Keywords: Pseudomonas aeruginosa, protein-protein inetarctions(PPIs), XL-MS, siderophores and enterobactin
Infectious bacterium and how to detect the structure of its molecules
Pseudomonas aeruginosa is a Gram-negative bacterium which is characterized by a rod-shaped body. It is considered as an opportunistic pathogen that causes severe and persistent infections in immunocompromised living organisms (plants and animals encompassing humans). The key Pseudomonas aeruginosa virulence, the development of MDR (multi-drug resistant) strains, and also the treatment-resistant phenotype of severe Pseudomonas infections call for the development of innovative therapeutic decisions. Almost biological roles in the living cells are controlled by the PPIs (protein-protein interactions). Therefore, advanced comprehending of the PPIs may offer new options for effective treatments of Pseudomonas aeruginosa. Even though, it is possible for large-scale interaction data to be held in the development of the drug. The information on PPIs of Pseudomonas aeruginosa is restricted to a few manually curated associations identified using the small-scale directed experiments.
Currently, the genome-scale assessment of the protein-protein interactions in Pseudomonas aeruginosa incorporated several genome characteristics of the Pseudomonas aeruginosa genes including products of their proteins. These features include co-expression, domain-domain association, and co-operon and colocalization among others. These features are against the assembled reference data set generated from the recognized protein-protein interactions existing in a three closely interrelated Gram-negative bacteria. These databases offer an important initial step in determining significant protein-protein interactions in the pathogenicity of the bacteria and the drug resistance. Further, the determination of the PPIs is facilitated by the experimental measurements. Additionally, practical experiments can indicate the specific interactions present and more predictions. This is possible because two proteins which do not have a common operon and believed to exist in distinctive subcellular sites can colocalize and associate in vivo. Therefore, they cannot be determined with high confidence as interactors.
Protein-protein interaction mapping experimentally may be realized by using well-recognized techniques like the co-immunoprecipitation, and the tandem affinity purification. Furthermore, genome-level screening like the high-throughput yeast two-hybrid technique (Navare et al., 762) can also be employed. Nonetheless, most of these techniques require the isolation of the proteins. Moreover, the assay within the non-native circumstances is necessary to enhance false-negative and false- positive outcomes. As a result of this, chemical crosslinking in connection with the mass spectrometry has dominated as a crucial tool for mapping protein-protein interactions originating from protein complexes. Furthermore, the chemical cross-linking with the mass spectrometry offers information to the conventional techniques with various distinctive pros. These advantages encompass the capability to distinguish the protein-protein interaction networks that occur in vivo. These types of PPI network are formed by the proteins that are close together as a result of a direct association as a result of protein association with the third protein. In addition, the chemical cross-linking in connection with the mass-spectrometry may be used to give structural details of useful protein complexes. For instance, chemical crosslinking of complete contagious virus spots of the Potato Leafroll Virus indicated the interaction locations between the viral coats of the protein that permitted the determination of the interfacial surfaces according to the cross-linking location-directed docking (Navare et al., 762).
Another advantage, of in vivo chemical cross-linking together with the mass spectrometry is its capability to differentiate the structural characteristics and protein-protein interactions from the intact cells, encompassing intact cell with the membrane proteins. Usually, membrane proteins represent about 30% of the whole eukaryotic proteome. The percentage representation of membrane proteins includes the required drug targets. However, they are under-represented in the data bank of proteins. For example, there are a few hundred of their recognized crystal structures. Intact membrane protein crystallization is problematic because membrane proteins are unstable once they have been isolated from their innate environments. Though, in vivo chemical cross-linking surpasses these problematic stages by connecting membrane proteins in their innate forms leading to the capture of the protein complexes and associations as the proteins occur in the cells. These advantages have made the chemical cross-linking together with the mass spectrometry to find numerous applications to investigate protein complexes from various samples encompassing the intact bacteria, cell lysate, and mammalian cells and virus particles.
Based on the protein interaction networks’ value, in determining the innovative drug targets for the human opportunistic bacteria, this paper focuses on the mapping of protein-protein interaction networks in the cells of Pseudomonas aeruginosa by using chemical cross-linking together with the mass spectrometer (XL-MS).
For the detection of protein-protein interactions, the PIR (protein interaction reporter) cross-linker BDP must be prepared. The PIR cross-linker BDP is prepared with the Fmoc chemistry. It is further esterified to yield the activated NHP (n-hydroxyphthalamide) ester of the BDP.
Pseudomonas aeruginosa cells are crosslinked by employing the BDP-NHP and the processed. Initially, P. aeruginosa is washed in PBS, followed by washing in 170mM sodium biphosphate maintained at pH 7.4 buffer. The BDP-NHP is then added to the final concentration of 10mM. The first stage of the cross-linking interaction is allowed to run for an hour. This is followed by centrifuging the cross-linked cells. When the interaction is complete, the cells are extensively washed in PBS buffer. The cross-linking is conducted twice in order to add 20mM to provide ample time for the cellular exposure to a soluble cross-linker. The cross-linked produced is processed, followed by analysis using liquid mass spectrometry (LC-MS).
The LC-MS examination is conducted by employing the reverse-phase nano-ultra performance LC which is coupled to a Thermo Scientific (Velos-FTICR-MS). Step 1 samples are usually examined with data-dependent acquisition mode (Schweppe et al., 1524). On the other hand, the crosslinked products are investigated by running the MS in the ReACT mode. ReACT mode maximises the efficiency of the examination by concentrating on the ions of cross-linked peptides which fulfill the anticipated protein reporter interaction relationships; this followed by the succeeding fragmentation of every ion of the cross-linked peptides.
Cross-linked protein identification
The identification of cross-linked proteins is achieved by constructing cross-linked protein-enhanced stage 1 database. This is followed by searching the ReACT acquisition data alongside the stage 1 database employing SEQUEST with adjustable BDP modification stump mass of 197.032422Da on the lysine. The produced lysine matches are then filtered at the static 5% FDR. The reserved peptide matches are joined to their distinctive ReACT-distinguishable protein interaction reporter associations and the peptide matches satisfying particular criteria are retained.
Protein structure determination and docking
The monomeric structures of proteins are determined with I-TASSER employing the use of the user-definite distance constraints which equal to or less than 35. Further, the resulting models are accessed to determine short linear distance normally between the intra-cross-linked lysine Cx atoms. The top monomeric models’ docking is conducted by using the SymmDock to create the homodimeric models. The top docked models are clarified according to the nonlinear distances which are calculated using the Xwalk algorithm. These top docked models are later placed on the Model Archive location and appropriate accession numbers are provided for the predicted proteins that are deposited in the Model Archive structural database (Tenover, 421).
Navare, Chavez, Manoli, Bruce, and others used this approach to investigate the PIN (protein interaction network) of Pseudomonas aeruginosa cells by employing the use of chemical cross-linking together with mass spectrometry (Navare et al., 763). They discovered that all representative proteins cross-linked. Moreover, their corresponding percentages existing in stage 1 database were directly related to their constitutive figures in the whole PAO1 database. Hence, this indicates that under particular cross-linking reaction situations, the cross-linker (BDP-NHP cross-linker) is not inclined towards the proteins of the certain cellular compartment. Further, the identification of the cross-linked peptide sets of cytoplasmic proteins shows that protein interaction reporter penetrated into the Pseudomonas aeruginosa cells. The finding was in agreement with the previous results of the mass spectrometry, confocal microscopy of protein interaction reporter marked human cells, and also the immune-nano-gold electron microscopy imaging of additional protein interaction cross-linked cells of bacteria (Navare et al., 763). After the implementation of data processing stages, the researchers found six hundred and ninety distinctive cross-linked peptides making six hundred and twenty-six peptide pairs originating from two hundred and eleven proteins. 70% of these peptides were peptides of a lone protein consisting of non-overlapping and non-identical structures (Navare et al., 763). 8% of the predicted peptides matched the links among peptides of a lone protein possessing identical and overlapping consequences. The remaining 22% represented the cross-linked peptide sets which were among different proteins. Various techniques are available which help in the determination of the cross-linked and non-cleavable peptides (Navare et al., 763). There is the possibility of these approaches to show the improved determination of homodimer peptide sets because these two peptides share the same fragments in multifarious MS2 spectra. However, is not true for cleavable crosslinkers, ReACT and other analogous techniques. ReACT calculates produced cross-linked peptide masses as they produced cross-linked peptide are removed from the cross-linkers. ReACT further fragments peptides distinctively for peptide determination. This is an agreement with the heterodimeric peptide and homodimeric sets. This permits the determination of all kinds of cross-linked peptide sets without unfairness towards the homodimeric links.
The protein interactions reporter XL-MS classifies binary protein interactions since two linked peptides are employed to determine connectivity (Tenover, 419). Therefore, presently it is not direct to determine whether the crosslinking between various proteins show the existence of various binary interactions. Though, for the intra-protein-links among peptides with either non-overlapping or non-identical structure require more efforts to ascertain whether the two links are from one protein monomer or two monomers of a particular protein. Contrary to the aforementioned, when the cross-linked peptides have a common or overlapping amino acid structures, and also if the sequence of the peptide exists once within a certain protein; this shows the presence of a multimeric complex between not less than two identical monomers of a particular protein. Navare et al. used unambiguous homodimers, intra-cross-links and the inter-protein linked sites to develop monomeric protein models (764). They used developed monomeric protein models to filter both heterodimeric and homodimeric protein models (Navare et al., 764).
Other scientists have also used XL-MS technique to investigate the host-microbe protein interactions at the time of bacterial infection (Schweppe et al., 1521). They observed that correlations between of two peptides originating from similar protein dominated their identifications. For instance, two residues existing within certain protein were most probable to have a close physical contact with the cells (Schweppe et al., 1522). Hence, these allow more cross-linking thus making a higher percentage of associations in cross-linked datasets (Schweppe et al., 1522). Significantly, intraprotein associations characterize the close residues within a particular protein; hence, offer crucial structural coordinates for distinguished proteins, even if there is a paucity of known structures of that specific protein. On the other hand, Scheweppe et al. derived interprotein protein-protein interactions from cross-linked associations between peptides from two distinctive proteins (1522). They utilized intraprotein protein-protein interactions to create interaction networks and produced structural data relating to the interaction crossing points of protein complexes (Scheweppe et al., 1522). Within the intraprotein associations, they found 3.7% of peptide-peptide cross-links, while 6.4% of all protein-protein interactions were interspecies interactions (Scheweppe et al., 1522). Surprisingly, they did not observe interspecies or protein interactions in uninfected cells (cross-linked H292 cells). According to cross-linked correlations, Scheweppe and team developed the protein interaction network, which composed of host-pathogen, host-host and pathogen-pathogen protein-protein interactions (Scheweppe et al., 1523). Further, to confirm whether protein interaction reporter could determine physiological protein complexes and associations, they measured a number of human protein-protein interactions which were early interpreted in the protein-protein interactions’ database, IntAct, and which were intraprotein interactions (Scheweppe et al., 1523). Out of five hundred and sixty-six of distinctive human protein-protein interactions, 70% were as a result of observed in protein-protein interactions or intraprotein interactions. Furthermore, sub-networks of protein-protein interactions originated from the identified complexes encompassed ATP synthase, integrins, host cytoskeleton, cohesin, histones and heterogeneous ribonucleoproteins(Scheweppe et al., 1523). Therefore, the study of protein interaction network of organisms’ cells is very important.
Interspecies protein associations are fundamental determinants in pathogenesis and bacterial infections (Scheweppe et al., 1525). Enhanced knowledge on which interspecies protein-protein interactions occur and how the proteins associate can significantly boost people understanding of molecular mechanisms that take place in host invasion and can help in providing new opportunities for antibacterial treatments. As the threat of antibiotic-resistant pathogens expands, new approaches which may help in treating multidrug-resistant bacterial infections are necessary.
The identification of protein-protein interactions in cells of infectious bacteria cannot be fully understood without discussing how iron interacts between the bacteria and the host organisms (animals encompassing human). To circumvent the toxicity of iron within the human body, transferrin which is the human iron transporter protein keeps the concentration of free ferric ion at 10-24 M (Raymond et al., 3584). Astonishingly, pathogenic bacteria such as Pseudomonas aeruginosa should compete in conjunction with this thermodynamic limit to acquire iron from tissues or human serum (Raymond et al., 3584). It is impossible to underrate the importance of iron as the limiting nutrient in the development of bacteria. Excess iron upsurges organism’s virulence. For example, Yersinia enterocolitica virulence is improved ten million-fold by the peritoneal administering of ferric desferrioxamine(Raymond et al., 3584). The same effect is observed when desferrioxamine is supplied at the time of Klebsiella and Salmonella infections (Raymond et al., 3584). Moreover, there is a direct relationship between the LD50 of Vibrio vulnificans and iron attainability. Bacteria have developed antagonistic iron acquisition processes. Bacterial secreted selective and powerful siderophores in response to a deficiency of iron. In many bacteria species, there is an iron regulator (Fur-like protein). It controls the uptake of iron in many bacteria species. Certain Gram-positive bacteria use diphtheria toxin regulator (DtxR) protein (Abergel et al., 3; Clark, 15).
Most bacteria have the ability to produce more than one siderophores. The siderophores are grouped based on the chemical functional groups they apply in chelating iron. They include the catecholate-type, hydroxamate-type, and the third class of siderophores which uses n-hydroxy amino side chains containing an oxygen atom in chelating iron (Clark, 13). The catecholate-type employs end-to-end hydroxyl groups of catechol rings to bind Fe3+. Enterobactin is a typical example of catecholate type siderophore(Clark, 13). Enterobactin has the highest affinity for ferric ions (Fe3+). It has been observed in some nitrogen-fixing bacteria (Clark, 13). Hydroxamate-type siderophore uses nitrogen atoms of oxazoline and thiazoline to chelate ferric ion. Ferrichrome is an archetypal example of hydroxamate-type siderophore. The third class of siderophore, on top of using N-hydroxy amino side chain together with an oxygen atom, it combines all the three siderophore types and it utilizes three distinctive methods of the binding ferric ion.
The gram-positive bacteria contain outer membrane receptor proteins which exist in their outer membranes (Clark, 15). The outer membrane receptor proteins recognize, bind and then transport the complex (iron-siderophore complex) into the cell periplasm. The outer membrane transport proteins are divided into two, active and passive transporters. Active protein transporters have high binding affinity for ferric ion and ATP provides them with energy (Neilands, 26723). Contrary to active transporters, passive transporters have a low binding affinity for ferric ion and existing chemical gradient supplies them with the energy. The outer membrane receptor proteins consist of two unique domains, the -barrel, and plug regions. -Barrel of every protein that inserts across the outer membrane lipid bilayer is generated by twenty-two antiparallel -strands which are linked to the periplasm by external loops and short turns which extend above the surface of the cell. On the other hand, -barrel is completely enclosed by the plug region (Clark, 16). The outer protein membrane receptor protein identifies and binds the substrate (ferric ion), followed by undergoing main conformational alterations (Abergel et al., 2). The -helix which exists at the plug N-terminal is relaxed and it spreads to the flexible conformation (Neilands, 26725). Furthermore, the N-proximal end does not fit into the fixed structure. Though, it is thought to enable interaction between TonB complex and outer membrane protein receptor in order to provide the energy necessary for the active transport. Once the iron-siderophore complex is actively conveyed into the periplasm, an iron-siderophore complex is tightly bound to the periplasmic binding protein (Clark, 16).
Knowing how the siderophore bacteria use to bind to ferric ion (Fe3+) to aid in the deep understanding of the protein-protein interaction of bacteria in the human body. For instance, enterobactin is known to strongly bind to human serum. Therefore, to successfully investigate the protein-protein interactions of bacteria, in this case, Pseudomonas aeruginosa, one need to inactivate enterobactin in human serum since most of the bacteria utilize it to bind to ferric ion (Raymond et al., 3587) and this may interfere with the experimental results.
Abergel, J., Rebecca, Warner, A., Jeffrey, Shuh, K., David, Raymond, K., Kenneth. “Enterobactin protonation and iron release: structural characterization of the salicylate coordination shift in ferric enterobactin.” J Am Chem Soc., 128.27(2006): 1-31.
Clark, Lee, Brianne. Characterization of a catechol-type siderophore and the detection of a possible outer membrane receptor protein from Rhizobium leguminosarum strain IARI 312. Dissertation, East Tennessee State University, 2004.
Navare, T., Ariti, Manoli, Colin, Chavez, D., Juan, Zheng, Chunxing, Weisbrod, R., Chad, Eng, K., Jimmy, Siehnel, Richard, Singh, K., Pradeep & Bruce E., James. “Probing the protein interaction network of Pseudomonas aeruginosa cells by chemical cross-linking mass spectrometry.” Elsevier, 23(2015): 762-773.
Neilands, J., B. “Siderophores: structure and function of microbial iron transport compounds.” The Journal of Biological Chemistry, 270.45(1995): 26723-26726.
Raymond, N., Kenneth, Dertz, A., Emily, & Kim, S., Sanggoo. “Enterobactin: an archetype for microbial iron transport.” PNAS, 100.7(2003): 3584-3588.
Scheweppe, K., David, Harding, Christopher, Chavez, D., Juan, Wu, Xia, Ramage, Elizabeth, Singh, K., Pradeep, Manoli, Colin & Bruce E., James. Host-microbe protein interactions during bacterial infection.” Chemistry and Biology, 22(2015): 1521-1530.
Tenover, C., Fred. “Rapid detection and identification of bacterial pathogens using novel molecular technologies: infection control and beyond.” Medical Microbiology, 2006: 418-423.