Actinobacillus pleuropneumoniae, which are considered to be the primary pathogen in pigs [27]. Coinfection with these two pathogens is known to cause more severe disease than infection with either pathogen alone or with other agents [5].

A. pleuropneumoniae is a Gram-negative, facultative anaerobic bacterium that belongs to the family Pasteurellaceae. The organism is known to cause porcine pleuropneumonia, a severe contagious respiratory disease that often leads to very rapidly evolving pleuropneumonia, which is characterized by hemorrhagic necrotizing pneumonia and fibrinous pleuritis, and most commonly affects pigs aged 18 to 20 weeks [14]. The polysaccharide capsule and exotoxins of A. pleuropneumoniae are the major virulence determinants and are responsible for the pathogenesis of pleuropneumonia [2, 15].

Over the last two decades, increasing numbers of A. pleuropneumoniae strains have been isolated that are resistant to a number of commonly utilized drugs for treating pleuropneumonia in pigs [19, 37, 40]. This alarming increase in the incidence of antimicrobial resistance has initiated a search for new therapeutics against these ‘superbugs’. Current developments in the fields of genomics and proteomics have opened new avenues into the development of new antimicrobial agents and vaccine targets for combating drug resistance [31].

Different researchers have discussed the significance of in silico approaches for the identification of vaccine and drug targets. Morya et al. [24] identified drug targets in Staphylococcus aureus by analyzing its metabolic pathways. Likewise, vaccine and therapeutic drug targets in methicillin-resistant S. aureus [35], Mycobacterium abscessus [32], Mycobacterium ulcerans [3], Vibrio cholerae [4], and Mycoplasma hyopneumoniae [6] were identified by using the same approach.

However, there are no reported therapeutic target data available for the metabolic pathways in A. pleuropneumoniae. Herein, using the available genome sequences and genetic/proteomic database resources, we identified putative targets for antibiotics and vaccine therapy in A. pleuropneumoniae by undertaking an in silico comparative and subtractive genomic/proteomic metabolic pathway analysis approach.

In this study, A. pleuropneumoniae L20 (serotype 5b), with 104 identified pathways, was selected. The sequences of three different A. pleuropneumoniae strains are deposited in the NCBI-KEGG database. A. pleuropneumoniae JL03 (serotype 3) and A. pleuropneumoniae AP76 (serotype 7) have 105 and 106 pathways, respectively. We selected the strain with all of the common pathways for further analysis.

Comparison of the selected pathways in the pathogen to the 295 and 299 pathways in pig and human, respectively, showed that 29 pathways were unique to the pathogen. Twenty of these pathways were metabolic pathways (Table 1). More than 900 proteins were involved in the common and unique metabolic pathways; however, only 273 of them were non-homologous to both pig and human, based on the established KEGG cutoff value.

Table 1
Unique metabolic pathways of Actinobacillus pleuropneumoniae

• Click for larger image

We aligned these non-homologous proteins with the essential protein sequences of Haemophilus influenzae Rd KW20 (NC000907), which shares 85% genome similarity with A. pleuropneumoniae, in the DEG and identified 122 essential non-homologous proteins in A. pleuropneumoniae. These essential proteins were involved in 40 different metabolic pathways, and, of these 122 proteins, 13 were from 6 different unique metabolic pathways, which were only present in the pathogen.

These non-homologous proteins were compared with all of the essential proteins of the 37 bacteria in the DEG. The greatest numbers of homologous proteins were found for the essential proteins of Synechococcus elongate PCC 7942, Mycobacterium tuberculosis H37Rv II, and Acinetobacter baylyi ADP1, with 130, 127, and 126 hits, respectively. The smallest number of hits was for Salmonella enterica subsp. enterica serovar Typhimurium str. 14028S, with only nine homologous proteins (Fig. 2).

Fig. 2
Number of Actinobacillus pleuropneumoniae genes with essential gene hits in the Database of Essential Genes (DEG). The essential genes of A. pleuropneumoniae were identified by comparison to those of all 36 bacteria in the DEG.

• Click for larger image

The cellular localization of each non-homologous protein was determined by using the CELLO database, and the results showed 95 cytoplasmic, 11 transmembrane, 4 periplasmic, and 2 outer membrane proteins as well as 10 proteins with undetermined localization. In the unique pathways, there were 7 cytoplasmic, 2 transmembrane, 1 periplasmic, and 1 outer membrane proteins and 2 with undetermined cellular localization. However, the TMHMM server predicted 14 transmembrane proteins, and 9 (64.3%) of them matched with the CELLO prediction (Table 2). Furthermore, 11 of them showed an antigenicity prediction > 0.4 in VaxiJen v2.0 (Table 3).

Table 2
Cellular localization of non-homologous essential proteins of Actinobacillus pleuropneumoniae and their molecular weight (MW)

• Click for larger image

Table 3
Antigenic prediction of the transmembrane proteins of Actinobacillus pleuropneumoniae using Vaxijen v.20 [8]

• Click for larger image

DrugBank was used to categorize the druggability of the identified essential proteins. Of the non-homologous essential proteins, 24 had a hit in DrugBank of an approved, nutraceutical, investigational, or experimental drug (Table 4). Six genes, appser1_11470, accC, guaA, rpoA, asd, and murC, had hits in DrugBank, with an e-value limit of 10⁻²⁵. Each of these genes had at least a one hit of an approved, nutraceutical, or experimental drug.

Table 4
Similarity of the non-homologous proteins of Actinobacillus pleuropneumoniae to the binding proteins of FDA-approved drugs from DrugBank [38]

• Click for larger image

The molecular weight of each essential protein was determined by using UniProt. Among the 122 essential non-homologous proteins, 120 had a low molecular weight (< 110 kDa; Table 2).

The essential proteins of A. pleuropneumoniae were then cross-checked for similarity to the proteins of M. hyopneumoniae, which is a major component of PRDC. Three proteins with similarity to essential proteins of M. hyopneumoniae were identified: DNA-directed RNA polymerase subunit alpha, rpoA; methionine-tRNA ligase, metG; and glutamate-tRNA ligase, gltX.

Previously, the development of new antimicrobial agents and vaccine therapies was limited by our understanding of the biology of the target microbial agents. However, in this post-genomic era, advances in the fields of genomics and proteomics have allowed for in silico investigations of new drugs and vaccine targets, by using genomic and protein sequence resources [6]. A. pleuropneumoniae, a bacterium known to cause PRDC in pigs [27], is often resistant to most of the drugs commonly used to treat the disease [37]. Thus, alternative therapeutic agents are greatly needed. In this study, by using an in silico approach, we identified unique proteins in the metabolic pathways of A. pleuropneumoniae as potential targets for antimicrobial and vaccine therapy.

Multiple unique metabolic pathways were identified in A. pleuropneumoniae that are not present in their natural host. The presence of these unique pathways offers an opportunity to identify antimicrobials that specifically target the pathogen, and thus, should be safe. Targeting essential genes/proteins in unique metabolic pathways, which are required for survival and replication, provides an extra advantage for designing potent therapeutic agents, since they should interfere with the survival and/or replication of the pathogen [23, 42].

Determining the cellular localization of a protein is an essential step toward identifying it as a potential target for therapeutic intervention [33]. This leads to elucidation of the function of each protein, which helps to differentiate between targets for antimicrobial agents and those for vaccine therapy [12]. Most of the identified non-host essential cytoplasmic proteins could serve as targets for antimicrobial treatment, while the transmembrane proteins, as suggested by the TMHMM, could be potential vaccine targets [16]. These transmembrane proteins may be selected toxins/surface-exposed proteins and may be used for the production of a preventive vaccine that initiates antibody-mediated immunity [7].

Most of the identified essential proteins of A. pleuropneumoniae (120/122) had a low molecular weight, thereby providing a broad opportunity for selecting and utilizing these proteins as targets for therapeutic intervention. Low molecular weight proteins are likely to be soluble and purified easily, which makes them suitable drug targets [9]. Moreover, the existence of FDA-approved, nutraceutical, or experimental drugs with the ability to bind proteins similar to the identified essential proteins of the pathogen demonstrates the potential druggability of these proteins as therapeutic targets and offers the opportunity of using different combinations of drugs to treat Actinobacillus infections in pigs. In fact, five FDA-approved drugs with hits in essential non-host A. pleuropneumoniae proteins were identified in this study.

Penicillin-binding protein 2 (PBP2) and peptidoglycan synthetase, which are from a unique pathway in A. pleuropneumoniae, have been previously characterized as drug targets in other pathogens. In S. aureus, PBP2 is the only bifunctional penicillin-binding protein [13, 29], and the transpeptidase domain of the protein was reported to be critical for the survival and replication of the bacterium [28]. Peptidoglycan synthetase, FtsI, a cell division protein, is essential for the synthesis of peptidoglycan and catalyzes the synthesis of cross-linked peptidoglycan from lipid-linked precursors [25, 34]. Hence, inhibition of these proteins/enzymes using one or more drugs might reduce the infection rate and incidence of drug resistance in A. pleuropneumoniae.

The remaining transmembrane proteins, either from unique or common pathways, which showed antigenic and MHC cleavability, have been suggested as targets for vaccine therapy. Specifically, tetra-acyldisaccharide 4′-kinase and 3-deoxy-Dmanno-octulosonic-acid transferase, two transmembrane proteins in unique pathways, are involved in the synthesis of lipid A in the lipopolysaccharide layer [10]. Hence, these proteins could be targets for a host antibody response, since Gram-negative bacteria resist the host defense mechanism by upregulating their expression in the outer membrane. This upregulation results in an increased host response, which suggests the potential of these proteins as vaccine candidates [21, 26, 39].

Glycerate dehydrogenase, 3-isopropylmalate dehydratase large subunit, 3-isopropylmalate dehydrogenase, D-3-phosphoglycerate dehydrogenase, D-alanine--ligase, acyl-[acylcarrier-protein]--UDP-N-acetylglucosamine O-acyltransferase, UDP-3-O-acylglucosamine N-acyltransferase, and D-alanyl-D-alanine carboxypeptidase/D-alanyl-D-alanine-endopeptidase are among the cytoplasmic proteins in unique metabolic pathways, and these can be targeted for the development of novel antimicrobial agents against A. pleuropneumoniae.

The cytoplasmic enzyme glycerate dehydrogenase (ldhA) is an essential protein from a unique pathway; moreover, it is involved in eight different metabolic pathways, but mainly in glyoxylate and dicarboxylate metabolism, which is crucial for the synthesis of carbohydrates in the anabolic pathway by converting acetyl CoA. Thus, blockage of this specific enzyme or the pathway might lead to bacterial cell death due to carbohydrate limitation. Therefore, the glyoxylate and dicarboxylate metabolic pathway could be a useful drug target for A. pleuropneumoniae.

Furthermore, three of the essential proteins of A. pleuropneumoniae are also essential for M. hyopneumoniae, which is another major component of PRDC [6]. The DNA-directed RNA polymerase subunit alpha rpoA, which is involved in purine and pyrimidine metabolism, had a hit with an FDA-approved drug. Methionine-tRNA ligase (metG), which is an essential protein in selenocompound metabolism, had a hit in DrugBank for an experimental drug. Glutamate-tRNA ligase (gltX), an essential protein in porphyrin and chlorophyll metabolism and aminoacyl-tRNA biosynthesis, catalyzes the binding of glutamate to tRNA (Glu) in a two-step reaction. Phosphorylation of this enzyme by HipA, a toxin and serine/threonine kinase, results in amino acid starvation; prevents replication, transcription, translation, and cell-wall synthesis, and it inhibits growth, leading to multidrug resistance and persistence [11, 18, 30]. Targeting these proteins in both pathogens could be beneficial for preventing antimicrobial resistance and pig loss due to PRDC.

In conclusion, in silico approaches are of paramount importance when identifying target proteins and metabolic pathways as potential drug and vaccine therapy targets. In this study, we identified the glyoxylate and dicarboxylate pathways and glycerate dehydrogenase as putative targets for antimicrobial therapy against A. pleuropneumoniae; moreover, we identified tetra-acyldisaccharide 4′-kinase and 3-deoxy-D-mannooctulosonic-acid transferase as prospective vaccine targets. In addition, we identified three non-host essential proteins that are common to both A. pleuropneumoniae and M. hyopneumoniae; proteins that could serve as targets for antimicrobial therapy against both pathogens. However, although an in silico-based approach involves a series of screens for proteins that can be used as potential drug targets and vaccine candidates, the method has a major limitation in that the identified target proteins require experimental confirmation of their potential. In addition, non-protein vaccine and drug targets cannot be identified by applying this method [31]. Thus, a study should be undertaken to identify any unlisted pathogenic genes of A. pleuropneumoniae and determine the practicability of using the proteins and pathways identified as targets for drug and vaccine therapies.